Effect of long range forces in cnt interaction

[Zexter]

Dear all
       In a carbon nanotube with capped end facing each other the interaction is complex with a multitude of forces acting.So when calculated with DFT calculations the interaction energy shows only the energy contribution from only the non-dispersion corrected part.the contribution from the dispersion forces is missed out. Is there any possible way of including this long range dispersion forces so that I can effectively compute the interaction between the capped cnt ends.

The DFT calculations underestimate the equilibrium positions in such a configuration since the dispersion equations are not included.Is this opinion is correct?

[kstokbro]

It is correct that dispersion forces are not described correctly in DFT. However, DFT can in some cases give rise to semi long range interactions which by coincidence agree with the dispersion forces, see forinstance :

K. Stokbro, E. Nielsen, E. Hult, Y. Andersson, and B. I. Lundqvist, Nature of bonding forces between two hydrogen passivated silicon wafers, Phys. Rev. B. 58, 16118 (1998).

For the case of carbon materials you may check how well gga and lda describes the experimental interlayer distance in graphene. According to my memory lda does the best job, and you can expect the same level of accuracy when using lda to simulate your carbon system.

[Zexter]

The paper you are refering gives good idea, but still I have a doubt. The paper you mentioned deals only with planar structure, but the nanotube case, the structure is actually a hemisphere with a curvature coming into picture. But still is it possible to expect a DFT to give some correct result?

I tried to do but the atom count 0f 200 make the calculation time consuming.

[Zexter]

Dear all,
    I got the following values for the total energy (without subtracting the energy from the individual nanotubes).I think the minimum value is close around 3 Angstrom matching the MD result of 3.345 Angstrom.

But one thing that intrigues me is that the total energy values of the nanotube doesn't vary much especially left to the decimal point. Is this change acceptable or is it wrong. I have attached a screen shot of my system. The total energy of the system is plotted for various separation distances between the two nanotubes.

[Nordland]

In order to be sure, I think I would decrease the tolerance to 1e-6 or so, in order to see if this is properly converged.
What tolerance have you used ?

[Zexter]

Sir I got it.I will reduce the tolerance from my current value of 4e-5 to 1e-6 and reply back.

[Zexter]

I tried with a tolerance level of 1e-6, but it doesn't show any improvement in values for the total energy of the individual nanotubes. The value of total energy of an isolated nanotube is around -15711.2 for a tolerance level of 4e-5, and the same value -15711.2 for a level of 1e-6.The same value occurs even changing the tolerance limit. So is this because I am computing the energy of isolated nanotube the tolerance limit is not having any effect

[ugglebot]

I don't understand why you say it doesn't change. Clearly there is a minimum somewhere around separation 2.5-3 Ang. Of course the total energy will not vary much at all once they are completely separated, out around 5-6 Ang. Just focus on the area between 2 and 3 Ang and scan in steps of 0.1 Ang or small, and you will most likely have a very nicely defined minimum.

[Zexter]

I will do it. thanks all