Dears
I have a problem about using Buckingham potential. The input as follows:
echo both
boundary p p p
atom_style atomic
units metal
read_data 32Z180.data
mass 1 114.818001 # In
mass 2 15.999400 # O
mass 3 118.709999 # Sn
group In type 1
group O type 2
group Sn type 3
pair_style buck 2.5
pair_coeff 1 2 2719.77 3.4281 0.0 # A, rho, C
pair_coeff 2 3 2180.89 3.25097 0.0
pair_coeff 1 3 0.0 1.5 0.0
pair_coeff 1 1 0.0 1.5 0.0
pair_coeff 2 2 0.0 1.5 0.0
pair_coeff 3 3 0.0 1.5 0.0
neighbor 2.0 bin
As you can find, I have a system that has included In, O and Sn. There are not any connection between 1 - 3 types as well as 1 - 1 types also 2 - 2 and 3 - 3 a well. It means we have only connections between In and O as well as Sn and O. In addition, since rho must be greater than 1, I set it as 1.5. The parameters of A are 0.0 for others instead of 1-2 and 2-3. Here, A, rho and C are Buckingham parameters. The run has not any error.
One wonder thing is the inputs for pair_coeffs are correct or NOT!
Best,
Muhammad
Please just look at the formula for the buckingham potential. What is its value when you set A and C to zero? And how does that depend on rho?
Now whether your model is correct (i.e. only having interactions between some atom types and not between others), I cannot tell. That is a matter of researching the published literature or checking whatever source you got the parameters you have from.
Also I am wondering if the 2.5 angstrom cutoff is correct. It is very short for metal units. A recommended cutoff should be given with the parameters that you have looked up.
Axel.
Dear Axel,
Many thanks
I understand your explanation completely.
So, the Buckingham parameters are correct in my research.Also, the cut-off is 10.0 A according other papers.
Now, the problem here is about “temperature”.
Starting simulation, temperature is 0 while it has set 300K for equilibrium part.
Please see the output:
Step Temp E_pair TotEng E_bond KinEng Press Volume Enthalpy c_Pvirial c_Pke c_pe_In c_pe_O c_pe_Sn c_MSDeq[1] c_MSDeq[2] c_MSDeq[3] c_MSDeq[4] Lx Ly Lz Atoms
0 0 2.3397455e+08 2.3397455e+08 0 0 4199894.8 19511816 2.8512221e+08 4199894.8 0 88241163 1.1698728e+08 28746113 4.0211895e-26 3.3229168e-27 4.2635997e-25 4.6989478e-25 329.24 329.24 180 81920
2000 0 0 0 0 0 0 27708326 0 0 0 0 0 0 139.15373 139.17071 0.34361753 278.66806 370.06964 370.06964 202.32212 81920
4000 0 0 0 0 0 0 27707647 0 0 0 0 0 0 139.13313 139.15011 0.3435839 278.62683 370.06662 370.06662 202.32047 81920
6000 0 0 0 0 0 0 27706968 0 0 0 0 0 0 139.11254 139.12951 0.34355028 278.5856 370.0636 370.0636 202.31882 81920
8000 0 0 0 0 0 0 27706290 0 0 0 0 0 0 139.09194 139.10891 0.34351666 278.54437 370.06058 370.06058 202.31717 81920
The inputs are as follows:
#equilibration ========================================================================
fix firstEQ all npt temp 300.0 300.0 0.010 iso 1.0 1.0 1
timestep 0.001
thermo 2000
thermo_style custom step temp epair etotal ebond ke press vol enthalpy c_Pvirial c_Pke c_pe_In c_pe_O c_pe_Sn c_MSDeq[1] c_MSDeq[2] c_MSDeq[3] c_MSDeq[4] lx ly lz atoms
fix avequ all ave/time 1 1000 1000 v_Temp v_Pe v_Press v_Enthalpy v_Etotal v_Eperatom c_Pvirial c_Pke c_pe_In c_pe_O c_pe_Sn c_MSDeq[1] c_MSDeq[2] c_MSDeq[3] c_MSDeq[4] file avequ.txt
dump firstEQ all xyz 1000 equilibrium.xyz
run 100000
unfix firstEQ
unfix avequ
undump firstEQ
you need to look at what happens during the first 2000 time steps.
two things are very obvious to note: 1) your initial potential energy is very large, 2) your volume increases almost 7-fold
that means your system was initially not in a geometry that is consistent with the potential parameters and settings and then it changed a lot.
please also note that the shape of the buckingham potential is such that when you have a bad geometry things can get very bad.
aren’t there also charges associated with your atom types? if yes, are they considered in your pair style?
please note that this is not leaving the realm of what is specific to LAMMPS and entering general MD simulation methodology and understanding.
that is something that you need to consult with experienced colleagues and textbooks. it is considered off-topic for this mailing list, which is about the “technical” side of using a modifying LAMMPS and only peripherally about understanding MD and doing it well.
axel.