📄 chemistry

Atomic Ehrenfest forces

This study systematically evaluates the convergence of atomic Ehrenfest forces calculated via pair density and stress tensor methods across various bonding types, revealing that the pair density approach is highly sensitive to basis-set size—potentially yielding qualitatively incorrect directions for ionic systems with small bases—while exhibiting weak dependence on the level of theory, a behavior that parallels Hellmann-Feynman forces due to their shared electron-nuclear origin.

Original authors: Aldo J. Mortera-Carbonell, Samuel García-García, Evelio Francisco, Jesús Hernández-Trujillo, Ángel Martín Pendás

Published 2026-07-29

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Aldo J. Mortera-Carbonell, Samuel García-García, Evelio Francisco, Jesús Hernández-Trujillo, Ángel Martín Pendás

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine you are trying to understand why a Lego castle stays together. You could look at the energy stored in the plastic bricks, but there's another way: you could look at the invisible hands pushing and pulling on every single brick. In the microscopic world of atoms, scientists have been using "force" as a new lens to see how molecules hold together, moving beyond just calculating energy. Two famous ideas help here. First, the Hellmann-Feynman force is like calculating the tug-of-war between the heavy, positive nuclei (the anchors) and the cloud of negative electrons. Second, the Ehrenfest force looks at the net push and pull felt by the electrons themselves as they dance around the nuclei. While energy tells us if a molecule is stable, these forces tell us how the atoms are holding hands, offering a more physical, mechanical picture of chemical bonding. But here's the catch: in the quantum world, our calculations are only as good as the tools we use to measure them. If our "rulers" (mathematical tools called basis sets) aren't long enough or detailed enough, the forces we calculate might point in the completely wrong direction, making a friendly hug look like a violent shove.

This paper takes a deep dive into one of these force calculators: the Atomic Ehrenfest force. The authors, a team from Mexico and Spain, wanted to see how reliable this force is when we change the "tools" we use to calculate it. Specifically, they tested two different ways to compute the force on an atom within a molecule. The first method uses a mathematical shortcut involving the "stress" of the electron cloud (like measuring the tension in a stretched rubber sheet). The second method tries to calculate the force by looking at how pairs of electrons interact with each other, which requires a massive amount of data called the "pair density." They tested these methods on a variety of molecules, from simple hydrogen bonds to ionic salts and even weak van der Waals interactions, using different sizes of mathematical toolkits (basis sets) and different levels of theoretical complexity.

What they found is a bit like discovering that a cheap map can lead you to the wrong city. When they used the "pair density" method with small, simple toolkits (small basis sets), the results were often wildly wrong. For ionic molecules like Lithium Fluoride (LiF), the force on the Lithium atom would point away from its partner, suggesting the atoms were repelling each other, when in reality, they should be attracting. It was as if the map said "Go North" when you needed to go South. However, as they upgraded to larger, more detailed toolkits (specifically those with "triple-zeta" quality or better), the map corrected itself. The force suddenly flipped to point in the right direction, matching the results from the more stable "stress-tensor" method. Interestingly, changing the complexity of the theory (how they treated electron interactions) didn't mess things up nearly as much as using a small toolkit did. The authors also confirmed a beautiful mathematical symmetry: the total force pulling on all the electrons in a molecule is exactly equal in strength (but opposite in direction) to the total force pulling on the nuclei, a rule that holds true regardless of how imperfect the calculation is.

The study suggests that while the pair-density method is powerful, it is extremely sensitive to the quality of the mathematical tools used. If you use a small basis set, you might get a qualitatively incorrect picture of chemical bonding, especially for ionic systems where the direction of the force is crucial. The "stress-tensor" method, which relies on simpler data, is much more forgiving and stable, even with smaller toolkits. The authors conclude that to get a trustworthy picture of these atomic forces using the pair-density route, you absolutely need large, flexible basis sets. Without them, the "hands" of the atoms might appear to be pushing apart when they are actually pulling together, leading to a fundamental misunderstanding of how the molecule is built.

Technical Summary: Atomic Ehrenfest Forces

Problem Statement
The paper addresses the need for force-based descriptions of chemical bonding as a complement to traditional energy-centric perspectives. While the Hellmann-Feynman (HF) force has long served as a descriptor for nuclear interactions, it suffers from marked sensitivity to basis-set incompleteness, often yielding non-physical results for approximate wavefunctions. Similarly, the atomic Ehrenfest force, which describes the net force acting on electrons within a specific atomic basin (defined by the Quantum Theory of Atoms in Molecules, QTAIM), can be calculated via two distinct routes: (1) the divergence of the Pauli kinetic stress tensor, or (2) the averaging of the electronic force operator over the wavefunction, which requires the pair density ( $\rho_2$ ). While the stress-tensor route is computationally efficient and relies on the first-order reduced density matrix, the pair-density route is theoretically rigorous but computationally demanding. The authors note that the behavior of the pair-density route regarding basis-set dependence and its potential for yielding qualitatively incorrect results (particularly for ionic systems) had not been systematically investigated.

