extrernal potential

LAMMPS LAMMPS Mailing List Mirror
1

Dear Users and Developers,

I would like to apply a fixed potential [v(x,y,z)] on my atoms which depends on the absolute x, y and z coordinates of the individual atoms. The potential can be expressed with simple functions (exp, cos, sqrt) of the Cartesian coordinates or can be tabulated.

[the potential is defined in eq 3. http://dx.doi.org/10.1038/nnano.2015.106]

Is there a way to apply this external potential to a system by defining analytically or by reading in tabulated form in LAMMPS?

Thank you for your kind answer in anticipation!

Best regards,

Janos

2

Dear Users and Developers,

I would like to apply a fixed potential [v(x,y,z)] on my atoms which depends on the absolute x, y and z coordinates of the individual atoms. The potential can be expressed with simple functions (exp, cos, sqrt) of the Cartesian coordinates or can be tabulated.

[the potential is defined in eq 3. http://dx.doi.org/10.1038/nnano.2015.106]

Is there a way to apply this external potential to a system by defining analytically or by reading in tabulated form in LAMMPS?

Yes. You can try fix addforce with atom style variables.

Axel

3

And if it needs to be faster, you could easily write a fix

which encodes your formulas in C code to add force

as a function of the coords of each atom.

Steve

4

Dear Steve and Axel,

Thank you for the nice solutions!

In case of the C code: is there some minimalistic example available somewhere
(eg. is there an already implemented feature written in the same way in the code)?

Just one remark to make the question more precise: I would need to apply the force on each atom of a given type.

So I hope to implement something like:
add_my_potential 3

where 3 is the atom type identifier.

Thanks in anticipation,
Janos

5

Any fix (there are many) which defines a post_force() method

is typically adding some force to some atoms. E.g. fix_efield.cpp/h.

In the post_force() method you simply need to do:

loop over atoms:

if atom not in group: continue

compute force on atom at x,y,z and type itype

add force to atom

store an energy contribution (if you wish)

Steve

6

Dear all, especially dear Axel and Steve,

I have posted here previously about adding external one body potential V(x,y,z) for a system.
There was two suggestion and in the end I took Axel’s suggestion to implement it through
fix addforce and atom-style variables.

When I tried to test it I found that:
a) either I do not understand how to print out the potential energy of the system
b) or fix addforce does not add the energy to the potential energy for some reason
c) I have messed up something in the syntax of adding the potential to the system through “energy”

Please help me find out which is the case! Please note that when I print out the potential
it gives reasonable values, but potential energy with the “pe” variable always gives 0.
I do not ask you to check the details of the potential it is only provided for illustration.

Please find the relevant input bellow.

Thank you for your help in anticipation!

Janos

#input:
include “system.in.init”

#(…)

PARAMETERS

variable pi equal 3.14159
variable b equal 4.83045
variable Ea equal 0.00698237
variable Ec equal 3.6749e-05
variable bb equal 0.423342
variable y0 equal 37.9585

#calculating the potential: PROBABLY NOT IMPORTANT ONLY FUNCTIONS OF x, y and z
variable Axz atom ({Ea}-{Ec}(6.0-4.0cos(2.0*${pi}x/b)*cos(2.0*{pi}z/b/sqrt(3.0))-2.0*cos(4.0*{pi}z/b/sqrt(3.0)))/9.0) variable By atom (exp(-2.0*{bb}(y-{y0})-2.0*exp(-{bb}(y-${y0}))))
variable Vxyz atom (v_Axzv_By)

#calculating derivatives: PROBABLY NOT IMPORTANT ONLY FUNCTIONS OF x, y and z
variable CuFx atom (-1.0*{Ec}*8.0/9.0*{pi}/b*cos(2.0*{pi}z/b/sqrt(3.0))*sin(2.0*{pi}x/b)*v_By) variable CuFy atom (-1.0*v_Axz*(-2.0*{bb}exp(-2.0{bb}*(y-{y0}))+2{bb}*exp(-1*{bb}(y-{y0})))) variable CuFz atom (-1.0*{Ec}8.0/9.0${pi}/sqrt(3.0)/b*sin(4.0*{pi}*z/sqrt(3.0)/$b)*v_By)

#Apply the force on the Xe atom: PLEASE CHECK THIS PART
fix field Xe addforce v_CuFx v_CuFy v_CuFz energy v_Vxyz

#Printing out quantities: PLEASE CHECK THIS PART
compute ener all reduce sum v_Vxyz
variable e equal c_ener
variable p equal pe
thermo 1
thermo_style custom pe c_ener #this gives 0 and some meaningful number for c_ener
fix extra all print 1 “{p} {e}” file ener.dat #this one does the same

#one note:
#the system is set up with no other potential then the one body V(x,y,z)
#eg in system.in.init

units electron
atom_style full
neigh_modify exclude type 1 1
kspace_style none
atom_modify sort 1 1.0
pair_style none

7

Dear all, especially dear Axel and Steve,

I have posted here previously about adding external one body potential
V(x,y,z) for a system.
There was two suggestion and in the end I took Axel's suggestion to
implement it through
fix addforce and atom-style variables.

When I tried to test it I found that:
a) either I do not understand how to print out the potential energy of the
system
b) or fix addforce does not add the energy to the potential energy for some
reason
c) I have messed up something in the syntax of adding the potential to the
system through "energy"

d) if you want a fix to add its energy to the total energy, you have
to use fix_modify energy.
it says so explicitly at the end of the documentation for fix addforce

axel.