Pool model: a mass preserving multi particle aggregation process

This paper introduces and analyzes the Pool model, a mass-preserving, rotationally symmetric analogue of Multi-Particle Diffusion-Limited Aggregation in $\mathbb{R}^2$ where particles performing random walks are absorbed by an expanding circular pool, utilizing a specialized version of Kurtz's theorem to characterize the particle field as an independent non-homogeneous Poisson point process conditioned on the pool's growth.

The Gist

Imagine a vast, empty ocean where tiny, invisible droplets of water are floating around, drifting randomly in the wind. In the very center of this ocean, there is a growing circular pond. This is the "Pool Model" described in the paper.

Here is the story of how this pond behaves, told through simple analogies.

The Setup: The Drifting Droplets

Think of the ocean as a 2D plane (a flat sheet).

• The Droplets: There are millions of tiny water droplets scattered everywhere. They don't have a destination; they just wander aimlessly, like drunk people walking home.
• The Pool: In the middle, there is a circular pool. Initially, it's small.
• The Rule: When a wandering droplet bumps into the edge of the pool, it doesn't bounce off. It gets swallowed. Once swallowed, it becomes part of the pool, adding its mass to the water.
• The Growth: Because the pool gains mass, it must get bigger to hold the new water. As it gets bigger, its surface area expands, and the circle grows outward.

The big question the scientists asked is: How fast does this pool grow? Does it grow slowly, quickly, or does it suddenly become infinitely huge in a split second?

The Three Scenarios (The Phases)

The behavior of the pool depends entirely on how crowded the ocean is (the density of droplets, denoted by $\lambda$). The paper finds three distinct "personalities" for the pool:

1. The Sparse Ocean ($\lambda < 1$): The Slow Crawler

Imagine the ocean is very empty. The droplets are far apart.

• What happens: The pool grows, but it has to wait a long time for a droplet to wander close enough to fall in.
• The Result: The pool grows, but very slowly. It's like a snail trying to cross a desert. The paper proves that in this case, the pool's size grows roughly like the square root of time (very slow). It never explodes; it just keeps getting bigger at a steady, manageable pace.

2. The Crowded Ocean ($\lambda > 1$): The Exploding Balloon

Imagine the ocean is packed shoulder-to-shoulder with droplets.

• What happens: As soon as the pool grows a tiny bit, it immediately swallows a massive wave of droplets that were waiting right next to it. This makes the pool grow even bigger, which immediately swallows even more droplets.
• The Result: This creates a runaway effect. The pool grows so fast that it reaches an infinite size in a finite amount of time.
• The Analogy: It's like a snowball rolling down a hill that suddenly hits a massive snowdrift. It doesn't just get bigger; it becomes a mountain instantly. The paper calls this an "Explosion."

3. The "Goldilocks" Ocean ($\lambda = 1$): The Critical Balance

This is the most interesting and tricky case. The density is just right—not too empty, not too crowded.

• What happens: The pool grows faster than the "Slow Crawler" but doesn't explode like the "Exploding Balloon."
• The Result: It grows in a weird, jagged way. Sometimes it stalls, and sometimes it jumps forward.
• The Mystery: The paper proves it definitely doesn't explode. However, it grows faster than any simple power law (like $t^2$ or $t^{0.5}$).
• The Conjecture: The authors suspect that in the long run, this pool might actually grow at a linear speed (a straight line), but it does so with huge, unpredictable "jumps." Imagine a car that drives mostly at a steady speed but occasionally hits a ramp and flies forward, then lands and drives steadily again.

The Secret Weapon: Kurtz's Theorem

To solve this puzzle, the authors used a mathematical tool called Kurtz's Theorem.

• The Metaphor: Imagine you are watching a movie of the pool growing. You want to know where the next droplet will come from.
• The Problem: The droplets that have already been swallowed are gone. The droplets that are still wandering are affected by the fact that the pool is moving and growing. It's a chaotic mess.
• The Solution: Kurtz's Theorem acts like a "magic filter." It allows the mathematicians to say: "If we ignore the history of the pool and just look at the remaining droplets right now, they behave like a perfectly random, independent cloud."
• This simplifies the chaotic problem into a manageable one, allowing them to calculate the odds of the pool exploding or growing slowly.

