ChemWiki - User contributions [en]

Running a HP-DFT calculation with CP2K

2023-10-27T15:45:55Z

2023-06-28T20:13:49Z

Mb620:

<div style="border:0px solid #90C0FF; background:#ffffff; width:90%; padding:40px">
=Nano Electrochemistry Group=

==Clotilde Cucinotta (PI)==

==Chris Ahart==
[[File:ChrisAhartProfile.png|frameless|bottom|200px|Chris Ahart]]

Research Associate (postdoc) August 2022 - present.

Dynamical modelling of electrochemical systems under applied potential

Chris is working to enable the dynamical modelling of electrochemical systems under applied potential. This is made possible by through a new interface between the DFT code CP2K and the NEGF code SMEAGOL, developed in collaboration with Dr Sergey Chulkov from the University of Lincoln. Once completed, CP2K+SMEAGOL will be used to unravel fundamental electro-catalytic phenomena at electrified interfaces, such as water splitting, corrosion, and energy storage/conversion.
==Frederik==

==Songyuan==

==Margherita Buraschi==
[[File:Margherita Buraschi photo.jpg|frameless|bottom|200px|Margherita Buraschi]]

PhD in Chemistry July 2020-present.

Computational study of electrochemical nanojunctions with open-boundaries description of the electrons.

Margherita is investigating the charge transfer properties of electrochemical nanojunctions (ECNJs) by means of computational studies. ECNJs are three-terminal functional devices of interest for the field of molecular electronics <ref> Huang et al., Chem. Soc. Rev., '''44''', 889-901 (2015) </ref> as they present the same functionality of common electronic units but at the molecular scale. Many factors define the conductance of ECNJs; amongst these factors are the structure of the functional unit, the type of metal as well as the EC transformations occurring in solutions. In order to obtain an atomistic understanding of all these factors and their effects on the overall conductance of the junction, ab initio molecular dynamics under bias and an explicit open‐boundary description of the electrons are performed during this project. To this end, the hair-probes formalism <ref> Horsfield, Boleininger, D'Agosta, Iyer, Thong, Todorov and White, Phys. Rev. B, '''94''', 075118 (2016) </ref><ref>Zauchner, Horsfield and Todorov, Physical Review B, '''97''', 045116 (2018) </ref>, an open‐boundary formalism suitable to describe multi-terminal EC systems, is being interfaced within the DFT scheme implemented in the CP2K computational package <ref>Kühne et al., The Journal of Chemical Physics, '''152'''(19), 194103 (2020) </ref>.
==Rashid Al-Heidous==
[[File:Rashid Personal Photo.jpeg|frameless|bottom|200px|Rashid Al-Heidous]]

==Zehui Duan==
[[File:Zehui Duan photo.jpg|frameless|bottom|200px|Zehui Duan]]

MRes Oct 2020-Sept2021.
Using density functional theory (DFT) as implemented in CP2K code, Zehui is investigating the thermodynamics of the CO<sub>2</sub> reduction reaction (CO2RR), to guide the rational design for dual-atom catalysts (DAC) with both high efficiency and good selectivity.

==Lorraine Wang==
[[File:Lorraine Wang photo.jpg|frameless|bottom|200px|Lorraine Wang]]

MRes Oct 2020-Sept2021.
In her MRes project, Lorraine is studying dual-metal-atom catalysts (DAC) by means of density functional theory (DFT) as implemented in CP2K code. Dual-metal-atom catalysts with synergistic effect <ref>X. Zhu, J. Yan, M. Gu, T. Liu, Y. Dai, Y. Gu and Y. Li, J. Phys. Chem. Lett., '''10''', [tel:7760–7766 7760–7766] (2019) </ref> between the two metal atoms can enhance the catalytic oxygen reduction reaction (ORR). They usually consist of a substrate (doped graphene layers) and anchored metal atoms. Using DFT simulations implemented within the CP2K code can help to build a fundamental understanding of the structure. By studying the thermodynamics of various patterns, we can propose a thermodynamically favourable pathway of DAC synthesis. In the future, the project can further extend to study catalysts phase stability, catalytic performance and kinetics with more complexed models.

