temperature profile

Dear LAMMPS user,

I am using LAMMPS-11Aug2017. I want to model the heat conduction through the liquid argon confined between two parallel walls according to the following paper:

“M. Barisik, A. Beskok, Boundary treatment effects on molecular dynamics simulations of interface thermal resistance, J. Comput. Phys. 231 (2012) 7881–7892. doi:10.1016/j.jcp.2012.07.026."

I have written the code according to the paper to be able to compare the result.

I have modeled the Case 2 of Table 2 on the fourth page of the paper where the temperature of the walls are 140 and 90.

The wall in our code is the same as what was used in the paper. Each wall consists of 23 layers of the atom. Actually, the outmost layer of each wall is fixed. On the three next, the Langevin thermostat is applied and the rest layers are free to interact.

It is expected that the temperature distribution is the same as the yellow diamond in the Fig3 of the paper. It should be a temperature jump between the thermostated region and other wall layers. In the third page of the paper, it is observed that Langevin damping factor is considered as 0.1.

I have used the same damping factor in the lines 181 and 189 of the code. It is observed that for 0.1, there is no temperature jump between the thermostated region and other wall layers. As we decrease the Langevin damping factor to 0.01 and 0.003, the temperature jump is appeared but is not the same as the paper as it is shown in the “temperature profile” distribution.

Consequently, the heat flux is different from what is reported in Table 2 of the paper. It should be around “5.91 E+08” but it is 4.16, 4.32 and 3.92 for the damping factor of 0.1, 0.01 and 0.003 respectively.

It seems that the inaccuracy in the prediction of temperature distribution leads to this differences between the calculated heat fluxes.

I will be really grateful if someone would check the code to see where the problem arises. I have checked it several times but unfortunately, I have not found any mistake. Maybe I have used something wrong and I am not aware of that.

The code for damping factor 0.1 and its corresponding averaging output is attached to the email.



AgAr0.1.txt (4.68 KB)

averaging1.txt (478 KB)


It seems like you’ve done a careful job of replicating
what is in the paper. I don’t think the mail list
is going to do any better. I suggest you contact
the authors and ask their advice. If they used
LAMMPS, maybe then can send you their script.