Animations from A. Donev's work

I have placed some non-VRML animations related to my work on hard-particle packings on this webpage. They are pretty much a random collection of animations. After a long search for a good movie format for scientific animations, I discovered the truly wonderful MNG format (like animated GIF but for PNG):
http://www.libpng.org/pub/mng
I use mngplay under Linux and IrfanView for Windows.
On this webpage there are also older animated GIFs and MPEG 2 animations, which should be standard for any system. They are not nearly as nice as the MNG ones however.

Collision-Driven MD

MD of hard-sphere FCC crystal at phi=0.70 (MPEG) and of hard-sphere dense liquid at phi=0.49 (MPEG). Also hard-disk crystal (MPEG) and hard-disk liquid (MPEG).
Illustration of the Lubachevsky-Stillinger algorithm for ellipses (MNG) and for ellipsoids (MMs!) inside a spherical container (MNG) and for superellipsoidal cubes (MNG).
Here is an MRI scan of a real (experimental) ellipsoid packing (MNG).
Illustration of the use of near-neighbor lists (NNLs) in the MD algorithm (MNG).
Illustration of the formation of force chains in the jamming limit for a binary hard-disk packing (MNG).
Compressing monodisperse disks with a fixed unit cell produces nearly triangular polycrystalline packings with line deffects (GIF).
Including deformations of the unit cell (Parinello-Rahman-type MD) leads to perfect triangular packings with some vacancies (GIF).
Here is an illustration of the working of the free-energy BCMD algorithm for disks (MNG) and for ellipses (MNG).

Jamming

Illustration of the concept of jamming as configurational trapping: A disk is locally jammed among three fixed disks (MNG), or it is not jammed and an unjamming motion exists (MNG).
A hard-disk packing was produced by freezing some of the disks (purple). It has a low density phi~0.84 and appears jammed at first sight. However, running MD reveals that it is not and an unjamming motion becomes apparent---the new jamming density is closer to phi~0.9 (GIF).
A collectively jammed disk packing (with rattlers removed) is not strictly jammed, as apparent when running MD with a deformable unit cell (lattice vectors) (GIF).
For ellipsoids the picture is more complicated: This ellipse is locally jammed, but barely so (MNG).
A collectively jammed disk packing is stretched to make an ellipse packing. It is no longer jammed, as MD demonstrates (GIF).
The strictly jammed triangular disk crystal is stretched to make an ellipse packing. It is no longer strictly jammed, since it can be continuously sheared (GIF).

Dominos

The density of a random domino tiling is slowly increased from liquid to jamming: N=5000 (GIF) and N=1250 (GIF).
The density of a random domino tiling is slowly reduced from close packing to liquid densities using MD (GIF).
A nematic domino tiling changes density from close packing to liquid (GIF).

Binary hard disks

Melting of binary phase-separated hard-disk crystal:
At phi=0.775 melting is incomplete (MNG), but
at phi=0.763 melting is complete, though slow (MNG)
For binary hard disks phi~0.8 is the kinetic glass transition density, and relaxatio dynamics of a supercooled liquid is very slow (MNG), but at density phi=0.775 the dynamics is liquid-like, with long-time diffusion and complete mixing (ergodicity?) (MNG).
Similarly, a partially-clustered (micro-segregated) nonequilibrium binary disk system at phi=0.8 is stable over long MD runs (MNG), but at a density phi=0.775 it demixes completely (though slowly) (MNG).
Illustration of the conversion of a partitioning of the monodisperse triangular crystal into small and large disks into a binary packing of disks: no clustering (MNG), medium clustering (MNG), lots of clustering (MNG).