Regarding MC simulations with CASM

[darjaved]

i have carried out the MC simulations at temperature of 5 K, i am getting the results.json file as:

{
 "<atom_frac(Na)>" : [ 0.371831216496, 0.500000000000, 0.500000000000, 0.500000000000, 0.575350322421, 0.593002909594, 0.612579571759, 0.620607827597, 0.642320172594, 0.687500000000, 0.687500000000, 0.705047768891, 0.750000000000 ],
 "<atom_frac(Sn)>" : [ 0.628168783504, 0.500000000000, 0.500000000000, 0.500000000000, 0.424649677579, 0.406997090406, 0.387420428241, 0.379392172403, 0.357679827406, 0.312500000000, 0.312500000000, 0.294952231109, 0.250000000000 ],
 "<clex_hull_dist(casm_learn_input,comp)>" : [ -0.068404327316, -0.000957283596, -0.000957283596, -0.000957283596, -0.028529665079, -0.033709611181, -0.037318901596, -0.037224163492, -0.037995706874, -0.023744743678, -0.023744743678, -0.024143933864, -0.014415744778 ]

What is that clex_hull_dist here? Can i get the idea of how is my CE?. Is it predicting new ground state structures?

[xivh]

Do you get any new ground states if you enumerate additional structures and test your cluster expansion (before Monte Carlo)?

[darjaved]

I have enumerated only upto 2*2*2 supercell. For that this is what i am getting and 2*2*2 supercell has around 400 structures in my expansion. Any suggestions?

…

On Sat, 15 Jun 2024 at 4:19 AM, xivh ***@***.***> wrote: Do you get any new ground states if you enumerate additional structures and test your cluster expansion (before Monte Carlo)? — Reply to this email directly, view it on GitHub <#363 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/A6AKEKUV5CQQFWHH5BO4MWDZHNXQZAVCNFSM6AAAAABHR5SUXKVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDCNRYHA2TGNBRGY> . You are receiving this because you authored the thread.Message ID: ***@***.***>

[xivh]

You should try enumerating larger supercells and predicting their energy with the cluster expansion to see if they break your cluster expansion convex hull.

[darjaved]

What if i do my MC with the existing CE? How much will it differ. Because i think i have to take the very large supercell to see if it breaks the convex hull ( there some inter metallics for this system with primitive cell having large number of atoms)

…

On Sat, 6 Jul 2024 at 12:13 AM, xivh ***@***.***> wrote: You should try enumerating larger supercells and predicting their energy with the cluster expansion to see if they break your cluster expansion convex hull. — Reply to this email directly, view it on GitHub <#363 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/A6AKEKQ4MA4H7BZX2UQ5JLLZK3SNPAVCNFSM6AAAAABHR5SUXKVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDEMJRGI3TSMBWGU> . You are receiving this because you authored the thread.Message ID: ***@***.***>