Below a pore's condensation pressure, gas molecules only build up a thin liquid film on the pore wall. The Kelvin equation gives the pore (meniscus) radius rk in equilibrium with relative pressure P/P₀:
ln(P/P0) = -k·γ·Vm / (r_k·R·T)
k = 1 cylindrical meniscus (filling / adsorption branch)
k = 2 hemispherical meniscus (emptying / desorption branch)
γ = liquid surface tension, Vm = molar volume, T = temperature
Because filling and emptying involve menisci of different curvature, the two branches trigger at different pressures for the same pore radius — this is the physical origin of the adsorption/desorption hysteresis loop seen in Type IV isotherms of real mesoporous solids (silica gels, MCM-41, activated alumina).
- Rising P — a film of thickness t(P) coats the wall until the shrunken core radius rc = rp − t satisfies the k=1 equation; the pore then snaps to fully liquid-filled (capillary condensation).
- Falling P — the filled pore stays condensed until the k=2 equation is satisfied at the full pore radius rp; it then empties back to a thin film.
- Pore radius slider — smaller pores condense/evaporate at lower P/P₀; this radius-dependence is exactly what BJH analysis inverts to recover a pore-size distribution from a measured isotherm.
- Gas selector — N₂ at 77 K and Ar at 87 K are the two adsorptives actually used in commercial porosimeters; they shift the loop because γ, Vm and T differ.
Film thickness here uses a simplified monotonic statistical-thickness approximation (t grows smoothly with P/P₀, capped below rp) rather than a specific literature t-curve fit — it is illustrative of the real multilayer-adsorption trend, not a metrology-grade fit.