Is there a fair future for computational theoretical chemistry?

Note_20130830_194536_03-1Computational theoretical chemistry is amazing, but it is a career dead-end. Today, hordes of grad students are in the field doing technical work with little scientific innovation. They will earn a doctor title and then move to a completely different field. The system needs them to keep up to the high-production demands, but is it fair? Maybe the future of CompChem is in outsourcing.

I am a professional in computational theoretical chemistry (although my background is in physics). Many people have never heard of this field, which consists of investigating chemical processes through computational simulations; and developing methods and computer programs to do such simulations.

Maybe the field will become a bit more sexy now that the Nobel Prize in Chemistry 2013 was awarded to three scientists in it. But for the chemistry community, computational theoretical chemistry, with its branches into fields as far apart as molecular biology and material sciences, has been part of the scientific routine for decades.

(Just for curiosity, a couple of illustrious names who once contributed to the field are Peter ‘Boson’ Higgs and – serious – the German chancellor Angela Merkel.)

There is an elephant in the room, anybody wants to talk about it

The problem with computational theoretical chemistry is that it is a career dead-end. After earning a doctorate in the field, the young researcher will find out that the job market is saturated. If he is clever enough, he will quickly move to a completely different area (like Merkel did), otherwise he risks haunting chemistry departments for years, jumping between precarious temporary contracts.

The reason the job market is so bad is a basic population ecology problem: too many people for too little resources. Any research group to survive must recruit hordes of graduate students to produce loads of scientific papers. This is just normal in hard sciences, and it is not generally a problem for most of chemistry fields because industry will absorb those young professionals. The particular problem with theoretical computational chemistry is that positions out of the academy are rare, creating a great surplus of people with a useless doctorate title.

From a cold population analysis, every established professor should be educating in average not more than the number of professionals that the job market will be able to absorb. In a field like catalysis, where professionals are largely required by industry, this may allow a professor to award few doctorates a year. In a field like computational theoretical chemistry, however, this may allow to award only a few doctorate titles during the whole professor’s career.

Right now, the situation edges the ridiculous: professors in the field often have half-dozen simultaneous students. I have colleagues who, even without tenure, have already few doctorate students. (And in a couple of years they will be competing with their pupils for a position!)

There is nothing that those senior researchers can do, as they need the students to keep the projects running, but I cannot avoid asking: Is it fair to let students specialize for years in a field that they will most probably have to completely abandon? Is it the better use of scholarship resources investing them in people who will not act in the field?

My two cents to move the elephant out

Research on computational theoretical chemistry should be deeply reformulated.

First of all, the number of graduate students in theoretical computational chemistry needs to be strongly reduced. To compensate the shortage of people, most activities in computational theoretical chemistry should be outsourced to technical departments and companies.

Much of the work in the field are technical and routine activities. If the research group needs simulations of the thermochemistry or a benchmark of vertical excitations for a new compound, this could be perfectly done by a technical staff. This data should be requested to a technical department in the same way we request NMR measurements.

If the group needs maintenance of their computer cluster, they should call the local IT department or have budget prediction to hire a company to do the service.

If the group needs to compute a property that standard commercial softwares can still not provide, their budget should allow to call their favorite CompChem company and hire them to implement it. In fact, if the group is developing a robust software in the field, it should be stimulated to spin-off from academy, as Gaussian or Turbomole successfully did.

Right now, armies of graduate students are buried into doing DFT, MD, CC, CI, MP2 simulations (make up a random acronym, probably it is already in use), writing codes, administrating computer systems. They think they are doing science. No, they are doing technical well-stablished routine work with little scientific innovation. The science happens afterwards, when those data flowing out of the computers clusters are taken, analysed and used to model reactions, discover new processes and understand nature.

Outsourcing is the key for a fair future for computational theoretical chemistry, where professionals have real working contracts and career perspectives; where studentship fundings are not wasted to educate people who will ending up working on a completely disconnected field.


