Quantum-Espresso_10_8_14
cjfoss
6,790 views
32 slides
Mar 24, 2016
Slide 1 of 32
1
2
3
4
5
6
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
About This Presentation
No description available for this slideshow.
Slide Content
INTRO TO QUANTUM ESPRESSO AND
RESEARCH UPDATE
BY
CAMERON FOSS
NANOELECTRONIC THEORY AND
SIMULATION LAB
First-Principle Calculations
with Quantum Espresso
[email protected]
RESEARCH UPDATE
BY
CAMERON FOSS
NANOELECTRONIC THEORY AND
SIMULATION LAB
First-Principle Calculations
with Quantum Espresso
[email protected]
What is QE?
First-Principle (“ab initio”)
Suite of C and Fortran codes based on DFT, PW,
Pseudopotentials.
Approximations:
Theory: Born Oppenheimer, DFT, Pseudopotentials.
Computational: FFT, cell size (nat), finite energy cutoff for
PW(ecutwfc), discrete BZ sampling (k-points) ...convergence!
Results:
Ground-state energy, atomic forces, stresses, MD, etc.
Structural optimizations (“minimum energy structures”), Potential
Energy Superfaces, Vibrational Spectroscopy
Alloys, supercells, metals, insulators, semi’s
P. Giannozzi, Quantum Simulations of materials using Quantum Espresso. Univ. of Udine (01/26/2012)
First-Principle (“ab initio”)
Suite of C and Fortran codes based on DFT, PW,
Pseudopotentials.
Approximations:
Theory: Born Oppenheimer, DFT, Pseudopotentials.
Computational: FFT, cell size (nat), finite energy cutoff for
PW(ecutwfc), discrete BZ sampling (k-points) ...convergence!
Results:
Ground-state energy, atomic forces, stresses, MD, etc.
Structural optimizations (“minimum energy structures”), Potential
Energy Superfaces, Vibrational Spectroscopy
Alloys, supercells, metals, insulators, semi’s
P. Giannozzi, Quantum Simulations of materials using Quantum Espresso. Univ. of Udine (01/26/2012)
Using Quantum Espresso
Linux
Terminal based(but there is an interface for pw.x, PWGUI)
Input files
Defining the structure/material you want
Unit cells and alloys
Pseudopotentials
Convergence
Some results
Batch scripting
Using clusters
Linux
Terminal based(but there is an interface for pw.x, PWGUI)
Input files
Defining the structure/material you want
Unit cells and alloys
Pseudopotentials
Convergence
Some results
Batch scripting
Using clusters
Input Files
Main packages of QE are based off of the PWscf and
PHonon packages:
pw.x (total energy), ph.x(phonons), q2r.x(IFC), matdyn.x(BZ),
plotband.x
These are executables that have the same general
execution format:
~/bin/”cmd”.x <“input file name”.in> “output file name”.out
And these input files are based off several necessary
(and optional) “cards” and “namelists”
Main packages of QE are based off of the PWscf and
PHonon packages:
pw.x (total energy), ph.x(phonons), q2r.x(IFC), matdyn.x(BZ),
plotband.x
These are executables that have the same general
execution format:
~/bin/”cmd”.x <“input file name”.in> “output file name”.out
And these input files are based off several necessary
(and optional) “cards” and “namelists”
Execution Pipeline
q2r.x
matdyn.x
Self-consistent Total Energy
Calculation
Phonon frequencies and
displacement patterns via Density
Functional Perturbation Theory
Interatomic Force Constants
in real space
Produces phonon frequencies
for dispersion curves.
band structure, structure
relaxations, total energy
Single points and Monkhorst-
Pack grid of q-points
Phonon frequencies over
desired paths.
Most time is spent editing pw.x input files for different materials and strucutres, and matdyn
input files for different paths and strained materials.
Plotband MATLAB
pw.x
ph.x
q2r.x
matdyn.x
Self-consistent Total Energy
Calculation
Phonon frequencies and
displacement patterns via Density
Functional Perturbation Theory
Interatomic Force Constants
in real space
Produces phonon frequencies
for dispersion curves.
band structure, structure
relaxations, total energy
Single points and Monkhorst-
Pack grid of q-points
Phonon frequencies over
desired paths.
Most time is spent editing pw.x input files for different materials and strucutres, and matdyn
input files for different paths and strained materials.
Plotband MATLAB
pw.x
ph.x
Namelists and Cards
&control
simulation specific namelist, contains calculation type and paths for
pseudopotential files and output files.
/
&system
structure specific namelist, can specify desired Bravais lattice, lattice constants, number of
atoms in unit cell, number of species in unit cell, cutoff energies.
/
&electrons
parameters for convergence and self-consistency.
/
ATOMIC_SPECIES
define atomic species
ATOMIC_POSITIONS
define positioning of atoms in unit cell. (BEWARE!)
K_POINTS
define a grid of k-points with which the code uses to define other matrices within the codes.
Optional cards include: &ions, &cell, CELL_PARAMETERS, CONSTRAINTS, etc.
&control
simulation specific namelist, contains calculation type and paths for
pseudopotential files and output files.
/
&system
structure specific namelist, can specify desired Bravais lattice, lattice constants, number of
atoms in unit cell, number of species in unit cell, cutoff energies.
/
&electrons
parameters for convergence and self-consistency.
