Delhi's winter PM2.5 routinely hits 500 µg/m³ — twenty times the WHO limit. The smog sits trapped under a temperature inversion, a 300-meter lid of warm air pressing cold, dirty air against the ground. What if we punched through it with a giant smoke ring?

Vortex rings are absurdly efficient at moving fluid over long distances. A dolphin's bubble ring holds shape for meters. Volcanic vortex rings from Etna have been tracked at 200 m altitude. The physics: a toroidal vortex carries its own low-pressure core, entraining surrounding fluid and self-propelling until viscous decay wins. Unlike a jet, which dissipates via turbulent mixing within ~10 diameters, a laminar vortex ring can travel 50-100 diameters before breaking up.

The machine: A 200-meter-diameter vertical cylinder, 500 m tall, open at the top. A piston (or pulsed fan array) at the bottom shoves a slug of air upward. As it exits the orifice, the shear layer rolls up into a torus. Aim it at the inversion layer.

Sizing the ring. Optimal vortex-ring formation number is L/D ≈ 4 (Gharib et al., 1998) — push a slug longer than that and the extra fluid trails as a jet, wasting energy. So for a 200 m orifice, piston stroke ≈ 800 m of air column. With ejection velocity U = 30 m/s, ring translational velocity ≈ 0.6U ≈ 18 m/s. It reaches the 300 m inversion in ~17 seconds.

Energy per shot. Slug volume: π(100)² × 800 ≈ 2.5 × 10⁷ m³. Mass at ρ=1.2 kg/m³ ≈ 3 × 10⁷ kg. Kinetic energy at 30 m/s: ½mv² = 1.35 × 10¹⁰ J = 3,750 kWh per ring. At $0.10/kWh, that's $375 per shot — trivial compared to Delhi's smog-related mortality costs (~$36 billion/year per Lancet).

Does it actually clear anything? Here's the beautiful part. The ring doesn't just push a hole — it entrains a coaxial column of air roughly 3× its own volume as it travels. Punching through the inversion creates a temporary chimney: warm ground-level air (buoyantly unstable once released) rushes up through the breach. One 200 m ring opens a ~600 m diameter mixing corridor for perhaps 10 minutes before the inversion re-seals.

The catch — Reynolds number. Re = UD/ν = 30 × 200 / (1.5 × 10⁻⁵) ≈ 4 × 10⁸. That's brutally turbulent. Real vortex rings at Re > 10⁵ develop azimuthal instabilities (Widnall waves) and shed vorticity chaotically. Lab-scale coherence numbers (50-100 diameters travel) collapse to maybe 2-3 diameters at these scales. So 600 m of travel, not 20 km. Still enough to breach a 300 m inversion — barely.

Structural reality. A 200 m cylinder must survive the reaction force from ejecting 3 × 10⁷ kg at 30 m/s in ~25 seconds: F = ma ≈ 3.6 × 10⁷ N, or ~3,600 tonnes of thrust. That's a Saturn V's F-1 engine, pointed downward, inside a stadium. The foundation must anchor against uplift on every reverse stroke. Manageable — offshore wind monopiles handle comparable cyclic loads.

Scaling up: Ring 20 of these around Delhi's periphery, fire in coordinated sequences during peak inversion hours. Total power draw during a firing cycle: ~500 MW — half a nuclear reactor for the world's largest cough drop.