====== Calculation of surface energies of Au ======
TO USE THE FUNCTION LIBRARY (VERSION UP TO DATE) IN THE INTERACTIVE SHELL:
you@eulerX ~$ module load courses mmm vmd
you@eulerX ~$ mmm-init
In this exercise, we will build three small slabs based on the following figures; the cell is an **orthorombic** cell with repetition along z.
{{ :exercises:2016_ethz_mmm:fcc_surf_2.png?1000 |}}
Figure out what is (in nearest neighbor units) the length of the three vectors Lx, Ly, Lz for the three high-symmetry surfaces.
{{ screen_shot_2014-03-07_at_8.47.53_am.png?direct |}}
Take the above table as reference for the orders of magnitude of surface energies.
* Download all the necessary files for computing the surface energies of all 3 high symmetry faces of gold from from the wiki: {{e4.2.zip|e4.2.zip}} (**all inputs are commented**) in your home directory and unzip it:
you@eulerX ~$ wget http://www.cp2k.org/_media/exercises:2017_ethz_mmm:e4.2.zip
you@eulerX ~$ unzip exercises:2017_ethz_mmm:e4.2.zip
you@eulerX ~$ cd exercise_4.2
If you cannot download with **wget**, you can copy the zip file directly from my directory
you@eulerX ~$ cp /cluster/home/danielep/2017/e4.2.zip .
you@eulerX ~$ unzip e4.2.zip
you@eulerX ~$ cd exercise_4.2
In principle all reconstructed phases should be with a lower surface energy, but this EAM potential fails for some of them.
The input files are:
- 100_bulk.inp for determining the energy per atom in the bulk
- 100_unr.inp for the bulk termination (100)
- 100_rec.inp for the hexagonally reconstructed (100)
- 110_unr.inp for the (110) bulk termination
- 110_rec.inp for the (2x1) reconstruction
- 111.inp for the unreconstructed (111).
Before, however, to be able to run the cp2k program, we need to build the **xyz** files for all cases.
First, we build the slabs, using this chain of commands (**First assignment**).
Use for the following:
- (100) surface: use (in first neighbor units) LX LY LZ = ???????? ???????? 1.4142135 ; NX NY NZ = 4 5 5
- (110) surface: use (in first neighbor units) LX LY LZ = ???????? ???????? 1.0000000 ; NX NY NZ = 6 4 5
- (111) surface: use (in first neighbor units) LX LY LZ = ???????? ???????? 2.4494897 ; NX NY NZ = 4 3 3
- Find the correct values of lx ly for the three cases.
- use ** D = 2.885 ** (correct value for gold)
Example:
you@eulerxx $ m_lattice LX LY LZ NX NY NZ < 100.unit | m_xyzrescale D | m_xyzrefold 1 1 1 | m_xyzcenter 1 1 1 | m_xyzsort > 100_unr.xyz
you@eulerxx $ m_lattice LX LY LZ NX NY NZ < 110.unit | m_xyzrescale D | m_xyzrefold 1 1 1 | m_xyzcenter 1 1 1 | m_xyzsort > 110_unr.xyz
you@eulerxx $ m_lattice LX LY LZ NX NY NZ < 111.unit | m_xyzrescale D | m_xyzrefold 1 1 1 | m_xyzcenter 1 1 1 | m_xyzsort > 111_unr.xyz
Substitute the LX LY LZ NX NY NZ and D by their appropriate values. As a result of the command execution you will obtain several slab structures. Open them with a text editor and get the unit cell parameters from the second line of each XYZ file. Use this parameters in the ABC section of the *inp files. Do not forget to **ADD SOME VACUUM IN THE Z DIRECTION**. Check with vmd that the cell is OK.
For the bulk case, simply copy 100_unr.xyz into 100_bulk.xyz
Run the cp2k code as
you@eulerxx $ bsub cp2k.popt -i 100_unr.inp -o 100_unr.out
you@eulerxx $ bsub cp2k.popt -i 110_unr.inp -o 110_unr.out
you@eulerxx $ bsub cp2k.popt -i 111_unr.inp -o 111_unr.out
you@eulerxx $ bsub cp2k.popt -i 100_bulk.inp -o 100_bulk.out
The compute_surf* scripts give an idea how to compute the surface energies.
In the next step, we work with 2 reconstructed surfaces given in the 100_rec.xyz and 110_rec.xyz files. For this calculations you will need to use the same unit cell parameters as for the corresponding unreconstracted surface structures that you have built previously.
you@eulerxx $ bsub cp2k.popt -i 100_rec.inp -o 100_rec.out
you@eulerxx $ bsub cp2k.popt -i 110_rec.inp -o 110_rec.out
Assignments:
- Mark surface energies and relaxations of the first 2-3 top layers (use vmd or editor)
- Try to color the surfaces in vmd in a way that makes the height evident (ask the assistant) and produce snapshots (render)
- Comment on the (2x1) reconstruction: what do you see?
- Copy the 110_unr.xyz into a new file 110_my_rec.xyz
- Edit the file and remove one line each two on the two surfaces
- Check that the result is the same as in 110_rec.xyz
- Plot the same slabs, but adding periodic images in vmd: remember pbc set { a b c 90 90 90 }.
Hint: you can save a configuration file in vmd and reuse it for a new sample! Ask the assistant!