VASP Tutorial:

A bit of surface science

University of Vienna,
Faculty of Physics and Center for Computational Materials Science,
Vienna, Austria
 Setting up a VASP calculation

VASP requires 4 input files to run a calculation:

• INCAR
• POSCAR
• KPOINTS
• POTCAR
 I: The INCAR file
The INCAR file contains the input parameters that steer the calculation:

• The default values set by VASP itself are a clever choice for most
standard calculations

• These standard settings may be modified to specify:

• What kind of calculation you want to do:

SCF calculation, DOS, dielectric properties, …

• Basic inputs concering the required precision, the requested

level of convergence, ...

For a list of all INCAR-tags have a look at:

• The VASP manual: http://cms.mpi.univie.ac.at/vasp/vasp/vasp.html
Index: http://cms.mpi.univie.ac.at/vasp/vasp/Index.html
• The VASP wiki: http://cms.mpi.univie.ac.at/wiki/index.php/Main_page
INCAR-tags: http://cms.mpi.univie.ac.at/wiki/index.php/Category:INCAR
 II: The POSCAR file
In the POSCAR file you specify the structure (Bravais lattice and basis):

fcc: Ni Header (comment)

3.53 Overall scaling constant
0.5 0.5 0.0 Bravais matrix
0.0 0.5 0.5
0.5 0.0 0.5
Ni Name(s) of atomic type(s)
1 Number of atoms (of each type)
Selective Dynamics (optional: selective dynamics)
Cartesian Cartesian or Direct coordinates
0 0 0 (T T T) positions of the atoms
 III: The KPOINTS file
In the KPOINTS file you specify the points VASP will use to sample the
first Brillouin zone in reciprocal space

Automatic mesh Header (comment)

0 Nk=0: automatic mesh generation
G (M) 𝝘-centered (G) mesh or Monkhorst-Pack (M) grid
4 4 4 # of subdivisions 𝑁𝑖 along 𝑏Ԧ𝑖
0. 0. 0. Optionally shift the mesh (𝑠𝑖 )
 IV: The POTCAR file
The POTCAR file has to contain the PAW datasets for all atomic types you have
specified in your POSCAR file:

VASP comes with a library of PAW datasets, (one or more) for most elements of the
periodic table:

• Each individual PAW data set starts with a descriptive section, specifying
amongst other things:
• Parameters that were required to generate the dataset:
• Number of valence electrons
• Atomic mass
• Default energy cutoffs

• When your unit cell contains more than one type of atom you have to
concatenate the corresponding PAW datasets in the same order as you have
specified the different atomic types in your POSCAR file.

• You should not mix PAW datasets generated with different exchange-correlation
functionals.
 OUTPUT files
OUTCAR
• detailed output of a VASP run, including:
• a summary of the input parameters
• information about the individual electronic steps:
total energy, Kohn-Sham eigenvalues, Fermi-energy.
• stress tensors
• forces in the atoms
• local charges, magnetic moments
• dielectric properties
• ... and a great many things more ...
• The amount of output written onto OUTCAR can be chosen by
means of the NWRITE-tag in the INCAR file.

OSZICAR & stdout

• give a short summary of the self-consistency-cycle
• chosen SCF algorithm
• convergence of energy and charge density
• free energies, total magnetic moment of the cell
 OUTPUT files
CONTCAR & XDATCAR
• CONTCAR: updated geometry data at the end of a run
• lattice parameter
• Bravais matrix
• ionic positions
• velocities
• the format of the CONTCAR is the same as for POSCAR:
It can be directly be used for a continuation run (copy
CONTCAR to POSCAR)
• XDATCAR: updated ionic positions of each ionic step

DOSCAR, CHGCAR & WAVECAR

• DOSCAR: total DOS and integrated DOS, (local partial DOS)
• CHGCAR: the charge density
• WAVECAR: plane wave coefficients of the orbitals.
Can be used to restart from a previous run
 Documentation

• The VASP manual (http://cms.mpi.univie.ac.at/vasp/vasp/vasp.html)

Index: http://cms.mpi.univie.ac.at/vasp/vasp/Index.html

• The VASP wiki (http://cms.mpi.univie.ac.at/wiki/index.php/Main_page)

INCAR-tags: http://cms.mpi.univie.ac.at/wiki/index.php/Category:INCAR
 A bit of surface science
Examples:

