Energy-consistent pseudopotentials for the 5d elements--benchmark calculations for oxides, nitrides, and Pt(2).

Area of Science:

• Computational chemistry
• Quantum chemistry
• Relativistic effects in chemistry

Background:

• Relativistic energy-consistent pseudopotentials are essential for accurately describing heavy elements.
• Previous pseudopotential methods often required empirical adjustments.
• Accurate theoretical models are needed for predicting molecular properties of 5d elements.

Purpose of the Study:

• To evaluate the performance of newly developed relativistic energy-consistent pseudopotentials for 5d elements.
• To assess the accuracy of these pseudopotentials in benchmark calculations for diatomic molecules.
• To determine the reliability of these pseudopotentials for predicting spectroscopic constants.

Main Methods:

• Utilized coupled cluster (CC) and multireference configuration interaction (MRCI) benchmark calculations.
• Employed basis sets of up to quintuple-zeta quality for high precision.
• Adjusted pseudopotentials to atomic valence spectra from multiconfiguration Dirac-Hartree-Fock (MCDHF) calculations.

Main Results:

• Achieved high accuracy for oxides and nitrides (HfO, TaO, WO, ReN, OsN, IrN) with corrections for pseudopotential errors (e.g., 4f shell correlation).
• Obtained bond length accuracy of 0.3 pm and harmonic vibrational frequency accuracy of 17 cm(-1) (1.5%).
• Reproduced spectroscopic constants for Pt(2) with deviations of 3 pm for bond length and 1 cm(-1) for vibrational frequency, even without pseudopotential error corrections.

Conclusions:

• The new relativistic pseudopotentials demonstrate excellent performance for 5d elements in molecular calculations.
• These pseudopotentials, particularly when accounting for minor error contributions, provide highly accurate spectroscopic constants.
• The study validates the use of these pseudopotentials for reliable theoretical predictions in heavy element chemistry.