SCF not convergence, I can't optimize my structure in solution environment which was optimized in vacuum

Questions
#1

SCF not convergence, I can’t optimize my structure in solution environment which was optimized in vacuum. This is my input(structure was optimized in vacuum environment):
$molecule
1 1
Cl 1.5746745308 0.0070770309 -0.0289671039
N 4.5855687067 -0.0029326501 -0.0235189175
N -4.8998317514 0.0113217088 0.0425701874
C -0.0715926499 1.9929355693 -2.2153032733
C -0.0502589085 -2.8763924230 -0.6038172432
C -0.0357830409 0.9010986116 2.8043819804
H -0.0356494869 1.9542602131 2.5669540335
C -0.0334259356 3.8608401344 -0.0149781104
C -0.0468406709 -1.9636597204 -3.3435003841
C -0.0125031709 -1.9344174212 3.3442993869
H -0.0126789508 4.5479408731 0.8191404478
O 3.0139196483 4.2598740515 0.6637116342
O -3.0904947923 4.3198298082 0.7017125927
O 2.9368580327 1.7785842865 2.6154288771
O -2.9742944222 1.6264928742 2.1357901998
C 2.5516649517 3.5719879636 -0.2048778967
C 2.5425793626 -1.9679714406 -3.0248154209
C 2.5763872778 -1.6417986363 3.1679517358
C -2.6232616662 3.5337189884 -0.0711555802
C -2.6340963859 -1.8470423751 -3.0018547875
C -2.6063925596 -1.7016618587 3.1242263492
C 1.1023753939 2.5133338622 -1.7004552415
C 1.1132959990 -2.7066360132 -1.3314780477
C 1.1338835098 0.1880752170 2.9933953887
C -1.2199428683 2.4035179119 -1.5650465178
C -1.2064978272 -2.5340340592 -1.2784241990
C -1.1888587416 0.1414361781 2.8519866595
C 2.5005885461 2.1072352725 -2.0146712893
C 2.5157451653 -2.7625974729 -0.8359479215
C 2.5319561845 0.6558672673 2.7912179288
C -2.6014540066 1.9082314966 -1.7419300387
C -2.5813947592 -2.4299895185 -0.7461914391
C -2.5701236581 0.5496589706 2.5222879218
C 1.1265314576 3.3980377124 -0.6255695916
C 1.1220200612 -2.2448501241 -2.6450437528
C 1.1522137566 -1.1837319852 3.2303562011
C -1.2129272587 3.3286258828 -0.5149990842
C -1.2173817199 -2.1173777351 -2.6146809949
C -1.1896114847 -1.2280031298 3.1418754365
N 3.2951771853 2.7230324886 -1.0473311561
N 3.2989389087 -2.2532822799 -1.8706612702
N 3.3222010812 -0.4914372224 2.8418854304
N -3.3859420552 2.5881640108 -0.8161059594
N -3.3808342611 -1.9965234720 -1.7969833410
N -3.3644353391 -0.5794155494 2.6788595120
C 4.7079753946 2.4580214121 -0.8838078553
C 4.7101514635 -1.9670432257 -1.7371412750
C 4.7349572794 -0.4984507476 2.5294935984
C -4.8265630666 2.4723237457 -0.7602753387
C -4.8215169007 -1.8985921263 -1.7085611126
C -4.8066350422 -0.5532016167 2.5675385816
H 5.1863073140 3.3824312563 -0.5596037215
H 5.1789612305 -2.1369457527 -2.7064129946
H 5.2156692411 -1.2430091132 3.1640787435
H 5.1148781153 2.1986431188 -1.8578290472
H 5.1333553927 -2.6806350121 -1.0348336852
H 5.1397848076 0.4755220338 2.7922433410
H -5.1912275341 2.2843193856 -1.7682367363
H -5.1786438940 -2.6864919852 -1.0485458519
H -5.1590274052 0.4104883532 2.9308558027
C 5.0410585673 1.4247814719 0.1999334097
C 5.0304631132 -0.5110932421 -1.3783546608
C 5.0592735942 -0.9130124491 1.0891517493
C -5.3572016666 1.4479852462 0.2458008558
C -5.3524007976 -0.5217223371 -1.3076101535
C -5.3586253814 -0.8860800370 1.1801780843
H 6.1248665380 1.3941284249 0.3211275347
H 6.1119823960 -0.3765120917 -1.4234530983
H 6.1419493672 -0.9982175193 0.9885586210
H 4.5898247344 1.7508170071 1.1321207636
H 4.5624042434 0.1313927917 -2.1178977087
H 4.6070501767 -1.8830406770 0.9068966400
H -5.0322757074 1.7338451502 1.2391366362
H -5.0219548036 0.2021458234 -2.0425241909
H -5.0493998561 -1.8901505651 0.9164010893
H 3.5154942827 -0.0039320425 -0.0109829682
H -3.8773417564 0.0106798947 0.0503383885
H -0.0790023161 1.2646825220 -3.0124552610
H -0.0414069819 -3.1772060661 0.4330344809
H -0.0382812449 -1.6037058780 -4.3625602879
H 0.0026390978 -2.9988830092 3.5309681461
O 2.9927966817 -1.5704096330 -4.0641338477
O 3.0355785792 -2.7374952609 3.3384079062
O -3.1159208519 -1.5920620781 -4.0678186721
O -3.0799662051 -2.7562787839 3.4351240165
O 2.9020569676 1.4005196594 -2.9058648916
O 2.9280911234 -3.1532881349 0.2285590061
O -3.0116681247 1.0488459912 -2.4934441029
O -2.9774097563 -2.6216234211 0.3846357228
H -5.2224057945 3.4355794858 -0.4363581949
H -5.2236110430 -2.0890034101 -2.7041764548
H -5.1989414770 -1.3234107663 3.2325504269
H -6.4453567788 1.4288260296 0.2022179264
H -6.4407507888 -0.5466155495 -1.2726488446
H -6.4459342007 -0.8252759489 1.1967054517
$end
$rem
JOBTYPE OPT
METHOD B3LYP
BASIS def2-TZVPP
SOLVENT_METHOD SMD
DFT_D D3
SCF_ALGORITHM l_bfgs
MEM_TOTAL 40000
MEM_STATIC 10000
SCF_MAX_CYCLES 200
GEOM_OPT_MAX_CYCLES 400

$end

$smx
solvent water
$end

and my output:

…

A restricted SCF calculation will be
performed using Roothaan-Hall, L-BFGS

Cycle Energy DIIS error

1   -3438.6055624981      1.16e-02  Roothaan Step

L-BFGS subspace size: 20

Cycle Energy Error

1   -3421.1203574123      6.65e-03  LineSearch Step

gen_scfman_exception: I think this should be impossible

Q-Chem fatal error occurred in module libgscf/gen_scfman/gen_scfman_main.C, line 223:

Error in gen_scfman

Please submit a crash report at Q-Chem Crash Reporter

#2

Why did you choose to use SCF_ALGORITHM = L_BFGS? This appears to be something having to do with the bonded EDA (not described in the manual even though there is one sample job that uses it). However it’s not listed as one of the options for SCF_ALGORITHM. I suggest that you try again with the default value of SCF_ALGORITHM (which is DIIS, although you can get the default by simply not specifying this keyword).

Note also that your molecule is large, PCM cavity surface (needed for SMD) contains 6,380 points. You might want to consider reducing the number of discretization points, at least for the geometry optimization, e.g., using

$pcm
HPoints 50
HeavyPoints 86
$end

which reduces the number of discretization points to about 3,500. (The cost here is just one Ngrid x Ngrid matrix diagonalization per SCF, so may be small compared to the cost of an SCF with 2,400 basis functions.) The solvation energy probably isn’t converged with that grid but the geom optimization is likely less sensitive, in the same way that geometry is less sensitive to basis set than are relative energy differences, so you optimize in small basis set and compute single points in large basis set.

