Let's assume a scenario:

We have extracted electron density F for ground electronic state S of molecule A and electron density G for a frank-condon excited state E of the same molecule A. We have extracted both F and G in a standard Gaussian cube format such that:

  1. The voxels of cubes are sufficiently fine to account for any significant distortion in electron cloud.
  2. The cube dimensions are sufficiently large that the electronic density lying outside the cube can be truncated to zero.

Having these conditions, are there any metrics via which we can differentiate and highlight difference between these cubes? We can project the charge density to partial charges on atoms of A or maybe use general physical features like number of nodes, phase, etc., but are there any other metrics for 3-D data that we can use to numerically quantify the difference between F and G?

(For ex.: For a 2-D distribution we can statistical dispersion to calculate skewness and variance of the dataset.)

  • 1
    $\begingroup$ Mathematically, the squared norm $||F-G||_2^2$ (i.e. the integral of |F-G|^2 over the real space) obviously counts as one, but I don't know if it suits your purpose. $\endgroup$
    – wzkchem5
    Dec 18, 2022 at 8:33
  • $\begingroup$ @wzkchem5 I am not sure how much unique that metric can be... Although it is reasonable choice. $\endgroup$
    – mykd
    Dec 18, 2022 at 8:44
  • $\begingroup$ Maybe you could compute multipole (dipole, quadrupole, etc.) moments to quantitatively describe the shapes of the charge density distributions. This sounds like what you’re interested in. You’ll have to take into account the atomic nuclear charges too so that the system is charge neutral. $\endgroup$
    – Stephen
    Dec 18, 2022 at 23:34
  • $\begingroup$ @StephenWeitzner Thanks for your suggestion! However I was looking for a methodology where I do not have to either decompose or project the charge density. I am also trying to compare the "exact density" between different density functionals, so I was looking for a general method to compare electron densities. $\endgroup$
    – mykd
    Dec 19, 2022 at 11:08
  • 1
    $\begingroup$ In case things start to get long here, don't accept the prompt to create a new chat room. Just continue the discussion in an existing room and link to it here. I can handle moving over/deleting existing comments. $\endgroup$
    – Tyberius
    Dec 28, 2022 at 0:06

1 Answer 1


If your CUBE file has the basis set info included, you can use the Multiwfn software. From the program site:

  • Outputting all supported real space functions as well as gradient and Hessian at a point.
  • Calculating real space function along a line and plot curve map (Electron density, Gradient norm of electron density, Laplacian of electron density, Value of orbital wavefunction, Electron spin density Hamiltonian kinetic energy density K(r), Lagrangian kinetic energy density G(r), Electrostatic potential, Electron localization function (ELF), Localized orbital locator (LOL), Local information entropy, Electrostatic potential (ESP), etc.).
  • Topology analysis for any real space function, such as electron density (AIM analysis).
  • Critical points (CPs) can be located, topology paths and interbasin surfaces.
  • Population analysis. Hirshfeld, Hirshfeld-I, VDD, Mulliken, Löwdin, Modified Mulliken (including three methods: SCPA, Stout & Politzer, Bickelhaupt), Becke, ADCH (Atomic dipole moment corrected Hirshfeld), CM5, 1.2*CM5, CHELPG, Merz-Kollmann, RESP (Restrained ElectroStatic Potential), RESP2, AIM (Atoms-In-Molecules), EEM (Electronegativity Equalization Method) and PEOE (Gasteiger).
  • Orbital composition analysis. Mulliken, Stout & Politzer, SCPA, Hirshfeld, Hirshfeld-I, Becke, natural atomic orbital (NAO) and AIM.
  • Bond order/strength analysis.
  • Charge decomposition analysis (CDA) and extended CDA analysis.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .