Tangent equations of motion for nonlinear response functions

This paper introduces a systematic framework based on tangent equations of motion (TEOM) that efficiently computes high-order nonlinear response functions directly from real-time dynamics, overcoming the factorial scaling and numerical instability of traditional methods by isolating perturbative orders through a closed hierarchy of derivative equations.

The Gist

Imagine you are trying to understand how a complex machine, like a car engine or a musical instrument, reacts when you push it.

The Problem: The "Push-and-Subtract" Mess
In physics, scientists study how materials react to light or magnetic fields. Usually, they look at simple reactions (like how a mirror reflects light). But to understand complex, "nonlinear" reactions (like how a material changes color under intense laser light), they need to measure how the system reacts to multiple pushes at once.

Traditionally, to find these complex reactions, scientists have used a clumsy method called "finite differences." Imagine trying to figure out exactly how a cake rises by baking it, then baking it again with a tiny bit more sugar, and subtracting the first cake from the second to see the difference.

• The issue: If you want to know the effect of five different ingredients at once, you have to bake thousands of cakes, subtracting them in a giant, messy math equation. The more ingredients you add, the more the math explodes into a nightmare of errors. It's like trying to hear a single whisper in a hurricane by shouting louder and louder; eventually, the noise drowns out the signal.

The Solution: The "Shadow" Method (TEOM)
The author, Atsushi Ono, introduces a brilliant new way to do this called Tangent Equations of Motion (TEOM).

Think of the physical system (the material) as a ball rolling down a hill.

• The Old Way: To see how the ball reacts to a tiny nudge, you roll the ball, then roll it again with a nudge, then roll it a third time with a different nudge, and compare the paths.
• The New Way (TEOM): Instead of rolling the ball multiple times, you imagine a "Shadow" that follows the ball perfectly.
  • As the ball rolls, the Shadow moves alongside it.
  • If you give the ball a tiny nudge, the Shadow instantly knows exactly how that nudge changes the ball's path.
  • If you give it two nudges, the Shadow has a "Shadow of a Shadow" that tracks the combined effect.
  • If you give it fifty nudges, the Shadow has a "Shadow of a Shadow of a Shadow..." that tracks the 50th-order effect instantly.

Because the Shadow is calculated mathematically alongside the ball, you don't have to subtract messy numbers. You just read the Shadow's position to know the answer. This avoids the "noise" and errors of the old method.

Why is this a Big Deal?

1. It's a Super-Organizer:
   Imagine a library where books are thrown into a pile. The old method was like trying to find a specific book by digging through the pile and hoping you don't lose pages. The TEOM method is like a librarian who sorts every single book into its exact slot as it arrives. You can pull out the "50th-order reaction" book instantly, without digging.

2. It Handles the "Impossible" Orders:
   The author tested this on a classic physics problem (a Duffing oscillator, which is like a spring that gets stiffer the more you stretch it).

   • Using old methods, calculating the reaction to the 5th or 6th "push" was already very hard.
   • Using TEOM, the author calculated the reaction to the 49th order. That's like predicting the outcome of a game of billiards where the balls bounce off each other 49 times in a row, with perfect accuracy.

3. It Works for Everything:
   This isn't just for quantum physics (tiny atoms). It works for classical physics (big springs), electronics, and even complex materials like superconductors. It's a universal tool for understanding how systems react to complex inputs.

The Real-World Impact
Why do we care about the 49th order?

• New Materials: It helps scientists design materials that can convert light into electricity more efficiently or create new types of lasers.
• Better Electronics: It helps us understand how fast computers can switch signals without breaking.
• Medical Imaging: It could improve how we use light to see inside the body without harming tissue.

In a Nutshell
The paper presents a new mathematical "Shadow" technique that lets scientists calculate how complex systems react to intense, multi-layered forces. Instead of doing messy, error-prone subtraction experiments, they run a parallel simulation that tracks the sensitivity of the system in real-time. This allows them to see details that were previously hidden by mathematical noise, opening the door to understanding and designing materials with incredibly complex behaviors.

Technical Summary

1. Problem Statement

Nonlinear response functions (formulated as multipoint correlation functions or Volterra kernels) are essential for understanding complex dynamical and spectroscopic properties in physical systems, ranging from high-harmonic generation in solids to collective modes in superconductors. However, their evaluation faces significant computational hurdles:

• Combinatorial Explosion: Traditional frequency-domain methods (e.g., sum-over-states, diagrammatic techniques) require enumerating time orderings and Feynman diagrams. The number of terms often scales factorially ($n!$) or exponentially with the response order $n$, making high-order calculations (e.g., $n > 5$) prohibitive.
• Numerical Instability in Time-Domain: While real-time evolution methods avoid explicit correlator construction, extracting specific perturbative orders ($n$) from the total nonlinear signal is difficult. Standard approaches rely on finite-difference approximations or subtracting lower-order contributions from simulations with different field amplitudes. As $n$ increases, these methods suffer from severe subtractive cancellation errors and ill-conditioning, limiting accuracy and the accessible response order.
• Gauge and Frequency Issues: In velocity gauge formulations, converting current responses to optical conductivities involves dividing by frequency ($1/\omega^n$), leading to numerical singularities and spurious oscillations near zero frequency.

