energy jump in SCCS

240 views
Skip to first unread message

Geng Sun

unread,
Dec 31, 2019, 12:25:44 PM12/31/19
to cp2k
Dear CP2K users,

I am trying to use the SCCS solvent model to optimize a H2O molecule in solvent and calculate the solvation energy.
I found that during the structure optimization, the total energy jumps up almost 3 eV, but the structure of H2O still looks good.
Finally, with the energy obtained in solvent ( -17.07615353442375 a.u.) and energy in vacuum (-17.22203206260123), the solvation energy is positive.
Therefore, there must be something wrong in the calculations, but I can not figure it out.


grep "Total energy:" cp2k.out gives:

  Total energy:                                               -17.22994560779211
 
Total energy:                                               -17.23016218028583
 
Total energy:                                               -17.23052945533816
 
Total energy:                                               -17.23057483762070
 
Total energy:                                               -17.23059795728501
 
Total energy:                                               -17.23061684603447
 
Total energy:                                               -17.23062875959604
 
Total energy:                                               -17.23063230147527
 
Total energy:                                               -17.23063606787884
 
Total energy:                                               -17.23458599053544
 
Total energy:                                               -17.07789974445440
 
Total energy:                                               -17.23050622895307
 
Total energy:                                               -17.23059409756609
 
Total energy:                                               -17.23153569898144
 
Total energy:                                               -17.23062905094530
 
Total energy:                                               -17.23063896643257
 
Total energy:                                               -17.06258564479369
 
Total energy:                                               -17.23061553671926
 
Total energy:                                               -17.23064419320881
 
Total energy:                                               -17.23017384159302
 
Total energy:                                               -17.09997174618866
 
Total energy:                                               -17.23053249119082
 
Total energy:                                               -17.23061888501270
 
Total energy:                                               -17.23062632201262
 
Total energy:                                               -17.10641501376824
 
Total energy:                                               -17.23063629047595
 
Total energy:                                               -17.23099984646603
 
Total energy:                                               -17.11100171504137
 
Total energy:                                               -17.10215507558170
 
Total energy:                                               -17.13946950681089
 
Total energy:                                               -17.05320236101690
 
Total energy:                                               -17.23214150951693
 
Total energy:                                               -17.14172122406513
 
Total energy:                                               -17.05319840556444
 
Total energy:                                               -17.06170695684858
 
Total energy:                                               -17.11231720859150
 
Total energy:                                               -17.07455263629729
 
Total energy:                                               -17.07623660559768
 
Total energy:                                               -17.07963295606425
 
Total energy:                                               -17.07859069368490
 
Total energy:                                               -17.07694563368373
 
Total energy:                                               -17.07637515123947
 
Total energy:                                               -17.07632103460653
 
Total energy:                                               -17.07605114342309
 
Total energy:                                               -17.07615353442375


Here is the input file:
&FORCE_EVAL
   
&DFT
      UKS
      BASIS_SET_FILE_NAME BASIS_MOLOPT_UCL
      BASIS_SET_FILE_NAME BASIS_MOLOPT
      POTENTIAL_FILE_NAME GTH_POTENTIALS
     
&MGRID
         CUTOFF
400
         COMMENSURATE
     
&END MGRID
     
&QS
         EXTRAPOLATION PS
         EXTRAPOLATION_ORDER
3
         EPS_DEFAULT  
1.0E-12
         MAP_CONSISTENT T
     
&END QS
     
&PRINT
         
&SCCS ON
           
&EACH
               QS_SCF
1
           
&END EACH
         
&END SCCS
     
&END PRINT
     
&SCCS
         ALPHA
[N*m^-1] 0.02
         GAMMA
[mN/m] 0.0
         BETA
[GPa] -0.08
         DELTA_RHO
2.0E-5
         DERIVATIVE_METHOD CD5
         DIELECTRIC_CONSTANT
6.02
         EPS_SCCS
1.0E-10
         EPS_SCF
0.03
         MAX_ITER
100
         METHOD
Andreussi
         MIXING
0.6
         
&ANDREUSSI
            RHO_MAX
0.003
            RHO_MIN
0.0003
         
&END ANDREUSSI
     
&END SCCS
     
&SCF
         EPS_SCF
1.0E-7
         MAX_SCF
31
         SCF_GUESS RESTART
         
&OT
            ENERGY_GAP
0.002
            LINESEARCH
2PNT
            PRECONDITIONER FULL_ALL
            MINIMIZER DIIS
         