1 Like
#3

Thanks a lot, I have try different SCF_ALGORITHM(DIIS, GDM,…), but all not work. Using RCA_DIIS method the output:

Cycle Energy DIIS Error

1   -3061.1858590837      2.80E-04 Done RCA. Switching to DIIS
2   -3061.1958241153      2.75E-04
3   -3061.0010782636      6.72E-04
4   -3061.2393283036      3.77E-05
5   -3061.2401502403      1.72E-05
6   -3061.2403151799      5.28E-06
7   -3061.2403288376      2.32E-06
8   -3061.2403315618      8.50E-07
9   -3061.2403319733      2.52E-07

10 -3061.2403320107 9.40E-08
11 17051.3935271557 6.33E+01
12 17051.3954629967 6.33E+01
13 17051.3966549415 6.33E+01
14 17051.3966551182 6.33E+01
15 17051.3966525154 6.33E+01
16 17051.3968958394 6.33E+01
17 17051.3966481966 6.33E+01
18 17051.3999126736 6.33E+01
19 17051.3978613451 6.33E+01
20 17051.3979242859 6.33E+01
21 17051.3997111765 6.33E+01
22 17051.3997742053 6.33E+01
23 17051.3997294048 6.33E+01
24 17051.3965932877 6.33E+01
25 17051.3933388254 6.33E+01
26-89727738.4343683124 3.36E-02
27-89727739.6784250885 8.22E-03
28-89727739.6780712903 8.22E-03
29-89727739.6779905111 8.22E-03

The DIIS output is similar.
Using GDM method the output:

Cycle Energy RMS Gradient

1   -3078.2327345373      1.30E-02 Done DIIS. Switching to GDM
2   16791.0462822197      1.09E+01 Descent step

GDM::BFGS: rebuilding Hinv to be positive definite
3 16570.3691913399 1.01E+01 Descent step
GDM::BFGS: rebuilding Hinv to be positive definite
4 16349.7262372271 9.27E+00 Descent step
5 16129.6051510085 7.11E+05 Line search: understep
6 14156.7690709846 7.15E+05 Line search: understep
7 1907.5843807434 1.35E+01 Descent step
GDM::BFGS: rebuilding Hinv to be positive definite
8 1736.2225449628 1.52E+01 Descent step
9 1563.5010078759 2.26E+05 Line search: understep
10 197.4103036297 2.29E+05 Line search: understep
11 -9013.2908704387 1.24E+01 Descent step
12 -9164.1073175218 1.04E+05 Line search: understep
13 -10517.2180498430 1.90E+01 Descent step
14 -10671.8900608645 1.27E+05 Line search: understep
15 -11826.9007521372 1.19E+01 Descent step
GDM::BFGS: rebuilding Hinv to be positive definite

Thanks for your advices, I may try to reduces the number of discretization points. And I don’t konw whether it will work to set BASIS2 a small basis?

#4

Okay, so the problem is that the SCF fails to converge, yes? That is usually a sign of a poor guess, and if I had to guess in your case I would say it might be due to the use of Cl (neutral) rather than Cl- in the superposition of atomic densities, which is the default SCF guess. In addition, your basis does not contain any diffuse functions so may have a problem knowing where to put the charge for Cl-.

Two general strategies in the case of SCF convergence failure are:
(1) BASIS2 = STO-3G. In my experience, there’s little reason to use any larger basis for BASIS2, except in rare circumstances.

(2) SCF_GUESS = AUTOSAD, which will generate an “in situ” SAD guess using your selected functional and basis set, rather than taking the SAD guess from a library of atomic HF calculations (which is the default behavior for SCF_GUESS=SAD.)

And a bonus strategy that might help in your specific case:
(3) You could use SCF_GUESS = FRAGMO to use a superposition of fragment densities, with Cl- as one fragment and the rest of the system as the other. The input for this would look something like

$molecule
1 1

1 1
Cl 1.5746745308 0.0070770309 -0.0289671039

0 1
N 4.5855687067 -0.0029326501 -0.0235189175
N -4.8998317514 0.0113217088 0.0425701874
C -0.0715926499 1.9929355693 -2.2153032733
…

One last comment is regarding whether you have thought carefully about your choice of basis set. As given, this is a fairly sizable calculation at 2,412 basis functions yet the B3LYP functional is not highly sensitive to basis set, especially for geometry optimizations (as suggested above). Moreover, other that the Cl- ion your system is quite rigid. I suggest that BASIS = 6-31+G* is probably just fine for this optimization. That reduces the calculation to 1,245 basis functions (and Pople basis sets are very fast in Q-Chem). You could also consider a mixed basis set where diffuse functions are used only for the Cl atom.

#5

Apologies, but I saw your Cl atom and as a result I misread your charge as -1 (meaning Cl-), but in fact you’ve set it to +1. Is that really what you mean, to make the system a cation?

In any case, all of my comments above remain valid except possibly that the way the FRAGMO guess is set up might need to change.

#6

Never mind, thank you for your advices.
I set the system a cation, because I think this structure is reasonable, and want to optimize the structure.
I will try your strategies later.

#7

That’s fine, should be possible but that may explain your SCF convergence failure insofar as it’s somewhat unusual, but one of my 3 strategies is likely to work. For the cation the diffuse functions are not as essential, but I was getting what looked like good convergence with 6-31+G*. (I didn’t run it to the end and I did set THRESH=14, as Q-Chem warns in that case because it’s a large system and the diffuse functions lead to a lot of numerical linear dependencies.)

#8

Still not work…
This is my input:

--------------------------------------------------------------
User input:
--------------------------------------------------------------
$molecule
0 1
O         3.0784040929   -1.2261703309   -4.2294494598
O         2.9288457148   -2.5781862292    0.1366110357
O         3.0753693384   -3.0512265548    3.1766770138
O         3.0788537158    4.2675857797    1.0553702256
O         2.9338834161    1.4077900694    2.1721170589
O         2.9339974473    1.1830052291   -2.3171265142
N         3.3685672106   -0.8378494310    2.5310545374
N         3.3676938791   -1.7730871137   -1.9891037697
N         4.9320230324   -0.0020542263    0.0048036377
N         3.3705550233    2.6058397803   -0.5428672338
C         2.6175259829   -1.9893884037    2.8543077270
C        -0.0011360213   -1.4119754432   -3.4594977130
C         2.5633316937    1.8973969104   -1.4142859784
C        -0.0010926975   -2.2883450528    2.9474205776
C         0.0017267103    1.7350496122   -1.6061436542
C        -1.1648096933   -2.0179973381   -1.4099648297
C         1.1760447737   -1.5674504805    2.7646145762
C         1.1654199933    2.2303356891   -1.0519301977
C         2.5589944562   -2.1640158241   -0.9378969825
C        -1.1774240657   -1.5662650399    2.7640876434
C        -1.1637134699   -0.2089696842    2.4508752221
C         0.0024961725    3.6955209411    0.5048129674
C         1.1758719229   -1.6134887049   -2.7425906393
C         2.6207149092    3.4607620363    0.2936688164
C         0.0004301734    0.5181456126    2.2995863622
C         5.3874657755    1.3603584824    0.3129052203
C         2.6186172981   -1.4784744125   -3.1497514115
C         4.8196627155    2.4958091016   -0.5506877283
C         2.5620678738    0.2698510053    2.3457363724
C         1.1616967446   -2.0187620773   -1.4097052731
C         4.8166778951   -1.7279581928   -1.8859965543
C         5.3901825233   -0.9509508822    1.0285844305
C         5.3861282783   -0.4144803262   -1.3299692890
C         4.8178975309   -0.7781631300    2.4437524326
C         1.1637726385   -0.2101383449    2.4514522722
C        -0.0017110600   -2.2501106199   -0.7031075406
C         1.1790336875    3.1771052339   -0.0299326978
H        -0.0009094607   -1.1068612623   -4.4964405594
H        -0.0016942860   -3.3391230083    3.2006348581
H         0.0014131097    1.0121566039   -2.4062385753
H         0.0028025552    4.4360400762    1.2922665660
H         0.0010458762    1.5731670329    2.0783201969
H         6.4869136659    1.4338977471    0.2350257209
H         5.1183354958    1.5798404015    1.3405909540
H         5.2149989217    3.4329685384   -0.1612224113
H         5.1297927310    2.4083938162   -1.5908248742
H         5.2141305397   -1.8560739186   -2.8918252134
H         5.1237919017   -2.5876675296   -1.2926408778
H         6.4890957119   -0.9148310895    1.1340856277
H         5.1281223809   -1.9506826837    0.6996320267
H         6.4852454834   -0.5206024360   -1.3553450499
H         5.1183194093    0.3677429338   -2.0320624452
H         5.2090665201   -1.5889900266    3.0566201222
H         5.1300419572    0.1627103852    2.8932064793
H        -0.0019369788   -2.5843502888    0.3219368575
C        -1.1784412073   -1.6128094940   -2.7428842605
C        -1.1615638827    2.2311208781   -1.0517817432
C        -2.5623079445   -2.1621553788   -0.9383973991
C        -2.6193459863   -1.9868118476    2.8531907996
C        -2.5614989600    0.2723586187    2.3444137910
C        -1.1744502476    3.1777986960   -0.0296994811
C        -2.6210120988   -1.4769985053   -3.1503907449
C        -2.5597221088    1.8993142446   -1.4141866462
N        -3.3705041559   -1.7706629360   -1.9897741516
O        -2.9326875323   -2.5755108211    0.1362443926
O        -3.0783481782   -3.0480777996    3.1757967914
N        -3.3691529417   -0.8344615573    2.5299903461
O        -2.9321444032    1.4106694311    2.1706910558
C        -2.6159076843    3.4622593871    0.2942003567
O        -3.0804108619   -1.2243802491   -4.2301811072
O        -2.9308841255    1.1858848503   -2.3175751195
N        -3.3664452566    2.6084808566   -0.5428858420
C        -4.8194317290   -1.7238753073   -1.8867353527
C        -4.8184310995   -0.7732161706    2.4428930454
O        -3.0734146822    4.2695806012    1.0557537381
C        -4.8157052566    2.5002032968   -0.5512810146
H        -5.2170504294   -1.8515847220   -2.8925502914
H        -5.1275204552   -2.5831762095   -1.2932928907
C        -5.3872350500   -0.4096986210   -1.3307249252
C        -5.3910991541   -0.9457777479    1.0278661367
H        -5.1294603516    0.1681347026    2.8921140794
H        -5.2104239493   -1.5834312730    3.0560432243
C        -5.3852929858    1.3654304074    0.3120366665
H        -5.1256018173    2.4133421699   -1.5915338466
H        -5.2099796228    3.4378130440   -0.1618253477
H        -5.1188217554    0.3721080911   -2.0330357673
H        -6.4864830090   -0.5145930751   -1.3556454514
N        -4.9320021702    0.0023353125    0.0037941694
H        -5.1302281064   -1.9459019787    0.6991519267
H        -6.4899530479   -0.9083355529    1.1335329012
H        -5.1157858635    1.5843341904    1.3397520319
H        -6.4846283725    1.4406632454    0.2342170789
$end
$rem
JOBTYPE OPT
METHOD B3LYP
BASIS STO-3G
SOLVENT_METHOD PCM
MEM_TOTAL 40000
MEM_STATIC 10000
SCF_MAX_CYCLES 200
GEOM_OPT_MAX_CYCLES 400
DFT_D D3
SCF_GUESS AUTOSAD
$end
$pcm
Theory CPCM
Method SWIG
Solver Inversion
HeavyPoints 50
HPoints 86
$end
$solvent
Dielectric 11
$end

and my output may be this:

 ---------------------------------------
  Cycle       Energy         DIIS error
 ---------------------------------------
    1   23561.3073182285      1.09e+01  
    2   -2655.6759344286      3.17e-01  
    3   -3892.3745145957      3.96e-01  
    4   -4297.8528289031      1.54e-01  
    5   -4420.0545884845      7.39e-02  
    6   -4068.4232053575      3.93e-02  
    7   -4146.0471280038      4.98e-02  
    8   -4416.6884426266      1.50e-02  
    9   -4407.6611426318      2.03e-02  
   10   -4383.5073482271      7.02e-03  
   11   -4490.0789693446      9.95e-03  
   12   -4403.0516669613      5.60e-03  
   13   -4472.1965409302      6.82e-03  
   14   -4467.3260160408      5.52e-03  
   15   -4042.5159271330      1.15e-02  
   16   -5479.5254427111      2.12e-01  
   17   -5480.2518979779      2.12e-01  
   18   -5480.2308823542      2.12e-01  
   19   -5480.2533364137      2.12e-01  
   20   -5480.2844213875      2.12e-01  
   21   -5480.5702674468      2.12e-01  
   22   -5480.9335248639      2.13e-01  
   23   -5481.9938900247      2.12e-01  
   24   -5481.5801906991      2.12e-01  
   25   -5484.1008214850      2.12e-01  
   26   -5413.2461619114      2.11e-01  
   27   -5441.7338788932      2.08e-01  
   28   -5461.7404558110      2.11e-01  
   29   -5468.0856756965      2.07e-01  
   30   -5470.0214174112      2.29e-01  
   31   -5657.9660707319      3.93e-02  
   32   -5884.9464645655      5.03e-02  
   33   -6059.2029708410      1.44e-02  
   34   -5989.6153761497      4.57e-03  
   35   -6045.4516727735      5.43e-03  
   36   -6022.4670245501      3.53e-03  
   37   -6023.9383790059      2.39e-03  
   38   -6035.7759739386      4.09e-03  
   39   -6005.3344294152      2.94e-03  
   40   -6028.0951289908      4.45e-03  
   41   -6023.3494394924      2.18e-03  
   42   -6022.2863394015      2.44e-03  
   43   -6032.2167511230      3.51e-03  
   44   -6031.7274566998      5.67e-03  
   45   -6024.7005014584      2.79e-03  
   46   -6024.3039271940      2.87e-03  
   47   -6036.2142306388      2.54e-03  
   48   -6036.2144971528      2.50e-03  
   49   -5991.6290219955      1.97e-03  
   50   -5991.5557789211      2.03e-03  
   51   -6022.9693232523      2.23e-03  
   52   -6023.1320622906      2.18e-03  
   53   -6023.1083136852      2.18e-03  
   54   -6022.9534555884      2.12e-03  
   55   -6005.2801896777      2.53e-03  
   56   -5991.3718315631      2.19e-03  
   57   -6005.0573262729      2.83e-03  
   58   -6030.0739856446      3.19e-03  
   59   -5991.7825743464      2.01e-03  
   60   -6021.1531255577      2.60e-03  
   61   -6004.9475921686      3.60e-03  
   62    3792.7154897172      1.98e-01  
   63    3793.2389225953      1.97e-01  
   64    3792.1838526924      2.00e-01  
   65    3785.8622387749      2.14e-01  
   66    3783.9787918395      2.14e-01  
   67    3780.7576430616      2.16e-01  
   68    3787.9018426442      2.10e-01  
   69    3705.1683641772      1.82e-01  
   70    3671.7119076877      1.67e-01  
   71    3633.9521550419      1.78e-01  
   72    3635.0589268623      1.82e-01  
   73    3634.6436347766      1.80e-01  
   74    3634.5602129370      1.80e-01  
   75    3633.9388886967      1.77e-01  
   76    3636.1330513271      1.74e-01

and also may be this:

 ---------------------------------------
  Cycle       Energy         DIIS error
 ---------------------------------------
    1   12686.2179030483      4.68e+00  
    2               -nan           nan  
solve(c,A,vect) failed Check on this
The B Matrix
           0.  -1.0000e+00  -1.0000e+00  -1.0000e+00
  -1.0000e+00   2.5266e+06          nan          nan
  -1.0000e+00          nan          nan          nan
  -1.0000e+00          nan          nan          nan
    3                nan     1.80e+308  
    4                nan           nan  
    5                nan           nan  
solve(c,A,vect) failed Check on this
The B Matrix
       0.  -1.0000  -1.0000  -1.0000
  -1.0000      nan      nan      nan
  -1.0000      nan      nan      nan
  -1.0000      nan      nan      nan
    6                nan     1.80e+308  
    7                nan           nan  
    8                nan           nan  
solve(c,A,vect) failed Check on this
The B Matrix
       0.  -1.0000  -1.0000  -1.0000
  -1.0000      nan      nan      nan
  -1.0000      nan      nan      nan
  -1.0000      nan      nan      nan
    9                nan     1.80e+308  
   10                nan           nan  
   11                nan           nan  
solve(c,A,vect) failed Check on this
The B Matrix
       0.  -1.0000  -1.0000  -1.0000
  -1.0000      nan      nan      nan
  -1.0000      nan      nan      nan
  -1.0000      nan      nan      nan
   12                nan     1.80e+308  
   13                nan           nan  
   14                nan           nan

I really feel frustrated about this.

#9

Look at your SCF energy on the first step: this is clearly the result of a bad guess, and the SCF simply doesn’t recover from this. Please post the complete input file that led to this output and I will do some tests, because I did try a couple of the things that I suggested above and while I didn’t run them through to completion, they looked well-behaved and on their way to convergence.

#10

I have tried " SCF_GUESS = AUTOSAD", but still not work.
Sometimes the SCF not converge in the first iteration, sometimes may in next iteration.
Although the SCF energy may be right at the first step, but still can change to the two situation I posted above gradually.

This my input file:

$molecule
0 1
  O    3.1010993   -1.8742236   -3.9562137
  O    2.9338588   -2.7721020    0.5238388
  O    3.0375184   -2.5277588    3.6387110
  O    3.0770321    4.3865157    0.4132705
  O    2.9269569    1.7939303    2.1557348
  O    2.9703071    0.9708905   -2.6217350
  N    3.3373323   -0.4157666    2.7159377
  N    3.3708866   -2.1467516   -1.6640328
  N    4.8843862   -0.0095137    0.0409503
  N    3.3751159    2.5403019   -0.9671556
  C    2.5837277   -1.5041009    3.2046591
  C    0.0214542   -2.0458150   -3.2048201
  C    2.5818403    1.7888638   -1.8204146
  C   -0.0326173   -1.7700371    3.3564388
  C    0.0198719    1.6769110   -2.0929821
  C   -1.1618924   -2.4717324   -1.1219308
  C    1.1459558   -1.0720231    3.1106088
  C    1.1807699    2.2055596   -1.5618438
  C    2.5619903   -2.4956770   -0.5932486
  C   -1.2052864   -1.0689175    3.0916864
  C   -1.1883360    0.2536122    2.6559431
  C    0.0078380    3.8017117   -0.1486000
  C    1.1905780   -2.1799905   -2.4619065
  C    2.6229749    3.5109923   -0.2719133
  C   -0.0219213    0.9661283    2.4544296
  C    5.3436858    1.3841202    0.1384500
  C    2.6328918   -2.0261846   -2.8607685
  C    4.8212357    2.3891740   -0.9022397
  C    2.5407484    0.6903584    2.4645703
  C    1.1667991   -2.4743361   -1.1010652
  C    4.8178614   -2.0284950   -1.5543330
  C    5.3337919   -0.7876714    1.2050187
  C    5.3462315   -0.6258399   -1.2106871
  C    4.7862926   -0.3858882    2.5845366
  C    1.1392241    0.2503211    2.6740067
  C   -0.0042080   -2.6584694   -0.3908462
  C    1.1867118    3.2341435   -0.6230895
  H    0.0314829   -1.8245477   -4.2630651
  H   -0.0367136   -2.7945075    3.7020698
  H    0.0244014    0.8911403   -2.8339483
  H    0.0032595    4.6046737    0.5755529
  H   -0.0177541    1.9947796    2.1262090
  H    6.4454944    1.4352905    0.0747841
  H    5.0568148    1.7596460    1.1154055
  H    5.2328244    3.3628031   -0.6378849
  H    5.1608425    2.1444740   -1.9067522
  H    5.2348224   -2.2977180   -2.5242953
  H    5.1438620   -2.7738163   -0.8311158
  H    6.4345595   -0.7536929    1.2956278
  H    5.0516810   -1.8231675    1.0425905
  H    6.4482434   -0.7088182   -1.2203775
  H    5.0622796    0.0280742   -2.0295670
  H    5.1878634   -1.0949031    3.3078575
  H    5.1113446    0.6112814    2.8763477
  H   -0.0141910   -2.9033842    0.6606267
  C   -1.1610978   -2.1767958   -2.4829219
  C   -1.1476358    2.2095720   -1.5807748
  C   -2.5659794   -2.4904332   -0.6388180
  C   -2.6454463   -1.4979419    3.1614107
  C   -2.5853185    0.6973648    2.4251347
  C   -1.1651631    3.2380132   -0.6420225
  C   -2.5955535   -2.0177797   -2.9067015
  C   -2.5457792    1.7979456   -1.8622979
  N   -3.3548108   -2.1377739   -1.7233562
  O   -2.9585204   -2.7678696    0.4709386
  O   -3.1085266   -2.5212286    3.5864283
  N   -3.3883801   -0.4074623    2.6612434
  O   -2.9640698    1.8023319    2.1121098
  C   -2.6058976    3.5192266   -0.3134968
  O   -3.0439377   -1.8622329   -4.0099270
  O   -2.9239911    0.9829819   -2.6715370
  N   -3.3500383    2.5509789   -1.0207281
  C   -4.8028211   -2.0112531   -1.6387162
  C   -4.8347097   -0.3743078    2.5030487
  O   -3.0679495    4.3961469    0.3645234
  C   -4.7976807    2.4057142   -0.9800048
  H   -5.2040718   -2.2764882   -2.6163940
  H   -5.1459490   -2.7557864   -0.9226792
  C   -5.3284968   -0.6058748   -1.3022840
  C   -5.3575620   -0.7726919    1.1130178
  H   -5.1630417    0.6230997    2.7902168
  H   -5.2513053   -1.0835830    3.2175301
  C   -5.3408696    1.4013745    0.0506710
  H   -5.1216368    2.1638552   -1.9903535
  H   -5.2094829    3.3806520   -0.7208533
  H   -5.0257228    0.0474773   -2.1148149
  H   -6.4306953   -0.6819185   -1.3320285
  N   -4.8855877    0.0061126   -0.0416612
  H   -5.0749431   -1.8085165    0.9540747
  H   -6.4597028   -0.7358904    1.1839542
  H   -5.0674981    1.7742541    1.0324776
  H   -6.4412956    1.4569956   -0.0304705
$end
$rem
JOBTYPE OPT
METHOD B3LYP
BASIS STO-3G
MEM_STATIC 10000
MEM_TOTAL 40000
DFT_D D3
SCF_MAX_CYCLES 200
GEOM_OPT_MAX_CYCLES 400
SYMMETRY FALSE
SYM_IGNORE TRUE
SOLVENT_METHOD PCM
$end
$pcm
Theory CPCM
Method SWIG
Solver Inversion
HeavyPoints 50
HPoints 86
$end
$solvent
Dielectric 4.6
$end

