exchange_val_hf.F90 23.5 KB
Newer Older
1 2 3 4 5 6
!--------------------------------------------------------------------------------
! Copyright (c) 2016 Peter Grünberg Institut, Forschungszentrum Jülich, Germany
! This file is part of FLEUR and available as free software under the conditions
! of the MIT license as expressed in the LICENSE file in more detail.
!--------------------------------------------------------------------------------

7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51
!     Calculates the HF exchange term 
!
!                                          s          s*          s            s*
!                                       phi    (r) phi     (r) phi     (r') phi    (r')
!                         occ.             n_1k       n'k+q       n'k+q        n_2k
!     exchange(n,q)  =  - SUM  INT INT  ------------------------------------------- dr dr'
!                         k,n'                           | r - r' |
!
!                         occ                  s          s    ~        ~       s         s
!                    =  - SUM  SUM  v     < phi      | phi     M    > < M    phi     | phi      >
!                         k,n' I,J   k,IJ      n'k+q      n_1k  q,I      q,J    n_2k      n'k+q
!
!     for the different combinations of n_1 and n_2 and where n' runs only over the valence states.     
!     ( n_1,n_2:  valence-valence, core-core,core-valence )
!
!
!     At the Gamma point (k=0) v diverges. After diagonalization of v at k=0 the divergence is
!     restricted to the head element I=1. Furthermore, we expand <...> with kp perturbation theory.
!     As a result, the total I=1 element is given by a sum of a divergent 1/k**2-term and an
!     angular dependent term. The former is separated from the numerical k-summation and treated
!     analytically while the latter is spherically averaged and added to the k=0 contribution of
!     the numerical k-summation. (A better knowledge of the integrand's behavior at the BZ edges
!     might further improve the integration.)
!
!     The divergence at the Gamma point is integrated with one of the following algorithms:
! (1) Switching-Off Function
!     In a sphere of radius k0=radshmin/2 a switching-off function g(k)=1-(k/k0)**n*(n+1-n*k/k0)
!     (n=npot) is defined. The 1/k**2 divergence is subtracted from the BZ integral in the form
!     g(k)/k**2 and integrated analytically. The non-divergent rest is integrated numerically.
! (2) Periodic Function (similar to the one used by Massidda PRB 48, 5058)
!     The function  F(k) = SUM(G) exp(-expo*|k+G|**3) / |k+G|**2  is subtracted from the BZ integral
!     and integrated analytically. The non-divergent rest is integrated numerically.
!     The parameter expo is chosen such that exp(-expo*q**3)=1/2
!     with q = radius of sphere with same volume as BZ.
! (3) Periodic Function (same as Massidda's) with expo->0
!     The function  F(k) = lim(expo->0) SUM(G) exp(-expo*|k+G|**2) / |k+G|**2  is subtracted from
!     the BZ integral and integrated analytically. The contribution to the BZ integral including
!     the "tail" is
!     vol/(8*pi**3) INT F(k) d^3k - P SUM(k) F(k)  ( P = principal value ) .
!     For expo->0 the two terms diverge. Therefore a cutoff radius q0 is introduced and related to
!     expo by exp(-expo*q0**2)=delta  ( delta = small value, e.g., delta = 1d-10 ) .
!     The resulting formula
!     vol/(4*pi**1.5*sqrt(expo)) * erf(sqrt(a)*q0) - sum(q,0<q<q0) exp(-expo*q**2)/q**2
!     converges well with q0. (Should be the default.)

52
MODULE m_exchange_valence_hf
53

54 55
   LOGICAL,PARAMETER:: zero_order=.false.,ibs_corr=.false.
   INTEGER,PARAMETER:: maxmem=600
56

57
CONTAINS
58

59 60 61
SUBROUTINE exchange_valence_hf(nk,kpts,nkpt_EIBZ,sym,atoms,hybrid,cell,dimension,input,jsp,hybdat,mnobd,lapw,&
                               eig_irr,results,parent,pointer_EIBZ,n_q,wl_iks,it,xcpot, noco,nsest,indx_sest,&
                               mpi,irank2,isize2,comm,mat_ex)
62

63 64 65 66 67 68 69 70
   USE m_types
   USE m_wrapper
   USE m_constants   
   USE m_trafo
   USE m_wavefproducts
   USE m_olap
   USE m_spmvec
   USE m_hsefunctional ,ONLY: dynamic_hse_adjustment
71
#if defined(CPP_MPI)&&defined(CPP_NEVER)
72 73
   USE m_mpi_work_dist
   USE m_mpi_tags
74
#endif
75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107
   USE m_io_hybrid
   USE m_kp_perturbation

   IMPLICIT NONE

   TYPE(t_hybdat),        INTENT(IN)    :: hybdat
   TYPE(t_results),       INTENT(IN)    :: results
   TYPE(t_xcpot_inbuild), INTENT(IN)    :: xcpot
   TYPE(t_mpi),           INTENT(IN)    :: mpi
   TYPE(t_dimension),     INTENT(IN)    :: dimension
   TYPE(t_hybrid),        INTENT(INOUT) :: hybrid
   TYPE(t_input),         INTENT(IN)    :: input
   TYPE(t_noco),          INTENT(IN)    :: noco
   TYPE(t_sym),           INTENT(IN)    :: sym
   TYPE(t_cell),          INTENT(IN)    :: cell
   TYPE(t_kpts),          INTENT(IN)    :: kpts
   TYPE(t_atoms),         INTENT(IN)    :: atoms
   TYPE(t_lapw),          INTENT(IN)    :: lapw
   TYPE(t_mat),           INTENT(INOUT) :: mat_ex

   ! scalars
   INTEGER,               INTENT(IN)    :: it,irank2,isize2,comm
   INTEGER,               INTENT(IN)    :: jsp
   INTEGER,               INTENT(IN)    :: nk,nkpt_EIBZ
   INTEGER,               INTENT(IN)    :: mnobd 

   ! arrays
   INTEGER,               INTENT(IN)    ::  n_q(nkpt_EIBZ)

   INTEGER,               INTENT(IN)    ::  parent(kpts%nkptf)
   INTEGER,               INTENT(IN)    ::  pointer_EIBZ(nkpt_EIBZ)
   INTEGER,               INTENT(IN)    ::  nsest(hybrid%nbands(nk))
   INTEGER,               INTENT(IN)    ::  indx_sest(hybrid%nbands(nk),hybrid%nbands(nk))
Daniel Wortmann's avatar
Daniel Wortmann committed
108 109

 
110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125
   REAL,                  INTENT(IN)    ::  eig_irr(dimension%neigd,kpts%nkpt)
   REAL,                  INTENT(IN)    ::  wl_iks(dimension%neigd,kpts%nkptf)    

   ! local scalars
   INTEGER                 ::  iband,iband1,ibando,ikpt,ikpt0
   INTEGER                 ::  i,ic,ix,iy,iz
   INTEGER                 ::  irecl_coulomb,irecl_coulomb1
   INTEGER                 ::  j
   INTEGER                 ::  m1,m2
   INTEGER                 ::  n,n1,n2,nn,nn2
   INTEGER                 ::  nkqpt
   INTEGER                 ::  npot
   INTEGER                 ::  ok
   INTEGER                 ::  psize
   REAL                    ::  rdum
   REAL                    ::  k0
Daniel Wortmann's avatar
Daniel Wortmann committed
126
     
127
   REAL , SAVE             ::  divergence
128

129 130
   COMPLEX                 ::  cdum,cdum1,cdum2 
   COMPLEX                 ::  exch0
Daniel Wortmann's avatar
Daniel Wortmann committed
131
 
132 133 134 135 136 137 138 139 140 141 142 143 144 145 146
   LOGICAL, SAVE           ::  initialize = .true.

   ! local arrays
   INTEGER              :: kcorner(3,8) = reshape((/ 0,0,0, 1,0,0, 0,1,0, 0,0,1, 1,1,0, 1,0,1, 0,1,1, 1,1,1 /), (/3,8/) )
   COMPLEX              :: exchcorrect(kpts%nkptf)
   COMPLEX              :: dcprod(hybrid%nbands(nk),hybrid%nbands(nk),3) 
   COMPLEX(8)           :: exch_vv(hybrid%nbands(nk),hybrid%nbands(nk))
   COMPLEX              :: hessian(3,3)
   COMPLEX              :: proj_ibsc(3,mnobd,hybrid%nbands(nk))
   COMPLEX              :: olap_ibsc(3,3,mnobd,mnobd)
   REAL                 :: carr1_v_r(hybrid%maxbasm1),carr1_c_r(hybrid%maxbasm1)
   COMPLEX              :: carr1_v_c(hybrid%maxbasm1),carr1_c_c(hybrid%maxbasm1)
   COMPLEX, ALLOCATABLE :: phase_vv(:,:)
   REAL,    ALLOCATABLE :: cprod_vv_r(:,:,:),cprod_cv_r(:,:,:), carr3_vv_r(:,:,:),carr3_cv_r(:,:,:)
   COMPLEX, ALLOCATABLE :: cprod_vv_c(:,:,:),cprod_cv_c(:,:,:), carr3_vv_c(:,:,:),carr3_cv_c(:,:,:)
147 148


149
#if defined(CPP_MPI)&&defined(CPP_NEVER)
150
   COMPLEX(8)           :: buf_vv(hybrid%nbands(nk),nbands(nk))
151 152
#endif

153 154 155 156 157 158
#if ( !defined CPP_NOSPMVEC && !defined CPP_IRAPPROX )
   REAL                 :: coulomb_mt1(hybrid%maxindxm1-1,hybrid%maxindxm1-1, 0:hybrid%maxlcutm1,atoms%ntype)       
   REAL                 :: coulomb_mt2_r(hybrid%maxindxm1-1,-hybrid%maxlcutm1:hybrid%maxlcutm1,0:hybrid%maxlcutm1+1,atoms%nat)
   REAL                 :: coulomb_mt3_r(hybrid%maxindxm1-1,atoms%nat,atoms%nat)
   COMPLEX              :: coulomb_mt2_c(hybrid%maxindxm1-1,-hybrid%maxlcutm1:hybrid%maxlcutm1,0:hybrid%maxlcutm1+1,atoms%nat)
   COMPLEX              :: coulomb_mt3_c(hybrid%maxindxm1-1,atoms%nat,atoms%nat)
159
#else
160 161
   REAL                 :: coulomb_r(hybrid%maxbasm1*(hybrid%maxbasm1+1)/2)
   COMPLEX              :: coulomb_c(hybrid%maxbasm1*(hybrid%maxbasm1+1)/2)
162 163 164
#endif

#ifdef CPP_IRCOULOMBAPPROX
165 166
   REAL                 :: coulomb_mtir_r((hybrid%maxlcutm1+1)**2*atoms%nat,&
                                          (hybrid%maxlcutm1+1)**2*atoms%nat+maxval(hybrid%ngptm))
167
#else
168 169
   REAL                 :: coulomb_mtir_r(((hybrid%maxlcutm1+1)**2*atoms%nat +maxval(hybrid%ngptm)) *&
                                          ((hybrid%maxlcutm1+1)**2*atoms%nat +maxval(hybrid%ngptm)+1)/2)
170 171 172
#endif

#ifdef CPP_IRCOULOMBAPPROX
173 174
   COMPLEX              :: coulomb_mtir_c((hybrid%maxlcutm1+1)**2*atoms%nat,&
                                          (hybrid%maxlcutm1+1)**2*atoms%nat+maxval(hybrid%ngptm))
175
#else
176 177
   COMPLEX              :: coulomb_mtir_c(((hybrid%maxlcutm1+1)**2*atoms%nat +maxval(hybrid%ngptm)) *&
                                          ((hybrid%maxlcutm1+1)**2*atoms%nat +maxval(hybrid%ngptm)+1)/2)
178 179
#endif

180
   LOGICAL              :: occup(dimension%neigd)
181
#if defined(CPP_MPI)&&defined(CPP_NEVER)
182 183 184
   INCLUDE "mpif.h"
   INTEGER              :: ierr,ierr2,length,rank
   CHARACTER(LEN=MPI_MAX_ERROR_STRING) :: errmsg
185
#endif
Daniel Wortmann's avatar
Daniel Wortmann committed
186
     
