Non-resonant Raman intensity

[twozniak28]

Dear Developers,

Is it possible to obtain non-resonant Raman intensitites with QERaman? In subroutine raman_tensor (file calc_raman_intensity.f90) I can see the following lines:

        ! Product of three matrix element
        ! [D(k,f2,i)*M_q(v,k,f1,f2)*D(k,i,f1)]
        ! Raman intensity should proportion to 1/E_L^4 since M_opt proportion to D/E_L
        rt = (dvec(ik,ibn2,ibi,i) / (elaser(iel)-eq(imode,iq))) &
            * matele_elph(imode,iq,ik,ibn1,ibn2) * (dvec(ik,ibi,ibn1,j) / elaser(iel))
        !rt = dvec(ik,ibn2,ibi,i) * matele_elph(imode,iq,ik,ibn1,ibn2) * dvec(ik,ibi,ibn1,j)
        ! Non-resonant
        !  rt = rt / ((eigv(ik,ibn1)-eigv(ik,ibi)) * (eigv(ik,ibn2)-eigv(ik,ibi)))
        ! Resonant condition
        rt = rt / ((elaser(iel)-(eigv(ik,ibn1)-eigv(ik,ibi))-im*gamma) &
            * (elaser(iel)-(eigv(ik,ibn2)-eigv(ik,ibi))-eq(imode,iq)-im*gamma))
        !make Raman tensor
        rtensor_k(iel,imode,ik,iq,i,j) = rtensor_k(iel,imode,ik,iq,i,j) + rt

Which ones should I comment/uncomment to obtain non-resonant intensities?

With kind regards,
Tomasz

[nguyen-group]

Dear Tomasz,

I am sorry for missing your previous question about the unit of Raman intensity.
So, let me answer both of your questions in this post.

1. In QERaman, the Raman intensity is an arbitrary unit. So you can not compare directly with non-resonant Raman in QERaman.
   It is noted that non-resonant Raman and resonant Raman are different mechanisms and cannot be compared with each other. The non-resonant Raman with laser energy < energy band gap and the resonant Raman with laser energy > energy band gap. Thus, non-resonant Raman does not depend on the laser energy.

2. Is it possible to obtain non-resonant Raman intensitites with QERaman?
   As explained above, non-resonant Raman does not depend on the laser energy (elaser(iel) in the code). So, by comment/uncomment, you can obtain non-resonant Raman. It also noted that this was for testing when we made the code. We have not yet validated it by experimental comparison. We would appreciate it if you could test it and let us know if it works on this discussion page.

[twozniak28]

Dear Nguyen,

Thanks for your answers.

ad. 2. Then these lines should look like that for non-resonant Raman?

        ! Product of three matrix element
        ! [D(k,f2,i)*M_q(v,k,f1,f2)*D(k,i,f1)]
        ! Raman intensity should proportion to 1/E_L^4 since M_opt proportion to D/E_L
        ! rt = (dvec(ik,ibn2,ibi,i) / (elaser(iel)-eq(imode,iq))) &
            * matele_elph(imode,iq,ik,ibn1,ibn2) * (dvec(ik,ibi,ibn1,j) / elaser(iel))
        rt = dvec(ik,ibn2,ibi,i) * matele_elph(imode,iq,ik,ibn1,ibn2) * dvec(ik,ibi,ibn1,j)
        ! Non-resonant
        rt = rt / ((eigv(ik,ibn1)-eigv(ik,ibi)) * (eigv(ik,ibn2)-eigv(ik,ibi)))
        ! Resonant condition
        ! rt = rt / ((elaser(iel)-(eigv(ik,ibn1)-eigv(ik,ibi))-im*gamma) &
            * (elaser(iel)-(eigv(ik,ibn2)-eigv(ik,ibi))-eq(imode,iq)-im*gamma))
        !make Raman tensor
        rtensor_k(iel,imode,ik,iq,i,j) = rtensor_k(iel,imode,ik,iq,i,j) + rt

[nguyen-group]

Hi,

It looks correct.

Nguyen

[twozniak28]

In that case, non-resonant Raman spectrum of 1L WS2 is wrong, wrt to QE:

1. QERaman provides high intensity for E'' mode (~300cm-1), which is very weak according to QE.
2. QERaman provides zero intensity for E' mode (~361cm-1), which is the strongest one according to QE.