Greetings,
I am trying to model Al2O3 nano-powder using Reax potential file using Al/AlO force field. I am also trying to fix temperature at “300 K” and fix pressure at “1 atm”. However the problem is that there is huge pressure fluctuations (like form -1000 atm to +1000 atm or even greater). I have also tried to use “this” potential for just modeling a cubic AL to find the problem, and I performed a NVT. (with a 0.2 fs timestep and 20 fs temperature damping parameter). In this case temperature is fixed after 5000 fs (5 ps) but pressure fluctuate around 2500 atm. How should I fix pressure while using Reax potential.
Thanks for your help.
Try to adjust pdamp properly.
temperature and pressure fluctuate at the size scale of MD simulations.
expecting them to be “fixed” when using a thermostat or barostat is a misinterpretation of what these fixes are doing.
it is well known (and has been discussed on this very mailing list many times), that pressure fluctuates much more than temperature, and pressure fluctuations depend on the compressibility of substances (so very hard, solid materials can have very large pressure fluctuations). both also depend on the size of the system. in fact, temperature and pressure are only well defined (and the macroscopic sense) in the limit of a very large system (where local(!!) fluctuations become negligible and average out).
trying to “correct” this through more aggressive choices of thermostat/barostat algorithms and settings, is rarely the correct way to deal with this situation. you are “manipulating” your dynamics and are likely to invalidate your results with it. it is more important to first determine, whether the (sufficiently long) time averaged property is converged to the desired value. …and before that one usually has to determine which are the proper MD simulation settings from running an equilibrated system without any thermostatting or barostatting to be able to properly monitor energy conservation.
axel.