Homework 3
Due: since the calculation takes longer than before, we’ll make it due on 02/13/2017, Monday
Email to:
Introduction
In this assignment, you will be using a commercial periodic code package called VASP (Vienna Ab initio Simulation Package). It is already installed in WAG machine. VASP computes an approximate solution to the many-body Schrödinger equation, either within density functional theory (DFT), solving the KohnSham equations, or within the Hartree-Fock (HF) approximation, solving the Roothaan equations. Hybrid functionals that mix the Hartree-Fock approach with density functional theory are implemented as well. Furthermore, Green's functions methods (GW quasiparticles, and ACFDT-RPA) and many-body perturbation theory (2nd-order Møller-Plesset) are available in VASP. If you encounter any problems, the online user manual is here:
Today, we are going to use periodic code to study surface chemistry, to be specific, surface catalysis. It is an important field, as the Nobel Prize in chemistry for 2007 is awarded to the study of fundamental molecular processes at the gas-solid interface by Gerhard Ertl. We are going to walk through one important reaction that Ertl studied: CO adsorption on Pt (110) surface. CO/Pt-surface is a system not only studied in lab, but also useful in industrial settings. Every car produced today has a catalyst system that converts carbon monoxide and hydrocarbons to carbon dioxide in the exhaust gases.
Technical Stuff
Every vasp calculations has to be in an individual directory. I suggest you make a new directory called HW3 first, and make several directories for CO, Pt surface, CO on Pt, and etc.
mkdir HW3
You need to prepare 5 input files for every vasp calculation, namely INCAR (specify which properties you want to calculate), POSCAR (input molecular/surface geometry), POTCAR (pseudopotential), KPOINTS (kspace points, useful for period crystal calculation), and pbs (the job-submission file). After the calculation is finished, you will notice several new files popping out in the same directory, such as OUTCAR, LOG and etc…
To use VASP on wag machine, add the following lines to your .cshrc file
> vi .cshrc
setenv PATH /project/source/VASP/vasp.5.3.5/bin:$PATH
setenv PATH /net/hulk/home3/jqian/scripts/vtstscripts-908:$PATH
use :wq to save and quit.
> source .cshrc
Part 1. CO molecular energy
In order to study CO adsorption on Pt surface, we need to know the absolute energy of CO first. Make a
new directory called CO
mkdir CO
Go to that directory CO
> cd CO
Now you need to create POSCAR, POTCAR, KPOINTS, INCAR and pbs. All the files are case sensitive.
For our CO molecule in a box, I’ll give you input POSCAR file but I’ll ask you later in write-up what these lines mean. Use
> vi POSCAR
type in the following:
Lines following # are just comments; you don’t need to copy it in order to run the code. After finishing,
use :wq to save and exit. Construct INCAR using vi
> vi INCAR
> vi KPOINTS
To generate POTCAR, type in:
>genPOTCAR.py PBE C O
(Usage of this python script: genPOTCAR.py [LDA/PBE] [element]...
Note the order of elements should be the same as that in POSCAR)
If you can’t use script genPOTCAR.py, copy it from Jin
cp /net/hulk/home3/jqian/HW_VASP/genPOTCAR.py .
To generate pbs, type in:
>genVASPpbs.py 1 4 24:00:00 CO
(Usage of this python script: genVASPpbs.py [# of nodes] [# of cpus per node] [walltime] [jobname]
[optional:gmo/soc, which we can ignore for now]
gmo is the speed-up gamma point only version for very large supercell
soc is the version with spin orbit coupling, mind that this consumes a lot of memory)
If you can’t use script genPOTCAR.py, copy it from Jin
cp /net/hulk/home3/jqian/HW_VASP/genVASPpbs.py .
Submit your job by
qsubpbs
It should only take minutes to finish, check energy in OUTCAR file, use the command:
grep energy OUTCAR, it should look like this:
Write up part 1:
1. Report the energy of CO molecule from OUTCAR
2. In POSCAR file, we specified that the unit cell size are 15Å x 15Å x15Å, what happens if we
change it to a small cell 3Å x 3Å x 3Å? Report the new energy (I suggest you make a new
directory for this) and why is it higher. (Hint: It’s a periodic system, meaning periodic in x, y, z
directions, consider van der waals curve.) Which energy is a better representation of the real
energy?
3. In INCAR, what does the ISMEAR tag mean? What should we use if our system is metal? (Hint:
check out user manual)
4. CO is a polar molecule, so a strict calculation should have dipole correction IDIPOL = 3. What
does this line mean? Rerun the calculation and report the new energy.
Part 2. Pt(110) slab energy
Similarily, we should prepare POSCAR, POTCAR, INCAR, KPOINTS and pbs for this calculation.
cp/net/hulk/home3/jqian/HW_VASP/Pt/POSCAR .
Here, selective dynamics means only relax some of the coordinates. In this case, we freeze (F) the
bottom Pt layers and relax (T) the top ones.
cp /net/hulk/home3/jqian/HW_VASP/Pt/INCAR .
cp /net/hulk/home3/jqian/HW_VASP/Pt/KPOINTS .
> genPOTCAR.py PBE Pt
> genVASPpbs.py 1 4 24:00:00 Pt110
qsubpbs
This calculation might take a day or two to finish, so make sure no error after ~5 minutes after
submission.
Write up part 2:
1. Pt is an FCC metal, scratch what 110 surface of fcc cubic look like.
2. Download and install a software called VESTA (visualization software for VASP), open POSCAR in
VESTA, and compare with what you have drawn.
3. Report the energy of this system.
4. After the run, a new file called CONTCAR (a file that contains the information of final geometry)
should pop out. Give a screenshot of CONTCAR, and compare with POSCAR on how molecules
moved. (If you prefer, open CONTCAR in VESTA and compare with POSCAR in VESTA).
Part 3. CO on Pt110
Consider CO adsorbs on top position of Pt110, such as the following (top view and side view generated
from VESTA, red is O, brown is C, and silver is Pt):
Construct your own input files for this run.
Hint: use the optimized CONTCAR file from previous Pt110 slab calculation and add CO on top position.
Hint: Pt-C distance is 1.84 Å, and C-O distance is 1.16 Å. Note that in CONTCAR, the coordinates are
direct coordinates instead of Cartesian coordinates, meaning you have to divide the distance by the
corresponding lattice parameter (z direction: 50 Å). Always visualize geometry(POSCAR) in VESTA before
submitting the job.
Write up part 3:
1. Report the energy of the system.
2. Eadsorption= Ept_co - Eco - EPt, report the adsorption energy.
3. What are other possible CO adsorption sites?
4. EXTRA POINTS: calculate CO adsorption energy for other adsorption sites.