Convergence issue with smear tag
Clicked A Few Times
2:35:16 PM PST - Tue, Nov 18th 2014
echo
title "cluster"
geometry autosym
.
.
.
.
.
.
end

basis "ao basis" cartesian
Pd library lanl2dz_ecp file /usr/local/NWChem/data/libraries/
end

ecp
Pd library lanl2dz_ecp file /usr/local/NWChem/data/libraries/
end

dft
vectors input cluster-3.movecs output cluster-smear.movecs

iterations 3000

direct

mult 1

XC xpbe96 cpbe96

convergence ncyds 3000 

grid fine
tolerances accqrad 20
smear 0.001

end

DRIVER
XYZ a.xyz

MAXITER 5000
END


task dft energy

I already have an optimized geometry which did not include any smearing. However when I use that optimized geometry in the above input file, the the scf cycles don't converge at all. Part of the output is shown below whose deltaE is always b/w D-01 and D+01 (even for ~ 100 cycles).

d= 0,ls=0.0,diis 1 -2563.1088837131 -1.07D+04 1.71D-02 1.02D+01 207.6
Grid integrated density:     367.999970645462

Requested integration accuracy:   0.10E-05

sigma= .10D-02 eFermi= -0.07880 T*S= -4.45E-05 nel(1)=  368.00 nel(

d= 0,ls=0.0,diis     2  -2562.7372814973  3.72D-01  5.69D-02  5.25D+00   389.4

Grid integrated density:     367.999952201265

Requested integration accuracy:   0.10E-05

sigma= .10D-02 eFermi= -0.14160 T*S= -1.62E-04 nel(1)=  368.00 nel(

d= 0,ls=0.0,diis     3  -2483.8130328019  7.89D+01  4.61D-02  3.72D+02   576.9

Grid integrated density:     367.999968493405

Requested integration accuracy:   0.10E-05

sigma= .10D-02 eFermi= -0.14870 T*S= -2.08E-04 nel(1)=  368.00 nel(

d= 0,ls=0.0,diis     4  -2550.5940778012 -6.68D+01  3.07D-02  8.44D+01   762.3

Grid integrated density:     367.999971195797

Requested integration accuracy:   0.10E-05

sigma= .10D-02 eFermi= -0.14698 T*S= -1.90E-04 nel(1)=  368.00 nel(

d= 0,ls=0.0,diis     5  -2545.8670568312  4.73D+00  2.81D-02  1.09D+02   947.5

Thanks for any guidance.

Forum Vet
4:17:15 PM PST - Fri, Nov 21st 2014
A way converge this is case is to first converge the closed-shell molecule with a plus one charge,
then use this orbitals for the neutral doublet.
A word of caution: the basis set you have chosen exhibits linear dependence.


charge +1
dft
 xc pbe0
 iterations 999
 vectors input atomic output Cu23_plus.movecs
 convergence lshift 0.2
end
set quickguess t
task dft 

charge 0
dft
 mult 2
 vectors input  Cu23_plus.movecs output Cu23_neutral.movecs
 end
task dft optimize

Clicked A Few Times
12:05:46 PM PST - Sun, Nov 23rd 2014
Thank you for your response.

Sorry, I forgot to mention that my system already is a closed shell with transition metal cluster having ethanol adsorbed. So do you still think I should make my closed shell system +1 charge and proceed with your given input file? Following is my complete input file if you want to look over it.


echo
title "cluster"

geometry autosym
  1. geo optimzied
Pd -3.22878706 2.81149077 0.75340399
Pd                   -0.45704722     3.20883783     0.77273295

