Dear all,
I intend to find thermal conductivity of CNT (17,0) using Muller-Plathe method. I am using fixed boundary condition in all three directions. Though I am able to obtain the expected temperature profile along the tube (please find it in the attachment), I have a couple of questions to ask regarding the same.
- Why is there a steep temperature drop on the extreme left?
- The two linear regions have an appreciable difference in their slopes. Since the conductivity should be same, the partitioning of applied heat flux must be unequal (from Fourier’s law). Why is it so?
Note: Regarding (2), In some published studies, conductivity had been calculated using an average value of the two temperature gradients.
I would appreciate any help.
Regards,
Sudhansu Srivastava
Final Year Undergraduate
IIT Bhubaneswar
Not only is the temperature low at the left cold sink, your temperature is also too high at the heat sink: you can see sudden increases in temperatures when converging from the left and right towards the middle.
Paste your input script so that I can confirm my suspicion.
Ray
units metal
atom_style full
boundary s s s
read_data CNT.dat
timestep 2.0e-4
pair_style tersoff
pair_coeff * * BNC.tersoff C
dump config all custom 100000 CNT.config.* id type x y z vx vy vz
dump_modify config format line “%8d %5d %25.18g %25.18g %25.18g %20.13g %20.13g %20.13g”
thermo 1000
thermo_style custom step cpu temp etotal lx ly lz press
velocity all create 300.0 665424 sum no mom yes rot yes dist gaussian
fix 1 all nve
fix mom all momentum 1 linear 1 1 1 rescale angular
run 200000
unfix 1
unfix mom
fix 1 all nvt temp 300.0 300.0 1
fix mom all momentum 1 linear 1 1 1 rescale angular
run 200000
unfix 1
unfix mom
fix 1 all nve
fix mom all momentum 1 linear 1 1 1 rescale angular
run 200000
#Flux
fix hss all thermal/conductivity 100 z 40 swap 1
fix hc all ave/time 500 1 500 f_hss file flux.dat
run 100000
#Temperature gradient
compute ke all ke/atom
variable temp atom c_ke/(1.5*8.617343e-5)
compute cc1 all chunk/atom bin/1d z lower 2
fix 101 all ave/chunk 1 400000 400000 cc1 v_temp file temp.profile.dat
run 3000000
CNT.dat (130 KB)
BNC.tersoff (5.37 KB)
You need to record the temperature profile BEFORE you do the kinetic energy swaps: meaning the compute chunk/atom and fix ave/chunk commands must come BEFORE the fix thermal/conductivity command.
Ray