Velocity Gauge for Oscillator Strength in $\Delta$SCF theory

This paper proposes using the velocity gauge within $\Delta$SCF theory to compute oscillator strengths, demonstrating that this approach naturally resolves the non-orthogonality and origin-dependence issues of standard methods to yield accurate, origin-independent results for conjugated chromophores, especially when combined with spin-purified singlet excitation energies.

The Gist

The Big Picture: Measuring the "Flash" of an Electron

Imagine a molecule as a tiny, complex city. Inside this city, electrons are the citizens. Sometimes, a citizen (an electron) gets a boost of energy and jumps from a low-rise apartment (the ground state) to a penthouse suite (the excited state).

When this happens, the molecule might absorb light or glow. Scientists want to know exactly how bright that glow will be. This brightness is called the Oscillator Strength.

For decades, scientists have used a method called $\Delta$SCF (Delta Self-Consistent Field) to calculate the energy required for this jump. It's like a reliable GPS that tells you exactly how high the penthouse is. However, this GPS has a major glitch: it cannot accurately tell you how bright the light will be when the electron jumps.

The Problem: The "Moving Origin" Glitch

To understand the glitch, imagine you are trying to measure the distance between two people in a room.

• The Standard Way (Length Gauge): You measure from the corner of the room.
• The Glitch: In the $\Delta$SCF method, the "ground state" person and the "excited state" person are calculated in two different simulations. Because they were calculated separately, they don't quite "fit" together perfectly. They are like two maps of the same city drawn by different cartographers who used different starting points (origins).

If you try to measure the distance between them using the standard method, the result changes depending on where you stand in the room (the "origin").

• Stand here? The distance is 5 meters.
• Stand there? The distance is 10 meters.

This is physically impossible. The brightness of a molecule shouldn't change just because you moved your chair. This makes the standard calculation "unphysical" and unreliable.

The Old Fixes: Trying to Force the Maps to Match

Scientists tried to fix this by forcing the two maps to align.

1. Symmetric Orthogonalization: They tried to mathematically "squash" the maps so they fit together perfectly. But this is like forcing two different puzzle pieces to fit by breaking the edges. It changes the shape of the pieces (the wavefunctions), which might introduce new errors.
2. Adding Nuclei: They tried adding the weight of the building's foundation (the atomic nuclei) to the calculation. This works great for neutral buildings (uncharged molecules), but if the building is floating in space (a charged ion), the fix doesn't work, and the measurement still wobbles.

The New Solution: The "Velocity Gauge" (The Speedometer)

The authors of this paper, Yang Shen, Yichen Fan, and Weitao Yang, found a clever workaround. Instead of trying to fix the "distance" (position) measurement, they switched to measuring speed (momentum).

The Analogy:
Imagine you are trying to measure how fast a car is going.

• Length Gauge: You try to measure the distance the car traveled from a specific starting line. If your starting line is shaky, your speed calculation is wrong.
• Velocity Gauge: You look at the car's speedometer. It doesn't matter where the car started or where the road is; the speedometer tells you the speed directly based on the engine's power.

In physics terms, the Velocity Gauge calculates the transition based on the momentum (how fast the electron is moving) rather than its position (where it is).

• The Magic: Momentum is "origin-independent." It doesn't matter where you place your coordinate system; the speed of the electron remains the same.
• The Result: By using the velocity gauge, the authors bypassed the "moving origin" problem entirely. They didn't need to force the maps to align or break the puzzle pieces. They just used a different ruler that didn't care about the starting point.

The Twist: Cleaning Up the "Spin"

There was one more catch. In the quantum world, electrons have a property called "spin" (like a tiny top spinning clockwise or counter-clockwise).

• The $\Delta$SCF method often creates a "messy" state where the electron spin is a confused mix of "up" and "down" (spin contamination).
• The authors found that if they "cleaned up" this spin mix (a process called Spin Purification) before doing the speed calculation, the results became even more accurate, especially for large, complex molecules like those used in solar cells or dyes.

Why This Matters

1. No More Guessing: This method allows scientists to calculate the brightness of light absorption/emission for charged molecules (ions) and large complex molecules with high accuracy, something previous methods struggled with.
2. Keep It Simple: You don't need to change the fundamental physics or force the wavefunctions to be orthogonal. You just switch the "gauge" (the tool you use to measure).
3. Better Materials: This helps in designing better organic solar cells, LEDs, and sensors, where knowing exactly how a molecule interacts with light is crucial.

Summary

The paper solves a decades-old problem where measuring the "brightness" of an electron jump was unreliable because the measurement depended on where you stood. The authors realized that instead of measuring the position (which was shaky), they should measure the speed (which was stable). By switching to the Velocity Gauge, they got accurate, reliable results without needing to force the math to behave unnaturally. It's a simple switch of perspective that yields powerful results.

Technical Summary

1. Problem Statement

The $\Delta$SCF (Delta Self-Consistent Field) method within Kohn-Sham Density Functional Theory (KS-DFT) is a widely used and computationally efficient approach for calculating electronic excitation energies, often outperforming Time-Dependent DFT (TDDFT) for core-level excitations and long-range charge transfer. However, calculating oscillator strengths (transition probabilities) in $\Delta$SCF remains a significant challenge due to the non-orthogonality of the ground-state and excited-state wavefunctions.

