# defect in GNR

Dear LAMMPS users,
I got thermal conductivity in GNR with L=75 nm and width=15nm equal 2973.9881803332687[W/mK] @ 300 k .but when i apply defect , the thermal conductivity will increase instead of reduse .the defect is delete 456 atom carbon in cube region
i must change values of Nevery,Nrepeat, when i apply defect?. I don’t know what numbers are true for these?

units metal
variable T equal 300
variable V equal vol
variable p equal 2000 # correlation length
variable s equal 2 # sample interval
variable d equal \$p*\$s # dump interval
variable d equal 2000*\$s
variable d equal 2000*2
variable dt equal 0.001

# convert from LAMMPS real units to SI

variable kB equal 1.3806504e-23 # [J/K] Boltzmann
variable ev2J equal 1.60217653e-19
variable A2m equal 1.0e-10
variable ps2s equal 1.0e-12
variable convert equal {ev2J}*{ev2J}/{ps2s}/{A2m}
variable convert equal 1.60217653e-019*{ev2J}/{ps2s}/{A2m} variable convert equal 1.60217653e-019*1.60217653e-019/{ps2s}/{A2m} variable convert equal 1.60217653e-019*1.60217653e-019/9.9999999999999998e-013/{A2m}
variable convert equal 1.60217653e-019*1.60217653e-019/9.9999999999999998e-013/1e-010

# setup problem

dimension 3
atom_style atomic
boundary p p p
orthogonal box = (0 -100 0) to (50 100 50)
1 by 1 by 1 MPI processor grid
44064 atoms

region right block 5 25 60 65 0 2 units box

group right region right
456 atoms in group right

mass 1 12.000
pair_style tersoff
pair_coeff * * sic.tersoff C

timestep 0.001
thermo \$d
thermo 4000

# equilibration and thermalization

velocity all create \$T 1234 mom yes rot yes dist gaussian
velocity all create 300 1234 mom yes rot yes dist gaussian

delete_atoms region right
Deleted 456 atoms, new total = 43608

fix NVT all nvt temp \$T \$T 0.001 drag 0.02
fix NVT all nvt temp 300 \$T 0.001 drag 0.02
fix NVT all nvt temp 300 300 0.001 drag 0.02

dump 1 all xyz 200 dump.xyz
run 4000
Memory usage per processor = 65.0334 Mbytes
Step Temp E_pair E_mol TotEng Press
0 300.16976 60020156 0 60021848 1.2682975e+008
4000 300.38051 -257973.14 0 -256280 175939.16
Loop time of 5376.78 on 1 procs for 4000 steps with 43608 atoms

Pair time () = 5348.28 (99.47) Neigh time () = 15.5844 (0.289847)
Comm time () = 2.23081 (0.0414896) Outpt time () = 2.4024 (0.0446811)
Other time (%) = 8.27801 (0.153958)

Nlocal: 43608 ave 43608 max 43608 min
Histogram: 1 0 0 0 0 0 0 0 0 0
Nghost: 27194 ave 27194 max 27194 min
Histogram: 1 0 0 0 0 0 0 0 0 0
Neighs: 0 ave 0 max 0 min
Histogram: 1 0 0 0 0 0 0 0 0 0
FullNghs: 2.0361e+006 ave 2.0361e+006 max 2.0361e+006 min
Histogram: 1 0 0 0 0 0 0 0 0 0

Total # of neighbors = 2036102
Ave neighs/atom = 46.691
Neighbor list builds = 116
Dangerous builds = 2

# thermal conductivity calculation, switch to NVE if desired

#unfix NVT
#fix NVE all nve
reset_timestep 0

compute myKE all ke/atom
compute myPE all pe/atom
compute myStress all stress/atom virial
compute flux all heat/flux myKE myPE myStress
variable Jx equal c_flux[1]/vol
variable Jy equal c_flux[2]/vol
variable Jz equal c_flux[3]/vol
fix JJ all ave/correlate \$s \$p \$d c_flux[1] c_flux[2] c_flux[3] type auto file J0Jt.dat ave running
fix JJ all ave/correlate 2 \$p \$d c_flux[1] c_flux[2] c_flux[3] type auto file J0Jt.dat ave running
fix JJ all ave/correlate 2 2000 \$d c_flux[1] c_flux[2] c_flux[3] type auto file J0Jt.dat ave running
fix JJ all ave/correlate 2 2000 4000 c_flux[1] c_flux[2] c_flux[3] type auto file J0Jt.dat ave running