This is the second iteration of SCF calcution in my output file:


 A restricted SCF calculation will be
 performed using Roothaan-Hall, DIIS
 SCF converges when DIIS error is below 1.0e-08
 ---------------------------------------
  Cycle       Energy         DIIS error
 ---------------------------------------
    1   -2921.3671321798      5.97e-04  Roothaan Step
 ---------------------------------------
  Cycle       Energy         DIIS error
 ---------------------------------------
    1   -2921.3912452980      4.84e-04  
    2   -2921.3820070774      9.16e-04  
    3   -2921.3964937101      1.03e-04  
    4   -2921.3966476320      4.93e-05  
    5   -2921.3966862354      1.40e-05  
    6   -1818.5623917072      6.68e-01  
    7   -1818.5711926781      6.68e-01  
    8   -1818.5711651703      6.68e-01  
    9   -1818.5709046176      6.68e-01  
   10   -1818.5705672915      6.68e-01  
   11   -1818.5705542099      6.68e-01  
   12   -1818.5705486272      6.68e-01  
   13   -1818.5705423656      6.68e-01  
   14   -1818.5705375221      6.68e-01  
   15   -1818.5705507027      6.68e-01  
   16   -1818.5704849982      6.68e-01  
   17   -1818.5357092036      6.68e-01  
   18   -1818.5356122914      6.68e-01  
   19   -1818.5634133509      6.68e-01  
   20   -1818.5640706482      6.68e-01  
   21   -2996.9726025395      4.79e-02  
   22   -3750.2040044034      6.43e-02  
   23   -3875.7947077231      1.07e-02  
   24   -3847.8270540597      7.21e-03  
   25   -3814.2149093005      5.74e-03  
   26   -3876.9361059375      6.51e-03  
   27   -3839.9780272810      4.17e-03  
   28   -3873.7526377314      5.66e-03  
   29   -3844.9574156648      3.63e-03  
   30   -3844.7294866317      4.82e-03  
   31   -3876.9519174025      3.87e-03  
   32   -3876.8707796161      4.27e-03  
   33   -3843.9111651216      4.29e-03  
   34   -3809.2146697831      3.51e-03  
   35   -3815.3616027987      4.28e-03  
   36   -3844.6541424310      3.99e-03  
   37   -3841.4685579817      3.12e-03  
   38   -3853.3414133315      5.86e-03  
   39   -3854.7985246341      1.04e-02  
   40   -3853.6885047667      4.64e-03  
   41   -3835.2549958884      4.47e-03  
   42   -3835.9595321918      4.68e-03  
   43   -3834.1717386881      4.94e-03  
   44   -3834.2237380768      4.87e-03  
   45   -3831.9533936008      4.05e-03  
   46   -3861.1986574832      5.87e-03  
   47   -3878.0646568023      4.71e-03  
   48   -3877.9978834354      4.65e-03  
   49   -3878.3519632037      5.11e-03  
   50   -3872.0560674893      6.53e-03  
   51   -3871.7343703559      5.62e-03  
   52   -3187.7179595546      2.69e-01  
   53   -3206.9737039458      2.72e-01  
   54   -3206.6961453014      2.72e-01  
   55   -3207.2388501284      2.73e-01  
   56   -3127.4349932304      2.59e-01  
   57   -3127.7586040676      2.59e-01  
   58   -3241.0751124138      2.61e-01  
   59   -3128.3426462101      2.60e-01  
   60   -3239.1527322269      2.65e-01  
   61   -3179.3038403142      2.93e-01  
   62   -3097.7046339065      2.74e-01  
   63   -3190.8419253934      2.82e-01  
   64   -3100.5099967546      2.64e-01  
   65   -3176.9009792460      2.97e-01  
   66   -3163.9552956084      2.88e-01  
   67   -8019.2931796766      2.19e-02  
   68   -8254.6797990780      1.17e-02  
   69   -8183.2691307395      4.57e-03  
   70   -8243.7196714302      5.34e-03  
   71   -8198.4698453943      4.29e-03  
   72   -8243.5320795334      5.06e-03  
   73   -8243.5190652841      4.94e-03  
   74   -8243.5107728680      4.87e-03  
   75   -8187.9543235586      2.17e-03  
   76   -8213.0957906475      4.91e-03  
   77   -2141.2633529856      2.94e-01  
   78   -2116.0348334061      2.65e-01  
   79   -2146.6363105241      2.82e-01  
   80   -2147.2102628429      2.80e-01  
   81   -2146.9693088202      2.81e-01  
   82   -2114.2290384486      2.68e-01  
   83   -2113.9182695899      2.68e-01  
   84   -2114.2180298851      2.68e-01  
   85   -2114.1995527636      2.68e-01  
   86   -2149.9395221665      3.11e-01  
   87   -2137.5312723969      2.98e-01  
   88   -2130.6712495740      2.85e-01  
   89   -2182.2271770659      2.53e-01  
   90   -2182.4516387840      2.53e-01  
   91   -2219.2211476695      2.65e-01  
   92   -4940.4152580200      6.48e-02  
   93   -5135.2947459534      3.25e-02  
   94   -5331.1520374532      1.38e-02  
   95   -5306.2698816302      1.07e-02  
   96   -5305.1641910639      4.76e-03  
   97   -5308.1625236079      9.98e-03  
   98   -5326.3643212017      4.01e-03  
   99   -5308.4960932614      9.95e-03  
  100   -5312.3474772859      9.59e-03  
  101   -5295.8787107587      1.02e-02  
  102   -5293.9972581038      9.80e-03  
  103   -5304.3718802683      9.80e-03  
  104   -5311.9967289206      9.75e-03  
  105   -5311.6935243432      9.72e-03  
  106   -5308.8221532237      1.00e-02  
  107   -5308.7733714265      1.03e-02  
  108   -5308.8522539868      1.01e-02  
  109   -5311.6705454322      9.74e-03  
  110   -5308.9135561003      9.99e-03  
  111   -5311.9193370508      9.70e-03  
  112   -5311.8858055773      9.69e-03  
  113   -5291.6980418846      9.86e-03  
  114   -5329.1088688884      6.11e-03  
  115   -5321.6837454718      4.09e-03  
  116   -5308.9657004466      9.77e-03  
  117   -5311.6215384916      9.76e-03  
  118   -5293.4754306083      9.81e-03  
  119   -5311.7353307269      9.81e-03  
  120   -5311.6729973547      9.85e-03  
  121   -5293.3114570127      1.00e-02  
  122   -5293.2430697858      9.92e-03  
  123   -5293.3828295656      9.88e-03  
  124   -5292.6398777658      1.07e-02  
  125   -5308.8564751119      9.83e-03  
  126   -5308.8253427666      9.84e-03  
  127   -5308.7976537187      9.84e-03  
  128   -5308.9729943874      9.79e-03  
  129   -5309.0211151421      9.82e-03  
  130   -5335.8285799924      1.18e-02  
  131   -5319.5183425383      6.39e-03  
  132   -5327.6813201348      4.72e-03  
  133   -5302.3592659316      1.03e-02  
  134   -5301.6771001665      1.01e-02  
  135   -5307.5990875719      1.03e-02  
  136   -5307.3525631522      1.02e-02  
  137   -5307.1387933267      1.02e-02  
  138   -5307.4047540425      1.01e-02  
  139   -5290.6283563406      9.81e-03  
  140   -5311.8182675916      9.78e-03  
  141   -5308.0859207152      9.92e-03  
  142   -5311.8380601972      9.81e-03  
  143   -5308.1182647538      9.84e-03  
  144   -5311.9933960189      9.59e-03  
  145   -5304.2334971340      1.14e-02  
  146   -5286.0723670135      9.97e-03  
  147   -5304.2402989895      9.91e-03  
  148   -5299.8941490224      5.27e-03  
  149   -5304.8205711996      9.46e-03  
  150   -5286.7310626205      9.54e-03  
  151   -5292.6078679871      9.09e-03  
  152   -5286.8897979379      9.48e-03  
  153   -5287.1040836046      9.43e-03  
  154   -5286.8713160831      9.46e-03  
  155   -5283.6510333650      9.71e-03  
  156   -5286.9072679340      9.49e-03  
  157   -1182.1879235233      2.18e-01  
  158   -1190.8342118774      2.15e-01  
  159   -1184.4584111257      2.13e-01  
  160   -1215.3498431977      2.27e-01  
  161   -1214.5890268930      2.26e-01  
  162   -1220.7051218902      2.28e-01  
  163   -1219.8198715649      2.28e-01  
  164   -1279.7247080301      2.07e-01  
  165   -1280.3267698985      2.09e-01  
  166   -1281.4574980306      1.99e-01  
  167   -1280.7403837053      2.06e-01  
  168   -1280.5991588417      2.07e-01  
  169   -1280.8015386364      2.06e-01  
  170   -1280.9932612483      2.06e-01  
  171   -1281.1928546478      2.04e-01  
  172   -3453.6543756687      1.68e-02  
  173   -3798.6276900473      1.57e-02  
  174   -3738.4099641831      6.26e-03  
  175   -3731.1577107587      4.16e-03  
  176   -3749.1618304756      5.54e-03  
  177   -3733.1807877464      3.17e-03  
  178   -3738.3570993589      4.37e-03  
  179   -3733.0341078237      2.28e-03  
  180   -3730.4610907661      3.95e-03  
  181   -3732.9045968803      2.84e-03  
  182    -892.7169138586      2.09e-01  
  183    -893.1881927239      2.09e-01  
  184    -893.1339327354      2.09e-01  
  185    -893.1524278476      2.09e-01  
  186    -893.1191993286      2.09e-01  
  187    -893.1148298802      2.09e-01  
  188    -892.9896847299      2.08e-01  
  189    -893.1070425905      2.08e-01  
  190    -893.2306707927      2.09e-01  
  191    -910.4540295446      1.99e-01  
  192    -901.3810819693      2.37e-01  
  193    -897.4631504043      2.30e-01  
  194    -898.0090366294      2.31e-01  
  195    -904.7254194670      2.50e-01  
  196    -926.3459532453      2.54e-01  
  197   -2082.5551195314      5.31e-02  
  198   -2175.4469588797      3.12e-02  
  199   -2367.3717642810      1.23e-02  
  200   -2341.0798041475      7.96e-03  
gen_scfman_exception: SCF failed to converge

