Chemical Forums

Chemistry Forums for Students => Physical Chemistry Forum => Topic started by: roo_sado on May 16, 2021, 03:13:17 PM

Title: Geometry Optimization of oxi-methylphenantrene using DFT in GAMESS
Post by: roo_sado on May 16, 2021, 03:13:17 PM
Hey! So, I'm not sure if this is supposed to happen, as my teacher hasn't replied to my queries, but my program has been running for well over 2 hours now, and yesterday I left it running for 11 hours and got nothing.

I'm supposed to do a vibrational analysis of the molecule, but for that I need the correct geometry. The teacher gave us a few specific functionals we should be using, and for the current optimization I'm running I'm using PBE exchange + PBE correlation.


Please help, I have to run 5 more files and at +12 hours per file I won't be done in time...

Thanks


P.D.
Here's the input file I used.

 $BASIS GBASIS=N21 NGAUSS=3 $END
 $CONTRL SCFTYP=UHF RUNTYP=OPTIMIZE DFTTYP=PBE MAXIT=99 $END
 $SYSTEM MWORDS=5 $END
 $STATPT OPTTOL=0.0001 NSTEP=20 HESS=CALC $END
 $FORCE METHOD=SEMINUM VIBANL=.FALSE. $END

 $DATA
Title
C1
C     6.0     3.13984     0.61327     0.30139
C     6.0     3.45596    -0.69886    -0.03575
C     6.0     2.43944    -1.60595    -0.32763
C     6.0     1.08918    -1.22237    -0.28903
C     6.0     0.72629     0.10535     0.05056
C     6.0     1.78990     0.99942     0.34164
C     6.0    -0.66926     0.46599     0.08019
C     6.0    -1.64470    -0.51708    -0.23204
C     6.0    -1.24719    -1.81496    -0.56145
C     6.0     0.10012    -2.16249    -0.58936
C     6.0    -1.15322     1.76017     0.40510
C     6.0    -2.51803     2.06763     0.42023
C     6.0    -3.43710     1.08145     0.10889
C     6.0    -3.01262    -0.19915    -0.21420
C     6.0     4.21383     1.61069     0.62039
H     1.0     4.49121    -1.02857    -0.07523
H     1.0     2.71398    -2.62615    -0.58873
H     1.0     1.57721     2.03109     0.60893
H     1.0    -1.98725    -2.57557    -0.80203
H     1.0     0.36862    -3.18388    -0.85096
H     1.0    -0.46536     2.56266     0.65544
H     1.0    -2.83244     3.07429     0.67576
H     1.0    -3.76043    -0.95148    -0.45378
O     8.0    -4.77724     1.32932     0.10809
H     1.0    -4.92104     2.25938     0.34780
H     1.0     5.21164     1.16702     0.53966
H     1.0     4.09254     1.98085     1.64346
H     1.0     4.16441     2.45366    -0.07636
 $END
Title: Re: Geometry Optimization of oxi-methylphenantrene using DFT in GAMESS
Post by: Orcio_Dojek on May 16, 2021, 03:47:20 PM
Have you see your compound ?

All hydrogens lies on a different plane than carbon atoms !!!

Besides - you do not need UHF if molecule is not radical (and I don't see any special multiplicity).

Also - molecule should be planar so application of Cs point group will reduce calculation time.
Title: Re: Geometry Optimization of oxi-methylphenantrene using DFT in GAMESS
Post by: roo_sado on May 16, 2021, 04:03:33 PM
Have you see your compound ?

All hydrogens lies on a different plane than carbon atoms !!!

Besides - you do not need UHF if molecule is not radical (and even for radicals sometimes ROHF is good enough).

Also - molecule should be planar so application of Cs point group will reduce calculation time.