Why Does This Matter?

This isn't just about imaginary pools. This model helps scientists understand real-world phenomena:

• Electrochemical Deposition: How metals build up on surfaces in batteries or circuits.
• Crystal Growth: How snowflakes or minerals form.
• The "Fractal" Question: In similar models (called MDLA), scientists have been arguing for decades about whether these shapes grow in a straight line or in a messy, fractal pattern. This paper shows that in a "mass-preserving" world (where nothing is destroyed, only added), the rules are different.

Summary

• Low Density: The pool grows slowly and steadily.
• High Density: The pool grows so fast it explodes instantly.
• Critical Density: The pool grows in a complex, jagged way that is faster than slow growth but doesn't explode.
• The Tool: They used a clever mathematical trick to turn a chaotic crowd of wandering particles into a predictable random cloud.

The paper essentially maps out the "personality" of a growing pool, showing us exactly when it will stay calm, when it will go crazy, and when it will dance in a complex, critical rhythm.

Technical Summary

1. Problem Statement and Motivation

The paper investigates a continuous-time, continuous-space aggregation model called the Pool Model, defined in $\mathbb{R}^2$. This model serves as a rotationally symmetric analogue to Multi-Particle Diffusion-Limited Aggregation (MDLA).

• Context: MDLA models electrochemical deposition where particles perform random walks and attach to an aggregate upon contact. In standard MDLA, particles at the contact site are annihilated (absorbed), and the aggregate grows by adding that single site.
• The Pool Model Difference: In the Pool model, the aggregate is a circular "pool" centered at the origin. When a diffusing particle (droplet) enters the pool, it is not annihilated; instead, its mass is added to the pool. Consequently, the pool expands to preserve total mass. The radius $E_t$ grows such that the area $\pi E_t^2$ equals the initial area plus the number of absorbed particles.
• Key Challenge: Unlike 1D MDLA or lattice-based models, the Pool model operates in $d=2$ with a continuously changing boundary. As the pool grows, the incremental radius added by a new particle decreases ($\Delta r \propto 1/\sqrt{N}$), creating a dynamic where the "search distance" for the next particle and the curvature effects change over time.
• Objective: To characterize the growth behavior of the pool radius $E_t$ as a function of the initial particle density $\lambda$, specifically determining conditions for explosion (infinite growth in finite time) and establishing growth rates in sub-critical and critical regimes.

2. Methodology and Tools

The authors employ a combination of probabilistic tools, stochastic analysis, and branching process theory.

• Kurtz's Theorem Adaptation: A central technical contribution is a rigorous proof of a version of Kurtz's Theorem for this specific model.
  • Theorem 1: Conditioned on the history of the pool's growth ($E_s$ for $s \leq t$), the field of active particles (those outside the pool) forms an independent non-homogeneous Poisson point process.
  • Significance: This allows the authors to treat the particle density outside the pool as a known intensity field, decoupling the complex interaction between particle motion and boundary growth. The proof involves discretizing time and using the Marking Theorem to handle the instantaneous engulfing procedure.
• Branching Process Analogy: The engulfing process is analyzed using Galton-Watson (GW) trees.
  • When the pool expands, the number of new particles captured in a shell is Poisson-distributed.
  • The total number of particles absorbed in a "generation" of growth follows a branching process structure where the offspring distribution is Poisson($\lambda$).
  • The condition for explosion corresponds to the survival probability of a supercritical GW process ($\lambda > 1$).
• Stochastic Domination: To prove non-explosion at criticality ($\lambda=1$), the authors construct a dominating process using i.i.d. critical GW trees and exponential waiting times, showing that the actual pool growth is stochastically bounded by a process that does not explode.
• Large Deviation and Concentration Inequalities: Used to bound the number of particles in expanding balls and to prove that the particle density does not deviate significantly from its expectation (Proposition 6.1).