 Q-Chem fatal error occurred in module libgscf/gen_scfman/gen_scfman_main.C, line 223:

 Error in gen_scfman


 Please submit a crash report at q-chem.com/reporter
#11

This is very strange because when I run your latest input file, with the current version of Q-Chem (pre-release copy of 5.4.2), on the 2nd cycle I get the following:

 A restricted SCF calculation will be
 performed using Roothaan-Hall, DIIS
 SCF converges when DIIS error is below 1.0e-08
 ---------------------------------------
  Cycle       Energy         DIIS error
 ---------------------------------------
    1   -2921.3674762229      5.84e-04  Roothaan Step
 ---------------------------------------
  Cycle       Energy         DIIS error
 ---------------------------------------
    1   -2921.3896862299      4.09e-04
    2   -2921.3831508938      7.74e-04
    3   -2921.3934220184      9.40e-05
    4   -2921.3935506838      4.34e-05
    5   -2921.3935805082      1.19e-05
    6   -2921.3935826682      4.28e-06
    7   -2921.3935829648      1.29e-06
    8   -2921.3935829944      1.90e-07
    9   -2921.3935830004      1.04e-07
   10   -2921.3935830024      5.83e-08
   11   -2921.3935830038      1.16e-08
   12   -2921.3935830010      5.75e-09  Convergence criterion met
 ---------------------------------------

It is almost the same as your energy at first (although not exactly the same), but notably I don’t see the bad behavior that you see starting in the 6th cycle. I did not modify your input file at all, and my optimization converges in 23 steps without a problem. What version of Q-Chem are you running?

If it is a version before v.5, would you try turning off DFT_D and see if that helps? There were some known problems with Grimme D3 that have now been fixed. (And it’s unlikely to make a difference to this structure, which is very rigid.)

#12

My version is 5.3.2
I try the same input again, and it fails at 10th step, sometimes the inital guess may be great like(This is the second step, and the result is approximately the same as your’s) :

  A restricted SCF calculation will be
 performed using Roothaan-Hall, DIIS
 SCF converges when DIIS error is below 1.0e-08
 ---------------------------------------
  Cycle       Energy         DIIS error
 ---------------------------------------
    1   -2921.3674769692      5.84e-04  Roothaan Step
 ---------------------------------------
  Cycle       Energy         DIIS error
 ---------------------------------------
    1   -2921.3896865390      4.09e-04  
    2   -2921.3831512194      7.74e-04  
    3   -2921.3934223144      9.40e-05  
    4   -2921.3935509827      4.34e-05  
    5   -2921.3935808069      1.19e-05  
    6   -2921.3935829649      4.28e-06  
    7   -2921.3935832613      1.29e-06  
    8   -2921.3935832937      1.90e-07  
    9   -2921.3935832944      1.04e-07  
   10   -2921.3935832945      5.83e-08  
   11   -2921.3935832946      1.16e-08  
   12   -2921.3935832946      5.75e-09  Convergence criterion met
 ---------------------------------------

Sometimes the inital guess is bad bu still converge(This is the 4th step):

 A restricted SCF calculation will be
 performed using Roothaan-Hall, DIIS
 SCF converges when DIIS error is below 1.0e-08
 ---------------------------------------
  Cycle       Energy         DIIS error
 ---------------------------------------
    1    9892.1567126910      8.00e+00  Roothaan Step
 ---------------------------------------
  Cycle       Energy         DIIS error
 ---------------------------------------
    1   -2740.1236730011      1.53e-01  
    2   -2803.3739353049      4.22e-02  
    3   -2395.8903479740      5.89e-02  
    4   -2673.7321953588      5.34e-02  
    5   -2911.1592609490      1.85e-02  
    6   -2883.7772020443      2.64e-02  
    7   -2911.5627651297      1.41e-02  
    8   -2918.9877493695      7.86e-03  
    9   -2919.3356695660      6.96e-03  
   10   -2920.8516461686      4.61e-03  
   11   -2921.3602166905      1.72e-03  
   12   -2921.3843156477      1.32e-03  
   13   -2921.4091399858      8.69e-04  
   14   -2921.4162662287      6.61e-04  
   15   -2921.4213897463      4.25e-04  
   16   -2921.4244745762      2.49e-04  
   17   -2921.4255258392      1.61e-04  
   18   -2921.4260197820      7.50e-05  
   19   -2921.4261192769      2.85e-05  
   20   -2921.4261328162      1.18e-05  
   21   -2921.4261346407      6.95e-06  
   22   -2921.4261355284      2.55e-06  
   23   -2921.4261356395      1.25e-06  
   24   -2921.4261356693      6.45e-07  
   25   -2921.4261356763      1.76e-07  
   26   -2921.4261356768      8.39e-08  
   27   -2921.4261356769      3.32e-08  
   28   -2921.4261356769      9.58e-09  Convergence criterion met
 ---------------------------------------

And the last step not converge with a bad inital guess(the 10th step):

