Dear experts,
Please find attached structures of doped graphene during and after geometry optimization. I have optimized graphene supercell with substitutional doping in the hexagonal ring of the graphene with different supercell sizes, and different input parameters in the INCAR of the vasp, but I have found that in each case the hexagonal arrangement of atoms is distorted to a large degree. However, similar studies have been published with hexagonal structure of doped graphene. Could anyone guide me how to solve the problem of geometry distortion to a large extend after geometry optimization/relaxation of the doped graphene?
Thank you.
I have used the following parameters in the INCAR, but got similar structure distortion.
ISTART = 0
INIWAV = 1
ICHARG = 2
NWRITE = 2
ENCUT = 520.00 eV
EDIFF = 1.0E-05 eV
PREC = Accurate
EDIFFG = -0.01
NSW = 500
IBRION = 1/2
ISIF = 2/3/4
ISMEAR = 0/-5
SIGMA = 0.05
GGA = PE
NEDOS = 1001
LREAL = A
Yes, I have tried all possible doping sites, but still I have distortion to a large extent in the supercell.
Hi,
on the first glance your parameters look good, however, often the details are important,
have you tested all your settings, k points, smearing,...? especially the XC functional could be a problem, I am not a graphene expert, but I believe GGAs sometimes cause problems here, maybe try LDA or functionals that include van der Waals corrections such as the opt functionals by Klimes et al. which are included in vasp or others
secondly, from what structure are you starting and what does your cell look like in the c direction, one layer of graphene and enough vacuum to decouple the periodically repeated cells?
Have you checked whether your pure graphene layer (without any dopants) is giving expected results (atomic spacing, bandstructure, ...)?
finally, you are mentioning similar studies, but how similar are they and are you expecting the same results or could some differences occur? are you using the same dopants, concentration, computational setup ...
hope this helps
cheers
Dear Dr. Gregor Feldbauer,
Thank you very much for your comment.
1- Yes, I have tested different k-points, smearing, sigma, relaxation algorithm, with and without symmetry, but all with GGA only, and I got all distorted structures at the end.
2- My structures are planar and remain planar during and after geometry optimization, might be due to symmetry forces. Moreover, I have used 15 to 25 angstrom vacuum in the z-direction, but same distorted structure at the end.
3- Yes, pure graphene planar structure gives me consistent results with the literature.
4- Well, the concentration of dopant is different, however, the same dopant with different concentration has been studied. However, If I keep the same coordination of the dopant in the graphene as already published in the literature, and use the same input parameters of vasp, even though I am not getting the same structure after geometry optimization as it has been already published.
Looking for more suggestions.
Hi,
I would really try some non GGA XC functionals, just to be sure
then, I would focus again on point 4, if the same setup and procedures result in differences (and this is the case, if I understand you correctly) there seems to be an error either in your or in the published work, which could both be the case, to be sure I would check again very carefully, are really all settings the same, are you starting from the same structures (e.g. maybe pre-distorted structures have been used to better accommodate the dopants,...), are there any experimental results, which suggest more reasonable input geometries, since I do not know the relevant literature it is a bit guessing the dark here :)
sorry that I can not be more helpful right now, but a diagnosis from the distance is always somewhat tricky
cheers
you need to put the dopants slightly off the plane, start with a small step (potim parameter?) and conjugate gradients algorithm. Gaussian smearing is always safe.
I am not sure what is nedos. First start with less restrictive convergence constraints, and look at the optimization steps to see if electronic iterations have converged. At each ionic step also look at the forces to see if they decrease or if you have something strange happening due to large forces and large atomic displacements. Then you can diagnose better.
- Badges
- Science topic
- Similar topics
- Physical Sciences
- Materials Science
- Materials
- Carbon Materials
- Graphite
- Graphene