==Felix==

==Anthony Liu==
[[File:Anthony.png|frameless|bottom|200px|Anthony Liu]]

MRes Student.

Machine learning potentials for solid-liquid interfaces.

Although accurate, DFT and AIMD calculations can sometimes be time-consuming. Anthony is currently working on using machine learning models to model the electric potential across the platinum-water interface at different electrolyte concentrations.

==Chien-Yu Kuo==
[[File:Chien-Yu.png|frameless|bottom|200px|Chien-Yu Kuo]]

MRes Student.

Enabling the modelling of the kinetics of electrochemical transformation under potential control with hairy probes formalism

Chien-Yu is co-working with PHD candidate Margherita Buraschi on potential-controlled Pt-water studies based on hairy-probes density functional theory (HP-DFT). The consecutive and better description about the electric double layer could be provided based on this novel development. Eventually, hoping this study could provide a more resonable simulation result for water splitting studies.

==Yuhong Ye==
[[File:Yuhong.png|frameless|bottom|200px|Yuhong Ye]]

MRes Student.

Rationalising the Effect of Electrical Double Layer Structure on the Oxygen Evolution Reaction.

Yuhong is currently conducting research on the specific characteristics of the double layer during oxygen evolution reactions. He is utilizing the DFT code CP2K to investigate the interface between the catalysts and the electrolytes.

=Computational Materials Science Group (CMSG)=

CMSG is a group of scientiests under Prof. Nicholas Harrison's supervison at Imperial College London.

==Nicholas Harrison (PI)==

==Paolo Restuccia==

==Huanyu Zhou==
[[File:Huanyu profile.jpg|frameless|bottom|200px|Huanyu Zhou]]

PhD in Chemistry 2021-present; MRes Nanomaterials 2020-2021

Huanyu's PhD research project is about the multi-scale modelling of thermodynamics of molecular crystals using ab initio methods. It involves accurate polymorphism prediction and modelling the solvation effects at solid-liquid interfaces of pharmaceutical crystals, which is a joint endeavour with Prof. Jerry Heng's group at the Department of Chemical Engineering.

During his master project co-supervised by Dr. Giuseppe Mallia, Huanyu investigated the monovacancy defect on compressed graphene by DFT. The correlation between magnetic moment and local geometry of monovacacies was identified, which lead to novel electronic and spintronic properties.

Huanyu is also one of the main authors of [https://crystal-code-tools.github.io/CRYSTALpytools/ CRYSTALpytools], the object-oriented python package for CRYSTAL, co-developed by CMSG Imperial and the Theoretical Chemistry Group (TEO) UniTo.

==Fei Gao==
[[File:Fei.jpeg|frameless|bottom|200px|Lorraine Wang]]

PhD in Chemistry 2021-present; MRes IMSE 2020-2021
</div>

File:Yuhong.png

2023-06-28T20:10:43Z

Mb620:

File:Chien-Yu.png

2023-06-28T20:09:41Z

Mb620:

Mb620:

Before starting a calculation, it is good practice to perform a series of convergence test to find the most appropriate set of parameters for the system under investigation; the objective is to find best input setting to obtain a high enough accuracy while, at the same time, using the computational resources efficiently.
This section contains a step-by-step tutorial on how to optimize these parameters for a metallic surface; platinum (Pt) is used as an example. To systematically find the best set of parameters, it is necessary to perform a series of single point energy calculations. To achieve this, it is much easier to use a set of scripts which can automate the process.
First, it is useful to write a template input file (''template.inp'') which has the same structure of a traditional CP2K input; in this file, however, the relevant parameters are replaced by “''markers''” (in this example: ''LT_basis_set'', ''LT_cutoff'', ''LT_rel_cutoff'', ''Kx'', ''Ky'', ''Kz''). The automated scripts will search for these markers and replace them with a keyword and/or value from a list previously established. The scripts will then copy the file to a new CP2K input file and move it to a specifically created directory.
It is important to remember that each parameter need to be optimized singularly; together, they then give the most suitable set of parameters for the system in exam.
All the scripts and files described in this page can be found in the group's gitlab repository [https://gitlab.doc.ic.ac.uk/rgc/metal-surfaces].
Additionally, this page contains workflows for extrapolating relevant parameters for a metallic system, such as the optimized lattice parameter, electronic structure and work function. Platinum is used as example for these sections as well.

