Harnessing dressed time-dependent density functional theory for the non-perturbative regime: Electron dynamics with double excitations

This paper demonstrates that response-reformulated time-dependent density functional theory (RR-TDDFT) enables the use of frequency-dependent kernels, originally developed for the perturbative regime to capture double excitations, to accurately model non-perturbative strong-field electron dynamics.

The Gist

The Big Picture: Predicting the Unpredictable

Imagine you are trying to predict how a complex machine (like a car engine or a musical instrument) will behave when you hit it with a massive, sudden force. In the world of atoms and electrons, this is called non-perturbative dynamics. It's when you hit the system so hard with a laser that it doesn't just wiggle a little; it gets completely shaken up, changing its state entirely.

For decades, scientists have used a powerful tool called Time-Dependent Density Functional Theory (TDDFT) to simulate these electron movements. Think of TDDFT as a "weather forecast" for electrons. It's usually very good at predicting light breezes (small changes) and sunny days (simple excitations).

The Problem:
However, when the "storm" hits (strong lasers), the standard forecast breaks down. It fails to predict specific, weird weather patterns called "double excitations."

• The Analogy: Imagine a trampoline. Usually, you can jump on it (a single jump). But sometimes, you need to jump and twist at the same time (a double jump). Standard TDDFT is like a trampoline that only knows how to handle straight jumps. If you try to do a twist, the math gets confused, and the prediction becomes nonsense.

The Old Solution vs. The New Solution

The Old Way (The "Adiabatic" Trap):
Traditionally, scientists tried to fix the forecast by using a "snapshot" approach. They looked at the electron system in its calm, resting state and assumed it would behave the same way even when it was being hammered by a laser.

• The Metaphor: It's like trying to predict how a car will drive in a hurricane by looking at how it drives on a flat parking lot. It works for slow speeds, but when the wind blows hard, the car flips over, and your parking-lot prediction is useless.

The New Breakthrough (RR-TDDFT + Dressed TDDFT):
This paper introduces a clever new strategy that combines two advanced ideas to fix the forecast.

1. The "Dressed" Kernel (The Specialized Tool):
   Scientists previously developed a special tool called "Dressed TDDFT" specifically for predicting those tricky "double jumps" (double excitations).

   • The Metaphor: Think of this as a specialized mechanic who is an expert at fixing twisted trampoline springs. They know exactly how the double-jump works. However, this mechanic usually only works in a quiet workshop (the "linear response" regime). They haven't been tested in a hurricane.

2. RR-TDDFT (The New Framework):
   The authors used a new framework called Response-Reformulated TDDFT (RR-TDDFT). Instead of trying to calculate the electron's path step-by-step through the chaos (which is where the old method fails), RR-TDDFT calculates the ingredients needed to build the solution first.

   • The Metaphor: Instead of trying to drive the car through the storm blindly, RR-TDDFT builds a detailed blueprint of the car's parts before the storm hits. It uses the "workshop" data (where the specialized mechanic is most accurate) to build a model that can survive the hurricane.

How They Combined Them

The genius of this paper is taking that "Specialized Mechanic" (the Dressed TDDFT tool) and putting them to work inside the "Blueprint Framework" (RR-TDDFT).

• The Result: They successfully simulated electrons being hit by strong lasers, accurately capturing the "double jumps" that previous methods missed.
• The Proof: They tested this on a simple model system (two electrons in a 1D box).
  • Old Method: Predicted the electrons would just wiggle back and forth.
  • New Method: Correctly predicted that the electrons would perform complex, synchronized "dance moves" (Rabi oscillations) involving those double excitations, matching the exact physics perfectly.

Why This Matters

This isn't just about fixing one specific calculation. It's a paradigm shift.

• The "Best of Both Worlds": For years, scientists had great tools for quiet, linear problems (Response Regime) and terrible tools for chaotic, strong-field problems (Non-perturbative Regime).
• The Bridge: This paper builds a bridge. It shows that you can take the sophisticated, accurate tools developed for the "quiet workshop" and use them to solve the "chaotic storm" problems, provided you use the RR-TDDFT framework.

Summary in One Sentence

The authors figured out how to take a highly accurate tool designed for calm, predictable electron behavior and successfully use it to predict the wild, chaotic dance of electrons during intense laser storms, specifically solving the long-standing mystery of "double excitations."

The Takeaway: We no longer have to choose between accuracy and handling strong forces. By changing how we ask the question (using RR-TDDFT), we can use our best answers to solve our hardest problems.

Technical Summary

1. The Problem

Time-Dependent Density Functional Theory (TDDFT) is the standard method for calculating electronic spectra and non-perturbative dynamics. However, it suffers from well-known failures, particularly in the non-perturbative (strong-field) regime.

• Adiabatic Approximation Failure: Standard TDDFT relies on the adiabatic approximation, where the exchange-correlation (xc) potential depends only on the instantaneous density. This neglects memory dependence (history of the density and initial states), leading to unphysical results in strong fields, such as incorrect Rabi oscillations and spurious peak shifting.
• The Double-Excitation Challenge: A specific and severe failure of adiabatic TDDFT is its inability to describe double excitations (states where two electrons are excited simultaneously). In the linear response regime, this manifests as missing excitation energies. In the non-perturbative regime, while standard TDDFT can technically access these states, the description is unreliable and unphysically dependent on field parameters.
• The Gap: While "dressed" TDDFT (DTDDFT) has successfully captured double excitations in the linear response regime using frequency-dependent xc kernels, these developments had not been extended to fully non-linear, non-perturbative dynamics.

