Correct TIP3P parameters

[Léon Luntadila Lufungula]

Dear all,

Just a quick question about the parameters for the TIP3P forcefield. In the example files (see attached water.inp) of the exercise on molecular simulation of water the TIP3P model is used to simulate water. In the input file, I noticed that the force constant of the H-O-H bend and the H-O bond are 110 rad^-2kcalmol and 900 angstrom^-2kcalmol, respectively. However, when looking online for the parameters like on the lammps documentation, I find that these values are actually half of what is defined in the input, i.e., 55 rad^-2kcalmol and 450 angstrom^-2kcalmol, respectively... Is this indeed an error in the input file or am I missing something here? Most water input files I see in the tests/Fist folder also contain the smaller values for the force constants.

Kind regards,

Léon

[Alin Marin Elena]

Hi Leon,

Lammps and cp2k use different definitions for constants. K Vs K/2 .

Regards,

Alin

--
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/2bed3d68-2c9d-4ceb-8672-208cfe7d0fd8n%40googlegroups.com.

[Krack Matthias]

Dear Léon

 

Note that LAMMPS uses a different potential definition. By contrast to CP2K, LAMMPS includes the prefactor ½ into the constant K.

 

HTH

 

Matthias

 

--

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/2bed3d68-2c9d-4ceb-8672-208cfe7d0fd8n%40googlegroups.com.

[Léon Luntadila Lufungula]

Hi Alin and Matthias,

Thanks for the quick reply, I didn't know that! If so, the 110/900 combination in the water.inp file should be correct then right? Still, in a lot of input files in the tests folder like for example cp2k/tests/Fist/regtest-4/H2O-1.inp (see attachment) the combination 55/450 is used. Can you explain why this is the case then?

Kind regards,

Léon

[Krack Matthias]

Hi Léon

 

Yes, 110 kcalmol*rad^-2 and 900 kcal*Angstrom^-2 are correct. Note, that the inputs in cp2k/tests are for code regression testing and the parameters used therein are not reliable by contrast to the example inputs given in the exercises. Unfortunately, many regression tests use this incorrect flexible TIP3P force field. This is most likely a copy-paste error which is indeed confusing for users. I will fix that.

 

Best

 

Matthias

 

.
To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/06168420-2bcc-4588-bd81-05bfcf6ac9den%40googlegroups.com.

[Léon Luntadila Lufungula]

Hi Matthias,

Thanks for clarifying this for me. It was indeed a bit confusing, but I'm happy that I have a clear answer now!

Kind regards,

Léon

[Krack Matthias]

Dear Léon

 

I have to correct my statement concerning the regression test inputs. The &BEND and &BOND section parameters in these inputs (e.g. in Fist/regtest-4/H2O-1.inp) are correct for the flexible TIP3P model, because the default value for the KIND keyword in these sections is CHARM and not HARMONIC as I wrongly assumed. So you have to take care which KIND of forcefield you are using when you set these parameters.

 

Best regards

 

Matthias

 

.
To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/591a7462-cdc1-4525-bd9a-9815690a32b4n%40googlegroups.com.

[Léon Luntadila Lufungula]

Dear Matthias,

Thanks for clarifying this for me, I'll keep it in mind in the future when setting up my forcefield parameters.

Kind regards,

Léon

.
To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/591a7462-cdc1-4525-bd9a-9815690a32b4n%40googlegroups.com.

[Krack Matthias]

Dear Léon

 

To avoid future confusion, I have added &FORCEFIELD sections for common 3-site water models to the cp2k database (see cp2k/data/forcefield_section/H2O) and the corresponding template files for 6x6x6 = 216 H2O water boxes are available as new regression tests. These template files can easily be modified (just adapt the @SET parameter na, nb, and nc lines in the top of each file) to generate arbitrary orthorhombic water boxes which can be equilibrated quickly.

 

Best

 

Matthias

 

To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/dc1bea19-4faf-4fa9-afed-303e9c1fb923n%40googlegroups.com.

[Léon Luntadila Lufungula]

Dear Matthias,

Thanks for adding this to the CP2K database! I think this will help a lot of people and avoid confusion about how to correctly implement these water models in a CP2K calculation.

Kind regards,

Léon