=Evaluating the numerical accuracy=
The objective of performing convergence tests, as mentioned, is to find the best combination of accuracy and efficiency; using parameters which are unnecessarily strict can result in more expensive calculations without any significant improvement in the accuracy.
To obtain a set of consistent parameters, the error associated to each of them needs be of the same order of magnitude as the others; generally, the error associated to the choice of the basis set is the one giving the biggest balance on the whole accuracy and could be taken as benchmark value.
To understand the concept of error associated to a parameter, let us consider the following: a convergence test consists in a series of calculation where a parameter is varied over a range of values. The energy of such a set of calculations is considered to have converged when, upon variation of the parameter, the energy difference between calculations is negligible. The error is defined as the difference between the energy obtained with a specific value of the parameter and the converged energy. The fully converged energy is generally the one obtained from a calculation performed using the higher level of theory (e.g. more expanded basis set) or a more accurate parameter (e.g. larger cutoff, finer kpoint grid, etc.).

==Convergence Tests==
Note: To use the scripts available on GitLab, remember to modify ''#path_to_basis_set'' ad ''#path_to_pseudopotential'' in ''template.inp''; these are the paths to the data files with the basis sets and the pseudopotentials.
===Basis set===
#Running the set of calculations: to create the input files and run the calculations type <code> ./Basis_set_run.sh </code>. Given a list of basis sets (e.g. SZV-GTH-LDA-q18, DZV-GTH-LDA-q18, TZV-GTH-LDA-q18...), the script generates, for each of them, a directory containing a CP2K input and the pbs script. The input files are created by substituting the marker ''LT_basis_set'' in ''template.inp'' with the one of the basis sets in the list. The other parameters indicated by a marker are also set to a value provided in the script. The script then submits the calculations with each basis set from their respective directories.
#Analysing the results: to analyse the results of this set of calculations, type <code> ./Basis_set_analyse.sh </code>. This script extracts the total energy from the CP2K output file in each directory and automatically converts it from Hartree to electronvolts. The values of energy obtained for each basis set are then printed into a new file, ''basis_set_TotEn.dat'', which can be plotted to observe the variation of energy as a function of the basis set.
===CUTOFF and REL_CUTOFF===
#Running the set of calculations: to create the input files and run the calculations type <code>./Cutoff_run.sh </code>. Given a range of cutoff values (e.g 50, 100, 150...), the script generates, for each of them, a directory (e.g. ''cutoff_50'', ''cutoff_100'',...) containing a CP2K input and the pbs script. The input files are created by substituting the marker ''LT_cutoff'' in ''template.inp'', setting it to one of the values in the given range. The other parameters indicated by a marker are also set to a value provided in the script. The script then runs a calculation for each given value of cutoff.
#Analysing the results: to analyse the results of this set of calculations, type <code> ./Cutoff_analyse.sh </code>. This script extracts the total energy the CP2K output file in each directory and automatically converts it from Hartree to electronvolts. The values of energy obtained for each cutoff are then printed into a new file, ''cutoff_TotEn.dat'', which can be plotted to observe the variation of energy as a function of the CUTOFF.

To perform the convergence test for REL_CUTOFF follow the same steps using <code> ./Rel_cutoff_run.sh </code> and <code> ./Rel_cutoff_analyse.sh </code> instead.
For a more detailed description of the CUTOFF and REL_CUTOFF keywords in CP2K, visit this [https://www.cp2k.org/howto:converging_cutoff| page].