Ni(100):
• surface relaxation
• surface energy
• LDOS
• surface bandstructure

Ni(111)
• clean surface
• CO adsorption
• LDOS
• workfunction (change)
• frequencies

STM of graphite and graphene

 Ni(100) surface relaxation (ex.: Ni100clean_rel)
POSCAR:
fcc (100) surface Header (comment)
3.53 Ni lattice constant
.50000 .50000 .00000 lattice vector a(1)
-.50000 .50000 .00000 lattice vector a(2)
.00000 .00000 5.00000 lattice vector a(3)
5 Number of atoms
Selective Dynamics Switch on ”selective dynamics”
Cartesian Positions in cartesian coordinates
.00000 .00000 .000000 F F F positions of the atoms, and the
.00000 .50000 .500000 F F F specfications whether or not they
.00000 .00000 1.00000 F F F are allowed to move during the
.00000 .50000 1.50000 T T T relaxation.
.00000 .00000 2.00000 T T T

• A 𝑝(1 × 1) surface cell: 1 Ni atom per layer POTCAR:

• 5 Ni layers Ni GGA PAW potential
• first two layers (on one side) are relaxed
• 3 × 3.53 = 10.59 Å vacuum
 Ni(100) surface relaxation (ex.: Ni100clean_rel)
INCAR:
SYTEM = clean Ni(100) surface Name of the calculation
ISTART = 0 initial wave functions: random numbers
ICHARG = 2 initial charge density: overlapping atoms
ENCUT = 270 cutoff energy 270 eV (from POTCAR)
ALGO = Fast use RMM-DIIS for electronic optimization
EDIFF = 1E-6 electronic convergence: energy change < 10-6 eV
ISMEAR = 2 2nd order Methfessel-Paxton smearing (metal!)
SIGMA = 0.2 smearing width 𝜎 = 0.2 eV
ISPIN = 2 spin-polarized calculation
MAGMOM = 5*1 initial magnetic moment on each Ni = 1 𝜇 𝐵
IBRION = 1 ionic relaxation
NSW = 100
POTIM = 0.8

KPOINTS:
K-Points Header (comment)
0 Nk=0: automatic mesh generation
Monkhorst-Pack Monkhorst-Pack grid
9 9 1 # of subdivisions 𝑁𝑖 along 𝑏Ԧ𝑖 (odd: centered on 𝝘)
0. 0. 0. Optionally shift the mesh (𝑠𝑖 )

• 15 k-points in the IBZ

• 1 k-point along the z-direction!
 Ni(100) surface relaxation (ex.: Ni100clean_rel)
Forces in the first and last step of the relaxation (in OUTCAR)
POSITION TOTAL-FORCE (eV/Angst)
-----------------------------------------------------------------------------------
0.00000 0.00000 0.00000 0.000000 0.000000 0.391783
0.00000 1.76500 1.76500 0.000000 0.000000 -0.397182
0.00000 0.00000 3.53000 0.000000 0.000000 0.005392
0.00000 1.76500 5.29500 -0.000000 -0.000000 0.390888
0.00000 0.00000 7.06000 0.000000 0.000000 -0.390881
-----------------------------------------------------------------------------------
total drift: 0.000000 -0.000000 0.016601

POSITION TOTAL-FORCE (eV/Angst)

-----------------------------------------------------------------------------------
0.00000 0.00000 0.00000 -0.000000 0.000000 0.396242
0.00000 1.76500 1.76500 0.000000 0.000000 -0.394948
0.00000 0.00000 3.53000 0.000000 -0.000000 0.000021
0.00000 1.76500 5.30249 0.000000 0.000000 -0.000262
0.00000 0.00000 6.98837 -0.000000 -0.000000 -0.001054
-----------------------------------------------------------------------------------
total drift: 0.000000 -0.000000 0.000212
 Ni(100) surface relaxation (ex.: Ni100clean_rel)
Energy Convergence
• Energy changes during relaxation
-25.554
from -25.556 eV to -25.572 eV -25.556 1 2 3 4 5
⟹ 𝐸rel = −16 meV -25.558
-25.560

TOTEN (EV)
-25.562
• Surface energy: -25.564
-25.566
-25.568
𝜎 = 12 𝐸surf − 𝑁atoms × 𝐸bulk
-25.570
-25.572
-25.574
STEPS
• Surface energy of unrelaxed surface:
⟹ 𝜎 unrel =12 −25.556 − 5 × −5.458 = 0.867 eV

