Diffusive hydrodynamics of hard rods from microscopics

✨

This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine a crowded hallway filled with people trying to walk in different directions. In physics, we call this a "many-body system." Usually, when we try to predict how this crowd moves, we use a set of rules called hydrodynamics. Think of these rules like a weather forecast: instead of tracking every single person's step, we just look at the "density" of the crowd and the "wind" (momentum) to predict if the crowd will move left or right.

For a long time, scientists believed they had the perfect forecast for a specific type of crowd: Hard Rods. These are like rigid sticks or spheres that can't pass through each other. They bounce off one another perfectly.

The standard forecast (called Navier-Stokes hydrodynamics) said: "If the crowd is moving, it will eventually smooth out and settle down, just like heat spreading out in a room. This process is irreversible; you can't un-mix the crowd."

The Big Surprise
This paper, written by a team of physicists, says: "Hold on. The standard forecast is wrong for the long run."

They discovered that the crowd doesn't just smooth out. Instead, the people in the crowd start developing a secret, long-distance "telepathy" or long-range correlation.

The Analogy: The "Echo" in the Hallway

Here is a simple way to understand what they found:

The Old View (The Standard Forecast):
Imagine you are in a hallway. You bump into someone, they bump into someone else, and eventually, the whole line shuffles forward. The old theory assumed that once you bump into your neighbor, you forget everything about the person three spots down the line. The crowd acts like a fluid where only immediate neighbors matter. This leads to "diffusion," where the crowd slowly spreads out and loses its structure.
The New View (The Hard Rod Discovery):
The authors found that because these "rods" are rigid and interact in a very specific way, a bump here causes a ripple that travels all the way down the line.
- The Metaphor: Imagine a line of people holding a very long, taut rope. If Person A pulls the rope, Person Z (who is far away) feels it instantly. Even if they are far apart, they are "correlated."
- In the hard rod gas, if a group of rods is crowded in one spot, it changes how a rod miles away will move in the future. They aren't just bumping into neighbors; they are "talking" to each other across the whole system.

Why This Changes Everything

The paper reveals two major shifts in how we understand this physics:

1. You Need Two Maps, Not One
The old theory used one map: "Where is the crowd density?"
The new theory requires two maps that talk to each other:

Map A: Where the crowd is right now.
Map B: How the crowd members are "connected" or correlated across long distances.
You can't predict the future movement of the crowd using just Map A. You must know the secret connections (Map B) to get the answer right.

2. The "Time Machine" Effect (Reversibility)
The old theory said time has an arrow: you can mix milk into coffee, but you can't un-mix it. This is because the old theory assumed the crowd "forgets" its past as it spreads out.
The new theory says: Time is reversible.
Because the crowd members remember their long-distance connections, if you played the movie of the crowd moving backward, the physics would still work perfectly. The "entropy" (disorder) doesn't have to keep increasing. The system doesn't necessarily "thermalize" (settle into a boring, uniform state) in the way we thought. It can stay in a complex, structured state forever.

The "Jump" in the Math

The authors found a weird mathematical "glitch" that saves the day.

The old theory predicted a certain amount of "spreading" (diffusion).
The new theory found that the long-distance connections create a "jump" in the math.
The Magic: This jump exactly cancels out the spreading predicted by the old theory!
The Result: The remaining spreading is driven only by the new, subtle long-distance connections. It's a completely different mechanism than anyone expected.

Why Should You Care?

This isn't just about sticks in a hallway.

Integrable Systems: This applies to a whole class of special physical systems (like certain quantum computers or super-cold atoms) that are "perfectly" predictable in theory but behave strangely in reality.
The Limits of Thermodynamics: It challenges our understanding of how things reach equilibrium. It suggests that in some perfect systems, things might never truly "settle down" into a boring, uniform state because they keep "remembering" their past interactions.

In a Nutshell:
Scientists thought they had the ultimate guide for how rigid particles move. They realized they missed a hidden layer of "long-distance friendship" between the particles. Because of this friendship, the crowd doesn't just spread out and forget; it remembers, it reverses, and it follows a much more complex, beautiful set of rules than we ever imagined.

1. Problem Statement

The paper addresses a fundamental gap in the theoretical understanding of non-equilibrium statistical mechanics: the derivation of diffusive hydrodynamic equations from microscopic dynamics in integrable systems.

Context: Generalized Hydrodynamics (GHD) successfully describes the Euler-scale (ballistic) dynamics of integrable 1D systems like the hard rod gas. However, the standard extension to the diffusive scale (Navier-Stokes GHD) relies on the assumption that the system remains in a local Generalized Gibbs Ensemble (L-GGE) at all times, implying that correlations decay rapidly and entropy increases monotonically.
The Conflict: Recent observations in integrable systems (e.g., in harmonic traps) suggest that standard thermalization arguments fail, and anomalous fluctuations occur at diffusive scales.
The Specific Gap: It was unclear whether the standard Navier-Stokes GHD equation (Eq. 3 in the paper) remains valid beyond infinitesimally short times for the hard rod model. The authors hypothesize that the emergence of long-range correlations (scaling as $1/\ell$ , where $\ell$ is the ratio of macroscopic to microscopic scales) invalidates the standard derivation, which assumes uncorrelated local equilibrium states.

2. Methodology

The authors perform an exact microscopic derivation starting from the known solution for the hard rod gas, avoiding phenomenological assumptions like the maximum entropy principle.