I have seen the molecule, the only hydrogen atoms that seem to be off the plane are the ones in the methyl group, how would I align those?
Title: Re: Geometry Optimization of oxi-methylphenantrene using DFT in GAMESS
Post by: Orcio_Dojek on May 16, 2021, 04:13:52 PM
Quote from: roo_sado
$DATA
Title
C1
C     6.0     3.13984     0.61327     0.30139
C     6.0     3.45596    -0.69886    -0.03575
C     6.0     2.43944    -1.60595    -0.32763
C     6.0     1.08918    -1.22237    -0.28903
C     6.0     0.72629     0.10535     0.05056
C     6.0     1.78990     0.99942     0.34164
C     6.0    -0.66926     0.46599     0.08019
C     6.0    -1.64470    -0.51708    -0.23204
C     6.0    -1.24719    -1.81496    -0.56145
C     6.0     0.10012    -2.16249    -0.58936
C     6.0    -1.15322     1.76017     0.40510
C     6.0    -2.51803     2.06763     0.42023
C     6.0    -3.43710     1.08145     0.10889
C     6.0    -3.01262    -0.19915    -0.21420
C     6.0     4.21383     1.61069     0.62039
H     1.0     4.49121    -1.02857    -0.07523
H     1.0     2.71398    -2.62615    -0.58873
H     1.0     1.57721     2.03109     0.60893
H     1.0    -1.98725    -2.57557    -0.80203
H     1.0     0.36862    -3.18388    -0.85096
H     1.0    -0.46536     2.56266     0.65544
H     1.0    -2.83244     3.07429     0.67576
H     1.0    -3.76043    -0.95148    -0.45378
O     8.0    -4.77724     1.32932     0.10809
H     1.0    -4.92104     2.25938     0.34780
H     1.0     5.21164     1.16702     0.53966
H     1.0     4.09254     1.98085     1.64346
H     1.0     4.16441     2.45366    -0.07636
 $END
You need to correct every one hydrogen + oxygen atom, not just methyl group.
Title: Re: Geometry Optimization of oxi-methylphenantrene using DFT in GAMESS
Post by: roo_sado on May 16, 2021, 04:24:22 PM

You need to correct every one hydrogen + oxygen atom, not just methyl group.


Right, I'm sorry, this is how the molecule looks right now. Are you saying I should be building the molecule aligned relative to, say XY?
Title: Re: Geometry Optimization of oxi-methylphenantrene using DFT in GAMESS
Post by: Orcio_Dojek on May 16, 2021, 04:31:31 PM
Application of Cs point group will reduce time of calculation by twice.

Application of RHF instead of UHF will also reduce time of calculation.

Still it will take many hours, as computing IR spectrum is the most time consuming.
Title: Re: Geometry Optimization of oxi-methylphenantrene using DFT in GAMESS
Post by: roo_sado on May 16, 2021, 04:40:20 PM
Application of Cs point group will reduce time of calculation by twice.

Application of RHF instead of UHF will also reduce time of calculation.

Still it will take many hours, as computing IR spectrum is the most time consuming.

Okay, I'll try to change those. What do I need to run it with the CS point group? I never learned anything about this, we always used C1...

Also I really appreciate you taking the time to reply.
Title: Re: Geometry Optimization of oxi-methylphenantrene using DFT in GAMESS
Post by: Orcio_Dojek on May 16, 2021, 04:44:04 PM
Quote
$DATA
Title
C1
Instead "C1" type "Cs".

But you will get error if molecule do not fits Cs symmetry.
Title: Re: Geometry Optimization of oxi-methylphenantrene using DFT in GAMESS
Post by: Orcio_Dojek on May 16, 2021, 05:12:43 PM
Increasing Mwords will also help (1 mword = only 8 mb of allocated memory).
Title: Re: Geometry Optimization of oxi-methylphenantrene using DFT in GAMESS
Post by: roo_sado on May 16, 2021, 05:18:50 PM
Quote
$DATA
Title
C1
Instead "C1" type "Cs".

