This semester (Spring 2013) I taught a new 400-level course here at Biola University: a second-semester Physical Chemistry II class (with laboratory) that will expand our previous one-semester Physical Chemistry offering into a two-semester option for students majoring in chemistry, and for others interested in an upper division elective.
Of course, the pedagogy of statistical thermodynamics hasn't changed much since I was an undergraduate: students learn to generalize microscopic states to bulk thermodynamic properties by studying the canonical ensemble, partition functions, etc. I recall wading through this material as an undergraduate, and occasionally becoming confused as to where all the theory was headed.
This time around, I took a lab period, and one lecture, to cover the application of statistical thermodynamics to the study of chemical systems by molecular dynamics (MD) methods. The excellent open-source Ballview software ran well on my students' computers (MAC, Windows, and Linux ports are available), and since most of this crop of students are biochemistry majors, I took the opportunity to acquaint them with the wealth of structural data online at rcsb.org, as well: a tool I hope they will put to good use in their senior year biochemistry courses and beyond.
In one portion of this semester's lab, my students performed simple molecular dynamics calculations to probe a phase change of interest to biochemists: the melting temperature of an RNA "hairpin" (actually a slightly modified data set derived from Ihle, et al,"A novel cGUUAg tetraloop structure with a converved yYNMGg-type backbone conformation from cloverleaf 1 of bovine enterovirus 1 RNA" (2005) Nucleic Acids Res. 33:2003-2011, rcsb record 1Z30). I supported this lab by taking students through the mathematics well described in an assigned supplemental reading of Furio Ercolessi's "A Molecular Dynamics Primer"
Calculations in Ballview required about 40-60 minutes per temperature, and were run in parallel by assigning each student a temperature to simulate. While calculations were underway, students were instructed to estimate the Tm for the structure using thermodynamic methods (see Richard Owczarzy, Peter M. Vallone, Frank J. Gallo, Teodoro M. Paner, Michael J. Lane, and Albert S. Benight, Predicting sequence-dependent melting stability of short duplex DNA oligomers, Biopolymers, 1997, Vol. 44, No. 3, pp 217-239). This exercise was a good review in its own right.
Aside from the opportunity to compare Tm values computed using MD and thermodynamic methods, the excellent visualization tools in Ballview permitted students to appreciate qualitatively the normal modes of macromolecules, the partition of vibrational energy between light atoms and heavier ones, the complexity of solvent shell interactions, and the degree to which the H-bonding of nucleic acid base pairs is influenced by thermal kinetic energy.
Purely as an example of the simple MD calculations possible on a student's laptop in Ballview, I made the video below to illustrate the valiant struggle of the H-bonds as they hold this small RNA fragment in "double strand" form against a thermodynamic temperature of 350K. I believe this brief tour through MD gave my students:
a tool for visualizing biomolecules that is clearly superior to their 2-D textbooks
an understanding of when and how classical physics can be brought to bear at the molecular level
a quantitative and qualitative understanding of the relevance of statistical thermodynamics as the "glue" between the microscopic and macroscopic.