Biomolecules+metal simulation

LAMMPS LAMMPS Mailing List Mirror
1

Hi everyone!

I have a few questions about my dynamics.
My system consist of a biomolecule (generated by charmm2lammps tool) plus a silver cluster, solvated in water. I treated silver atoms with EAM, the biomolecule by lj/charmm/coul/long and between those a lj/cut, by hybrid/overlay pair style. For that, i’ve generated the solvated biomolecule data file by charmm2lammps and added manually the silver atoms positions. Because i assigned the EAM potential to silver atoms in my data file, i’ve put the lj/cut interaction in the input file, is there a better way to do it?

The thing is, when i run de dynamics, i get a warning: “WARNING: Using a manybody potential with bonds/angles/dihedrals and special_bond exclusions (…/pair.cpp:211)”. Is
that because i’ve assigned EAM to silver atom and then i tell LAMMPS to compute a two-body interaction with the biomolecule?
And other question is, in the case i want to put some charge state to my metal cluster (let’s assume the charge redistribute equally upon the atoms in the cluster), +5 for example, it is okey to asigne a simple EAM+coul interaction between silver atoms?

In general, the dynamic is okey? Waht should i check to see if i can trust the result?

Here is part of my input and data files:

----------------------------------------- INPUT --------------------------------------------------

Created by charmm2lammps v1.8.1 on jue may 21 18:06:18 ART 2015

processors 2 2 2 # nxnynz = N procesadores
units real
boundary p p p
neigh_modify delay 0 every 1 check yes
pair_style hybrid/overlay eam lj/charmm/coul/long 8.0 10.0 coul/long 10.0
atom_style full
bond_style harmonic
angle_style charmm
dihedral_style charmm
improper_style harmonic

read_data Green1_wb_ions.data

type 39 - Ag+

type 40 - Ag0 (o con alguna carga)

pair_coeff 40 40 eam Ag_u3.eam # eam Ag-Ag
pair_coeff 40 40 coul/long # coul Ag-Ag
pair_coeff 1 40 lj/charmm/coul/long 0.445729 1.50001
pair_coeff 2 40 lj/charmm/coul/long 0.810506 2.87529
pair_coeff 3 40 lj/charmm/coul/long 0.445729 1.50001
pair_coeff 4 40 lj/charmm/coul/long 0.445729 1.50001
pair_coeff 5 40 lj/charmm/coul/long 0.445729 2.27999
pair_coeff 6 40 lj/charmm/coul/long 0.445729 1.50001
pair_coeff 7 40 lj/charmm/coul/long 0.30825 2.47599
pair_coeff 8 40 lj/charmm/coul/long 0.347753 2.4938
pair_coeff 9 40 lj/charmm/coul/long 0.321957 2.4938
pair_coeff 10 40 lj/charmm/coul/long 0.657191 2.99271
pair_coeff 11 40 lj/charmm/coul/long 0.657191 2.99271
pair_coeff 12 40 lj/charmm/coul/long 0.657191 2.99271
pair_coeff 13 40 lj/charmm/coul/long 0.623466 2.99271
pair_coeff 14 40 lj/charmm/coul/long 0.623466 2.99271
pair_coeff 15 40 lj/charmm/coul/long 0.569144 2.99271
pair_coeff 16 40 lj/charmm/coul/long 0.569144 2.99271
pair_coeff 17 40 lj/charmm/coul/long 0.569144 2.99271
pair_coeff 18 40 lj/charmm/coul/long 0.293905 3.32679
pair_coeff 19 40 lj/charmm/coul/long 0.293905 3.32679
pair_coeff 20 40 lj/charmm/coul/long 0.491797 3.09071
pair_coeff 21 40 lj/charmm/coul/long 0.491797 3.09071
pair_coeff 22 40 lj/charmm/coul/long 0.580415 3.11743
pair_coeff 23 40 lj/charmm/coul/long 0.929408 2.94816
pair_coeff 24 40 lj/charmm/coul/long 0.929408 2.94816
pair_coeff 25 40 lj/charmm/coul/long 0.929408 2.94816
pair_coeff 26 40 lj/charmm/coul/long 0.929408 2.94816
pair_coeff 27 40 lj/charmm/coul/long 0.929408 2.94816
pair_coeff 28 40 lj/charmm/coul/long 0.929408 2.94816
pair_coeff 29 40 lj/charmm/coul/long 0.929408 2.94816
pair_coeff 30 40 lj/charmm/coul/long 0.929408 2.94816
pair_coeff 31 40 lj/charmm/coul/long 0.719917 2.81453
pair_coeff 32 40 lj/charmm/coul/long 0.719917 2.81453
pair_coeff 33 40 lj/charmm/coul/long 0.810506 2.87689
pair_coeff 34 40 lj/charmm/coul/long 0.719917 2.81453
pair_coeff 35 40 lj/charmm/coul/long 0.810506 2.87689
pair_coeff 36 40 lj/charmm/coul/long 0.810506 2.87689
pair_coeff 37 40 lj/charmm/coul/long 1.58953 3.21543
pair_coeff 38 40 lj/charmm/coul/long 0.450068 2.51496
pair_coeff 39 40 lj/charmm/coul/long 0.000000 0.00000

2

Hi everyone!

I have a few questions about my dynamics.
My system consist of a biomolecule (generated by charmm2lammps tool) plus a
silver cluster, solvated in water. I treated silver atoms with EAM, the
biomolecule by lj/charmm/coul/long and between those a lj/cut, by
hybrid/overlay pair style. For that, i've generated the solvated biomolecule
data file by charmm2lammps and added manually the silver atoms positions.
Because i assigned the EAM potential to silver atoms in my data file, i've
put the lj/cut interaction in the input file, is there a better way to do
it?

because you are using hybrid/overlay, you have to be careful and
monitor if all parameters that you expect to be generated through
mixing are actually generated. with hybrid overlay things can get
complicated rather quickly and then LAMMPS will give up quickly. i
would in this case prefer to have all parameters given explicitly in
the input file.

The thing is, when i run de dynamics, i get a warning: "WARNING: Using a
manybody potential with bonds/angles/dihedrals and special_bond exclusions
(../pair.cpp:211)". Is
that because i've assigned EAM to silver atom and then i tell LAMMPS to
compute a two-body interaction with the biomolecule?

that should be ok, for as long as you have no bonds declared between
your silver atoms.

And other question is, in the case i want to put some charge state to my
metal cluster (let's assume the charge redistribute equally upon the atoms
in the cluster), +5 for example, it is okey to asigne a simple EAM+coul
interaction between silver atoms?

it is most certainly an inconsistent model, since EAM will assume
neutral atoms and not adjust the charge density used for computing the
embedding energy according to the charge state of your cluster. how
far off this is, i cannot say. this requires very careful testing and
comparing against methods that can handle this case more accurately
(e.g. DFT or TB).

axel.

3

Along Axel’s lines you may want to keep in mind how well your model can mimic the process you intend to study. A metal cluster with open sites embedded in water and no explicit consideration of a polarizable model… If your reasons for the current choice are clear then great but is never bad practice to sit back and consider pros and cons. I understand models can get complicated quickly, that many times error cancellation occurs, that more complexity does not necessary add more insight yet boots computation effort, that not all models are implemented within lammps, etc. But like men, not all models are created equal and therefore will deliver proportionally to their capacities.

Carlos