Single point energy diverges in large box, but converges in smaller box? P_Mix problem?

[Andreas Döll]

Hallo,

I want to calculate RESP charges for a COF structure with pbc in x and y direction. For ease I put vacuum on top to get the charges of a singular layer. This procedure worked out for another structure with 66 atoms. 

For this structure which has 132 atoms, the single point energy calculation FAILS! The energy quickly diverges if I consider a large box in z direction, but converges if I have a smaller box. What is the reason for that? At the P_Mix step it diverges, I cannot even get to DIIS minimizer!

I have attached the input file resp.inp, the only thing changed between runs is the cell in z direction from 16 A to 10 A as well as the xyz structure and the output for the large cell. The working output.dat is also attached. Note again that the only thing which I changed is the smaller cell.

Any input is appreciated, as I do not understand why the cell size makes the energy diverge!

Sincerely, 

Andreas Doell

[Marcella Iannuzzi]

Hi Andreas, 

This is a structure problem. Having a triclinic cell with c vector  1.625     1.123    10.481 

you cannot expect to add something to the z component and maintain the same symmetry. 

The angles alpha and beta are change and probably the box is not anymore consistent with the coordinates.

Regards

Marcella

CELL_TOP| Vector a [angstrom    37.453     0.000     0.000   |a| =    37.452754

 CELL_TOP| Vector b [angstrom   -20.749    29.823     0.000   |b| =    36.331227

 CELL_TOP| Vector c [angstrom     1.625     1.123    10.481   |c| =    10.665429

 CELL_TOP| Angle (b,c), alpha [degree]:                                90.034600

 CELL_TOP| Angle (a,c), beta  [degree]:                                81.235440

 CELL_TOP| Angle (a,b), gamma [degree]:                               124.828009

CELL_TOP| Vector a [angstrom    37.453     0.000     0.000   |a| =    37.452754

 CELL_TOP| Vector b [angstrom   -20.749    29.823     0.000   |b| =    36.331227

 CELL_TOP| Vector c [angstrom     1.625     1.123    16.481   |c| =    16.598865

 CELL_TOP| Angle (b,c), alpha [degree]:                                90.022232

 CELL_TOP| Angle (a,c), beta  [degree]:                                84.381360

 CELL_TOP| Angle (a,b), gamma [degree]:                               124.828009

[Andreas Döll]

you are 100% right, my mistake! 

But why is it then exploding in the original box with 16 A, but works in the smaller box?

[Marcella Iannuzzi]

you need to find yourself what the right box is

sorry, we cannot help with that

Marcella