Pattern Formation in Laplacian Growth: Theory

TL;DR

This paper develops a statistical, Hamiltonian-based theory for two-dimensional Laplacian growth, predicting pattern formation, fractality, and interface morphology through conformal mapping and many-body dynamics.

Contribution

It introduces a novel Hamiltonian framework for Laplacian growth, incorporating surface effects as particle interactions, and connects particle statistics to interface fractality and morphology.

Findings

Hamiltonian dynamics describe interface evolution.

Surface effects modeled as repulsive particle interactions.

Explicit relation between particle statistics and fractal dimension.

Abstract

A first-principles statistical theory is constructed for the evolution of two dimensional interfaces in Laplacian fields. The aim is to predict the pattern that the growth evolves into, whether it becomes fractal and if so the characteristics of the fractal pattern. Using a time dependent map the growing region is conformally mapped onto the unit disk and the problem is converted to the dynamics of a many-body system. The evolution is argued to be Hamiltonian, and the Hamiltonian is shown to be the conjugate function of the real potential field. Without surface effects the problem is ill-posed, but the Hamiltonian structure of the dynamics allows introduction of surface effects as a repulsive potential between the particles and the interface. This further leads to a field representation of the problem, where the field's vacuum harbours the zeros and the poles of the conformal map as particles and antiparticles. These can be excited from the vacuum either by fluctuations or by surface effects. Creation and annihilation of particles is shown to be consistent with the formalism and lead to tip-splitting and side-branching. The Hamiltonian further allows to make use of statistical mechanical tools to analyse the statistics of the many-body system. I outline the way to convert the distribution of the particles into the morphology of the interface. In particular, I relate the particles statistics to the distributions of the curvature and the growth probability along the physical interface and to the fractal dimension. If the pattern turns fractal the latter distribution gives rise to a multifractal spectrum, which can be explicitly calculated for a given particles distribution. A `dilute boundary layer approximation' is discussed, which allows explicit calculations and shows emergence of an algebraically long tail