===k-point grid===
There are several schemes to define the k-point grid; nowadays, one of the most widely used is the Monkhorst-Pack grid <ref>Hendrik J. Monkhorst and James D. Pack, Phys. Rev. B, '''13''', 5188</ref>, which is an unbiased method of defining the k-points set for sampling the Brillouin zone:

:     <math> \bold{k}_{n_1,n_2,n_3} = \sum_{i} \frac{2n_i - N_i - 1}{2N_i} \bold{G}_i </math> <ref> S. H. Simon, ''The Oxford Solid State Basics'',
Oxford University Press, 2013 </ref>

where '''G'''<sub>i</sub> are the primitive vectors of the reciprocal lattice and ''n<sub>i</sub> = 1, 2, .., N<sub>i</sub>'' is a set of uniform points in each direction ''i''. To perform a k-points calculation using a Monkhorst-Pack grid in CP2K, it is necessary to add the following section:

&KPOINTS
SCHEME MONKHORST-PACK N<sub>x</sub> N<sub>y</sub> N<sub>z</sub>
&END KPOINTS

where N<sub>x</sub>, N<sub>y</sub> and N<sub>z</sub> (i.e. the number of points in all three dimensions) define an evenly spaced rectangular grid of dimension N<sub>x</sub> N<sub>y</sub> N<sub>z</sub>. More details and other k-points schemes available can be found in the dedicated [https://manual.cp2k.org/trunk/CP2K_INPUT/FORCE_EVAL/DFT/KPOINTS.html page ] of the CP2K manual. To perform the convergence test:
#Running the set of calculations: to create the input files and run the calculations type <code>./Kpoints_run.sh </code>. Given a range of k-point values (e.g 2, 4, 6...), the script generates, for each of them, a directory (e.g. ''kpoints_2'', ''kpoints_4'',...) containing a CP2K input and the pbs script. The input files are created by substituting the markers ''Kx'', ''Ky'' and ''Kz'' in ''template.inp'', setting them to one of the values in the given range. The other parameters indicated by a marker are also set to a value provided in the script. The script then runs a calculation for each k-points grid.
#Analysing the results: to analyse the results of this set of calculations, type <code> ./Kpoints_analyse.sh </code>. This script extracts the value of total energy from the CP2K output file in each directory and automatically converts it from Hartree to electronvolts. The values of energy obtained for each kpoint grid are then printed into a new file, ''kpoints_TotEn.dat'', which can be plotted to observe the variation of energy as a function of the k-points grid size.

=Extrapolation of parameters=
==Equilibrium lattice parameter==
In this section it will be shown how to calculate the equilibrium lattice parameter of a metallic system using the [https://en.wikipedia.org/wiki/Murnaghan_equation_of_state Murnaghan equation of states]. The Murnaghan equation relates the energy of a system to its volume. For this reason, the volume of the unit cell is changed by a few percentage points; the energy of the system for each new value of volume (i.e. lattice parameter) is calculated. The resulting data is fit using the EOS.
To calculate the equilibrium lattice parameter you'll need the following files: ''template_lattice_param.inp'', ''lattice_param_run.sh'', ''lattice_param_analyse.sh'', ''murn.x'' and, of course, ''job.pbs''.
#Setting up and running the set of calculations: to create the input files and run the calculations type <code> ./lattice_param_run.sh </code>. Given a range of values, the script generates a set of inputs with different lattice parameters. Each input is then moved into a newly created directory (e.g. change_*), together with a pbs script. From each directory the calculation with the new value of lattice parameter is launched.
#Analysing the results: to analyse the results of this set of calculations you need to proceed in two steps:
## type <code> ./lattice_param_analyse.sh </code>. This script collects all the values of lattice parameters and energy from each cp2k output file and automatically generates two new files: ''energies.dat'' and ''murn.inp''. ''energies.dat'' contains the values of lattice parameter (Å), volume (Å<sup>3</sup>) and energy (eV) obtained for each percentage change. ''murn.inp'' contains the data necessary to fit the values of energy vs. volume using the [https://en.wikipedia.org/wiki/Murnaghan_equation_of_state Murnaghan equation of states].
## To find the equilibrium lattice parameter type <code> ./murn.x < murn.inp > murn.out </code>. This will produce the file ''murn.out'' which contains the value of equilibrium lattice parameter. You can obtain it by typing <code> grep -A3 "alat=" murn.out </code>.

