abinit/doc/developers/psp1_data.txt

354 lines
15 KiB
Plaintext

The present file contains informations about the complete set of
pseudopotentials generated by D.C. Allan and A. Khein,
available on the ABINIT Web site.
These pseudopotentials are single projector, ordinary norm conserving,
based on the Troullier-Martins method.
All these pseudopotentials have been tested against ghost states.
Their cut-off radii follows reasonable trends across the
periodic table. Some of them have been the subject of systematic
testing. Many others have been used since the release of this
table, during 1995, as this set has been provided
with the plane_wave code commercialized by Biosym (BiosymII),
before being used in connection with ABINIT.
This file present a list of results obtained with these pseudopotentials,
and provide some comparison with LAPW and experimental results.
Comparison with LAPW is meaningful, as the LDA is common to both,
and the convergence parameters have been chosen identical when possible.
WARNING : The comparison with experimental results is NOT indicative
of the accuracy of the pseudopotential for numerical work,
since it will depend first on the accuracy of the LDA, and second,
on the numerical parameters of the tests, in particular the number
of k points: in most cases this was NOT lead to full convergence.
Experimental data (often without ref !) are provided to discuss
potential problems with the pseudopotentials or numerical convergence.
In these tests, one will compare lattice constant - acell (bohr) - and
bulk modulus - b0 (hartree/bohr^3, with 1 hartree/bohr^3 = 29421.033 GPa).
The cut-off energy - ecut (Hartree) is an important parameter for planewave basis
convergence, and the behaviour of acell and b0 is provided as a function of ecut.
Thus, the practical information that one can gain from the present data are:
- an estimation of the cut-off energy to be used to start convergence studies
for other materials, with these pseudopotentials
- some warnings about the importance of semi-core states : some pseudopotentials
are inappropriate for use in a strongly electronegative environment,
for example in bonds with Oxygen atoms (!)
There are 20 crystals represented below (in alphabetical order) :
BaTiO3, C, CeO2, CuBr, GaAs, Ge, InSb, KCl, KI, KNbO3, LiF, MgS,
NaCl, PbZrO3, RbI, Si, TlCl, Yb, ZnS, and ZnSe.
Thus, this file provides information on the pseudopotentials for the
following 27 elements : As, Ba, Br, C, Ce, Cl, Cu, F, Ga, Ge, I, In, K,
Li, Mg, Na, Nb, O, Pb, Rb, S, Sb, Se, Si, Ti, Tl, Zn.
For each crystal, one finds first the result (acell, b0) using the present
pseudopotential table (0), then experimental data (1), then
lapw (2), then eventual other relevant data.
Then some additional information : the numerical parameters of the
present calculation, or bibliographical infos.
Then a convergence study, then some comments.
****************************************
BaTiO3
acell b0
(0) 7.90583 .02229689 Present, 10 sp, ecut=35
(1) 7.5778 Exp.
(2) 7.45 lapw (from King-Smith+Vanderbilt PRB 49, 5828 (1994))
Convergence of (0)
ecut acell Etot d2edx2 b0 b0'
35 7.90583 -71.75573246 .39661955 .02229689 4.61792795
NOTE : the bad lattice parameter wrt experiment can be attributed to the lack
of semi-core 3s and 3p state for the Titanium pseudopotential. Semicore
states are important in this case, because the strongly electronegative
oxygen pump the electrons of Ti ...
****************************************
C
acell b0
(0) 6.69171 .015498 Present, 2sp, ecut=50
(1) 6.741 .01502 Exp
Convergence of (0)
ecut acell Etot d2edx2 b0 b0'
10 6.76518 -11.83805645 .25577589 .01680343 7.65504363
15 6.72330 -11.99928675 .18756893 .01239927 4.76660316
20 6.70916 -12.06037543 .23429604 .01552082 3.64182667
25 6.69190 -12.07143817 .23520357 .01562110 3.64991544
30 6.69246 -12.07299436 .23333254 .01549555 3.64052864
35 6.69251 -12.07331477 .23335813 .01549713 3.63707511
40 6.69194 -12.07359373 .23347528 .01550623 3.64308353
45 6.69174 -12.07379469 .23333287 .01549724 3.64019181
50 6.69171 -12.07392088 .23334942 .01549841 3.63946513
****************************************
CeO2
acell b0
(0) 12.02537 .00250819 Present, 2 sp, ecut=35
(1) 10.225 Exp.
