NPT box size not reaching equilibrium with multiple atomic constraints

[Sean Larmore]

Hi CP2K community, 

I am trying to run an npt simulation on a molecule containing Boron with the generalized amber force field (GAFF). GAFF does not contain parameters for Boron, so I had set up these simulations to keep the Boron atom (and it's connected bonds, angles, and dihedrals) constrained/frozen. However, I quickly noticed that the box size would not equilibrate and hovered around the initial box size, leading to significant void spaces in my solvent, poor solvent density, and solvent distribution. 

I set up some control tests on a different molecule without Boron, so I knew the FF would not cause issues, and found that the issue seems to be caused by atomic constraints. As you can see in the attached graphs, with no constraints the edge length of the box reaches an equilibrium value of ~23A, with one constraint it also goes to ~23A, with two constraints it goes to ~24.5A, and with three constraints it goes to ~26.5A (3.5A larger than the box without atomic constraints). This essentially makes it impossible to have more than 1 or 2 constraints on a system as it goes through an npt, as the box size will never reach the correct equilibrium size. This also precludes me from my goal of simulating the boron containing molecule I'm currently interested, as I need a minimum of 7 constraints. 

Does anyone have any insight into this problem with npt simulations and constraints and/or a potential work around in CP2K? I've also attached my input and output files for one of the constrained jobs. 

Thanks in advance,

Sean

[Krack Matthias]

Hi Sean

 

Are you sure that the FF is set up properly, in particular, did you check if the values for EI_SCALE14 and VDW_SCALE14 are correct?

 

Best

 

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 visit https://groups.google.com/d/msgid/cp2k/6768d237-6879-45e0-985f-55a250183f41n%40googlegroups.com.

[Sean Larmore]

Hi,

Sorry for the delay in response here. I did add EI_SCALE14 and VDW_SCALE14 values, which I had not included before, however this does not change the ultimate outcome which is that NPT simulations with large numbers of constraints fail to converge to the correct box size for the system. Since I am only changing the number of constraints in my tests, I doubt that this issue has anything to do with the FF parameters/scaling. As I mentioned above, without constraints the box size contracts to 23A, while with 3 constraints (this being the only difference between my tests) it only contracts to 26.5A.

Any further help/insight would be appreciated. 

Thanks,

Sean