The file, POSCAR.xyz, has a particle of 32 atoms fairly distributed in a cubical space, 0.588 x 0.599 nm. The cubical boundary is made up of 28 atoms with only four atoms located at the vertex of the cube, and the remaining four atoms are located in the center of the cube. The distance of an atom to its nearest neighboring atom is 0.282nm.

In the file, movie.xyz, as the energy was minimized, the structure of the particle has changed from a perfect cube into a slightly spherical shape. The distances between atoms have changed from 0.282nm to 0.236nm, 0.298nm, and more. Before, there were only angles 60, 90, 180; but now the angles are 88.3, 174.0, etc. (Degree Celsius)

-145.291327-32 x -0.176408= -139.646271 eV (Binding energy of Pt32) My hypothesis is that the binding energy of the Pt nanoparticle should be less than the cohesive energy of the bulk Pt, because bulk Pt is so much larger in size and more atoms are bounded at all sides. * The actual cohesive energy of bulk Pt is 134.9kcal/mol, thus it is smaller than the binding energy of Pt32.

I had used VMD to rearrange the atoms, changing structure of the Pt32 from a cube into a sphere while leaving the center atoms(4 of them)intact. The calculation is as followed: -135.954495-32 x -0.176408= -130.309439 eV (Binding energy of low energy Pt32) * The low energy Pt32 required 4.590561eV less energy than the bulk Pt, thus making it more stable.