This is the same Knudsen-effect physics as the 3D aerogel model, but represented two ways that are both genuinely 2D — not a flattened camera view of a 3D box. The top canvas is a real 2D hard-disk gas: molecules undergo elastic disk-disk collisions and, unlike a simple bounding box, also collide elastically with a schematic lattice of fixed silica "pillars" (a 2D cross-section of the pore-wall skeleton) whose spacing shrinks as you shrink the pore diameter slider. The bottom canvas plots the analytic suppression curve from kinetic theory:
λ = k_B T / (√2 π σ² P)
Kn = λ / d
κ_gas(d) = κ_gas,0 / (1 + 2β·Kn), β ≈ 2 (accommodation)
As you shrink the pore, the pillar lattice in the top canvas gets denser, so the live-measured "pillar-wall hits/s" genuinely rises relative to "gas-gas hits/s" — an emergent result of the 2D particle geometry, not a hardcoded number. The bottom canvas's marker slides down the same analytic curve using the real λ, Kn and κ formulas above. Watching both together shows the same causal chain the 3D model shows: shrink the pore (or evacuate the gas) past the mean free path, and wall collisions start to dominate over gas-gas collisions, collapsing the gas's contribution to heat flow.
- Pore diameter — sets d directly and controls the pillar lattice spacing in the top canvas.
- Pressure — stretches λ (fewer molecules to collide with) in the analytic model; a lower-pressure aerogel panel reaches the Knudsen regime even with wider pores.
- Temperature — raises λ slightly and raises the intrinsic gas κ₀ as (T/300)^0.75; the two partly offset.
- Gas species — a larger molecule (bigger collision cross-section σ) has a shorter λ, so it needs an even smaller pore to reach the same suppression.