Electric transition dipole moment in pre-Born–Oppenheimer molecular structure theory

Benjamin Simmen, Edit Mátyus, Markus Reiher

TL;DR

This paper presents a novel quantum mechanical method for calculating electric transition dipole moments in molecules without relying on the Born-Oppenheimer approximation. This approach treats electrons and nuclei on an equal footing, enabling more accurate calculations for systems where the traditional approximation breaks down, demonstrated here for the H₂ molecule.

What Problem Does This Paper Address?

The Born-Oppenheimer (BO) approximation is a cornerstone of molecular quantum chemistry, simplifying calculations by treating nuclei as stationary while electrons move around them. However, for certain phenomena, particularly those involving light molecules or strong electron-nuclear coupling, this separation becomes inaccurate. Calculating electric transition dipole moments, which govern how molecules interact with light, is crucial for understanding spectroscopy and photochemistry. While highly accurate calculations exist within the BO framework, extending these to a fully quantum mechanical, non-BO treatment has been a significant challenge. This paper addresses this by developing and applying a method to compute electric transition dipole moments without the BO approximation, treating all particles (electrons and nuclei) simultaneously.

Methods

The authors employ a pre-Born-Oppenheimer (pre-BO) quantum mechanical framework. Key to their approach is the use of explicitly correlated Gaussian (ECG) basis functions combined with the global vector representation (GVR). This combination ensures that the resulting wave functions are eigenfunctions of the total spatial angular momentum and parity operators.

Crucially, the calculations are performed in laboratory-fixed Cartesian coordinates, avoiding the need to separate the center-of-mass motion. Translational contamination, a common issue in such approaches, is eliminated through the derivation of translationally invariant integral expressions. The paper details the derivation of these integrals for the electric transition dipole moment operator in both the length and velocity representations. These representations are equivalent for exact wave functions, and the authors show how to evaluate them using a three-step procedure involving generating functions, derivatives, and angular integrals.

Datasets

Not discussed in the paper. The calculations are performed for the H₂ molecule using parameter sets for basis functions, which are provided in the supporting information.

Key Insights and Ideas

  • Simultaneous Treatment of Electrons and Nuclei: The core innovation is treating all particles on an equal quantum mechanical footing, moving beyond the BO approximation's separation of nuclear and electronic motion. This is essential for systems where electron-nuclear correlation is significant.
  • Translationally Invariant Integrals: Developing methods to eliminate translational contamination from integrals in laboratory-fixed coordinates is a significant technical achievement, simplifying the computational approach.
  • Equivalence of Length and Velocity Representations: The paper demonstrates the derivation of transition dipole moment integrals in both length and velocity forms, highlighting their equivalence for exact wave functions and their potential for assessing the accuracy of approximate wave functions.
  • First Pre-BO Calculation of Transition Dipole Moment: This work represents the first evaluation of the electric transition dipole moment within a full quantum mechanical, non-BO framework, opening doors for more accurate spectroscopic predictions.

Key Results

The authors successfully calculated the electric transition dipole moments for transitions between the lowest rovibronic levels of the H₂ molecule, specifically between the X ¹Σ⁺g and B ¹Σ⁺u electronic states. They report converged values for both the length and velocity representations, indicating the robustness of their method. The results are presented with high precision, showing the first differing digit.

Limitations and Future Work

The authors acknowledge that some sensitivity to the approximation of the wave functions was observed, which is expected in any approximate quantum mechanical treatment. They suggest that their developed scheme for eliminating translational contamination could be extended to other internal molecular properties. Future work could explore the interpretation of the permanent dipole moment in the pre-BO framework and investigate LFCC-TICC hybrid methods where specific correlation paths are parameterized.

Summary updated

Abstract

This paper presents the calculation of the electric transition dipole moment in a pre-Born-Oppenheimer framework. Electrons and nuclei are treated equally in terms of the parametrization of the non-relativistic total wave function, which is written as a linear combination of basis functions constructed from explicitly correlated Gaussian functions and the global vector representation. The integrals of the electric transition dipole moment are derived corresponding to these basis functions in both the length and the velocity representation. The calculations are performed in laboratory-fixed Cartesian coordinates without relying on coordinates which separate the center of mass from the translationally invariant degrees of freedom. The effect of the overall motion is eliminated through translationally invariant integral expressions. The electric transition dipole moment is calculated between two rovibronic levels of the H2 molecule assignable to the lowest rovibrational states of the X (1)Σ(g)(+) and B (1)Σ(u)(+) electronic states in the clamped-nuclei framework. This is the first evaluation of this quantity in a full quantum mechanical treatment without relying on the Born-Oppenheimer approximation.
BibTeX
@article{simmen2014electric,
  author = {Benjamin Simmen and Edit Mátyus and Markus Reiher},
  title = {{Electric transition dipole moment in pre-Born–Oppenheimer molecular structure theory}},
  journal = {The Journal of Chemical Physics},
  year = {2014},
  doi = {10.1063/1.4897632},
}
Figure