# Computing total tangential force for CNT filled with water

Dear All,

I am a beginner with Lammps and want to calculate a friction coefficient for a carbon nanotube (CNT) filled with water.

To do this, it needs to compute the autocorrelation of total tangential force acting along the axial direction of the CNT.

Hereby, my question is here: how is possible to compute total tangential force acting along the axial direction of CNT filled with water?

You can use the out-of-equilibrium method, i.e. impose a net flux of water in the CNT and measure the force exerted by the water on the CNT.

If you go for the equilibrium method, just extract the force exerted by the fluid on the CNT at regular interval and run a long simulation. Then you can simply do the autocorrelation using internal LAMMPS commands. However I am not sure that equilibrium procedures are a good idea for water/carbon interfaceâ€¦ the friction coefficient may be too small, I donâ€™t know if the result converges in a realistic time.

I would go for the out-of-equilibrium method, its an easy method to implement.

Simon

1 Like

In the case of equilibrium condition: Is it true to use the following command for extracting the perpendicular force excerted by the water on the CNT?

Fix wall H2O wall/region myCNT lj126 0.1132 3.27 4

No, this command is used to create a wall from a regionâ€¦

1 Like

Moreover, any comments regarding the command for calculating the perpendicular force excerted by the water on the CNT will be appreciated as well.

Hi @Nostalgic,

The force exerted on your CNT can be printed for each atoms with a `dump 1 myCNT custom N dumpfilename fx fy fz` and the total force can be obtained using `compute group/group` command. Assuming you have two groups `myCNT` and `water` (or any name) containing your atoms, this would give you the total force vector acting between the two.

The tangential/normal components will depend on your CNT orientation so there are no command for that. Concerning the simulation procedure, refer to the comments from @simongravelle.

As you said youâ€™re a beginner, I suggest you also take a good look at the how-to output section of the manual, especially the section concerning dump, global and local vectors and fixes that produces files. Also try to find more experimented users around you if you can, to ask them directly questions on MD and LAMMPS. Thatâ€™s the best way to go.

1 Like

Calculating the friction coefficient between water and a CNT is not trivial, and as a beginner you run a significant risk of burning up all your simulation time without sensible results at the end. I would strongly recommend that you look up a suitable paper and try replicating it first.

Obtaining the radial force between the water and CNT will also be very difficult. You will need to at least divide the CNT up into (at least) quadrants and then calculate a `compute force/force` between the water and each quadrant. That radial force will in turn be a small but consistent underestimate.

(To see why, consider a uniform radial force density acting on the interior of the CNT. If you integrate that force density along one quadrant, what will the total magnitude of the resultant force be? What if you integrate it along the entire CNTâ€™s interior circumference?)

1 Like