Page 1 of 1

small EDIFF

Posted: Mon Dec 29, 2025 4:30 am
by jasius

Dear all, in the attached system I optimizing Co3O4 slab. This and related jobs suffer from a very peculiar problem - the energy difference goes to zero but convergence is not signalled and I get fatal error in bracketing.

I would like an advice on why this may arise as all of the jobs of this system suffer from the same problem

sorry I can't attach OUTCAR, too large

thank you

167 F= -.61977686E+03 E0= -.61976030E+03 d E =-.338499E-05 mag= 66.6723
curvature: -0.00 expect dE=-0.345E-08 dE for cont linesearch -0.498E-10
ZBRENT: bisectioning
opt : 0.0741 next Energy= -619.776858 (dE=-0.339E-05)
bond charge predicted
N E dE d eps ncg rms rms(c)
DAV: 1 -0.619776857920E+03 -0.33731E-07 -0.49539E-09 1836 0.321E-04 0.258E-04
RMM: 2 -0.619776857971E+03 -0.51274E-07 -0.34171E-09 425 0.334E-04 0.355E-03
RMM: 3 -0.619776857929E+03 0.41622E-07 -0.30044E-09 423 0.277E-04 0.834E-04
RMM: 4 -0.619776857939E+03 -0.96843E-08 -0.25570E-10 222 0.161E-04
168 F= -.61977686E+03 E0= -.61976030E+03 d E =-.340931E-05 mag= 66.6723
curvature: -0.00 expect dE=-0.222E-08 dE for cont linesearch -0.443E-10
ZBRENT: interpolating
opt : 0.0741 next Energy= -619.776858 (dE=-0.338E-05)
bond charge predicted
N E dE d eps ncg rms rms(c)
DAV: 1 -0.619776857939E+03 -0.10055E-07 -0.18674E-09 1902 0.259E-04 0.215E-04
RMM: 2 -0.619776857965E+03 -0.25091E-07 -0.52866E-10 208 0.279E-04 0.178E-03
RMM: 3 -0.619776857941E+03 0.23276E-07 -0.26129E-10 216 0.194E-04 0.940E-04
RMM: 4 -0.619776857950E+03 -0.83091E-08 -0.93057E-11 168 0.136E-04
169 F= -.61977686E+03 E0= -.61976030E+03 d E =-.341981E-05 mag= 66.6723
curvature: -0.00 expect dE=-0.205E-09 dE for cont linesearch -0.296E-11
ZBRENT: can not reach accuracy
opt : 0.0000 next Energy= 0.000000 (dE= 0.620E+03)
bond charge predicted
N E dE d eps ncg rms rms(c)
DAV: 1 -0.619776854074E+03 0.38671E-05 -0.20522E-04 3456 0.324E-02 0.882E-03
RMM: 2 -0.619776856671E+03 -0.25966E-05 -0.53409E-06 2728 0.739E-03 0.268E-02
RMM: 3 -0.619776857088E+03 -0.41757E-06 -0.21248E-06 2040 0.481E-03 0.317E-02
RMM: 4 -0.619776856550E+03 0.53867E-06 -0.18349E-06 2043 0.462E-03
170 F= -.61977686E+03 E0= -.61976030E+03 d E =-.201989E-05 mag= 66.6723
curvature: 0.00 expect dE= 0.108-320 dE for cont linesearch 0.598-321
ZBRENT: fatal error in bracketing
please rerun with smaller EDIFF, or copy CONTCAR
to POSCAR and continue


Re: small EDIFF

Posted: Mon Dec 29, 2025 9:22 am
by pedro_melo

Dear jasius,

Thank you for your report. Before we proceed, I would like clear some inconsistencies in your INCAR, between the comments and the variables you are using.

You have "SYSTEM = Chabazite_01Ns0, PBE, PAW, ISPIN=1, ISIF=2, PREC=NORMAL, ENCUT=400eV, Symmetry is OFF, single K-point". While VASP does not assign values to any tag in this line, below it you use ISPIN = 2 and and you set "ISYM = 2 ! Symmetry is ON". Also your KPOINTS file has more than a single k-point. It seems that there is some inconsistency between the type of run you declare in your SYSTEM tag and the values you ended up using. Could you clarify if this is what you intended to do?