/
ATOMIC_SPECIES
define atomic species
ATOMIC_POSITIONS
define positioning of atoms in unit cell. (BEWARE!)
K_POINTS
define a grid of k-points with which the code uses to define other matrices within the codes.
Optional cards include: &ions, &cell, CELL_PARAMETERS, CONSTRAINTS, etc.
Example pw.x input file
&control
calculation=‘scf’ !(‘nscf’, ‘bands’, ‘vc-relax’, ’relax’, ‘md’)
restart_mode=‘from_scratch’ !(‘restart’)
prefix=‘Si’
pseudo_dir=‘/home/espresso-5.1/pseudo/’
outdir=‘/home/espresso-5.1/out/’
!outdir=‘/*output folder for large runs*/’
/
&system
ibrav=2, celldm(1)=10.263, nat=2, ntyp=1, ecutwfc=16.0
/
&electrons
mixing_beta=0.7, conv_thr= 1.0d-9
/
ATOMIC_SPECIES
Si 28.0855 Si.pz-vbc.UPF
ATOMIC_POSITIONS {alat} !(bohr | angstrom | crystal)
Si 0.00 0.00 0.00
Si 0.25 0.25 0.25
K_POINTS {automatic}
4 4 4 1 1 1
Things worth noting:
•celldm in units of bohr
•ibrav
•pseudopotentials
•k-points
&control
calculation=‘scf’ !(‘nscf’, ‘bands’, ‘vc-relax’, ’relax’, ‘md’)
restart_mode=‘from_scratch’ !(‘restart’)
prefix=‘Si’
pseudo_dir=‘/home/espresso-5.1/pseudo/’
outdir=‘/home/espresso-5.1/out/’
!outdir=‘/*output folder for large runs*/’
/
&system
ibrav=2, celldm(1)=10.263, nat=2, ntyp=1, ecutwfc=16.0
/
&electrons
mixing_beta=0.7, conv_thr= 1.0d-9
/
ATOMIC_SPECIES
Si 28.0855 Si.pz-vbc.UPF
ATOMIC_POSITIONS {alat} !(bohr | angstrom | crystal)
Si 0.00 0.00 0.00
Si 0.25 0.25 0.25
K_POINTS {automatic}
4 4 4 1 1 1
Things worth noting:
•celldm in units of bohr
•ibrav
•pseudopotentials
•k-points
Example pw.x input file
&control
calculation=‘scf’ !(‘nscf’, ‘bands’, ‘vc-relax’, ’relax’, ‘md’)
restart_mode=‘from_scratch’ !(‘restart’)
prefix=‘Si’
pseudo_dir=‘/home/espresso-5.1/pseudo/’
outdir=‘/home/espresso-5.1/out/’
!outdir=‘/*output folder for large runs*/’
/
&system
ibrav=2, celldm(1)=10.263, nat=2, ntyp=1, ecutwfc=16.0
/
&electrons
mixing_beta=0.7, conv_thr= 1.0d-9
/
ATOMIC_SPECIES
Si 28.0855 Si.pz-vbc.UPF
ATOMIC_POSITIONS {alat} !(bohr | angstrom | crystal)
Si 0.00 0.00 0.00
Si 0.25 0.25 0.25
K_POINTS {automatic}
4 4 4 1 1 1
Other possible options necessary
for many other types of
calculations. For instance,
specifying calculation= ‘bands’
will perform an energy band
calculation and return data for
plotting the band structure of the
material.
You can also change how the
values in ATOMIC_POSITIONS
are represented. This will become
clearer later on.
&control
calculation=‘scf’ !(‘nscf’, ‘bands’, ‘vc-relax’, ’relax’, ‘md’)
restart_mode=‘from_scratch’ !(‘restart’)
prefix=‘Si’
pseudo_dir=‘/home/espresso-5.1/pseudo/’
outdir=‘/home/espresso-5.1/out/’
!outdir=‘/*output folder for large runs*/’
/
&system
ibrav=2, celldm(1)=10.263, nat=2, ntyp=1, ecutwfc=16.0
/
&electrons
mixing_beta=0.7, conv_thr= 1.0d-9
/
ATOMIC_SPECIES
Si 28.0855 Si.pz-vbc.UPF
ATOMIC_POSITIONS {alat} !(bohr | angstrom | crystal)
Si 0.00 0.00 0.00
Si 0.25 0.25 0.25
K_POINTS {automatic}
4 4 4 1 1 1
Other possible options necessary
for many other types of
calculations. For instance,
specifying calculation= ‘bands’
will perform an energy band
calculation and return data for
plotting the band structure of the
material.
You can also change how the
values in ATOMIC_POSITIONS
are represented. This will become
clearer later on.
Example pw.x input file
&control
calculation=‘scf’
restart_mode=‘from_scratch’
prefix=‘Si’
pseudo_dir=‘/home/espresso-5.1/pseudo/’
outdir=‘/home/espresso-5.1/out/’
!outdir=‘/*output folder for large runs*/’
/
&system
ibrav=2, celldm(1)=10.263, nat=2, ntyp=1, ecutwfc=16.0
/
&electrons
mixing_beta=0.7, conv_thr= 1.0d-9
/
ATOMIC_SPECIES
Si 28.0855 Si.pz-vbc.UPF
ATOMIC_POSITIONS {alat}
Si 0.00 0.00 0.00
Si 0.25 0.25 0.25
K_POINTS {automatic}
4 4 4 1 1 1
Calculation:
Generally remains the same when
concerning phonons. ‘relax’ and ‘vc-
relax’ are helpful in applying strains and
in performing structure relaxations.
