Problem with Implementing DPD Thermostat

Dear Lammps users,

I am having some trouble with the implementation of DPD thermostat. I have tried looking at mailing archive for similar questions, But couldnt find something that was helping me in this context. I would really appreciate it if someone could give me comments on this. Below you will find the information about my system and Lammps input file I use.

__________ frozen wall ___________


__________ frozen wall ___________


units real
atom_style full
dimension 3
boundary p p f

pair_style hybrid/overlay lj/cut/coul/long 10.0 lj/cut 10.0 dpd/tstat 300.0 300.0 10.0 13245
comm_modify vel yes
bond_style harmonic
angle_style harmonic

kspace_style pppm 0.00001
kspace_modify slab 3.0
pair_modify tail yes

read_data data.atoms
include forcefield.SPCE
include forcefield.wall

pair_coeff * * dpd/tstat 38.0 10.0

fix fShake water shake 1.0e-4 100 0 b 1 a 1
fix rigidwall wall_atoms rigid single force 1 off off off torque 1 off off off
fix integrate water nve # I am integrating only the water and not the walls
fix constrain_COM all recenter INIT INIT INIT
run 100000


The temperature slowly increases like it would with regular NVE (without any thermostat). This problem comes only when I run simulations longer than 100ps after equilibration. Also, If i use Nose-Hoover for the same system, it works fine. Hence I assumed the issue is in my implementation of DPD thermostat.

I have used frictional coefficient and cutoff from Here they use Reaction field(RF) method for Electrostatics but I have seen other papers where they implement PPPM method too. So that shouldnt be the reason either.

Maybe I am missing something here, do you have any comments that could help me?

Please Let me know if you need more information about the system or if I am not clear in explaining myself.

Kind Regards,

Vishal Dadhich

PhD student


please remove fix recenter. it makes no sense in a system with immobilized atoms and it can cause unphysical relative motion of those immobilized atoms.