187 188 189 190 191
   IF(initialize) THEN !it .eq. 1 .and. nk .eq. 1) THEN
      call calc_divergence(cell,kpts,divergence)
      PRINT *,"Divergence:",divergence
      initialize = .false.
   END IF
Daniel Wortmann's avatar
Daniel Wortmann committed
192
   
193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209
   ! calculate valence-valence-valence-valence, core-valence-valence-valence
   ! and core-valence-valence-core exchange at current k-point
   ! the sum over the inner occupied valence states is restricted to the EIBZ(k)
   ! the contribution of the Gamma-point is treated separately (see below)

   ! determine package size loop over the occupied bands
   if (mat_ex%l_real) THEn
      rdum  = hybrid%maxbasm1*hybrid%nbands(nk)*4/1048576.
   else
      rdum  = hybrid%maxbasm1*hybrid%nbands(nk)*4/1048576.
   endif
   psize = 1
   DO iband = mnobd,1,-1
      ! ensure that the packages have equal size
      IF(modulo(mnobd,iband).eq.0) THEN
         ! choose packet size such that cprod is smaller than memory threshold
         IF(rdum*iband.le.maxmem) THEN
210 211
            psize = iband
            EXIT
212
         END IF
213
      END IF
214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239
   END DO

   IF(psize.ne.mnobd) THEN
      WRITE(6,'(A,A,i3,A,f7.2,A)') ' Divide the loop over the occupied hybrid%bands in packages',&
                                   ' of the size',psize,' (cprod=',rdum*psize,'MB)'
   END IF
   ALLOCATE( phase_vv(psize,hybrid%nbands(nk)),stat=ok )
   IF(ok.ne.0) STOP 'exchange_val_hf: error allocation phase'
   phase_vv=0
   IF(ok.ne.0) STOP 'exchange_val_hf: error allocation phase'

   if (mat_ex%l_real) THEN
      ALLOCATE( cprod_vv_c(0,0,0), carr3_vv_c(0,0,0))
      ALLOCATE( cprod_vv_r(hybrid%maxbasm1,psize,hybrid%nbands(nk)),stat=ok )
      IF( ok .ne. 0 ) STOP 'exchange_val_hf: error allocation cprod'
      ALLOCATE( carr3_vv_r(hybrid%maxbasm1,psize,hybrid%nbands(nk)),stat=ok )
      IF( ok .ne. 0 ) STOP 'exchange_val_hf: error allocation carr3'
      cprod_vv_r = 0 ; carr3_vv_r = 0 
   ELSE
      ALLOCATE( cprod_vv_r(0,0,0), carr3_vv_r(0,0,0))
      ALLOCATE( cprod_vv_c(hybrid%maxbasm1,psize,hybrid%nbands(nk)),stat=ok )
      IF( ok .ne. 0 ) STOP 'exchange_val_hf: error allocation cprod'
      ALLOCATE( carr3_vv_c(hybrid%maxbasm1,psize,hybrid%nbands(nk)),stat=ok )
      IF( ok .ne. 0 ) STOP 'exchange_val_hf: error allocation carr3'
      cprod_vv_c = 0 ; carr3_vv_c = 0
   END IF
Daniel Wortmann's avatar
Daniel Wortmann committed
240
         
241
   exch_vv = 0
242

243 244
   DO ikpt = 1,nkpt_EIBZ
      ikpt0 = pointer_EIBZ(ikpt)
245

246 247 248
      n  = hybrid%nbasp + hybrid%ngptm(ikpt0)
      IF( hybrid%nbasm(ikpt0).ne.n) STOP 'error hybrid%nbasm'
      nn = n*(n+1)/2
249

250
      ! read in coulomb matrix from direct access file coulomb
251
#if( !defined CPP_NOSPMVEC && !defined CPP_IRAPPROX )
252 253 254 255 256
      IF (mat_ex%l_real) THEN
         CALL read_coulomb_spm_r(kpts%bkp(ikpt0),coulomb_mt1,coulomb_mt2_r,coulomb_mt3_r,coulomb_mtir_r)
      ELSE
         CALL read_coulomb_spm_c(kpts%bkp(ikpt0),coulomb_mt1,coulomb_mt2_c,coulomb_mt3_c,coulomb_mtir_c)
      END IF
257
#else
258
	   call read_coulomb(kpts%bkp(ikpt0),coulomb)
259 260
#endif