Convergence of (0)
ecut acell Etot d2edx2 b0 b0'
20 11.37152 -36.29574319 .18407408 .00719435 2.97906597
25 12.18576 -36.50506602 .05591076 .00203920 4.21893854
30 12.01906 -36.53740553 .06858205 .00253605 4.09821618
35 12.02537 -36.54430893 .06786434 .00250819 4.11053753
NOTE : the bad lattice parameter wrt to experiment could come from
a too soft pseudopotential, or a lack of k points, or even from the LDA !
****************************************
CuBr
acell b0
(0) 10.44645 .00211778 Present, 10 sp, ecut=40
(1) 10.75 Exp.
Convergence of (0)
ecut acell Etot d2edx2 b0 b0'
20 10.31035 -67.77801084 .05189193 .00223689 4.68626483
25 10.40127 -68.96085814 .05154285 .00220242 5.41383681
30 10.43451 -69.23610234 .05035934 .00214499 5.40905077
35 10.44565 -69.28622321 .04969485 .00211443 5.32420911
40 10.44645 -69.29245307 .04977741 .00211778 5.32823650
NOTE : the 3d electrons of Cu demand a high ecut.
****************************************
GaAs
acell b0
(0) 10.32069 .00281133 Present, 2 sp, ecut=10
(1a)10.662 .00266 Exp., see Nielsen & Martin, Phys. Rev. B 32, 3792 (1985).
(1b)10.683 .00268 Exp., see Ihm & Joannopoulos, Phys. Rev. B 24, 4191 (1981).
(2) 10.63525 .00242491 lapw, 2 sp (Alex Khein)
(3) 10.31637 .00280981 Other TM pseudopotential, 2 sp, ecut=30
Convergence of (3)
ecut acell Etot d2edx2 b0 b0'
05 10.32849 -10.72338064 .06144027 .00264383 4.16068683
10 10.32068 -10.73655972 .06528396 .00281135 4.59015040
20 10.31663 -10.73997682 .06523807 .00281048 4.62708169
30 10.31637 -10.74010831 .06522073 .00280981 4.62679512
Convergence of (0) is same as that of (3).
ecut acell Etot d2edx2 b0 b0'
10 10.32069 -10.73683231 .06528346 .00281133 4.59013355
****************************************
Ge
acell b0
(0) 10.45124 .00264655 Present, 10 sp, ecut=40.
(1a)10.681 .00262 Exp., see Yin and Cohen, Phys. Rev. B 26, 5668 (1982).
(1b)10.677 .00261 Exp., see Nielsen & Martin, Phys. Rev. B 32, 3792 (1985).
Convergence of (0)
ecut acell Etot d2edx2 b0 b0'
05 10.42147 -10.06921548 .06120818 .00261034 4.95801315
10 10.44798 -10.10946927 .06331623 .00269339 4.63888046
15 10.44766 -10.11513681 .06262526 .00266409 4.70669933
20 10.45100 -10.11694086 .06223257 .00264653 4.67345001
30 10.45121 -10.11724831 .06224021 .00264681 4.67049786
40 10.45124 -10.11732594 .06223447 .00264655 4.67182070
****************************************
InSb
acell b0
(0) 11.57070 .00190571 Present, 10 sp, ecut=25
(1) 12.243 Exp.
