Geometric theory of constrained Schrödinger dynamics with application to time-dependent density-functional theory on a finite lattice

This paper establishes a general geometric framework for constrained Schrödinger dynamics to revisit the mathematical foundations of time-dependent density-functional theory (TDDFT) on finite lattices, revealing a novel, purely geometric evolution equation that leads to new Kohn–Sham schemes enforced by imaginary potentials or nonlocal Hermitian operators.

The Gist

Imagine you are trying to guide a very fast, chaotic dancer (an electron) across a stage. In the world of quantum physics, this dancer follows strict rules called the Schrödinger equation. Usually, the dancer moves freely, responding only to the music (the external forces or electric fields).

However, scientists often need to force this dancer to follow a specific path or keep a specific formation (like keeping the "density" of electrons in a certain spot). This is the challenge of Time-Dependent Density-Functional Theory (TDDFT). It's the tool we use to predict how materials react to light, electricity, or chemical changes.

The problem? The current mathematical rules for forcing the dancer to stay on the path are a bit shaky, especially when things move very fast or change suddenly.

This paper proposes a new way of thinking about how to guide this dancer. Instead of just "pushing" the dancer with a stick, the authors suggest looking at the geometry of the dance floor itself.

Here is the breakdown of their new ideas using simple analogies:

1. The Two Old Ways to Guide the Dancer

The paper compares two existing methods for forcing the electron to stay on its path:

• The "Variational" Method (The Traditional Approach):

  • The Analogy: Imagine the dancer is trying to minimize the "effort" or "friction" of their movement while staying on the path. This is like finding the smoothest, most efficient route.
  • The Flaw: This method works great for slow, gentle movements. But if you try to make the dancer spin wildly or change direction instantly, the math breaks down. It's like trying to drive a car with a broken steering wheel; you can't make sharp turns without crashing. In physics terms, it fails when the electron density changes too quickly (non-adiabatic situations).

• The "Geometric" Method (The New Approach):

  • The Analogy: Imagine the dancer is on a curved surface (a manifold). Instead of worrying about "effort," we simply ask: "What is the closest possible move the dancer can make to their natural instinct, without stepping off the path?"
  • The Magic: This method projects the dancer's natural movement onto the allowed path. It's like a magnetic rail that gently nudges the dancer back onto the track if they drift, but allows them to move freely along the track.
  • The Benefit: This method is much more robust. It can handle wild, fast changes that the old method couldn't. It's like having a high-tech guide rail that works even on a rollercoaster.

2. The "Imaginary" Potentials (The Secret Sauce)

In the traditional method, to keep the electron on track, scientists add a "real" force (like a real electric field).

In the new Geometric Method, the math reveals something strange: to keep the electron on the track, you have to add a force that looks "imaginary" (in the mathematical sense, not the "fake" sense).

• The Analogy: Think of it like a video game character. If the character tries to walk off a cliff, the game code doesn't push them back with a physical wall. Instead, it instantly teleports them back or changes their speed in a way that feels like a "ghost" force.
• Why it's cool: This "imaginary" force isn't a real physical field you can measure with a voltmeter. It's a mathematical trick that ensures the electron stays in the right place. The authors show that this trick is actually a very stable, mathematically sound way to describe reality.

3. The "Oblique" Principle (The Best of Both Worlds)

The authors also introduce a middle ground called the Oblique Principle.

• The Analogy: Imagine you are walking down a hallway.
  • The Variational method says: "Walk straight, but if you hit a wall, slide along it."
  • The Geometric method says: "Walk straight, but if you hit a wall, bounce off it at a perfect angle."
  • The Oblique method says: "You can choose any angle you want between sliding and bouncing."
• This gives scientists a "dial" to tune their simulations. If they are in a calm situation, they can turn the dial to the traditional method. If things get chaotic, they can turn it to the geometric method.

4. The Test: The "Hubbard Dimer"

To prove their theory works, the authors tested it on a tiny, simple model called the Hubbard Dimer.