261
      IF(kpts%bkp(ikpt0).ne.ikpt0) THEN
262
#if( !defined CPP_NOSPMVEC && !defined CPP_IRAPPROX )
263
         IF((kpts%bksym(ikpt0).gt.sym%nop).and.(.not.mat_ex%l_real)) THEN
Daniel Wortmann's avatar
Daniel Wortmann committed
264
            coulomb_mt2_c = conjg(coulomb_mt2_c)
265 266
            coulomb_mtir_c = conjg(coulomb_mtir_c)
         END IF
267
#else
268 269 270
         if (.not.mat_ex%l_real) THEN
            IF( kpts%bksym(ikpt0) .gt. sym%nop ) coulomb = conjg(coulomb)
         endif
271
#endif
272
      END IF
273

274
      DO ibando = 1, mnobd, psize
275

276
         IF (mat_ex%l_real) THEN
277
#ifdef CPP_IRAPPROX
278 279
            CALL wavefproducts_inv(1,hybdat,dimension,jsp,atoms,lapw,obsolete,kpts,nk,ikpt0,&
                                   mnobd,hybrid,parent,cell,sym,nkqpt,cprod_vv)
280
#else
281 282 283
            CALL wavefproducts_inv5(1,hybrid%nbands(nk),ibando,ibando+psize-1,dimension,input,jsp,atoms,&
                                    lapw,kpts,nk,ikpt0,hybdat,mnobd,hybrid,parent,cell,hybrid%nbasp,sym,&
                                    noco,nkqpt,cprod_vv_r)
284
#endif
285
         ELSE
286
#ifdef CPP_IRAPPROX
287 288
            CALL wavefproducts_noinv(1,hybdat,nk,ikpt0,dimension,jsp,cell,atoms,hybrid, 
                                     kpts,mnobd,lapw,sym,nkqpt,cprod_vv)
289
#else
290 291
            CALL wavefproducts_noinv5(1,hybrid%nbands(nk),ibando,ibando+psize-1,nk,ikpt0,dimension,input,jsp,&!jsp,&
                                      cell,atoms,hybrid,hybdat,kpts,mnobd,lapw,sym,hybrid%nbasp,noco,nkqpt,cprod_vv_c)
292
#endif
293
         END IF
294

295 296 297 298 299
         ! The sparse matrix technique is not feasible for the HSE
         ! functional. Thus, a dynamic adjustment is implemented
         ! The mixed basis functions and the potential difference
         ! are Fourier transformed, so that the exchange can be calculated
         ! in Fourier space
300
#ifndef CPP_NOSPMVEC
301
         IF (xcpot%is_name("hse").OR.xcpot%is_name("vhse")) THEN
Daniel Wortmann's avatar
Daniel Wortmann committed
302
            iband1  = hybrid%nobd(nkqpt)
303 304 305 306 307 308 309 310 311 312

            exch_vv = exch_vv +&
                      dynamic_hse_adjustment(atoms%rmsh,atoms%rmt,atoms%dx,atoms%jri,atoms%jmtd,kpts%bkf(:,ikpt0),ikpt0,&
                                             kpts%nkptf,cell%bmat,cell%omtil,atoms%ntype,atoms%neq,atoms%nat,atoms%taual,&
                                             hybrid%lcutm1,hybrid%maxlcutm1,hybrid%nindxm1,hybrid%maxindxm1,hybrid%gptm,&
                                             hybrid%ngptm(ikpt0),hybrid%pgptm(:,ikpt0),hybrid%gptmd,hybrid%basm1,&
                                             hybrid%nbasm(ikpt0),iband1,hybrid%nbands(nk),nsest,ibando,psize,indx_sest,&
                                             atoms%invsat,sym%invsatnr,mpi%irank,cprod_vv_r(:hybrid%nbasm(ikpt0),:,:),&
                                             cprod_vv_c(:hybrid%nbasm(ikpt0),:,:),mat_ex%l_real,wl_iks(:iband1,nkqpt),n_q(ikpt))
         END IF
313 314
#endif

