search

prisms-center / CASMcode

First-principles statistical mechanical software for the study of multi-component crystalline solids
Other
104 stars 69 forks source link

control chemical potential in GC LTE #258

Open sjtuzhanglei opened 2 years ago

sjtuzhanglei commented 2 years ago

Dear CASM developers,

I tried running a GC LTE for my system. By specifying the param_chem_pot(a), the oxide configuration falls off the neutral configuration. I am wondering if by possibly specifying chem_pot(SPECIES) or other ways, we can strict the grand canonical LTE into only the neutral sub-space.

I only have one independent composition parametric component in the composition definition, wondering why it cannot strict the calculation along this neutral direction.

cat lte1gc.json { "comment" : "This is a sample input file. Unrecognized attributes (like the ones prepended with '') are ignored.", "debug" : true, "ensemble" : "grand_canonical", "method" : "lte1", "model" : { "formation_energy" : "formation_energy" }, "supercell" : [ [2, 0, 0], [0, 2, 2], [0, -4, 4]
], "data" : { "storage" : { "write_observations" : true, "write_trajectory" : true, "output_format" : ["csv", "json"] } }, "driver" : { "mode" : "incremental", "dependent_runs" : false, "motif" : { "configname" : "restricted_auto", "_configname" : "SCEL1_1_1_1_0_0_0/6", "_configdof" : "path/to/final_state.json" }, "initial_conditions" : { "param_chem_pot" : { "a" : -1.0 }, "temperature" : 300.0, "tolerance" : 0.001 }, "final_conditions" : { "param_chem_pot" : { "a" : 0.0 }, "temperature" : 300.0, "tolerance" : 0.001 }, "incremental_conditions" : { "param_chem_pot" : { "a" : 0.01 }, "temperature" : 0.0, "tolerance" : 0.001 } } }

The LTE gives me a BaZrO0.75 config, which is neither the neutral config nor the calculated config.

configname T phi_LTE Beta gs_potential_energy gs_formation_energy param_ch

em_pot(a) gs_comp(a) gs_comp_n(Ba) gs_comp_n(Zr) gs_comp_n(Y) gs_comp_n(Va) gs_comp_n(O) SCEL4_1_4_1_2_0_0/195 300.00000000 -26.79776305 38.68168650 -26.79776305 -27.69776305 -1 .00000000 0.90000000 1.00000000 1.00000000 0.00000000 2.25000000 0.75000000

sjtuzhanglei commented 2 years ago

I know that in canonical ensemble, we can control the comp_n(SPECIES) or comp(a) to let the run only in neutral space, but in the semi-grand_canonical ensemble, chem_pot is the first derivative of Phi with comp. If there is no control of comp in neutral space, only specifying chem_pot won't help, right?

bpuchala commented 2 years ago

The current grand canonical Monte Carlo release cannot be restricted to a composition subspace.