Also, I noticed that you wrote "ISIF = 2 ! The cell shape and the volume is relaxed". But with ISIF=2 the cell chape and cell volume are not degrees of freedom, so they are kept the same. Is this also intended?

Kind regards,
Pedro


Re: small EDIFF

Posted: Mon Dec 29, 2025 1:50 pm
by jasius

Dear Pedro, everything in INCAR is correct. Annotations/annotations are old and irrelevant, sorry for the confusion. This is spin unrestricted system with with ISYM=2 and ISIF=2. I do not understand why all of these cobalt oxide calculations behave this strange way in my case, e.g. they all run out into small EDIFF, can you help me out please?


Re: small EDIFF

Posted: Mon Dec 29, 2025 2:38 pm
by pedro_melo

Dear jasius,

Thank you for the clarification.

In your INCAR the stopping criteria for geometry optimisation is set by EDIFFG= -0.005, so VASP will only stop when the magnitude of all forces is below 0.005, not when the difference in energy between two consecutive ionic steps is below 0.005. Your energy differences seem to have reached that threshold, but maybe your forces are still beyond that.

Could you upload a file with the forces on each atom for the ionic steps 169 to 170, i.e. the same steps you pasted in your initial message? Just the section from the OUTCAR file which has that information would be OK.

Kind regards,
Pedro


Re: small EDIFF

Posted: Mon Dec 29, 2025 5:08 pm
by jasius

See attached several end steps and thenk you for your help


Re: small EDIFF

Posted: Tue Dec 30, 2025 1:37 am
by jasius

Update: ADDGRID does not help


Re: small EDIFF

Posted: Tue Dec 30, 2025 10:42 am
by pedro_melo

Dear jasius,

Looking at the section from the OUTCAR file that you sent I think the issue is that your system is stuck in a configuration that is a minimum in energy and unable to move away from there. The forces on some atoms are still larger than the value you set for EDIFFG, so VASP tries to continue to reduce them, but with no success.

Could you try with a different value for EDIFFG, maybe EDIFF= -0.02, and see if the calculation is able to finish? If that happens then you can lower the value of EDIFFG and restart the job using the WAVECAR you just got and see if you can reduce the forces.

Kind regards,
Pedro


Re: small EDIFF

Posted: Tue Dec 30, 2025 12:42 pm
by jasius

Hi Pedro, I brute forced the optimization convergence with PREC=ACCURATE and ENCUT=520

But now I am facing another problem. I understand that those structures are not global minima so I calculate frequencies and observe negative vibration, displace along it and repeat the process. The problem is that the converged structure does not converge SCF in frequency calculations. Runs out of 100 cycles. Anything I can do to improve SCF convergence (I find it strange that SCF does not converge in the first SCF frequency cycle after optimization converged all the cycles)

