Validity of symmetry breaking for energetic accuracy in strongly correlated systems

Establish whether using symmetry-broken ground states to transform strong correlation in a symmetric state into normal correlation necessarily yields correct total energies within density functional theory for strongly correlated systems such as the chromium dimer, where the symmetric singlet ground state may be approximated by an antiferromagnetically ordered broken-symmetry state.

Background

The chromium dimer is a challenging strongly correlated system where many-body methods have historically struggled. The authors discuss a common strategy—symmetry breaking—to reduce strong correlation by describing a symmetry-broken state that a reliable density functional can handle.

They note that while this approach often works, it lacks a general proof guaranteeing correct energies, highlighting a foundational uncertainty about the validity of symmetry breaking for energetic predictions in such systems.