did you put scf=maxcycle=1000 ? Maybe your calculation requires more cycles to converge than the default cycles of 256 (I dont remember). Aslo 6-31+G* might be a bit difficult to converge. Especially when you opt with HF/Sto-3G . I think the input structure is still not good enough. Try to increase the basis set gradually. Try B3LYP/6-31G first. If none of that work, you can opt using scf=qc.   It is mostly for difficult conversion cases using a quadratically convergent SCF procedure. It might take lot of time and much slower but the results are more accurate

Generally, there are two options that could be applied to improve convergence error

First, scf=fermi option is generally used to improve convergence failure

Second, you could obtain normal termination using a low convergence cut off value, after that, you can perform default optimization using of optimized structure of low cut off value

Of course, scf = qc option is strong tool to improve convergence error but it takes much time (not efficient)

Hope this help

Pre optimize using a force field. If you do not have one try and optimize with PM6 than proceed with higher levels. Avogadro  is a free too for a force field optimization.

First: what is the point of doing DFT geometry optimization for this molecule? Do You just want to have a reasonable ligand structure and then dock it to a protein? If so, I don't think going to such high-level methods is necessary, since the conformation of ligand in solvent (non-interacting) is rarely identical with the docked conformation.

If you docking protocol allows fro ligand flexibility, I would just optimize the ligand at a much lower level, something like PM6 should do just fine. If you're using rigid docking, I will do conformational search on even lower level (MM) and make sure that I have a good set of diverse conformations, rather than having one, well optimized conformer.

If you, however, insist on QM: don't do HF and don't go below 6-31G basis set, it doesn't make any sense. There is a set of gaussian commands that usually work in cases, where convergence is difficult:

Int=UltraFine (basically mandatory for DFT

SCF=QC

SCF=XQC

opt=z-matrix

opt=cartesian

Try different combinations of these and see if it works. If it doesn't - preoptimize the ligand using PM6 (as suggested above) or using other DFT functional.

Thank you all... I did the dockings using PM 6.

why convergence error appears during the optimization of Methyl ammonium Lead Tri iodide (CH3 NH3PbI3) crystal in g09? The following error appears:

Convergence failure -- run terminated.

Any response is highly appreciated.

Thank you

Hi Everyone,

Can anyone tell me the reason of error for the DFT of a molecule containing Sr atom? I had tried scf=xqc also and without integral too.

Input:

%nprocshared=12

%rwf=xyz.rwf

%nosave

%mem=2500MB

%chk=xyz.chk

#p opt b3lyp/gen pseudo=cards scf=tight integral(grid=ultrafine) geom=nocrowd geom=connectivity

0 1

.........

.......

C H O N 0

6-31G

****

Sr 0

LANL2DZ

****

Sr 0

LANL2DZ

Output

>>>>>>>>>> Convergence criterion not met.

SCF Done: E(RB3LYP) = -5406.32747855 A.U. after 129 cycles

Convg = 0.1694D-05 -V/T = 2.0091

KE= 5.357699770761D+03 PE=-6.589180508508D+04 EE= 2.851595118398D+04

Convergence failure -- run terminated.

Regards

If there is no problem about starting geometry and/or spin charge values then you can force gaussian to continue calculation with scf=qc keyword. if still you are getting error message you can add use=l506 to route section as well . Here is an example I used and got rid off error message (reference: gaussian manual and/or web site)

%nprocshared=4

# opt pm6 geom=connectivity scf=qc use=l506

Molecule Name

0 2

C 0.27600000 -3.35400000 -0.73200000

C -0.55800000 -4.05200000 -1.62500000

Also if you are calculating MK charges you can use UHF/6-31G∗∗+LanL2DZ level

reference for calculating MK charges with this level: Article Evaluation of AMBER force field parameters for copper(II) wi...

Nice to share an outstanding summary about "How to solve the error message in Gaussian" as follow:

https://docs.computecanada.ca/wiki/Gaussian_error_messages

please,i want to know which basis to use,if i want to doppe C60 with Ag ? polarisation? what it can be my input?

I want to do an optimization for C60 with transition metal (complexes)

thank u

There are other possibilities present to solve the convergence message

use

scf = xqc

scf = qc use 1506

opt = Z-matrix

opt = cartesian