Bootstrapping transport in the Drude-Kadanoff-Martin model

This paper establishes sharp constraints on the parameters of the Drude-Kadanoff-Martin model by deriving upper bounds on the charge density retarded Green's function, demonstrating that the model fails at microscopic scales and proving a Mott-Ioffe-Regel-type bound that forbids conventional Drude peaks in systems where the collective mean free path is significantly shorter than the lattice spacing.

The Gist

Imagine you are trying to understand how electricity flows through a metal, or how heat moves through a wall. In physics, we often use a simple, smooth model to describe this flow, kind of like how we might describe traffic on a highway as a smooth river of cars. This specific paper focuses on a famous, simple model called the Drude-Kadanoff-Martin (DKM) model. It treats electricity like a fluid that slows down due to friction (relaxation) and spreads out (diffusion).

However, the real world isn't a smooth river; it's a bumpy, pixelated landscape made of atoms (a lattice). The authors of this paper ask a crucial question: How far can this smooth, simple model actually go before it breaks down?

To answer this, they use a clever mathematical strategy called "bootstrapping." Think of it like this: Imagine you are trying to guess the shape of a hidden object by only looking at its shadow. You know certain rules about how shadows must behave (they can't be infinitely wide, they can't appear out of nowhere). By knowing the rules of the "shadow" (the math of the model) and the rules of the "object" (the real atomic world), you can figure out strict limits on what the object can look like.

Here is the breakdown of their findings using everyday analogies:

1. The "Smooth River" vs. The "Pixelated World"

The DKM model is like a smooth, continuous river. But the actual material is like a grid of stepping stones (a lattice).

• The Problem: The smooth river model predicts that if you look at very high speeds (high frequencies), the flow just tapers off slowly, like a gentle slope.
• The Reality: In a real atomic grid, if you try to move things too fast, the "pixels" of the grid stop you. The flow doesn't just taper off; it gets crushed and disappears exponentially fast (like a light being switched off instantly).
• The Conclusion: The smooth river model cannot describe the world at very high speeds. It breaks down before it reaches the speed of the atomic grid. The authors prove that the model must stop working at a specific energy level, otherwise, it would violate the fundamental rules of the atomic grid.

2. The "Speed Limit" of the Mean Free Path

The paper focuses on a specific measurement called the mean free path ($\ell$). Imagine a pinball machine. The "mean free path" is the average distance a ball travels before hitting a bumper.

• The Old Rule: Physicists have long suspected that a ball can't travel a distance shorter than the size of the bumper itself. If the ball hits a bumper every inch, but the bumpers are 10 inches apart, the model is broken. This is known as the Mott-Ioffe-Regel (MIR) bound.
• The New Proof: The authors use their "shadow" method to prove this rule mathematically. They show that if the "smooth river" model (DKM) is to work, the pinball must travel a distance at least as long as the size of the atomic "bumpers" (the lattice spacing).
• The Catch: If a material is so "bad" at conducting electricity that the pinball hits a bumper more often than the bumpers are spaced apart (a mean free path shorter than the lattice), then the smooth river model cannot exist for that material. The material isn't a "metal" in the traditional sense; it's something else entirely (like an insulator or a "bad metal").

3. The "Bad Metal" Paradox

There are materials called "bad metals" where electricity seems to flow very poorly, and the pinball seems to bounce around chaotically, hitting things faster than the atomic spacing allows.

• The Paper's Verdict: The authors say, "If you see a 'bad metal' where the pinball is bouncing faster than the grid allows, you cannot use the standard smooth river model to describe it."
• Why it matters: This confirms that these strange materials are doing something fundamentally different. They aren't just "normal metals that are dirty"; they are operating under different rules where the simple idea of a "particle traveling a distance" stops making sense.

4. The "Bootstrap" Method

How did they prove this without solving every single atom in the universe?

• They used a technique borrowed from particle physics. They assumed the "smooth river" model is true for slow, low-energy movements.
• Then, they looked at the "high-energy" rules (the atomic grid) which say, "You can't have infinite energy, and you can't move faster than the grid allows."
• By forcing the "smooth river" to respect the "atomic grid" rules, they found that the river's parameters (like how fast it flows or how far it goes) are trapped in a cage. The cage is the MIR bound. If the parameters try to escape the cage (by getting too short), the model collapses.

Summary

In simple terms, this paper proves that you cannot have a standard, smooth flow of electricity if the particles are bouncing around so fast that they hit obstacles more often than the obstacles are spaced apart.

If you see a material where the "bouncing" is that chaotic, the standard textbook description of electricity (the Drude model) is wrong. The material is likely an insulator or a "bad metal" that requires a completely new way of thinking. The authors didn't just guess this; they used strict mathematical "shadow" rules to prove that the standard model simply cannot exist in those extreme conditions.

Technical Summary

Technical Summary: Bootstrapping Transport in the Drude-Kadanoff-Martin Model

Problem Statement
The paper addresses the challenge of connecting low-energy transport phenomena to the microscopic "high-energy" physics of underlying lattice models, particularly in regimes beyond weak interactions where standard Boltzmann transport theory fails. While fundamental bounds on transport coefficients (such as the Mott-Ioffe-Regel (MIR) bound on mean free paths or Planckian dissipation bounds) are widely discussed, rigorous proofs based on general physical principles remain elusive. The authors aim to derive sharp, quantitative constraints on the parameters of effective low-energy transport theories by treating them as "bootstrap" problems, where the consistency of a low-energy effective description with a specific class of microscopic completions (local, bounded lattice Hamiltonians) imposes strict limits on the effective parameters.