 A restricted SCF calculation will be
 performed using Roothaan-Hall, DIIS
 SCF converges when DIIS error is below 1.0e-08
 ---------------------------------------
  Cycle       Energy         DIIS error
 ---------------------------------------
    1   13109.2362546465      9.51e+00  Roothaan Step
 ---------------------------------------
  Cycle       Energy         DIIS error
 ---------------------------------------
    1   -2535.0512650874      3.04e-01  
    2   -2781.6933873836      4.68e-02  
    3   -2503.2676556427      5.20e-02  
    4   -2747.6567679220      6.05e-02  
    5   -2857.2241488242      3.93e-02  
    6   -2845.4059463735      4.10e-02  
    7   -2873.0456511175      3.35e-02  
    8   -2069.2289479849      3.56e-01  
    9   -2023.4598865329      3.40e-01  
   10   -2018.6223093795      3.39e-01  
   11   -2017.2377981616      3.38e-01  
   12   -2013.4850537009      3.36e-01  
   13   -2010.0251709510      3.35e-01  
   14   -2009.3181546857      3.34e-01  
   15   -2012.0415162839      3.36e-01  
   16   -2012.0143847332      3.36e-01  
   17   -2013.3028711112      3.37e-01  
   18   -1880.4079550436      3.23e-01  
   19   -2148.3821513591      2.10e-01  
   20   -2753.6482152836      1.94e-01  
   21   -2824.3530275347      1.95e-01  
   22   -2730.8476077244      1.90e-01  
   23   -2947.4965708218      3.01e-02  
   24   -3174.3898309565      3.79e-02  
   25   -3264.2280857778      1.42e-02  
   26   -3259.3894362046      1.11e-02  
   27   -3249.4709120044      7.03e-03  
   28   -3283.7267942361      5.91e-03  
   29   -3256.7817909434      4.42e-03  
   30   -3282.9580800217      3.56e-03  
   31   -3262.7826712188      4.04e-03  
   32   -3255.9528208792      3.14e-03  
   33   -3282.7749030364      3.59e-03  
   34   -3282.7775433973      3.58e-03  
   35   -3282.8120012657      3.58e-03  
   36   -3282.8086659317      3.58e-03  
   37   -3282.7970190584      3.57e-03  
   38   -3282.9080605099      3.51e-03  
   39   -3261.6591189642      1.77e-03  
   40   -3278.6152726475      4.37e-03  
   41   -3278.6179209698      4.37e-03  
   42   -3278.5046710591      4.35e-03  
   43   -3278.3208748361      4.23e-03  
   44   -3276.9427329564      4.89e-03  
   45   -3280.2615146947      4.13e-03  
   46   -3275.3257811312      3.48e-03  
   47   -3287.4133374067      4.18e-03  
   48   -3278.1477105100      4.58e-03  
   49   -3275.4883152227      4.39e-03  
   50   -3277.8635886309      4.86e-03  
   51   -3282.6066165370      3.76e-03  
   52   -3278.3272639446      4.68e-03  
   53   -3283.1623708654      3.66e-03  
   54   -3277.4419475859      4.84e-03  
   55   -3241.5112670314      5.32e-03  
   56   -3265.8948975487      2.82e-03  
   57   -3262.9432410025      2.30e-03  
   58   -3283.6897844728      3.61e-03  
   59   -3285.6918854314      3.88e-03  
   60   -3281.9403399028      4.58e-03  
   61   -3281.8459038294      4.55e-03  
   62   -3283.1109844621      3.66e-03  
   63   -3274.0871441439      3.47e-03  
   64   -3274.5116686751      3.52e-03  
   65   -3279.0074109700      4.35e-03  
   66   -3283.7287999811      5.13e-03  
   67   -3279.0620679060      4.95e-03  
   68   -3283.4595969020      4.91e-03  
   69   -3283.1328597113      4.71e-03  
   70   -3284.0918282275      3.67e-03  
   71   -3280.9716267967      4.78e-03  
   72   -3276.9331880018      5.10e-03  
   73   -3236.2379091324      4.19e-03  
   74   -3256.6657451325      3.31e-03  
   75   -3278.1609472705      3.66e-03  
   76   -3278.5493869571      4.01e-03  
   77   -3278.4352534907      3.97e-03  
   78   -3283.5559123530      3.78e-03  
   79   -3283.4759830305      3.80e-03  
   80   -3283.4175943262      3.77e-03  
   81   -3267.8162121666      3.32e-03  
   82   -3276.7716960236      3.61e-03  
   83   -3283.1186252966      4.18e-03  
   84   -3283.1394106233      4.34e-03  
   85   -3283.3559796230      4.52e-03  
   86   -3283.3723452730      4.52e-03  
   87   -3283.4147539213      4.56e-03  
   88   -3283.3483814110      4.68e-03  
   89   -3281.9299909615      5.50e-03  
   90   -3277.2968454642      3.96e-03  
   91   -3274.6527139707      3.77e-03  
   92   -3282.9902207375      4.16e-03  
   93   -3260.1154290855      2.62e-03  
   94   -3257.5275034451      3.17e-03  
   95   -3261.7275584639      1.47e-03  
   96   -3277.3444378119      4.94e-03  
   97   -3278.0917830126      4.91e-03  
   98   -3277.9982644021      4.98e-03  
   99   -3278.3486715527      4.55e-03  
  100   -3278.4376697517      4.49e-03  
  101   -3278.4512884858      4.48e-03  
  102   -3278.5002510546      4.37e-03  
  103   -3278.2640858280      4.17e-03  
  104   -3278.1635067690      4.16e-03  
  105   -3275.6524240061      4.33e-03  
  106   -3278.6848194735      4.04e-03  
  107   -3269.8096378099      5.82e-03  
  108   -3286.2157662694      3.99e-03  
  109   -3281.9804252311      4.23e-03  
  110   -3278.7932580454      4.66e-03  
  111   -3257.5195945274      3.73e-03  
  112   -3264.8189842121      2.03e-03  
  113   -3273.7731789675      3.71e-03  
  114   -3275.8304464561      3.37e-03  
  115   -3275.8905887870      3.44e-03  
  116   -3276.0846679737      3.46e-03  
  117   -3276.7865619103      3.76e-03  
  118   -3277.7951712963      4.27e-03  
  119   -3277.1733484544      3.80e-03  
  120   -3277.6579806301      4.12e-03  
  121   -3277.6692598544      4.11e-03  
  122   -3277.6189347200      4.18e-03  
  123   -3277.9707654391      4.35e-03  
  124   -3276.7446694539      3.86e-03  
  125   -3276.9442848446      3.81e-03  
  126   -3276.9495511169      3.83e-03  
  127   -3276.8448048775      3.88e-03  
  128   -3266.4510865297      3.29e-03  
  129   -3265.5850777465      3.29e-03  
  130   -3276.1301621357      3.37e-03  
  131   -3275.1150469053      3.76e-03  
  132   -3274.8094024475      3.70e-03  
  133   -3277.7431796292      3.86e-03  
  134   -3277.7565436404      3.92e-03  
  135   -3278.1892755923      4.08e-03  
  136   -3277.7259656183      4.00e-03  
  137   -3278.1536725209      4.22e-03  
  138   -3277.7094511799      3.88e-03  
  139   -3277.1479963186      3.78e-03  
  140   -3278.0776691458      4.28e-03  
  141   -3278.2949767961      4.34e-03  
  142   -3277.7073696697      4.01e-03  
  143   -3277.6682811620      4.08e-03  
  144   -3277.4121119742      3.98e-03  
  145   -3272.4318644177      4.06e-03  
  146   -3261.1691046337      2.77e-03  
  147   -3261.4463897837      1.96e-03  
  148   -3266.1264759169      2.53e-03  
  149   -3278.0837834678      4.06e-03  
  150   -3277.2947329703      3.70e-03  
  151   -3277.1251351610      3.69e-03  
  152   -3274.9979061932      3.59e-03  
  153   -3276.1600744676      4.00e-03  
  154   -3277.4569875260      5.09e-03  
  155   -3277.9149759919      4.74e-03  
  156   -3275.0434779455      3.83e-03  
  157   -3277.7919199973      4.94e-03  
  158   -3278.0688705411      4.80e-03  
  159   -3278.0668904715      4.81e-03  
  160   -3277.7106990689      4.69e-03  
  161   -3278.0100704344      4.69e-03  
  162   -3273.1758328669      4.43e-03  
  163   -3279.5570946620      4.26e-03  
  164   -3265.3583346263      3.39e-03  
  165   -3265.6241820647      3.12e-03  
  166   -3278.1282780392      4.39e-03  
  167   -3277.5874645700      4.11e-03  
  168   -3278.2935699929      4.34e-03  
  169   -3278.5913257816      4.20e-03  
  170   -3273.4240083520      5.21e-03  
  171   -3275.9290756587      3.48e-03  
  172   -3272.6445645786      4.30e-03  
  173   -3276.4251554039      3.60e-03  
  174   -3277.2020100289      3.95e-03  
  175   -3276.8002528943      3.82e-03  
  176   -3272.8209027125      4.17e-03  
  177   -3280.0085414794      4.21e-03  
  178   -3280.5144276334      4.30e-03  
  179   -3283.2030909158      3.54e-03  
  180   -3250.7204475449      4.60e-03  
  181   -3260.8464666356      2.46e-03  
  182   -3277.2066180918      3.99e-03  
  183   -3277.4202175855      4.06e-03  
  184   -3277.2917094781      3.95e-03  
  185   -3277.6277667094      3.78e-03  
  186   -3278.6177762785      4.08e-03  
  187   -3278.3662406251      3.98e-03  
  188   -3277.3685846188      3.73e-03  
  189   -3276.9197674240      3.63e-03  
  190   -3275.9499855221      3.50e-03  
  191   -3275.0160147240      3.54e-03  
  192   -3260.5799429416      4.96e-03  
  193   -3273.6097715546      3.78e-03  
  194   -3273.6711968801      3.79e-03  
  195   -3282.8932443838      3.82e-03  
  196   -3251.8352764316      3.20e-03  
  197   -3279.6984082958      5.22e-03  
  198   -3282.4311387944      3.48e-03  
  199   -3259.6369172366      2.74e-03  
  200   -3273.4964097811      3.56e-03  
gen_scfman_exception: SCF failed to converge

