Essentially, there are two ways to do this using VASP. In each case, you'll obtain the dielectric tensor for which the imaginary part is related to the absorption.

The first is to use the LOPTICS tag. If this is set to be true, then VASP will calculate the dielectric tensor following Kubo-Greenwood type formula (i.e. the imaginary part of the dielectric tensor is the sum over occupied and unoccupied bands of the dipole matrix elements, neglecting local field effects). Depending upon the system, this may be accurate enough. If the NPAR tag is set to 1, then the matrix elements will be written out to the OPTICS file which would allow you to use a program to recalculate the dielectric tensor with a rigid band shift to adjust the gap (for example, I have done this to match ellipsometry data for NbO2 and obtain an approximation for the the band gap).

http://cms.mpi.univie.ac.at/vasp/vasp/LOPTICS_frequency_dependent_dielectric_matrix.html

A more involved process is to do a GW quasi-particle calculation followed by the absorption with the Bethe-Salpeter Equation (BSE).  The first step corrects Kohn-Sham eigenvalues by a perturbation theory approach, while the second step then calculates the dielectric tensor accounting for electron-hole effects.  This can have the impact of correct oscillator strengths, i.e. relative peak heights (for example in silicon) to the calculation of excitons (i.e. a state in which an electron-hole pair has a binding energy that reduces the optical gap to be lower than the fundamental gap calculated through standard methods or measured via transport or IPS/UPS). This type of calculation is much more involved than the above; however, may be necessary depending upon the system.

http://cms.mpi.univie.ac.at/vasp/vasp/Frequency_dependent_GW_calculations.html

Looks like a job for TD-DFT. Don't know for extended solids but works for single molecules

Essentially, there are two ways to do this using VASP. In each case, you'll obtain the dielectric tensor for which the imaginary part is related to the absorption.

The first is to use the LOPTICS tag. If this is set to be true, then VASP will calculate the dielectric tensor following Kubo-Greenwood type formula (i.e. the imaginary part of the dielectric tensor is the sum over occupied and unoccupied bands of the dipole matrix elements, neglecting local field effects). Depending upon the system, this may be accurate enough. If the NPAR tag is set to 1, then the matrix elements will be written out to the OPTICS file which would allow you to use a program to recalculate the dielectric tensor with a rigid band shift to adjust the gap (for example, I have done this to match ellipsometry data for NbO2 and obtain an approximation for the the band gap).

http://cms.mpi.univie.ac.at/vasp/vasp/LOPTICS_frequency_dependent_dielectric_matrix.html

A more involved process is to do a GW quasi-particle calculation followed by the absorption with the Bethe-Salpeter Equation (BSE).  The first step corrects Kohn-Sham eigenvalues by a perturbation theory approach, while the second step then calculates the dielectric tensor accounting for electron-hole effects.  This can have the impact of correct oscillator strengths, i.e. relative peak heights (for example in silicon) to the calculation of excitons (i.e. a state in which an electron-hole pair has a binding energy that reduces the optical gap to be lower than the fundamental gap calculated through standard methods or measured via transport or IPS/UPS). This type of calculation is much more involved than the above; however, may be necessary depending upon the system.

http://cms.mpi.univie.ac.at/vasp/vasp/Frequency_dependent_GW_calculations.html

I follow Andrew's response. Just be aware that VASP, for example, will only output the real and imaginary part of the dielectric function, so is your job to script a code that calculates optical properties (refractive index and extension coefficient).

Thanks for your valuable suggestions.

Dear David, 

Could you please guide me how to script the code to calculate optical properties?

Dear Muhammad,

You can use vaspkit to calculate optical properties including absorption coefficient, refractive coefficient, reflectivity coefficient, extinction coefficient and energy-loss function as a function of photon energy.

http://vaspkit.sourceforge.net/

yes, you can

https://www.researchgate.net/publication/289406883_First-principles_calculations_of_electronic_optical_and_thermodynamic_properties_of_borophene

Article First-principles calculations of electronic, optical, and th...

Dear Peng,

Have you used any external post processing tool for calculating absorption spectra, reflectivity, dielectric function and electron energy-loss function ?

All these properties can we calculate only by VASP  output  by setting appropriate INCAR tags ? Extracting these results from VASP output is easy or little bit complicated to make good quality plot.

Can you elaborate little much about these calculation and how does extract data and plotting of these data.

I know PHONOPY for lattice thermodynamic properties calculations.

Thank You

Shilendra

Hi Shilendra,

   I think you can use the real and imaginary parts of dielectric function as inputs, and calculate other optical properties directly. Vaspkit may be OK, or you can use MATLAB or even Excel. I do not know why it should be a problem.

Best,

Bo

Hi,

Once I have done the calculation with LOPTICS and I have used vaspkit I have several files (IMAG.in, ABSORB.dat, EXTINCTION.dat). I have two questions:

1. Which one should I represent to compare with experimental absorption spectra?

2. How should I treat the six tensor components? I mean which of them I should consider or if I should do the average of the six columns to compare with experimental espectra.

Thanks in advance,

Rocío

Dear All,

How can we know dielectric constant by command from VASP OUTCAR after static dielectric calculation.

Hi Kuntal,

PHONOPY is a free code. I am using in my linux (Ubuntu) system. In Linux, you can install PHONOPY by command or by sourse code also. Very easy installation by one command that is available on PHONOPY website and inside PHONOPY documetattion.

Hi, Kuntal Kabra

Ideally you should install VASPKIT and phonopy on Linux OS. It is much easier to use, or you can install Virtual Machine on your Windows OS, then install linux in it.

Hi All

I have a question as:

I've used vaspkit to calculate the linear optics of crystaline structure so what is unit in ABSORB.dat at column 2-7 for component of I(omega) xx yy zz xy yz zx.

Thanks,