Categories: Computational Theoretical Chemistry, Science Policy, Scientific Culture

Tags: , , , ,

37 replies

  1. My impression is that big industrial companies are acutally starting to hire compuational chemists. The idea is that it should save them a lot of money if they are able to go for rational design processes, rather than pure trial and error syntheses. For example here in Heidelberg several colleagues are actually taking off to big companies or at least get funding from them, for organic electronics.
    But I agree that computational chemistry needs to become a real world science. There should be more focus on producing usable results, rather than all this formal talk about methodological details (like multi-reference coupled cluster). And ultimatly a lot of the work should be done by technical staff just as this is also the case for synthesis. But on the other hand, one should remember that it is a field which is maybe 10-15 years old (before that there was just not the necessary computational power available). So there is certainly still a lot of room for improvement

  2. I also want to throw my two cents opinion.
    In general, I agree on the analysis of the current job market. I am one of those precarious people with a temporary contract and I can tell you how frustrating it is to live at the edge, with a very indefinite future. Even getting a project funded becomes really complicated when the market is so saturated and nested with more and more frequent Matthew effect episodes.
    On the other hand, I disagree on some of the solutions you provide. The idea of outsourcing is definitely the way to go, however I do not think that professors should limit the number of students to educate. As in other fields of chemistry, the industry should absorb the working force, however in a way that should stimulate the creation of spin-off companies dedicated to the solution of problems that the industry rises. From what I understand from your statements, you either train computational chemists to solve problems that also technicians can do (simple vertical excitation spectra), or you have novel methodological problems that need to be implemented (and will be provided by spin-off ChemComp company).
    I actually think this view there is lots of room in the middle. There are many problems (for example, the study of the interface between nanomaterials and the environment) that require refined skills in computational modelling and at the same time a deep knowledge of the chemistry involved. In these cases, it is very likely that new methodologies need NOT to be implemented, however basic computational skills are also not sufficient to solve these problems. Actually, you need a well prepared computational chemist, capable to work side-by-side with experimentalists, capable to also understand how the industry works, and finally capable of providing (even not definitive) results in a fast manner (otherwise they are useless) that are of importance for the industry.
    As I see it, graduate students should be trained as they are now in academy, where they will obtain a PhD degree either with (a) an application computational chemistry profile or (b) as method developers.
    The former could find job in a (spin-off) company that, for example, does computational chemistry calculations for several industries. The latter can either work in a Quantum Chemistry Software Company like Gaussian, Turbomole, ADF, or also for a (spin-off) company that can be dedicated to both application and method development. This last point is very important, because many new developments are needed to solve problems that the simple applications cannot solve, but that cannot be predicted if you only focus on developing new methodologies.
    In other words, I see that the future of computational chemistry lies in the creation of these “intermediate” companies that provide solutions to industry with excellent and well-prepared personnel.
    In this way, the stress of finding work in academia will be lifted. Fundings will probably raise in this field, new professors can be hired in academia and private companies, and graduate and postdoc students will not live anymore with the terror of not knowing what the heck their future will be.

  3. Just quick comment about “After earning a doctorate in the field, the young researcher will find out that the job market is saturated. If he is clever enough, he will quickly move to a completely different area (like Merkel did), otherwise he risks haunting chemistry departments for years, jumping between precarious temporary contracts.”

    While Merkel is clever, her career change choice might have been also ‘helped’ by the US denying her a visa to do her postdoc and the bad climate for women (specially theoreticians) at that time (and even now)…

  4. I’m currently a graduate student looking to switch group (possibly from analytical chemistry to computational)

    This post really helped put things into perspective!

    Thank you!

  5. I can only make the following simple comment: I could not agree more with this posting. I would even go further in saying that the system is producing far too many doctors without any perspectives. I will be even more extreem in doubting whether the doctorate has not lost its value in general. But that might be also somewhat extreem…

    • Doctorate is still very relevant. It is during this time that the professional acquires scientific independence to conduct the research. For this reason, companies looking for senior researchers will allways give preference to those with doctorate.

