Accuracy of density functional theory methods for the calculation of magnetic exchange couplings in binuclear iron(III) complexes

[Display omitted] We assess the performance of commonly used hybrid density functional approximations for the prediction of magnetic exchange couplings in FeIII oxo-bridged complexes. We show that expected statistical errors are as small as 8%, which is significantly smaller than errors expected for a general metal-metal case. We assess the performance of popular hybrid density functional approximations for the prediction of magnetic exchange couplings (J) in FeIII complexes. To this end, eleven dinuclear oxo-bridged iron(III) complexes with accurately determined J values were chosen to cover cases of weak and strong antiferromagnetic couplings ranging from −6 cm-1 to −132 cm-1. Seven representative exchange-correlation functionals ranging from the hybrids PBEh and B3LYP to meta hybrids (M06, M062X), range separated functionals (HSE, LC-ωPBE), and LSDA with 30% of HF exchange were used for this assessment. These functionals have been suggested as good performers for this particular property in transition metal complexes and hence this assessment aims to narrow the interval of confidence for the particular case of Fe complexes. In addition, we examined the performance of an alternative method based on differential local spin rotations to calculate magnetic exchange coupling parameters in these complexes. Our results show that, although some of these functionals perform on par, the M06 and HSE functionals yield magnetic exchange couplings in closer agreement with experimental data, with mean absolute percentage errors of 8.6 and 9.2%, respectively.