Methodology
The authors adapt the "bootstrap" logic traditionally used in scattering amplitudes (specifically the S-matrix bootstrap) to the context of charge transport. The core methodology involves:

1. Defining the Object of Study: The retarded Green's function for charge density, $G_R(\omega, \mathbf{k})$, is treated analogously to partial waves in scattering theory.
2. Establishing Microscopic Bounds: The authors derive two upper bounds on the integrated spectral weight, $\int \frac{d\Omega}{\pi} \int_\Sigma \frac{d^dk}{(2\pi)^d} \frac{\text{Im } G_R(\Omega, \mathbf{k})}{\Omega}$, over a sub-region of the Brillouin zone and a frequency range $[\omega, \Lambda]$.
   • Bound 1: Derived using positivity of the spectral function and thermal expectation values, analogous to the $f$-sum rule but restricted to a finite frequency window.
   • Bound 2: Derived using the locality and boundedness of the microscopic Hamiltonian. This bound relies on operator growth arguments (commutators of the Hamiltonian with the charge density) and exhibits an exponential suppression at high frequencies ($\omega \gg \varepsilon$, where $\varepsilon$ is the characteristic energy scale of local interactions).
3. Applying to an Effective Model: These bounds are applied to the Drude-Kadanoff-Martin (DKM) model, a minimal effective theory for charge transport characterized by charge compressibility ($\chi$), diffusivity ($D$), and a current relaxation time ($\tau$). The DKM model predicts a specific form for the spectral weight that decays as a power law at high frequencies.
4. Comparing Scales: The authors compare the power-law decay of the DKM spectral weight against the exponential decay required by the microscopic Bound 2. This comparison forces the effective theory to have a cutoff $\Lambda$ below the microscopic energy scale $\varepsilon$.
5. Deriving Constraints: By evaluating the bounds within the DKM framework (assuming the Drude peak is fully contained within the low-energy description), the authors derive inequalities relating the collective mean free path $\ell \equiv \sqrt{D\tau}$ to the microscopic lattice spacing $a$.

Key Contributions and Results

• Inconsistency at Microscopic Scales: The paper proves that the DKM model cannot be a valid description at microscopic energy scales. The exponential suppression of spectral weight required by local lattice models (Bound 2) is incompatible with the power-law falloff of the DKM model. Consequently, the DKM model must break down at frequencies $\Lambda \lesssim \varepsilon$.
• Lower Bound on Collective Mean Free Path: Under the assumption that the DKM model describes the low-energy physics up to a cutoff $\Lambda \lesssim \varepsilon$, the authors derive a lower bound on the collective mean free path $\ell$ in terms of the lattice spacing $a$. The bound takes the form:
  $$\beta_d \frac{1 - e^{-1/(T\tau)}}{1} \leq \left(\frac{\ell}{a}\right)^d \frac{\tau}{\chi a^d \langle n_0^2 \rangle_T}$$
  where $\beta_d$ is a numerical prefactor.
• Recovery of Mott-Ioffe-Regel (MIR) Bound: In various physical regimes, this derived inequality implies a version of the MIR bound ($\ell \gtrsim a$):
  • Low Temperature (Disordered): For systems with short mean free paths (violating $\ell \gtrsim a$), the DKM model cannot support a conventional Drude peak. This is consistent with such systems being insulators rather than metals.
  • Fermi Liquids & Planckian Scattering: In regimes where $\ell \gg a$ (Fermi liquids) or $\ell \sim a$ (Planckian metals up to $T \sim E_F$), the bounds are satisfied, and conventional Drude peaks are consistent with the effective description.
  • High Temperature "Bad Metals": The authors show that "bad metals" where the mean free path decreases below the lattice spacing ($\ell < a$) at high temperatures are incompatible with a conventional Drude peak within the DKM framework. This suggests that such systems must exhibit non-Drude spectral weight redistribution (e.g., Gaussian peaks or multi-scale relaxation) rather than a single relaxation timescale.
• Extension to Imaginary Frequencies: The paper outlines a method to derive bounds using the Green's function on the imaginary frequency axis, $G_R(i\alpha, \mathbf{k})$, which avoids integrals over real frequencies and may be more robust for complex effective theories.

Significance and Claims
The authors claim that their work provides a rigorous, "bootstrap-style" derivation of transport bounds based on general principles of locality and bounded interactions in lattice models.

• Precision: Unlike phenomenological arguments, the derived bounds are numerically precise subject to the specific assumption that the DKM model holds over a defined range of frequencies and wavelengths.
• Clarification of "Bad Metals": The results offer a theoretical explanation for why "bad metals" (systems violating MIR) do not exhibit standard Drude peaks; the violation of the derived bound implies the breakdown of the single-timescale relaxation assumption inherent in the DKM model.
• Methodological Shift: The paper demonstrates that the bootstrap program, typically applied to high-energy scattering amplitudes, can be successfully adapted to condensed matter transport problems by utilizing the analyticity and positivity properties of retarded Green's functions.
• Limitations: The authors explicitly state that their approach does not prove that a specific microscopic Hamiltonian generates the DKM model; rather, it determines what constraints must be satisfied if the DKM model is the correct effective description. They also note that many real materials exhibit multiple timescales, requiring extensions beyond the simple DKM model, which they discuss as a future direction.