Convergence of (0)
ecut acell Etot d2edx2 b0 b0'
5 11.61529 -10.03348129 .04569867 .00174860 4.58884406
10 11.57566 -10.05720398 .05080794 .00195076 5.07212776
15 11.57075 -10.06415054 .04960653 .00190544 4.98618778
20 11.57081 -10.06459277 .04957256 .00190412 4.97977831
25 11.57070 -10.06490900 .04961350 .00190571 4.98595553
NOTE : semicore 4d states might be needed for In.
****************************************
KCl
acell b0
(0) 12.07417 .00068937 Present, 2 sp, ecut=30.
(1a)11.89 .000670 Exp.
(1b)11.79 Exp.
(2) 11.56660 .00082554 lapw, 2 sp (Alex Khein)
Convergence of (0)
ecut acell Etot d2edx2 b0 b0'
30 12.07417 -33.84009477 .01872814 .00068937 5.07275107
****************************************
KI
acell b0
(0) 13.41731 .00046229 Present, 2 sp, ecut=30.
(1) 13.36 .000398 Exp.
(2) 12.91145 .00054908 lapw, 2 sp (Alex Khein)
Convergence of (0)
ecut acell Etot d2edx2 b0 b0'
10 13.53076 -31.57524273 .01424581 .00046793 4.30704897
15 13.43027 -31.61368086 .01385196 .00045840 4.92372641
20 13.41736 -31.61665632 .01395020 .00046209 5.01409895
25 13.41748 -31.61721674 .01396052 .00046243 5.00743101
30 13.41731 -31.61767432 .01395610 .00046229 5.00902509
****************************************
KNbO3
acell b0
(0) 8.03339 .02175020 Present, 10 sp, ecut=30
(1) 7.58 Exp. (from King-Smith+Vanderbilt PRB 49, 5828 (1994))
(2) 7.488 lapw (from King-Smith+Vanderbilt PRB 49, 5828 (1994))
Convergence of (0)
ecut acell Etot d2edx2 b0 b0'
20 8.04244 -73.18829985 .41473429 .02291921 4.05925771
25 8.04202 -73.42412542 .38551325 .02130550 4.27473026
30 8.03339 -73.48444944 .39313748 .02175020 4.24896923
NOTE : semicore states might be needed for the Nb pseudopotential in the
presence of electronegative oxygen.
****************************************
LiF
acell b0
(0) 7.70575 .00276340 Present, 10 sp, ecut=40
(1) 7.597 Exp.
Convergence of (0)
ecut acell Etot d2edx2 b0 b0'
15 7.73441 -24.33099404 .05014266 .00288136 6.16938411
20 7.80452 -24.87218438 .06540478 .00372461 2.35027217
25 7.74858 -25.10005721 .04259643 .00244325 3.93631029
30 7.71322 -25.18066480 .04903335 .00282536 4.07016856
35 7.70682 -25.20271478 .04790631 .00276271 4.16814366
40 7.70575 -25.20691894 .04791162 .00276340 4.21731835
NOTE : the 2p electrons of Fluorine demand a high cut-off energy.
****************************************
MgS
acell b0
(0) 9.73912 .00283529 Present, 10 sp, ecut=20
(1) 9.8266 Exp.
Convergence of (0)
ecut acell Etot d2edx2 b0 b0'
10 9.74238 -12.38271718 .06193946 .00282566 4.14005348
15 9.73865 -12.38619663 .06219016 .00283818 4.11473786
20 9.73912 -12.38815996 .06212977 .00283529 4.09445283
****************************************
NaCl
acell b0
(0) 10.47177 .00117877 Present, 2 sp, ecut=20
(1a)10.582 .000904 Exp.
(1b)10.658 Exp.
(2) 10.44243 .00105384 lapw, 2 sp (Alex Khein)
Convergence of (0)
ecut acell Etot d2edx2 b0 b0'
10 10.47416 -22.98001420 .02765559 .00117349 5.27012543
20 10.47177 -22.99094114 .02777356 .00117877 5.19754969
****************************************
PbZrO3
acell b0
(0) 8.34436 .01874107 Present, 10 sp, ecut=30
(1) 7.7668 Exp.
