How To Arrive at Accurate Benchmark Values for Transition Metal Compounds: Computation or Experiment?

Area of Science:

• Quantum Chemistry
• Computational Chemistry
• Spectroscopy

Background:

• Benchmarking quantum chemical methods requires reliable reference data.
• Transition metal compounds present unique challenges for computational modeling.
• Experimental data may not always be sufficiently accurate for rigorous benchmarking.

Purpose of the Study:

• To analyze appropriate reference data for benchmarking quantum chemical approaches for transition metal compounds.
• To develop and apply a composite computational method for accurate property prediction.
• To assess the impact of reference data choice on the evaluation of approximate methods.

Main Methods:

• A dataset of 60 transition metal diatomic molecules with known experimental properties was compiled.
• A composite computational approach using coupled-cluster theory with basis set extrapolation and relativistic/multireference corrections was employed.
• Internally contracted multireference coupled-cluster (icMRCC) theory was used for multireference corrections.

Main Results:

• The composite computational approach yielded accurate dissociation energies and spectroscopic constants for the studied molecules.
• Coupled cluster singles and doubles with perturbative triples (CCSD(T)) at the complete basis set limit often provides accurate reference values.
• Multireference corrections enhanced accuracy in cases where CCSD(T) was insufficient, though some experimental values may require revision.

Conclusions:

• Accurate computational methods can serve as a valuable alternative to experimental data for benchmarking.
• The choice of reference data (experimental vs. computational) had minimal impact on the relative performance ranking of DFT functionals.
• Careful revision of experimental data is suggested for certain transition metal compounds where discrepancies persist.