[gmx-users] Viscosity calculations

[Nicholas Michelarakis]

Dear everyone,

I am trying to use simulations to calculate the viscosity of a polymer
solution. Although I have some experience with MD, this is a new area for
me so apologies for the naive questions. Having gone through the mailing
list and after reading Hess' paper (2002), I am a bit confused about the
current best practices.

1. If I have understood correctly, the -vis flag in gmx energy is not the
best way to calculate viscosity due to its sensitivity to the system set-up
and it should be removed. I am assuming this is still the case?

2. Assuming one wants to use this method, what is the best (shifted)
potential and electrostatic cut-offs to use?

For using a non-equilibrium method, one has to use the acceleration options
in the mdp file to set up a shear flow, if I have understood it correctly.

3. When performing NEMD, can gmx energy and the -vis flag be used to
calculate the viscosity? The last line of the manual implies so (
http://manual.gromacs.org/documentation/2019-rc1/reference-manual/special/viscosity-calculation.html)
but just want to make sure, given the criticisms above. If not, this then
only leaves the gmx tcaf option?

4. When choosing groups to be accelerated to create a flow in acc-grps,
should one choose both the solvent and the solute (polymer in this case) or
just the solvent?

5. I think understand how the deform option works, but is this necessary,
and if so why, for NEMD?

Thank you very much in advance!

Best,
Nick