K Points

Sampling the Brillouin-zone:
Andreas EICHLER
Materialphysik and
Institut fur
Center for Computational Materials Science
Universitat Wien, Sensengasse 8, A-1090 Wien, Austria
b-initio
ackage
ienna
A. E ICHLER , S AMPLING
THE
B RILLOUIN - ZONE
imulation
Page 1
Overview
introduction
k-point meshes
Smearing methods
What to do in practice
THE
B RILLOUIN - ZONE
Page 2
Introduction
For many properties
(e.g.: density of states, charge density, matrix elements, response functions, . . . )
integrals (I) over the Brillouin-zone are necessary:
dk
F nk
1
BZ
BZ
To evaluate computationally
weighted sum over special k-points
integrals

1
BZ
BZ
ki
THE
B RILLOUIN - ZONE
Page 3
k-points meshes - The idea of special points

Chadi, Cohen, PRB 8 (1973) 5747.
function f k with complete lattice symmetry

introduce symmetrized plane-waves (SPW):
ekR

Am k
Cm
sum over symmetry-equivalent R

Cm
SPW
Cm
shell of lattice vectors
develope f k in Fourier-series (in SPW)

m 1
f m Am k
f0
f k
THE
B RILLOUIN - ZONE
Page 4
evaluate integral (=average) over Brillouin-zone

12

f0
BZ
Am k dk
BZ
f k dk
2 3
with:
2 3
taking n k-points with weighting factors k so that

n

i 1
ki
ki A m
f = weighted sum over k-points for variations of f that can be described within the
shell corresponding to CN .
THE
B RILLOUIN - ZONE
Page 5
Monkhorst and Pack (1976):

Idea: equally spaced mesh in Brillouin-zone.
Construction-rule:
ur b2
us b3
u p b1

qr

12

2r qr 1
2qr

ur
k prs
bi
reciprocal lattice-vectors
qr
determines number of
k-points in r-direction
THE
B RILLOUIN - ZONE
Page 6
r
b

Example:
quadratic 2-dimensional lattice

k
k
k r
b

16 k-points
q2
q1
1
4
1
4
1
2
IBZ
BZ
k3
THE
1
2F
k2
1
4F
k1
1
4F
BZ
F k dk
1
8
3
8
1
8
1
BZ
k3
k2
k1
1
8
3
8
3
8
IBZ)
only 3 inequivalent k-points (
B RILLOUIN - ZONE
Page 7
Interpretation:
representation of functionF k on a discrete equally-spaced mesh
E
N
n 0
an cos 2nk
more Fourier-components
higher accuracy
density of mesh
Common meshes :
Two choices for the center of the mesh

centered on (
belongs to mesh).
centered around . (can break symmetry !!)
THE
B RILLOUIN - ZONE
Page 8
Algorithm:
calculate equally spaced-mesh
shift the mesh if desired
apply all symmetry operations of Bravaislattice to all k-points

extract the irreducible k-points (
IBZ)
calculate the proper weighting
THE
B RILLOUIN - ZONE
Page 9
Smearing methods
Problem: in metallic systems Brillouin-zone integrals over functions that are

discontinuous at the Fermi-level.
high Fourier-components
dense grid is necessary.
Solution: replace step function by a smoother function.
Example:
bandstructure energy
nk
k nk
"
1 x
x
with:
E
e
nk
0 x
nk
nk
k nk f
necessary: appropriate function f
f equivalent to partial occupancies.
THE
B RILLOUIN - ZONE
Page 10
nk
#!
exp
nk
Fermi-Dirac function
consequence: energy is no longer variational with respect to the partial occupacies f .
f ln 1
f ln f
S fn

S f
kB T
F free energy.
new variational functional - defined by (1).
entropy
of a system of non-interacting electrons at a finite temperature T.
S f
smearing parameter.
can be interpreted as finite temperature via (3).

calculations at finite temperature are possible (Mermin 1965)

THE
B RILLOUIN - ZONE
Page 11
Consistency:

ln 1 f f

ln 1
ln f
'
&
Sfn
1
fn
ln 1 fnfn
nk
,
exp
(
fn

(
exp

)(

fn
(
kB T

E
fn
S
f
E
fn
)(
f ln 1
fn
f ln f
S f
S fn
THE
B RILLOUIN - ZONE
Page 12
Gaussian smearing
broadening of energy-levels with Gaussian function.

f becomes an integral of the Gaussian function:
nk
erf
1
1
2
&
'
nk
no analytical inversion of the error-function erf exists

entropy and free energy cannot be written in terms of f .
exp
has no physical interpretation.

variational functional F differs from E 0 .

forces are calculated as derivatives of the variational quantity (F ).

not necessarily equal to forces at E 0 .
THE
B RILLOUIN - ZONE
Page 13
E 0
E 0

E
B RILLOUIN - ZONE
F
1
2
2
THE

2
2

(

E 0
3
2
0.
Improvement: extrapolation to
Page 14
Method of Methfessel and Paxton (1989)
Idea:
expansion of stepfunction in a complete set of orthogonal functions
term of order 0 = integral over Gaussians
generalization of Gaussian broadening with
functions of higher order.
THE
B RILLOUIN - ZONE
Page 15
erf x
m 1
x2
xe
Am H2m
f0 x
fN x
1
2
f0 x
xe
An
x2
with:
1
2 AN H2N
1n
n!4n
SN x
HN : Hermite-polynomial of order N
advantages:
extrapolation for
deviation of F from E 0 only of order 2+N in

(
0 usually not necessary, but also possible:

1
E
1F
N 2
E
2
E 0
THE
B RILLOUIN - ZONE
Page 16
The significance of N and
MP of order N leads to a negligible error, if X is representable as a polynomial of

degree 2N around F .
linewidth can be increased for higher order to obtain the same accuracy
if S few meV
F
then E
n SN fn ) describes deviation of F from E .

entropy term (S
5
E 0.
2
3
2
2
3
forces correct within that limit.
in practice: smearings of order N=1 or 2 are sufficient
THE
B RILLOUIN - ZONE
Page 17
Linear tetrahedron method
Idea:
1. dividing up the Brillouin-zone into tetrahedra
2. Linear interpolation of the function to be integrated
Xn within these tetrahedra
3. integration of the interpolated function Xn
THE
B RILLOUIN - ZONE
Page 18
ad 1.
How to select mesh for tetrahedra
map out the IBZ
use special points
r
b
r
b
1
3 r
1 b
3
3
THE
B RILLOUIN - ZONE
1
4 r
1 b
Page 19
ad 2. interpolation
Seite 1 von 1
c j k Xn k j
Xn k
j .......... k-points
THE
B RILLOUIN - ZONE
Page 20
ad 3. k-space integration: simplification by Blo chl (1993)
remapping of the tetrahedra onto the k-points

dkc j k f n k
n j
1
BZ
BZ
effective weights n j for k-points.
k-space summation:
nj
THE
kj
n j Xn
B RILLOUIN - ZONE
Page 21
Drawbacks:
tetrahedra can break the symmetry of the Bravaislattice
at least 4 k-points are necessary
must be included
linear interpolation under- or overestimates
the real curve
THE
B RILLOUIN - ZONE
Page 22
Corrections by Blochl (1993)

Idea:
linear interpolation under- or overestimates the real curve
for full-bands or insulators these errors cancel
for metals: correction of quadratic errors is possible:
kn
j 1
F jn
kn
1
40 DT
DT
DOS for the tetrahedron T at F .
corners (k-point) of the tetrahedronT

THE
B RILLOUIN - ZONE
Page 23
Result:
best k-point convergence for energy
forces:
with Blochl corrections the new effective partial occupancies do not minimize the
groundstate total energy
variation of occupancies nk w.r.t. the ionic positions would be necessary
with US-PP and PAW practically impossible
THE
B RILLOUIN - ZONE
Page 24
Convergence tests
(from P.Blo chl, O. Jepsen, O.K. Andersen, PRB 49,16223 (1994).)
bandstructure energy of silicon:

conventional LT -method vs.
LT+Blochl corrections
bandstructure energy vs. k-point spacing :
THE
B RILLOUIN - ZONE
Page 25
What to do in practice
energy/DOS calculations:
linear tetrahedron method with Blochl corrections
ISMEAR=-5
calculation of forces:
semiconductors: Gaussian smearing (ISMEAR=0; SIGMA=0.1)
metals : Methfessel-Paxton (N=1 or 2)
always: test for energy with LT+Blochl-corr.
in any case:
careful checks for k-point convergence are necessary
THE
B RILLOUIN - ZONE
Page 26
The KPOINTS - file:

1>
2>
3>
4>
5>
k-points for a metal

0
Gamma point
9 9 9
0 0 0
1st line: comment

2nd line: 0 (
automatic generation)
3rd line: Monkhorst or Gammapoint (centered)

4th line: mesh parameter
5th line: 0 0 0 (shift)
THE
B RILLOUIN - ZONE
Page 27
mesh parameter
determine the number of intersections in each direction

longer axes in real-space
shorter axes in k-space
less intersections necessary for equally spaced mesh
Consequences:

molecules, atoms (large supercells)

1 1 1
is enough.
2-D Brillouin-zone)
surfaces (one long direction
x y 1 for the direction corresponding to the long direction.

9 /atom
4 /atom

B RILLOUIN - ZONE
THE
semiconductors:
typical values (never trust them!):
metals:
Page 28
Example - real-space/ reciprocal cell

r
b
r
b
;
G
44
E
D CBA
? >=<
@
J >IH
@
IBZ
BZ
k1
k2
r
b
F
N MLK
E
24
k1
k2
k3 r
b
IBZ
BZ
doubling the cell in real space halves the reciprocal cell

zone boundary is folded back to
same sampling is achieved with halved mesh parameter
THE
B RILLOUIN - ZONE
Page 29
Example - hexagonal cell

before
after
symmetrization
shifted to G
c Tba
V
Y TXW
V
U TSR
V
\ [Z
]
r
`_^
QPO
in certain cell geometries (e.g. hexagonal cells) even meshes break the symmetry
symmetrization results in non equally distributed k-points
Gamma point centered mesh preserves symmetry
THE
B RILLOUIN - ZONE
Page 30
Convergence tests

with respect to
and number of k-points in the IBZ
G.Kresse, J. Furthmuller, Computat. Mat. Sci. 6, 15 (1996).

THE
B RILLOUIN - ZONE
Page 31

K Points

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

K Points

Uploaded by

Copyright:

Available Formats

Sampling the Brillouin-zone:

k-points meshes - The idea of special points

function f k with complete lattice symmetry

sum over symmetry-equivalent R

shell of lattice vectors

develope f k in Fourier-series (in SPW)

evaluate integral (=average) over Brillouin-zone

taking n k-points with weighting factors k so that

Monkhorst and Pack (1976):

quadratic 2-dimensional lattice

only 3 inequivalent k-points (

centered around . (can break symmetry !!)

extract the irreducible k-points (

calculate the proper weighting

Problem: in metallic systems Brillouin-zone integrals over functions that are

necessary: appropriate function f

f equivalent to partial occupancies.

consequence: energy is no longer variational with respect to the partial occupacies f .

calculations at finite temperature are possible (Mermin 1965)

broadening of energy-levels with Gaussian function.

no analytical inversion of the error-function erf exists

has no physical interpretation.

variational functional F differs from E 0 .

forces are calculated as derivatives of the variational quantity (F ).

Method of Methfessel and Paxton (1989)

deviation of F from E 0 only of order 2+N in

0 usually not necessary, but also possible:

The significance of N and

MP of order N leads to a negligible error, if X is representable as a polynomial of

n SN fn ) describes deviation of F from E .

forces correct within that limit.

in practice: smearings of order N=1 or 2 are sufficient

Linear tetrahedron method

use special points

ad 3. k-space integration: simplification by Blo chl (1993)

remapping of the tetrahedra onto the k-points

effective weights n j for k-points.

Corrections by Blochl (1993)

DOS for the tetrahedron T at F .

corners (k-point) of the tetrahedronT

(from P.Blo chl, O. Jepsen, O.K. Andersen, PRB 49,16223 (1994).)

bandstructure energy of silicon:

The KPOINTS - file:

k-points for a metal

1st line: comment

3rd line: Monkhorst or Gammapoint (centered)

longer axes in real-space

shorter axes in k-space

less intersections necessary for equally spaced mesh

molecules, atoms (large supercells)

typical values (never trust them!):

Example - real-space/ reciprocal cell

doubling the cell in real space halves the reciprocal cell

Example - hexagonal cell

and number of k-points in the IBZ

G.Kresse, J. Furthmuller, Computat. Mat. Sci. 6, 15 (1996).

You might also like