LDA part: xc-table for Pade appr. of Perdew
POSCAR, INCAR and KPOINTS ok, starting setup
FFT: planning ...
WAVECAR not read
entering main loop
N E dE d eps ncg rms rms(c)
DAV: 1 0.343503089776E+04 0.34350E+04 -0.12368E+05 4200 0.131E+03
DAV: 2 0.501894265296E+03 -0.29331E+04 -0.26686E+04 4500 0.368E+02
DAV: 3 -0.214801016205E+03 -0.71670E+03 -0.67029E+03 4944 0.201E+02
DAV: 4 -0.275225100313E+03 -0.60424E+02 -0.57789E+02 4452 0.625E+01
DAV: 5 -0.277348125304E+03 -0.21230E+01 -0.21073E+01 4788 0.130E+01 0.831E+01
RMM: 6 -0.296891296923E+03 -0.19543E+02 -0.62642E+02 4200 0.773E+01 0.930E+01
RMM: 7 -0.259026011420E+03 0.37865E+02 -0.13065E+02 4252 0.334E+01 0.688E+01
RMM: 8 -0.317738142442E+03 -0.58712E+02 -0.50943E+01 4293 0.209E+01 0.953E+01
RMM: 9 -0.273024562575E+03 0.44714E+02 -0.98202E+01 4218 0.313E+01 0.585E+01
RMM: 10 -0.258673968083E+03 0.14351E+02 -0.52005E+01 4206 0.223E+01 0.613E+01
RMM: 11 -0.245795984721E+03 0.12878E+02 -0.20631E+01 4207 0.141E+01 0.496E+01
RMM: 12 -0.261493394823E+03 -0.15697E+02 -0.10308E+01 4222 0.909E+00 0.625E+01
RMM: 13 -0.247343444076E+03 0.14150E+02 -0.63063E+00 4221 0.768E+00 0.551E+01
RMM: 14 -0.245845583517E+03 0.14979E+01 -0.64050E+00 4228 0.787E+00 0.311E+01
RMM: 15 -0.243632823112E+03 0.22128E+01 -0.23533E+00 4209 0.488E+00 0.166E+01
RMM: 16 -0.246276234566E+03 -0.26434E+01 -0.47401E-01 4250 0.206E+00 0.193E+01
RMM: 17 -0.247132848062E+03 -0.85661E+00 -0.23430E-01 4223 0.152E+00 0.207E+01
RMM: 18 -0.247215508503E+03 -0.82660E-01 -0.50442E-02 4323 0.684E-01 0.210E+01
RMM: 19 -0.247233593981E+03 -0.18085E-01 -0.11469E-01 4289 0.936E-01 0.212E+01
RMM: 20 -0.247565336233E+03 -0.33174E+00 -0.73704E-01 4260 0.213E+00 0.207E+01
RMM: 21 -0.248412505750E+03 -0.84717E+00 -0.35824E-01 4443 0.161E+00 0.225E+01
RMM: 22 -0.248214105604E+03 0.19840E+00 -0.13957E-01 4255 0.113E+00 0.222E+01
RMM: 23 -0.247459674611E+03 0.75443E+00 -0.61479E-01 4238 0.205E+00 0.172E+01
RMM: 24 -0.248201035097E+03 -0.74136E+00 -0.33834E-01 4364 0.147E+00 0.162E+01
RMM: 25 -0.248784902568E+03 -0.58387E+00 -0.16994E-01 4240 0.100E+00 0.163E+01
RMM: 26 -0.248863731950E+03 -0.78829E-01 -0.65123E-02 4245 0.712E-01 0.165E+01
RMM: 27 -0.248475450001E+03 0.38828E+00 -0.15289E-01 4280 0.110E+00 0.173E+01
RMM: 28 -0.247773070243E+03 0.70238E+00 -0.30369E-01 4298 0.144E+00 0.189E+01
RMM: 29 -0.247857003796E+03 -0.83934E-01 -0.45804E-02 4270 0.598E-01 0.179E+01
RMM: 30 -0.248512900455E+03 -0.65590E+00 -0.17570E-01 4220 0.121E+00 0.164E+01
RMM: 31 -0.248919383806E+03 -0.40648E+00 -0.86478E-01 4245 0.259E+00 0.139E+01
RMM: 32 -0.250078408977E+03 -0.11590E+01 -0.71686E-01 4245 0.237E+00 0.135E+01
RMM: 33 -0.248430735474E+03 0.16477E+01 -0.19069E+00 4243 0.378E+00 0.149E+01
RMM: 34 -0.248428810456E+03 0.19250E-02 -0.35378E-01 4322 0.167E+00 0.138E+01
RMM: 35 -0.249046562084E+03 -0.61775E+00 -0.94682E-02 4428 0.822E-01 0.133E+01
RMM: 36 -0.249354615905E+03 -0.30805E+00 -0.92204E-02 4234 0.736E-01 0.122E+01
RMM: 37 -0.249710944553E+03 -0.35633E+00 -0.10291E-01 4243 0.716E-01 0.123E+01
