A nice summary (in pdf format) by Michael Francis at Accelrys about the use of solid-state density functional theory (DFT) in the assignment of the HMX THz spectrum published in J. Phys. Chem. A earlier this year. Also, I'm on record again extolling the virtues of DMol3, having now used it in a number of inelastic neutron scattering studies, this crazy terahertz business, and for (an upcoming) examining the crystal environment of small polyoxometalates (which I wouldn't have even considered as early as two years ago given the state of computers way back when). Three more papers on the subject have come out since this case study was presented, all of which are linked and available in the terahertz spectroscopy category of this site.
In press, available from ChemPhysChem. Wasting no time putting it all on the site. This is the sister paper to the solid-state density functional theory (DFT) paper on HMX from J. Phys. Chem. A (including the complete assignment of modes and motions of the PETN crystal using the DMol3 program) that made pretty good press. As for the determined agreement with experiment in these projects, two points certainly make a line. At the very least, the futility of isolated-molecule calculations for the assignment of crystalline THz spectra is certainly evident in the analysis. There's quite a bit more work to be had for this project, of course. The good chaps at Teraview, Ltd., the supplier of the spectra, have recently gone cryogenic, which is going to do wonders for the theoretical assignments.
D. G. Allis and T. M. Korter
Abstract: The experimental solid-state terahertz (THz) spectrum (3 to 120 cm-1) of the high explosive pentaerythritol tetranitrate (PETN, C5H6N4O12) has been modeled using solid-state density functional theory (DFT) calculations. Solid-state DFT employing the BP density functional is in best qualitative agreement with the features in the previously reported THz spectrum. The crystal environment of PETN includes numerous intermolecular hydrogen-bonding interactions that contribute to large (up to 80 cm-1) calculated shifts in molecular normal mode positions in the solid-state. Comparison of the isolated-molecule and solid-state normal mode calculations for a series of density functionals reveals the extent to which the inclusion of crystal packing interactions and the relative motions between molecules are required for correctly reproducing the vibrational structure of solid-state THz spectra. The THz structure below 120 cm-1 is a combination of both intermolecular (relative rotations and translations) and intramolecular (torsions, large amplitude motions) vibrational motions. Vibrational mode analyses indicate that the first major feature (67.2 cm-1) in the PETN THz spectrum contains all of the optical rotational and translational cell modes and no internal (molecular) vibrational modes.