3. Key Results

The paper establishes a phase transition based on the intensity parameter $\lambda$:

A. Explosion Condition (Theorem 2)

• Supercritical Case ($\lambda > 1$): The pool explodes almost surely in finite time ($T_E < \infty$).
  • Mechanism: The particle density is high enough that the "offspring" (newly absorbed particles) in each growth step form a supercritical branching process, leading to an infinite cascade of absorption in finite time.
• Critical Case ($\lambda = 1$): The pool does not explode almost surely ($T_E = \infty$).
  • Mechanism: The associated branching process is critical. While it can grow large, the probability of infinite growth in finite time is zero.

B. Growth Rates (Theorem 3)

• Sub-critical Case ($\lambda < 1$): The growth is diffusive but with logarithmic corrections.
  • Asymptotically almost surely: $\sqrt{t} \log^{-1+\epsilon}(t) \leq E_t \leq \sqrt{t} \log(t)$.
  • This indicates the pool grows roughly like a standard diffusion process ($\sim \sqrt{t}$), slowed slightly by the low density of particles.
• Critical Case ($\lambda = 1$): The growth is super-diffusive (faster than any polynomial sub-linear growth) but sub-linear.
  • For any $\zeta < 1$, $\limsup_{t \to \infty} E_t / t^\zeta = \infty$ almost surely.
  • The paper provides an upper bound of $E_t \leq 2\exp(Ct)$, though simulations suggest a linear growth rate might be the true behavior (see Open Problems).

4. Significance and Contributions

1. First Rigorous Proof of Kurtz's Theorem for MDLA-like Models: While the theorem was attributed to Kurtz in private communications regarding MDLA, this paper provides the first written, rigorous proof for a continuous-space, mass-preserving variant. This tool is essential for analyzing the conditional distribution of particles in aggregation models.
2. Resolution of the Linear Speed Conjecture for Low Intensity: The authors demonstrate that the conjecture that MDLA grows at a linear speed for any $\lambda$ (proven for high $\lambda$) fails for low intensity in the Pool model. The sub-critical regime exhibits diffusive ($\sqrt{t}$) rather than linear growth.
3. Mass Preservation vs. Annihilation: By contrasting with standard MDLA (where particles are annihilated), the Pool model highlights how mass preservation leads to different explosion dynamics. The model is physically relevant for liquid pooling of diffusing droplets.
4. Critical Phenomena in 2D: The paper provides a detailed analysis of the critical case ($\lambda=1$) in two dimensions, a regime where standard mean-field approximations often fail due to fluctuations.

5. Open Problems and Future Directions

The paper concludes with several conjectures and open questions:

• Conjecture 1 (Linear Growth at Criticality): Simulations suggest that at $\lambda=1$, the pool might actually grow linearly ($E_t \sim \xi t$) despite the theoretical lower bound being super-linear. The authors suspect the "jumps" seen in simulations become rarer as the system evolves.
• Brownian Motion vs. Random Walk: The proof for non-explosion at $\lambda=1$ relies on the discrete nature of the random walk. If particles follow Brownian motion, the proof fails because Brownian paths can move arbitrarily fast in small time intervals. It is conjectured that the Brownian Pool model also does not explode, but this remains unproven.
• Annihilating Pool Model: What happens if the Pool model is modified so that particles are annihilated upon entry (like standard MDLA)? The authors conjecture that at criticality, this version grows at a linear speed.
• Engulfing MDLA: A variant where excess particles inside the aggregate perform random walks until hitting the boundary is proposed, with conjectures regarding its limiting shape and growth speed.

Summary

This paper introduces a mathematically tractable, rotationally symmetric model of multi-particle aggregation. By leveraging a novel application of Kurtz's theorem and branching process theory, the authors rigorously classify the model's behavior into three distinct phases: diffusive growth for low density, non-explosive but rapid growth at criticality, and finite-time explosion for high density. The work bridges the gap between lattice-based DLA models and continuous physical processes, offering new insights into the critical behavior of aggregation in two dimensions.