I am very new with CASTEP simulation. Can you please help me out with the general steps for geometry optimization?
(1) Build unit cell > geometry optimization > cleave > supercell > optimize again> band structure and optical calculation
(2) Build unit cell > geometry optimization > cleave > supercell > band structure and optical calculation
(3) Build unit cell > geometry optimization > band structure and optical calculation If I go with (1) the simulation runs indefinitely. With (2), after oftimization, cleaving and supercell when I try to run bandstructure and optical calculation, an error message show saying "There are no CASTEP input or output files available. CASTEP input and output files from a previous run are required in order to perform a properties run." With (3) everything runs perfectly but this is only a unit structure and not a supercell.
Please advice
If you just want to get the band structure and optical properties, it is meaningless to run them on a supercell.
Normally, you need follow (1) to get results. In your case, your structure is too huge to finish. Please consider increase the number of processers of decrease k points and cut-off energy. Did you run this in MS? If so, the software will require geometry optimization before property calculation. This problem might be aroused by the missing of opt.
Sihan Zhang thanks for the reply. Is a supercell not necessary for simulation of bulk crystal lattice or is the unit cell representative of the lattice?
I tried with a 4x4 superstructure, cut off 500 eV and k-points 8 8 5. Do you think this is too much?
It will depend on your system. Common properties of crystals can be gained from unit cell, because it is run in inverted space. These backgrounds can be found in relative books. Supercell is used in doped materials or surfaces. In these cases, you cannot describe them in simple cells. Like the adsorption of H+, maybe only 25% if the sites are occupied by H. Therefore, you need a supercell to represent this.
For the parameters, I cannot say too much. The best way is to run convergency test and choose a set of parameters with acceptable errors.
Some additions. Keeping parameters in supercells many lose accuracy. Please check basic theories of KS equations. The integral is run by grid method. Cut off means how many wave functions are used in simulation. Increasing the size needs the increase of these two parameters.
Sihan Zhang thanks for the great pointers. It's a great help for me.
Minhaz Mahmood The main purpose of geometry optimization is to achieve the structure with minimum energy. It's not necessary to build a supercell if you aim to simulate certain physical properties, such as electronic band structure, density of states (DOS), elastic constants, optical parameters, etc. The following links to a published paper and a YouTube video might be helpful to you. Thank you.
Paper: Article Physical properties of niobium based intermetallics (Nb3B; B...
Tutorial: https://youtu.be/IMvzznBhEns?si=pBvApId4IYK0d053
Mst. Irin Naher thank you for the input. I have seen some of your videos in YouTube. I am having an issue recurrently, after geometry optimization, when I go to simulate the material properties such as band structure the following message comes: "There are no CASTEP input or output files available. CASTEP input and output files from a previous run are required in order to perform a properties run."
Minhaz Mahmood I have not faced this problem yet. I am not sure why you are facing it. Best of luck.
Minhaz Mahmood you need to make sure that you have copied the files back from wherever you did the CASTEP calculation, and you need to perform the band structure (or whatever calculation you want) on the geometry-optimised structure.