In conventional fluids, surface waves push particles in the direction of wave propagation. This is Stokes drift — a net transport that arises because circular particle orbits don't quite close, creeping forward with each cycle. It's one of the oldest results in fluid mechanics, derived by Stokes himself.
Du Plessis and collaborators show that odd viscosity reverses it.
Odd viscosity is the parity-breaking component of the viscous stress tensor. It appears in chiral fluids — active biological suspensions, electron fluids under magnetic fields, rotating shallow flows — and it does not dissipate energy. Instead, it deflects motion sideways, like a Coriolis force at the molecular level.
When odd viscosity acts on capillary waves, it splits them into two branches: a dispersive branch (familiar) and a quasi-acoustic branch that has no counterpart in ordinary fluids. The two branches propagate at different speeds and attenuate at different rates. Standing waves, which require equal-and-opposite traveling components, become impossible — they spontaneously resolve into traveling waves because the two directions are no longer equivalent.
Above a threshold odd viscosity, the accumulated vorticity near the surface reverses the shear current. Particles drift opposite to the wave propagation direction. This anti-Stokes drift has no counterpart in any conventional fluid — it requires the breaking of parity at the level of the constitutive law.
The mechanism is nonlinear accumulation: odd viscosity generates vorticity at each cycle, and because the vorticity production is asymmetric between the two wave branches, it doesn't cancel over a period. The net vorticity drives a mean flow against the wave. The surface provides the boundary condition that traps the vorticity and focuses the effect.
One-way fluidic waveguides become possible. A chirality-programmed interface could route waves and transport mass in a direction that no amount of conventional viscous manipulation could produce.