Putting error bars on the ab initio theoretical estimates of the magnetic coupling constants: the parent compounds of superconducting cuprates as a case study

The influence of the basis set size and computational method in the calculation of the magnetic coupling constant J is evaluated using a series of cuprate superconductor parent compounds as a case study. The variational DDCI method and an iterative modification, the IDDCI method, are tested, as well as the perturbative CASPT2 method, with two different reference wave functions. Results show that the DDCI magnetic coupling constant is in rather good agreement with the experiment, although it shows a moderate basis set dependency. The IDDCI results are less dependent on the size of the basis set, but slightly overestimate the magnetic coupling constant. CASPT2 results are nearly independent of the chosen basis set. With a minimal active space values are obtained that are about 20% smaller than the DDCI results. The experimental coupling constant can be reproduced when an extended reference wave function is used.