• 𝜎 = 𝜎 unrel + 𝐸rel = 0.867 − 0.016 = 0.851 eV

N.B.: you will find the setup for the calculation of the “bulk” energy in the
Ni100clean_rel/bulk subdirectory.
 Ni(100) surface relaxation (ex.: Ni100clean_rel)

Start p4vasp:
> p4v [vasprun.xml]

Step. 1) Go to:
Convergence/Energy

Step. 2) “update” graph

 Ni(100) surface relaxation (ex.: Ni100clean_rel)
Final geometry from CONTCAR (or OUTCAR) file:
fcc (100) surface
3.53000000000000
0.5000000000000000 0.5000000000000000 0.0000000000000000
-0.5000000000000000 0.5000000000000000 0.0000000000000000
0.0000000000000000 0.0000000000000000 5.0000000000000000
Ni
5
Selective dynamics
Direct
0.0000000000000000 0.0000000000000000 0.0000000000000000 F F F
0.5000000000000000 0.5000000000000000 0.1000000000000014 F F F
0.0000000000000000 0.0000000000000000 0.2000000000000028 F F F
0.5000000000000000 0.5000000000000000 0.3004245271852446 T T T
0.0000000000000000 -0.0000000000000000 0.3959414474619545 T T T

0.00000000E+00 0.00000000E+00 0.00000000E+00

0.00000000E+00 0.00000000E+00 0.00000000E+00
0.00000000E+00 0.00000000E+00 0.00000000E+00
0.00000000E+00 0.00000000E+00 0.00000000E+00
0.00000000E+00 0.00000000E+00 0.00000000E+00

• Inward/outward relaxation of the surface layers:

Δ𝑑54 = 0.3959 − 0.3004 − 0.1 Τ0.1 × 100 = −4.5%
Δ𝑑43 = 0.3004 − 0.2000 − 0.1 Τ0.1 × 100 = +0.4%
 Ni(100) surface relaxation (ex.: Ni100clean_rel)
Start p4vasp:> p4v [vasprun.xml] Possibly
Step 1.) Show structure “replicate” cells

Step 2.) Select “initial” or “final” positions

 Ni(100) LDOS (ex.: Ni100clean_LDOS)

INCAR:
SYTEM = clean Ni(100) surface Name of the calculation
ENCUT = 270 cutoff energy 270 eV (from POTCAR)
ALGO = Normal use block-Davidson for electronic minimization

ISMEAR = -5 Tetrahedron method with Blöchl corrections

ISPIN = 2 spin-polarized calculation
MAGMOM = 5*1 initial magnetic moment on each Ni = 1 𝜇 𝐵
LORBIT = 11 LM-decomposed site resolved density of states

N.B.: we want to use the optimized structure from Ni100clean_rel:

Normally this would mean copying Ni100clean_rel/CONTCAR to

POSCAR in the directory where you want to run Ni100clean_LDOS.

In this case, however, that has already been taken care of, and the
POSCAR file in Ni100clean_LDOS is the correct one.
 Ni(100) LDOS (ex.: Ni100clean_LDOS)

total charge
# of ion s p d tot
• At the end of the OUTCAR file
------------------------------------------ information on local charge and
1 0.466 0.326 8.314 9.106
2 0.490 0.481 8.333 9.304 magnetization is given.
3 0.494 0.482 8.338 9.313
4 0.500 0.501 8.350 9.351
5 0.478 0.346 8.345 9.169 • Instead of LORBIT=11, one might
--------------------------------------------------
tot 2.427 2.135 41.681 46.244 use LORBIT=1 and set RWIGS
appropriately.
magnetization (x)
# of ion s p d tot • As is clearly shown, the local
------------------------------------------
1 -0.003 -0.019 0.725 0.703
magnetic moments at the surface
2 -0.008 -0.024 0.613 0.582 are enhanced.
3 -0.008 -0.024 0.611 0.579
4 -0.008 -0.024 0.605 0.573
5 -0.004 -0.020 0.703 0.680
--------------------------------------------------
• The central layers behave
tot -0.030 -0.111 3.257 3.116 “bulk”-like.
 Ni(100) LDOS (ex.: Ni100clean_LDOS)

• Projection onto surface

and bulk layers.

