Van der Waals radius consideration during the insert-molecules

GROMACS version: 2020.2
GROMACS modification: Yes

Hello everybody,

I am inserting C60 and TAPC molecules onto a Graphene sheet. It is mentioned in GROMACS user guide in the gmx-insert that "Molecules are not inserted where the distance between any existing atom and any atom of the inserted molecule is less than the sum based on the van der Waals radii of both atoms. A database ( vdwradii.dat ) of van der Waals radii is read by the program, and the resulting radii scaled by -scale . If radii are not found in the database, those atoms are assigned the (pre-scaled) distance -radius".

I tried to insert the C60 molecules at a height of 0.52 nm from the Graphene sheet but it was not inserted.
I tried to insert it ad a height of 0.53 nm and it was successfully inserted but there is a bond between the C60 molecule and the Graphene sheet as shown in image 1.
I have inserted the C60 molecule at a height of 0.55 nm and it was inserted successfully without any bond as shown in image 2.
The van der waals radius for the C60 molecule is in the range of 0.5 to 0.55 depending on the reference.

Is it accepted to insert the molecules near the graphene sheet in a position where there is a bond between the C60 molecule and the graphene sheet as in image 1 ?

Should I consider the van der waals radius of the C60 molecule (0.55) while inserting the molecules by using the -radius 0.55 ?

I am also inserting TAPC (4,4’-cyclohexylidenebis(n,n-bis(p-tolyl)aniline)) molecules, with a united atom topology. I found that in some places the C60 and the TAPC molecules are very close to each other (almost joined) as in image 3. Is that accepted ?

The reason I suspected that may be I have a problem in the molecules insertion. Is that the energy minimization has stopped in step 9000 with a minimum energy more than of 4e+11 as shown in image 4 and the NVT equilibiration has stopped with segmentation fault at step 0.

Many Thanks

3 2 1

Those additional “bonds” you are seeing is most likely inserted by the visualisation software, it is not a bond contained within the topology. What software are you using? I do know that VMD guesses where bonds are based on the distance between atoms, it does not know about the topology.

I am using Pymol for visualization

The same principle applies - visualization software has no idea what the topology specifies as connectivity, so it guesses based on interatomic distance. Bonds rendered in VMD, PyMOL, etc. are not necessarily real. What is real is what is specified in the topology.