Unitcell relaxiation non MD- Quantum espresso series

Dear friends,

In this blog, I will have posted about unitcell optimization. Let us go to the input file, as usual, we used to keep the calculation as ‘vc-relax’- which states that vary cell parameters. Previously we may found some files like “wfc” that. To avoid such interruption we are going to use the prefix and outdir controls. The outdir directory only meant for temporary file storage. Then etot_conv_thr and forc_conv_thr are the energy and force convergence threshold with respectively. Both are the accuracy controls for our calculation. The force convergence should be less than the energy convergence. Otherwise, our computation cost will drain.

In the system controls we are going to provide the cell dimension using celldm or a . Here we are using the simple cubic Si crystal structure. So I give a = 7  (unoptimized lattice parameter.) but the reference is we known as a = 5.4683 Å.

If you want to allow the atoms to move inside the unitcell you can use the &ions input. Mainly this used in the molecular dynamic case such as BFGS. Here we do not want such confusions. So let it be empty life make easy…!!

In the &cell input we have to implement the freedom of unitcell parameter. I used the unitcell can optimize according to the ibrav. So the degree of freedom is restricted no more many body problems. { cell_dofree = ‘ ibrav’ }

We have to implement the same thing in the movement of the atoms. So better we convert the atomic positions into ( alat ) atomic lattice from angstrom.  The restriction for the movement of the atoms along the axis can be controlled by the last three digits of  0   0   0 in the atomic position.

Note: we are not treating and care of K_points but if you are using 8 8 8 0 0 0 points for scf calculation you must use as 4 4 4 0 0 0 points.

Now we can run the program following commands.

$ mpiexec –np 2 pw.exe < relax.in > relax.out

Here, the input file saved as in the name of relax.in. We can wait for the program has to run.

Suppose if you have WSL (windows system linux) you can see the output file by following command in WSL $ tail –f relax.out .  Now our cell dimension can be seen in the output file as celldm(1) = 10.34723695. 

If you are feel not well the use WSL cmd  as $ grep “celldem(1) =” relax.out . I have used the command for searching the pressure variation $ grep “ P=” relax.out. if the pressure is in negative scale then our unitcell volume is large similarly for small volume of unitcell pressure will be high.

Now let us interpret our results to check the following figure. The calculated values obtained from the relax.out file.

I hope this blog may help you.  Thank you for supporting my blog.....!!🥰

Molecular Dynamics( MD - Optimization ) Parallel Execution in QUANTUM ESPRESSO

Dear Friends, 

        In this session, we are going to see the unitcell optimization using the "Verlet ion dynamics  Parrinello-Rahman algorithm of lagrangian". Before executing groundstate studies we have to optimize the structure first. In this session, I used the same Silicon structure by changing some parameters in the input file.  In the calculation, I have defined as "vc_relax" which stands for structure relaxation. nstep has the control of iteration. we must have to fix the parameter of convergent accuracy in the iteration. and we have chosen MD optimization so we also have to define the mixing_beta this will have the control of volume or XYZ variation. 

we force the silicon ions to move along MD trajectory by applying the pressure. so our unitcell also a variable parameter. These workflows can be controlled by the following input parameters.

&IONS

  ion_dynamics  = 'damp'

/

&CELL

  cell_dynamics = 'damp-pr'        #damp-pr denotes Parrinello-Rahman algorithm of lagrangian

  press_conv_thr = 0.05

/

🔺 Note: we can also trace the ions on various the temperature by following notations in the input file.

&ions

ion_temperature='nose'         

          tempw=4000      

          fnosep=3.0  

Now, we can proceed to the parallel calculation. Open your command prompt by using Windows key + R →type CMD press Enter. Then go to your mpi source code directory by the cd command. Then execute the mpiexec.exe run command. now it will show you the cores, threads usage command for your PC. I have 4 core processors so its showing mpiexec -n 4. Then go to your program directory by cd commands. 

Now we can execute the mpi parallel execution run command as mpiexec -n 4 pw.exe < Si.in > Si.out 

Wait for a while the program as to run completely. Then Open your  Si.out file using notepad, scroll down and find the " End final coordinates" this is your optimized atomic lattice parameter and optimized structure.

We have to copy this ed final coordinates and paste them in the input file. Now, this input file can be visualize the crystal structure using the Xcrysden package. Thank you for supporting my blog.....!!🥰

Quantum Espresso input file preparation

Dear Beginners,

In this session, I am gonna explain to you the steps to create a simple input file. QE input file must contain six parameters they are ...

• &CONTROL
• &SYSTEM
• &ELECTRONS
• K_POINTS
• ATOMIC_SPECIES
• ATOMIC_POSITIONS

Here the &CONTROL first, we have to define the calculation such as ' scf ' or ' relax '. Next, we have to define the pseudopotential file directory by pesudo_dir. In the example file, we have denoted the pesudo_dir = ' . ' which means that pseudopotentials are on the same directory of the input file. 

input file

In the &SYSTEM, we have to specify the crystal system. Here I downloaded the crystal information file from COD http://www.crystallography.net/cod/search.html. Our CIF file belongs to the cubic system. We have the lattice parameter as A=B=C= 5.46873Å. So we were given a = 5.46873. Then we have to specify The ecutwfc (Kinetic Energy cutoff for wave function). I have given define as 25 Ry. ibrav denotes our bravais lattice index, our crystal system belong to P1(simple cubic). So I choose 1 here; if it is FCC ibrav=2 similarly for other cases, https://www.quantum-espresso.org/Doc/INPUT_PW.html#idm199 (referred from). 

 nat is the total number of atoms in the unitcell.

 ntyp denotes total types of atoms in the unitcell. 

K_POINTS chooses to be automatic. We have a full symmetric system. It will not work for symmetry broken structures. ATOMIC_SPECIES must have the atomic mass and pseudopotential file. ATOMIC_POSITIONS copied from the CIF file, each position should be named by the atomic label. This is the simple input file format for quantum espresso. The file has to be saved in the specified directory as .in format [E-g: Si.in]. Moreover, it should contain the Pseudopotential file (you can download it from Quantum espresso official site).

let's do the calculation  by the following  command

 $ pw < Si.in > Si.out              # and press enter key

▲This will take time to finish, don't close the command prompt until the process to complete. 

Run command

File directory ad CMD after the run command

Check the Si.out file, it can be accessed by the help of Notepad.
 link for pseudopotentials https://www.quantum-espresso.org/pseudopotentials

=------------------------------------------------------------------------------=

                                         JOB DONE.

=------------------------------------------------------------------------------=