This is a genuine two-compartment membrane transport model, not a decorative animation. Five molecule species — O₂, CO₂, H₂O, glucose and Na⁺ — each carry their own permeability coefficient set by real size and charge, so small nonpolar gases cross the lipid bilayer easily while the charged ion and bulky sugar barely leak at all. Glucose and Na⁺ instead rely on channel proteins modelled as real bind → hold → release state machines, which is why their transport rate saturates exactly like a real GLUT transporter or ion channel once every channel stays constantly occupied.
Fick's law of simple diffusion emerging from independent per-particle crossing probabilities, and Michaelis-Menten-style saturation kinetics emerging from a finite pool of channel proteins with a fixed translocation time.
Pick a molecule to graph, set channel/transporter counts, choose a starting gradient and temperature, then watch the live concentration-vs-time chart converge as both compartments equilibrate.
Because channels have a hard throughput ceiling (Vmax), doubling the concentration gradient across a saturated membrane barely changes the flow — the only way to move more is to add more channels, which is exactly what insulin does with GLUT4.
The 3D version models one shared pool of particles crossing through geometric channel slots or a single toggled pump, with no distinction between molecule species beyond water/ion. This 2D version simulates five distinct real species simultaneously, each with its own physically-motivated permeability, and gives its two facilitated species genuine channel-protein kinetics (bind, hold, release) that produce measurable Michaelis-Menten saturation rather than just a geometric bottleneck.
Every free particle attempts to cross the membrane each frame with the same fixed probability, regardless of which side it's on. Because the more crowded side simply has more particles making that same attempt, the net flow is automatically proportional to the concentration difference — which is precisely Fick's law, arising from independent per-particle behaviour rather than being programmed in directly.
Each channel is a small state machine: it can bind one molecule from whichever side offers it (with a probability proportional to that side's population), hold it for a fixed translocation time, then release it on the far side. At low concentration, more molecules mean more binding events and throughput rises linearly. Once every channel is permanently occupied, throughput hits a hard ceiling set only by channel count and translocation time — that rise-then-plateau shape is the defining signature of Michaelis-Menten kinetics.
Glucose is a large, polar molecule and Na⁺ carries a full charge, and the lipid bilayer's hydrophobic core strongly resists both size and charge. Their simple-diffusion permeability is set near zero in this model, mirroring real cell membranes, so they need dedicated transporters and channels to cross at any useful rate.
It shows how many of the available channels for the currently graphed species are mid-transport right now. When that number sits near 100%, the membrane is running at its Michaelis-Menten Vmax — adding channels, not raising concentration further, is the only way to move more molecules per second.
Net flux is the difference between left-to-right and right-to-left crossings per second, not the total crossing activity. Near equilibrium, particles still hop back and forth constantly in both directions, but because the concentrations are nearly equal the two directions almost cancel, so the net (the quantity Fick's law actually predicts) approaches zero even though individual molecules keep moving.
It is a genuine mechanistic model — the diffusion and saturation behaviours are real physics and enzyme-kinetics concepts, not scripted animation — but the specific rate constants are tuned for a clear, watchable demonstration rather than fitted to a particular real membrane or transporter's measured Vmax and Km.