Hello, I have written a simple python program to compute the total and local density of states (DOS) of molecular systems from the PROCAR file. The reason was that it is very difficult to catch all the molecular DOS peaks using VASP no matter how fine the energy grid (EMIN, EMAX, NEDOS) is chosen. However, some doubt arise when comparing the total and local DOS resulting from my program. I am aware that the wavefunction characters in PROCAR were computed inside the Wigner-Seitz sphere of each atom and do not represent the total component of the wavefunction. If I understand correctly, when computing the local DOS one has to multiply the lorentzian energy function at band i by the squared value of the character Cijk of the i-th wavefuction at site j and orbital k. Is it correct? Or the characters written to PROCAR are already the squared values Cijk * Cijk?

Thanks in advance for your help.