  6. I think this article is very very true. I was active in this field since early nineties and have gone through a number of temporary contracts that forced me to drag my family through various places around the world.
    Dear readers, please be warned that commercial spin-off can also prove to be a career trap, as it happened in my case: the company, after investing many-million funding into combined, experimental and theoretical study, filed for bankruptcy before concluding the research, the management drove off with their brand-new Porsches and I, among other former employees, was left with the feeling of disgust and yet another gap in my CV. I then tried REALLY hard to change my career track. My goal is a career in IT, but it is not as easy as some people might think. Contrary to computational chemistry, job interviews tend to be hard, and you are always confronted with the questions like ‘are you going to go back to the academia?’ or ‘all the time you did this computation stuff, why this sudden career change?’
    I’m not writing this to discourage people to try their luck in spin-offs or start-ups, but just want to stress that this path has been walked before, and it is precarious one, too. Computational chemists, especially after turning 40, need stability (they are normal people after all!), and my opinion is that choosing to change one’s career path at a more ‘advanced’ age will most probably be final and decisive for the rest of the professional life. It might be better to look for opportunities where there are more jobs overall, but more applicants. Today’s science is governed by economy, global financial factors are influencing country’s state budgets. And pure research is financed from taxpayer money. On the other hand, investors, whether business angels, venture capital, or banks, have skilled analysts who will no doubt determine if there is market for computational chemistry services, before deciding on the funding (and this is this very funding which is going to pay your rent or electricity bills or your kids’ school). I’ve faced investors before and believe me, it is not an easy task to convince them your research is going to bring revenue, and they don’t care about the Schrödinger equation!
    I think one has to stay realistic, but some optimism will not hurt. Fingers crossed for all the hard working computational chemists.

  7. Hi Mario, great article! Hope you are still at the Fem-Ex and I can catch you tomorrow morning to discuss more. Just saw you yesterday watching football.

  8. How I wish I’d read this article six years ago, before writing three theses (B. Sc., M. Sc., Ph. D., where I come from) in Computational Chemistry. The future really does look bleak for our kind.

    I have literally lost count of the number of cover letters and CV’s I have sent out. First I tried to look for placement in my own field (Computational Materials Chemistry), regardless of whether a position was being advertised or not. I wasn’t picky: Europe, USA, Canada, world class institutions like MIT and Oxford and obscure little Universities in towns I’d never even heard of, as long as they did something vaguely similar to what I’d been working on, they were all fair game. Many never even bothered to reply, some were kind enough to let me know that they had no vacancies, one even shortlisted me for an interview, but competition was stiff and I didn’t pass. After a while I became discouraged, and as my Ph.D. approached its conclusion and the prospect of unemployment drew nearer, I started sending out applications to any company that happened to be looking for a chemist: I tried many different fields, including but not limited to oil, renewable energies, paper, paint, cosmetics, food. No one showed the slightest bit of interest: understandably, they were only looking for people with lab-experience, preferably in their specific sector.

    Eventually, after reading through the 1000th-or-so job posting list, I took the hint and realized that pharmaceutical companies are practically the only ones outside of Academia that hire people with a background similar to my own (more or less). So when one slow Sunday afternoon I saw a position being advertised for MD modelling of proteins at a respectable University I applied right away: I was interviewed less than two weeks later and I was able to land a three year contract.

    Which is nice, all things considered, but I still have the distinct feeling that I have only postponed the problem: by making myself marketable to Big Pharma, my chances at finding a job have increased slightly, but what if they don’t take me? I guess what really scares me is that in this career there doesn’t seem to be room for any Plan Bs: if worse comes to worst, experimental chemists can always swallow their pride and recycle themselves as lab technicians, scrubbing beakers and running tests on the local product to make sure it meets the quality standards. Computational chemists don’t have that option: unless you manage to become a professor or a researcher at a big company, you (and your spouse and children) are doomed to a nomadic life of one-to-three year post doc contracts at different cities, countries, even continents, unable to make any sort of long term planning and with the fear of unemployment constantly looming over your life. Like Gregg, I have also considered the IT path, but so far I only know Python, and then, I’ve never coded anything longer than a few hundred lines. I was thinking of taking programming classes and maybe pick up another language, but after reading Gregg’s post, I’m not so sure it would be worth the effort any more.