==Electronic structure==
===DOS===

===Band Structure===
CP2K offers the possibility to calculate the band structure of solids; for a detailed description of a CP2K input to get the band structure visit this [https://www.cp2k.org/exercises:2016_uzh_cmest:band_structure_calculation| page]. This input produces an additional output file named ''[name_of_your_job].bs''. In this example, we will still refer to a platinum slab; in this case we will refer to the band structure file as ''platinum.bs''. To convert ''platinum.bs'' to a file which can be plotted you can use the script ''bandstructure.out'', which can be found in the GitLab repository. Before running the script, you need to know the Fermi energy of the system, which can be obtained as explained above. Once obtained the Fermi energy, you can run the script as follow:

./bandstructure.out

While running, the script will ask for the names of the input and output files and for the Fermi energy; in the case of the platinum slab, for example, it will go as follows:

Enter input file name:
> platinum.bs
Enter output file name:
> bandstructure.dat
Enter Fermi Energy (Hartree):
> 0.5619

The script rearranges the energy values in ''platinum.bs'' in a way that can be plotted and automatically subtract the Fermi energy.

==Work Function==
The [https://en.wikipedia.org/wiki/Work_function#:~:text=From%20Wikipedia,%20the%20free%20encyclopedia%20In%20solid-state%20physics,,in%20the%20vacuum%20immediately%20outside%20the%20solid%20surface.| work function] of a surface, ''W'', is the minimum energy necessary to remove an electron from a solid to a point in the vacuum immediately outside the solid surface. Mathematically it can be defined as follows:
:<math>W = -e\phi - E_{\rm F} </math>
where ''-e'' is the charge of an [[electron]], ''ϕ'' is the [[electrostatic potential]] in the vacuum nearby the surface, and ''E<sub>F</sub>'' is the [[Fermi level]] ([[electrochemical potential]] of electrons) inside the material. The term ''-eϕ'' is the energy of an electron at rest in the vacuum nearby the surface.

To calculate the work function using CP2K, first you need to add, in the SCF section of your input file, the following [https://manual.cp2k.org/trunk/CP2K_INPUT/FORCE_EVAL/DFT/PRINT/V_HARTREE_CUBE.html| section]:

&PRINT
&V_HARTREE_CUBE
...
&END V_HARTREE_CUBE
&END PRINT

This will print an additional output file: a cube file containing the values of electrostatic potential generated by the total density. Now run the script ''aver.out'', which you may find in the GitLab repository. To do this you you will need to:
[[File:POT 4layers noGhost.png|left|thumb|200px| Potential energy of a 4 layer Pt slab.]]
<ol>
<li> Create the files ''input'' and ''cube'' from the cube file generated by CP2K. Check ''input-commented'' to understand how to write them. Remember to have both this files in the same directory as ''aver.out''; </li>
<li> Run the script by typing: <code> ./aver.out < input </code>. ''aver.out'' converts the data stored in ''cube'' into six new files which can be plotted. These files are: ''aver_X.dat'', ''aver_Y.dat'', ''aver_Z.dat'', which contain the values of electrostatic potential in the X, Y, and Z direction and ''avermacro_X.dat'', ''avermacro_Y.dat'', ''avermacro_Z.dat'', which contain the macroscopic average of the potential in the X, Y, Z directions. If your surface is perpendicular to the Z axis, the data along Z will be the ones you are interested in for the calculation of the work function; </li>
<li> From the value of Fermi energy and the value of electrostatic potential energy in the vacuum region (flat region in the plot) you can now calculate the work function. The Fermi energy can be obtained from the CP2K output file (e.g ''cp2k.out'') by typing: <code> grep 'Fermi energy:' cp2k.out </code>.