Restart_Mode:
Really only need to set ‘restart’ if
simulation is very long (days/weeks)
Prefix:
Randomly chosen by user.
Pseudo_dir:
path to pseudopotential folder
Outdir:
output folder path
&control
calculation=‘scf’
restart_mode=‘from_scratch’
prefix=‘Si’
pseudo_dir=‘/home/espresso-5.1/pseudo/’
outdir=‘/home/espresso-5.1/out/’
!outdir=‘/*output folder for large runs*/’
/
&system
ibrav=2, celldm(1)=10.263, nat=2, ntyp=1, ecutwfc=16.0
/
&electrons
mixing_beta=0.7, conv_thr= 1.0d-9
/
ATOMIC_SPECIES
Si 28.0855 Si.pz-vbc.UPF
ATOMIC_POSITIONS {alat}
Si 0.00 0.00 0.00
Si 0.25 0.25 0.25
K_POINTS {automatic}
4 4 4 1 1 1
Calculation:
Generally remains the same when
concerning phonons. ‘relax’ and ‘vc-
relax’ are helpful in applying strains and
in performing structure relaxations.
Restart_Mode:
Really only need to set ‘restart’ if
simulation is very long (days/weeks)
Prefix:
Randomly chosen by user.
Pseudo_dir:
path to pseudopotential folder
Outdir:
output folder path
Example pw.x input file
&control
calculation=‘scf’
restart_mode=‘from_scratch’
prefix=‘Si’
pseudo_dir=‘/home/espresso-5.1/pseudo/’
outdir=‘/home/espresso-5.1/out/’
!outdir=‘/*output folder for large runs*/’
/
&system
ibrav=2, celldm(1)=10.263, nat=2, ntyp=1, ecutwfc=16.0
/
&electrons
mixing_beta=0.7, conv_thr= 1.0d-9
/
ATOMIC_SPECIES
Si 28.0855 Si.pz-vbc.UPF
ATOMIC_POSITIONS {alat}
Si 0.00 0.00 0.00
Si 0.25 0.25 0.25
K_POINTS {automatic}
4 4 4 1 1 1
Ibrav=2 specifies fcc unit cell
Celldm(1)=10.263 bohr=5.431 angs
(note: cubic so celldm(1)= celldm(2)=
celldm(3))
Nat=number of atoms in Primitive unit cell
ntyp= # of species in unit cell
ecutwfc=energy cutoff (Ry)for plane
waves (note: ecutwfc is one half to
convergence testing, increasing will also
increase simulation time)
(there is also ecutrho which is the cutoff
energy in Ry for charge density.)
&control
calculation=‘scf’
restart_mode=‘from_scratch’
prefix=‘Si’
pseudo_dir=‘/home/espresso-5.1/pseudo/’
outdir=‘/home/espresso-5.1/out/’
!outdir=‘/*output folder for large runs*/’
/
&system
ibrav=2, celldm(1)=10.263, nat=2, ntyp=1, ecutwfc=16.0
/
&electrons
mixing_beta=0.7, conv_thr= 1.0d-9
/
ATOMIC_SPECIES
Si 28.0855 Si.pz-vbc.UPF
ATOMIC_POSITIONS {alat}
Si 0.00 0.00 0.00
Si 0.25 0.25 0.25
K_POINTS {automatic}
4 4 4 1 1 1
Ibrav=2 specifies fcc unit cell
Celldm(1)=10.263 bohr=5.431 angs
(note: cubic so celldm(1)= celldm(2)=
celldm(3))
Nat=number of atoms in Primitive unit cell
ntyp= # of species in unit cell
ecutwfc=energy cutoff (Ry)for plane
waves (note: ecutwfc is one half to
convergence testing, increasing will also
increase simulation time)
(there is also ecutrho which is the cutoff
energy in Ry for charge density.)
Example pw.x input file
&control
calculation=‘scf’
restart_mode=‘from_scratch’
prefix=‘Si’
pseudo_dir=‘/home/espresso-5.1/pseudo/’
outdir=‘/home/espresso-5.1/out/’
!outdir=‘/*output folder for large runs*/’
/
&system
ibrav=2, celldm(1)=10.263, nat=2, ntyp=1, ecutwfc=16.0
/
&electrons
mixing_beta=0.7, conv_thr= 1.0d-9
/
ATOMIC_SPECIES
Si 28.0855 Si.pz-vbc.UPF
ATOMIC_POSITIONS {alat}
Si 0.00 0.00 0.00
Si 0.25 0.25 0.25
K_POINTS {automatic}
4 4 4 1 1 1
Mixing_beta= variable for self-
consistency
Conv_thr=convergence threshold for self-
consistency.