    I wish I’d not had to learn all this the hard way. Science is my vocation in life and even if I could go back, I would still choose to be a scientist and take a Ph.D. But if someone had warned me, I would have taken a different route and opted for something perhaps not as intellectually titillating as DFT or CASSCF, but with more career options and a better chance to provide some stability for my family. I never expected to become rich working in Science, but this feels downright unfair.

    I also wish someone had told me this: “They think they are doing science. No, they are doing technical well-stablished routine work with little scientific innovation”. That pretty much sums up the entirety of my Ph.D. work. And now that I have finished writing this comment, I feel even more robbed.

  9. Reblogged this on The Peers of Beinan series on Word Press and commented:
    Whereas I am still only an amateur scientist — a factor of that accident that also took my eyesight — I found this discussion quite interesting and applicable to many segments of academia.

  10. When googling “computational chemistry jobs” this post appears on the first page. Now, I am not contrary to cautioning the expectations of students entering the field of computational chemistry. However, a bit more faith in the field’s demand and applicability outside academia would be nice. I haven’t even graduated in the field yet and already have a good industry offer for a computational chemistry position. Reading this post and its pessimistic tone nearly discouraged me a while ago to pursue a career in the field I love. I sure hope this doesn’t happen to others like me. Surely, that is not the intention of the contributor.

  11. I am also a computational chemist, and I also feel the same as most of you mentioned here. Most of the computational Chemistry jobs are in US, few are in Europe and some other countries. Middle East and Gulf are lacking computational chemists, nevertheless, they are not even interested to hire these people in most of the cases. I feel that a computational chemist must also be trained in experimental techniques. This way they might get a better opportunity in finding a good job.

  12. I am going to select my PhD research group. I (Was) thinking about a computational group of my university to join. As like as Bob, when I did google about computational chemistry, this page came at first. Now I am thinking, I shouldn’t think about computational chemistry anymore. Boring organic synthesis lab work is far more better than computational chemistry.

  13. With all of the direness that has been posted about Phd’s being a waste of time I think it should really force those of us still in the pipeline to make sure we have our backs covered. I am finishing a double major in chem-physics now and planning to go onto postgrad in theoretical chem next year because I love it but I will not go in blind. I think its important to develop as many cross disciplinary hard skills as possible, things like being about to program in multiple languages (python, c++, java, R), high level statistical analysis, solid mathematical problem solving skills. These 3 sets of skills form a toolbox that will allow entry into all kinds of other fields if needed (all of which are higher paying that theo chem haha).

    Data science is crying out for people, there are freelancers on elance doing 20k USD consulting contracts based on computational and mathematics backgrounds. So while we all want tenure and the ability to run our own lab, win a nobel and benefit humanity in the future it is always nice to know that between temporary contracts perhaps you can sneak in some freelance work using hard won skills during the grind of Phd work. I also think with next generation discovery big Pharma is going to be going more and more towards computational methods. Look at bioinformatics and how it has blown up in the past few years. I would say most graduating with just a masters are nothing but glorified techies there is going to be a major need for principals to run the show down the line. Perhaps I am naive and will still be a bartender 10 years from now? =)

  14. absolutely agree with Donnie.

  15. is their any who can answer the question….?how we can solve the problem in computational chemistry..

  16. I have the feeling that the demand for computational chemists in industry is growing. But admittedly, it is still a very tiny niche.
    But coming to your post, I think the true problem either goes deeper or is rooted somewhere else.

    In my opinion, universities are not intended to properly qualify students or PhDs for industry jobs. At least, they do not see this as their core task. Rather, they want to prepare people for “pure science”, for the acacemic carreer – which, at least here in Germany, is more risky, worse paid and offers even less perspective than trying to switch into industry as you only get (often times 50%) limited contracts and then, after PhD and most likely habilitation or junior professorship (without tenure, of course!) – either a professorship (to be realistic: haha) or you get kicked out because of “Wissenschaftszeitvetragsgesetz”, which states you are allowed to work only 6 years on limited contracts and there are basically no unlimited positions apart from professorships. Sounds great, doesn’t it?

