Modeling Ostwald Ripening Dynamics in Porous Microstructures

TL;DR

This paper introduces an advanced image-based pore-network model (iPNM) that accurately simulates Ostwald ripening in porous microstructures, capturing complex topological changes and flow dynamics beyond previous models.

Contribution

The iPNM removes limitations of existing models by encoding pore shapes via curvature-saturation curves and coupling flow, transport, and ripening in a unified, efficient framework.

Findings

iPNM agrees well with microfluidic experiments without adjustable parameters

It captures population statistics and individual ganglion curvatures

It is computationally efficient compared to direct numerical simulation

Abstract

Partially miscible ganglia trapped in a porous medium evolve through Ostwald ripening, driven by differences in interfacial curvature. In practice, ganglia can span multiple pores and undergo discrete capillary events - invasion, snap-off, retraction, fragmentation, coalescence, and dislocation - that alter their topology and induce local flow. Existing pore-network models (PNMs) for ripening are limited to single-pore ganglia, assume idealized pore shapes, and operate under quasi-static conditions that preclude flow. We present an image-based pore-network model (iPNM) that removes these limitations. Unlike existing PNMs, iPNM requires no idealization of pore shapes, as the effect on capillarity is encoded locally in curvature-saturation curves computed via the pore-morphology method. iPNM couples two-phase flow, solute transport, and Ostwald ripening within a unified framework. We first verify iPNM against a prior quasi-static PNM, then validate it against recent high-resolution microfluidic experiments of hydrogen ripening in a sandstone-patterned micromodel over 15-24 days at 40C and 80C. Good agreement is obtained without adjustable parameters. Comparison with a continuum model shows that while macroscopic saturation is captured by both approaches, iPNM uniquely resolves population statistics, individual ganglion curvatures, and pre-equilibrium ripening dynamics within a representative elementary volume. Its computational efficiency over direct numerical simulation makes it suitable for guiding the development of improved theories of ripening in confined geometries.