Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Molecular thermodynamics vs. quantum mechanics

  1. Apr 9, 2003 #1
    Hello people,

    What are the advantages and disadvantages in using molecular thermodynamics instead of quantum mechanics with macro molecules?

    thank you.
  2. jcsd
  3. Apr 9, 2003 #2
    Um...um....how long do you have? People have spent enormous amounts of their professional lives working on this topic. :smile:

    The short version: if the system of interest is "macro" enough, you can usually get away with using classical mechanics-based methods (e.g. molecular mechanics, molecular dynamics) and thermodynamic approaches. This would be like seeing if you can fold an extended polymer into some type of globule and to see if it remains as such. If you're interested in some type of distinct microscopic effect (whether or not a hydrogen bonding interaction crops up and its characteristics, some type of chemical reaction), some type of quantum mechanically based calculation will be needed. The thing is that there are often varying degrees of accuracy and specificity in both areas and that it will depend on the amount of error you can tolerate in your work. (For example, DFT was used for years by physicists without too much problem but it wasn't until the 1980s that chemists began using it since the thermodynamic data they got was often far beyond what was needed for chemical accuracy with the advent of Becke's work in functionals & GGAs as well as others.)

    The ultimate arbiter tends to be money and CPU time (not to mention just how competitive your area happens to be). You can run immensely complicated and accurate simulations, but they'll take a miserably long time and eat up all sorts of CPU capacity, but you'll get beautifully accurate results (ideally). Or you can run reasonably accurate simulations, but get them in a fraction of the time, and sacrifice some accuracy as well.

    Added after lunch: I guess the quick and dirty guide would be as follows...

    QM: Pros - high accuracy, can often treat intramolecular interactions more precisely, can treat reactive processes reasonably well (given certain circumstances). Cons - Can only reasonably treat 100 to 300 atoms for most things, can be very expensive in terms of time & processing power.

    Thermo/Classical: Pros - can treat very large systems w/o too much agony, can model dynamics of large macromolecular assemblies/polymer chains. Cons - will lack detailed atomic/molecular physical details, may not be suitable for certain types of systems (e.g. nanotech applications to nanocrystals).

    There are also hybrid methods, both in terms of molecular dynamics force fields (where you'll often plug in parametrized values obtained either ab initio or semi-empirically and then run a classical simulation) as well as mixed QM/MM methods (e.g. where you'll treat the active site of a protein with QM, have an interface of frozen molecular orbitals, and then the rest is classical in nature).

    A good book that you may be interested in digging up at your nearest library is Molecular Modeling 2nd Ed by Andrew Leach. Covers a whole boatload of methods from quantum mechanical in nature to classical in nature to thermodynamic in nature, and a variety of systems from solid state to chemical to biological.

    This is the quick, rough, and dirty guide to the topic (which means I'm sure someone will take issue with what I've said here, and they're probably right as I didn't feel like writing an immensely long and tedious post with all the details and all the qualifying comments). If you have more specific questions, feel free to ask, I do (in principle) do this stuff for a living (at least while I'm in grad school).
    Last edited: Apr 9, 2003
  4. May 7, 2003 #3
    Does anybody know of any good sites that talk about this!?!?!
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook