problem with SQS

Alloy Theoretic Automated Toolkit (ATAT)
1

Hi, everybody
I am trying to build a cation-disordered delta-phase compound Y6UO12.
when I used the command "mcsqs -n=171", I got the following message
Impossible to match point correlations due to incompatible supercell size.
my input file is

6.5297 6.5297 6.5297 99.047 99.047 99.047
1 0 0
0 1 0
0 0 1
0.500000000 0.000000000 0.000000000 U=0.14285714284714 Y=0.85714285714286
0.106500000 0.318100005 0.145899996 U=0.14285714284714 Y=0.85714285714286
0.893499970 0.681900024 0.854099989 U=0.14285714284714 Y=0.85714285714286
0.181899995 0.854099989 0.393500000 U=0.14285714284714 Y=0.85714285714286
0.818099976 0.145900011 0.606500030 U=0.14285714284714 Y=0.85714285714286
0.354099989 0.393500000 0.681900024 U=0.14285714284714 Y=0.85714285714286
0.645900011 0.606500030 0.318099976 U=0.14285714284714 Y=0.85714285714286
0.414999992 0.041999999 0.691999972 O
0.585000038 0.958000004 0.308000028 O
0.458000004 0.307999998 0.085000001 O
0.541999996 0.692000031 0.915000021 O
0.192000002 0.915000021 0.041999999 O
0.807999969 0.084999979 0.958000004 O
0.321000010 0.574999988 0.407999992 O
0.679000020 0.425000012 0.592000008 O
0.075000003 0.407999992 0.820999980 O
0.925000012 0.592000008 0.179000020 O
0.092000000 0.179000005 0.425000012 O
0.907999992 0.820999980 0.574999988 O

I think my supercell is big enough. when I tried "mcsqs -rc", I got the same error message.
How can I solve this problem?

thank u.

2

It is not simply a question of the supercell being big enough, there are a limited number of acceptable sizes. Because you have a 6:1 ratio of Y:U occupancy, each symmetry-inequivalent site must occur a multiple of seven times in your supercell - no more, no less. If the supercell contains, for example, 16 copies of one site, it is impossible for the program to fill that site with a 6:1 ratio of cations - instead it must have either 13:3 or 14:2, and neither of these will allow the correct correlations. It would need to have either 14 copies of that site, or 21.

I would suggest you use the cellcvrt utility with the -s option to first find the smallest possible cell, which should allow you to identify your inequivalent sites, and then choose a supercell size such that each of them appears a multiple of seven times.

3

thank u steven, I finally understand how that works.