And I tried a 3rd time, it behaves normal in 20 steps without a problem. The total energy here declines with small ampitude:

 Total energy in the final basis set =    -2921.2937349402
 Total energy in the final basis set =    -2921.3935832946
 Total energy in the final basis set =    -2921.4393525208
 Total energy in the final basis set =    -2921.4544374699
 Total energy in the final basis set =    -2921.4531068491
 Total energy in the final basis set =    -2921.4575736581
 Total energy in the final basis set =    -2921.4536115344
 Total energy in the final basis set =    -2921.4551289011
 Total energy in the final basis set =    -2921.4574585515
 Total energy in the final basis set =    -2921.4596110190
 Total energy in the final basis set =    -2921.4585099152
 Total energy in the final basis set =    -2921.4602367289
 Total energy in the final basis set =    -2921.4602858263
 Total energy in the final basis set =    -2921.4604568527
 Total energy in the final basis set =    -2921.4602922321
 Total energy in the final basis set =    -2921.4604786410
 Total energy in the final basis set =    -2921.4604477830
 Total energy in the final basis set =    -2921.4605150857
 Total energy in the final basis set =    -2921.4604682740
 Total energy in the final basis set =    -2921.4605292532

So it really puzzles me about why this problem happens, is it truly depends on the inital guess and the set of good inital guess is so small that makes it hard to find a good inital guess? And for a large basis, it becomes even more difficult to find a good inital guess. However, this can not explain why sometimes starts with a good inital guess but changes suddenly in 6th cycle just like I post above.

Another question is about a warning in output onetime, it is:

 Size of previous SCF MO coefficient file: 231880 (expected); 835924 (actual). 

 Q-Chem warning in module libgscf/gen_scfman/rhf.C, line 652:

 Inconsistent size for SCF MO coefficient file.  Please check basis and PURECART setting.

I don’t know whether it is because the MO coefficient file read from the disk don’t match the MO coefficent you need in calcution(i.e. read def2-TZVPP coefficient but need STO-3G)?

By the way, how many cores and memory you use in the calculation?

Thanks for your reply.

#13

Hi Hanbin,
A couple of things here. That warning message indicates that the SCF coefficient file is much smaller than what Q-Chem is expecting. This warning can sometimes appear when there are changes in the linear dependencies from one step to the next but in those cases the message is that Q-Chem is looking for a slightly different number of coefficients than what it finds, and usually those warnings are not a problem. In your case, something is clearly wrong. The discrepancy is so large that I think this has to be the result of trying to read a small BASIS2 (maybe STO-3G) when Q-Chem is expecting the large target basis (maybe def2-TZVPP). That said, it’s not clear to me what combination of input options got you to this point, however, because BASIS2 does work with geometry optimizations and does not normally generate this error. Note that it should not be necessary to use SCF_GUESS=READ anywhere in order to use, e.g., BASIS2=STO-3G, so if you are specifying SCF_GUESS=READ then that is likely your problem.

Second, if I understand correctly you are getting inconsistent behavior in different basis sets. (Your various outputs must be different basis sets because the total energies are quite different.) I’m not sure what is going on here because I’m unable to reproduce that behavior. With BASIS=STO-3G the optimization converges without issue, that’s what I said last night. I just ran BASIS=DEF2-TZVPP, both with and without BASIS2=STO-3G, and on 28 cores it doesn’t quite make it through one SCF cycle in an hour but it definitely looks like it’s on it’s way to convergence. (The nodes that I can get without waiting are time-limited at one hour.) I used the same memory settings, etc. as in your original input file, changed only BASIS and/or BASIS2.

Finally, I would reiterate that I don’t understand your desire to use def2-TZVPP here. That’s usually overkill for DFT geometry optimizations, and your system is particularly rigid. At a minimum, even if I wanted that answer (say, to compare to something in the literature), I would pre-optimize with a smaller basis set. Using BASIS=6-31G (without BASIS2), I can get 8 steps in one hour on 28 cores, again using your input file as given except for the change in basis. That calculation proceeds without problem:

   15   -2959.6375197983      3.89e-09  Convergence criterion met
   12   -2959.6453431238      7.77e-09  Convergence criterion met
   12   -2959.6462627051      7.39e-09  Convergence criterion met
   12   -2959.6443251407      8.60e-09  Convergence criterion met
   12   -2959.6479709568      7.74e-09  Convergence criterion met
   12   -2959.6460471442      7.08e-09  Convergence criterion met
   12   -2959.6478102965      5.12e-09  Convergence criterion met
   12   -2959.6472449032      7.27e-09  Convergence criterion met
#14

I may find the reason…
I use 32 cores this time comparing with 16 cores before, and it behaves normal in every calculation.
I don’t konw why this can solve the problem, but I find this is always effective.
Thank you so much for discussing with me!

#15

If you are getting different answers on different numbers of cores then that seems like a bug and you should report it to Q-Chem support.

#16

This job completed for me with Q-Chem 5.3.2 in 23 cycles without any SCF issues whatsoever. @hanbin could you please send the full output to support@q-chem.com for inspection?

#17

My list of energies above (8 steps in one hour on 28 cores) were for 6-31G(d), the website formatting ate the “star”. This SCF converges w/o issue for me in that basis and several others.