&control
calculation=‘scf’
restart_mode=‘from_scratch’
prefix=‘Si’
pseudo_dir=‘/home/espresso-5.1/pseudo/’
outdir=‘/home/espresso-5.1/out/’
!outdir=‘/*output folder for large runs*/’
/
&system
ibrav=2, celldm(1)=10.263, nat=2, ntyp=1, ecutwfc=16.0
/
&electrons
mixing_beta=0.7, conv_thr= 1.0d-9
/
ATOMIC_SPECIES
Si 28.0855 Si.pz-vbc.UPF
ATOMIC_POSITIONS {alat}
Si 0.00 0.00 0.00
Si 0.25 0.25 0.25
K_POINTS {automatic}
4 4 4 1 1 1
Mixing_beta= variable for self-
consistency
Conv_thr=convergence threshold for self-
consistency.
Example pw.x input file
&control
calculation=‘scf’
restart_mode=‘from_scratch’
prefix=‘Si’
pseudo_dir=‘/home/espresso-5.1/pseudo/’
outdir=‘/home/espresso-5.1/out/’
!outdir=‘/*output folder for large runs*/’
/
&system
ibrav=2, celldm(1)=10.263, nat=2, ntyp=1, ecutwfc=16.0
/
&electrons
mixing_beta=0.7, conv_thr= 1.0d-9
/
ATOMIC_SPECIES
Si 28.0855 Si.pz-vbc.UPF
ATOMIC_POSITIONS {alat}
Si 0.00 0.00 0.00
Si 0.25 0.25 0.25
K_POINTS {automatic}
4 4 4 1 1 1
ATOMIC_POSITIONS
Species atomic mass pseudopotential
^^ ^^
(the number of species here must agree with
the ntyp amount specified in &system)
ATOMIC_POSITIONS
species x y z
^^^^^^^^
(now the number of atoms here, must agree
with “nat” specified earlier)
K_POINTS
2 options: you can list your own set of k-
points –OR– you can do an automatic grid
that uniformly spaces k-points over the BZ.
More k-points = longer simulation.
This is also the second half to convergence.
&control
calculation=‘scf’
restart_mode=‘from_scratch’
prefix=‘Si’
pseudo_dir=‘/home/espresso-5.1/pseudo/’
outdir=‘/home/espresso-5.1/out/’
!outdir=‘/*output folder for large runs*/’
/
&system
ibrav=2, celldm(1)=10.263, nat=2, ntyp=1, ecutwfc=16.0
/
&electrons
mixing_beta=0.7, conv_thr= 1.0d-9
/
ATOMIC_SPECIES
Si 28.0855 Si.pz-vbc.UPF
ATOMIC_POSITIONS {alat}
Si 0.00 0.00 0.00
Si 0.25 0.25 0.25
K_POINTS {automatic}
4 4 4 1 1 1
ATOMIC_POSITIONS
Species atomic mass pseudopotential
^^ ^^
(the number of species here must agree with
the ntyp amount specified in &system)
ATOMIC_POSITIONS
species x y z
^^^^^^^^
(now the number of atoms here, must agree
with “nat” specified earlier)
K_POINTS
2 options: you can list your own set of k-
points –OR– you can do an automatic grid
that uniformly spaces k-points over the BZ.
More k-points = longer simulation.
This is also the second half to convergence.
Example Input files
# Calculation of the force constants on a grid of q-points for dispersion curves with ph.x
Phonons of Si
&inputph
tr2_ph=1.0d-16,
prefix=‘Si’,
amass(1)=28.0855,
outdir=‘/home/espresso-5.1/out/’
ldisp=.true.,
nq1=4, nq2=4, nq3=4
fildyn=‘si.dyn’
/
# Calculation for phonon DOS with matdyn.x
&input
asr=‘simple’
amass(1)=28.0855,
flfrc=‘si.444.fc’
flfrq=si.dos.freq’
dos=.true.
fldos=‘si.dos’
deltaE=1.d0
nk1=4, nk2=4, nk3=4,
/
Note: “!” Comments out a line in input files.
Note this is not the
general form of most
matdyn.x input files!!
Tr2_ph= convergence threshold
amass=atomic mass
ldisp=set ‘true’ if want phonon frequency over
Monkhorst-Pack grid of k-points
“Nq’s”= specify the size of the Monhorst- Pack grid”
(repeated terms have same definitions as before)
# Calculation of the force constants on a grid of q-points for dispersion curves with ph.x
Phonons of Si
&inputph
tr2_ph=1.0d-16,
prefix=‘Si’,
amass(1)=28.0855,
outdir=‘/home/espresso-5.1/out/’
ldisp=.true.,
nq1=4, nq2=4, nq3=4
fildyn=‘si.dyn’
/
# Calculation for phonon DOS with matdyn.x
&input
asr=‘simple’
amass(1)=28.0855,
flfrc=‘si.444.fc’
flfrq=si.dos.freq’
dos=.true.
fldos=‘si.dos’
deltaE=1.d0
nk1=4, nk2=4, nk3=4,
/
Note: “!” Comments out a line in input files.
Note this is not the
general form of most
matdyn.x input files!!
Tr2_ph= convergence threshold
amass=atomic mass
ldisp=set ‘true’ if want phonon frequency over
Monkhorst-Pack grid of k-points
“Nq’s”= specify the size of the Monhorst- Pack grid”
(repeated terms have same definitions as before)
Defining Crystal Structures
There are a few ways to define structures and materials in QE.
Most commonly we use ibrav=1,..,14.
ATOMIC_POSITIONS
Tells QE where atoms sit
relative to chosen ibrav.
ibrav=0
CELL_PARAMETERS
(ie lattice vectors)
Alloys and Supercell
Definitions!
