[Pw_forum] CuO, LDA+U, Pseudopots, again

Paul M. Grant w2agz at pacbell.net
Thu Jan 10 06:22:20 CET 2008


To All:

Please refer to a previous pw_forum thread "CuO, LDA+U, Pseudopots,
etc."

I'm continuing to try to achieve a pwscf calculation which will yield a
"Hubbard gap" in any compound containing Cu++ and O--.  My efforts to
date have been unsucessful.  My "model" has been a pseudo-rocksalt CuO
structure based on "proxy" lattice constants near those of FeO, NiO and
MgO.

I strongly suspect my problem involves use (or misuse or unavailability)
of a proper Cu pseudopotential.  The issue of Cu++ bonded to O-- is
central to the physics of high temperature superconductivity, and a
failure of pwscf to properly resolve this difficulty would be
detrimential to the future deployment of quantum-espresso to this very
important and critical area of computational materials physics, not
just HTSC, but possibly GMR structures as well.  Note in what follows is
that I changed occ_loc in tabd.f90 from 10 to 9.

I  attach  below  1)  the  list  of  PPs  I've  tried  and their d-level
configurations.  
None  really contain the  generally assumed 3d9  Hund's
Rule configuration for CuO complexes.   Following this is the  output 2)
of a "hand written  grep-er" of the scf.out  file for a typical  example
(one of about 20) combinations of Cu PPs and the "starting_ns_eigenvalue
suggestions" to escape the "minimum energy trap."  I selected my "best
attempt" at this effort.  The critical (IMHO) lines are pointed to by a
"-->" marker.  BTW, I didn't try any Wannier projections since the
eigenvalue forcings didn't work, and my lesson from Example 25 is that
these have to be done successfully first.

Any of those in the pwscf community aware of a "better" PP for my
problem, please e-mail such to me.  In the meantime, I'm experimenting
"rolling my own" with the tools available at the Berlin-FHI site,
trying 3d9-4s2 as a first attempt--stay tuned.

Sorry for all my American slang.  We didn't invent the language...the
Brits did, and shoved it down our throats.  We just made English more user
friendly.

-Paul
----------------------------------
1)
Cu.blyp-n-van_ak.UPF
 nl pn  l   occ
 3D  3  2  9.50
 4S  4  0  1.00
 4P  4  1  0.50

Cu.pbe-d-rrkjus.UPF
 nl pn  l   occ
 3D  3  2 10.00
 3D  3  2  0.00
 4P  2  1 -1.00
 4S  1  0  1.00

Cu.pbe-n-van_ak.UPF
 nl pn  l   occ
 3D  3  2  9.50
 4S  4  0  1.00
 4P  4  1  0.50

Cu.pw91-n-van_ak.UPF
 nl pn  l   occ
 3D  3  2  9.50
 4S  4  0  1.00
 4P  4  1  0.50

Cu.pz-d-rrkjus.UPF
 nl pn  l   occ
 3D  3  2 10.00
 3D  3  2  0.00
 4P  2  1 -1.00
 4S  1  0  1.00

Cu.pz-n-van_ak.UPF
 nl pn  l   occ
 3D  3  2  9.50
 4S  4  0  1.00
 4P  4  1  0.50
---------------------------------------
2)
Today is Wed Jan  9 19:07:17 PST 2008
Now is Wed Jan  9 19:10:07 PST 2008
SCF Run preparatory to post-processing
Script is ./CuO_3.scf.sh, Output is CuO_3.scf.data

     kinetic-energy cutoff     =      30.0000  Ry
     charge density cutoff     =     240.0000  Ry
     convergence threshold     =      1.0E-06

-->    pwscf_exec          = /home/pmpgrant/O-sesame/PW/pw.x
    a_Ang               = 4.200
    a_Bohr              = 7.93950
    nbnd                = 20,
    nelec               = ,
    occupations         = 'smearing',
    smearing            = gauss,
    degauss             = 0.01,
    la2F                = .false.,
    diagonalization     = 'david',
    starting_magnetization(1) = 0.0
    starting_magnetization(2) = 0.5
    starting_magnetization(3) = -0.5
    nspin           = 2,
    lda_plus_u      = .true.,
-->    Hubbard_U(2)    = 5.0
-->    Hubbard_U(3)    = 5.0
-->    U_projection_type             = norm-atomic
-->    starting_ns_eigenvalue(1,2,2) =  0.0
    starting_ns_eigenvalue(2,2,2) =
    starting_ns_eigenvalue(3,2,2) =
    starting_ns_eigenvalue(4,2,2) =
    starting_ns_eigenvalue(5,2,2) =
-->    starting_ns_eigenvalue(1,1,3) =  0.0
    starting_ns_eigenvalue(2,1,3) =
    starting_ns_eigenvalue(3,1,3) =
    starting_ns_eigenvalue(4,1,3) =
    starting_ns_eigenvalue(5,1,3) =
-->    TM_pseudopot    = Cu.pz-d-rrkjus.UPF
-->    A_pseudopot     = O.LDA.US.RRKJ3.UPF