315 316 317 318 319 320 321 322 323 324 325 326 327 328
         ! the Coulomb matrix is only evaluated at the irrecuible k-points
         ! bra_trafo transforms cprod instead of rotating the Coulomb matrix
         ! from IBZ to current k-point
         IF( kpts%bkp(ikpt0) .ne. ikpt0 ) THEN
            CALL bra_trafo2(mat_ex%l_real,carr3_vv_r(:hybrid%nbasm(ikpt0),:,:),cprod_vv_r(:hybrid%nbasm(ikpt0),:,:),&
                            carr3_vv_c(:hybrid%nbasm(ikpt0),:,:),cprod_vv_c(:hybrid%nbasm(ikpt0),:,:),&
                            hybrid%nbasm(ikpt0),psize,hybrid%nbands(nk),kpts%bkp(ikpt0),ikpt0,kpts%bksym(ikpt0),sym,&
                            hybrid,kpts,cell,atoms,phase_vv)
            IF (mat_ex%l_real) THEN
               cprod_vv_r(:hybrid%nbasm(ikpt0),:,:) = carr3_vv_r(:hybrid%nbasm(ikpt0),:,:)
            ELSE
               cprod_vv_c(:hybrid%nbasm(ikpt0),:,:) = carr3_vv_c(:hybrid%nbasm(ikpt0),:,:)
            ENDIF
         ELSE
329
            phase_vv(:,:) = (1d0,0d0)
330
         END IF
331

332
         ! calculate exchange matrix at ikpt0
333

334
         DO n1=1,hybrid%nbands(nk)
335
            DO iband = 1,psize
336 337 338
               IF((ibando+iband-1).gt.hybrid%nobd(nkqpt)) CYCLE

               cdum  = wl_iks(ibando+iband-1,nkqpt) * conjg(phase_vv(iband,n1))/n_q(ikpt)
339 340

#if( !defined CPP_NOSPMVEC && !defined CPP_IRAPPROX )
341 342 343 344 345 346 347 348 349
               IF (mat_ex%l_real) THEN
                  carr1_v_r(:n) = 0 
                  CALL spmvec_invs(atoms,hybrid,hybdat,ikpt0,kpts,cell,coulomb_mt1,coulomb_mt2_r,coulomb_mt3_r,&
                                   coulomb_mtir_r,cprod_vv_r(:n,iband,n1),carr1_v_r(:n))
               ELSE
                  carr1_v_c(:n) = 0 
                  CALL spmvec_noinvs(atoms,hybrid,hybdat,ikpt0,kpts,cell,coulomb_mt1,coulomb_mt2_c,coulomb_mt3_c,&
                                     coulomb_mtir_c,cprod_vv_c(:n,iband,n1),carr1_v_c(:n))
               END IF
350
#else
351 352 353 354 355
               IF (mat_ex%l_real) THEN
                  carr1_v_r(:n) = matvec( coulomb_r(:nn),cprod_vv_r(:n,iband,n1) )
               ELSE
                  carr1_v_r(:n) = matvec( coulomb_c(:nn),cprod_vv_c(:n,iband,n1) )
               END IF
356 357
#endif

358 359 360 361 362 363 364 365 366 367 368 369 370
               IF (mat_ex%l_real) THEN
                  DO n2=1,nsest(n1)!n1
                     nn2 = indx_sest(n2,n1)
                     exch_vv(nn2,n1) = exch_vv(nn2,n1) + cdum*phase_vv(iband,nn2) *&
                                                         dotprod(carr1_v_r(:n),cprod_vv_r(:n,iband,nn2))
                  END DO !n2
               ELSE
                  DO n2=1,nsest(n1)!n1
                     nn2 = indx_sest(n2,n1)
                     exch_vv(nn2,n1) = exch_vv(nn2,n1) + cdum*phase_vv(iband,nn2) *&
                                                         dotprod(carr1_v_c(:n),cprod_vv_c(:n,iband,nn2))
                  END DO !n2
               END IF
371
            END DO
372 373 374
         END DO  !n1
      END DO !ibando
   END DO  !ikpt
375

376
   ! add contribution of the gamma point to the different cases (exch_vv,exch_cv,exch_cc)
377