2. Methodology: Tangent Equations of Motion (TEOM)

The author proposes a systematic framework based on the Gateaux derivative with respect to the external field in function space. Instead of approximating derivatives via finite differences, the method propagates the derivatives themselves as independent dynamical variables.

• Core Concept: If a system state $\rho[f](t)$ evolves under an equation of motion (EOM) driven by a field $f(t)$, the directional functional derivatives of the state with respect to $f$ (denoted as $D_g \rho$) obey their own closed set of evolution equations. These are the Tangent Equations of Motion (TEOM).
• Hierarchical Structure:
  • The TEOM form a closed hierarchy. To compute the $n$-th order response, one propagates the state $\rho$ alongside its mixed directional derivatives up to order $n-1$.
  • The number of auxiliary variables scales as $2^{n-1}$ in the general multivariable case (exponential but significantly better than factorial).
  • In the symmetric case (all perturbation directions identical), the hierarchy collapses to $n$ variables, yielding polynomial complexity ($O(n^2)$ or $O(n^3)$ depending on the nonlinearity).
• Reconstruction Protocol:
  • Infinitesimal Limit: By integrating the TEOM, the method evaluates derivatives in the strict $\epsilon \to 0$ limit, eliminating the need for finite-difference step-size tuning.
  • Frequency Resolution: By employing tailored perturbation fields (cosine/sine or complex exponentials) with specific frequencies and phases, the method reconstructs multivariable frequency-resolved kernels $\bar{\chi}^{(n)}(\omega_1, \dots, \omega_n)$. This allows access to off-diagonal frequency slices, not just harmonic (diagonal) responses.
  • Gauge Stability: The framework computes optical conductivities directly in the time domain, avoiding the $1/\omega$ singularities associated with converting current responses in the frequency domain.

3. Key Contributions

• General EOM Formulation: Derivation of a closed TEOM hierarchy applicable to both linear and nonlinear generators of dynamics (including mean-field, density-matrix, and classical nonlinear ODEs). It handles observables that explicitly depend on the external field (e.g., current in the velocity gauge).
• Order-Resolved Extraction: A protocol to isolate specific perturbative orders without subtractive cancellation, enabling high-precision extraction of high-order responses.
• Term-by-Term Decomposition: The TEOM framework naturally separates physical contributions (e.g., state variation vs. operator variation), allowing for the identification of specific mechanisms like injection currents vs. shift currents.
• Scalability: Demonstrates that while general multivariable scaling is exponential, the symmetric (diagonal) case is polynomial, making very high-order calculations feasible.

4. Numerical Results and Validation

The paper validates the TEOM framework across quantum and classical systems:

• Rice-Mele Model (Ferroelectrics):
  • Successfully computed 2nd and 3rd order optical responses, matching perturbation theory and demonstrating convergence with respect to window functions and field amplitudes.
  • Showed that direct time-evolution calculation of optical conductivity avoids the spurious low-frequency divergences seen in frequency-domain conversions.
• Rice-Mele-Hubbard Model (Mean-Field Dynamics):
  • Applied to interacting systems using Hartree-Fock mean-field dynamics.
  • Validated that TEOM correctly captures interaction-induced spectral features compared to tensor-network (iTEBD) results.
  • Used TEOM decomposition to diagnose that long-lived DC-like components in mean-field approximations arise specifically from the state-variation channel, providing a tool to distinguish physical effects from approximation artifacts.
• 2D Four-Band Topological Model:
  • Computed 5th-order frequency-resolved response functions for a solid-state electron model.
  • Generated 2D frequency maps ($\omega_1, \omega_5$) of the 5th-order kernel, revealing complex resonance structures.
  • Quantitatively reproduced the polarization-angle dependence of 5th-harmonic generation intensity, a stringent test requiring coherent phase sensitivity.
• Classical Duffing Oscillator:
  • Demonstrated the method's ability to reach extremely high orders.
  • Calculated response functions up to the 49th order with relative errors below $10^{-6}$.
  • Confirmed that in the symmetric (diagonal) setting, the computational cost scales polynomially ($O(n^2)$), whereas the general case scales exponentially ($O(4^n)$), both avoiding the factorial scaling of diagrammatic methods.

5. Significance

• Overcoming the "High-Order Barrier": The TEOM framework makes it practically feasible to compute nonlinear response functions of orders ($n=5, 10, \dots, 49$) that were previously inaccessible due to combinatorial or numerical constraints.
• Unified Framework: It provides a single, systematic approach applicable to diverse systems: quantum many-body systems (via tensor networks, DMRG, or mean-field), open quantum systems, and classical nonlinear dynamics.
• Physical Insight: By providing a term-by-term decomposition of the response, it offers new avenues for interpreting complex nonlinear phenomena (e.g., distinguishing between different current mechanisms in topological materials).
• Numerical Robustness: The method eliminates the numerical instability of finite-difference subtraction and the singularities of frequency-domain gauge conversions, offering a stable route to high-precision spectroscopy and transport calculations.

In summary, this work establishes Tangent Equations of Motion as a powerful, general, and numerically stable tool for computing high-order nonlinear response functions, bridging the gap between real-time dynamics and frequency-resolved spectroscopy.