NOTE : semicore states are definitely needed for Pb as well as Zr, see
the comment on BaTiO3.
****************************************
RbI
acell b0
(0) 13.95057 .00042382 Present, 10 sp, ecut=20
(1) 13.871 Exp.
Convergence of (0)
ecut acell Etot d2edx2 b0 b0'
10 13.95143 -30.03817339 .01330642 .00042390 5.16176399
15 13.95080 -30.04863071 .01331246 .00042411 5.14966285
20 13.95057 -30.05128579 .01330325 .00042382 5.14618128
****************************************
Si
acell b0
(0) 10.21572 .00320012 Present, 2sp, ecut=20.
(1a)10.259 .00336 Exp., see Yin and Cohen, Phys. Rev. B 26, 5668 (1982).
(1b)10.263 .00337 Exp., see Nielsen & Martin, Phys. Rev. B 32, 3792 (1985).
Convergence of (0)
ecut acell Etot d2edx2 b0 b0'
5 10.28812 -8.84901608 .07113187 .00307288 3.96857170
10 10.21457 -8.86725003 .07373241 .00320816 4.22989617
15 10.21547 -8.86984610 .07356532 .00320060 4.20822798
20 10.21572 -8.87029473 .07355601 .00320012 4.20750185
****************************************
TlCl
acell b0
(0) 7.71721 .00385530 Present, 10 sp, ecut=30
(1) 7.2314 Exp.
Convergence of (0)
ecut acell Etot d2edx2 b0 b0'
10 7.72014 -18.64741166 .06704874 .00385996 4.29268140
15 7.71898 -18.65955033 .06700461 .00385800 4.27495739
20 7.71787 -18.66217719 .06709814 .00386394 4.26516400
25 7.71734 -18.66458783 .06693242 .00385466 4.27668788
30 7.71721 -18.66531958 .06694247 .00385530 4.27578066
NOTE : the error in the cell parameter is to be attributed to the
lack of semicore states.
****************************************
Yb
acell b0
(0) 9.95275 .00052866 Present, 2 sp, ecut=40
(1) 10.358 Exp.
Convergence of (0)
ecut acell Etot d2edx2 b0 b0'
35 9.95726 -84.41873943 .01168913 .00052175 1.85748564
40 9.95275 -84.42082730 .01183857 .00052866 2.01091840
NOTE : the error on the lattice parameter might be due to the
very inaccurate k-point sampling, since Yb is a metal...
****************************************
ZnS
acell b0
(0) 10.07018 .00286110 Present, 2 sp, ecut=50.
(1) 10.2123 .00261 Exp. (ref. in Martins et al PRB 43, 2213 (1991)).
(2a)10.1157 .00296 lapw (Martins et al PRB 43, 2213 (1991)).
(2b)10.02671 .00293672 lapw, 10 sp, Alex Khein
Convergence of (0)
ecut acell Etot d2edx2 b0 b0'
30 10.07122 -72.61951511 .06478546 .00285899 4.39154224
40 10.07014 -72.62231669 .06485548 .00286239 4.42473262
50 10.07018 -72.62410724 .06482658 .00286110 4.42808050
NOTE : the 3d electrons of Zn require a high cut-off energy
****************************************
ZnSe
acell b0
(0) 10.55093 .00236365 Present, 2 sp, ecut=60
(1a)10.677 Exp. (Kittel)
(1b)10.709 Exp. (Pankove)
(1c)10.711 Exp. (Wang and Klein after correction)
Convergence of (0) (e99.0=19.7, e99.9=25.1 (Zn(3d)))
30 10.55014 -72.62842490 .05611463 .00236393 4.56988631
40 10.55091 -72.63123392 .05608830 .00236265 4.58916452
50 10.55092 -72.63302367 .05612227 .00236408 4.59065156
60 10.55093 -72.63346840 .05611220 .00236365 4.58993535
NOTE: the 3d electrons of Zn require a high cut-off energy