378 379 380 381 382 383 384 385 386 387 388 389 390 391 392
   ! valence-valence-valence-valence exchange

   IF ((.not.xcpot%is_name("hse")).AND.(.not.xcpot%is_name("vhse"))) THEN ! no gamma point correction needed for HSE functional
      IF( zero_order .and. .not. ibs_corr ) THEN
         WRITE(6,'(A)') ' Take zero order terms into account.'
      ELSE IF( zero_order .and.  ibs_corr ) THEN
         WRITE(6,'(A)') ' Take zero order terms and ibs-correction into account.'
      END IF

      IF(zero_order) THEN
         CALL dwavefproducts(dcprod,nk,1,hybrid%nbands(nk),1,hybrid%nbands(nk),.false.,atoms,hybrid,&
                             cell,hybdat,kpts,kpts%nkpt,lapw,dimension,jsp,eig_irr)

         ! make dcprod hermitian
         DO n1 = 1, hybrid%nbands(nk)
393
            DO n2 = 1,n1
394 395
               dcprod(n1,n2,:) = (dcprod(n1,n2,:) - conjg(dcprod(n2,n1,:)))/2   
               dcprod(n2,n1,:) = -conjg(dcprod(n1,n2,:))
396
            END DO
397 398 399 400 401 402 403
         END DO

         IF(ibs_corr) THEN
            CALL ibs_correction(nk,atoms,dimension,input,jsp,hybdat,hybrid,lapw,kpts,kpts%nkpt,cell,mnobd,&
                                sym,proj_ibsc,olap_ibsc)
         END IF
      END IF
Daniel Wortmann's avatar
Daniel Wortmann committed
404
        
405 406 407 408
      !This should be done with w_iks I guess!TODO
      occup = .false.
      DO i=1,hybrid%ne_eig(nk)
         IF (results%ef.ge.eig_irr(i,nk)) THEN
409
            occup(i) = .true.
410 411 412 413
         ELSE IF ((eig_irr(i,nk)-results%ef).le.1E-06) THEN
            occup(i) = .true.
         END IF
      END DO
414

415 416
      DO n1 = 1, hybrid%nbands(nk)
         DO n2 = 1, nsest(n1)!n1
417 418
            nn2 = indx_sest(n2,n1)
            exchcorrect = 0
419 420 421
            exch0 = 0

            ! if zero_order = .true. add averaged k-dependent term to the numerical integration at Gamma-point contribution
422 423 424 425 426 427

            ! if we start with a system with a small DFT band gap (like GaAs), the contribution
            ! of the highest occupied and lowest unoccupied state in Hessian is typically
            ! large; a correct numerical integration requires a dense k-point mesh, so
            ! we don't add the contribution exchcorrect for such materials 

428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446
            IF(zero_order) THEN
               hessian = 0
               IF(occup(n1).and.occup(nn2)) THEN
                  DO i = 1,3
                     j = i
                     DO iband = 1, hybrid%nbands(nk)
                        IF(occup(iband)) THEN
                           hessian(i,j) = hessian(i,j) + conjg(dcprod(iband,n1,i)) *dcprod(iband,nn2,j)
                        END IF
                        hessian(i,j) = hessian(i,j) - dcprod(iband,nn2,i) * conjg(dcprod(iband,n1,j))
                     END DO

                     ! ibs correction
                     IF(ibs_corr) THEN 
                        hessian(i,j) = hessian(i,j) - olap_ibsc(i,j,n1,nn2)/cell%omtil
                        DO iband = 1,hybrid%nbands(nk)
                           hessian(i,j) = hessian(i,j) + conjg(proj_ibsc(i,nn2,iband)) * proj_ibsc(j,n1,iband)/cell%omtil
                        END DO
                     END IF
447
                  END DO
448 449 450 451 452 453 454 455
               ELSE
                  DO i = 1,3
                     j = i 
                     DO iband = 1, hybrid%nbands(nk)
                        IF(occup(iband)) THEN
                           hessian(i,j) = hessian(i,j) + conjg(dcprod(iband,n1,i)) * dcprod(iband,nn2,j)
                        END IF
                     END DO
456
                  END DO
457
               END IF
Daniel Wortmann's avatar
Daniel Wortmann committed
458
 