variable scale equal {convert}/{kB}/\$T/\$T/\$V*s*{dt}
variable scale equal 2.566969633282841e-016/\${kB}/\$T/\$T/\$V*s*{dt}
variable scale equal 2.566969633282841e-016/1.3806504000000001e-023/\$T/\$T/\$V*s*{dt}
variable scale equal 2.566969633282841e-016/1.3806504000000001e-023/300/\$T/\$V*s*{dt}
variable scale equal 2.566969633282841e-016/1.3806504000000001e-023/300/300/\$V*s*{dt}
variable scale equal 2.566969633282841e-016/1.3806504000000001e-023/300/300/500000*s*{dt}
variable scale equal 2.566969633282841e-016/1.3806504000000001e-023/300/300/5000002{dt} variable scale equal 2.566969633282841e-016/1.3806504000000001e-023/300/300/500000*2*0.001 variable k11 equal trap(f_JJ[3])*{scale}
variable k11 equal trap(f_JJ[3])8.2633184517250094e-007
variable k22 equal trap(f_JJ[4])
{scale} variable k22 equal trap(f_JJ[4])*8.2633184517250094e-007 variable k33 equal trap(f_JJ[5])*{scale}
variable k33 equal trap(f_JJ[5])*8.2633184517250094e-007
thermo_style custom step temp v_Jx v_Jy v_Jz v_k11 v_k22 v_k33
run 8000
Memory usage per processor = 78.2298 Mbytes
Step Temp Jx Jy Jz k11 k22 k33
0 300.38051 0.009232223 0.002334846 -0.014331305 8.8039399 0.56309409 21.214654
4000 299.58903 -0.00084436811 -0.0014041688 -0.0084586756 99.231854 -1.7423237 1181.3462
8000 299.42427 0.0016410622 0.0047518503 -0.003760334 352.15429 181.80008 17456.413
Loop time of 4967.16 on 1 procs for 8000 steps with 43608 atoms

Pair time () = 4917.86 (99.0075) Neigh time () = 11.4036 (0.22958)
Comm time () = 3.55681 (0.0716065) Outpt time () = 4.82041 (0.0970456)
Other time (%) = 29.5196 (0.594296)

Nlocal: 43608 ave 43608 max 43608 min
Histogram: 1 0 0 0 0 0 0 0 0 0
Nghost: 21630 ave 21630 max 21630 min
Histogram: 1 0 0 0 0 0 0 0 0 0
Neighs: 0 ave 0 max 0 min
Histogram: 1 0 0 0 0 0 0 0 0 0
FullNghs: 1.95125e+006 ave 1.95125e+006 max 1.95125e+006 min
Histogram: 1 0 0 0 0 0 0 0 0 0

Total # of neighbors = 1951252
Ave neighs/atom = 44.7453
Neighbor list builds = 93
Dangerous builds = 0
variable k equal (v_k11+v_k22+v_k33)/3.0
variable ndens equal count(all)/vol
print “average conductivity: \$k[W/mK] @ T k, {ndens} /A^3”
average conductivity: 5996.7889867439699[W/mK] @ 300 k, 0.087216000000000002 /A^3

Dear LAMMPS users,
I got thermal conductivity in GNR with L=75 nm and width=15nm equal
2973.9881803332687[W/mK] @ 300 k .but when i apply defect , the thermal
conductivity will increase instead of reduse .the defect is delete 456
atom carbon in cube region
i must change values of Nevery,Nrepeat, when i apply defect?. I don't
know what numbers are true for these?

​that is the smallest of your problems. first you have to learn about
statistics and convergence.

axel.​

​you quote a number (2973.9881803332687) with 18 digits of accuracy. do you
think that your result is *that* accurate? if not, what is the
(statistical) error on it? ...and how do you know the uncertainty of this
kind of result?

i compare my answers with paper answer(control of thermal and electronic transport in defect-enginered graphene nanoribbons)
the answer was 2300W/mK .and i found that answer is corect. do u mean i have to the second part of program in NVE mode??
thanks alooooot

transport in defect-enginered graphene nanoribbons)

​i don't understand what you are saying here. *what* do you compare to
what? *how* do you assess correctness of what?

do u mean i have to the second part of program in NVE mode??

​this doesn't answer any of the questions i posed and confirms that you
have not spent any time worrying about these issues, but that you should.
there have been extensive discussions on this very mailing list about that
subject convergence and accuracy in general and with special reference to
computing thermal conductivity.

there is no point in continuing this discussion until you have studied this
and the necessary background information from the adequate text books. this
has nothing to do with some detail of the simulation, but is a very general
concern that you currently are ignoring.

axel.

Where should we know the system has converged ?The output should be how do I know it was right?
these values( c_flux[1]*c_flux[1] c_flux[2]*c_flux[2] c_flux[3]*c_flux[3]) ​​must be what I am to understand correctly the answer?

I know I ask lots of questions, really sorry but I do not anyone to ask questions and engage with these issues in mind…sorry

Where should we know the system has converged ?

​that is exactly why i suggested that you should update your knowledge
about statistics and how you determined the statistical error of
measurements.​

The output should be how do I know it was right?

​how so? the two numbers you quoted where significantly different. also,
have you considered that the numbers could be "right" just by chance.
again, that refers back to my previous statement.​

these values( c_flux[1]*c_flux[1] c_flux[2]*c_flux[2]
c_flux[3]*c_flux[3]) ​​must be what I am to understand correctly the answer?

I know I ask lots of questions, really sorry but I do not anyone to ask
questions and engage with these issues in mind...sorry

​i don't understand what this is supposed to mean. how about actually
following the suggestions you were given?​ is it too much effort?

axel.