Run_Type        npk     Ef(eV)          Total E(Ry)       CPU_Time
Wall_Time
scf                 8    11.1457        -239.16751151     1m 8.04s
2m50.40s

d-orbital occupations
138:eigenvalues:  1.0000000 1.0000000 1.0000000 1.0000000 1.0000000
==>	152:eigenvalues:  0.8000000 0.8000000 0.8000000 0.8000000 0.8000000
==>	167:eigenvalues:  0.8000000 0.8000000 0.8000000 0.8000000 0.8000000
181:eigenvalues:  1.0000000 1.0000000 1.0000000 1.0000000 1.0000000
213:eigenvalues:  0.9976937 0.9976937 0.9984132 0.9984132 1.0008045
227:eigenvalues:  0.5164626 0.5164626 0.9680621 0.9823458 0.9823458
242:eigenvalues:  0.5164626 0.5164626 0.9680621 0.9823458 0.9823458
256:eigenvalues:  0.9976937 0.9976937 0.9984132 0.9984132 1.0008045
277:eigenvalues:  0.9976937 0.9976937 0.9984132 0.9984132 1.0008045
--> 291:eigenvalues:  0.0000000 0.5164626 0.9680621 0.9823458 0.9823458
--> 306:eigenvalues:  0.0000000 0.5164626 0.9680621 0.9823458 0.9823458
320:eigenvalues:  0.9976937 0.9976937 0.9984132 0.9984132 1.0008045
353:eigenvalues:  0.9936756 0.9936756 0.9968054 0.9968054 0.9991184
--> 367:eigenvalues:  0.7870114 0.7870114 0.9938818 0.9938818 0.9969944
--> 382:eigenvalues:  0.7867963 0.7867963 0.9938806 0.9938806 0.9969906
396:eigenvalues:  0.9936787 0.9936787 0.9968055 0.9968055 0.9991185
429:eigenvalues:  0.9914062 0.9914062 0.9971611 0.9971611 0.9994970
443:eigenvalues:  0.7968500 0.7968500 0.9944826 0.9944826 0.9975186
458:eigenvalues:  0.7971215 0.7971215 0.9944915 0.9944915 0.9975167
472:eigenvalues:  0.9914126 0.9914126 0.9971603 0.9971603 0.9994966
507:eigenvalues:  0.9667233 0.9667233 0.9969086 0.9969086 0.9992909
521:eigenvalues:  0.8723219 0.8723219 0.9959798 0.9959798 0.9984415
536:eigenvalues:  0.8713559 0.8713559 0.9959649 0.9959649 0.9984298
550:eigenvalues:  0.9663176 0.9663176 0.9968918 0.9968918 0.9992743
585:eigenvalues:  0.9525509 0.9525509 0.9963181 0.9963181 0.9986638
599:eigenvalues:  0.9127300 0.9127300 0.9961748 0.9961748 0.9985388
614:eigenvalues:  0.9146296 0.9146296 0.9962191 0.9962191 0.9985819
628:eigenvalues:  0.9537915 0.9537915 0.9963783 0.9963783 0.9987221
663:eigenvalues:  0.9458265 0.9458265 0.9962868 0.9962868 0.9986542
677:eigenvalues:  0.9104092 0.9104092 0.9959371 0.9959371 0.9983246
692:eigenvalues:  0.9047312 0.9047312 0.9957990 0.9957990 0.9982026
706:eigenvalues:  0.9417229 0.9417229 0.9961282 0.9961282 0.9984984
741:eigenvalues:  0.9354950 0.9354950 0.9960520 0.9960520 0.9984263
755:eigenvalues:  0.9174694 0.9174694 0.9958534 0.9958534 0.9982404
770:eigenvalues:  0.9179431 0.9179431 0.9958601 0.9958601 0.9982469
784:eigenvalues:  0.9358828 0.9358828 0.9960625 0.9960625 0.9984347
819:eigenvalues:  0.9341625 0.9341625 0.9960316 0.9960316 0.9984028
833:eigenvalues:  0.9222122 0.9222122 0.9959062 0.9959062 0.9982848
848:eigenvalues:  0.9216331 0.9216331 0.9958882 0.9958882 0.9982692
862:eigenvalues:  0.9336231 0.9336231 0.9960127 0.9960127 0.9983843
897:eigenvalues:  0.9327282 0.9327282 0.9960266 0.9960266 0.9983951
911:eigenvalues:  0.9284646 0.9284646 0.9959858 0.9959858 0.9983585
926:eigenvalues:  0.9283145 0.9283145 0.9959811 0.9959811 0.9983538
940:eigenvalues:  0.9325413 0.9325413 0.9960218 0.9960218 0.9983904
975:eigenvalues:  0.9318453 0.9318453 0.9960271 0.9960271 0.9983966
989:eigenvalues:  0.9297349 0.9297349 0.9960062 0.9960062 0.9983772
1004:eigenvalues:  0.9299729 0.9299729 0.9960136 0.9960136 0.9983841
1018:eigenvalues:  0.9320457 0.9320457 0.9960350 0.9960350 0.9984044
1053:eigenvalues:  0.9311035 0.9311035 0.9960273 0.9960273 0.9983963
1067:eigenvalues:  0.9316853 0.9316853 0.9960330 0.9960330 0.9984016
1082:eigenvalues:  0.9316968 0.9316968 0.9960336 0.9960336 0.9984023
1096:eigenvalues:  0.9310994 0.9310994 0.9960275 0.9960275 0.9983966
--> 1131:eigenvalues:  0.9314650 0.9314650 0.9960338 0.9960338 0.9984027
--> 1145:eigenvalues:  0.9312033 0.9312033 0.9960312 0.9960312 0.9984004
--> 1160:eigenvalues:  0.9311989 0.9311989 0.9960309 0.9960309 0.9984001
--> 1174:eigenvalues:  0.9314517 0.9314517 0.9960337 0.9960337 0.9984026

Paul M. Grant, PhD
Principal, W2AGZ Technologies
Visiting Scholar, Applied Physics, Stanford University
EPRI Science Fellow (Retired)
IBM Research Staff Member Emeritus
w2agz at pacbell.net
http://www.w2agz.com
 
 


-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.democritos.it/pipermail/pw_forum/attachments/20080109/1ca8de8a/attachment-0001.htm 


More information about the Pw_forum mailing list