2. Methodology

The authors propose a hybrid framework that bridges the gap between response theory and real-time dynamics by combining Response-Reformulated TDDFT (RR-TDDFT) with Dressed TDDFT (DTDDFT).

A. Response-Reformulated TDDFT (RR-TDDFT)

Instead of solving the time-dependent Kohn-Sham (TDKS) equations for orbitals, RR-TDDFT solves a system of ordinary differential equations (ODEs) for the coefficients ($C_m(t)$) of the many-body wavefunction expanded in unperturbed states ($\Psi_m$).

• Key Advantage: The xc functional is only required in the ground state and response regimes (linear and quadratic), not in the fully non-equilibrium domain. This allows the use of highly accurate response functionals that would otherwise fail in the TDKS equation.
• Inputs Required:
  • Excitation energies ($E_m$).
  • Transition densities ($\rho_{mn}$) between ground and excited states, and between excited states.
  • State densities ($\rho_{mm}$).

B. Dressed TDDFT (DTDDFT) Kernel

To address the double-excitation problem, the authors utilize a frequency-dependent xc kernel derived in the response regime.

• Small-Matrix Approximation (SMA): They employ a "dressed" SMA (DSMA) where a single excitation ($q$) mixes with a double excitation ($d$).
• The Kernel: The xc kernel includes a frequency-dependent term that "dresses" the single excitation, effectively folding the double excitation into the single excitation manifold. This generates corrected excitation energies and redistributes transition densities to account for the mixed single/double character.
• Approximation for Couplings: Since a full "dressed" quadratic response (for excited-state to excited-state couplings) is not yet fully developed, the authors approximate transition densities between excited states using auxiliary wavefunctions weighted by the eigenvectors of the response matrix.

C. The Combined Approach

The authors integrate the DSMA/AEXX (Adiabatic Exact Exchange) functional into the RR-TDDFT framework.

1. Calculate ground-state and response properties (energies, densities) using the dressed kernel.
2. Feed these properties into the RR-TDDFT ODEs.
3. Propagate the system in real-time under strong external fields.

3. Key Contributions

• Extension of DTDDFT: The paper successfully extends the application of dressed TDDFT (previously limited to linear response) to fully non-perturbative dynamics.
• Validation of RR-TDDFT: It demonstrates that RR-TDDFT can exploit high-accuracy response functionals (developed for linear response) to solve non-equilibrium problems, bypassing the need for non-equilibrium xc potentials.
• Accurate Double-Excitation Dynamics: The method accurately captures the dynamics of systems involving double excitations, a regime where standard adiabatic TDDFT fails completely.
• Generalizability: The work establishes a general protocol: any functional developed in the response regime can be utilized for non-equilibrium dynamics via RR-TDDFT.

4. Results

The method was tested on a 1D model system of two soft-Coulomb interacting electrons in an anharmonic oscillator.

• Rabi Oscillations (Single State):

  • When driving the system at the frequency of a double-excitation character state, standard TDKS with adiabatic functionals failed to capture Rabi oscillations.
  • RR-TDDFT with standard adiabatic functionals showed oscillations but with incorrect densities (lacking the double-excitation component).
  • RR-TDDFT with DSMA/AEXX accurately reproduced both the dipole oscillations and the electron density at the half-Rabi period, matching the exact Time-Dependent Schrödinger Equation (TDSE) solution.

• Multi-State Coherent Superposition (Pulse 1):

  • A non-resonant pulse drove a superposition of four states.
  • Standard TDKS and RR-TDDFT (without dressing) failed to capture the "beating" pattern and overestimated dipole magnitudes.
  • RR-TDDFT with DSMA accurately captured the magnitude and beating pattern of the dipole moment, confirming that the error was due to the lack of double-excitation character in the functional, not the truncation of the Hilbert space.

• Strong Population Transfer (Pulse 2):

  • A pulse near the first excited state induced significant population in multiple states.
  • Standard TDKS failed to capture the decay of dipole oscillations (free evolution of the superposition).
  • RR-TDDFT with DSMA accurately tracked the population transfer and dipole dynamics, showing close agreement with exact TDSE results.

5. Significance

This work represents a significant step forward in computational quantum chemistry and physics:

1. Solving the Memory Problem: It circumvents the difficulty of constructing memory-dependent functionals for the non-equilibrium domain by shifting the complexity to the response regime, where such functionals are better understood and more successful.
2. Reliability for Strong Fields: It provides a reliable, computationally efficient method for simulating strong-field electron dynamics involving complex electronic states (like double excitations) that are currently inaccessible to standard TDDFT.
3. Future Applications: The framework opens the door for applying a wide range of advanced response-functionals (e.g., for charge transfer, dissipation, and double excitations in real molecules) to non-equilibrium phenomena, potentially revolutionizing the simulation of ultrafast spectroscopy and strong-field physics.

In conclusion, the paper demonstrates that by decoupling the dynamics solver (RR-TDDFT) from the xc potential evaluation (using response theory), one can harness the accuracy of advanced response functionals to solve non-perturbative problems with high fidelity.