Dear all,

I’m having trouble with the temperature dependent mean square displacement msd. I want to calculate the diffusion coefficient for varying temperature of molecules with Knudsen conditions inside CNT forests.

However the diffusion coefficient doesn’t change with varying temperature!

My simulation starts with two overlying CNT’s, which are then displaced in x and y direction to get a periodic CNT forest in all 3 dimensions.

Then I follow the movement of 25 atoms, where forces between these atoms are turned off (Knudsen condition).

I have applied NVE ensemble with temp/rescale and an initial velocity according to the temperature, see below.

The calculated diffusion coefficient fluctuates without a clear trend between 1E-4 and 1E-5 for temperatures between -150 and 1300 °C.

Any help or comments are appreciated.

Here is a simplified input script:

dimension 3

boundary p p p

atom_style atomic

units metal

read_data run.data

group CNT1 type 1

**etc.**

neighbor 2 bin

neigh_modify delay 1

displace_atoms CNT1 move 9.6 0 0 units box

displace_atoms CNT2 move 0 9.6 0 units box

#Potential forces

pair_style hybrid lj/cut 15 airebo 2 1 1

pair_coeff * * airebo CH.airebo C C NULL NULL NULL NULL NULL

pair_coeff 3* 3* none

pair_coeff 1 3* lj/cut 0.005549 3.605

pair_coeff 2 3* lj/cut 0.005549 3.605

**velocity CNT create $t 12345**

**velocity CH4 create $t 131183**

**fix 1 all nve**

timestep 0.0005

**fix 8 all temp/rescale 40 ${t2} $t 0.1 1**

#coordination number and msd for every atom

compute 1 1 coord/atom ${cutoff}

compute 2 1 reduce sum c_1

variable comp equal c_2

variable step equal step

fix 9 1 msd 200 diff_1.txt

fix 10 1 print 60 “{step} {comp}” file comp_1.txt

**etc.**

run 4000000

Best Regards

Thomas Siegfried