RMM: 38 -0.249793625120E+03 -0.82681E-01 -0.23087E-02 4322 0.374E-01 0.124E+01
RMM: 39 -0.249799320932E+03 -0.56958E-02 -0.44971E-03 4369 0.189E-01 0.125E+01
RMM: 40 -0.249884394360E+03 -0.85073E-01 -0.71457E-03 4262 0.251E-01 0.129E+01
RMM: 41 -0.250061178266E+03 -0.17678E+00 -0.14911E-02 4253 0.325E-01 0.132E+01
RMM: 42 -0.250183687880E+03 -0.12251E+00 -0.28707E-02 4233 0.374E-01 0.136E+01
RMM: 43 -0.249997090219E+03 0.18660E+00 -0.64486E-02 4233 0.628E-01 0.141E+01
RMM: 44 -0.249950553845E+03 0.46536E-01 -0.80361E-03 4252 0.247E-01 0.142E+01
RMM: 45 -0.249953514819E+03 -0.29610E-02 -0.16886E-03 4612 0.109E-01 0.143E+01
RMM: 46 -0.250098818415E+03 -0.14530E+00 -0.63556E-03 4229 0.226E-01 0.142E+01
RMM: 47 -0.250050296950E+03 0.48521E-01 -0.34046E-03 4223 0.172E-01 0.143E+01
RMM: 48 -0.249951914818E+03 0.98382E-01 -0.11574E-02 4230 0.297E-01 0.145E+01
RMM: 49 -0.246320601333E+03 0.36313E+01 -0.11576E+00 4213 0.292E+00 0.169E+01
RMM: 50 -0.244496979382E+03 0.18236E+01 -0.59632E-01 4289 0.210E+00 0.179E+01
RMM: 51 -0.244629160755E+03 -0.13218E+00 -0.49307E-01 4368 0.178E+00 0.173E+01
RMM: 52 -0.240846878013E+03 0.37823E+01 -0.10007E+00 4213 0.265E+00 0.215E+01
RMM: 53 -0.242006952124E+03 -0.11601E+01 -0.19990E-01 4265 0.123E+00 0.250E+01
RMM: 54 -0.242760017152E+03 -0.75307E+00 -0.17309E-01 4299 0.115E+00 0.255E+01
RMM: 55 -0.242283246498E+03 0.47677E+00 -0.11287E-01 4392 0.861E-01 0.260E+01
RMM: 56 -0.245756529316E+03 -0.34733E+01 -0.10017E+00 4239 0.245E+00 0.231E+01
RMM: 57 -0.245788968170E+03 -0.32439E-01 -0.79678E-02 4292 0.816E-01 0.229E+01
RMM: 58 -0.245247701531E+03 0.54127E+00 -0.62103E-02 4293 0.647E-01 0.234E+01
RMM: 59 -0.245328463171E+03 -0.80762E-01 -0.29088E-02 4439 0.397E-01 0.234E+01
RMM: 60 -0.245656479248E+03 -0.32802E+00 -0.15315E-02 4281 0.327E-01 0.231E+01
RMM: 61 -0.243639494498E+03 0.20170E+01 -0.20325E+00 4236 0.345E+00 0.191E+01
RMM: 62 -0.247926138057E+03 -0.42866E+01 -0.33357E+00 4328 0.378E+00 0.155E+01
RMM: 63 -0.248363326844E+03 -0.43719E+00 -0.11107E+00 4317 0.207E+00 0.154E+01
RMM: 64 -0.248391539023E+03 -0.28212E-01 -0.54277E-01 4316 0.133E+00 0.181E+01
RMM: 65 -0.248922359905E+03 -0.53082E+00 -0.16320E-01 4268 0.960E-01 0.191E+01
RMM: 66 -0.251812996285E+03 -0.28906E+01 -0.18535E+00 4222 0.365E+00 0.137E+01
RMM: 67 -0.253181051309E+03 -0.13681E+01 -0.89258E-01 4217 0.205E+00 0.160E+01
RMM: 68 -0.254064630609E+03 -0.88358E+00 -0.71362E-01 4269 0.160E+00 0.189E+01
RMM: 69 -0.254001956303E+03 0.62674E-01 -0.93925E-02 4243 0.779E-01 0.221E+01
RMM: 70 -0.254528671774E+03 -0.52672E+00 -0.16687E-01 4208 0.952E-01 0.232E+01
RMM: 71 -0.253827107567E+03 0.70156E+00 -0.79789E-01 4202 0.232E+00 0.194E+01
RMM: 72 -0.254859845385E+03 -0.10327E+01 -0.15992E-01 4218 0.115E+00 0.212E+01
RMM: 73 -0.255332890535E+03 -0.47305E+00 -0.65103E-02 4250 0.628E-01 0.222E+01
RMM: 74 -0.255551225459E+03 -0.21833E+00 -0.61896E-02 4249 0.674E-01 0.224E+01
RMM: 75 -0.255588699158E+03 -0.37474E-01 -0.88055E-02 4230 0.741E-01 0.225E+01