    If you complain that computational chemistry does not properly prepare its members for a job in industry: what do you think of social science, theology, history, arts or archaeology and such? In these fields, it is sometimes even hard to find a somehow funded PhD position, whereas in chemistry, you quite regularly get 50% positions (at least in Germany). So, if you say “university should only be allowed to train as many people as there are positions in industry” and maybe “plus open academic positions”, the consequence would basically be to completely and radically shut down most of the departments apart from engineering and (natural) science. But I do not want to live in a world without music, philosophy and such!

    And besides, I think it is part of globalisation etc. that you have to be prepared to do “something else”. Even if you find and get “your favourite industry position”, you most likely will not be doing this until you retire. Indeed, I think it is part of your personal responsibility to shape and refine your profile to something that can make your living. This may sound a little harsh and most likely, you will need some luck and good intuition on when to do what but you can learn a lot of things – isn’t this partially what a PhD in science is about?

    Besides all this, there is one major thing I agree: I think that people should be disillusioned profoundly and early. It is not good if people come to realize “o sh**, I cannot find any job with my training” after 8 years of study, a PhD, several postdocs and some time of vain job search, but never gave it a thought before. It should be part of your choice of study, PhD and so on that you actively and knowingly choose the track in a field in Germany called “brotlos” (meaning, not providing your living). In this sense, I also think professors should tell their students honestly about these issues and not promise them unrealistic things – but alas, how should they know? They studied, did a PhD and got a professorship, so for them, “it worked” – I guess, most of them did not even do a single internship outside of academia.
    What might be a good idea would be to include coures on economics, how to build your profile etc. as obligatory part of your PhD, so you are forced to think about these issues. But would it work? Who would pay it? Who would even start and plan it? Kind of a hen-egg-problem.

    So, my personal plan is: finish my PhD and get out of the ivory tower of academia (after studies of chemistry and physics, master thesis in CompChem and PhD in protein simulation). Hopefully, it works with “my favourite industry job” (process still running after interview), but even if not, I think I should be able to find a job in either software delevopment or – if needed, I honestly do not want it – consulting. Maybe not where I want it, maybe not how I want it, maybe I need to to some other training – but I don’t think I will end up washing dishes or driving a taxi.
    Good luck to you all!

    • Hi Matthias

      I am in the second year of my PhD and the project involves protein simulations. It’s been five months since your post,I am wondering if you could provide update if you found a job or not? Also, can you tell a little about what skills did you manage to learn when running protein simulations?

      I am trying to enhance my skills in Python and learn at least one other tool like MATLAB or R to be able to qualify for data scientist jobs following my Phd (hopefully within next three years). But my project doesn’t require any of these languages as I run simulations with programs which were configured some 20-25 years ago and are used as they are with least modifications. If you managed to learn any other skills (which were helpful during the job application procedure), and you don’t mind sharing them, they might guide me in a direction I am not able to see so far.

  17. It’s a sad age for science. I knew that even before I signed up for BSc Chemistry. I also met a lot of people, mostly postgraduate students, who despise academia, who call it “ivory tower”, who hate it for lacking job prospects and call for everyone to join industry.

    Natural science is a career dead end career, but I just love it, more than anything, more than survival. So I made a cyanide capsule to “motivate” myself: If I cannot do science for living, what’s the point of living at all ? The only two things I enjoy the most are science and classical music, since they don’t make classical music anymore, there is nothing to live for if I cannot afford doing science.

    I’m going to finish my PhD in theoretical chemistry, and I wont stop there. I will continue learning condensed matter, particle physics, theoretical physics, cosmology… I will not leave academia. The best thing I can hope for is professorship or a place in research institutes.

    I still remember a friend of mine, my classmate in secondary school, who share my enthusiasm. We use to skip P.E. classes to talk about the light cone in the special relativity and the Schrödinger equation for hours, even we know little about it at that time. He chose cosmology as his major, which has a even grimmer future. He is doing something like Lambda-CDM model I think, I really wish him well.

    To me, learning is not a mean to accomplish something, to find some jobs, its the end. Being an line engineer, doing consulting or driving a bus.. is not an option.

