mirror of
https://github.com/triqs/dft_tools
synced 2024-08-26 22:21:45 +02:00
![Manuel Zingl](/assets/img/avatar_default.png)
This adds another option and a mode flag in dmftproj: * third value in window line defines now the mode * new option to provide an energy window where all bands which are within the window (at least at one k-point) are taken into account for the projectors. * updates to documention to reflects those changes
750 lines
33 KiB
Fortran
750 lines
33 KiB
Fortran
|
|
c ******************************************************************************
|
|
c
|
|
c TRIQS: a Toolbox for Research in Interacting Quantum Systems
|
|
c
|
|
c Copyright (C) 2011 by L. Pourovskii, V. Vildosola, C. Martins, M. Aichhorn
|
|
c
|
|
c TRIQS is free software: you can redistribute it and/or modify it under the
|
|
c terms of the GNU General Public License as published by the Free Software
|
|
c Foundation, either version 3 of the License, or (at your option) any later
|
|
c version.
|
|
c
|
|
c TRIQS is distributed in the hope that it will be useful, but WITHOUT ANY
|
|
c WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
|
|
c FOR A PARTICULAR PURPOSE. See the GNU General Public License for more
|
|
c details.
|
|
c
|
|
c You should have received a copy of the GNU General Public License along with
|
|
c TRIQS. If not, see <http://www.gnu.org/licenses/>.
|
|
c
|
|
c *****************************************************************************/
|
|
|
|
SUBROUTINE set_projections(e1,e2)
|
|
C %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|
|
C %% %%
|
|
C %% This subroutine sets up the projection matrices in the energy %%
|
|
C %% window [e1,e2]. If proj_mode is 1 or 2 then e1 and e2 are not %%
|
|
C %% energies, but directly used as band indices. %%
|
|
C %% Two types of projection can be defined : %%
|
|
C %% - The projectors <u_orb|ik,ib,is> for the correlated orbital %%
|
|
C %% only. (orb = iatom,is,m) %%
|
|
C %% (They are stored in the table pr_crorb) %%
|
|
C %% - The Theta projectors <theta_orb|k,ib> for all the orbitals %%
|
|
C %% (They are stored in the table pr_orb) %%
|
|
C %% %%
|
|
C %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|
|
|
|
C Definiton of the variables :
|
|
C ----------------------------
|
|
C
|
|
USE almblm_data
|
|
USE common_data
|
|
USE prnt
|
|
USE projections
|
|
USE reps
|
|
USE symm
|
|
IMPLICIT NONE
|
|
C
|
|
REAL(KIND=8) :: e1, e2
|
|
INTEGER :: iorb, icrorb, ik, is, ib, m, l, lm, nbbot, nbtop
|
|
INTEGER :: isrt, n, ilo, iatom, i, imu, jatom, jorb,isym, jcrorb
|
|
LOGICAL :: included,param
|
|
COMPLEX(KIND=8), DIMENSION(:), ALLOCATABLE :: coeff
|
|
COMPLEX(KIND=8), DIMENSION(:,:), ALLOCATABLE :: tmp_mat
|
|
COMPLEX(KIND=8), DIMENSION(:,:), ALLOCATABLE :: tmp_matbis
|
|
COMPLEX(KIND=8), DIMENSION(:,:,:), ALLOCATABLE :: tmp_matn
|
|
C
|
|
C
|
|
C
|
|
WRITE(buf,'(a)')'Creation of the projectors...'
|
|
CALL printout(0)
|
|
C
|
|
C
|
|
C ======================================================================
|
|
C Selection of the bands which lie in the chosen energy window [e1;e2]
|
|
C or if proj_mode = [1,2] e1 and e2 are directly used as band indices:
|
|
C ======================================================================
|
|
C
|
|
|
|
IF(proj_mode.gt.0) THEN
|
|
C The same number of bands are included at each k-point
|
|
kp(:,:)%included=.TRUE.
|
|
C Use directly e1 and e2 as band indices. Set to nbmin or
|
|
C nbmax if too small or too large
|
|
DO is=1,ns
|
|
DO ik=1,nk
|
|
IF(e1 > kp(ik,is)%nbmin) THEN
|
|
kp(ik,is)%nb_bot=INT(e1)
|
|
ELSE
|
|
kp(ik,is)%nb_bot=kp(ik,is)%nbmin
|
|
ENDIF
|
|
IF(e2 < kp(ik,is)%nbmax) THEN
|
|
kp(ik,is)%nb_top=INT(e2)
|
|
ELSE
|
|
kp(ik,is)%nb_top=kp(ik,is)%nbmax
|
|
ENDIF
|
|
ENDDO
|
|
ENDDO
|
|
ELSE
|
|
kp(:,:)%included=.FALSE.
|
|
C the field kp%included = boolean which is .TRUE. when there is at least one band
|
|
C at this k-point whose energy eignevalue is in the energy window.
|
|
DO is=1,ns
|
|
DO ik=1,nk
|
|
included=.FALSE.
|
|
DO ib=kp(ik,is)%nbmin,kp(ik,is)%nbmax
|
|
IF(.NOT.included.AND.kp(ik,is)%eband(ib) > e1.AND.
|
|
& kp(ik,is)%eband(ib).LE.e2) THEN
|
|
C If the energy eigenvalue E of the band ib at the k-point ik is such that e1 < E =< e2,
|
|
C then all the band with ib1>ib must be "included" in the computation and kp%nb_bot is initialized at the value ib.
|
|
included=.TRUE.
|
|
kp(ik,is)%nb_bot=ib
|
|
ELSEIF(included.AND.kp(ik,is)%eband(ib) > e2) THEN
|
|
C If the energy eigenvalue E of the current band ib at the k-point ik is such that E > e2 and all the previous
|
|
C band are "included", then the field kp%included = .TRUE. and kp%nb_top = ib-1 (the index of the previous band)
|
|
kp(ik,is)%nb_top=ib-1
|
|
kp(ik,is)%included=.TRUE.
|
|
EXIT
|
|
C The loop on the band ib is stopped, since all the bands after ib have an energy > that of ib.
|
|
ELSEIF(ib==kp(ik,is)%nbmax.AND.kp(ik,is)%eband(ib)
|
|
& > e1.AND.kp(ik,is)%eband(ib).LE.e2) THEN
|
|
C If the energy eigenvalue E of the last band ib=kp%nbmax at the k-point ik is such that e1 < E =< e2 and all the
|
|
C previous bands are "included", then the band ib must be "included" and kp%nb_bot is initialized at the value kp%nbmax.
|
|
kp(ik,is)%nb_top=ib
|
|
kp(ik,is)%included=.TRUE.
|
|
ENDIF
|
|
C If the eigenvalues of the bands at the k-point ik are < e1 and included=.FALSE.
|
|
C of if the eigenvalues of the bands at the k-point ik are in the energy window [e1,e2] and included=.TRUE.,
|
|
C nothing is done...
|
|
ENDDO ! End of the ib loop
|
|
C If all the eigenvalues of the bands at the k-point ik are not in the window,
|
|
C then kp%included remains at the value .FALSE. and the field kp%nb_top and kp%nb_bot are set to 0
|
|
IF (.not.kp(ik,is)%included) THEN
|
|
kp(ik,is)%nb_bot=0
|
|
kp(ik,is)%nb_top=0
|
|
ENDIF
|
|
ENDDO ! End of the ik loop
|
|
ENDDO ! End of the is loop
|
|
ENDIF
|
|
C ---------------------------------------------------------------------------------------
|
|
C Checking of the input files if spin-polarized inputs and SO is taken into account:
|
|
C There should not be any difference between up and dn limits for each k-point.
|
|
C Printing a Warning if this is not the case.
|
|
C -------------------
|
|
C
|
|
IF (ifSP.AND.ifSO) THEN
|
|
param=.TRUE.
|
|
DO ik=1,nk
|
|
param=param.AND.(kp(ik,1)%included.eqv.kp(ik,2)%included)
|
|
param=param.AND.(kp(ik,1)%nb_bot==kp(ik,2)%nb_bot)
|
|
param=param.AND.(kp(ik,1)%nb_top==kp(ik,2)%nb_top)
|
|
IF (.not.param) EXIT
|
|
C For a valid compoutation, the same k-points must be included for up and dn states,
|
|
C and the upper and lower limits must be the same in both case.
|
|
ENDDO
|
|
IF (.not.param) THEN
|
|
WRITE(buf,'(a,a)')'A Spin-orbit computation for this',
|
|
& ' compound is not possible with these input files.'
|
|
CALL printout(0)
|
|
WRITE(buf,'(a)')'END OF THE PRGM'
|
|
CALL printout(0)
|
|
STOP
|
|
ENDIF
|
|
ENDIF
|
|
C ---------------------------------------------------------------------------------------
|
|
C
|
|
C
|
|
C ==================================================================
|
|
C Orthonormalization of the radial wave functions for each orbital :
|
|
C ==================================================================
|
|
C
|
|
C This step is essential for setting the Theta projectors.
|
|
IF(.NOT.ALLOCATED(norm_radf)) THEN
|
|
ALLOCATE(norm_radf(norb))
|
|
C norm_radf is a table of "ortfunc" elements, its size ranges from 1 to norb.
|
|
DO iorb=1,norb
|
|
l=orb(iorb)%l
|
|
isrt=orb(iorb)%sort
|
|
norm_radf(iorb)%n=nLO(l,isrt)+2
|
|
n=norm_radf(iorb)%n
|
|
ALLOCATE(norm_radf(iorb)%s12(n,n,ns))
|
|
C norm_radf%n = size of the matrix s12
|
|
C norm_radf%s12 = matrix of size n*n (one for spin up, one for spin down, if necessary)
|
|
DO is=1,ns
|
|
norm_radf(iorb)%s12(1:n,1:n,is)=0d0
|
|
norm_radf(iorb)%s12(1,1,is)=1d0
|
|
norm_radf(iorb)%s12(2,2,is)=u_dot_norm(l,isrt,is)
|
|
C Initialization of the matrix norm_radf%s12 for each orbital (l,isrt).
|
|
C We remind tha it is assumed that nLO has the value 0 or 1 only !!
|
|
DO ilo=1,nLO(l,isrt)
|
|
norm_radf(iorb)%s12(2+ilo,2+ilo,is)=1d0
|
|
norm_radf(iorb)%s12(2+ilo,1,is)=
|
|
= ovl_LO_u(ilo,l,isrt,is)
|
|
norm_radf(iorb)%s12(1,2+ilo,is)=
|
|
= ovl_LO_u(ilo,l,isrt,is)
|
|
norm_radf(iorb)%s12(2+ilo,2,is)=
|
|
= ovl_LO_udot(ilo,l,isrt,is)
|
|
norm_radf(iorb)%s12(2,2+ilo,is)=
|
|
= ovl_LO_udot(ilo,l,isrt,is)
|
|
ENDDO
|
|
C Computation of the square root of norm_radf:
|
|
CALL orthogonal_r(norm_radf(iorb)%
|
|
& s12(1:n,1:n,is),n,.FALSE.)
|
|
C the field norm_radf%s12 is finally the C matrix described in the tutorial (or in equation (3.63) in my thesis)
|
|
ENDDO
|
|
ENDDO
|
|
ENDIF
|
|
C
|
|
C =====================================
|
|
C Creation of the projection matrices :
|
|
C =====================================
|
|
C
|
|
IF(.NOT.ALLOCATED(pr_orb)) THEN
|
|
ALLOCATE(pr_crorb(ncrorb,nk,ns))
|
|
ALLOCATE(pr_orb(norb,nk,ns))
|
|
ENDIF
|
|
C pr_crorb = table of "proj_mat" elements for the correlated orbitals (size from 1 to ncrorb, from 1 to nk, from 1 to ns)
|
|
C pr_orb = table of "proj_mat_n" elements for all the orbitals (size from 1 to norb, from 1 to nk, from 1 to ns)
|
|
DO is=1,ns
|
|
DO ik=1,nk
|
|
C Only the k-points with inlcuded bands are considered for the projectors.
|
|
IF(.NOT.kp(ik,is)%included) CYCLE
|
|
C ------------------------------------------------
|
|
C Wannier Projectors for the correlated orbitals :
|
|
C ------------------------------------------------
|
|
DO icrorb=1,ncrorb
|
|
l=crorb(icrorb)%l
|
|
iatom=crorb(icrorb)%atom
|
|
isrt=crorb(icrorb)%sort
|
|
C Case of l=0 :
|
|
C -------------
|
|
IF (l==0) THEN
|
|
IF(ALLOCATED(pr_crorb(icrorb,ik,is)%mat)) THEN
|
|
DEALLOCATE(pr_crorb(icrorb,ik,is)%mat)
|
|
ENDIF
|
|
ALLOCATE(pr_crorb(icrorb,ik,is)%
|
|
% mat(1,kp(ik,is)%nb_bot:kp(ik,is)%nb_top))
|
|
C pr_crorb%mat = the projection matrix with 1 line and (nb_top-nb_bot) columns
|
|
DO ib=kp(ik,is)%nb_bot,kp(ik,is)%nb_top
|
|
pr_crorb(icrorb,ik,is)%mat(1,ib)=
|
|
= kp(ik,is)%Alm(1,iatom,ib)
|
|
DO ilo=1,nLO(l,isrt)
|
|
pr_crorb(icrorb,ik,is)%mat(1,ib)=
|
|
= pr_crorb(icrorb,ik,is)%mat(1,ib)+
|
|
+ kp(ik,is)%Clm(ilo,1,iatom,ib)*
|
|
* ovl_LO_u(ilo,l,isrt,is)
|
|
ENDDO ! End of the ilo loop
|
|
ENDDO ! End of the ib loop
|
|
C prcrorb(icrorb,ik,is)%mat(1,ib)= <ul1(icrorb,1,is)|psi(is,ik,ib)> = Alm+Clm*ovl_LO_u
|
|
C
|
|
C Case of any other l :
|
|
C ---------------------
|
|
ELSE
|
|
lm=l*l
|
|
C Since the correlated orbital is the l orbital, the elements range from l*l+1 to (l+1)^2
|
|
C the sum from 0 to (l-1) of m (from -l to l) is l^2.
|
|
IF(ALLOCATED(pr_crorb(icrorb,ik,is)%mat)) THEN
|
|
DEALLOCATE(pr_crorb(icrorb,ik,is)%mat)
|
|
ENDIF
|
|
ALLOCATE(pr_crorb(icrorb,ik,is)%
|
|
% mat(-l:l,kp(ik,is)%nb_bot:kp(ik,is)%nb_top))
|
|
C pr_crorb%mat = the projection matrix with (2*l+1) lines and (nb_top-nb_bot) columns
|
|
DO m=-l,l
|
|
lm=lm+1
|
|
DO ib=kp(ik,is)%nb_bot,kp(ik,is)%nb_top
|
|
pr_crorb(icrorb,ik,is)%mat(m,ib)=
|
|
= kp(ik,is)%Alm(lm,iatom,ib)
|
|
DO ilo=1,nLO(l,isrt)
|
|
pr_crorb(icrorb,ik,is)%mat(m,ib)=
|
|
= pr_crorb(icrorb,ik,is)%mat(m,ib)+
|
|
+ kp(ik,is)%Clm(ilo,lm,iatom,ib)*
|
|
* ovl_LO_u(ilo,l,isrt,is)
|
|
ENDDO ! End of the ilo loop
|
|
ENDDO ! End of the ib loop
|
|
ENDDO ! End of the m loop
|
|
C prcrorb(icrorb,ik,is)%mat(m,ib)= <ul1(icrorb,m,is)|psi(is,ik,ib)> = Alm+Clm*ovl_LO_u
|
|
ENDIF ! End of the if l=0 if-then-else
|
|
ENDDO ! End of the icrorb loop
|
|
C
|
|
C ---------------------------------------
|
|
C Theta Projectors for all the orbitals :
|
|
C ---------------------------------------
|
|
DO iorb=1,norb
|
|
l=orb(iorb)%l
|
|
n=norm_radf(iorb)%n
|
|
iatom=orb(iorb)%atom
|
|
C Case of l=0 :
|
|
C -------------
|
|
IF (l==0) THEN
|
|
IF(ALLOCATED(pr_orb(iorb,ik,is)%matn)) THEN
|
|
DEALLOCATE(pr_orb(iorb,ik,is)%matn)
|
|
ENDIF
|
|
ALLOCATE(pr_orb(iorb,ik,is)%
|
|
% matn(1,kp(ik,is)%nb_bot:kp(ik,is)%nb_top,n))
|
|
ALLOCATE(coeff(1:n))
|
|
C pr_orb%matn = the projection matrix with 1 line and (nb_top-nb_bot) columns for the n (size of s12) coefficients
|
|
C coeff = table of size n which will contain the decomposition of the Bloch state |psi_ik,ib,is>
|
|
C as in equation 22 of the tutorial (Alm, Blm, and Clm )
|
|
DO ib=kp(ik,is)%nb_bot,kp(ik,is)%nb_top
|
|
coeff(1)=kp(ik,is)%Alm(1,iatom,ib)
|
|
coeff(2)=kp(ik,is)%Blm(1,iatom,ib)
|
|
coeff(3:n)=kp(ik,is)%Clm(1:n-2,1,iatom,ib)
|
|
coeff=MATMUL(coeff,norm_radf(iorb)%s12(1:n,1:n,is))
|
|
C coeff = coefficients c_(j,lm) of the decomposition of the state |psi> in the orthogonalized basis |phi_j>
|
|
C as defined in the tutorial (equation 25)
|
|
pr_orb(iorb,ik,is)%matn(1,ib,1:n)=coeff(1:n)
|
|
ENDDO
|
|
DEALLOCATE(coeff)
|
|
C pr_orb(iorb,ik,is)%matn(m,ib,1:n) is then the Theta projector as defined in equation 26 of the tutorial.
|
|
C
|
|
C Case of any other l :
|
|
C ---------------------
|
|
ELSE
|
|
lm=l*l
|
|
C As the orbital is the l orbital, the elements range from l*l+1 to (l+1)^2
|
|
C the sum from 0 to (l-1) of m (from -l to l) is l^2.
|
|
IF(ALLOCATED(pr_orb(iorb,ik,is)%matn)) THEN
|
|
DEALLOCATE(pr_orb(iorb,ik,is)%matn)
|
|
ENDIF
|
|
ALLOCATE(pr_orb(iorb,ik,is)%
|
|
% matn(-l:l,kp(ik,is)%nb_bot:kp(ik,is)%nb_top,n))
|
|
ALLOCATE(coeff(1:n))
|
|
C pr_orb%matn = the projection matrix with (2*l+1) lines and (nb_top-nb_bot) columns for the n (size of s12) coefficients
|
|
C coeff = table of size n which will contain the decomposition of the Bloch state |psi_ik,ib,is>
|
|
C as in equation 22 of the tutorial (Alm, Blm, and Clm )
|
|
DO m=-l,l
|
|
lm=lm+1
|
|
DO ib=kp(ik,is)%nb_bot,kp(ik,is)%nb_top
|
|
coeff(1)=kp(ik,is)%Alm(lm,iatom,ib)
|
|
coeff(2)=kp(ik,is)%Blm(lm,iatom,ib)
|
|
coeff(3:n)=kp(ik,is)%Clm(1:n-2,lm,iatom,ib)
|
|
coeff=MATMUL(coeff,
|
|
& norm_radf(iorb)%s12(1:n,1:n,is))
|
|
C coeff = coefficients c_(j,lm) of the decomposition of the state |psi> in the orthogonalized basis |phi_j>
|
|
C as defined in the tutorial (equation 25)
|
|
pr_orb(iorb,ik,is)%matn(m,ib,1:n)=coeff(1:n)
|
|
ENDDO
|
|
ENDDO ! End of the m loop
|
|
DEALLOCATE(coeff)
|
|
C pr_orb(iorb,ik,is)%matn(m,ib,1:n) is then the Theta projector as defined in equation 26 of the tutorial.
|
|
ENDIF ! End of the if l=0 if-then-else
|
|
ENDDO ! End of the iorb loop
|
|
C
|
|
ENDDO ! End of the loop on ik
|
|
ENDDO ! End of the loop on is
|
|
C
|
|
C
|
|
C ==========================================================================
|
|
C Multiplication of the projection matrices by the local rotation matrices :
|
|
C ==========================================================================
|
|
C
|
|
C ------------------------------------------------
|
|
C Wannier Projectors for the correlated orbitals :
|
|
C ------------------------------------------------
|
|
C
|
|
DO jcrorb=1,ncrorb
|
|
jatom=crorb(jcrorb)%atom
|
|
isrt=crorb(jcrorb)%sort
|
|
l=crorb(jcrorb)%l
|
|
C
|
|
C The case l=0 is a particular case of "non-mixing" basis.
|
|
C --------------------------------------------------------
|
|
IF (l==0) THEN
|
|
C For the s orbital, no multiplication is needed, since the matrix representation of any rotation
|
|
C (and thus Rloc) is always 1.
|
|
DO ik=1,nk
|
|
DO is=1,ns
|
|
C Only the k-points with inlcuded bands are considered for the projectors.
|
|
IF(.NOT.kp(ik,is)%included) CYCLE
|
|
nbtop=kp(ik,is)%nb_top
|
|
nbbot=kp(ik,is)%nb_bot
|
|
IF(ALLOCATED(pr_crorb(jcrorb,ik,is)%mat_rep)) THEN
|
|
DEALLOCATE(pr_crorb(jcrorb,ik,is)%mat_rep)
|
|
ENDIF
|
|
ALLOCATE(pr_crorb(jcrorb,ik,is)
|
|
& %mat_rep(1,nbbot:nbtop))
|
|
pr_crorb(jcrorb,ik,is)%mat_rep(1,nbbot:nbtop)=
|
|
= pr_crorb(jcrorb,ik,is)%mat(1,nbbot:nbtop)
|
|
C As a result, prcrorb%matrep = prcrorb%mat
|
|
ENDDO
|
|
ENDDO
|
|
C
|
|
C If the basis representation needs a complete spinor rotation approach (matrices of size 2*(2*l+1) )
|
|
C ---------------------------------------------------------------------------------------------------
|
|
ELSEIF (reptrans(l,isrt)%ifmixing) THEN
|
|
C If this option is used, then ifSO=.TRUE. (because of the restriction in set_ang_trans.f)
|
|
C Moreover ifSP=.TRUE. (since ifSO => ifSP in this version)
|
|
C As a result, we know that nb_bot(up)=nb_bot(dn) and nb_top(up)=nb_top(dn)
|
|
DO ik=1,nk
|
|
C Only the k-points with inlcuded bands are considered for the projectors.
|
|
IF(.NOT.kp(ik,1)%included) CYCLE
|
|
nbbot=kp(ik,1)%nb_bot
|
|
nbtop=kp(ik,1)%nb_top
|
|
C In this case, the projection matrix will be stored in prcrorb%matrep with is=1.
|
|
IF(ALLOCATED(pr_crorb(jcrorb,ik,1)%mat_rep)) THEN
|
|
DEALLOCATE(pr_crorb(jcrorb,ik,1)%mat_rep)
|
|
ENDIF
|
|
ALLOCATE(pr_crorb(jcrorb,ik,1)%
|
|
% mat_rep(1:2*(2*l+1),nbbot:nbtop))
|
|
C The element prcrorb%matrep for is=2 is set to 0, since all the matrix will be stored in the matrix matrep for is=1
|
|
IF(.not.ALLOCATED(pr_crorb(jcrorb,ik,2)%mat_rep)) THEN
|
|
ALLOCATE(pr_crorb(jcrorb,ik,2)%mat_rep(1,1))
|
|
pr_crorb(jcrorb,ik,2)%mat_rep(1,1)=0.d0
|
|
ENDIF
|
|
C Creation of a matrix tmp_mat which "concatenates" up and dn parts of pr_crorb.
|
|
ALLOCATE(tmp_mat(1:2*(2*l+1),nbbot:nbtop))
|
|
tmp_mat(1:(2*l+1),nbbot:nbtop)=
|
|
= pr_crorb(jcrorb,ik,1)%mat(-l:l,nbbot:nbtop)
|
|
C The first (2l+1) lines are the spin-up part of the projection matrix prcrorb%mat.
|
|
C
|
|
C ---------------------------------------------------------------------------------------
|
|
C Interruption of the prgm if there is no dn part of pr_orb.
|
|
C -------------------------
|
|
C
|
|
IF(.not.ifSP) THEN
|
|
WRITE(buf,'(a,a,i2,a)')'The projectors on ',
|
|
& 'the dn states are required for isrt = ',isrt,
|
|
& ' but there is no spin-polarized input files.'
|
|
CALL printout(0)
|
|
WRITE(buf,'(a)')'END OF THE PRGM'
|
|
CALL printout(0)
|
|
STOP
|
|
ENDIF
|
|
C ---------------------------------------------------------------------------------------
|
|
C
|
|
C The last (2l+1) lines are the spin-dn part of the projection matrix prcrorb%mat.
|
|
tmp_mat((2*l+2):2*(2*l+1),nbbot:nbtop)=
|
|
= pr_crorb(jcrorb,ik,2)%mat(-l:l,nbbot:nbtop)
|
|
C
|
|
C Multiplication by the local rotation matrix ; Up and dn parts are treated independently
|
|
C since in lapw2 (-alm) the coefficients Alm, Blm and Clm were calculated in the local frame
|
|
C but without taking into account the spinor-rotation matrix.
|
|
ALLOCATE(tmp_matbis(1:(2*l+1),nbbot:nbtop))
|
|
tmp_matbis(1:(2*l+1),nbbot:nbtop)=
|
|
= tmp_mat(1:(2*l+1),nbbot:nbtop)
|
|
CALL rot_projectmat(tmp_matbis,
|
|
& l,nbbot,nbtop,jatom,isrt)
|
|
tmp_mat(1:(2*l+1),nbbot:nbtop)=
|
|
= tmp_matbis(1:(2*l+1),nbbot:nbtop)
|
|
tmp_matbis(1:(2*l+1),nbbot:nbtop)=
|
|
= tmp_mat(2*l+2:2*(2*l+1),nbbot:nbtop)
|
|
CALL rot_projectmat(tmp_matbis,
|
|
& l,nbbot,nbtop,jatom,isrt)
|
|
tmp_mat(2*l+2:2*(2*l+1),nbbot:nbtop)=
|
|
= tmp_matbis(1:(2*l+1),nbbot:nbtop)
|
|
DEALLOCATE(tmp_matbis)
|
|
C
|
|
C Putting pr_crorb in the desired basis associated to (l,isrt)
|
|
C
|
|
pr_crorb(jcrorb,ik,1)%mat_rep(1:2*(2*l+1),nbbot:nbtop)=
|
|
= MATMUL(reptrans(l,isrt)%transmat
|
|
& (1:2*(2*l+1),1:2*(2*l+1)),
|
|
& tmp_mat(1:2*(2*l+1),nbbot:nbtop))
|
|
C pr_crorb%mat_rep = proj_{new_i} = reptrans*proj_{lm} = <new_i|lm>*proj_{lm}
|
|
DEALLOCATE(tmp_mat)
|
|
ENDDO ! End of the ik loop
|
|
C
|
|
C If the basis representation can be reduce to the up/up block (matrices of size (2*l+1) only)
|
|
C --------------------------------------------------------------------------------------------
|
|
ELSE
|
|
DO ik=1,nk
|
|
DO is=1,ns
|
|
C Only the k-points with inlcuded bands are considered for the projectors.
|
|
IF(.NOT.kp(ik,is)%included) CYCLE
|
|
C In this case, nb_top(up) and nb_bot(up) can differ from nb_top(dn) and nb_bot(dn)
|
|
nbbot=kp(ik,is)%nb_bot
|
|
nbtop=kp(ik,is)%nb_top
|
|
IF(ALLOCATED(pr_crorb(jcrorb,ik,is)%mat_rep)) THEN
|
|
DEALLOCATE(pr_crorb(jcrorb,ik,is)%mat_rep)
|
|
ENDIF
|
|
ALLOCATE(pr_crorb(jcrorb,ik,is)
|
|
& %mat_rep(-l:l,nbbot:nbtop))
|
|
pr_crorb(jcrorb,ik,is)%mat_rep(-l:l,nbbot:nbtop)=
|
|
= pr_crorb(jcrorb,ik,is)%mat(-l:l,nbbot:nbtop)
|
|
C
|
|
C Multiplication by the local rotation matrix
|
|
C since in lapw2 (-alm) the coefficients Alm, Blm and Clm were calculated in the local frame
|
|
CALL rot_projectmat(pr_crorb(jcrorb,ik,is)
|
|
& %mat_rep(-l:l,nbbot:nbtop),l,nbbot,nbtop,jatom,isrt)
|
|
C
|
|
C Putting pr_crorb in the desired basis associated to (l,isrt)
|
|
pr_crorb(jcrorb,ik,is)%mat_rep(-l:l,nbbot:nbtop)=
|
|
= MATMUL(reptrans(l,isrt)%transmat(-l:l,-l:l),
|
|
& pr_crorb(jcrorb,ik,is)%mat_rep(-l:l,nbbot:nbtop))
|
|
C pr_crorb%mat_rep = proj_{new_i} = reptrans*proj_{lm} = <new_i|lm>*proj_{lm}
|
|
ENDDO ! End of the is loop
|
|
ENDDO ! End of the ik loop
|
|
ENDIF ! End of the if mixing if-then-else
|
|
ENDDO ! End of the jcrorb loop
|
|
C
|
|
C ---------------------------------------
|
|
C Theta Projectors for all the orbitals :
|
|
C ---------------------------------------
|
|
C
|
|
DO jorb=1,norb
|
|
jatom=orb(jorb)%atom
|
|
isrt=orb(jorb)%sort
|
|
n=norm_radf(jorb)%n
|
|
l=orb(jorb)%l
|
|
C
|
|
C The case l=0 is a particular case of "non-mixing" basis.
|
|
C --------------------------------------------------------
|
|
IF (l==0) THEN
|
|
C For the s orbital, no multiplication is needed, since the matrix representation of any rotation
|
|
C (and therefore Rloc) is always 1.
|
|
DO ik=1,nk
|
|
DO is=1,ns
|
|
C Only the k-points with inlcuded bands are considered for the projectors.
|
|
IF(.NOT.kp(ik,is)%included) CYCLE
|
|
nbtop=kp(ik,is)%nb_top
|
|
nbbot=kp(ik,is)%nb_bot
|
|
IF(ALLOCATED(pr_orb(jorb,ik,is)%matn_rep)) THEN
|
|
DEALLOCATE(pr_orb(jorb,ik,is)%matn_rep)
|
|
ENDIF
|
|
ALLOCATE(pr_orb(jorb,ik,is)%matn_rep
|
|
& (1,nbbot:nbtop,1:n))
|
|
pr_orb(jorb,ik,is)%matn_rep(1,nbbot:nbtop,1:n)=
|
|
= pr_orb(jorb,ik,is)%matn(1,nbbot:nbtop,1:n)
|
|
C As a result, prorb%matnrep = prorb%matn
|
|
ENDDO
|
|
ENDDO
|
|
C
|
|
C If the basis representation needs a complete spinor rotation approach (matrices of size 2*(2*l+1) )
|
|
C ---------------------------------------------------------------------------------------------------
|
|
ELSEIF (reptrans(l,isrt)%ifmixing) THEN
|
|
C If this option is used, then ifSO=.TRUE. (restriction in set_ang_trans.f)
|
|
C Moreover ifSP=.TRUE. (since ifSO => ifSP)
|
|
C As a result, we know that nb_bot(up)=nb_bot(dn) and nb_top(up)=nb_top(dn)
|
|
DO ik=1,nk
|
|
C Only the k-points with inlcuded bands are considered for the projectors.
|
|
IF(.NOT.kp(ik,1)%included) CYCLE
|
|
nbbot=kp(ik,1)%nb_bot
|
|
nbtop=kp(ik,1)%nb_top
|
|
C In this case, the projection matrix will be stored in prorb%matnrep with is=1.
|
|
IF(ALLOCATED(pr_orb(jorb,ik,1)%matn_rep)) THEN
|
|
DEALLOCATE(pr_orb(jorb,ik,1)%matn_rep)
|
|
ENDIF
|
|
ALLOCATE(pr_orb(jorb,ik,1)%
|
|
% matn_rep(1:2*(2*l+1),nbbot:nbtop,1:n))
|
|
C The element prorb%matnrep for is=2 is set to 0, since all the matrix will be stored in the matrix matnrep for is=1
|
|
IF(.not.ALLOCATED(pr_orb(jorb,ik,2)%matn_rep)) THEN
|
|
ALLOCATE(pr_orb(jorb,ik,2)%matn_rep(1,1,1))
|
|
pr_orb(jorb,ik,2)%matn_rep(1,1,1)=0.d0
|
|
ENDIF
|
|
C Creation of a matrix tmp_matn which "concatenates" up and dn parts of pr_orb
|
|
ALLOCATE(tmp_matn(1:2*(2*l+1),nbbot:nbtop,1:n))
|
|
tmp_matn(1:(2*l+1),nbbot:nbtop,1:n)=
|
|
= pr_orb(jorb,ik,1)%matn(-l:l,nbbot:nbtop,1:n)
|
|
C The first (2l+1) lines are the spin-up part of the projection matrix prorb%matn.
|
|
C
|
|
C ---------------------------------------------------------------------------------------
|
|
C Interruption of the prgm if there is no dn part of pr_orb.
|
|
C -------------------------
|
|
C
|
|
IF(.not.ifSP) THEN
|
|
WRITE(buf,'(a,a,i2,a)')'The projectors on ',
|
|
& 'the down states are required for isrt = ',isrt,
|
|
& ' but there is no spin-polarized input files.'
|
|
CALL printout(0)
|
|
WRITE(buf,'(a)')'END OF THE PRGM'
|
|
CALL printout(0)
|
|
STOP
|
|
ENDIF
|
|
C ---------------------------------------------------------------------------------------
|
|
C
|
|
C The last (2l+1) lines are the spin-dn part of the projection matrix prorb%matn.
|
|
tmp_matn(2*l+2:2*(2*l+1),nbbot:nbtop,1:n)=
|
|
= pr_orb(jorb,ik,2)%matn(-l:l,nbbot:nbtop,1:n)
|
|
C
|
|
DO i=1,n
|
|
C Multiplication by the local rotation matrix ; Up and dn parts are treated independently
|
|
C since in lapw2 (-alm) the coefficients Alm, Blm and Clm were calculated in the local frame
|
|
C but without taking into account the spinor-rotation matrix.
|
|
ALLOCATE(tmp_matbis(1:(2*l+1),nbbot:nbtop))
|
|
tmp_matbis(1:(2*l+1),nbbot:nbtop)=
|
|
= tmp_matn(1:(2*l+1),nbbot:nbtop,i)
|
|
CALL rot_projectmat(tmp_matbis,
|
|
& l,nbbot,nbtop,jatom,isrt)
|
|
tmp_matn(1:(2*l+1),nbbot:nbtop,i)=
|
|
= tmp_matbis(1:(2*l+1),nbbot:nbtop)
|
|
tmp_matbis(1:(2*l+1),nbbot:nbtop)=
|
|
= tmp_matn(2*l+2:2*(2*l+1),nbbot:nbtop,i)
|
|
CALL rot_projectmat(tmp_matbis,
|
|
& l,nbbot,nbtop,jatom,isrt)
|
|
tmp_matn(2*l+2:2*(2*l+1),nbbot:nbtop,i)=
|
|
= tmp_matbis(1:(2*l+1),nbbot:nbtop)
|
|
DEALLOCATE(tmp_matbis)
|
|
C Putting pr_orb in the desired basis associated to (l,isrt)
|
|
pr_orb(jorb,ik,1)%matn_rep
|
|
& (1:2*(2*l+1),nbbot:nbtop,i)=
|
|
= MATMUL(reptrans(l,isrt)%
|
|
& transmat(1:2*(2*l+1),1:2*(2*l+1)),
|
|
& tmp_matn(1:2*(2*l+1),nbbot:nbtop,i))
|
|
C pr_orb%matn_rep = proj_{new_i} = reptrans*proj_{lm} = <new_i|lm>*proj_{lm}
|
|
ENDDO ! End of the i-loop
|
|
DEALLOCATE(tmp_matn)
|
|
ENDDO ! End of the ik loop
|
|
C
|
|
C If the basis representation can be reduce to the up/up block (matrices of size (2*l+1) only)
|
|
C --------------------------------------------------------------------------------------------
|
|
ELSE
|
|
DO ik=1,nk
|
|
DO is=1,ns
|
|
C Only the k-points with inlcuded bands are considered for the projectors.
|
|
IF(.NOT.kp(ik,is)%included) CYCLE
|
|
C In this case, nb_top(up) and nb_bot(up) can differ from nb_top(dn) and nb_bot(dn)
|
|
nbbot=kp(ik,is)%nb_bot
|
|
nbtop=kp(ik,is)%nb_top
|
|
IF(ALLOCATED(pr_orb(jorb,ik,is)%matn_rep)) THEN
|
|
DEALLOCATE(pr_orb(jorb,ik,is)%matn_rep)
|
|
ENDIF
|
|
ALLOCATE(pr_orb(jorb,ik,is)%
|
|
& matn_rep(-l:l,nbbot:nbtop,1:n))
|
|
pr_orb(jorb,ik,is)%matn_rep(-l:l,nbbot:nbtop,1:n)=
|
|
= pr_orb(jorb,ik,is)%matn(-l:l,nbbot:nbtop,1:n)
|
|
C
|
|
DO i=1,n
|
|
C Multiplication by the local rotation matrix
|
|
C since in lapw2 (-alm) the coefficients Alm, Blm and Clm were calculated in the local frame
|
|
CALL rot_projectmat(pr_orb(jorb,ik,is)
|
|
& %matn_rep(-l:l,nbbot:nbtop,i),
|
|
& l,nbbot,nbtop,jatom,isrt)
|
|
C Putting pr_orb in the desired basis associated to (l,isrt)
|
|
pr_orb(jorb,ik,is)%matn_rep(-l:l,nbbot:nbtop,i)=
|
|
= MATMUL(reptrans(l,isrt)%transmat(-l:l,-l:l),
|
|
& pr_orb(jorb,ik,is)%matn_rep(-l:l,nbbot:nbtop,i))
|
|
C pr_orb%matn_rep = proj_{new_i} = reptrans*proj_{lm} = <new_i|lm>*proj_{lm}
|
|
ENDDO ! End of the i loop
|
|
ENDDO ! End of the is loop
|
|
ENDDO ! End of the ik loop
|
|
ENDIF ! End of the if mixing if-then-else
|
|
ENDDO ! End of the jorb loop
|
|
C
|
|
C
|
|
C =============================================================================
|
|
C Printing the projectors with k-points 1 and nk in the file fort.18 for test :
|
|
C =============================================================================
|
|
DO icrorb=1,ncrorb
|
|
iatom=crorb(icrorb)%atom
|
|
isrt=crorb(icrorb)%sort
|
|
l=crorb(icrorb)%l
|
|
WRITE(18,'()')
|
|
WRITE(18,'(a,i4)') 'icrorb = ', icrorb
|
|
WRITE(18,'(a,i4,a,i4)') 'isrt = ', isrt, ' l = ', l
|
|
IF (l==0) THEN
|
|
WRITE(18,'(a,i4)') 'ik = ', 1
|
|
DO ib = kp(1,1)%nb_bot,kp(1,1)%nb_top
|
|
WRITE(18,*) pr_crorb(icrorb,1,1)%mat_rep(:,ib)
|
|
IF (ifSP)
|
|
& WRITE(18,*) pr_crorb(icrorb,1,2)%mat_rep(:,ib)
|
|
WRITE(18,'()')
|
|
ENDDO
|
|
WRITE(18,'(a,i4)') 'ik = ', nk
|
|
DO ib = kp(nk,1)%nb_bot,kp(nk,1)%nb_top
|
|
WRITE(18,*) pr_crorb(icrorb,nk,1)%mat_rep(:,ib)
|
|
IF (ifSP)
|
|
& WRITE(18,*) pr_crorb(icrorb,nk,2)%mat_rep(:,ib)
|
|
WRITE(18,'()')
|
|
ENDDO
|
|
ELSEIF (reptrans(l,isrt)%ifmixing) THEN
|
|
WRITE(18,'(a,i4)') 'ik = ', 1
|
|
DO ib = kp(1,1)%nb_bot,kp(1,1)%nb_top
|
|
WRITE(18,*) pr_crorb(icrorb,1,1)%mat_rep(:,ib)
|
|
WRITE(18,'()')
|
|
ENDDO
|
|
WRITE(18,'(a,i4)') 'ik = ', nk
|
|
DO ib = kp(nk,1)%nb_bot,kp(nk,1)%nb_top
|
|
WRITE(18,*) pr_crorb(icrorb,nk,1)%mat_rep(:,ib)
|
|
WRITE(18,'()')
|
|
ENDDO
|
|
ELSE
|
|
WRITE(18,'(a,i4)') 'ik = ', 1
|
|
DO ib = kp(1,1)%nb_bot,kp(1,1)%nb_top
|
|
WRITE(18,*) pr_crorb(icrorb,1,1)%mat_rep(:,ib)
|
|
IF (ifSP)
|
|
& WRITE(18,*) pr_crorb(icrorb,1,2)%mat_rep(:,ib)
|
|
WRITE(18,'()')
|
|
ENDDO
|
|
WRITE(18,'(a,i4)') 'ik = ', nk
|
|
DO ib = kp(nk,1)%nb_bot,kp(nk,1)%nb_top
|
|
WRITE(18,*) pr_crorb(icrorb,nk,1)%mat_rep(:,ib)
|
|
IF (ifSP)
|
|
& WRITE(18,*) pr_crorb(icrorb,nk,2)%mat_rep(:,ib)
|
|
WRITE(18,'()')
|
|
ENDDO
|
|
ENDIF
|
|
ENDDO
|
|
C
|
|
DO iorb=1,norb
|
|
iatom=orb(iorb)%atom
|
|
isrt=orb(iorb)%sort
|
|
l=orb(iorb)%l
|
|
n=norm_radf(iorb)%n
|
|
WRITE(18,'()')
|
|
WRITE(18,'(a,i4)') 'iorb = ', iorb
|
|
WRITE(18,'(a,i4,a,i4)') 'isrt = ', isrt, ' l = ', l
|
|
IF (l==0) THEN
|
|
WRITE(18,'(a,i4)') 'ik = ', 1
|
|
DO i=1,n
|
|
WRITE(18,'(i4)') i
|
|
DO ib = kp(1,1)%nb_bot,kp(1,1)%nb_top
|
|
WRITE(18,*) pr_orb(iorb,1,1)%matn_rep(:,ib,i)
|
|
IF (ifSP)
|
|
& WRITE(18,*) pr_orb(iorb,1,2)%matn_rep(:,ib,i)
|
|
WRITE(18,'()')
|
|
ENDDO
|
|
ENDDO
|
|
WRITE(18,'(a,i4)') 'ik = ', nk
|
|
DO i=1,n
|
|
WRITE(18,'(i4)') i
|
|
DO ib = kp(nk,1)%nb_bot,kp(nk,1)%nb_top
|
|
WRITE(18,*) pr_orb(iorb,nk,1)%matn_rep(:,ib,i)
|
|
IF (ifSP)
|
|
& WRITE(18,*) pr_orb(iorb,nk,2)%matn_rep(:,ib,i)
|
|
WRITE(18,'()')
|
|
ENDDO
|
|
ENDDO
|
|
ELSEIF(reptrans(l,isrt)%ifmixing) THEN
|
|
DO i=1,n
|
|
WRITE(18,'(i4)') i
|
|
DO ib = kp(1,1)%nb_bot,kp(1,1)%nb_top
|
|
WRITE(18,*) pr_orb(iorb,1,1)%matn_rep(:,ib,i)
|
|
WRITE(18,'()')
|
|
ENDDO
|
|
ENDDO
|
|
WRITE(18,'(a,i4)') 'ik = ', nk
|
|
DO i=1,n
|
|
WRITE(18,'(i4)') i
|
|
DO ib = kp(nk,1)%nb_bot,kp(nk,1)%nb_top
|
|
WRITE(18,*) pr_orb(iorb,nk,1)%matn_rep(:,ib,i)
|
|
WRITE(18,'()')
|
|
ENDDO
|
|
ENDDO
|
|
ELSE
|
|
DO i=1,n
|
|
WRITE(18,'(i4)') i
|
|
DO ib = kp(1,1)%nb_bot,kp(1,1)%nb_top
|
|
WRITE(18,*) pr_orb(iorb,1,1)%matn_rep(:,ib,i)
|
|
IF (ifSP)
|
|
& WRITE(18,*) pr_orb(iorb,1,2)%matn_rep(:,ib,i)
|
|
WRITE(18,'()')
|
|
ENDDO
|
|
ENDDO
|
|
WRITE(18,'(a,i4)') 'ik = ', nk
|
|
DO i=1,n
|
|
WRITE(18,'(i4)') i
|
|
DO ib = kp(nk,1)%nb_bot,kp(nk,1)%nb_top
|
|
WRITE(18,*) pr_orb(iorb,nk,1)%matn_rep(:,ib,i)
|
|
IF (ifSP)
|
|
& WRITE(18,*) pr_orb(iorb,nk,2)%matn_rep(:,ib,i)
|
|
WRITE(18,'()')
|
|
ENDDO
|
|
ENDDO
|
|
ENDIF
|
|
ENDDO
|
|
C
|
|
RETURN
|
|
END
|
|
|