Pd                    2.31469261     3.60618489     0.79206191

Pd                   -4.27032308     0.21267956     0.70668192

Pd                   -1.50388322     0.61576568     0.73287763

Pd                    1.27574771     1.01724638     0.75517366

Pd                    4.04489640     1.40472073     0.76466880

Pd                   -2.54011930    -1.98878459     0.67928880

Pd                    0.23964499    -1.58776657     0.74282917

Pd                    3.00336038    -1.19409048     0.71794673

Pd                   -0.80991550    -4.19024875     0.65189569

Pd                    1.96182433    -3.79290170     0.67122464

Pd                   -1.71747674     2.24694320    -1.53526968

Pd                    1.05426309     2.64429025    -1.51594073

Pd                   -2.75901278    -0.35186802    -1.58199176

Pd                    0.01272706     0.04547904    -1.56266280

Pd                    2.78446689     0.44282610    -1.54333384

Pd                   -1.02880898    -2.55333217    -1.60938488

Pd                    1.74293085    -2.15598512    -1.59005591

C                    -0.38788185    -0.51148103     3.77152084

C                    -0.77881090    -0.79465830     5.21257262

H                    -1.23393133    -0.07986040     3.19679913

H                     0.45362467     0.19317299     3.70666471

H                    -1.08517192     0.13922353     5.70993533

H                     0.06533156    -1.22529683     5.77159028

H                    -1.62871698    -1.49436465     5.26747987

H                    -0.64824848    -2.35386007     3.05825989

O                     0.08896011    -1.71126376     3.10195854
end


basis "ao basis" cartesian
Pd library lanl2dz_ecp file /usr/local/NWChem/data/libraries/
C library 6-311G** file /usr/local/NWChem/data/libraries/
O library 6-311G** file /usr/local/NWChem/data/libraries/
H library 6-311G** file /usr/local/NWChem/data/libraries/
end

dft
vectors input cluster-3.movecs output cluster-smear.movecs

iterations 3000

direct

mult 1

XC xpbe96 cpbe96

convergence ncyds 3000

smear 0.001

grid fine
tolerances accqrad 20
end

DRIVER
XYZ a.xyz

MAXITER 5000
END

task dft energy

Clicked A Few Times
6:23:09 PM PST - Fri, Dec 12th 2014
So I was able to get the calculation going for a similar system but after 20 ionic steps it again stops converging; change in energy between each scf cycle (deltaE) keeps shuttling between d+1 and d-1 as shown in the first post.
It should also be mentioned that for these cases doing a vacuum runs okay (basically for different systems, it sometimes converges when using smear and sometimes when I don't use smear it converges; no idea why). But for SMD, the already converged vacuum wavefunctions doesn't help converge the COSMO gas phase calculations and most often (here also) it is the COSMO gas phase that stops converging.
Input:
echo
title "cluster"
start cluster
geometry autosym
Pd 1.70943005 -2.14308979 0.89662505
Pd -1.06559264 -2.51414462 0.94528092
Pd 2.77348016 0.40093843 0.41029405
Pd -0.00175805 0.04718270 0.54937384
Pd -2.77656521 -0.34117121 0.50760582
Pd 1.06250759 2.57391181 -0.02738105
Pd -1.71251512 2.20285699 0.02127484
Pd 0.23318027 -1.97942240 -1.47712719
Pd 1.29723038 0.56460581 -1.96345816
Pd -1.47779230 0.19355100 -1.91480229
C 0.40882077 1.91696584 3.35410663
C -0.25397607 2.54962114 4.55944859
H 0.43806335 2.62591555 2.50097303
H 1.44010328 1.60533040 3.57263399
H 0.30156279 3.45405187 4.85532314
H -0.26743193 1.85650399 5.41503530
H -1.28948898 2.84430813 4.32565940
H -1.21072521 0.95787029 2.75550044
O -0.28187350 0.70205456 2.93749940
end
basis "large" cartesian
Pd library lanl2dz_ecp file /usr/local/NWChem/data/libraries/
C library 6-311G** file /usr/local/NWChem/data/libraries/
O library 6-311G** file /usr/local/NWChem/data/libraries/
H library 6-311G** file /usr/local/NWChem/data/libraries/
end
ecp
Pd library lanl2dz_ecp file /usr/local/NWChem/data/libraries/
end
set geometry:actlist 4 11:19
set "ao basis" "large"
dft
vectors output pbe0.movecs ?#?tried? SCF, B3LYP too which didn't converge but PBE0 went on to give me 20 ionic steps
iterations 500
direct
XC PBE0
convergence ncyds 500
smear 0.001
end
task dft
dft
vectors input pbe0.movecs output dft.movecs
iterations 1000
direct
mult 1
XC xpbe96 cpbe96 ?#?pbe96? is the final XC I want to use
convergence ncyds 1000
smear 0.001
end
DRIVER
XYZ a.xyz
MAXITER 500
END
cosmo
do_cosmo_smd true
solvent water
dielec 80.4
end
set cosmo:map cosmo-bondi.par
task dft optimize

Clicked A Few Times
6:25:06 PM PST - Fri, Dec 12th 2014
I could not post the whole thing together. So when I do "grep @ output_file"

