Quantum and Computational Chemistry Discussion

  Interpret at your peril.

  The not-so delicate balance between truncation error and roundoff error.

  Motivated by various conversations, e.g., https://sourcegraph.com/blog/the-death-of-the-junior-developer. Will add more later when I can find the links/motivation... An amusing analogy appears in startups. A Corin Wagen piece  linked to some common founder mistakes , and one of the many is choosing an obscure niche to avoid competition. So competition is good!? Ah, I am not a free market fundamentalist, but that is another topic. An interesting question is how this applies to academia: Derek Lowe commented  that interdisciplinary science is where many breakthroughs are had (and many will corroborate that) but there is also a fear that such pioneers, say "chemical biologist" grad students, will fail to achieve expertise in either field. But where is the argument? In a time when funding is especially tough, is there a pattern to reduced funding (aside the obviously targeted fields)? For instance, will interdisciplinary groups (say, enzymatic catalysis + ML) be bashed...

  Many professors in chemistry departments, especially theoretical faculty, have a PhD in physics. Some tend to like orders of magnitude, and also strange units, e.g., the  erg . Will elaborate later. 

  Ah, diffuse basis sets are important for the energies of anionic systems, but they may not even be that helpful for the geometries and cause a host of other problems. Frank Neese complains about this every now and then, and Tian Lu (aka Sobereva) also has a strong opinion .  In short, don't put diffuse functions on hydrogen (ahem, 6-311++G**) and don't use diffuse functions for cations. Oh, and if you're going to benchmark a bunch of functionals, don't add diffuse functions instead   of polarization; I will elaborate the roast later. 

See for yourselves! Gaussian: https://gaussian.com/man/ ORCA: https://www.faccts.de/docs/orca/6.1/manual/

  There is an amusing pattern in many forms of learning: " U-shaped development ." You learn a little something and are good (the real Newbie's Luck?), you learn more and your skills decline, then finally you get better again. Perhaps here is similar: you learn to calculate (manually) the standard enthalpy of formation (\Delta H_f), then  in quantum/computational chemistry you forget, then you need to calculate/report it for a collaborative work with synthetic chemists and learn it better... 

The taboo around asking coworkers' salaries has somewhat assuaged, but these topics remain rather private. The academic analogy in my view, especially for public  universities, is not salaries, but start-up funds , the money a new professor gets to "start-up" their lab. Some fields are inherently more expensive than others, at least in procuring initial equipment (sometimes, the grad students cost more... ); when one receives that wonted offer (or two??) for a tenure-track position, an important concern for the 5-7 year deadline. More grad students = more papers, but don't want to run dry before the next grant comes in. Ah, just accept students without guaranteed funding. Worst case?  They get some valuable teaching experience; industry loves that, right? If anyone has more numbers, please do send... 

  "Fortuitous error cancellation" is the premise of DFT. Albeit,  not only do two wrongs make a right; three can, and so forth. In general, some "lower" levels of theory perform better for this; in Sherill's words , these may represent so-called " Pauling Points " due to the aforementioned coincidence. Alas, even given the exact same level of theory, results may vary across computational software (e.g., Gaussian versus ORCA) for reasons that are nigh impossible to discern. Sometimes, performing the exact same calculation in the exact same program will lead to different results; JCIM mandated triplicate MD simulations for this reason, albeit they are indeed a journal focused on methods development (where sub 1 kcal/mol "chemical accuracy" doesn't quite cut it for many). So when espousing the wonders of DFT, as someone  like  Kieron Burke does, one ought to be careful, lest feed the " computational nihilism "  that besets our fie...

In reference to a famous saying by  Kṣitigarbha : "If I do not descend into hell, who will?" Indeed, us academics are unified by our spirit of self-sacrifice.  

  In ONIOM, "M" stands for "Molecular." We often represent the  number of atoms or system size with "N," so... same difference?

[ 1 ]