• The Core Issue: In $\Delta$SCF, the ground and excited states are obtained from separate SCF optimizations, resulting in two different Kohn-Sham (KS) determinants that are not orthogonal.
• Consequence in Length Gauge: When calculating transition dipole moments (TDM) using the standard length gauge ($\hat{\mu} = \sum \hat{r}_i$), the non-orthogonality leads to origin-dependent results. As shown in the paper's Eq. 1, translating the coordinate system arbitrarily changes the TDM, rendering the physical quantity meaningless.
• Limitations of Existing Solutions:
  • Symmetric Orthogonalization: Forces orthogonality but alters the ground or excited state determinants, potentially introducing errors.
  • Nuclear Contribution Correction: Adding nuclear terms to the perturbation Hamiltonian works for neutral systems but fails to fully cancel origin dependence in charged systems.
  • Configuration Interaction (CI) expansions: Often require additional assumptions or are computationally expensive.

2. Methodology

The authors propose a novel solution: utilizing the velocity gauge to compute oscillator strengths directly from the unmodified $\Delta$SCF KS wavefunctions.

• Velocity Gauge Formulation: Instead of the position operator ($\hat{r}$), the velocity gauge employs the momentum operator ($\hat{p}$) in the interaction Hamiltonian. The oscillator strength ($f_{VG}$) is defined as:
  $$f_{VG} = \frac{2}{3} \frac{|\langle \Phi_0 | \sum \hat{p}_i | \Phi_m \rangle|^2}{E_m - E_0}$$
• Theoretical Advantage: The momentum operator $\hat{p}$ is translationally invariant. Consequently, the velocity gauge naturally accounts for the non-orthogonality of the $\Delta$SCF determinants without requiring the wavefunctions to be orthogonalized or the origin to be fixed. This eliminates the origin-dependence issue inherently.
• Spin Purification: The paper investigates the impact of spin contamination in open-shell singlet states.
  • Standard $\Delta$SCF often yields a mixture of singlet and triplet states.
  • The authors test two approaches: using the raw (unpurified) $\Delta$SCF energy and using a spin-purified singlet energy (calculated via $E_S = 2E_{mixed} - E_{triplet}$).
  • Crucially, they compute the transition momentum using the raw, single-determinant $\Delta$SCF wavefunctions (without projecting onto multi-configurational states), arguing that the single determinant is the direct approximation of the interacting system wavefunction in this framework.
• Implementation: The transition momentum is evaluated using a wavefunction-like approach involving the overlap matrix ($S_{0m}$) between occupied orbitals of the ground and excited states (Eq. 15). The code was implemented in the PySCF package.

3. Key Contributions

1. Origin-Independence without Modification: Demonstrated that the velocity gauge provides origin-independent transition properties for $\Delta$SCF calculations without altering the KS determinants or adding ad-hoc corrections (like nuclear terms).
2. Applicability to Charged Systems: Showed that while length-gauge corrections fail for charged systems (due to incomplete cancellation of origin dependence), the velocity gauge remains robust and accurate for both neutral and charged species.
3. Spin-Purification Strategy: Identified that while the velocity gauge works with unpurified energies, adopting spin-purified excitation energies significantly improves the accuracy of oscillator strength predictions, particularly for conjugated chromophores where singlet-triplet splitting is small but non-negligible.
4. Validation: Validated the method against high-level benchmarks (EOM-CCSD) for small molecules and TDDFT for large conjugated systems.

4. Results

The study tested the method on two sets of molecules: small organic molecules and large conjugated chromophores.

• Small Molecules (Neutral):
  • The velocity gauge results (using unpurified energies) showed Mean Absolute Errors (MAE) comparable to length-gauge results that required symmetric orthogonalization.
  • Spin-purified energies yielded slightly larger errors for small molecules but were statistically insignificant.
• Charged Systems:
  • For the $CH_3^-$ anion, the length-gauge TDM showed strong linear dependence on the coordinate origin (slope $\approx$ net charge $\times$ overlap).
  • The velocity gauge TDM was effectively independent of the origin (standard deviation $\approx 10^{-10}$ a.u.), confirming its theoretical robustness for charged systems.
• Conjugated Chromophores:
  • Standard $\Delta$SCF in the length gauge systematically overestimated oscillator strengths compared to TDDFT.
  • The velocity gauge with unpurified energies replicated this overestimation.
  • Critical Finding: Using spin-purified singlet excitation energies in the velocity gauge significantly reduced the error, bringing the results into close agreement with TDDFT benchmarks (MAPE for oscillator strength dropped from ~100% to ~17%).

5. Significance

This work provides a rigorous and practical pathway to calculate oscillator strengths within the $\Delta$SCF framework, a method previously limited by non-orthogonality issues.

• Theoretical Rigor: It leverages the gauge flexibility of Maxwell's equations to bypass the non-orthogonality problem rather than "fixing" the wavefunctions, preserving the physical meaning of the $\Delta$SCF determinants as approximations to the many-body wavefunction.
• Practical Utility: The method is computationally efficient (comparable to ground-state DFT) and requires no additional assumptions or complex orthogonalization schemes.
• Broad Applicability: It is uniquely suited for systems where TDDFT fails (e.g., core excitations, charge transfer) and is the first robust method for calculating oscillator strengths in charged $\Delta$SCF systems.
• Recommendation: The authors conclude that for general $\Delta$SCF applications, especially involving conjugated systems, the combination of the velocity gauge and spin-purified excitation energies offers the most reliable prediction of oscillator strengths.