Density Functional Theory (DFT) combined with the Broken Symmetry (BS) method is today widely used in the field of molecular magnetism for the computation of exchange coupling constants. But this method (DFT-BS) remains semi-quantitative as it suffers from a series of drawbacks already discussed in the literature. In order to better understand the origin of such problems, we reexamined the physical contributions acting in the exchange phenomenon. We then propose alternative analytical expressions built along two complementary approaches (molecular orbitals and valence bond). During our analysis, we found a new problem linked to the broken symmetry state as it comes out of a DFT calculation. Our models will be applied to copper(II) dimers and we will show how to quantify the different parameters involved in order to reconstruct the coupling constants. Moreover, our work allows for an original quantitative correspondence between the two above-mentioned approaches.