Microscopic Model: The system consists of classical hard rods of diameter $a$ in 1D. The exact trajectory of a tracer particle $i$ is given by:
$x_i(t) = x_i + p_i t + a \sum_{j \neq i} [\theta(\hat{x}_i + p_i t - \hat{x}_j - p_j t) - \theta(\hat{x}_i - \hat{x}_j)]$
where $\hat{x}$ are contracted coordinates.
Scaling Limit: They consider the hydrodynamic limit where the number of particles $N$ , observation length $\ell$ , and time $T$ go to infinity ( $N \sim \ell \sim T \to \infty$ ), keeping ratios fixed.
Expansion Strategy:
1. They express the evolved quasi-particle density $\rho(t, x, p)$ as an average over the initial state.
2. They perform a cumulant expansion (large deviation theory) up to order $O(1/\ell)$ .
3. Crucially, they do not assume the state is a local GGE. Instead, they explicitly track the evolution of the connected two-point correlation function $C(t, x, p, y, q)$ .
4. They decompose the correlation function into a singular part (GGE-like, $\delta(x-y)$ ) and a long-range part ( $C_{LR}$ ), which is piecewise smooth but exhibits a jump discontinuity at $x=y$ .
Derivation of Time Evolution: By analyzing the time derivative of the density at $t \to 0^+$ (after the $\ell \to \infty$ limit), they derive the evolution equation for the density, incorporating contributions from both the local GGE correlations and the newly formed long-range correlations.

3. Key Contributions

The paper makes several significant theoretical breakthroughs:

Coupled System of Equations: The authors demonstrate that the diffusive dynamics cannot be described by a single equation for the density $\rho(t, x, p)$ . Instead, it requires a closed system of two coupled equations:
- One for the one-point function (density).
- One for the two-point correlation function.
  The evolution of the density depends on the two-point function, and the evolution of the two-point function depends on the one-point function.
Cancellation of Standard Diffusion: A major finding is that the standard Navier-Stokes diffusion term (derived from local GGE correlations) is exactly cancelled by the contribution from the jump discontinuity in the long-range correlations.
- The standard term: $\frac{1}{2\ell} \partial_x \int D(p,q) \partial_x \rho$ .
- The long-range correction: Arises from $C_{LR}$ and cancels the above term.
- Result: The net diffusive current is driven solely by the continuous (symmetric) part of the long-range correlations, not by the local equilibrium fluctuations.
Time-Reversibility: Unlike the standard Navier-Stokes equations which imply an arrow of time (monotonic entropy increase), the new derived equations are time-reversible.
- The system does not intrinsically thermalize to a GGE on the diffusive scale because the long-range correlations preserve information about the initial state's history.
- Entropy is not guaranteed to increase, challenging the standard view that diffusion implies thermalization in integrable systems.
Dynamically Stable Manifold: The authors show that while a local equilibrium state (with no long-range correlations) is a valid initial condition, it is unstable. Immediately ( $t \to 0^+$ ), the system develops a specific structure of long-range correlations (a "jump" at $x=y$ ) that places it on a "dynamically stable manifold." The hydrodynamic equations must be evaluated on this manifold, not on the initial local equilibrium state.

4. Main Results

The primary result is the new diffusive hydrodynamic equation for the hard rod gas (Eq. 7 and Eq. 9 in the paper):

$\partial_t \rho(t, x, p) + \partial_x (v_{eff} \rho) = -\frac{1}{\ell} \partial_x \left[ \frac{a}{(2\pi)^2} \int dq \frac{p-q}{1_{dr}} C_{LR, sym}(t, x, p, q) \right] + O(1/\ell^2)$

Where:

$v_{eff}$ is the effective velocity.
$1_{dr} = 1 - a \int dq \rho$ .
$C_{LR, sym}$ is the symmetric part of the long-range correlation function of normal modes.
The standard diffusion matrix $D(p,q)$ is absent because it is cancelled by the correlation jump.

Numerical Verification:
The authors performed extensive numerical simulations of the hard rod gas ( $N \sim 10^7$ realizations).

They compared the $O(1/\ell)$ corrections to the Euler dynamics against their theoretical predictions.
Figure 5 shows that the new theory (Eq. 51) matches the numerical time derivatives of the moments perfectly.
In contrast, the standard Navier-Stokes GHD (Eq. 3) shows significant deviations from the numerical data once long-range correlations have developed (i.e., for $t > 0$ ).

5. Significance

Resolution of the Thermalization Paradox: The paper explains why integrable systems in certain potentials (like harmonic traps) fail to thermalize to a GGE. The standard diffusive mechanism (entropy production) is suppressed by the specific structure of long-range correlations in integrable systems.
Validation of General Theory: This work provides the first exact microscopic derivation confirming the general framework proposed in the authors' companion paper [1] regarding diffusion in linearly degenerate hydrodynamics.
New Paradigm for Hydrodynamics: It establishes that for integrable systems, hydrodynamics at the diffusive scale is inherently a multi-point correlation problem. One cannot close the hierarchy at the one-point function level; the two-point function is essential.
Implications for Transport: The findings suggest that transport in integrable systems is fundamentally different from non-integrable systems, where local equilibrium assumptions usually hold. The "arrow of time" is not generated by diffusive corrections in these systems, implying that thermalization must occur via other mechanisms (e.g., integrability breaking).

In summary, the paper rigorously proves that the standard Navier-Stokes GHD is insufficient for describing the long-time diffusive behavior of hard rods. The correct description requires a coupled system of equations accounting for dynamically generated long-range correlations, leading to time-reversible dynamics that do not inherently drive the system toward thermal equilibrium.