My Community

Dear VASP users,

I am trying to use mBSE for the dielectric function calculation, for this I am looking at the examples related to mBSE.
I am using the following link for the reference,

https://www.vasp.at/wiki/index.php/Impr ... c_function

Here, I am trying to fit the epsilon^-1 at small wave vector (q) with respect to |q+G| where, G is the reciprocal vector, to calculate the lambda.
After fitting I am not getting 1.26 as mention in the example. Please help me out to fit the same so that I can use the same for my calculations.

---
Chandan

Hi,

I am not quite sure if it is the fitting that is the problem or the calculated dielectric function. Do you get the same dielectric function as in the example on the wiki? What dielectric constant do you get?

In the example file they have provided all the inputs along with the calculated dielectric file, i.e.
dieG_g6x6x6-GW0.dat file. I have taken the static dielectric function from the input itself which is 0.088.

Then I fitted this dieG_g6x6x6-GW0.dat data with the model local function. For your reference I am attaching the
gnuplot script which I am using to fit the dielectric data with the model function, please have a look and let me know.

######### Gnuplot script #############
set terminal postscript color solid enhanced font "Times-Roman, 16"
set style line 1 lw 3 lc rgb '#990042' ps 2 pt 12 pi 3
set style line 2 lw 3 lc rgb '#0044a5' ps 2 pt 6 pi 3
set style function linespoints

set output "fig.eps"
set size square
set ylabel "{/Symbol e}^{-1}(G)" font "Times-Roman, 28"
set xlabel "|G|" font "Times-Roman, 28"
set xtics 0, 1, 7 font "Times-Roman, 24"
set ytics 0, 0.2, 1 font "Times-Roman, 24"
set key at 0.05, 7.68
set key spacing 1.5
f(x) = 1-(1-0.088)*exp(-(x)**2/4*z**2)
fit f(x) "dieG_g6x6x6-GW0.dat" using 1:2 via z

plot [0:7] [0:1] "dieG_g6x6x6-GW0.dat" using 1:2 w l ls 2 , f(x) w l ls 1 notitle

########################################

---
Chandan

Your script is fitting a different function. It should be

Thanks Alexey, It worked!!!