• Each spin component is

plotted separately.

• Band narrowing at the

surface.

• Exchange splitting is
larger at the surface.
 Ni(100) LDOS (ex.: Ni100clean_LDOS)
Select spin channel
Select: “Electronic/Local DOS+Bands control”

Select atom,
and orbital character

Inver y-axis (optional)

Place pointer over

atom, and press
space to
select/deselect.

The selection
should appear in
the Electronic
Control applet
above
Press “Show” to show structure
Ni(100) surface bandstructure (ex.: Ni100clean_band)

INCAR:
SYTEM = clean Ni(100) surface Name of the calculation
ENCUT = 270 cutoff energy 270 eV (from POTCAR)
ALGO = Normal use block-Davidson for electronic minimization

ISMEAR = 2 2nd order Methfessel-Paxton smearing (metal!)

SIGMA = 0.2 smearing width 𝜎 = 0.2 eV
ISPIN = 2 spin-polarized calculation
MAGMOM = 5*1 initial magnetic moment on each Ni = 1 𝜇 𝐵
LORBIT = 11 LM-decomposed site resolved density of states
ICHARG = 11 Read initial charge from CHGCAR (ICHARG=1), and
keep it fixed (ICHARG=ICHARG+10) during the
subsequent calculation

N.B.: You need to copy the self-consistent charge density (CHGCAR) from
Ni100clean_LDOS to the directory where you want to run Ni100clean_band.

You need to do this: if VASP can not read the CHGCAR file, the run will terminate.
Ni(100) surface bandstructure (ex.: Ni100clean_band)
KPOINTS:
• 13 k-points along Γത − ഥ X−M ഥ − Γത
• Explicitly specfied, in reciprocal
kpoints for band-structure G-X-M-G
13 coordinates
reziprok • All points with weight 1
.00000 .00000 .00000 1
.12500 .00000 .00000 1
.25000 .00000 .00000 1
.37500 .00000 .00000 1
.50000 .00000 .00000 1

.50000 .12500 .00000 1

.50000 .25000 .00000 1
.50000 .37500 .00000 1
.50000 .50000 .00000 1

.37500 .37500 .00000 1

.25000 .25000 .00000 1
.12500 .12500 .00000 1
.00000 .00000 .00000 1
Ni(100) surface bandstructure (ex.: Ni100clean_band)
...
Static calculation
charge density remains constant during run
spin polarized calculation • status message in OUTCAR
...
Bandstructure (projected) on actual job:
⇒ non-selfconsistent calc.

• Bandstructure consist mainly

of bulk-like bands.

• Dots mark localization at

surface layer.

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6

Ni(100) surface bandstructure (ex.: Ni100clean_band)
Select: “Electronic/Local DOS+Bands control”

Select atom,
and orbital character

Place pointer over

atom, and press
space to
select/deselect.

The selection
should appear in
the Electronic
Control applet
Select “Bands” above
Press “Show” to show structure
 Ni(111) surface relaxation (ex.: Ni111clean_rel)
INCAR:
SYTEM = clean Ni(100) surface Name of the calculation
ISTART = 0 initial wave functions: random numbers
ICHARG = 2 initial charge density: overlapping atoms
ENCUT = 270 cutoff energy 270 eV (from POTCAR)
ALGO = Fast use RMM-DIIS for electronic optimization
EDIFF = 1E-6 electronic convergence: energy change < 10-6 eV
ISMEAR = 2 2nd order Methfessel-Paxton smearing (metal!)
SIGMA = 0.2 smearing width 𝜎 = 0.2 eV
IBRION = 1 ionic relaxation
NSW = 100
POTIM = 0.8

• Essentially the same INCAR file as used in Ni100clean_rel.

• Spin polarization neglected.
 Ni(111) surface relaxation (ex.: Ni111clean_rel)
POSCAR:
Ni - (111) Header (comment)
3.53 Ni lattice constant
.70710678 .0000000 .000000 lattice vector a(1)
-0.35355339 0.6123724 .000000 lattice vector a(2)
.000000 .000000 5.196152 lattice vector a(3)
5 Number of atoms
selective dynamics Switch on ”selective dynamics”
direct Positions in cartesian coordinates
.00000000 .00000000 .00000000 F F F positions of the atoms, and the
.33333333 .66666667 .11111111 F F F specfications whether or not they
.66666667 .33333333 .22222222 F F F are allowed to move during the
.00000000 .00000000 .33333333 T T T relaxation.
.33333333 .66666667 .44444444 T T T

Similar setup as for (100) surface:

• again 5 layers, 2 relaxed
• 1 − .444 × 5.196 × 3.53 ≈ 10.2 Å of vacuum
 Ni(111) surface relaxation (ex.: Ni111clean_rel)

Geometry and surface energy:

POSITION TOTAL-FORCE (eV/Angst)
-----------------------------------------------------------------------------------
0.00000 0.00000 0.00000 0.000000 -0.000000 0.179780
-0.00000 1.44112 2.03805 0.000000 -0.000000 -0.064012
1.24804 0.72056 4.07609 -0.000000 0.000000 -0.116688
0.00000 -0.00000 6.09140 0.000000 0.000000 0.000253
-0.00000 1.44112 8.09550 -0.000000 0.000000 0.000667
-----------------------------------------------------------------------------------
total drift: -0.000032 0.000002 -0.018784

• Forces at the beginning already rather small

⟹ small relaxations for compact surfaces

• For surface energy we need non-spin-polarized bulk nickel as reference:

the setup for the calculation of the “bulk” energy you’ll find in the
Ni111clean_rel/bulk subdirectory.

𝜎 unrel = 12 −25.737 − 5 × −5.407 =0.649

• (111) surface more stable than the (100) surface

 CO@Ni(111) (ex.: COonNi111_rel)
POSCAR:
Ni - (111)
• Compared to previous examples,
3.53 we now have two additional atomic
.70710678 .0000000 .000000 types (C & O) ⟹ change POTCAR!
-0.35355339 0.6123724 .000000
.000000 .000000 5.1961524
5 1 1 • Concatenate nickel, carbon,
selective dynamics
direct and oxygen PAW datasets.
.00000000 .00000000 .00000000 F F F
.33333333 .66666667 .11111111 F F F
.66666667 .33333333 .22222222 F F F
.00000000 .00000000 .33333333 T T T
.33333333 .66666667 .44444444 T T T
.33333333 .66666667 .54029062 T T T
.33333333 .66666667 .60298866 T T T

CO molecule put above surface atom ⟹ “on-top”:

• 𝑧𝐶 = .540 − .444 × 5.196 × 3.53 ≈ 1.76 Å

• 𝑧𝐶𝑂 = .603 − .540 × 5.196 × 3.53 ≈ 1.16 Å
 CO@Ni(111) (ex.: COonNi111_rel)
Geometry
POSITION TOTAL-FORCE (eV/Angst)
-----------------------------------------------------------------------------------
0.00000 0.00000 0.00000 0.000000 -0.000000 0.174586
-0.00000 1.44112 2.03805 0.000000 -0.000000 -0.107681
1.24804 0.72056 4.07609 -0.000000 -0.000000 -0.073588
-0.00000 0.00000 6.10111 0.000000 -0.000000 -0.000361
-0.00000 1.44112 8.15449 -0.000000 -0.000000 -0.002450
-0.00000 1.44112 9.91089 0.000000 -0.000000 0.014846
-0.00000 1.44112 11.06589 -0.000000 0.000000 -0.005352
-----------------------------------------------------------------------------------
total drift: -0.000177 -0.000361 0.013511

• Small outward relaxation of the surface due to adsorption:

∆𝑑12 = 8.154 − 6.101 Τ2.038 × 100 − 100 = 0.74%

• CO geometry:
𝑑𝐶𝑂 = 11.066 − 9.911 = 1.155 Å
𝑧𝐶𝑁𝑖 = 9.911 − 8.154 = 1.757 Å
Ni(111) with higher cutoff (ex.: Ni111clean_400eV)
• the PAW datasets for carbon and oxygen used in example COonNi111_rel
require a plane wave energy cutoff of 400 eV (see C and O POTCAR files).
• The previous calculation of the Ni(111) clean surface was done with an
energy cutoff of 270 eV

⟹ To be able to calculate the adsorption energy we need to run a single point

calculation for the Ni(111) clean surface at a higher cutoff (400eV).
SYTEM = clean Ni(111) surface Name of the calculation
ISTART = 0 initial wave functions: random numbers
ICHARG = 2 initial charge density: overlapping atoms
ENMAX = 400 cutoff energy 400 eV
ALGO = Fast use RMM-DIIS for electronic optimization
EDIFF = 1E-6 electronic convergence: energy change < 10-6 eV
ISMEAR = 2 2nd order Methfessel-Paxton smearing (metal!)
SIGMA = 0.2 smearing width 𝜎 = 0.2 eV
LVHAR = .TRUE. Write electrostatic potential to LOCPOT file, needed
# LVTOT = .TRUE. to calculate the work function

• Change in cutoff lowers total energy: N.B.: the setup for the calculation
ENCUT=270 eV: 𝐸 𝜎 → 0 = −25.737 of “ECO”, you will find in the
Ni111clean_400eV/CO subdirectory
ENCUT=400 eV: 𝐸 𝜎 → 0 = −25.742
• Adsorption energy: 𝐸ads = 𝐸total − 𝐸clean − 𝐸CO
𝐸ads = −40.829 + 25.742 + 14.835 = −0.252 eV
Ni(111) with higher cutoff (ex.: Ni111clean_400eV)

• Use p4vasp to show the planar

average of the potential (LOCPOT).

• vacuum-potential 𝐸vac = 5.45 eV

• Fermi-level 𝜀𝐹 = 0.226 eV
(in OUTCAR)

• Φ = 𝐸vac − 𝜀𝐹 = 5.22 eV
Ni(111) with higher cutoff (ex.: Ni111clean_400eV)
Go to “Electronic/Local potential”
Select “Z - direction”

“Measure” point
in graph

Zoom to area of
interest

Start p4vasp:
> p4v [vasprun.xml]
 CO@Ni(111) LDOS&Workfunction
(ex.: COonNi111_LDOS)
INCAR:
SYSTEM = CO adsorption on Ni(111) Name of the calculation
ENMAX = 400 cutoff energy 400 eV
ISMEAR = -5 tetrahedron method with Blöchl corrections
ALGO = Fast use RMM-DIIS for electronic optimization
LORBIT = 11 LM-decomposed site resolved density–of-states
IDIPOL = 3 Enable dipole corrections in direction 3
LDIPOL = .TRUE. Switch on dipole corrections to potential
(=diplole layer)
LVHAR = .TRUE. Write Hartree part of local potential to LOCPOT
#LVTOT = .TRUE. Write total local potential to LOCPOT

N.B.: We want to use the optimized structure of COonNi111_rel:

In principle this would mean copying the CONTCAR of COonNi111_rel to
POSCAR in the directory where you want to run COonNi111_LDOS.
This, however, has already been taken care of, and the POSCAR file in
COonNi111_LDOS is already the correct one.
 CO@Ni(111) LDOS&Workfunction
(ex.: COonNi111_LDOS)

• lm-decomposed DOS helps

analyse the bonding

• CO: 5𝜎, 1𝜋, 2𝜋 ∗

• From comparison with

substrate LDOS:
- hybridisation with Ni-𝑑3𝑧 2 −𝑟 2
- no interaction with 𝑑𝑥𝑦
⟹ from symmetry

Visualisation: use p4vasp as explained for example Ni100clean_LDOS.

 CO@Ni(111) LDOS&Workfunction
(ex.: COonNi111_LDOS)
Due to dipole corrections
• Fermi-level 𝜀𝐹 = 1.68 eV
(in OUTCAR)

• vacuum-potential at 8.24/6.77 eV:

ΦCO = 8.24 − 1.68 = 6.56 eV

Φclean = 6.77 − 1.68 = 5.09 eV

From Ni111clean_400eV, however:

Φclean = 5.22 eV

This discrepancy is due to a

too small vacuum!

Visualisation: use p4vasp as explained for example Ni111clean_400eV.

 CO@Ni(111) vibrational frequencies
(ex.: COonNi111_freq)
INCAR:
SYSTEM = CO on Ni (111) frequencies
ENMAX = 400 cutoff energy 400 eV
ISMEAR = 2 2nd order Methfessel-Paxton smearing
SIGMA = 0.2 smearing width 𝜎 = 0.2 eV
ALGO = Fast use RMM-DIIS for electronic optimization
EDIFF = 1E-6 tight convergence needed to compute
vibrational properties
MAXMIX = 60 (Default: MAXMIX=45) Store more mixing steps,
save time but costs memory
NSW = 100

IBRION = 5 Compute vibrational spectrum using finite

differences (displacements: 0.015 Å)

NFREE = 2 Use two displacements per degree-of-freedom

(i.e., a 2nd order finite difference stencil)
 CO@Ni(111) vibrational frequencies
(ex.: COonNi111_freq)
POSCAR: Ni - (111) + CO on-top
3.53000000000000
0.7071067800000000 0.0000000000000000 0.0000000000000000
-0.3535533900000000 0.6123724000000000 0.0000000000000000
0.0000000000000000 0.0000000000000000 5.1961523999999999
Ni C O
5 1 1
Selective dynamics
Direct
0.0000000000000000 0.0000000000000000 0.0000000000000000 F F F
0.3333333300000021 0.6666666699999979 0.1111111100000031 F F F
0.6666666699999979 0.3333333300000021 0.2222222199999990 F F F
-0.0000000000000000 0.0000000000000000 0.3326227833039623 F F F
0.3333333300000021 0.6666666699999979 0.4445699380869117 F F F
0.3333333300000021 0.6666666699999979 0.5403264650180125 F F T
0.3333333300000021 0.6666666699999979 0.6032949698060487 F F T

0.00000000E+00 0.00000000E+00 0.00000000E+00 0.00000000E+00

0.00000000E+00 0.00000000E+00 0.00000000E+00 0.00000000E+00
0.00000000E+00 0.00000000E+00 0.00000000E+00 0.00000000E+00
0.00000000E+00 0.00000000E+00 0.00000000E+00 0.00000000E+00
0.00000000E+00 0.00000000E+00 0.00000000E+00 0.00000000E+00
0.00000000E+00 0.00000000E+00 0.00000000E+00 0.00000000E+00
0.00000000E+00 0.00000000E+00 0.00000000E+00 0.00000000E+00

• Frequencies only for CO molecule, and only in the z-direction

(z and (x,y) are independent).

Use the POSCAR file in COonNi111_freq, or copy COonNi111_rel/CONTCAR to POSCAR

 CO@Ni(111) vibrational frequencies
(ex.: COonNi111_freq)
Additional output in the OUTCAR file for frequency calculations via finite differences,
e.g.:
Finite differences progress:
Degree of freedom: 1/ 2
Displacement: 1/ 2
Total: 1/ 4

• After the initial calculation for the equilibrium geometry, NFREE displacements
(±POTIM) are performed for each degree of freedom; from the forces that
are induced by these displacements the dynamical matrix is set up and diagonalized.

• At the end of the OUTCAR file the

- forces,
- the dynamical matrix, and finally
- the eigenfrequencies and
- eigenvectors (first normalized and then mass-weighted)
are listed.
 CO@Ni(111) vibrational frequencies
(ex.: COonNi111_freq)

Eigenvectors and eigenvalues of the dynamical matrix

----------------------------------------------------
1 f = 63.847137 THz 401.163391 2PiTHz 2129.711174 cm -1 264.050627 meV
X Y Z dx dy dz
0.000000 0.000000 0.000000 0 0 0
-0.000000 1.441116 2.038046 0 0 0
1.248043 0.720558 4.076093 0 0 0
0.000000 0.000000 6.101106 0 0 0
-0.000000 1.441116 8.154488 0 0 0
-0.000000 1.441116 9.910894 0 0 -0.781357
-0.000000 1.441116 11.065888 0 0 0.624084
CO stretch

2 f = 12.457302 THz 78.271540 2PiTHz 415.530869 cm -1 51.519280 meV

X Y Z dx dy dz
0.000000 0.000000 0.000000 0 0 0
-0.000000 1.441116 2.038046 0 0 0
1.248043 0.720558 4.076093 0 0 0
0.000000 0.000000 6.101106 0 0 0
-0.000000 1.441116 8.154488 0 0 0
-0.000000 1.441116 9.910894 0 0 -0.624084
-0.000000 1.441116 11.065888 0 0 -0.781357
CO-metal
 STM (ex.: Graphite_STM and Graphene_STM)
Chose a plane roughly 1 Å
above the surface atoms
Step 1.) select “STM”

Replicate the unit cell a few

times in both directions
Show the structure as well
STM (ex.: Graphite_STM and Graphene_STM)