Ice-likeness along a short nucleation run

Ice-likeness here is the mean Lechner-Dellago \(\bar q_6\). A liquid sits near 0.2. A crystal sits near 0.5. A nucleus in a liquid pulls the mean up as it grows, which is why the score is the whole-frame average and not a single molecule.

The example trajectory has three frames. GenIce ice Ih is shaken into a liquid. The basal half of the original lattice is the nucleus. Frame 0 is the shaken liquid. Frame 1 puts half of that nucleus back on its lattice sites. Frame 2 puts the rest back. OpenMM relaxes each frame for two steps with those sites restrained, so the liquid can move and the nucleus cannot dissolve before you measure it.

The three frames give

nucleation frame 0 q6bar 0.276023
nucleation frame 1 q6bar 0.320504
nucleation frame 2 q6bar 0.334254

The mean rises at each stage.

Full run

The paper trajectory is figshare 10.6084/m9.figshare.11448702.v1: 148 frames of 4096 molecules, md5 c28fc209a2854d2ba49e8f9b55accbe8. Run seams steinhardt on every frame.

nucleation full paper molecules 4096 frames 148
nucleation full frame 1 q6bar 0.26063
nucleation full frame 148 q6bar 0.287256
nucleation full lines 148

Frame 1 of that file is the shipped input/traj/nucleation_f1.lammpstrj. The mean \(\bar q_6\) on frame 1 is 0.26063. Frame 148 is 0.287256.