</li>
</ol>

{{reflist}}

Nano Electrochemistry Group

2021-09-02T08:56:20Z

Mb620:

Optimization of metallic surfaces parameters

2021-08-27T15:24:47Z

Mb620:

<div style="padding: 10px; font-family: Arial; font-size: 95%;">

Before starting a calculation, it is good practice to perform a series of convergence test to find the most appropriate set of parameters for the system under investigation; the objective is to find best input setting to obtain a high enough accuracy while, at the same time, using the computational resources efficiently.
This section contains a step-by-step tutorial on how to optimize these parameters for a metallic surface; platinum (Pt) is used as an example. To systematically find the best set of parameters, it is necessary to perform a series of single point energy calculations. To achieve this, it is much easier to use a set of scripts which can automate the process.
First, it is useful to write a template input file (''template.inp'') which has the same structure of a traditional CP2K input; in this file, however, the relevant parameters are replaced by “''markers''” (in this example: ''LT_basis_set'', ''LT_cutoff'', ''LT_rel_cutoff'', ''Kx'', ''Ky'', ''Kz''). The automated scripts will search for these markers and replace them with a keyword and/or value from a list previously established. The scripts will then copy the file to a new CP2K input file and move it to a specifically created directory.
It is important to remember that each parameter need to be optimized singularly; together, they then give the most suitable set of parameters for the system in exam.
All the scripts and files described in this page can be found in the group's gitlab repository [https://gitlab.doc.ic.ac.uk/rgc/metal-surfaces].
Additionally, this page contains workflows for extrapolating relevant parameters for a metallic system, such as the optimized lattice parameter, electronic structure and work function. Platinum is used as example for these sections as well.

=Evaluating the numerical accuracy=
The objective of performing convergence tests, as mentioned, is to find the best combination of accuracy and efficiency; using parameters which are unnecessarily strict can result in more expensive calculations without any significant improvement in the accuracy.
To obtain a set of consistent parameters, the error associated to each of them needs be of the same order of magnitude as the others; generally, the error associated to the choice of the basis set is the one giving the biggest balance on the whole accuracy and could be taken as benchmark value.
To understand the concept of error associated to a parameter, let us consider the following: a convergence test consists in a series of calculation where a parameter is varied over a range of values. The energy of such a set of calculations is considered to have converged when, upon variation of the parameter, the energy difference between calculations is negligible. The error is defined as the difference between the energy obtained with a specific value of the parameter and the converged energy. The fully converged energy is generally the one obtained from a calculation performed using the higher level of theory (e.g. more expanded basis set) or a more accurate parameter (e.g. larger cutoff, finer kpoint grid, etc.).

==Convergence Tests==
Note: To use the scripts available on GitLab, remember to modify ''#path_to_basis_set'' ad ''#path_to_pseudopotential'' in ''template.inp''; these are the paths to the data files with the basis sets and the pseudopotentials.
===Basis set===
#Running the set of calculations: to create the input files and run the calculations type <code> ./Basis_set_run.sh </code>. Given a list of basis sets (e.g. SZV-GTH-LDA-q18, DZV-GTH-LDA-q18, TZV-GTH-LDA-q18...), the script generates, for each of them, a directory containing a CP2K input and the pbs script. The input files are created by substituting the marker ''LT_basis_set'' in ''template.inp'' with the one of the basis sets in the list. The other parameters indicated by a marker are also set to a value provided in the script. The script then submits the calculations with each basis set from their respective directories.
#Analysing the results: to analyse the results of this set of calculations, type <code> ./Basis_set_analyse.sh </code>. This script extracts the total energy from the CP2K output file in each directory and automatically converts it from Hartree to electronvolts. The values of energy obtained for each basis set are then printed into a new file, ''basis_set_TotEn.dat'', which can be plotted to observe the variation of energy as a function of the basis set.
===CUTOFF and REL_CUTOFF===
#Running the set of calculations: to create the input files and run the calculations type <code>./Cutoff_run.sh </code>. Given a range of cutoff values (e.g 50, 100, 150...), the script generates, for each of them, a directory (e.g. ''cutoff_50'', ''cutoff_100'',...) containing a CP2K input and the pbs script. The input files are created by substituting the marker ''LT_cutoff'' in ''template.inp'', setting it to one of the values in the given range. The other parameters indicated by a marker are also set to a value provided in the script. The script then runs a calculation for each given value of cutoff.
#Analysing the results: to analyse the results of this set of calculations, type <code> ./Cutoff_analyse.sh </code>. This script extracts the total energy the CP2K output file in each directory and automatically converts it from Hartree to electronvolts. The values of energy obtained for each cutoff are then printed into a new file, ''cutoff_TotEn.dat'', which can be plotted to observe the variation of energy as a function of the CUTOFF.