• The Analogy: This is like testing a new car engine on a small, controlled track before putting it on a highway. It's a system with just two "sites" (like two rooms) where electrons can hop back and forth.
• The Result: When they simulated a situation where the electrons had to move very fast (a "non-adiabatic" event, like a sudden charge transfer), the old method failed or gave weird results. The new Geometric Method worked perfectly, keeping the electrons exactly where they were supposed to be, even during the chaos.

Why Does This Matter?

Currently, when scientists simulate how molecules react to light (like in solar cells or photosynthesis), they often use approximations that fail when things happen too fast. This paper suggests that by using this geometric view, we can build better, more accurate simulations.

In a nutshell:
The authors realized that the old way of forcing electrons to follow rules was like trying to herd cats with a stick. Their new method is like building a custom-shaped track that the cats naturally want to stay on. It's mathematically cleaner, handles fast changes better, and opens the door to understanding complex quantum behaviors that were previously too difficult to calculate.

Technical Summary

1. Problem Statement

Time-Dependent Density-Functional Theory (TDDFT) is a cornerstone for studying electronic dynamics, yet its mathematical foundations remain less rigorous than those of ground-state DFT.

• Core Issue: The standard formulation of TDDFT relies on the Runge–Gross theorem, which assumes the time-analyticity of the external potential and wavefunction. This assumption fails for singular potentials (like Coulomb) and limits the mathematical robustness of the theory.
• Specific Challenge: Enforcing time-dependent constraints (e.g., a prescribed electron density $\rho(t)$) on the Schrödinger evolution is not unique. The standard approach uses a Variational Principle (stationarity of the action), but for commuting observables (like density operators), this leads to existence and uniqueness issues, particularly requiring specific initial conditions and failing for rapidly changing densities.
• Goal: To revisit the foundations of TDDFT within a finite-dimensional setting by developing a general geometric framework for Schrödinger dynamics subject to prescribed expectation values. The authors aim to identify alternative, mathematically robust definitions of constrained dynamics that do not rely on the restrictive assumptions of the standard variational approach.

2. Methodology

The authors develop a general geometric theory for constrained quantum dynamics on a manifold $\mathcal{M}$ defined by fixed expectation values of observables $\langle \psi, O_m \psi \rangle = o_m(t)$.

A. Geometric Framework

The state space is treated as a real manifold embedded in $\mathbb{C}^d$. The authors distinguish between the tangent space $T_\psi$ (directions preserving constraints) and the normal space $N_\psi$ (directions changing constraints).
They propose three distinct principles to define the constrained evolution $i\partial_t \psi = (H + F)\psi$:

1. Variational Principle (VP):

   • Based on the stationarity of the action functional (Dirac-Frenkel principle).
   • Requires the residual of the Schrödinger equation to be orthogonal to the tangent space in the real inner product sense: $\text{Re}\langle h, i\partial_t \psi - H\psi \rangle = 0$.
   • Result: Leads to a modified Hamiltonian with real Lagrange multipliers (potentials) $v_m(t)$.
   • Limitation: For commuting observables (standard TDDFT), this requires the invertibility of a matrix $K_\Psi$ involving double commutators $[O_n, [H, O_m]]$. This imposes strict constraints on the initial state and the rate of change of the density (velocity constraints), often making the problem ill-posed for fast dynamics.

2. Geometric Principle (GP):

   • Based purely on geometry: The time derivative $\partial_t \psi$ is the orthogonal projection of the unconstrained velocity $-iH\psi$ onto the tangent space $T_\psi$.
   • Condition: $\partial_t \psi + iH\psi \in N_\psi$.
   • Result: Leads to a modified Hamiltonian with imaginary potentials $w_m(t)$ (or equivalently, a nonlocal Hermitian operator).
   • Advantage: Requires only the invertibility of the overlap matrix $S_\Psi$ (related to the linear independence of constraints). It does not impose velocity constraints on the density, allowing for the representation of arbitrarily fast time-dependent densities.

3. Oblique Principle (OP):

   • An interpolation between VP and GP using an angle parameter $\theta$.
   • As $\theta \to 0$, it recovers the Variational Principle (often singularly). For $\theta \neq 0$, it behaves like the Geometric Principle.

