Dear lammps community,I have a puzzle about the water model of SPC/E.In my simulation ,when I use "pair_coeff lj/cut/coul/cut" to calculate the coulombic pairwise interactionhen ,the self diffusion of water is 0.6X10-5 cm2/s ,the self diffusion of water is unreasonable. And the simulation speed is fast.Here is my input file:dimension 3 boundary p p p atom_style full units metal pair_style lj/cut/coul/cut 9 9 bond_style harmonic angle_style harmonic neighbor 2.0 bin neigh_modify every 10 read_data D9.86onewater.data #kspace_style pppm 1.0e-6 minimize 1.0e-8 1.0e-12 1000 10000 group water type 1 2 velocity all create 300.0 4928459 rot yes mom yes velocity all zero linear velocity all zero angular compute stress_all all stress/atom NULL compute atom_pe all pe/atom compute 1 water temp compute 2 water msd com yes variable Mises atom 0.707106781186548*sqrt((c_stress_all[1]-c_stress_all[2])^2+(c_stress_all[2]-c_stress_all[3])^2+(c_stress_all[3]-c_stress_all[1])^2+6*((c_stress_all[4])^2+(c_stress_all[5])^2+(c_stress_all[6])^2)) fix 2 water nvt temp 300.0 300.0 0.5 run_style verlet fix 3 water shake 0.0001 20 0 b 1 a 1 timestep 0.002 dump 1 all custom 250 d9.86onewater.lammpstrj id type x y z v_Mises c_atom_pe c_stress_all[1] c_stress_all[2] c_stress_all[3] c_stress_all[4] c_stress_all[5] c_stress_all[6] thermo_style custom step temp c_1 c_2[4] thermo 25 reset_timestep 0 run 10000While, when I use "pair_coeff lj/cut/coul/long" to calculate the coulombic pairwise interactionhen ,the self diffusion of water is 2.92X10-5 cm2/s ,the self diffusion of water is reasonable.But the simulation speed is slow.Here is my input file:dimension 3 boundary p p p atom_style full units metal pair_style lj/cut/coul/long 9 9 bond_style harmonic angle_style harmonic neighbor 2.0 bin neigh_modify every 10 read_data D9.86onewater.data kspace_style pppm 1.0e-6 minimize 1.0e-8 1.0e-12 1000 10000 group water type 1 2 velocity all create 300.0 4928459 rot yes mom yes velocity all zero linear velocity all zero angular compute stress_all all stress/atom NULL compute atom_pe all pe/atom compute 1 water temp variable Mises atom 0.707106781186548*sqrt((c_stress_all[1]-c_stress_all[2])^2+(c_stress_all[2]-c_stress_all[3])^2+(c_stress_all[3]-c_stress_all[1])^2+6*((c_stress_all[4])^2+(c_stress_all[5])^2+(c_stress_all[6])^2)) fix 2 water nvt temp 300.0 300.0 0.5 run_style verlet fix 3 water shake 0.0001 20 0 b 1 a 1 timestep 0.002 dump 1 all custom 250 d9.86onewater.lammpstrj id type x y z v_Mises c_atom_pe c_stress_all[1] c_stress_all[2] c_stress_all[3] c_stress_all[4] c_stress_all[5] c_stress_all[6] thermo_style custom step temp c_1 thermo 25 reset_timestep 0 run 10000So I want to know why pair_coeff of"lj/cut/coul/cut"don't have a reasonable self diffusion and pair_coeff of"lj/cut/coul/long"don't have a fast calculation speed.Thanks for any advices!Best regards, Zhongzhong,Cui
When you use lj/cut/coul/long, you are using a kspace style (PPPM) and are including the long-range Coulombic interactions. This is more expensive but also can be more realistic. When you use lj/cut/coul/cut, you are neglecting the long-range interactions (i.e. no kspace style), so it runs faster but gives a different answer.