Normal modes visualisation with molden

[woody]

Hello,

I want to calculate some IR-Spectra, so I did a normal mode
(VIBRATIONAL_ANALYSIS) on my molecule. I also got normal modes but I
cannot visualise them with molden, because molden cannot read the
output format.

I need molden, because with molden I can get the frequencies and
intesities so I can plot them with gnuplot.

Has anyone a suggestion how molden can read my output?

Or does anyone know another visualisation tool which can do the same
job?

Thanks for Help

woody

[Teodoro Laino]

Woody,

the output of the normal modes at the end of cp2k resembles the one
of Gaussian.
At the moment the output is not supported by any kind of graphical
application (as far as I know)..
Two possibilities:

1) you can play with a fake output Gaussian (to be built by yourself)
cut&pasting normal modes and
frequenciesin order to allow molden to read the frequencies and
normal modes

2)wait a couple of days: i will check in the molden format (will send
a message on this forum when it's ready).. this assumes
that you've to redo the calculations then..

Teo

[Teodoro Laino]

Anyway if you want the molden output only for:

On 19 Oct 2007, at 10:30, woody wrote:

I need molden, because with molden I can get the frequencies and

intesities so I can plot them with gnuplot.

you can just read them at the end of the output.. Beware, intensities are

not yet provided..

Cheers

teo

[woody]

Hi Theo,

thanks for the quick answear.

> 1) you can play with a fake output Gaussian (to be built by yourself)
> cut&pasting normal modes and
> frequenciesin order to allow molden to read the frequencies and
> normal modes

Yes, I still work on that but I havn't figured it out yet.

> 2)wait a couple of days: i will check in the molden format (will send
> a message on this forum when it's ready).. this assumes
> that you've to redo the calculations then..

That would be great. With QS my calculations don't take as long as
with gaussian, so thats no problem.

Thanks for your Help

Woody

[woody]

> you can just read them at the end of the output.. Beware, intensities
> are
> not yet provided..

Hm, thats not good...
Actually I haven't noticed that yet.
There are only Frequencies, Reduced Masses and Force constants
available, right?

You mean that there is no way (maybe another keyword?) to get the
intensities at this moment?

woody

[Teodoro Laino]

for intensities you need the derivatives of the dipole moment w.r.t.
normal modes.
We can compute the dipole moment.. so in principle you can always
compute by
hand the derivatives of the dipole moment once you know the normal
modes.

For sure we will implement them in the next weeks (anyway it has a
very low priority)..
teo

[woody]

>the output of the normal modes at the end of cp2k resembles the one of Gaussian.

The output of QS ist similar to Gaussian.
Molden by the way reads the exact Gaussian output.

> For sure we will implement them in the next weeks (anyway it has a
> very low priority)..

If it doesn't produce much extra work it would be gread if you could
modify the QS normalmode and xyz output to Gaussian.

Will you post a message anywhere when this is done?

Thanks a lot.
woody

[flo]

Hi everybody,

I just finished a Davidson algorithm for mode selctive calculations.
On this way I teached cp2k to write a molden output File (only for
calculations without frozen atoms yet). Intensities are still not
available. Maybe I will add them later on, while improving the
algorithm.

flo