Please follow link below for all ibrav
definitions and all details on pw.x input
files.
http://www.quantum-espresso.org/wp-content/uploads/Doc/INPUT_PW.html
There are a few ways to define structures and materials in QE.
Most commonly we use ibrav=1,..,14.
ATOMIC_POSITIONS
Tells QE where atoms sit
relative to chosen ibrav.
ibrav=0
CELL_PARAMETERS
(ie lattice vectors)
Alloys and Supercell
Definitions!
Please follow link below for all ibrav
definitions and all details on pw.x input
files.
http://www.quantum-espresso.org/wp-content/uploads/Doc/INPUT_PW.html
Unit cells to Supercells
(1) (2) (3)
ibrav=2
CELL_PARAMETERS
-0.5 0.0 0.5
0.0 0.5 0.5
-0.5 0.5 0.0
ibrav=1
CELL_PARAMETERS
1.0 0.0 0.0
0.0 1.0 0.0
0.0 0.0 1.0
ibrav=0
CELL_PARAMETERS
2.0 0.0 0.0
0.0 1.0 0.0
0.0 0.0 1.0
What about celldm!? CELL_PARAMETERS{ alat | bohr | angstrom}
(1) (2) (3)
ibrav=2
CELL_PARAMETERS
-0.5 0.0 0.5
0.0 0.5 0.5
-0.5 0.5 0.0
ibrav=1
CELL_PARAMETERS
1.0 0.0 0.0
0.0 1.0 0.0
0.0 0.0 1.0
ibrav=0
CELL_PARAMETERS
2.0 0.0 0.0
0.0 1.0 0.0
0.0 0.0 1.0
What about celldm!? CELL_PARAMETERS{ alat | bohr | angstrom}
Defining the Supercell
(3) is a supercell defined for a SiGe alloy where 11/36(x~=30.55%) of the atoms are Ge and 25/36 are Si.
The input file may look like this (where ibrav=0, and the lattice constant is the result of this equation:
5.431+0.20�+0.027�
2
��?????? ?????? ????????????
1−�??????�
� ??????���� )(1)
ATOMIC_SPECIES
Si 28.0855 Si.pz.vbc.UPF
Ge 72.64 Ge.pz-bhs.UPF
ATOMIC_POSITIONS
Si 0 0 0
Si 0.5 0.5 0
Si 0.5 0 0.5
Ge 0 0.5 0.5
Ge 0.25 0.25 0.25
Ge 0.75 0.75 0.25
Ge 0.75 0.25 0.75
Ge 0.25 0.75 0.75
Si 1 0.5 0.5
Si 1 1 0
Ge 1.25 0.25 0.25
Ge 1.5 0 0.5
Ge 1.5 0.5 0
Si 1.75 0.75 0.25
Si 1.75 0.25 0.75
Si 1.25 0.75 0.75
CELL_PARAMETERS { alat}
2 0 0
0 1 0
0 0 1
(1) http://www.ioffe.rssi.ru/SVA/NSM/Semicond/SiGe/basic.html
(3) is a supercell defined for a SiGe alloy where 11/36(x~=30.55%) of the atoms are Ge and 25/36 are Si.
The input file may look like this (where ibrav=0, and the lattice constant is the result of this equation:
5.431+0.20�+0.027�
2
��?????? ?????? ????????????
1−�??????�
� ??????���� )(1)
ATOMIC_SPECIES
Si 28.0855 Si.pz.vbc.UPF
Ge 72.64 Ge.pz-bhs.UPF
ATOMIC_POSITIONS
Si 0 0 0
Si 0.5 0.5 0
Si 0.5 0 0.5
Ge 0 0.5 0.5
Ge 0.25 0.25 0.25
Ge 0.75 0.75 0.25
Ge 0.75 0.25 0.75
Ge 0.25 0.75 0.75
Si 1 0.5 0.5
Si 1 1 0
Ge 1.25 0.25 0.25
Ge 1.5 0 0.5
Ge 1.5 0.5 0
Si 1.75 0.75 0.25
Si 1.75 0.25 0.75
Si 1.25 0.75 0.75
CELL_PARAMETERS { alat}
2 0 0
0 1 0
0 0 1
(1) http://www.ioffe.rssi.ru/SVA/NSM/Semicond/SiGe/basic.html
Pseudopotentials
Not as important
when considering
phonons.
Things to notice:
Suggested cutoff
values
Authors
empirically fitted to
handle different
cases.
Not as important
when considering
phonons.
Things to notice:
Suggested cutoff
values
Authors
empirically fitted to
handle different
cases.
Convergence!!!
Minimum Total Energy
Convergence:
ecutwfc – kinetic energy cutoff for plane waves
K-points – Discrete Monkhorst-Pack grid
Structure optimization
Relaxations, MD, variable cell dynamics
Hydrostatic strains, relaxing into other strains.
Batch scripts…
1 Ry = 13.605698066 eV
Minimum Total Energy
Convergence:
ecutwfc – kinetic energy cutoff for plane waves
K-points – Discrete Monkhorst-Pack grid
Structure optimization
Relaxations, MD, variable cell dynamics
Hydrostatic strains, relaxing into other strains.
Batch scripts…
1 Ry = 13.605698066 eV
Convergence of ecutwfc
(2) N. Zabaras. Quantum-Espresso. Cornell University (2/29/2012)
•For simple materials that don’t require the specification of ecutrho this may suffice for
ecutwfc.
