Viscous Fingering: The Physics of a Branching Interface

The Hele-Shaw cell

Squeeze fluid between two flat plates separated by a very thin gap — a Hele-Shaw cell — and inertia becomes irrelevant: viscosity dominates so completely that the gap-averaged flow obeys Darcy's law, exactly the same equation used to describe flow through porous rock. Inject a low-viscosity fluid, like air or water, into the center of a high-viscosity fluid, like glycerin or oil, and watch the growing bubble.

Why the interface can't stay circular

Philip Saffman and Sir Geoffrey Taylor worked out the instability mechanism in 1958. Any small bump on the growing interface pokes forward into a region of the viscous fluid ahead where the resistance to flow is, for a moment, slightly lower — so the local pressure gradient there is steeper, and the bump is pushed forward even faster than its neighbors. That is a self-reinforcing feedback loop with nothing to stop it: the bump grows into a finger, and the finger grows faster than the rest of the interface. It is the same basic feedback logic as the Rayleigh-Taylor instability, but here the driving force is a viscosity contrast under an imposed pressure gradient, not gravity.

u = -(b² / 12μ) · ∇p        (Darcy's law, gap-averaged Hele-Shaw flow)

b = plate separation (gap thickness)
μ = fluid viscosity
∇p = pressure gradient driving the flow

pressure in each fluid obeys the Laplace equation:  ∇²p = 0

Surface tension sets the finger width

Without surface tension the linear theory predicts fastest growth at infinitely short wavelengths, which is unphysical and a sign the model is incomplete. Surface tension resists sharply curving the interface, stabilizing the shortest wavelengths and setting a preferred, finite finger width. The balance between viscous forces and surface tension is captured by the capillary number, Ca = μU/σ. In a channel geometry, Saffman and Taylor found — using an elegant conformal-mapping technique that made this one of the first nonlinear pattern-forming problems ever solved exactly — that at low capillary number a single dominant finger settles at almost exactly half the width of the channel, a famous result now called the McLean-Saffman relation.

Fractal fingering and diffusion-limited aggregation

Inject from a single point into an open radial cell instead of a channel, and at high enough injection rate the fingers themselves sprout side branches, which sprout further branches, producing a delicate, fractal-looking pattern. This is not a coincidence: because pressure obeys the Laplace equation in both fluids, the growth problem is mathematically the same kind of Laplacian growth process as diffusion-limited aggregation (DLA), where particles undergoing random walks stick to a growing cluster — both problems are governed by a harmonic field whose gradient at the boundary sets the local growth rate, which is why viscous-fingering patterns and DLA clusters look so visually similar despite arising from completely different physics.

Where it shows up

Viscous fingering is mostly an unwanted nuisance in engineering. In enhanced oil recovery, injecting water or CO₂ into a reservoir to push out the remaining oil works poorly precisely because the low-viscosity injected fluid fingers through the oil rather than displacing it uniformly, letting the injected fluid break through to the production well while oil stays trapped behind. The same instability threatens groundwater remediation and industrial coating processes, where an unwanted fingered edge is a manufacturing defect. On the theoretical side, because the underlying Laplacian-growth mathematics reappears in electrodeposition and dielectric breakdown patterns, viscous fingering remains one of the standard model systems used across condensed-matter and nonequilibrium physics to study pattern formation.

Frequently asked questions

Try it live

Everything above runs in your browser — open Viscous Fingering and change the parameters while it is running. Nothing is installed, nothing is uploaded, the whole model lives in one tab.