“…The choice of the electronic basis functions, {Φ i }, deserves a separate discussion. In many prior works on NA-MD simulations, such functions are chosen as either KS or MO orbitals, ,,,− or as the Slater determinants (SDs). ,− ,,− Although the adoption of the SD basis has been helpful in enabling modeling coupled electron–nuclear dynamics, for example, in Auger processes, such functions still miss an important physical constraint of being the eigenfunctions of the spin operator Ŝ 2 . A proper solution is the use of the spin-adapted configurations (SACs).…”