Answers are grounded in this paper and cite their sources.
We show that a naive treatment of open-shell systems in hybrid density functional calculations ignoring the spin dependence causes significant errors due to a kind of self interaction that is not emerged in spin-dependent calculations. As numerical examples, we compare the results of the LDA, GGA, and PBE0 calculations on the ionization potential and electron affinity of C₆₀ molecule and the GGA and HSE calculations on the singly charged monovacancy in crystalline Si.
No takes yet. Share an insight, caveat, or question.
Iwata et al. (2016) studied this question.