3. Input files and output files

3.1. UPOSCAR

Information about the unit cell of crystal lattice. The format of UPOSCAR is the same as POSCAR of VASP.

  • Sample

    rocksalt-lattice
1.0
    4.157000000      0.000000000      0.000000000   # axis
    0.000000000      4.157000000      0.000000000
    0.000000000      0.000000000      4.157000000
4 4 # number of lattice sites
Direct
    0.000000000      0.000000000      0.000000000   # position of lattice site
    0.500000000      0.500000000      0.000000000
    0.500000000      0.000000000      0.500000000
    0.000000000      0.500000000      0.500000000
    0.000000000      0.000000000      0.500000000
    0.500000000      0.000000000      0.000000000
    0.000000000      0.500000000      0.000000000
    0.500000000      0.500000000      0.500000000

3.2. UPOTCAR

  • Sample

    VRHFIN =Mg:
VRHFIN =O:

3.3. position.out

Lattice sites of clusters (translational vector and basis). This file is automatically generated by cluster. And this file is used in correlation.

3.4. of.out

Lattice sites of clusters (translational vector and basis) and combination of onsite cluster functions for multicomponent systems. This file is automatically generated by correlation. And this file is used in subsequent programs, gss, cmc, gcmc, search_poscar_variance and lte.

3.5. ENERGY

  • 1st column : energy. It is preferable that unit of energy is set to ‘per unit cell’ in order to avoid any confusion.

  • 2nd column : comments (e.g. index of structure)

  • Sample

    0.8934986 1
2.4398790 2
4.3428997 3
3.7830942 4
4.8734704 5

3.6. CORRELATION

  • CORRELATION is generated by performing the following three steps for each structures.

    1. Inserting a line of comment for the structure in CORRELATION.

    2. Calculating correlation functions using correlation for the structure.

    3. Streaming standard outputs by correlation into CORRELATION after the line of comment.

  • An example of shell script for generating CORRELATION. $WORKDIR is a directory for calculating correlation functions.

    for i in {1..$n};do
    cd $WORK_DIR/$i
    echo "structure "$i >> $WORK_DIR/CORRELATION
    correlation >> $WORK_DIR/CORRELATION
done

  • Sample

    structure 1
1
0
0.5
structure 2
1
0.5
0.2
structure 3
1
0.8
-0.3

3.7. WEIGHT

  • Weights for structures when performing the weighted least squares fitting.

  • 1st column : weight

  • 2nd column : comments (e.g. structure index)

  • Sample

    1.0 1
2.0 2
1.0 3
3.0 4
5.0 5

3.8. ECI

  • ECI can be generated by erasing lines below “sum of square error” from a standard output of lsf.

  • The unit of ECIs must be set to “per unit cell (UPOSCAR)” when performing Monte Carlo simulations.

  • ECI is used in gss, cmc, gcmc, lte.

  • Sample

    0    7.9025344
1    -2.5040100
6    -0.51742919
8    3.2831494

3.9. VARIANCE

  • Covariance matrix of correlation funcitons of structures in the population.

    \bm{\Sigma}_{\alpha_1, \alpha_2} = \frac{1}{N_{\rm all}} \sum_{i=1}^{N_{\rm all}} \left( \varphi_{\alpha_1} (i) - \bm{\mu}_{\alpha_1} \right) \left( \varphi_{\alpha_2} (i) - \bm{\mu}_{\alpha_2} \right)

    where N_{\rm all} denotes the number of structures in the population. \varphi_\alpha (i) is the correlation function of cluster \alpha of structure i. \bm{\mu}_\alpha is the mean correlation function of cluster \alpha for structures in the population.

  • VARIANCE is automatically generated by variance_stpop (under development).

  • VARIANCE can be made by generating of random structures by mkposcar and calculating correlation functions of the random structures.

  • Sample

    // S_00 S_10 S_20 S_30 S_40
// S_01 S_11 S_21 S_31 S_41
// S_02 S_12 S_22 S_32 S_42
// S_03 S_13 S_23 S_33 S_43
// S_04 S_14 S_24 S_34 S_44

0 0 0 0 0
0 0.0310019 0.000310993 0.000844725 0.000183406
0 0.000310993 0.0219471 0.00205187 -0.000421977
0 0.000844725 0.00205187 0.0217157 -0.000413628
0 0.000183406 -0.000421977 -0.000413628 0.00530565

3.10. MEAN

  • Mean correlation functions of structures in the population.

    \bm{\mu}_\alpha = \frac{1}{N_{\rm all}} \sum_{i=1}^{N_{\rm all}} \varphi_\alpha (i)

    where, N_{\rm all} denotes the number of structures in the population. \varphi_\alpha (i) is the correlation function of cluster \alpha of structure i.

  • MEAN is automatically generated by variance_stpop (under development).

  • MEAN can be made by generating of random structures by mkposcar and calculating correlation functions of the random structures.

  • Sample

    1           // No. 0
0.002075    // No. 1
0.013292    // No. 2
0.006850    // No. 3
-0.007910   // No. 4

3.11. path_data

  • Energy and composition along paths for the thermodynamic integration.

  • Staring from the left, specify chemical potential, temperature, composition and energy per atom (different from ENERGY).

  • Insert comments into the beginning of a path.

  • Sample

    path 1
0.000 10 0.500000 -0.0163385
0.000 15 0.500000 -0.0163385
0.000 20 0.499997 -0.0161024
0.000 25 0.499976 -0.0161020
0.000 30 0.499911 -0.0163363
path 2
0.000 30 0.499911 -0.0163363
0.002 30 0.499648 -0.0160955
0.004 30 0.499227 -0.0160875
0.006 30 0.498583 -0.0160735
0.008 30 0.497998 -0.0162939
0.010 30 0.495421 -0.0162372
0.012 30 0.488967 -0.0160975
0.014 30 0.466447 -0.0156363
0.016 30 0.266050 -0.0128446
0.018 30 0.256566 -0.0127752
0.020 30 0.252923 -0.0127505
0.022 30 0.251300 -0.0127397
0.024 30 0.250589 -0.0127351
0.026 30 0.250278 -0.0127330
0.028 30 0.250123 -0.0127320
0.030 30 0.250053 -0.0127314