A zeolite crystal is riddled with uniform-diameter nanopores running through an otherwise closed aluminosilicate framework. Catalytically active sites sit inside those channels, not on the outer surface, so a reactant molecule can only react if it can physically fit through the pore mouth first. Linear (rod-shaped, cyan), branched (amber) and bulky (purple, spherical/caged) molecules of increasing effective width all approach the crystal face. Whichever ones are narrower than the current pore diameter thread the channel, reach the internal active site, react, and emerge as product on the far side. Anything too wide simply bounces off the exterior wall, unreacted — the sieve is size-selective, not reactivity-selective.
enters pore if molecule_width < pore_diameter
reacts at the internal active site (mid-channel)
exits as product on the far side
bounces off exterior wall if molecule_width ≥ pore_diameter
- Pore diameter — the sieve cutoff. Narrow pores convert only the smallest, linear molecules; widen them and progressively bulkier shapes start getting through too, until nothing is excluded.
- Zeolite vs unconfined toggle — switch to an unconfined (non-porous) catalyst surface and every molecule that touches it reacts on contact and bounces back out, regardless of shape or size. There is no sieving at all — conversion is 100% for every shape, at the cost of losing product selectivity.
- Feed rate — how often new molecules are launched at the crystal face.
This is exactly how industrial zeolite catalysts (e.g. ZSM-5) achieve shape-selective cracking and isomerization: the pore geometry itself does the separating, favoring straight-chain hydrocarbons over their bulkier branched or aromatic isomers without any extra separation step.