In NPT simulations with a deformable (tilted) box, will the orientation of rigid bodies change as the box shape evolves?

[miaoblah]

I found that HOOMD-blue supports simulating rigid bodies while allowing the simulation box to change shape with npt integrator. I know that when the box shape is modified (e.g., due to barostats), particle positions are usually rescaled accordingly. However, I’m unsure whether the orientation of rigid bodies also changes in this process.

Could you clarify whether the orientation changes or remains fixed when the box deforms, and explain the reasoning behind it?

[joaander]

Yes, orientation remains fixed when the box is deformed.

The ConstantPressure documentation references many papers which explain the how the constant pressure thermostat is derived:

• G. J. Martyna, D. J. Tobias, M. L. Klein 1994
• S. E. Feller, Y. Zhang, R. W. Pastor, B. R. Brooks 1995
• M. E. Tuckerman et. al. 2006
• T. Yu et. al. 2010