To perform the convergence test for REL_CUTOFF follow the same steps using <code> ./Rel_cutoff_run.sh </code> and <code> ./Rel_cutoff_analyse.sh </code> instead.
For a more detailed description of the CUTOFF and REL_CUTOFF keywords in CP2K, visit this [https://www.cp2k.org/howto:converging_cutoff| page].

===k-point grid===
There are several schemes to define the k-point grid; nowadays, one of the most widely used is the Monkhorst-Pack grid <ref>Hendrik J. Monkhorst and James D. Pack, Phys. Rev. B, '''13''', 5188</ref>, which is an unbiased method of defining the k-points set for sampling the Brillouin zone:

:     <math> \bold{k}_{n_1,n_2,n_3} = \sum_{i} \frac{2n_i - N_i - 1}{2N_i} \bold{G}_i </math> <ref> S. H. Simon, ''The Oxford Solid State Basics'',
Oxford University Press, 2013 </ref>

where '''G'''<sub>i</sub> are the primitive vectors of the reciprocal lattice and ''n<sub>i</sub> = 1, 2, .., N<sub>i</sub>'' is a set of uniform points in each direction ''i''. To perform a k-points calculation using a Monkhorst-Pack grid in CP2K, it is necessary to add the following section:

&KPOINTS
SCHEME MONKHORST-PACK N<sub>x</sub> N<sub>y</sub> N<sub>z</sub>
&END KPOINTS

where N<sub>x</sub>, N<sub>y</sub> and N<sub>z</sub> (i.e. the number of points in all three dimensions) define an evenly spaced rectangular grid of dimension N<sub>x</sub> N<sub>y</sub> N<sub>z</sub>. More details and other k-points schemes available can be found in the dedicated [https://manual.cp2k.org/trunk/CP2K_INPUT/FORCE_EVAL/DFT/KPOINTS.html page ] of the CP2K manual. To perform the convergence test:
#Running the set of calculations: to create the input files and run the calculations type <code>./Kpoints_run.sh </code>. Given a range of k-point values (e.g 2, 4, 6...), the script generates, for each of them, a directory (e.g. ''kpoints_2'', ''kpoints_4'',...) containing a CP2K input and the pbs script. The input files are created by substituting the markers ''Kx'', ''Ky'' and ''Kz'' in ''template.inp'', setting them to one of the values in the given range. The other parameters indicated by a marker are also set to a value provided in the script. The script then runs a calculation for each k-points grid.
#Analysing the results: to analyse the results of this set of calculations, type <code> ./Kpoints_analyse.sh </code>. This script extracts the value of total energy from the CP2K output file in each directory and automatically converts it from Hartree to electronvolts. The values of energy obtained for each kpoint grid are then printed into a new file, ''kpoints_TotEn.dat'', which can be plotted to observe the variation of energy as a function of the k-points grid size.

=Extrapolation of parameters=
==Equilibrium lattice parameter==
To calculate the equilibrium lattice parameter you'll need the following files: ''template_lattice_param.inp'', ''lattice_param_run.sh'', ''lattice_param_analyse.sh'', ''murn.x'' and, of course, ''job.pbs''.
#Setting up and running the set of calculations: to create the input files and run the calculations type <code> ./lattice_param_run.sh </code>. Given a range of values (which you can modify in line **) the script generates a set of inputs of inputs with different lattice parameters; then moves each of them into a newly created directory (e.g. change_*), together with a pbs script.