&END OT
         
&OUTER_SCF
            EPS_SCF
1.0E-7
            MAX_SCF
20
         
&END OUTER_SCF
     
&END SCF
     
&XC
         
&XC_FUNCTIONAL PBE
         
&END XC_FUNCTIONAL
     
&END XC
   
&END DFT
   
&SUBSYS
     
&KIND H
         BASIS_SET TZVP
-MOLOPT-GTH
         POTENTIAL GTH
-PBE-q1
     
&END KIND
     
&KIND O
         BASIS_SET TZVP
-MOLOPT-GTH
         POTENTIAL GTH
-PBE-q6
     
&END KIND
     
&COORD
         O
7.424704500000000706e+00 9.621645000000000891e+00 1.106072519999999848e+01
         H
7.912573799999999657e+00 1.000790999999999897e+01 1.180680414000000056e+01
         H
6.473416799999999860e+00 9.760409999999998476e+00 1.119866495999999856e+01
     
&END COORD
     
&CELL
         PERIODIC XYZ
         A
1.641000000000000014e+01 0.000000000000000000e+00 0.000000000000000000e+00
         B
0.000000000000000000e+00 1.650000000000000000e+01 0.000000000000000000e+00
         C
0.000000000000000000e+00 0.000000000000000000e+00 1.824599999999999866e+01
     
&END CELL
   
&END SUBSYS
   
&PRINT
     
&FORCES ON
         LOG_PRINT_KEY T
     
&END FORCES
   
&END PRINT
&END FORCE_EVAL



I greatly appreciate any suggestions on fixing this.

Thank you very much.

Best Regards,

Geng

Krack Matthias (PSI)

unread,
Jan 7, 2020, 1:40:51 PM1/7/20
to cp...@googlegroups.com

Hi Geng

 

Such fluctuations are usually caused by numerical noise. A larger cutoff, e.g. 800 instead of 400Ry, will most likely resolve that issue.

 

HTH

 

Matthias

--
You received this message because you are subscribed to the Google Groups "cp2k" group.
To unsubscribe from this group and stop receiving emails from it, send an email to cp2k+uns...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/be04892b-30f6-4654-b51a-5acc8228e885%40googlegroups.com.

Tzu-Yao HSU

unread,
Jun 4, 2020, 10:44:04 AM6/4/20
to cp2k
Hi all,

I got similar problem as Geng, but my calculation is pure DFT energy evaluation.
During the SCF run, the polarisation energy abruptly jumped up to a positive value and the SCF convergence jumped down at certain convergence level .
Then the polarisation energy always stayed positive until the SCF run converged.

Here are my input file (.inp) and reduced output file (.out, grep 'Diag.' and 'Polarisation energy' then paste) .
My molecular structure (optimized.dat) was obtained by performing GEO_OPT in vacuum.

Further increasing the MGRID CUTOFF, EPS_SCF in SCCS section, simulation box size, and decreasing CUTOFFs in XC section do not change the converged energy.
Since the converged polarisation energy is always positive, I still wonder how may I improve my input file?

I appreciate any response to this problem.
Thank you very much
Best regards
Tzu-Yao


Matthias Krack於 2020年1月7日星期二 UTC+1下午7時40分51秒寫道:

To unsubscribe from this group and stop receiving emails from it, send an email to cp...@googlegroups.com.

neutral_L30_e1500.inp
neutral_L30_e1500.out
optimized.dat

Krack Matthias (PSI)

unread,
Jun 4, 2020, 4:39:13 PM6/4/20
to cp2k

Hi Tzu-Yao

 

You may try smaller values for RHO_MAX and RHO_MIN. A cut-off of 600 Ry (maybe even 400 Ry) should large enough.

 

HTH

 

Matthias

To unsubscribe from this group and stop receiving emails from it, send an email to cp2k+uns...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/9ba5de66-d129-4f24-8c09-139d9f834286%40googlegroups.com.

Tzu-Yao HSU

unread,
Jun 9, 2020, 2:41:03 PM6/9/20
to cp2k
Thank you very much Matthias!

I made it work by lowering RHO_MIN. (Obtained negative polarisation energy instead of positive value)
However, if I use a cut-off of 600 Ry, I would still get a positive polarisation energy.

Tzu-Yao

Matthias Krack於 2020年6月4日星期四 UTC+2下午10時39分13秒寫道:
Reply all
Reply to author
Forward
0 new messages