459 460
               exchcorrect(1) = fpi_const/3 * (hessian(1,1)+hessian(2,2)+hessian(3,3))
               exch0 = exchcorrect(1)/kpts%nkptf
461 462 463 464 465 466 467 468 469
            END IF

            ! tail correction/contribution from all other k-points (it  goes into exchcorrect )

            ! Analytic contribution

            cdum2 = 0
            !multiply divergent contribution with occupation number;
            !this only affects metals 
470
            IF (n1.eq.nn2) THEN
Daniel Wortmann's avatar
Daniel Wortmann committed
471
               cdum2 = fpi_const/cell%omtil * divergence * wl_iks(n1,nk)*kpts%nkptf
472 473 474 475
            END IF

            ! due to the symmetrization afterwards the factor 1/n_q(1) must be added

476
            IF(n1.EQ.nn2) hybrid%div_vv(n1,nk,jsp) = REAL(cdum2) 
477 478
            exch_vv(nn2,n1)  = exch_vv(nn2,n1) + (exch0 + cdum2)/n_q(1)

479 480 481
         END DO !n2
      END DO !n1
   END IF ! xcpot%icorr .ne. icorr_hse
482 483


484 485 486 487
   IF (.not.mat_ex%l_real) THEN
      IF(any(abs(aimag(exch_vv)).gt.1E-08)) CALL judft_warn('unusally large imaginary part of exch_vv',&
                                                            calledby='exchange_val_hf.F90')
   END IF
488

489 490 491 492 493 494 495
   ! write exch_vv in mat_ex
   CALL mat_ex%alloc(matsize1=hybrid%nbands(nk))
   IF (mat_ex%l_real) THEN
      mat_ex%data_r=exch_vv
   ELSE
      mat_ex%data_c=exch_vv
   END IF
Daniel Wortmann's avatar
Daniel Wortmann committed
496
     
497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512
END SUBROUTINE exchange_valence_hf




SUBROUTINE calc_divergence(cell,kpts,divergence)

   USE m_util, ONLY: cerf
   USE m_types
   USE m_constants

   IMPLICIT NONE

   TYPE(t_cell), INTENT(IN)  :: cell
   TYPE(t_kpts), INTENT(IN)  :: kpts
   REAL,         INTENT(OUT) :: divergence
Daniel Wortmann's avatar
Daniel Wortmann committed
513
        
514 515 516 517
   INTEGER :: ix,iy,iz,sign,n
   logical :: found
   REAL    :: expo,rrad,k(3),kv1(3),kv2(3),kv3(3),knorm2
   COMPLEX :: cdum
Daniel Wortmann's avatar
Daniel Wortmann committed
518
        
519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554
   expo       = 5d-3
   rrad       = sqrt(-log(5d-3)/expo)
   cdum       = sqrt(expo)*rrad
   divergence = cell%omtil / (tpi_const**2) * sqrt(pi_const/expo) * cerf(cdum)
   rrad       = rrad**2
   kv1        = cell%bmat(1,:)/kpts%nkpt3(1)
   kv2        = cell%bmat(2,:)/kpts%nkpt3(2)
   kv3        = cell%bmat(3,:)/kpts%nkpt3(3)
   n          = 1
   found      = .true.

   DO WHILE(found)
      found = .false.
      DO ix = -n,n
         DO iy = -(n-abs(ix)),n-abs(ix)
            iz     = n - abs(ix) - abs(iy)
            DO sign=-1,1,2
               iz=sign*iz
               k(1) = ix*kv1(1) + iy*kv2(1) + iz*kv3(1)
               k(2) = ix*kv1(2) + iy*kv2(2) + iz*kv3(2)
               k(3) = ix*kv1(3) + iy*kv2(3) + iz*kv3(3)
               knorm2 = k(1)**2 + k(2)**2 + k(3)**2
               IF(knorm2.lt.rrad) THEN
                  found = .true.
                  divergence = divergence - exp(-expo*knorm2)/knorm2 / kpts%nkptf
               END IF
               IF(iz==0) exit
            END DO 
         END DO
      END DO
      n = n + 1
   END DO

END SUBROUTINE calc_divergence

END MODULE m_exchange_valence_hf