RMM: 76 -0.254552962170E+03 0.10357E+01 -0.42446E-01 4210 0.178E+00 0.207E+01
RMM: 77 -0.254597708144E+03 -0.44746E-01 -0.40420E-02 4236 0.585E-01 0.209E+01
RMM: 78 -0.254663574012E+03 -0.65866E-01 -0.46102E-02 4220 0.617E-01 0.208E+01
RMM: 79 -0.254684719662E+03 -0.21146E-01 -0.23773E-02 4242 0.404E-01 0.197E+01
RMM: 80 -0.254756421166E+03 -0.71702E-01 -0.28185E-02 4229 0.478E-01 0.188E+01
RMM: 81 -0.252786356060E+03 0.19701E+01 -0.18710E+00 4212 0.388E+00 0.133E+01
RMM: 82 -0.253664633422E+03 -0.87828E+00 -0.73784E-01 4213 0.262E+00 0.122E+01
RMM: 83 -0.254015143213E+03 -0.35051E+00 -0.21413E-01 4217 0.128E+00 0.120E+01
RMM: 84 -0.254384062014E+03 -0.36892E+00 -0.26522E-01 4242 0.141E+00 0.130E+01
RMM: 85 -0.254607069978E+03 -0.22301E+00 -0.11257E-01 4252 0.943E-01 0.155E+01
RMM: 86 -0.254709610011E+03 -0.10254E+00 -0.10305E-01 4270 0.873E-01 0.173E+01
RMM: 87 -0.254617343667E+03 0.92266E-01 -0.82591E-02 4214 0.717E-01 0.169E+01
RMM: 88 -0.254660629362E+03 -0.43286E-01 -0.23723E-02 4329 0.353E-01 0.163E+01
RMM: 89 -0.254801283431E+03 -0.14065E+00 -0.22950E-02 4214 0.398E-01 0.158E+01
RMM: 90 -0.254915446594E+03 -0.11416E+00 -0.19103E-02 4237 0.337E-01 0.152E+01
RMM: 91 -0.254631363642E+03 0.28408E+00 -0.95554E-02 4213 0.884E-01 0.137E+01
RMM: 92 -0.254639726430E+03 -0.83628E-02 -0.76083E-02 4218 0.748E-01 0.132E+01
RMM: 93 -0.254644848418E+03 -0.51220E-02 -0.37149E-02 4304 0.496E-01 0.132E+01
RMM: 94 -0.254603718906E+03 0.41130E-01 -0.52843E-03 4215 0.214E-01 0.132E+01
RMM: 95 -0.254633726919E+03 -0.30008E-01 -0.24667E-03 4302 0.131E-01 0.132E+01
RMM: 96 -0.254627923415E+03 0.58035E-02 -0.92589E-02 4237 0.842E-01 0.134E+01
RMM: 97 -0.254311666386E+03 0.31626E+00 -0.45758E-02 4222 0.594E-01 0.132E+01
RMM: 98 -0.254269129946E+03 0.42536E-01 -0.52547E-03 4469 0.178E-01 0.132E+01
RMM: 99 -0.254296705589E+03 -0.27576E-01 -0.10252E-03 4260 0.100E-01 0.132E+01
RMM: 100 -0.254528741087E+03 -0.23204E+00 -0.27638E-02 4219 0.508E-01
1 F= -.25452874E+03 E0= -.25452232E+03 d E =-.128330E-01 mag= 24.9444
Finite differences POTIM= 0.01500 DOF= 132
bond charge predicted
N E dE d eps ncg rms rms(c)
DAV: 1 -0.255392572162E+03 -0.10959E+01 -0.54779E+00 6216 0.106E+01 0.142E+01
RMM: 2 -0.285319564182E+03 -0.29927E+02 -0.14064E+01 4241 0.933E+00 0.718E+01
RMM: 3 -0.249646481456E+03 0.35673E+02 -0.14900E+01 4201 0.114E+01 0.186E+01
RMM: 4 -0.253329454178E+03 -0.36830E+01 -0.30847E+00 4256 0.521E+00 0.120E+01
RMM: 5 -0.254341911389E+03 -0.10125E+01 -0.12822E+00 4224 0.363E+00 0.758E+00
RMM: 6 -0.254738791641E+03 -0.39688E+00 -0.34930E-01 4263 0.162E+00 0.554E+00
RMM: 7 -0.255014189990E+03 -0.27540E+00 -0.22433E-01 4282 0.126E+00 0.371E+00
RMM: 8 -0.255131885344E+03 -0.11770E+00 -0.12071E-01 4241 0.102E+00 0.251E+00
RMM: 9 -0.255198187262E+03 -0.66302E-01 -0.11204E-01 4267 0.897E-01 0.184E+00
RMM: 10 -0.255282953406E+03 -0.84766E-01 -0.55105E-02 4255 0.655E-01 0.198E+00
RMM: 11 -0.255305049127E+03 -0.22096E-01 -0.435


Re: small EDIFF

Posted: Mon Jan 05, 2026 4:58 pm
by jasius

Hi Pedro,

I really struggle to converge this system in frequency calculations. It goes through the optimization but then when I do frequencies, SCF does not converge.
First question is why it gives me the warning below when I read from WAVECAR

found WAVECAR, reading the header
number of bands has changed, file: 222 present: 246
trying to continue reading WAVECAR, but it might fail
number of k-points has changed, file: 25 present: 34
trying to continue reading WAVECAR, but it might fail

I just optimized, copied WAVECAR with ISTART=1 and immediately got this message. Still, even reading from the WAVECAR SCF cycle on the previously converged system fails, see attached and thank you

see attached and thank you


Re: small EDIFF

Posted: Wed Jan 07, 2026 10:47 am
by pedro_melo

Dear jasius,

Did you change any parallelisation parameters in between runs? Or any other parameters in your INCAR file, for that matter. From the warnings it seems that something changed in between runs.

For example: VASP adds a few empty states to help with convergence, but the number of extra states has to be divisible by the number of cores. See https://vasp.at/wiki/NBANDS#Parallelization for more information.

Could you post here the input files for both runs?

Kind regards,
Pedro


Re: small EDIFF

Posted: Wed Jan 07, 2026 2:41 pm
by jasius

I just copied INCAR file and changed form IBRION=2 or IBRION=5... In general, how can I improve convergence of that particular Co3O4 (100) slab, even with reading WAVECAR it fails SCF convergence? What I do not understand is why converged system at the last step of the optimization does not converge when copied/pasted into IBRION=5 run


Re: small EDIFF

Posted: Mon Jan 12, 2026 10:04 am
by pedro_melo

Dear jasius,

Thank you for your reply. I'm looking into this issue, but don't really have an answer. I will speak with my colleagues to know if they have seen such behaviour before and how to solve it.

I'll get back with an answer as soon as I can.

Kind regards,
Pedro