LCPQ/dft_tools
3
0
Fork 0
mirror of https://github.com/triqs/dft_tools synced 2024-11-07 06:33:48 +01:00
Code Issues Releases Wiki Activity

Updated documentation of the hk converter

Browse Source
This commit is contained in:
aichhorn 2016-08-29 10:27:33 +02:00
parent d419f1a37d
commit c4b4620b36
3 changed files with 69 additions and 12 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

66
doc/guide/conversion.rst
View File

@ -188,8 +188,8 @@ A general H(k)
In addition to the more complicated Wien2k converter,
:program:`DFTTools` contains also a light converter. It takes only
one inputfile, and creates the necessary hdf outputfile for
the DMFT calculation. The header of this input file has to have the
following format:
the DMFT calculation. The header of this input file has a defined
format, an example is the following:
.. literalinclude:: images_scripts/case.hk
@ -197,10 +197,64 @@ The lines of this header define
#. Number of :math:`\mathbf{k}`-points used in the calculation
#. Electron density for setting the chemical potential
#. Number of correlated atoms in the unit cell
#. The next line contains four numbers: index of the atom, index
of the correlated shell, :math:`l` quantum number, dimension
of this shell. Repeat this line for each correlated atom.
#. Number of total atomic shells in the hamiltonian matrix. In short,
this gives the number of lines described in the following. IN the
example file give above this number is 2.
#. The next line(s) contain four numbers each: index of the atom, index
of the equivalent shell, :math:`l` quantum number, dimension
of this shell. Repeat this line for each atomic shell, the number
of the shells is given in the previous line.
In the example input file given above, we have two inequivalent
atomic shells, one on atom number 1 with a full d-shell (dimension 5),
and one on atom number 2 with one p-shell (dimension 3).
Other examples for these lines are:
#. Full d-shell in a material with only one correlated atom in the
unit cell (e.g. SrVO3). One line is sufficient and the numbers
are `1 1 2 5`.
#. Full d-shell in a material with two equivalent atoms in the unit
cell (e.g. FeSe): You need two lines, one for each equivalent
atom. First line is `1 1 2 5`, and the second line is
`2 1 2 5`. The only difference is the first number, which tells on
which atom the shell is located. The second number is the
same in both lines, meaning that both atoms are equivalent.
#. t2g orbitals on two non-equivalent atoms in the unit cell: Two
lines again. First line is `1 1 2 3`, second line `2 2 2 3`. The
difference to the case above is that now also the second number
differs. Therefore, the two shells are treated independently in
the calculation.
#. d-p Hamiltonian in a system with two equivalent atoms each in
the unit cell (e.g. FeSe has two Fe and two Se in the unit
cell). You need for lines. First line `1 1 2 5`, second
line
`2 1 2 5`. These two lines specify Fe as in the case above. For the p
orbitals you need line three as `3 2 1 3` and line four
as `4 2 1 3`. We have 4 atoms, since the first number runs from 1 to 4,
but only two inequivalent atoms, since the second number runs
only form 1 to 2.
Note that the total dimension of the hamiltonian matrices that are
read in is the sum of all shell dimensions that you specified. For
example number 4 given above we have a dimension of 5+5+3+3=16. It is important
that the order of the shells that you give here must be the same as
the order of the orbitals in the hamiltonian matrix. In the last
example case above the code assumes that matrix index 1 to 5
belongs to the first d shell, 6 to 10 to the second, 11 to 13 to
the first p shell, and 14 to 16 the second p shell.
#. Number of correlated shells in the hamiltonian matrix, in the same
spirit as line 3.
#. The next line(s) contain six numbers: index of the atom, index
of the equivalent shell, :math:`l` quantum number, dimension
of the correlated shells, a spin-orbit parameter, and another
parameter defining interactions. Note that the latter two
parameters are not used at the moment in the code, and only kept
for compatibility reasons. In our example file we use only the
d-shell as correlated, that is why we have only one line here.
#. The last line contains several numbers: the number of irreducible
representations, and then the dimensions of the irreps. One
possibility is as the example above, another one would be 2

5
doc/guide/images_scripts/case.hk
View File

@ -1,5 +1,8 @@
64 ! number of k-points
1.0 ! Electron density
1 ! number of correlated atoms
2 ! number of total atomic shells
1 1 2 5 ! iatom, isort, l, dimension
2 2 1 3 ! iatom, isort, l, dimension
1 ! number of correlated shells
1 1 2 5 0 0 ! iatom, isort, l, dimension, SO, irep
1 5 ! # of ireps, dimension of irep