Help with LAMMPS : comp_chem

Thing I want to do –

1. Simulated Annealing of a Polymer

Methodology I am following –

1. Create a Polymer chain of desired length.

2. Use PACKMOL to pack desired number of chains in a conveniently huge box.

3. Using LAMMPS follow the general equilibration protocol, NVE, NVT, NPT dynamics.

4. Shrink the box size using NPT dynamics to get desired density. Output density using ‘thermo_style custom’ command.

Issues –

1. When I shrink the box size with ‘fix npt’ with Pstart and Pstop values as 1 1 and run the simulation for enough period of time, I am nowhere close to my desired density.

2. To overcome this I increase the pressure being applied on the system. With trial and error, I figure out that using 1GPa of pressure on the system brings me closer to desired density of the system.

3. Everything’s good. Pstart and Pstop are 1GPa and the density outputted at the end of the run is close to the expected density of the system.

4. Now when I try to equilibrate the system, the system volume increase and density drops down rapidly. Now I realise that is because the pressure on the system has been released.

5. A NPT at 1GPa followed by the NVT with no barostatting, it is obvious that system pressure will drop down and eventually lead to decrease in the density.

Problem –

1. I want to perform Simulated Annealing at atmospheric pressure and not 1GPa. How do I maintain the density of the system even after the pressures is released ?

2. Am I following the right protocol ? Or does it make no sense ?

3. Also, please point me towards understanding forcefields.

Edit – Providing more data, Polymer – Polyethylene. Force field – Dreiding Potential. Not a liquid system, just the polymer.