Exponential and Weibull models for spherical and spherical-shell diffusion-controlled release systems with semi-absorbing boundaries

TL;DR

This paper develops simple exponential and Weibull models for particle diffusion in spherical and shell systems with various boundary conditions, accurately approximating the proportion of particles remaining over time.

Contribution

It introduces a moment matching approach to derive low-parameter models for diffusion systems with complex geometries and boundary conditions, extending previous spherical models.

Findings

Models agree well with stochastic and continuum simulations.

Explicit parameter dependence on system properties.

Effective approximation of complex diffusion behaviors.

Abstract

We consider the classical problem of particle diffusion in $d$-dimensional radially-symmetric systems with absorbing boundaries. A key quantity to characterise such diffusive transport is the evolution of the proportion of particles remaining in the system over time, which we denote by $\mathcal{P}(t)$. Rather than work with analytical expressions for $\mathcal{P}(t)$ obtained from solution of the corresponding continuum model, which when available take the form of an infinite series of exponential terms, single-term low-parameter models are commonly proposed to approximate $\mathcal{P}(t)$ to ease the process of fitting, characterising and interpreting experimental release data. Previous models of this form have mainly been developed for circular and spherical systems with an absorbing boundary. In this work, we consider circular, spherical, annular and spherical-shell systems with absorbing, reflecting and/or semi-absorbing boundaries. By proposing a moment matching approach, we develop several simple one and two parameter exponential and Weibull models for $\mathcal{P}(t)$, each involving parameters that depend explicitly on the system dimension, diffusivity, geometry and boundary conditions. The developed models, despite their simplicity, agree very well with values of $\mathcal{P}(t)$ obtained from stochastic model simulations and continuum model solutions.