Frequencies with frozen atoms
Just Got Here
2:25:36 PM PDT - Mon, Jul 2nd 2012
HI!

I am new at NWCHEM and i have been running some tests on optimizing and running numerical frequencies with frozen atoms. And surprizingly I get very different results ( different frequencies) if I do the following:

CASE A:

A1. i generate a file.nw in which I specify geometry, charge basis set, active list ... and I put task dft optimize.

A2. I restart the calculation using the restart keyword and put task dft frequencies numerical

CASE B:

B1. I submit a job with the same molecule and all but I put both dft optimize and dft frequencies numerical in the same .nw, the frequency calculation didn't finish

B2. I restart the frequency calculation using restart.


I have pasted the .nw input files below ... could somebody tell me what is wrong? or what am I missing? is there some information I have to always put in the input file even if I restart the calculation?

Thanks!

A1----------------
echo

start b-optimizefroz

title 'b-optimizefroz'

charge 0

geometry units angstrom noautoz nocenter noautosym
C                 -0.86999022   -0.43499511    0.00000000

end

basis
H library def2-tzvp

N library def2-tzvp

C library def2-tzvp
end

driver
maxiter 200
end

dft
iterations 200

xc cpbe96

mulliken

direct
end

set geometry:actlist 14:25

task dft optimize

A2----------------
echo

restart b-optimizefroz

title 'b-optimizefroz'

task dft freq numerical

B1----------------

the same as A1 plus :

task dft freq numerical

B2----------------
echo

restart b-optimizefroz-freq

title 'b-optimizefroz'

charge 0

basis
H library def2-tzvp

N library def2-tzvp

C library def2-tzvp
end

driver
maxiter 200
end

dft
iterations 200

xc cpbe96

mulliken

direct
end

set geometry:actlist 14:25

task dft freq numerical

Forum Vet
2:42:40 PM PDT - Mon, Jul 2nd 2012
In A1/2. Did the DFT optimize converge? I.e. did you do a frequency restart using a converged geometry.

You may want to check the geometries used in A and B to see if the same geometries were used in the frequency calculation, and that the optimized geometries are actually the same.

If both items above are true, to reproduce, could you need to provide the full A1 input deck (with all atoms).

Thanks,

Bert


Quote:Noraplanas Jul 2nd 9:25 pm
HI!

I am new at NWCHEM and i have been running some tests on optimizing and running numerical frequencies with frozen atoms. And surprizingly I get very different results ( different frequencies) if I do the following:

CASE A:

A1. i generate a file.nw in which I specify geometry, charge basis set, active list ... and I put task dft optimize.

A2. I restart the calculation using the restart keyword and put task dft frequencies numerical

CASE B:

B1. I submit a job with the same molecule and all but I put both dft optimize and dft frequencies numerical in the same .nw, the frequency calculation didn't finish

B2. I restart the frequency calculation using restart.


I have pasted the .nw input files below ... could somebody tell me what is wrong? or what am I missing? is there some information I have to always put in the input file even if I restart the calculation?

Thanks!

A1----------------
echo

start b-optimizefroz

title 'b-optimizefroz'

charge 0

geometry units angstrom noautoz nocenter noautosym
C                 -0.86999022   -0.43499511    0.00000000

end

basis
H library def2-tzvp

N library def2-tzvp

C library def2-tzvp
end

driver
maxiter 200
end

dft
iterations 200

xc cpbe96

mulliken

direct
end

set geometry:actlist 14:25

task dft optimize

A2----------------
echo

restart b-optimizefroz

title 'b-optimizefroz'

task dft freq numerical

B1----------------

the same as A1 plus :

task dft freq numerical

B2----------------
echo

restart b-optimizefroz-freq

title 'b-optimizefroz'

charge 0

basis
H library def2-tzvp

N library def2-tzvp

C library def2-tzvp
end

driver
maxiter 200
end

dft
iterations 200

xc cpbe96

mulliken

direct
end

set geometry:actlist 14:25

task dft freq numerical

Just Got Here
10:28:02 AM PDT - Tue, Jul 3rd 2012
Hi,

1. Yes in both cases the geometry converged, but I cannot vizualize the outputfile since I am fighting to get ECCE installed on my mac computer ( any other simpler software could be used for that?)

2. Secondly, even if the geometries are different once I run frequencies I get 25 imaginary frequencies which makes me doubt the frequency calculation did indeed locate the frozen atoms since to my understanding I should get no computed frequencies for those right?

Thanks,

Nora

Forum Vet
3:30:42 PM PDT - Tue, Jul 3rd 2012
I'll need the full A1 input deck to try and reproduce and understand the issue.

Bert


Quote:Noraplanas Jul 3rd 5:28 pm
Hi,

1. Yes in both cases the geometry converged, but I cannot vizualize the outputfile since I am fighting to get ECCE installed on my mac computer ( any other simpler software could be used for that?)

2. Secondly, even if the geometries are different once I run frequencies I get 25 imaginary frequencies which makes me doubt the frequency calculation did indeed locate the frozen atoms since to my understanding I should get no computed frequencies for those right?

Thanks,

Nora


Forum >> NWChem's corner >> General Topics