==Electronic structure==
===DOS===

===Band Structure===
CP2K offers the possibility to calculate the band structure of solids; for a detailed description of a CP2K input to get the band structure visit this [https://www.cp2k.org/exercises:2016_uzh_cmest:band_structure_calculation| page]. This input produces an additional output file named ''[name_of_your_job].bs''. In this example, we will still refer to a platinum slab; in this case we will refer to the band structure file as ''platinum.bs''. To convert ''platinum.bs'' to a file which can be plotted you can use the script ''bandstructure.out'', which can be found in the GitLab repository. Before running the script, you need to know the Fermi energy of the system, which can be obtained as explained above. Once obtained the Fermi energy, you can run the script as follow:

./bandstructure.out

While running, the script will ask for the names of the input and output files and for the Fermi energy; in the case of the platinum slab, for example, it will go as follows:

Enter input file name:
> platinum.bs
Enter output file name:
> bandstructure.dat
Enter Fermi Energy (Hartree):
> 0.5619

The script rearranges the energy values in ''platinum.bs'' in a way that can be plotted and automatically subtract the Fermi energy.

==Work Function==
The [https://en.wikipedia.org/wiki/Work_function#:~:text=From%20Wikipedia,%20the%20free%20encyclopedia%20In%20solid-state%20physics,,in%20the%20vacuum%20immediately%20outside%20the%20solid%20surface.| work function] of a surface, ''W'', is the minimum energy necessary to remove an electron from a solid to a point in the vacuum immediately outside the solid surface. Mathematically it can be defined as follows:
:<math>W = -e\phi - E_{\rm F} </math>
where ''-e'' is the charge of an [[electron]], ''ϕ'' is the [[electrostatic potential]] in the vacuum nearby the surface, and ''E<sub>F</sub>'' is the [[Fermi level]] ([[electrochemical potential]] of electrons) inside the material. The term ''-eϕ'' is the energy of an electron at rest in the vacuum nearby the surface.

To calculate the work function using CP2K, first you need to add, in the SCF section of your input file, the following [https://manual.cp2k.org/trunk/CP2K_INPUT/FORCE_EVAL/DFT/PRINT/V_HARTREE_CUBE.html| section]:

&PRINT
&V_HARTREE_CUBE
...
&END V_HARTREE_CUBE
&END PRINT

This will print an additional output file: a cube file containing the values of electrostatic potential generated by the total density. Now run the script ''aver.out'', which you may find in the GitLab repository. To do this you you will need to:
[[File:POT 4layers noGhost.png|left|thumb|200px| Potential energy of a 4 layer Pt slab.]]
<ol>
<li> Create the files ''input'' and ''cube'' from the cube file generated by CP2K. Check ''input-commented'' to understand how to write them. Remember to have both this files in the same directory as ''aver.out''; </li>
<li> Run the script by typing: <code> ./aver.out < input </code>. ''aver.out'' converts the data stored in ''cube'' into six new files which can be plotted. These files are: ''aver_X.dat'', ''aver_Y.dat'', ''aver_Z.dat'', which contain the values of electrostatic potential in the X, Y, and Z direction and ''avermacro_X.dat'', ''avermacro_Y.dat'', ''avermacro_Z.dat'', which contain the macroscopic average of the potential in the X, Y, Z directions. If your surface is perpendicular to the Z axis, the data along Z will be the ones you are interested in for the calculation of the work function; </li>
<li> From the value of Fermi energy and the value of electrostatic potential energy in the vacuum region (flat region in the plot) you can now calculate the work function. The Fermi energy can be obtained from the CP2K output file (e.g ''cp2k.out'') by typing: <code> grep 'Fermi energy:' cp2k.out </code>.

</li>
</ol>

{{reflist}}

2021-07-22T13:00:07Z

Mb620:

File:Zehui Duan photo.jpg

2021-07-22T12:56:49Z

Mb620:

File:Margherita Buraschi photo.jpg

2021-07-22T12:53:04Z

Mb620: