how to define cut-off for large cell

Alloy Theoretic Automated Toolkit (ATAT)
1

Hello,

I am not sure how to define the cut-off for a 30 atoms unite cell when calculating the correlation functions. For example, for the first two shells of nearest neighbors are to be considered for pair correlation functionl, what value should I choose for -2=?

corrdump -l=rndstr.in -ro -noe -nop -clus -2=?

Below is the nearest neighbor table generated by VASP.

ion position nearest neighbor table
1 0.018 0.348 1.000- 28 2.59 23 2.60 19 2.62 21 2.65 11 3.72
2 0.652 0.670 0.333- 20 2.59 18 2.60 17 2.62 30 2.65 3 3.72
3 0.301 0.308 0.472- 24 2.63 14 2.63 30 2.67 18 2.84 16 2.89 2 3.72 12 3.90 9 3.90

4 0.699 0.692 0.972- 23 2.63 28 2.63 19 2.67 15 2.84 29 2.89 5 3.72 10 3.90 11 3.90

5 0.348 0.330 0.833- 26 2.59 15 2.60 25 2.62 19 2.65 4 3.72
6 0.670 0.018 0.166- 29 2.59 27 2.60 21 2.62 17 2.65 12 3.72
7 0.330 0.982 0.666- 16 2.59 22 2.60 13 2.62 25 2.65 10 3.72
8 0.982 0.652 0.500- 14 2.59 24 2.60 30 2.62 13 2.65 9 3.72
9 0.993 0.301 0.639- 22 2.63 16 2.63 13 2.67 24 2.84 26 2.89 8 3.72 10 3.90 3 3.90

10 0.692 0.993 0.806- 15 2.63 26 2.63 25 2.67 22 2.84 28 2.89 7 3.72 9 3.90 4 3.90

11 0.007 0.699 0.139- 27 2.63 29 2.63 21 2.67 23 2.84 20 2.89 1 3.72 12 3.90 4 3.90

12 0.308 0.007 0.306- 18 2.63 20 2.63 17 2.67 27 2.84 14 2.89 6 3.72 11 3.90 3 3.90

13 0.954 0.652 0.636- 7 2.62 8 2.65 9 2.67
14 0.338 0.965 0.453- 8 2.59 3 2.63 12 2.89
15 0.737 0.657 0.828- 5 2.60 10 2.63 4 2.84
16 0.373 0.338 0.619- 7 2.59 9 2.63 3 2.89
17 0.698 0.046 0.303- 2 2.62 6 2.65 12 2.67
18 0.263 0.343 0.328- 2 2.60 12 2.63 3 2.84
19 0.348 0.302 0.969- 1 2.62 5 2.65 4 2.67
20 0.965 0.627 0.286- 2 2.59 12 2.63 11 2.89
21 0.046 0.348 0.136- 6 2.62 1 2.65 11 2.67
22 0.657 0.919 0.661- 7 2.60 9 2.63 10 2.84
23 0.081 0.737 0.994- 1 2.60 4 2.63 11 2.84
24 0.919 0.263 0.494- 8 2.60 3 2.63 9 2.84
25 0.302 0.954 0.803- 5 2.62 7 2.65 10 2.67
26 0.035 0.373 0.786- 5 2.59 10 2.63 9 2.89
27 0.343 0.081 0.161- 6 2.60 11 2.63 12 2.84
28 0.662 0.035 0.953- 1 2.59 4 2.63 10 2.89
29 0.627 0.662 0.119- 6 2.59 11 2.63 4 2.89
30 0.652 0.698 0.469- 8 2.62 2 2.65 3 2.67

Regards
Vera

2

The cutoff has nothing to do with the unit cell size.
To be rigorous, you just need to check how the property you want (e.g. energy) converges as the cutoff(s) are increased.
Researchers often go the other way, decide on a supercell size that can live with and then find an sqs that matches as many correlations as possible for that size.