    Its either science, or science in the after life.

    Wish you all can find a place where you get paid while doing what you like.

    • I admire your passion for science, I also, like science very dearly. What people do not realize is that artificial intelligence and automation will eliminate major parts of the global economy, but it will also create a new section of the job market. Computational chemistry will be one of the areas of wonder that artificial intelligence will not eliminate, instead it will help human computational chemists run more accurate simulations of molecules, of supra-molecular systems, and ultimately of entire cells and organisms. In fact, all of scientific experimentation will be replaced by computational simulations. This will help computational chemists discover new drug compounds faster and with greater certainty, it will allow us to replace animal testing and clinical trials with computer simulations, and it will give us great depth of knowledge into the inner workings of the cell and the workhorse of the cell, proteins. I predict that by the year 2030, computational chemists would help cure almost any illness possible, find all organic compounds that could ever exist, and rapidly describe the function of all molecules in the animal kingdom. Time will only bring more power to the computational chemist and the future will be far more advanced than any science fiction film or novel ever made.

  18. I’m the guy who commented on July 9th 2014. I just wanted to share an update on my situation. After my postdoc, I found a job as data analyst by sheer luck through a friend who knew the CEO of a medium sized company, otherwise I would still be unemployed: out of approximately 300 job applications (and about as many tailor made cover letters), only four employers ever bothered contacting me for an interview.

    My advice to students: leave computational chemistry alone, get the hell out of academia as fast as you can, and get some practical, real world experience. Look for internships, traineeships, anything that you can put in your CV to show that you have e-x-p-e-r-i-e-n-c-e. Let this word be your guiding light: employers don’t care that you graduated cum laude, nor are they going to be impressed by your publications or h-index, what they care about is whether or not you have some experience that is relevant to the opening they are looking to fill. I know it’s ugly, I understand you want to do something you feel is noble and intellectually stimulating, but somewhere down the road you will find that in this world, ideals alone aren’t going to put a roof over your head.

    Oh and one more thing: take one good look at the people who are above you in the academic hierarchy, because that is what your future will look like if you are successful in academia. Look at the people who made it, and look at what they do most of the time: what portion of their day is dedicated to doing actual science, compared to the time they spend between fund raising, paperwork, and lessons? Answer that question, and draw your own conclusions.

    Good luck to all.

    “Now the hardness of this world
    Slowly grinds your dreams away
    Making a fool’s joke
    Out of the promises we make

    And what once seemed black and white
    Turns to so many shades of gray
    We lose ourselves in work to do
    Work to do and bills to pay”

  19. Well said. Good guidelines for dean’s of science. But for those poor fellows who did not have a chance to read this at the beginning of his/her graduate career what is the strategy ?

  20. Dear Mario,
    It is a quite interesting blog. I am a Ph.D. student in Computational Chemistry.
    I saw two main points of you but I would like to mention that in my country Universities budget is always small. If they have chemicals, they don’t have access to characterization instruments.
    In my country, Computational Chemists are always successful because they need a one-time license, Computer and generator/UPS. Paper (Research article with high impact) is the most important thing here.
    I am planning to move to Germany, Canada or USA for sandwich program. you can also tell me shortly, is it a good move or not? We don’t have access to clusters, we use normal Cori 7 PCs with 8 processors and a few with 24 processors.

  21. I do not agree and think your remarks are both short sighted and incorrect. Do you think that computational theoretical Physics is also not necessary?
    You mimic the people who wanted to close the patent office around the turn of the century because all of the good ideas had already been discovered.
    Are you currently in academia?

    • The author didn’t mean computational theoretical chemistry should not exist now. I think his point is if there is a fair future for the most people in this area/


  1. Who are the paper’s authors? | Much Bigger Outside
  2. And the First “Much Bigger Outside” Award goes to … | Much Bigger Outside
  3. How much $papers is a scientist worth? | Much Bigger Outside
  4. Let Science Manage Scientists | Much Bigger Outside
  5. Science Across the Silk Road – Much Bigger Outside
  6. PhD: Less Is Better – Much Bigger Outside

A penny for your thoughts...

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: