mirror of
https://github.com/triqs/dft_tools
synced 2024-08-17 01:31:44 +02:00
594 lines
26 KiB
FortranFixed
594 lines
26 KiB
FortranFixed
![]() |
|
||
|
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 orthogonal_wannier
|
||
|
C %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|
||
|
C %% %%
|
||
|
C %% This subroutine orthonormalizes the Wannier-like functions %%
|
||
|
C %% obtained with the projectors P(icrorb,ik,is), in order to %%
|
||
|
C %% get a set of "true" Wannier orbitals. %%
|
||
|
C %% %%
|
||
|
C %% Only the correlated orbitals are treated here. %%
|
||
|
C %% %%
|
||
|
C %% THIS VERSION CAN NOT BE USED WITH SPIN-ORBIT %%
|
||
|
C %% (since the calculation is made independently for up/dn states) %%
|
||
|
C %% THIS VERSION CAN BE USED WITH SPIN-POLARIZED INPUT FILES. %%
|
||
|
C %% %%
|
||
|
C %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|
||
|
|
||
|
C Definiton of the variables :
|
||
|
C ----------------------------
|
||
|
USE almblm_data
|
||
|
USE common_data
|
||
|
USE prnt
|
||
|
USE projections
|
||
|
USE reps
|
||
|
IMPLICIT NONE
|
||
|
COMPLEX(KIND=8), DIMENSION(:,:), ALLOCATABLE :: Dmat, D_orth, D
|
||
|
INTEGER :: is, ik, l, nbnd, ndim, isrt, nbbot, nbtop
|
||
|
INTEGER :: icrorb, ind1, ind2, ib, iatom
|
||
|
INTEGER :: m1, m2, irep
|
||
|
C
|
||
|
WRITE(buf,'(a)')'Orthonormalization of the projectors...'
|
||
|
CALL printout(0)
|
||
|
CALL printout(0)
|
||
|
C
|
||
|
IF(ncrorb==0) RETURN
|
||
|
C
|
||
|
C =====================================
|
||
|
C Creation of the overlap matrix Dmat :
|
||
|
C =====================================
|
||
|
C
|
||
|
C -----------------------------------------------------------
|
||
|
C Determination of the dimension ndim of the overlap matrix :
|
||
|
C -----------------------------------------------------------
|
||
|
ndim=0
|
||
|
C Loop on the correlated orbitals
|
||
|
DO icrorb=1,ncrorb
|
||
|
isrt=crorb(icrorb)%sort
|
||
|
l=crorb(icrorb)%l
|
||
|
C Since this subroutine is used only in the case without SO,
|
||
|
C the correlated ireps can be considered if there are any. (ifsplit=.TRUE.)
|
||
|
IF(crorb(icrorb)%ifsplit) THEN
|
||
|
C the value of l can not be 0 here, because ifsplit is necessary .FALSE.
|
||
|
C for s-orbital (restriction in dmftproj.f)
|
||
|
DO irep=1,reptrans(l,isrt)%nreps
|
||
|
IF(crorb(icrorb)%correp(irep))
|
||
|
& ndim=ndim+reptrans(l,isrt)%dreps(irep)
|
||
|
C The dimension of the irep is added to ndim.
|
||
|
ENDDO
|
||
|
ELSE
|
||
|
C If no particular irep is considered (ifsplit=.FALSE.),
|
||
|
C The whole matrix of the representation is considered.
|
||
|
ndim=ndim+2*l+1
|
||
|
ENDIF
|
||
|
ENDDO
|
||
|
C ------------------
|
||
|
C Creation of Dmat :
|
||
|
C ------------------
|
||
|
ALLOCATE(Dmat(1:ndim,1:ndim))
|
||
|
C
|
||
|
C =====================================================================
|
||
|
C Computation of the orthonormalized Wannier functions and projectors :
|
||
|
C =====================================================================
|
||
|
C The computation is performed for each k_point and each spin-value independently
|
||
|
C because they are good quantum numbers.
|
||
|
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
|
||
|
nbnd=kp(ik,is)%nb_top-kp(ik,is)%nb_bot+1
|
||
|
nbbot=kp(ik,is)%nb_bot
|
||
|
nbtop=kp(ik,is)%nb_top
|
||
|
ALLOCATE(D(1:ndim,1:nbnd))
|
||
|
C
|
||
|
C --------------------------------
|
||
|
C Initialization of the D matrix :
|
||
|
C --------------------------------
|
||
|
C This D matrix of size ndim*nbnd is the complete "projector matrix"
|
||
|
C which enables to go from the Wannier-like basis |u_orb> to the Bloch states |ik,ib>.
|
||
|
ind1=0
|
||
|
DO icrorb=1,ncrorb
|
||
|
isrt=crorb(icrorb)%sort
|
||
|
l=crorb(icrorb)%l
|
||
|
C If l=0, there only possible irep is the whole matrix itself.
|
||
|
IF (l==0) THEN
|
||
|
D(ind1+1,1:nbnd)=pr_crorb(icrorb,ik,is)%
|
||
|
& mat_rep(1,nbbot:nbtop)
|
||
|
ind1=ind1+1
|
||
|
ELSE
|
||
|
C the projectors of the correlated ireps are considered if there are any. (ifsplit=.TRUE.)
|
||
|
IF(crorb(icrorb)%ifsplit) THEN
|
||
|
C the value of l can not be 0 here, because ifsplit is necessary .FALSE.
|
||
|
C for s-orbital (restriction in dmftproj.f)
|
||
|
m1=-l-1
|
||
|
DO irep=1,reptrans(l,isrt)%nreps
|
||
|
IF(crorb(icrorb)%correp(irep)) THEN
|
||
|
m2=m1+reptrans(l,isrt)%dreps(irep)
|
||
|
ind2=ind1+reptrans(l,isrt)%dreps(irep)
|
||
|
C Since there is no SO, prcrorb%matrep is of size 2*l+1, from -l to l
|
||
|
C (the basis which mix up/dn states are not possible here.)
|
||
|
C The states range from m1+1 to m2 in the irep.
|
||
|
C The corresponding projector is stored from the line (ind1+1) to the line ind2, in the D matrix.
|
||
|
D(ind1+1:ind2,1:nbnd)=pr_crorb(icrorb,ik,is)%
|
||
|
& mat_rep(m1+1:m2,nbbot:nbtop)
|
||
|
ind1=ind2
|
||
|
ENDIF
|
||
|
m1=m1+reptrans(l,isrt)%dreps(irep)
|
||
|
ENDDO
|
||
|
ELSE
|
||
|
C The projectors of the whole correlated representation is considered. (ifsplit=.FALSE.)
|
||
|
ind2=ind1+2*l+1
|
||
|
C Since there is no SO, prcrorb%matrep is of size 2*l+1, from -l to l.
|
||
|
C (the basis which mix up/dn states are not possible here.)
|
||
|
C The corresponding projection matrix is stored from the line (ind1+1) to the line ind2, in the D matrix.
|
||
|
D(ind1+1:ind2,1:nbnd)=pr_crorb(icrorb,ik,is)%
|
||
|
& mat_rep(-l:l,nbbot:nbtop)
|
||
|
ind1=ind2
|
||
|
ENDIF ! End of the ifsplit if-then-else
|
||
|
ENDIF ! End of the l=0 if-then-else
|
||
|
ENDDO ! End of the icrorb loop
|
||
|
C
|
||
|
C ----------------------------------------
|
||
|
C Computation of the overlap matrix Dmat :
|
||
|
C ----------------------------------------
|
||
|
C The overlap matrix is stored in Dmat = D*transpose(conjugate(D))
|
||
|
CALL ZGEMM('N','C',ndim,ndim,nbnd,DCMPLX(1.D0,0.D0),
|
||
|
& D,ndim,D,ndim,DCMPLX(0.D0,0.D0),Dmat,ndim)
|
||
|
C
|
||
|
C -------------------------------------------
|
||
|
C Computation of the matrix S = Dmat^{-1/2} :
|
||
|
C -------------------------------------------
|
||
|
CALL orthogonal_h(Dmat,ndim,.TRUE.)
|
||
|
C This matrix is stored in Dmat.
|
||
|
C
|
||
|
C -----------------------------------------------
|
||
|
C Computation of the orthonormalized projectors :
|
||
|
C -----------------------------------------------
|
||
|
C The calculation performed is the following : P=O^(-1/2)*P_tilde.
|
||
|
C Its value is stored in the matrix D_orth (of size ndim*nbnd)
|
||
|
|
||
|
ALLOCATE(D_orth(1:ndim,1:nbnd))
|
||
|
CALL ZGEMM('N','N',ndim,nbnd,ndim,DCMPLX(1.D0,0.D0),
|
||
|
& Dmat,ndim,D,ndim,DCMPLX(0.D0,0.D0),D_orth,ndim)
|
||
|
DEALLOCATE(D)
|
||
|
C
|
||
|
C --------------------------------------------------------------------------------
|
||
|
C Storing the value of the orthonormalized projectors in the pr_crorb structures :
|
||
|
C --------------------------------------------------------------------------------
|
||
|
ind1=0
|
||
|
DO icrorb=1,ncrorb
|
||
|
isrt=crorb(icrorb)%sort
|
||
|
l=crorb(icrorb)%l
|
||
|
C If l=0, there only possible irep is the whole matrix itself.
|
||
|
IF (l==0) THEN
|
||
|
pr_crorb(icrorb,ik,is)%mat_rep
|
||
|
& (1,nbbot:nbtop)=D_orth(ind1+1,1:nbnd)
|
||
|
ind1=ind1+1
|
||
|
ELSE
|
||
|
C the projectors of the correlated ireps are considered if there are any. (ifsplit=.TRUE.)
|
||
|
IF(crorb(icrorb)%ifsplit) THEN
|
||
|
C the value of l can not be 0 here, because ifsplit is necessary .FALSE.
|
||
|
C for s-orbital (restriction in dmftproj.f)
|
||
|
m1=-l-1
|
||
|
DO irep=1,reptrans(l,isrt)%nreps
|
||
|
IF(crorb(icrorb)%correp(irep)) THEN
|
||
|
m2=m1+reptrans(l,isrt)%dreps(irep)
|
||
|
ind2=ind1+reptrans(l,isrt)%dreps(irep)
|
||
|
C prcrorb%matrep is of size 2*l+1, from -l to l (the basis which mix up/dn states are not possible here.)
|
||
|
C In the D_orth matrix, the corresponding part of the projection matrix ranges from the line (ind1+1) to the line ind2.
|
||
|
C The projector associated to the ireps is stored in the prcrorb%matrep from m1+1 to m2.
|
||
|
pr_crorb(icrorb,ik,is)%
|
||
|
& mat_rep(m1+1:m2,nbbot:nbtop)=
|
||
|
& D_orth(ind1+1:ind2,1:nbnd)
|
||
|
ind1=ind2
|
||
|
ENDIF
|
||
|
m1=m1+reptrans(l,isrt)%dreps(irep)
|
||
|
ENDDO
|
||
|
ELSE
|
||
|
C The projectors of the whole correlated representation is considered. (ifsplit=.FALSE.)
|
||
|
ind2=ind1+2*l+1
|
||
|
C Since there is no SO, prcrorb%matrep is of size 2*l+1, from -l to l.
|
||
|
C (the basis which mix up/dn states are not possible here.)
|
||
|
C In the D_orth matrix, the projection matrix ranges from the line (ind1+1) to the line ind2.
|
||
|
C The projector is stored in the pr_crorb%matrep (from -l to l).
|
||
|
pr_crorb(icrorb,ik,is)%mat_rep
|
||
|
& (-l:l,nbbot:nbtop)=D_orth(ind1+1:ind2,1:nbnd)
|
||
|
ind1=ind2
|
||
|
ENDIF ! End of the ifsplit if-then-else
|
||
|
ENDIF ! End of the l=0 if-then-else
|
||
|
ENDDO ! End of the icrorb loop
|
||
|
C prcrorb%matrep contains now the orthonormalized projectors.
|
||
|
DEALLOCATE(D_orth)
|
||
|
ENDDO ! End of the loop on is
|
||
|
ENDDO ! End of the loop on ik
|
||
|
DEALLOCATE(Dmat)
|
||
|
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)') 'apres othonormalizsation'
|
||
|
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
|
||
|
RETURN
|
||
|
END
|
||
|
|
||
|
|
||
|
|
||
|
SUBROUTINE orthogonal_wannier_SO
|
||
|
C %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|
||
|
C %% %%
|
||
|
C %% This subroutine orthonormalizes the Wannier-like functions %%
|
||
|
C %% obtained with the projectors P(icrorb,ik,is), in order to %%
|
||
|
C %% get a set of "true" Wannier orbitals. %%
|
||
|
C %% %%
|
||
|
C %% Only the correlated orbitals are treated here. %%
|
||
|
C %% %%
|
||
|
C %% THIS VERSION MUST BE USED WITH SPIN-ORBIT %%
|
||
|
C %% (since the calculation for up/dn states is made simultaneously) %%
|
||
|
C %% THIS VERSION CAN NOT BE USED WITHOUT SPIN-POLARIZED INPUT FILES.%%
|
||
|
C %% %%
|
||
|
C %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|
||
|
|
||
|
C Definiton of the variables :
|
||
|
C ----------------------------
|
||
|
USE almblm_data
|
||
|
USE common_data
|
||
|
USE prnt
|
||
|
USE projections
|
||
|
USE reps
|
||
|
IMPLICIT NONE
|
||
|
COMPLEX(KIND=8), DIMENSION(:,:), ALLOCATABLE :: Dmat, D_orth, D
|
||
|
INTEGER :: is, ik, l, nbnd, ndim, isrt, nbbot, nbtop
|
||
|
INTEGER :: icrorb, ind1, ind2, iatom, ib
|
||
|
INTEGER :: m1, m2, irep
|
||
|
C
|
||
|
WRITE(buf,'(a)')'Orthonormalization of the projectors...'
|
||
|
CALL printout(0)
|
||
|
CALL printout(0)
|
||
|
C ---------------------------------------------------------------------------------------
|
||
|
C Interruption of the prgm if there is no dn part of pr_crorb.
|
||
|
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 =====================================
|
||
|
C Creation of the overlap matrix Dmat :
|
||
|
C =====================================
|
||
|
C
|
||
|
C -----------------------------------------------------------
|
||
|
C Determination of the dimension ndim of the overlap matrix :
|
||
|
C -----------------------------------------------------------
|
||
|
ndim=0
|
||
|
C Loop on the correlated orbitals
|
||
|
DO icrorb=1,ncrorb
|
||
|
isrt=crorb(icrorb)%sort
|
||
|
l=crorb(icrorb)%l
|
||
|
C The case l=0 is a particular case of "non-mixing" basis.
|
||
|
C --------------------------------------------------------
|
||
|
IF (l==0) THEN
|
||
|
C Since this subroutine is used only in the case with SO,
|
||
|
C the only irep possible for s-orbital is the matrix itself.
|
||
|
ndim=ndim+2
|
||
|
C If the basis representation needs a complete spinor rotation approach (basis with "mixing" ).
|
||
|
C ---------------------------------------------------------------------------------------------
|
||
|
ELSEIF (reptrans(l,isrt)%ifmixing) THEN
|
||
|
C the projectors of the correlated ireps are considered if there are any. (ifsplit=.TRUE.)
|
||
|
IF(crorb(icrorb)%ifsplit) THEN
|
||
|
DO irep=1,reptrans(l,isrt)%nreps
|
||
|
IF(crorb(icrorb)%correp(irep)) THEN
|
||
|
ndim=ndim+reptrans(l,isrt)%dreps(irep)
|
||
|
ENDIF
|
||
|
C The dimension of the irep is added to ndim.
|
||
|
ENDDO
|
||
|
ELSE
|
||
|
C If no particular irep is considered (ifsplit=.FALSE.),
|
||
|
C The whole matrix of the representation is considered.
|
||
|
ndim=ndim+2*(2*l+1)
|
||
|
ENDIF
|
||
|
C If the basis representation can be reduce to the up/up block (basis without "mixing").
|
||
|
C --------------------------------------------------------------------------------------
|
||
|
ELSE
|
||
|
C Since this subroutine is used only in the case with SO,
|
||
|
C the only irep possible for this orbital is the matrix itself.
|
||
|
ndim=ndim+2*(2*l+1)
|
||
|
ENDIF
|
||
|
ENDDO
|
||
|
C ------------------
|
||
|
C Creation of Dmat :
|
||
|
C ------------------
|
||
|
ALLOCATE(Dmat(1:ndim,1:ndim))
|
||
|
C
|
||
|
C =====================================================================
|
||
|
C Computation of the orthonormalized Wannier functions and projectors :
|
||
|
C =====================================================================
|
||
|
C The computation is performed for each k_point independently
|
||
|
C because they are still good quantum numbers.
|
||
|
DO ik=1,nk
|
||
|
C Only the k-points with inlcuded bands are considered for the projectors.
|
||
|
IF(.NOT.kp(ik,1)%included) CYCLE
|
||
|
nbnd=kp(ik,1)%nb_top-kp(ik,1)%nb_bot+1
|
||
|
nbbot=kp(ik,1)%nb_bot
|
||
|
nbtop=kp(ik,1)%nb_top
|
||
|
C it was checked that nbtop(up)=nbtop(dn) and nbbot(up)=nbbot(dn)
|
||
|
C for a computation with SO [in set_projections.f]
|
||
|
ALLOCATE(D(1:ndim,1:nbnd))
|
||
|
C
|
||
|
C --------------------------------
|
||
|
C Initialization of the D matrix :
|
||
|
C --------------------------------
|
||
|
C This D matrix of size ndim*nbnd is the complete "projector matrix"
|
||
|
C which enables to go from the Wannier-like basis |u_orb> to the Bloch states |ik,ib>.
|
||
|
ind1=0
|
||
|
DO icrorb=1,ncrorb
|
||
|
isrt=crorb(icrorb)%sort
|
||
|
l=crorb(icrorb)%l
|
||
|
C The case l=0 is a particular case of "non-mixing" basis.
|
||
|
C --------------------------------------------------------
|
||
|
IF (l==0) THEN
|
||
|
C the only irep possible for s-orbital is the matrix itself.
|
||
|
DO is=1,ns
|
||
|
C D(ind1,1:nbnd)=
|
||
|
C Bug correction 8.11.2012
|
||
|
D(ind1+1,1:nbnd)=
|
||
|
& pr_crorb(icrorb,ik,is)%mat_rep(1,nbbot:nbtop)
|
||
|
ind1=ind1+1
|
||
|
ENDDO
|
||
|
C If the basis representation needs a complete spinor rotation approach (basis with "mixing" ).
|
||
|
C ---------------------------------------------------------------------------------------------
|
||
|
ELSEIF (reptrans(l,isrt)%ifmixing) THEN
|
||
|
C In this case, the projection matrix is stored in prcrorb%matrep with is=1.
|
||
|
C the projectors of the correlated ireps are considered if there are any. (ifsplit=.TRUE.)
|
||
|
IF (crorb(icrorb)%ifsplit) THEN
|
||
|
m1=0
|
||
|
DO irep=1,reptrans(l,isrt)%nreps
|
||
|
IF (crorb(icrorb)%correp(irep)) THEN
|
||
|
m2=m1+reptrans(l,isrt)%dreps(irep)
|
||
|
ind2=ind1+reptrans(l,isrt)%dreps(irep)
|
||
|
C The states range from m1+1 to m2 in the irep.
|
||
|
C The corresponding projector is stored from the line (ind1+1) to the line ind2, in the D matrix.
|
||
|
D(ind1+1:ind2,1:nbnd)=pr_crorb(icrorb,ik,1)%
|
||
|
& mat_rep(m1+1:m2,nbbot:nbtop)
|
||
|
ind1=ind2
|
||
|
ENDIF
|
||
|
m1=m1+reptrans(l,isrt)%dreps(irep)
|
||
|
ENDDO
|
||
|
ELSE
|
||
|
C The projectors of the whole correlated representation is considered. (ifsplit=.FALSE.)
|
||
|
ind2=ind1+2*(2*l+1)
|
||
|
C The corresponding projection matrix is stored from the line (ind1+1) to the line ind2, in the D matrix.
|
||
|
D(ind1+1:ind2,1:nbnd)=pr_crorb(icrorb,ik,1)%
|
||
|
& mat_rep(1:2*(2*l+1),nbbot:nbtop)
|
||
|
ind1=ind2
|
||
|
ENDIF ! End of the ifsplit if-then-else
|
||
|
C If the basis representation can be reduce to the up/up block (basis without "mixing").
|
||
|
C --------------------------------------------------------------------------------------
|
||
|
ELSE
|
||
|
C the only irep possible for such an orbital is the matrix itself.
|
||
|
DO is=1,ns
|
||
|
ind2=ind1+2*l+1
|
||
|
D(ind1+1:ind2,1:nbnd)=
|
||
|
& pr_crorb(icrorb,ik,is)%mat_rep(-l:l,nbbot:nbtop)
|
||
|
ind1=ind2
|
||
|
ENDDO
|
||
|
ENDIF ! End of the ifmixing if-then-else
|
||
|
ENDDO ! End of the icrorb loop
|
||
|
C
|
||
|
C ----------------------------------------
|
||
|
C Computation of the overlap matrix Dmat :
|
||
|
C ----------------------------------------
|
||
|
C The overlap matrix is stored in Dmat = D*transpose(conjugate(D))
|
||
|
CALL ZGEMM('N','C',ndim,ndim,nbnd,DCMPLX(1.D0,0.D0),
|
||
|
& D,ndim,D,ndim,DCMPLX(0.D0,0.D0),Dmat,ndim)
|
||
|
C
|
||
|
C -------------------------------------------
|
||
|
C Computation of the matrix S = Dmat^{-1/2} :
|
||
|
C -------------------------------------------
|
||
|
CALL orthogonal_h(Dmat,ndim,.TRUE.)
|
||
|
C This matrix is stored in Dmat.
|
||
|
C
|
||
|
C -----------------------------------------------
|
||
|
C Computation of the orthonormalized projectors :
|
||
|
C -----------------------------------------------
|
||
|
C The calculation performed is the following : P=O^(-1/2)*P_tilde.
|
||
|
C Its value is stored in the matrix D_orth (of size ndim*nbnd)
|
||
|
ALLOCATE(D_orth(1:ndim,1:nbnd))
|
||
|
CALL ZGEMM('N','N',ndim,nbnd,ndim,DCMPLX(1.D0,0.D0),
|
||
|
& Dmat,ndim,D,ndim,DCMPLX(0.D0,0.D0),D_orth,ndim)
|
||
|
DEALLOCATE(D)
|
||
|
C
|
||
|
C --------------------------------------------------------------------------------
|
||
|
C Storing the value of the orthonormalized projectors in the pr_crorb structures :
|
||
|
C --------------------------------------------------------------------------------
|
||
|
ind1=0
|
||
|
DO icrorb=1,ncrorb
|
||
|
isrt=crorb(icrorb)%sort
|
||
|
l=crorb(icrorb)%l
|
||
|
C The case l=0 is a particular case of "non-mixing" basis.
|
||
|
C --------------------------------------------------------
|
||
|
IF (l==0) THEN
|
||
|
C the only irep possible for s-orbital is the matrix itself.
|
||
|
DO is=1,ns
|
||
|
pr_crorb(icrorb,ik,is)%mat_rep(1,nbbot:nbtop)=
|
||
|
& D_orth(ind1+1,1:nbnd)
|
||
|
ind1=ind1+1
|
||
|
ENDDO
|
||
|
C If the basis representation needs a complete spinor rotation approach (basis with "mixing" ).
|
||
|
C ---------------------------------------------------------------------------------------------
|
||
|
ELSEIF (reptrans(l,isrt)%ifmixing) THEN
|
||
|
C the projectors of the correlated ireps are considered if there are any. (ifsplit=.TRUE.)
|
||
|
IF(crorb(icrorb)%ifsplit) THEN
|
||
|
m1=0
|
||
|
DO irep=1,reptrans(l,isrt)%nreps
|
||
|
IF (crorb(icrorb)%correp(irep)) THEN
|
||
|
m2=m1+reptrans(l,isrt)%dreps(irep)
|
||
|
ind2=ind1+reptrans(l,isrt)%dreps(irep)
|
||
|
C In the D_orth matrix, the corresponding part of the projection matrix ranges from the line (ind1+1) to the line ind2.
|
||
|
C The projector associated to the ireps is stored in the prcrorb%matrep from m1+1 to m2.
|
||
|
pr_crorb(icrorb,ik,1)%mat_rep(m1+1:m2,nbbot:nbtop)
|
||
|
& =D_orth(ind1+1:ind2,1:nbnd)
|
||
|
ind1=ind2
|
||
|
ENDIF
|
||
|
m1=m1+reptrans(l,isrt)%dreps(irep)
|
||
|
ENDDO
|
||
|
ELSE
|
||
|
C The projectors of the whole correlated representation is considered. (ifsplit=.FALSE.)
|
||
|
ind2=ind1+2*(2*l+1)
|
||
|
C The corresponding projection matrix is stored from the line (ind1+1) to the line ind2, in the D matrix.
|
||
|
pr_crorb(icrorb,ik,1)%mat_rep(1:2*(2*l+1),nbbot:nbtop)
|
||
|
& =D_orth(ind1+1:ind2,1:nbnd)
|
||
|
ind1=ind2
|
||
|
ENDIF ! End of the ifsplit if-then-else
|
||
|
C If the basis representation can be reduce to the up/up block (basis without "mixing").
|
||
|
C --------------------------------------------------------------------------------------
|
||
|
ELSE
|
||
|
C the only irep possible for this orbital is the matrix itself.
|
||
|
DO is=1,ns
|
||
|
ind2=ind1+2*l+1
|
||
|
pr_crorb(icrorb,ik,is)%mat_rep(-l:l,nbbot:nbtop)
|
||
|
& =D_orth(ind1+1:ind2,1:nbnd)
|
||
|
ind1=ind2
|
||
|
ENDDO
|
||
|
ENDIF ! End of the ifmixing if-then-else
|
||
|
ENDDO ! End of the icrorb loop
|
||
|
DEALLOCATE(D_orth)
|
||
|
ENDDO ! End of the loop on ik
|
||
|
DEALLOCATE(Dmat)
|
||
|
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)') 'apres othonormalizsation'
|
||
|
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
|
||
|
RETURN
|
||
|
END
|
||
|
|
||
|
|