# [lammps-users] I have several questions regarding the compute heat/flux

Deal All:

There’re some points that I do not understand regarding the command of compute heat/flux.
Let me describe it with the example given by the manual:

1. The atom_style should not be dqd, right? coz the new version of lammps does not include this type.

2. Why should the heat flux divided by volume “V” ?
variable J equal c_flux[1]/vol
thermo_style custom step temp v_J

Besides, there’s also a “V” in the numerator, so is the volume term cancelled after multiplying the heat flux with V/(kB T^2) ? According to some other articles, the formula of the thermal conductivity k should include the volume “V” at the denominator, but why it’s expressed different in Lammps?

1. I’ve tried to output the volume during the simulation. It’s 10290.076 in this case which is a little different from the domain defined by “region box block 0 4 0 4 0 4”. So what’s exact the meaning of the volume?

2. Is anybody can help me derive the exact coeff. of 2917703220.0 for me? In the manual, it said “The integral of the correlation needs to be multiplied by V/(kB T^2) times the sample interval and the appropriate unit” conversion factor". I appreciate it very much if someone can help me figure out how to derive this coeff.

#### Sample LAMMPS input script

``````atom_style      dpd
units 		real
dimension	3
boundary	p p p
lattice 	fcc  5.376  orient x 1 0 0 orient y 0 1 0 orient z 0 0 1
region  	box block 0 4 0 4 0 4
create_box 	1 box
create_atoms 	1 box
mass 		1 39.948
pair_style	lj/cut 13.0
pair_coeff	* * 0.2381 3.405
group 		every region box
velocity 	all create 70 102486 mom yes rot yes dist gaussian
timestep 	4.0
thermo	        10

``````
``````# ------------- Equilibration and thermalisation ----------------

``````
``````fix 		NPT all npt 70 70 10 xyz 0.0 0.0 100.0 drag 0.2
run 		8000
unfix           NPT

``````
``````# --------------- Equilibration in nve -----------------

``````
``````fix 		NVE all nve
run 		8000

``````
``````# -------------- Flux calculation in nve ---------------

``````
``````reset_timestep  0
compute	        myPE all pe/atom pair
compute 	flux all heat/flux myPE
log     	flux.log
variable        J equal c_flux[1]/vol
thermo_style 	custom step temp v_J
run 	        100000

``````

Answers below.

Steve

2010/3/7 Lingqi Yang <[email protected]...>:

Deal All:

There're some points that I do not understand regarding the command of
compute heat/flux.
Let me describe it with the example given by the manual:

1. The atom_style should not be dqd, right? coz the new version of lammps
does not include this type.

yes - that is outdated, it should be

atom_style atomic
communicate single vel yes

2. Why should the heat flux divided by volume "V" ?
variable J equal c_flux[1]/vol
thermo_style custom step temp v_J

As the doc page states:

The vector values calculated by this compute are "extensive". They
should be divided by the appropriate volume to get a flux. The vector
values will be in energy*velocity "units"_units.html.

Besides, there's also a "V" in the numerator, so is the volume term
cancelled after multiplying the heat flux with V/(kB T^2) ? According to
some other articles, the formula of the thermal conductivity k should
include the volume "V" at the denominator, but why it's expressed different
in Lammps?

Because V is not always an obvious value. If you invoke this compute
on a subset of atoms, what is V? It's up to you to answer that question,
not LAMMPS.

3. I've tried to output the volume during the simulation. It's 10290.076 in
this case which is a little different from the domain defined by "region
box block 0 4 0 4 0 4". So what's exact the meaning of the volume?

For a periodic system, it's the size of the box, in distance^3 units.
Your region
command appears to be in lattice units which are different.

4. Is anybody can help me derive the exact coeff. of 2917703220.0 for me?
In the manual, it said "The integral of the correlation needs to be
multiplied by V/(kB T^2) times the sample interval and the appropriate
unit" conversion factor". I appreciate it very much if someone can help me
figure out how to derive this coeff.

Ask Reese Jones (rjones at sandia.gov)