Question on SQS

Alloy Theoretic Automated Toolkit (ATAT)
1

Dear Axel, dear all,

I want to produce an sqs structure which contains 2 Ti atoms, 10 Zr atoms and 6 Si, The following is the rndstr.in:

6.88781 6.88799 3.46046 90.0000 90.0000 120
1 0 0
0 1 0
0 0 1
0.0000073923427867 -0.0000091903537424 0.0000000000000000 Si=1
0.3333631028059582 0.6666508419502295 0.0000000000000000 Si=1
0.6666573468948027 0.3332940714063822 0.5000000000000000 Si=1
0.0692230822102227 0.3332948930985555 0.5000000000000000 Ti=0.16667, Zr=0.83333
0.9263057113991990 0.5929208867543792 0.0000000000000000 Ti=0.16667, Zr=0.83333
0.6666939660191561 0.0736702098850560 0.0000000000000000 Ti=0.16667, Zr=0.83333
0.2641685703628827 0.9307915616209347 0.5000000000000000 Ti=0.16667, Zr=0.83333
0.6666713807938354 0.7358703791969377 0.5000000000000000 Ti=0.16667, Zr=0.83333
0.4070522311711613 0.3333093154412584 0.0000000000000000 Ti=0.16667, Zr=0.83333
the following is the command lines:
corrdump -l=rndstr.in -ro -noe -nop -2=3
mcsqs -n=18

I don’t know why it shows "Impossible to match point correlations due to incompatible supercell size".
But if I set "Ti=0.5, Zr=0.5" for the last 6 line in rndstr.in , it can produce an SQS structure Ti6Zr6Si6. Who can tell me why ? And how can I build an SQS-structure for TixZr12-xSi6 (x=2,3,4…10)?

2

0.16667*12 is not an integer (to good decimal place). That error you got is commonly found if when you multiply any compositional ratio by the number of total atoms of that sublattice, you get a non integer value.

Try adding a couple more 6’s and it might work to the correct decimal place.

3

Thanks to Forrest for identifying the problem.