Is it necessary to perform frequency calculations for the optimized geometry in cp2k?

[Mohammad Shakiba]

Hello everyone,

When optimizing with CP2K, we set a force limit and a maximum step size (for example less than 10meV and 0.002Bohr) and when the calculations reached those limits it is done. Before using CP2K I used Gaussian 09. When the geometry optimization was fully performed, I performed frequency calculation in order to check that there is no imaginary frequency so the optimization is not stuck in the local minima. However, when it came to big molecules, the rwf file tended to be hundreds of GB, so the calculations was not possible. Now, my question is:

Is it necessary to perform frequency calculations (like phonon calculations using phonopy) for the optimized geometry in CP2K? I mean the BFGS algorithm for geometry optimization is guaranteed to not stuck in the local minima? For example my system contains a quantum dot with many organic ligands (near 300 atoms) and the optimization is fully performed but I don't know if it would be acceptable to allege that this is the optimum geometry!  To sum up: when can we say that this is optimum geometry for this specific geometry configuration?

Thanks in advance.

[Kee Choon Wee]

Hi Shakiba,

The purpose of a frequency calculation is to verify that the stationary point you found is not a saddle point (1 imaginary frequency, i.e a TS) or higher-order saddle point with (>1 imaginary frequency).

I don't think it has anything to do with getting struck in a local minima, In fact, if you have no imaginary frequency the structure you found is a minima (could be a global/local one).

You can try a short MD to see if there is a large transition within say a picosecond, this might give you a hint as to whether it is a minima or transition state.

--
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 post to this group, send email to cp...@googlegroups.com.
Visit this group at https://groups.google.com/group/cp2k.
To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/9ff9bed8-3d0a-4ca5-b833-5435e3b9e54c%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Mohammad Shakiba]

Dear Kee Chon Wee,

Thank you so much for your help.

On Tuesday, June 11, 2019 at 1:54:09 PM UTC+4:30, Kee Choon Wee wrote:

Hi Shakiba,

The purpose of a frequency calculation is to verify that the stationary point you found is not a saddle point (1 imaginary frequency, i.e a TS) or higher-order saddle point with (>1 imaginary frequency).

I don't think it has anything to do with getting struck in a local minima, In fact, if you have no imaginary frequency the structure you found is a minima (could be a global/local one).

You can try a short MD to see if there is a large transition within say a picosecond, this might give you a hint as to whether it is a minima or transition state.

On Mon, Jun 10, 2019 at 8:32 PM Mohammad Shakiba <mshakiba....@gmail.com> wrote:

Hello everyone,

When optimizing with CP2K, we set a force limit and a maximum step size (for example less than 10meV and 0.002Bohr) and when the calculations reached those limits it is done. Before using CP2K I used Gaussian 09. When the geometry optimization was fully performed, I performed frequency calculation in order to check that there is no imaginary frequency so the optimization is not stuck in the local minima. However, when it came to big molecules, the rwf file tended to be hundreds of GB, so the calculations was not possible. Now, my question is:

Is it necessary to perform frequency calculations (like phonon calculations using phonopy) for the optimized geometry in CP2K? I mean the BFGS algorithm for geometry optimization is guaranteed to not stuck in the local minima? For example my system contains a quantum dot with many organic ligands (near 300 atoms) and the optimization is fully performed but I don't know if it would be acceptable to allege that this is the optimum geometry!  To sum up: when can we say that this is optimum geometry for this specific geometry configuration?

Thanks in advance.

--
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 cp...@googlegroups.com.

[Kee Choon Wee]

You are welcome.

Regards,

Choon Wee

To unsubscribe from this group and stop receiving emails from it, send an email to cp2k+uns...@googlegroups.com.

To post to this group, send email to cp...@googlegroups.com.
Visit this group at https://groups.google.com/group/cp2k.

To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/26eedeb9-3c38-4975-8e6f-c47e4b3ffd71%40googlegroups.com.