Bias-free temperature profile in the case of Couette flow

Dear Lammps Users,

I want to obtain the bias-free temperature profile for the rarefied gas confined between two solid walls, in the case of isothermal Couette flow. As you can see in the attached input script, to obtain the bias-free temperature profile I used two different approaches:

  1. Compute temp/chunk + fix ave/time
  2. Compute temp/profile + fix ave/time
    First of all, I was expecting that both methods should give the same results, but as you can see in the attached figure the results from compute temp/profile is always higher. I would be really thankful if someone could tell me why this is happening here.
    Besides that, since the system is isothermal, I was expecting that the temperature in the central part of the channel should be around the wall temperature (around 300 K). But as you see here using both approaches the temperature in the central part is strangely higher than the near-wall temperature. I highly appreciate it if someone could tell me what I am doing wrong here.
    ##------------------------------------------Input script----------------------
    units metal
    dimension 3
    boundary p p p
    atom_style atomic
    lattice fcc {latparamAu} region whole block 0 100 0 155 0 100 units box create_box 2 whole region Bottom_wall block INF INF 10 19 INF INF units box #Bottom Wall region Top_wall block INF INF 135 145 INF INF units box #Top wall mass 1 196.97 mass 2 39.95 create_atoms 1 region Bottom_wall create_atoms 1 region Top_wall group BW region Bottom_wall group TW region Top_wall region Bottom_Wall_fix block INF INF 10 11 INF INF units box region Top_wall_fix block INF INF 144 145 INF INF units box group BWfix region Bottom_Wall_fix group TWfix region Top_wall_fix timestep 0.001 #----------------------------------------------FORCE FIELD---------------------------------------------------------- pair_style hybrid eam lj/cut {rc}

    #-------------------------------------------------------Adding Argon atoms



If you read the doc pages for the 2 commands,
they explain exactly how they compute temperature,
and it can be different. Chunks are more general than
spatial bins. I suggest you take a single snapshot (no
time averaging), dump the particles and velocities,
and the results of the 2 computes. You can calculate
the temperature yourself from the particle velocities,
using the description on the 2 doc pages. See
if you get the same answers as the 2 computes.