The ABINIT Users Mailing List ( CLOSED )

[verstraete@pcpm.ucl.ac.be]

Hello Corsin, a few tentative answers:

To determine the irreducible qpoints you need, do a normal abinit
calculation (GS) with the grid density you want, but for kpoints, eg 
ngkpt 2 2 2 and eventual shifts, and prtvol -1. In the output file you will 
have
a list of irred points which you can use for the phonon calculations. Once 
you have merged the ddbs, if you run anaddb with the same grid 
(222+shifts), the program will tell you if any perturbations or qpoints 
are missing (it will also tell you that the perturbations wrt electric 
field are missing, but this is normal). The qpoints you calculate are all 
along gamma-X. This is useless, as they will be interpolated in anaddb. 

With only the gamma point, you have only gamma point frequencies. This is 
enough for dielectric characterizations and raman+IR frequencies, but not 
for the full phonon band structure. Anaddb will interpolate the phonon 
band structure and DOS from the minimal set you have actually calculated. 
On these points, anaddb without ASR should give the same result as the 
abinit output files.


Matthieu

On Thu, 4 Mar 2004, Corsin Battaglia wrote:

> Dear abinit users
> 
> Well, I do not really have obtained a bandstructure for the phonons yet.
> But I am trying to calculate the phonon dispersion of the metallic compound 
> 1T-NbTe2. Since it is metallic, I am just using the phonon type 
> perturbation and neglecting the electric field type perturbation.
> My input file is based on test t26.in (phonon bandstructure of Al) and the 
> infos given in the respfn-help file.
> There are several problems:
> 
> 1) I get 6 accoustic phonon modes at q=0.
> 
> 2) The RF-SCF cycles do not converge (tolvrs* 1.0d-10) (dataset 5).
> 
> 3) I can not get rid of the feeling that I am calculating to many
> q-points and that their might be a much more economic way to get the
> bandstructure exploiting the DDB with the help of ifc=anaddb. In the
> header of the t26.in file it says that 6 dynamical matrices are needed
> for the computation of the full phonon bandstructure of Al. There are as
> well six different q-vector for which an RF calculation is performed.
> Does this also mean that I will get 6 DDB's which I have to merge and
> from which I can compute the eigenfrequencies using anaddb? How to
> decide which q-vectors must be included in the DDB? Horacio W. Leite
> Alves, in response to Kwan-Woo Lee's mail, mentions that only the
> gamma-point DDB is needed. Does this also apply to my case? Then why do
> we need the DDB merge tool? Are the eigenenergies I get from abinit the
> same as the one from anaddb (without imposing the accoustic sum rule)?
> 
> My input file is attached. Thanks in advance for your help. 
> 
> Corsin
> 
> P.S.: The electronic bandstructure is great now!!! 
> 
> 
> My input file:
> ********************************************************************
> 
> 
> # NbTe2 phonon band structure
> 
> 
>  ndtset 13
> 
>  nbdbuf 2
> 
> #Input that is common to most datasets, but not all
>   getden 1
>   getwfk 2
>   kptopt 3
>   nqpt 1
> 
> #Dataset 1 : SCF 
> 
>   getden1  0
>   getwfk1  0
>   kptopt1  1
>     nqpt1  0
>   prtden1  1
>     iscf1  5
>   tolvrs1  1.0d-13
>  
> #Dataset 2 : non-SCF for all k points
> 
>   getwfk2  1
>     iscf2 -2
>     nqpt2  0
>   tolwfr2  1.0d-22
> 
> #Dataset 3 : RF at q=0 0 0
>  
>   kptopt3  2
>     iscf3  3
>      qpt3  0 0 0
>  rfatpol3  1 3
>    rfdir3  1 1 1
>   rfphon3  1
>   tolvrs3  1.0d-10
> 
> #Dataset 4 : non-SCF at q=1/16 0 0
> 
>     iscf4 -2
>      qpt4  1/16 0.0 0.0
>   tolwfr4  1.0d-22
> 
> #Dataset 5 : RF at q=1/16 0 0
>  
>   getwfq5  4
>     iscf5  3
>      qpt5  1/16 0.0 0.0
>  rfatpol5  1 3
>    rfdir5  1 1 1
>   rfphon5  1
>   tolvrs5  1.0d-10
> 
> #Dataset 6 : non-SCF at q=1/8 0 0
> 
>     iscf6 -2
>      qpt6  1/8 0.0 0.0
>   tolwfr6  1.0d-22
> 
> #Dataset 7 : RF at q=1/8 0 0
> 
>   getwfq7  6
>     iscf7  3
>      qpt7  1/8 0.0 0.0
>  rfatpol7  1 3
>    rfdir7  1 1 1
>   rfphon7  1
>   tolvrs7  1.0d-10
> 
> #Dataset 8 : non-SCF at q=3/16 0 0
> 
>     iscf8 -2
>      qpt8  3/16 0.0 0.0
>   tolwfr8  1.0d-22
> 
> #Dataset 9 : RF at q=3/16 0 0
> 
>   getwfq9  8
>     iscf9  3
>      qpt9  3/16 0.0 0.0
>  rfatpol9  1 3
>    rfdir9  1 1 1
>   rfphon9  1
>   tolvrs9  1.0d-10
> 
> #Dataset 10 : non-SCF at q=1/4 0 0
> 
>     iscf10 -2
>      qpt10  1/4 0.0 0.0
>   tolwfr10  1.0d-22
> 
> #Dataset 11 : RF at q=1/4 0 0
> 
>   getwfq11  10
>     iscf11  3
>      qpt11  1/4 0.0 0.0
>  rfatpol11  1 3
>    rfdir11  1 1 1
>   rfphon11  1
>   tolvrs11  1.0d-10
> 
> #Dataset 12 : non-SCF at q=5/16 0 0
> 
>     iscf12 -2
>      qpt12  5/16 0.0 0.0
>   tolwfr12  1.0d-22
> 
> #Dataset 13 : RF at q=5/16 0 0
> 
>   getwfq13  12
>     iscf13  3
>      qpt13  5/16 0.0 0.0
>  rfatpol13  1 3
>    rfdir13  1 1 1
>   rfphon13  1
>   tolvrs13  1.0d-10
> 
> #Dataset 14 : non-SCF at q=3/8 0 0
> 
>     iscf14 -2
>      qpt14  3/8 0.0 0.0
>   tolwfr14  1.0d-22
> 
> #Dataset 7 : RF at q=3/8 0 0
> 
>   getwfq15  14
>     iscf15  3
>      qpt15  3/8 0.0 0.0
>  rfatpol15  1 3
>    rfdir15  1 1 1
>   rfphon15  1
>   tolvrs15  1.0d-10
> 
> #Dataset 16 : non-SCF at q=7/16 0 0
> 
>     iscf16 -2
>      qpt16  7/16 0.0 0.0
>   tolwfr16  1.0d-22
> 
> #Dataset 17 : RF at q=7/16 0 0
> 
>   getwfq17  16
>     iscf17  3
>      qpt17  7/16 0.0 0.0
>  rfatpol17  1 3
>    rfdir17  1 1 1
>   rfphon17  1
>   tolvrs17  1.0d-10
> 
> #Dataset 18 : non-SCF at q=1/2 0 0
> 
>     iscf18 -2
>      qpt18  1/2 0.0 0.0
>   tolwfr18  1.0d-22
> 
> #Dataset 19 : RF at q=1/2 0 0
> 
>   getwfq19  18
>     iscf19  3
>      qpt19  1/2 0.0 0.0
>  rfatpol19  1 3
>    rfdir19  1 1 1
>   rfphon19  1
>   tolvrs19  1.0d-10
> 
> #Common data
> 
> #Definition of the unit cell
> acell   3.68  3.68     6.61 angstrom
> #rprim  0.866 -0.500   0.000     # It is better to define
> #       0.000  1.000   0.000     # the primitive vectors
> #       0.000  0.000   1.000     # using rprim
> angdeg 90 90 120
> 
> #Definition of the atom types
> ntypat 2          # There are two type of atoms
> znucl 41 52       # The keyword "znucl" refers to the atomic number of the
>                   # possible type(s) of atom. The pseudopotential(s)
>                   # mentioned in the "files" file must correspond
>                   # to the type(s) of atom.
> 
> 
> #Definition of the atoms
> natom 3           # There are three atoms
> natrd 2           # Reads two atoms
> typat 1 2         # type 1 is Nb, type 2 is Te
> xred              # This keyword indicate that the location of the atoms
>                   # will follow, one triplet of number for each atom
>    0.0  0.0  0.0  # Triplet giving the REDUCED coordinate of atom 1.
>    1/3  2/3  1/4  # Triplet giving the REDUCED coordinate of atom 2.
>                   # Note the use of fractions (remember the limited
>                   # interpreter capabilities of ABINIT)
> 
> spgroup 164       # Spacegroup
> 
> #Definition of the occupation numbers
> occopt 4
> tsmear 0.01
> 
> #Read psp
> npsp 2            # Read 2 psp files
> ixc 1             # Nb is of type ixc 1. Te is of type ixc 1.
>                   # LDA. Nb contains semicores.
> 
> #Definition of the planewave basis set
> ecut 10.0         # Maximal kinetic energy cut-off, in Hartree
> 
> ngkpt 8 8 4       # Creates a 8x8x4 k-point grid
> 
> #Definition of the SCF procedure
> nstep 250         # Maximal number of SCF cycles
> 
> #diemac 12.0      # For metals, we use the default 10^6.
> nband 35          # nband=nb of electrons in unit cell/2+(20% for metals)
>                   # more bands are needed with semicore states
> 
> 
> 

-- 
===================================================================
Matthieu Verstraete               mailto:verstraete@pcpm.ucl.ac.be 
PCPM, Boltzmann, pl. Croix du Sud, 1            tel: 010/ 47 33 59
B-1348 Louvain-la-Neuve  Belgium                fax: 010/ 47 34 52