Power Spectrum. Help wanted.

[z sh]

Hello everyone,

I am trying to do vibrational analysis for gas phase system. I am following the tutorial " Computing Bulk Phase Vibrational Spectra with CP2k and TRAVIS (2018)". Initially i have done massive equilibration of methanol system for 6000 steps. After that i have done production run for 60000 steps. Then, I calculated the power spectrum with xyz. file. Unfortunately, O-H stretching vibration at 3700 cm-1 is not seen in the power spectrum. Would you please guide me why peak 3700 cm-1 is not seen?

 

 Here i am attaching my power spectrum.

 Please help me on this.

 Thanks.

[Lucas Lodeiro]

Hello, 

Computing a single molecule MD is not easy, and many things have to be checked. Running a single molecule NVT simulation is difficult due to the thermostat not being well behaved. Also it is important to check the angular momentum of the system is practically zero, because the rotational movement could make "fake signals", also the system linear momentum affect the power spectrum at 0 cm⁻¹. There are keywords to control both. If your power spectra shows a non zero value at cm⁻¹, probably you have one or both problems.

About the signal at 3700cm⁻¹, the high frequency signals needs a time step of 0.5 fs or at least 1.0 fs, also, if you use a short time step, you need to save the coordinates at each step. Also, if you can print the instantaneous velocities from the MD, you can use them directly and not approximate them by differentiation of coordinates, as done by TRAVIS when only the xyz file is used.

Remember the power spectra shows signals where their area is proportional to the number of modes at each frequency frame. Also, if your system has more than one methanol molecule, maybe there are H-bonds, which weakens the O-H bonds and "moves" those signals to lower frequencies.

Regards

--
You received this message because you are subscribed to the Google Groups "cp2k" group.
To unsubscribe from this group and stop receiving emails from it, send an email to cp2k+uns...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/53635676-500b-4caa-8e2b-f4c3a2ee7871n%40googlegroups.com.

[Thomas Kühne]

Beside the lack of the dangling OH peak there are a couple of more things at odds: 

(i) symmetric OH stretch shifted by ~350 cm-1

(ii) bending mode

Without providing your input file one can only speculate, maybe a resonance effect 

du to the thermostat, time step etc. or s.th. completely different … 

Best, 

Thomas 

<power.png>

--
You received this message because you are subscribed to the Google Groups "cp2k" group.
To unsubscribe from this group and stop receiving emails from it, send an email to cp2k+uns...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/53635676-500b-4caa-8e2b-f4c3a2ee7871n%40googlegroups.com.

<power.png>

==============================

Thomas D. Kühne

Dynamics of Condensed Matter

Chair of Theoretical Chemistry

University of Paderborn

Warburger Str. 100

D-33098 Paderborn

Germany

thomas...@upb.de

+49/(0)5251/60-5726

[z sh]

Dear Lucas

Thanks for your helping.

[z sh]

Dear tkuehne

 

Thanks for your helping.

Here i am attaching my input files of massive pre-equilibration and production run of methanol system.

Please help me on this.

Thanks.

[z sh]

Dear Thomas 

Thanks for your helping.

Here i am attaching my input files of massive pre-equilibration and production run of methanol system.

Please help me.

Thanks.

[Thomas Kühne]

OK, I misread your email. I thought you are computing a bulk phase vibrational spectra. 

In any case, I would suggest using the CSVR thermostat that is much more robust wrt. 

to dynamic quantities than the Nose-Hoover and moreover much less prone to technical 

issues such as ergodicity etc. 

To equilibrate you may you e.g. CSVR, MASSIVE, TIMECON 50.0 or even less (CSVR 

can also be used with smaller time constants!) and NVE, or if temperature control is 

desired for such a small system, CSVR, MASSIVE, TIMECON 500.0 for the production 

run. 

Best, 

Thomas

To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/9240b519-0247-4fbc-8374-4d9820e3def0n%40googlegroups.com.
<production-run.inp><massive equilibration-run.inp>

[z sh]

Dear Lucas Lodeiro

 

Hello,

Thank you for your help. As you said I calculated the instantaneous velocities of atoms of methanol from production run. Then, I calculated the power spectrum with velocity file. As this plot shows, the O-H stretching vibration at 3700 cm-1 is seen but unfortunately, the other vibrations from 0 cm-1 to 800 cm-1 are not seen in power spectrum. Would you please guide me why all of the peaks from 0 cm-1 to 800 cm-1 are not seen?

 

Here i am attaching my power spectrum that was calculated with velocity data.

Please help me on this.

Thanks.

On Sunday, June 27, 2021 at 4:58:04 PM UTC-7 Lucas Lodeiro wrote:

[Lucas Lodeiro]

Hi, it is not clear why your spectrum does not show the low frequency peaks... Without more information it is impossible to say anything. It could be problems on the dynamics (AIMD) or problems defining the settings on TRAVIS to obtain the power spectra.

Regards

To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/0595a3ca-37f1-4e7d-941b-411674fc3167n%40googlegroups.com.

[Long Wang]

Hi,

I have a question regarding the TRAVIS and power spectrum.

Is it require only the xyz trajectory to calculate the power spectrum?.

Could you provide the input for it.

Thank you for your help.

Best

Long