B. Application to Finite Lattices

The theory is applied to $N$ interacting fermions on a finite lattice (e.g., Hubbard model).

• Observables: Site density operators $N_m$.
• Kohn-Sham (KS) Schemes: The authors derive new KS equations where the exact interacting density is reproduced by non-interacting particles.
  • Standard TDKS: Uses the VP (real potential $v_{Hxc}$).
  • Geometric TDKS (TDGKS): Uses the GP. The correction term appears as a nonlocal Hermitian operator (commutator form) or an imaginary local potential $w(t)$.
  • Hybrid Schemes: The authors propose correcting standard adiabatic approximations with a geometric term to capture non-adiabatic effects.

3. Key Contributions

1. Unified Geometric Theory: The paper establishes a rigorous geometric framework distinguishing between the Variational and Geometric principles for constrained Schrödinger dynamics, clarifying why they differ for non-Kähler manifolds (like those defined by density constraints).
2. New TDDFT Formulation: It introduces a Geometric TDDFT formulation that is mathematically more robust than the standard variational approach. It proves that for commuting observables, the geometric principle guarantees existence and uniqueness of the potential under much weaker conditions (only requiring the invertibility of the overlap matrix $S_\Psi$).
3. Novel Kohn-Sham Equations:
   • Derivation of the Time-Dependent Geometric Kohn-Sham (TDGKS) equation.
   • Identification that the geometric correction term acts as a nonlocal Hermitian operator (rank-2N) or an imaginary potential that acts as a "source/sink" for probability amplitude to enforce density constraints, while preserving the norm.
4. Resolution of v-Representability Issues: The geometric approach resolves the "velocity constraints" problem inherent in standard TDDFT for commuting observables, showing that any smooth density trajectory is representable without the restrictive conditions required by the van Leeuwen equation.

4. Numerical Results (Hubbard Dimer)

The authors tested these theories on the Hubbard dimer (2 sites, 2 electrons) under time-dependent electric fields.

• Off-Resonant (Adiabatic) Regime: Both standard (TDKS) and geometric (TDGKS) approaches perform well. The non-adiabatic corrections are small for both.
• Resonant (Non-Adiabatic) Regime:
  • Standard TDKS: The exact potential $v_{Hxc}$ exhibits large, rapid oscillations and steps to reproduce the density, reflecting strong non-adiabatic correlation effects. The adiabatic approximation fails significantly.
  • Geometric TDKS: The geometric potential $w(t)$ (or the nonlocal operator) successfully reproduces the exact density. The structure of the geometric potential is qualitatively different from the standard potential; it does not show the same singular steps but rather smooth, oscillatory patterns.
  • Hybrid Correction: Adding a geometric correction term to an adiabatic approximation (TDeaKS+G) significantly improves the description of non-adiabatic charge transfer, outperforming the standard adiabatic approximation.
• Asymmetric Dimer (Charge Transfer): In a long-range charge transfer scenario, the standard adiabatic approximation completely fails (cannot transfer charge). The geometric approach successfully models the transfer, demonstrating its utility for strongly non-adiabatic processes.

5. Significance and Conclusion

• Mathematical Robustness: The geometric principle provides a mathematically sound alternative to the variational principle for TDDFT, particularly for systems with commuting observables where the standard approach is fragile.
• New Approximation Strategies: The identification of the geometric correction as a nonlocal Hermitian operator (or imaginary potential) opens a new avenue for constructing TDDFT approximations. Instead of trying to approximate the complex, singular non-adiabatic real potential, one might approximate the smoother geometric term.
• Beyond Adiabaticity: The results suggest that geometric TDDFT is naturally suited for describing highly non-adiabatic phenomena (like charge transfer and Rabi oscillations) where standard adiabatic approximations break down.
• Future Outlook: The authors note that this framework can be extended to continuous systems (companion paper) and mixed states, potentially leading to a new generation of TDDFT functionals that are more stable and accurate for dynamic electronic processes.

In summary, this work fundamentally rethinks the constraints of TDDFT through a geometric lens, offering a more flexible and robust mathematical foundation that yields new, effective computational schemes for simulating non-adiabatic quantum dynamics.