@ Step Energy Delta E Gmax Grms Xrms Xmax Walltime
@ ---- ---------------- -------- -------- -------- -------- -------- --------
@ 0 -1422.46927315 0.0D+00 0.00597 0.00119 0.00000 0.00000 8681.7
@ 1 -1422.46981948 -5.5D-04 0.00743 0.00228 0.02413 0.12849 17517.0
@ 2 -1422.47044524 -6.3D-04 0.00803 0.00190 0.02796 0.12531 26755.1
@ 3 -1422.47118389 -7.4D-04 0.00496 0.00162 0.03045 0.08617 39807.3
@ 4 -1422.47153302 -3.5D-04 0.00533 0.00139 0.02118 0.08575 68618.6
@ 5 -1422.47196273 -4.3D-04 0.00987 0.00282 0.03740 0.11470 75591.6
@ 6 -1422.47231723 -3.5D-04 0.00502 0.00113 0.01022 0.03380 84140.3
@ 7 -1422.47269121 -3.7D-04 0.00659 0.00174 0.03798 0.09107 92429.5
@ 8 -1422.47299636 -3.1D-04 0.00323 0.00101 0.01293 0.03605 105150.5
@ 9 -1422.47317946 -1.8D-04 0.00573 0.00169 0.02867 0.07629 117606.4
@ 10 -1422.47332189 -1.4D-04 0.00306 0.00080 0.00720 0.02312 160223.1
@ 11 -1422.47353210 -2.1D-04 0.00490 0.00135 0.03030 0.10080 167589.6
@ 12 -1422.47370703 -1.7D-04 0.00459 0.00109 0.01810 0.06483 201010.9
@ 13 -1422.47385470 -1.5D-04 0.00415 0.00122 0.03116 0.10145 214450.8
@ 14 -1422.47392198 -6.7D-05 0.00324 0.00082 0.00516 0.01597 218354.7
@ 15 -1422.47404269 -1.2D-04 0.00533 0.00131 0.03172 0.11620 226441.3
@ 16 -1422.47421307 -1.7D-04 0.00259 0.00065 0.01231 0.04168 235024.9
@ 17 -1422.47429355 -8.0D-05 0.00705 0.00138 0.03656 0.13893 279233.8
@ 18 -1422.47441467 -1.2D-04 0.00202 0.00061 0.00879 0.02289 286687.5
@ 19 -1422.47445428 -4.0D-05 0.00997 0.00174 0.06015 0.23781 301495.8
@ 20 -1422.47459086 -1.4D-04 0.00171 0.00056 0.02224 0.08821 310739.6
@ 21 -1422.47473261 -1.4D-04 0.00366 0.00086 0.03085 0.11152 318049.1

Similarly just for Pd cluster with ethanol, it went till 70 steps and then the same problem.

Forum Vet
9:22:58 PM PST - Tue, Dec 16th 2014
After looking at the 21 optimization steps, I don't see any problem in this geometry optimization. It just seems a hard case to optimize.

Clicked A Few Times
10:57:39 PM PST - Fri, Dec 19th 2014
Thak you for your response. So the 21 steps are fine. But for the next (22nd step) the electronic SCF does the following:

d= 0,ls=0.0,diis     1  -1422.4228059610 -4.24D+03  2.62D-02  4.72D-02324599.5

d= 0,ls=0.0,diis     2  -1419.3656492454  3.06D+00  1.96D-02  2.58D+01324638.6
.
.
d= 0,ls=0.0,diis    20  -1421.6291504436 -3.12D+00  1.57D-02  5.37D+00325277.5

d= 0,ls=0.0,diis    21  -1422.0141945978 -3.85D-01  1.37D-02  3.07D+00325316.7

d= 0,ls=0.0,diis    22  -1422.2203634481 -2.06D-01  1.45D-02  1.74D+00325355.1

d= 0,ls=0.0,diis    23  -1421.9293534908  2.91D-01  1.15D-02  3.34D+00325394.8

d= 0,ls=0.0,diis    24  -1422.3858426713 -4.56D-01  7.86D-03  3.69D-01325431.6

d= 0,ls=0.0,diis    25  -1422.2879766300  9.79D-02  9.54D-03  1.00D+00325467.3
.
.
.
d= 0,ls=0.0,diis 54 -1414.7170046919 2.52D+00 3.35D-02 5.40D+01326493.0
d= 0,ls=0.0,diis    55  -1414.5033236239  2.14D-01  3.04D-02  4.31D+01326527.7

