Dear Prof. Axel van de Walle,
Thank you for your ATAT program. It is really useful in study the phase behavior. Recently we are calculating phase diagram for CuPd binary alloy, involving BCC and FCC lattice. I use phb code in ATAT to trace the phase boundary. However, for some ground states(e.g. gs1=2 gs2=3), the phase boundary shift dramatically (from x=-0.25 to 1 within 10 K) . What happens to my calculation? Does it means one phase I consider is unstable and it undergoes phase transition?
Another question is, what is the energy of disordered phase? I know we can specify -gs=-1 to represent disordered phase. But how does ATAT know its energy as it does not show up in my convex hull.
Exactly. When a phase boundary jumps like that it is because a new phase appears.
When you use the disordered state as one of the phases, you need to specify the chemical potential explicitly with -mu=… (the code can’t determine it automatically). Typically this value would come from an earlier run where you saw a new phase appear that you suspect is the disordered phase. You would then replace whichever phase jumped by the phase "-1" and rerun specifying the -mu from the output file, just after the jump.
Wow, cool! I got it. But does it mean that this new phase appeared is always disordered phases? if the boundary shifts to x=-0.25 where other ordered phases are(e.g. gs=2 x=-0.25), does it means the gs=2 ordered phase appears?