But you will get error if molecule do not fits Cs symmetry.
I did get an error :/
Title: Re: Geometry Optimization of oxi-methylphenantrene using DFT in GAMESS
Post by: Orcio_Dojek on May 16, 2021, 05:23:55 PM
Molecule is planar so one of the coordinate (x, y, z) should be equal 0 for every atom, except of two hydrogen atoms in methyl group.
Title: Re: Geometry Optimization of oxi-methylphenantrene using DFT in GAMESS
Post by: roo_sado on May 16, 2021, 05:48:31 PM
Molecule is planar so one of the coordinate (x, y, z) should be equal 0 for every atom, except of two hydrogen atoms in methyl group.

So now it looks like this, but it still gives me the same error

 $BASIS GBASIS=N21 NGAUSS=3 $END
 $CONTRL SCFTYP=RHF RUNTYP=OPTIMIZE DFTTYP=PBEX MAXIT=99 $END
 $SYSTEM MWORDS=1000 $END
 $STATPT OPTTOL=0.0001 NSTEP=20 HESS=CALC $END
 $FORCE METHOD=SEMINUM VIBANL=.TRUE. $END

 $DATA
Title
CS
C     6.0   -12.73450    -2.10578     0.00000
C     6.0   -12.08405    -0.88474     0.00000
C     6.0   -10.68539    -0.85494     0.00000
C     6.0    -9.89978    -2.03709     0.00000
C     6.0   -10.60561    -3.26886     0.00000
C     6.0   -12.01003    -3.28890     0.00000
C     6.0    -8.45821    -2.05304     0.00000
C     6.0    -7.78171    -3.29905     0.00000
C     6.0    -8.51118    -4.49083     0.00000
C     6.0    -9.90283    -4.47600     0.00000
C     6.0    -7.64600    -0.88824     0.00000
C     6.0    -6.24323    -0.94145     0.00000
C     6.0    -5.61576    -2.18393     0.00000
C     6.0    -6.37855    -3.34957     0.00000
C     6.0    -5.42158     0.31362     0.00000
H     1.0    -6.05237     1.20859     0.00000
H     1.0    -4.78872     0.34983    -0.89255
H     1.0    -4.78872     0.34983     0.89255
H     1.0    -8.10848     0.09483     0.00000
H     1.0    -4.53066    -2.25440     0.00000
H     1.0    -5.86237    -4.30757     0.00000
H     1.0    -8.00113    -5.45189     0.00000
H     1.0   -10.43469    -5.42536     0.00000
H     1.0   -12.55161    -4.23193     0.00000
H     1.0   -12.63468     0.05029     0.00000
H     1.0   -10.21463     0.12391     0.00000
O     8.0   -14.09478    -2.18948     0.00000
H     1.0   -14.46100    -1.29001     0.00000               
 $END
Title: Re: Geometry Optimization of oxi-methylphenantrene using DFT in GAMESS
Post by: Orcio_Dojek on May 17, 2021, 09:34:43 AM
When using point group you need to leave an empty line between the symbol of point group (here "Cs") and coordinates of first atom (C).

Plus - you need to specify type of coordinates (coord=prinaxis) in $contr group.
Title: Re: Geometry Optimization of oxi-methylphenantrene using DFT in GAMESS
Post by: Orcio_Dojek on May 17, 2021, 10:17:23 AM
As of using Cs point group....

I was wondering is the molecule in fact planar (H9 hydrogen lies on the same plane or above), but somehow - planar structure is slightly less stable than structure with H9 out of plane, so yes - Cs point group is even better for purpose of calculations.
Title: Re: Geometry Optimization of oxi-methylphenantrene using DFT in GAMESS
Post by: roo_sado on May 17, 2021, 02:04:13 PM
Thanks a lot! I was able to finish by changing the HESS to GUESS. It still took about 8-10 hours of around constant 100% CPU and 100% memory allocated usage.
Title: Re: Geometry Optimization of oxi-methylphenantrene using DFT in GAMESS
Post by: Orcio_Dojek on May 17, 2021, 07:30:39 PM
I think that optimalization of geometry could also be omitted, as initial geometry looks good enough.

I mean - optimization is waste of time if you need only IR spectra.