d= 0,ls=0.0,diis    56  -1415.7564851070 -1.25D+00  3.14D-02  4.20D+01326561.7

d= 0,ls=0.0,diis    57  -1413.1572689243  2.60D+00  3.76D-02  6.64D+01326595.3

d= 0,ls=0.0,diis    58  -1401.2257986795  1.19D+01  4.13D-02  4.76D+01326630.3

d= 0,ls=0.0,diis    59  -1401.7042682279 -4.78D-01  3.50D-02  5.91D+01326668.2
.
.
.
d= 0,ls=0.0,diis 132 -1394.0935913883 -1.11D+01 5.51D-02 6.82D+01329417.5
d= 0,ls=0.0,diis   133  -1393.9825069211  1.11D-01  2.82D-02  1.60D+02329453.3

d= 0,ls=0.0,diis   134  -1403.5399975205 -9.56D+00  5.92D-03  1.17D+02329494.0

d= 0,ls=0.0,diis   135  -1400.4911030725  3.05D+00  5.00D-02  1.12D+02329533.0

d= 0,ls=0.0,diis   136  -1385.1628955159  1.53D+01  5.79D-02  1.41D+02329573.3
.
.
d= 0,ls=0.0,diis 155 -1325.1747422602 8.04D+01 6.48D-02 3.89D+02330299.1
d= 0,ls=0.0,diis   156  -1345.7669300301 -2.06D+01  4.55D-02  2.34D+02330331.6

d= 0,ls=0.0,diis   157  -1388.2766727841 -4.25D+01  6.26D-02  9.78D+01330364.6

d= 0,ls=0.0,diis   158  -1376.4337595372  1.18D+01  9.10D-02  1.67D+02330397.4
.
.
d= 0,ls=0.0,diis   244  -1396.9068848743  1.75D+01  4.19D-02  5.55D+01333354.4

d= 0,ls=0.0,diis   245  -1399.9061871473 -3.00D+00  3.54D-02  5.40D+01333387.5

d= 0,ls=0.0,diis   246  -1418.5992062226 -1.87D+01  3.07D-02  2.50D+01333420.8
.
.
d= 0,ls=0.0,diis   375  -1397.2718083660  1.77D+01  3.82D-02  5.91D+01338645.7

d= 0,ls=0.0,diis   376  -1412.8643753581 -1.56D+01  2.87D-02  4.62D+01338681.5

d= 0,ls=0.0,diis   377  -1417.4775320464 -4.61D+00  3.07D-02  2.98D+01338716.8

d= 0,ls=0.0,diis   378  -1418.5810586490 -1.10D+00  3.68D-02  2.66D+01338750.4
.
.
d= 0,ls=0.0,diis   382  -1405.6183330736  7.18D+00  3.90D-02  9.01D+01338889.0

d= 0,ls=0.0,diis   383  -1400.8098202537  4.81D+00  4.29D-02  5.17D+01338923.7

d= 0,ls=0.0,diis   384  -1395.4073309806  5.40D+00  4.29D-02  5.40D+01338959.6
.
.
d= 0,ls=0.0,diis   421  -1407.6589098064  1.00D+01  3.69D-02  7.70D+01340273.3

d= 0,ls=0.0,diis   422  -1399.3046011525  8.35D+00  3.08D-02  4.87D+01340311.1

d= 0,ls=0.0,diis   423  -1420.9746433833 -2.17D+01  3.30D-02  9.05D+00340353.4

d= 0,ls=0.0,diis   424  -1403.6049952363  1.74D+01  3.90D-02  7.17D+01340390.6

d= 0,ls=0.0,diis   425  -1406.3224562101 -2.72D+00  3.85D-02  6.70D+01340426.5
.
.

d= 0,ls=0.0,diis   515  -1398.6148617789 -1.50D+00  3.93D-02  6.10D+01344002.3

d= 0,ls=0.0,diis   516  -1398.0746506544  5.40D-01  4.00D-02  6.58D+01344043.0

d= 0,ls=0.0,diis   517  -1398.7168932119 -6.42D-01  4.79D-02  5.73D+01344081.3

Thank you for your time and help.

Forum Vet
11:17:26 AM PST - Tue, Dec 23rd 2014
My suggestion is the following:
Extract the last geometry from the output file and start the calculation from scratch


Forum >> NWChem's corner >> General Topics