•In any case it is necessary to compare this with the convergence of different k-point
grids (ie 4 4 4, 8 8 8, 12 12 12). Together these will lead to the structure with the least
total energy.
(2) Convergence for a
“minimum total
energy’ means that the
structure is at
equilibrium. In other
words, the structure is
at it’s optimal volume
which yields the
lowest total energy.
(2) N. Zabaras. Quantum-Espresso. Cornell University (2/29/2012)
•For simple materials that don’t require the specification of ecutrho this may suffice for
ecutwfc.
•In any case it is necessary to compare this with the convergence of different k-point
grids (ie 4 4 4, 8 8 8, 12 12 12). Together these will lead to the structure with the least
total energy.
(2) Convergence for a
“minimum total
energy’ means that the
structure is at
equilibrium. In other
words, the structure is
at it’s optimal volume
which yields the
lowest total energy.
My project: Effects of Strain
Want to know how strain effects heat transfer in silicon thin-films!
Strain a material, gather phonon frequencies, post proc with MATLAB
How does biaxial strain affect the unit cell
bond lengths and angles
Elastic Constants (stiffness constants)
??????
11 ??????
12 ??????
13
??????
21 ??????
22 ??????
23
??????
31 ??????
32 ??????
33
??????
44
Perpendicular effects
??????
⊥??????=??????
�??????�−�
??????��??????
??????��??????
??????∥??????−??????�??????
??????�??????
(3)
What does this mean for phonons exactly
Suspect that phonon group velocities may speed up or slow down in certain directions.
Z
E
R
O
S
Z E R O S
(3) Rieger, Vogl. Electronic-band parameters in strained ????????????
1−� ??????�
�alloys on ????????????
1−� ??????�
� substrates. PRB Vol. 48 #19 Nov. 1993
Want to know how strain effects heat transfer in silicon thin-films!
Strain a material, gather phonon frequencies, post proc with MATLAB
How does biaxial strain affect the unit cell
bond lengths and angles
Elastic Constants (stiffness constants)
??????
11 ??????
12 ??????
13
??????
21 ??????
22 ??????
23
??????
31 ??????
32 ??????
33
??????
44
Perpendicular effects
??????
⊥??????=??????
�??????�−�
??????��??????
??????��??????
??????∥??????−??????�??????
??????�??????
(3)
What does this mean for phonons exactly
Suspect that phonon group velocities may speed up or slow down in certain directions.
Z
E
R
O
S
Z E R O S
(3) Rieger, Vogl. Electronic-band parameters in strained ????????????
1−� ??????�
�alloys on ????????????
1−� ??????�
� substrates. PRB Vol. 48 #19 Nov. 1993
Effects of Strain in QE
&control
calculation=‘scf’
restart_mode=‘from_scratch’
prefix=‘Strained_Si’
pseudo_dir=‘/home/espresso-5.1/pseudo/’
outdir=‘/home/espresso-5.1/out/’
/
&system
ibrav=10, A=5.646.B=5.646,C=5.234,
cosAB=0, cosBC=0, cosAC=0,
nat=2, ntyp=1, ecutwfc=16.0
/
&electrons
mixing_beta=0.7, conv_thr= 1.0d-9
/
ATOMIC_SPECIES
Si 28.0855 Si.pz-vbc.UPF
ATOMIC_POSITIONS {crystal | alat | bohr | angstrom}
Si 0.00 0.00 0.00
Si 0.25 0.25 0.25
K_POINTS {automatic}
4 4 4 1 1 1
- BZ is morphed due to the
strain. (Be careful of what
kind of effect this is.)
[−
2??????
??????
,
2??????
??????
]
&control
calculation=‘scf’
restart_mode=‘from_scratch’
prefix=‘Strained_Si’
pseudo_dir=‘/home/espresso-5.1/pseudo/’
outdir=‘/home/espresso-5.1/out/’
/
&system
ibrav=10, A=5.646.B=5.646,C=5.234,
cosAB=0, cosBC=0, cosAC=0,
nat=2, ntyp=1, ecutwfc=16.0
/
&electrons
mixing_beta=0.7, conv_thr= 1.0d-9
/
ATOMIC_SPECIES
Si 28.0855 Si.pz-vbc.UPF
ATOMIC_POSITIONS {crystal | alat | bohr | angstrom}
Si 0.00 0.00 0.00
Si 0.25 0.25 0.25
K_POINTS {automatic}
4 4 4 1 1 1
- BZ is morphed due to the
strain. (Be careful of what
kind of effect this is.)
[−
2??????
??????
,
2??????
??????
]
Effects of Strain in QE
&control
calculation=‘scf’
restart_mode=‘from_scratch’
prefix=‘Strained_Si’
pseudo_dir=‘/home/espresso-5.1/pseudo/’
outdir=‘/home/espresso-5.1/out/’
/
&system
ibrav=10, A=5.646.B=5.646,C=5.234,
cosAB=0, cosBC=0, cosAC=0,
nat=2, ntyp=1, ecutwfc=16.0
/
&electrons
mixing_beta=0.7, conv_thr= 1.0d-9
/
ATOMIC_SPECIES
Si 28.0855 Si.pz-vbc.UPF
ATOMIC_POSITIONS {crystal | alat | bohr | angstrom}
Si 0.00 0.00 0.00
Si 0.25 0.25 0.25
K_POINTS {automatic}
4 4 4 1 1 1
Notice the changes in the &systems namelist
for strain materials:
•ibrav=10 specifies orthorhormbic fcc
•A,B,C, cosAB, cosBC, cosAC relate
directly to celldm(1-6) and allow for
full flexibility in specifying the
structure.
•A,B,C units are in angstroms.
•“cyrstal” is specified for
ATOMIC_POSITIONS this changes
how the values specified below it are
represented in the code. See
INPUT_PW.txt in ~/PW/doc .
– OR – the webpage at the bottom of
slide 14
&control
calculation=‘scf’
restart_mode=‘from_scratch’
prefix=‘Strained_Si’
pseudo_dir=‘/home/espresso-5.1/pseudo/’
outdir=‘/home/espresso-5.1/out/’
/
&system
ibrav=10, A=5.646.B=5.646,C=5.234,
cosAB=0, cosBC=0, cosAC=0,
nat=2, ntyp=1, ecutwfc=16.0
/
&electrons
mixing_beta=0.7, conv_thr= 1.0d-9
/
ATOMIC_SPECIES
Si 28.0855 Si.pz-vbc.UPF
ATOMIC_POSITIONS {crystal | alat | bohr | angstrom}
Si 0.00 0.00 0.00
Si 0.25 0.25 0.25
K_POINTS {automatic}
4 4 4 1 1 1
Notice the changes in the &systems namelist
for strain materials:
•ibrav=10 specifies orthorhormbic fcc
•A,B,C, cosAB, cosBC, cosAC relate
directly to celldm(1-6) and allow for
full flexibility in specifying the
structure.
•A,B,C units are in angstroms.
•“cyrstal” is specified for
ATOMIC_POSITIONS this changes
how the values specified below it are
represented in the code. See
INPUT_PW.txt in ~/PW/doc .
– OR – the webpage at the bottom of
slide 14
Results Context
Thermal conductivity (4)
??????
��=
ℏ
2
??????Ω??????
????????????
2
??????
�????????????
�????????????
??????
2
�
??????�
??????+1??????
????????????
c – group velocities(defined below), n – equilibrium populations, ?????? –
frequencies, N – size of q-point mesh, T- temp, Ω- cell volume, and
lambda refers to a particular q-vector and phonon branch(1-6).
Phonon Group Velocities
??????
??????�=??????
????????????(�)
(4) Garg, et al. Role of Disorder and Anharmonicity in the thermal conductivity of Silicon-Germanium Alloys: A first-principles study. PRL 106, 045901 (2011)
Thermal conductivity (4)
??????
��=
ℏ
2
??????Ω??????
????????????
2
??????
�????????????
�????????????
??????
2
�
??????�
??????+1??????
????????????
c – group velocities(defined below), n – equilibrium populations, ?????? –
frequencies, N – size of q-point mesh, T- temp, Ω- cell volume, and
lambda refers to a particular q-vector and phonon branch(1-6).
Phonon Group Velocities
??????
??????�=??????
????????????(�)
(4) Garg, et al. Role of Disorder and Anharmonicity in the thermal conductivity of Silicon-Germanium Alloys: A first-principles study. PRL 106, 045901 (2011)
Recent Results
??????
��=
ℏ
2
??????Ω??????????????????
2
??????
�????????????
�????????????
??????
2
�
??????�
??????+1??????
????????????
; ??????
??????�=??????
????????????(�)
??????
��=
ℏ
2
??????Ω??????????????????
2
??????
�????????????
�????????????
??????
2
�
??????�
??????+1??????
????????????
; ??????
??????�=??????
????????????(�)
Results
??????
��=
ℏ
2
??????Ω??????????????????
2
??????
�????????????
�????????????
??????
2
�
??????�
??????+1??????
????????????
; ??????
??????�=??????
????????????(�)
??????
��=
ℏ
2
??????Ω??????????????????
2
??????
�????????????
�????????????
??????
2
�
??????�
??????+1??????
????????????
; ??????
??????�=??????
????????????(�)
Results
When we look at the dispersions we are looking for two
effects:
Any change in height in the actual branches change in w(k)
And any change in slope of the branches change in group velocity
Either effect will carry through the thermal conductivity
equation and result in a overall change in conductivity.
We also want to see if the effects are different in different
directions.
Eg the plots on the prior two slides are plotted for the
100 and 001 directions (100 and 010 should be
equivalent in this case due to symmetry)
When we look at the dispersions we are looking for two
effects:
Any change in height in the actual branches change in w(k)
And any change in slope of the branches change in group velocity
Either effect will carry through the thermal conductivity
equation and result in a overall change in conductivity.
We also want to see if the effects are different in different
directions.
Eg the plots on the prior two slides are plotted for the
100 and 001 directions (100 and 010 should be
equivalent in this case due to symmetry)
Isosurfaces
Unstrained , branch 1, at w=2.5e13 same but branch 2
•Constant energy surfaces are good for visualizing the shapes of certain
branches at different energies.
•Also very good for showing asymmetric properties as a result of strain.
•The isosurfaces here are of Silicon at branch 1 and 2.
Unstrained , branch 1, at w=2.5e13 same but branch 2
•Constant energy surfaces are good for visualizing the shapes of certain
branches at different energies.