Methodology
The study systematically evaluates the atomic Ehrenfest force, $F(\Omega)$ , for a set of archetypal molecules representing covalent ( $H_2$ ), polar ( $H_2O$ ), ionic ($LiH$, $LiF$, $(NaF)_2$ ), and van der Waals ( $He_2$ ) interactions, as well as the $CNO^-$ anion. The calculations were performed at experimental geometries using various levels of theory (HF, DFT, CASSCF, MP2, CISD, FCI/MRCI) and a hierarchy of Dunning's basis sets (from cc-pVDZ to aug-cc-VQZ).

Two computational routes were compared:

Stress-Tensor Route: Calculated via the divergence of the Pauli kinetic stress tensor, dependent only on the first-order reduced density matrix.
Pair-Density Route: Calculated by integrating the electronic force operator over the QTAIM basin. This required the pair density $\rho_2(r, r_2)$ . For methods where the exact pair density is unavailable (e.g., MP2, DFT), the authors employed the Müller approximation, which reconstructs the exchange-correlation pair density using natural orbitals and occupation numbers. The validity of this approximation was benchmarked against exact pair-density calculations for selected cases.

The study also examined the dissociation curves of $H_2$ and $He_2$ and compared the total magnitudes of the Ehrenfest and Hellmann-Feynman forces across different basis sets and theoretical levels.

Key Results

Basis-Set Sensitivity: The atomic Ehrenfest force derived from the pair density exhibits severe sensitivity to basis-set size. For ionic systems (e.g., $LiH$, $LiF$), small basis sets (e.g., cc-pVDZ) yield a qualitatively incorrect force direction: the force on the electropositive atom is spuriously repulsive (pointing away from the bond) rather than attractive. As the basis set size increases to triple-zeta quality (aug-cc-pVTZ) and beyond, the force converges to the physically correct attractive direction and aligns with the values obtained from the stress-tensor route.
Stress-Tensor Robustness: The stress-tensor route is comparatively insensitive to basis-set incompleteness because it relies solely on the first-order density matrix. It consistently yields the correct force direction and magnitude across the tested basis sets.
Level of Theory Dependence: In contrast to basis-set size, the level of theory (and the treatment of electron correlation) has a comparatively weak influence on the atomic Ehrenfest force. The behavior parallels that of the Hellmann-Feynman force, where the description of electron density near the nuclei (governed by the cusp condition) is the dominant factor.
Müller Approximation: The Müller approximation for the pair density reproduces the trends and values of exact pair-density calculations closely, validating its use for correlated wavefunctions where exact $\rho_2$ is computationally prohibitive.
Force Sum Rules and Equality: The study confirms that the sum of atomic Ehrenfest forces over all topological atoms approaches zero as the basis set limit is reached, reflecting the fulfillment of the hypervirial theorem. Furthermore, the paper establishes an exact equality between the magnitudes of the total Ehrenfest force ( $|F_{Ehr}|$ ) and the total Hellmann-Feynman force ( $|F_{Feyn}|$ ) for any electron density and geometry. This equality arises because the electron-electron and nuclear-nuclear interaction terms vanish upon summation due to antisymmetry, leaving only the electron-nuclear integral, which is identical in magnitude but opposite in sign for the two forces.

Significance and Conclusions
The paper concludes that while the atomic Ehrenfest force is a valuable descriptor for interatomic interactions, its interpretation from pair-density calculations requires caution, particularly when using modest basis sets. For ionic systems, insufficient basis sets can lead to a complete reversal of the force vector, masking the true nature of the bonding. The authors assert that adequately flexible, augmented basis sets are essential for the pair-density route to yield reliable chemical conclusions.

The work highlights that the shared electron-nuclear origin of both the Ehrenfest and Hellmann-Feynman forces underpins their identical sensitivity to basis-set incompleteness. By demonstrating the exact equality of their total magnitudes and the convergence of the pair-density route to the stress-tensor route at the complete-basis-set limit, the study provides a rigorous framework for validating force-based chemical descriptors. The findings reinforce that force-based representations offer a tangible toolkit for visualizing molecular structure, provided the computational parameters are sufficient to satisfy the underlying physical constraints.

Technical Summary: Atomic Ehrenfest Forces

More like this