I am an undergrad researcher trying to study the melting temperature of gold using an MD simulation with LAMMPS. I'm trying to do some preliminary exercises to make sure that I understand how the program works. I'm working on finding the thermal expansion coefficient by heating up a 5 by 5 by 5 block of FCC gold (using the gold eam potential that came with LAMMPS) and measuring the lattice parameter at each temperature. Unfortunately, my sample seems to be melting at 100K . My units are set to real, which should mean that everything is measured in K. I've tried using nph and temp/rescale, npt, nve with press/brendensen and temp/rescale. All of these result in a complete loss of all crystal structure. I even tried just leaving the sample on NVT without modifying the pressure and it still melted. Any ideas as to why this is happening? Please help!