•Also very good for showing asymmetric properties as a result of strain.
•The isosurfaces here are of Silicon at branch 1 and 2.
Batch Scripts
Want to do convergence testing?
Need to run a bunch of executables but can’t wait
around running each one at a time?
Batch scripts solve this pretty easily.
Iterative convergence testing!
Getting dispersion for many paths!
Example…
Want to do convergence testing?
Need to run a bunch of executables but can’t wait
around running each one at a time?
Batch scripts solve this pretty easily.
Iterative convergence testing!
Getting dispersion for many paths!
Example…
Using Clusters
When do I use clusters and why?
Lengthy simulations: more than 20 hrs.
ph.x is the big time killer (other calculations are usually within
5-10 mins)
Interactive jobs versus queue submissions(qsub).
Interactive jobs are good for finding convergence of
resources(if necessary).
queue’d jobs are where most researchers conduct work these
are batch scripts containing envrionment variables and
commands for running a “number” of simulations.
GORDON
When do I use clusters and why?
Lengthy simulations: more than 20 hrs.
ph.x is the big time killer (other calculations are usually within
5-10 mins)
Interactive jobs versus queue submissions(qsub).
Interactive jobs are good for finding convergence of
resources(if necessary).
queue’d jobs are where most researchers conduct work these
are batch scripts containing envrionment variables and
commands for running a “number” of simulations.
GORDON
Resources
www.quantum-espresso.org
Here you’ll find pseudopotential for download, all other necessary downloads for
packages and extensions, and all links for tutorials and documentation.
www.quantum-espresso.org/wp-
content/uploads/Doc/INPUT_PW.html#__top__
Where I spend all my time
Pw_forum (need to subscribe, its free though)
This webpage has lots of tutorials and slides for describing both theory
and the use of QE: http://www.fisica.uniud.it/~giannozz/QE-Tutorial/
www.quantum-espresso.org
Here you’ll find pseudopotential for download, all other necessary downloads for
packages and extensions, and all links for tutorials and documentation.
www.quantum-espresso.org/wp-
content/uploads/Doc/INPUT_PW.html#__top__
Where I spend all my time
Pw_forum (need to subscribe, its free though)
This webpage has lots of tutorials and slides for describing both theory
and the use of QE: http://www.fisica.uniud.it/~giannozz/QE-Tutorial/
Downloading and installing QE (serial version)
Go to the webpage: www.quantum-espresso.org
Click the download tab, and find the highlighted link that says
“download page”
You should now be on a page that lists a lot of tar.gz files. This is
the download page.
Next, find an appropriate version of espresso (eg espresso-
5.1.tar.gz is what I downloaded)
Untar the folder and browse the “Doc” and “install” folders for
READ_ME.txt’s and user_guides.
Also check out the PW and PHonon folders for other helpful
Doc’s and user_guides.
Go to the webpage: www.quantum-espresso.org
Click the download tab, and find the highlighted link that says
“download page”
You should now be on a page that lists a lot of tar.gz files. This is
the download page.
Next, find an appropriate version of espresso (eg espresso-
5.1.tar.gz is what I downloaded)
Untar the folder and browse the “Doc” and “install” folders for
READ_ME.txt’s and user_guides.
Also check out the PW and PHonon folders for other helpful
Doc’s and user_guides.
More on installation
Next, from the terminal, enter the untarred folder and run these
two commands:
compile
make all
Note: make is the default cmd for compiling make files, which simply compile other
codes.
If ‘make’ throws an error, most likely the folders in the ‘archive’
folder did not download correctly. The ‘archive’ folder contains
backup libraries necessary for running some calculations. You can
fix this by manually re-downloading them from the download page.
Once ‘make’ is complete, you’re done and you can begin using QE!
For parallel version installation see the ‘install’ folder for files like
README.CRAY-XE6, the CRAY-XE6 refers to the cluster’s node
type.
Next, from the terminal, enter the untarred folder and run these
two commands:
compile
make all
Note: make is the default cmd for compiling make files, which simply compile other
codes.
If ‘make’ throws an error, most likely the folders in the ‘archive’
folder did not download correctly. The ‘archive’ folder contains
backup libraries necessary for running some calculations. You can
fix this by manually re-downloading them from the download page.
Once ‘make’ is complete, you’re done and you can begin using QE!
For parallel version installation see the ‘install’ folder for files like
README.CRAY-XE6, the CRAY-XE6 refers to the cluster’s node
type.
Tags
Categories
GeneralDownload
Download Slideshow Get the original presentation fileQuick Actions
Statistics
- Views 6,790
- Slides 32
- Favorites 8
- Age 3543 days
Related Slideshows
22
Pray For The Peace Of Jerusalem and You Will Prosper
RodolfoMoralesMarcuc
33 views
26
Don_t_Waste_Your_Life_God.....powerpoint
chalobrido8
36 views
31
VILLASUR_FACTORS_TO_CONSIDER_IN_PLATING_SALAD_10-13.pdf
JaiJai148317
33 views
14
Fertility awareness methods for women in the society
Isaiah47
30 views
35
Chapter 5 Arithmetic Functions Computer Organisation and Architecture
RitikSharma297999
29 views
5
syakira bhasa inggris (1) (1).pptx.......
ourcommunity56
30 views