DLA growth is governed by Laplace's equation for the concentration field of diffusing particles: the local growth probability at any surface point is proportional to the gradient of the concentration field there. This connection to Laplacian growth means DLA is mathematically related to viscous fingering (Hele-Shaw flow), dielectric breakdown, and electrostatic field-line problems. The fractal dimension can be derived (approximately) from conformal mapping theory; exact analytic results remain an open problem in mathematical physics.

Classic examples include copper electrodeposition from copper-sulfate solution (forming fractal metal trees on the cathode), the branching of lightning discharge channels through air, the growth of snowflake dendrites in supersaturated vapor, the infiltration patterns of water into dry sand or porous rock, and bacterial colony growth on nutrient-limited agar plates. The cosmic filament network seen in large-scale structure simulations also displays a DLA-like topology driven by gravitational collapse along density gradients.

While each individual DLA run is stochastic and no two clusters look identical, the statistical properties — fractal dimension, branch-tip density distribution, mass-radius scaling — are universal and highly reproducible. Large-scale structure emerges deterministically from the Laplacian growth rule: branch tips always have the highest probability of capturing the next particle because the diffusion field is concentrated there, driving a consistent tip-splitting and screening dynamic. The randomness determines fine detail; the fractal geometry is a robust emergent property.

DLA was introduced by Thomas A. Witten Jr. and Leonard M. Sander in their 1981 paper "Diffusion-Limited Aggregation, a Kinetic Critical Phenomenon" published in Physical Review Letters (volume 47, p. 1400). Witten and Sander were studying irreversible growth processes at the University of Michigan. Their model became one of the most-studied examples of fractal pattern formation, spawning thousands of follow-up papers across physics, chemistry, materials science, and mathematics over the following four decades.

DLA belongs to a broad family of Laplacian growth models. Closely related processes include dielectric breakdown model (DBM) fractal discharge, viscous fingering in Hele-Shaw cells, Eden growth (compact clusters from random-walk aggregation with no diffusion step), Ballistic Aggregation, and the broader field of iterated function systems and L-system fractals. On this site, the Mandelbrot Set and Julia Set explorers demonstrate deterministic fractal geometry, while the Random Walks article explains the underlying Brownian motion physics shared with DLA.

DLA models are applied in materials science to predict dendrite formation during metal solidification — short-circuiting dendrites are a critical failure mode in lithium-ion batteries and must be suppressed. In petroleum engineering, DLA-like simulations model how oil or gas migrates through fracture networks. In semiconductor fabrication, DLA has been used to understand thin-film growth and surface roughening. Biomedical researchers use DLA-inspired models to simulate tumor vascularization and the branching of vascular networks.

Active research topics include the exact analytic derivation of the 2D fractal dimension (still unproven rigorously), DLA in three dimensions (D_f approximately 2.50), multi-particle DLA variants where clusters interact, off-lattice and anisotropic DLA that captures crystalline symmetry, and DLA on curved or hyperbolic geometries. There is also growing interest in "inverse DLA" problems: given a fractal structure (e.g., a river network or a vascular tree), infer the growth rules that produced it. Machine learning methods are beginning to be applied to learn DLA dynamics from image data.