Argon lattice constant

[sumit agrawal]

Dear cp2k users,

I am trying to optimize the argon unit cell using KPOINTS. The reported cell constant for argon is 5.2229 A. But I am getting a cell constant of 5.5435 A. I have prepared a unit cell with the dimension of 5.2229 A. Then in the $CELL section I mention the same dimension.

Here I am attaching my input and output files.

Please give suggestions!!!

[Nicholas Winner]

Where is it reported as 5.2229A? On Materials Project, using PBE with VASP, they calculated 5.64A as the lattice constant. Considering you are using a dispersion correction, your calculation seems reasonably close to that. If you're sure that 5.2229 is the 'correct' answer using PBE+D3 then consider using a finer kpoint mesh.

[Lucas Lodeiro]

If you have the reference for that lattice constant, share it... I do a short search and I cannot find something similar. For example in https://doi.org/10.1063/1.1726009 says that the lattice constant is 5.3A. 

Nicholas comments about materials Project data, probably this: https://materialsproject.org/materials/mp-23155/

About materials Project data, is worth mentioning, the calculation was done using just the gamma point, which does not appear to be a converged calculation. Also the MP lattice constant is larger than experimental data.

Probably if you use a higher kpoint grid, higher cutoff and a little bit larger basis set, your calculation will get closer to experimental data, but do not think GGA DFT functionals (PBE among others) will give you exactly the correct lattice constant. Depending on what you want in terms of precision, maybe a hybrid functional will be necessary to improve the lattice constant, but remember that hybrid functional are more expensive than GGA ones... you have to define a proper balance between precision and cost.

regards

--
You received this message because you are subscribed to the Google Groups "cp2k" group.
To unsubscribe from this group and stop receiving emails from it, send an email to cp2k+uns...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/3aeed7b8-2782-4402-a3e4-d1bf789dbb24n%40googlegroups.com.

[sumit agrawal]

Dear Nicholas,

I also found it is 5.64 A on the Materials project. But i want to reproduce data from the following paper. In this paper they have calculated 5.2226 A. 

One query about kpoint mesh: I also use "SCHEME  MONKHORST-PACK  8 8 8". Here also I get the cell constant of 5.54 A, but I am not getting below 5.54 A.

PFA

Phys. Chem. Chem. Phys., 2015, 17, 17159-17168

--

[sumit agrawal]

Thanks Lucas,

Your suggestions are very helpful and informative.

I have the following queries :

Where it is mentioned that calculation is done using gamma points? In the calculation summary I haven't found this.

Second, can you suggest some good hybrid functional for such types of calculations?

Third, when I want to choose a lattice structure for my crystal (here for argon it is fcc which is cubic in nature) from a material project, which space group should I choose? I can see there are two space groups for agon cubic structure (m3m and 6/mmm). For m-3m i can see all the angles are 90 degree but for 6/mmm angles are not 90 degree then how could it be a cubic structure ? Should I go with m-3m only then ?

Thanks,

Sumit

[Lucas Lodeiro]

at the end of MP page, there is a table, # of K-points, when no kpoints are used, the calculation is computed on gamma point.

About the functional... it is important to search some literature about your system and used functionals... but HSE06 is a good one for almost all solid state calculations.

To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/CALSDoYaF6mKqeMOJg2WY0Eceb%2B_qbm-YhW8%2BiE6yv%3DMemAXhQA%40mail.gmail.com.

[sumit agrawal]

Thanks, Lucas....It's very helpful!!

To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/CAOFT4PJZkmiqDmjZ2%3D01EFsk-UpwnJR9vPH2PiEUQ4ryTpNLag%40mail.gmail.com.

[sumit agrawal]

I have one more query: When we construct a fcc lattice there will be 14 atoms (8 at the corner of the cube and 6 at center of the faces) in a 1x1x1 unit cell. But when we download an argon crystal structure file from a material project, there are only four atoms. Why is this so?  

Here is the download file.

[sumit agrawal]

This is understood. But if I want to prepare a fcc lattice for any small organic molecule, for example CH3OH or CCl4. Then how will I prepare the fcc lattice for these types of molecules? 

On Sun, May 23, 2021 at 2:34 AM Travis <polla...@gmail.com> wrote:

Hi,

Symmetry...

fcc crystal: 8 corners shared across 8 cells + 6 faces shared across 2 cells, reduced is 8*(1/8) + 6*(0.5) = 1 + 3 = 4 atoms. The CIF file has 4 atoms.

-T

To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/afb1d1bc-4ea7-4fdb-afe5-c74a81a372c0n%40googlegroups.com.