Please add:

shiftup 2.0

shiftdn 2.0

in the scf block. Also, you can decrease the energy threshold to 1e-7 or 1e-6 (the default value is 1e-8).

You can take a look to ORCA user's forum too:

https://orcaforum.cec.mpg.de/viewtopic.php?f=8&t=3147&hilit=SCF+converges

The last time one of my calculations showed this message " HUGE, UNRELIABLE STEP WAS ABOUT TO BE TAKEN" it was because my initial structure was not very well optimized. Apparently when the SCF procedure results in a huge step (High energy difference between configurations), the software is program to consider the run too unreliable, and possibly wrong. Try performing a better optimization before running your DFT opt.

Other observations:

Is you multiplicity 6 and the charge -3?

You should use tightSCF at least.

Why are you using !MORead? If this is the first time you are calculating it, there is no reason to use it. Use something like PMODEL instead.

Also, it doesn't really makes sense using SlowConv and NormalSCF. Just use TightSCF, and if you are having any issues when with convergence you can try to use SlowConv.

I think you don't need to use the command Grid4 since 4 is the default value. I'm not sure about Gridx4, but I think it is also the default value.

According to the manual of the last version of orca, the variable for CNVDIIS is either true or false, didn't find 1 as a value any where in the manual of this version. Same for CNVSOSCF. If you are not sure if you should change these settings, just use the default ones.

You input should be more like this I think(Check the -3 6 values):

! B3LYP OPT def2-SVP def2/J NoFinalGrid RijCosX Gridx4 TightSCF

! PMODEL PRINTBASIS

* xyz -3 6

coordinates

*

Narges Masoumi thanks for the suggestion. i will go through it.

@ Eduardo Maurina Morais thanks for the valuable input. i am trying to optimize Fe(III)tris catecholate complex in which it has -3 charge and 6 multiplicity (high spin). i started this using the manual instructions.

previously i have done with tightSCF which terminated with same error. that why started again the calculation with some input modifications (eg. Moread from previous cal and slowconv). i will try once with your suggestions.

may I know what is your suggestion for better optimization before DFT... should i start with HF or single point energy? or charge and multiplicity makes it difficult?

Maheshwaran Duraiyarasu The optimization depends of the kind of molecule. For organic molecules is quite simple, I usually start with simple MM such as UFF or MM94FF, then PM3, HF and finally DFT. But since I don't work with metals, I'm not sure if you can use the same methods. But you should definitively progressively improve your structure to avoid convergence errors. It would be a good idea to go straight from MM to DFT for instance, especially for complex molecules.

I agree with Eduardo. For the next steps (beyond optimization) you can take a look to the ORCA CASSCF tutorial. Please find the attachment.

Dear Eduardo Maurina Morais thanks again. I am also doing same what u said. I start with uff then dft and i have successful in gaussian for some metal complexes. this is my new try. the calculation going on as per ur suggestion. thanks to Narges Masoumi for the document. i am reading and trying to understand. I will get back to you in a short while.

Good luck!

Hello! I have some experience working with ORCA, so I believe I can be of some help here.

The problem is that ORCA is trying to use SOSCF as a SCF converger. SOSCF is lightning fast for closed shell systems but almost always will bring you problems if you have an open shell.

Get rid of CNVSOSCF 1 in your %scf block, put NOSOSCF in your keyword block (the !) and if the default DIIS does not perform the magic for you, you can try the KDIIS solver (almost always the best for open shell).

Best of luck.

Henrique C. S. Junior thanks for the input.

Dear Eduardo Maurina Morais finally i have successfully finished optimization calculation. thanks for the suggestion. however it took more than 10 days to finish the calculation in PC.

In the calculation it generates alpha as well as beta orbitals. but i can visualize only alpha orbitals in avogadro program. is there any option to visualize beta orbitals too? can you give some suggestions regarding this if you know?

Maheshwaran Duraiyarasu Happy to hear that! I think you can use Gabedit for beta orbitals.(http://gabedit.sourceforge.net/).

The easiest way to view your orbitals is using ChemCraft: https://www.chemcraftprog.com/download.html

Henrique C. S. Junior thanks for your input. right now I am using this software. it is very user friendly.

Dear Dr Duraiyarasu Maheshwaran

Can you share your last input file?

I am using ORCA in MOF. In my calculations ORCA finished by error "file orca tools/qcmsg.cpp line 458"

My input file keywords are:

! UKS OPT def2-TZVP b3lyp

!Normalprint

%output

Print[P_Basis]2

Print[P_MOs]1

end

* xyz 0 2

Thank you in advance

Dear Alisher Eshimbetov,

Please find the input details below,

! B3LYP OPT def2-SVP def2/J NoFinalGrid RijCosX Gridx4 TightSCF

%scf

MaxIter 500

CNVDIIS 1

CNVSOSCF 1

end

%output

print[p_mos] true

print[p_basis] 5

end

* xyz -3 6

coordinates

*

Thank you Dear Dr. Duraiyarasu Maheshwaran

Dear Dr. @Duraiyurasu I don't know if you have as for now solved your problem. I recommend you add the keyword NoSOSCF in the keyword line and remove CNVSOSCF 1 in the %scf block. This should work out. Wish you the best

Hi Tushar Pandit