References for Molecular Dynamics?

Question

I am new to the world of Molecular Dynamics (zero practical experience), and I know that a starting point to really go into the back-bone of MD is Statistical Mechanics (took a semester course in undergrad, loved it, will be the focus of my PhD).

My ultimate goal is to become a developer of advanced or modern MD, such as Many-body Molecular and Quantum MD methods (e.g. Path Integral MD). Example systems I am interested in being able to model are quantum fluids and super-critical fluids.

What are some essential books I need to get and read in order to become an avid MD theoretician? Are there any must read papers I need to dedicate my weekend mornings to?

Answer 1

A good place to start is the classical Allen-Tildesley book, Computer Simulation of Liquids, which covers the basics of molecular dynamics that hasn't really changed in a long time. The book can be supplemented with literature, such as review articles for whatever it is you want to do.

Software manuals are also often quite useful to find out how things are done nowadays. For instance, GROMACS has had very good documentation for a long time; here is a link to their current reference manual which covers a lot of topics and has references to the original literature.

Answer 2

Another extremely popular resource is Frenkel and Smit's textbook "Understanding Molecular Simulation". It covers all basics on molecular dynamics, Monte Carlo, some common enhanced sampling methods, free energies and it even derives the Ewald summation. This is basically a book that doesn't shy away from messy derivations and it gives you a lot of example code for many of the basic algorithms. Great for a first course in MD.

A personal favourite which is more theoretical is Tuckerman's "Statistical Mechanics: Theory and Molecular Simulation". While it is certainly not as practical as the first book, it is specifically oriented towards theory of molecular simulation and it includes many interesting topics, including path integral methods so this one is also very relevant.

Finally, a very good book on the aspect of integrators is Leimkuhler and Reich's "Simulating Hamiltonian Dynamics" textbook. I'd say it's much more advanced than the other two books and is strictly theoretical, so while it might not be of immediate use to you, it is definitely a good read.

Answer 3

Another good book that starts from the very beginning and it's very hands-on is "The art of molecular dynamics" by D. C. Rapaport. It is particularly useful if you want to code up a MD code yourself. This is something I always recommend, even if you want to go on and use one of the big MD packages out there, since you will learn the essentials and troubleshoot many of the classic issues that one encounters in MD simulations (setting the time step, coupling a thermostat/barostat, understanding how neighbour lists work, etc.).

Answer 4

Also worth mentioning Bill Hoover's excellent textbooks, Molecular Dynamics which provide an introduction and a slightly different perspective on molecular dynamics, along with Computational Statistical Mechanics. They are freely downloadable from his website.

I also second Tuckerman's "Statistical Mechanics: Theory and Molecular Simulation" as one of the most accessible introdutions to the topic and Rapaport "The art of molecular dynamics" for anyone who wants to write an MD code themself.

For non-equilibrium molecular dynamics, classics are the comprehensive but fairly steep Evans and Morriss "Statistical mechanics of nonequilibrium liquids" and the more accessible Todd and Daivis "Nonequilibrium Molecular Dynamics Theory, Algorithms and Applications"

Answer 5

Since you are interested in quantum simulation, I would actually suggest reading Feynman and Hibbs Quantum Mechanics and Path Integrals. It is very cheap which is a big plus and also remarkably readable as everything Feynman wrote is. Even though the book would be considered advanced because path integrals are considered advanced, much of what the book does is frame fairly basic statistical mechanics in terms of path integrals.

There are other very good reasons to read this book based on your question. One is that Feynman did some of the earliest work on very low temperature Helium which is about as quantum of a fluid as one can get. Helium turns into a super-fluid at very low temperatures, and one can understand and model this using path integrals quite naturally. This is actually discussed a bit in the book. Feynman also provides a nice introduction to some of the problems which are still major problems in quantum statistical mechanics today. Those being things like the "sign problem" (of which there are actually a few different ones).

As far as academic literature goes, this isn't molecular dynamics, but I would suggest reading the review on the theory of condensed helium, Rev. Mod. Phys. 67, 279, as he gives a very nice introduction to path integral monte carlo and actually closes the loop on a lot of the things Feynman discusses in the book I mentioned before. Also, there are a lot more conceptual similarities between path integral monte carlo and path integral molecular dynamics than between ordinary monte carlo and ordinary molecular dynamics.

Finally, you should absolutely read the following paper: Chandler, D., & Wolynes, P. G. (1981). Exploiting the isomorphism between quantum theory and classical statistical mechanics of polyatomic fluids. The Journal of Chemical Physics, 74(7), 4078-4095.

This paper uses the fact that the propagator of path integrals in imaginary time has the same form as the classical partition function of a classical system in an extended phase space.

Finally, for more modern work in PIMD, I would suggest looking at the work of Michele Ceriotti, who is one of the primarily developers of i-PI, which is a good package for doing PIMD simulations.