Hi,
I simulate a system once by Langevin thermostat and once by Nose-Hoover and I got different results for such as the size of the system. I know that the results by langevin thermostat are correct. In order to get the same results from Nose-Hoover, do I need to add any extra commands to the Lammps input script?
Thank in advance.
How do you define a “correct” result? Depending on the task you want to achieve, both thermostats could be an equally good choice. (Well, if you only need to pack molecules that are very far from thermal equilibrium, a Berendsen thermostat could be a very robust choice --although not if you need a proper statistical sampling).
You need to be more specific (eg, by sharing your input deck and elaborate on what are your goals) to get meaningful suggestions.
Thank you for the reply. For example, I calculate radius of gyration for a polymer from Monte Carlo and Langevin and I get the same results (within the error bars) but by Nose-Hoover, I find a different value (about 5% difference) which is not negligible. I need the results from llangevin and Nose-Hoover to be inconsistent with each other so I can compare my results. I used:
fix 1 all nvt temp T1 T2 Tdamp
fix 2 all recenter INIT INIT INIT
fix 3 all momentum 1 linear 1 1 1
This is insufficient information. It doesn’t tell us anything about the system or how you prepared it, how many atoms/bonds/etc there are, whether you have solvent or not and what kind, how you euquilibrated it, what kind of settings you use and over what time you collect data for your results.
But also, if your system has been properly set up and equilibrated you should need neither fix recenter nor fix momentum. Using fix momentum in every step almost guarantees to get bogus results with any kind of thermostat.
Thank you for the reply.