I have a low temperature single crystal X-ray structure of a compound that exhibits a reversible phase change by DSC at 500K. I would like to investigate this phase change to see how the structure changes at high temperature. Is there a way to set the temperature of the cell when performing a geometry optimization or is there a better way to go about determining the new structure? Also, how do you set the temperature for which thermal ellipsoids are calculated in a PhononDOS calculation? I can only seem to get them calculated at 298K.
Temperature plays no role in geometry optimization. It is simply trying to find the lowest energy structure. It sounds like your crystal has at least two minima in its energy surface. Typically, optimization will find the nearest one to your starting point.
The only place temperature plays a role is in molecular dynamics. Well, it also plays a role if you do phonons (see germanium tutorial).