Spin-State Splittings in 3d Transition-Metal Complexes Revisited: Toward a Reliable Theory Benchmark

A new composite method
for the calculation of spin–crossover
energies in 3d transition-metal complexes based on multireference
methods is presented. The method reduces to MRCISD+Q at the complete-basis-set
(CBS) level for atomic ions, for which it gives excitation energies
with a mean absolute error of only ca. 0.01 eV. For molecular complexes,
the CASPT2+δMRCI composite approach corresponds to a CASPT2/CBS
calculation augmented by a high-level MRCISD+Q-CASPT2 correction with
a smaller ligand basis set. For a set of [Fe(He)6]n+ test complexes, the approach reproduces
full MRCISD+Q/CBS results to within better than 0.04 eV, without depending
on any arbitrary IPEA shifts. The high-quality CASPT2+δMRCI
method has then been applied to a series of 3d transition-metal hexaqua
complexes in aqueous solution, augmented by an elaborate 3D-RISM-SCF
solvent treatment of the underlying structures. It provides unprecedented
agreement with experiment for the lowest-lying vertical spin–flip
excitation energies, except for the Fe3+ system. Closer
examination of the latter case provides strong evidence that the observed
lowest-energy excitation at 1.56 eV, which has been used frequently
for evaluating quantum-chemical methods, does not arise from the iron(III)
hexaqua complex in solution, but from its singly deprotonated counterpart, [Fe(H2O)5OH]2+.