Hi Vikas, It's a tough life, that of being a DFT guy... Sometimes we need zen-buddhist attitudes towards our calculations. :)
Now seriously. I see that you are using in-plane lattice parameters 25 ang long. I have used this much in an earlier work on carbon nanotubes, but this was because it was a charged cell. Of course the optimal length will depend on the diameter of your nanotube, but there are some things you can do to save time in your next calculations. 1) You can use a cell with hexagonal symmetry: 0.5000000 0.8660254 0.0000000 -0.5000000 0.8660254 0.0000000 0.0000000 0.0000000 0.xxxxxxx where xxxx is the value of c/a. The volume of such a cell is about 14% smaller than that of the cell you are using, so you get some saving there - by having a smaller mesh. Also, the symmetry will be closer to cylindrical (for isolated nanotubes) and to the symmetry of nanotube bundles, in case you want to study proximity effects. 2) If your MeshCutoff and Monkhorst-Pack grid are already well-converged, you can check how the total energy varies with the in-plane lattice vector length. The idea is to check how negligible the interaction between nanotubes will be. Generally the minimum distance between images (that is, the in-plane lattice vector lengths minus the diameter of your nanotube) would be between 7 and 10 angstroms, which might be enough to consider the system to be isolated. Nevertheless, you will have to check it by yourself, especially if you are going to study some kind of defect, charged or not. Maybe you'll find that your in-plane lattice parameters are overestimated and that you could have done it with a smaller one. 3) About the force tolerance. As I discussed recently in the list (and what I wrote is not necessarily authoritative, it's just my impressions on the subject), it will depend on what you want to get. Usually 0.04 eV/Ang is good enough, but some cases could need stricter tolerances. 0.01 eV/Ang is never bad, but I guess one of the reasons your calculation is time-consuming is because you seem to be performing variable-cell calculations. In VC calculations you have at least three extra degrees of freedom for the CG algorithm, respect to fixed-cell ones, which come from the lattice vectors. Since you have specified long in-plane lattice vectors, including these degrees of freedom in a VC calculation could make it longer and would be a waste of time: most probably your final in-plane lattice parameters are very close to the initial ones. Instead, it might take less time to perform a series of fixed-cell calculations varying the z-axis length, in which the positions of the current calculations are used as an input to the next one. In the end, you could fit the results to a Murnaghan equation of state, whose only meaningful results would be the equilibrium cell volume (from which you would extract the value of c) and total energy. Nothing that a little shell scripting could not handle. Best regards, Marcos > Dear Sir (Ricardo & Marcos) > Thank you for your response. Now it clear my doughts. One question i want > to ask is how much max. force tolerence is good enough for atomic > position relaxation . As i have specified it to be 0.01 eV/Ang but it is > very much time cosuming. > > Marcos Verissimo Alves <[EMAIL PROTECTED]> wrote: Vikas, > > The deviations of the angles are really so small that it won't really > matter if you change the angles to 90 degrees (think about it - ~1.e-4 > degrees... I guess that not even the best experimental measurements could > resolve such a small difference). However, if you still want to do it, > edit the .XV file and change the off-diagonal elements of the cell vectors > to zero, as you initially intended. In this way, you can use the XV file > directly in subsequent calculations. > > Marcos > >> Dear all, >> I relax my atomic coordinates for CNT up to max. force tolerance of 0.01 >> eV/Ang. The out cell is as below. I have CNT axis along Z axis. I need >> cell angles to be 90 90 90. Can i make change in this e.g. First line >> can i write 0.0000 25.420679 0.00000 and like wise for 2nd and and 3rd >> line. >> outcell: Unit cell vectors (Ang): >> 0.000458 25.420679 -0.000031 >> 25.427458 0.000458 0.000089 >> 0.000015 -0.000005 4.389456 >> outcell: Cell vector modules (Ang) : 25.420679 25.427458 >> 4.389456 >> outcell: Cell angles (23,13,12) (deg): 89.9996 90.0001 >> 89.9979 >> --------------------------------- >> Fussy? Opinionated? Impossible to please? Perfect. Join Yahoo!'s user >> panel and lay it on us. > > > -- > Dr. Marcos Verissimo Alves > Post-Doctoral Fellow > Condensed Matter and Statistical Physics Sector > International Centre for Theoretical Physics > Trieste, Italy > > -------- > > I have become so addicted to vi that I try to exit OpenOffice by typing > :wq! > > > > --------------------------------- > Boardwalk for $500? In 2007? Ha! > Play Monopoly Here and Now (it's updated for today's economy) at Yahoo! > Games. -- Dr. Marcos Verissimo Alves Post-Doctoral Fellow Condensed Matter and Statistical Physics Sector International Centre for Theoretical Physics Trieste, Italy -------- I have become so addicted to vi that I try to exit OpenOffice by typing :wq!
