[QE-users] Convergence problem of Mg solvated ions in SCCS model

[Li, Bingxin]

Dear QE-Environ users,

I am trying to relax a Mg 2+ ions (solvated with 6 water molecules, as shown in the following figure) in the implicit water (using the SCCS model implemented in the Environ packages). The input file of the SCF calculation is attached in the email and the SCCS model is listed as follows:

&ENVIRON
 !
 verbose = 0
 environ_thr = 5.d-4
 environ_type = 'input'
 env_static_permittivity = 80
 env_electrostatic = .true.
 env_surface_tension = 0.D0
 env_pressure = 0.D0
 !
/
&BOUNDARY
/
&ELECTROSTATIC
 !
 pbc_correction = 'parabolic'
 pbc_dim = 0
 !
 tol = 1.d-20
 mix = 0.6
 !
/

[cid:image001.png at 01D7140F.617C73D0]

However, I found the SCF is hard to get converged. Although I have tried using smaller tolerance in the polarisation calculations (tol = 1.d-20) and also smaller value of the mixing beta (0.05) in the outer SCF and even starting the SCCS calculation when the outer SCF has converge into a quite small threshold (5.d -4), the SCF still cannot can converged.  It seems that the SCF is failed at certain step, with the total energy changed tenuously (see the output file).

   iteration # 16     ecut=    70.00 Ry     beta= 0.05
     CG style diagonalization
     ethr =  4.12E-06,  avg # of iterations =  3.8

     negative rho (up, down):  9.571E-03 0.000E+00

     add environment contribution to local potential

     total cpu time spent up to now is     3170.9 secs

     total energy              =    -238.62974894 Ry
     Harris-Foulkes estimate   =    -238.63007614 Ry
     estimated scf accuracy    <       0.00019209 Ry
     electrostatic embedding   =       0.00036370 Ry
     correction to one-el term =       0.00000000 Ry

     iteration # 17     ecut=    70.00 Ry     beta= 0.05
     CG style diagonalization
     c_bands:  5 eigenvalues not converged
     ethr =  4.00E-07,  avg # of iterations = 12.1

     negative rho (up, down):  9.171E-03 0.000E+00

     total cpu time spent up to now is     3573.5 secs

     total energy              =    -369.87859582 Ry
     Harris-Foulkes estimate   =    -347.04441288 Ry
     estimated scf accuracy    <      53.24399064 Ry
     electrostatic embedding   =      -0.57585314 Ry
     correction to one-el term =     -22.25879718 Ry

Could you please give me any suggestions about this calculation and help me get over this convergence problem?

I am looking forward to your reply.


Best Regards,
Bingxin Li

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20210308/ecec3a1f/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image001.png
Type: image/png
Size: 121403 bytes
Desc: image001.png
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20210308/ecec3a1f/attachment.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Mg_ions.in
Type: application/octet-stream
Size: 1882 bytes
Desc: Mg_ions.in
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20210308/ecec3a1f/attachment.obj>