What are osmosis and diffusion?

Molecules never sit still. In water they jostle and wander at random, and over time that random wandering spreads things out evenly. That is diffusion: a drop of dye in still water slowly colours the whole glass with nobody stirring.

Now put a semipermeable membrane between pure water and sugar water. It is a sheet with holes big enough for water molecules but too small for sugar. The sugar can't spread into the pure side, so the water does the moving instead: more water molecules cross into the sugar side than come back. That one-way drift of water is osmosis.

The sugary side fills up and its level rises. The taller column pushes back, and the flow stops when the extra height's pressure ρgh equals the osmotic pressure π. Jacobus van 't Hoff found in 1886 that π follows a law just like a gas: π = i M R T. M is the concentration in moles per litre, T the temperature in kelvin, R the gas constant, and i the number of particles each unit breaks into: 1 for sugar, 2 for salt (Na⁺ and Cl⁻).

It is surprisingly strong. Seawater, about 0.6 M salt, has π ≈ 27 bar. To balance it you'd need a column of water about 280 m tall. Even 2 mM of sugar overflows this 50 cm tube.

Osmosis pulls water towards the salty side with a pressure π. So what if you push on the salty side harder than π? Water is squeezed the other way, out through the membrane, and the salt stays behind. That is reverse osmosis, the heart of an RO purifier (see ROClear).

The membrane only lets water through when the pump pressure beats the osmotic pressure: water flow ∝ P − π. Tap water with 1,000 ppm of salts has π ≈ 0.8 bar, so a small home pump at 5–6 bar wins easily. Seawater's π is about 27 bar, and the brine gets saltier as water leaves it, so desalination plants push 55–70 bar.

Not all the water can come out. The salt left behind must be flushed away as reject water, and the more water you take (higher recovery), the saltier and higher-π the leftover gets. A tiny bit of salt sneaks through too, so the clean side isn't perfectly pure: its TDS (total dissolved solids) drops by about 95–99%.

Your kidneys face the same maths. The saltiest urine they can make holds about 280 mmol/L of sodium, but seawater has 469. To get rid of the salt in one litre of seawater you'd have to pass about 1.7 litres of urine, so each glass costs you about 0.7 L of water. That is why castaways die of thirst surrounded by water.

A red blood cell is a bag of salty, protein-rich water in a thin skin that water crosses easily. Its inside is about 290 mOsm/L (milli-osmoles: the count of dissolved particles). Put it in liquid with the same count, isotonic, and water crosses both ways equally.

In hypotonic liquid (fewer particles, like pure water) water rushes in. The cell swells into a sphere and, since its skin can barely stretch, it bursts: lysis. In hypertonic liquid (more particles, like seawater) water leaves and the cell shrivels into a spiky crenated shape. That is why a hospital drip is 0.9% saline: 154 mmol/L of NaCl gives about 290 mOsm/L, matching blood exactly.

Your kidneys use osmosis to save water. Deep inside, the medulla is kept very salty, and the hormone ADH opens water channels so water is drawn out of the urine back into the body (see KidneyClear). Your gut absorbs about 8 litres of water a day the same way: it pumps in salt and sugar, and water follows (see IntestineClear).

That is the secret of ORS. A transporter called SGLT1 carries sugar in only together with sodium. So a drink with the right mix of salt and glucose pulls water in even during cholera. The WHO's 2002 formula is slightly weaker than blood, 245 mOsm/L, so water flows in more easily. Very sugary drinks, around 600 mOsm/L, do the opposite and draw water into the gut.

Oxygen gets into blood by plain diffusion across a barrier only about 0.5 µm thick, spread over some 70 m² of lung (see LungsClear). Fick's law: flow = D × area × pressure gap ÷ thickness. Blood takes 0.75 s to pass, and it is full of oxygen within 0.25 s.

Sprinkle salt on sliced cucumber and in a few minutes it sits in a puddle. The salt makes the outside far more concentrated than the cucumber's cells, so water leaves them by osmosis. That is the first step of an achaar, a pickle.

The same pull works on germs. A bacterium is a tiny bag of water too. In brine with 10–20% salt, or jam with 65% sugar, water is drawn out of any microbe that lands there, and it can't grow. Food scientists measure this with water activity, a_w: 1.0 for pure water, lower when more is dissolved. Most bacteria need a_w above 0.91, most yeasts 0.88 and most moulds 0.80. A little acid, oil and spice finish the job, but salt and sugar do the heavy lifting.

Plants run on osmosis. Their cells keep about 400 mOsm/L of sugars and salts inside, so water pours in until the stiff cell wall pushes back. That inward push, turgor pressure, around 10 bar, is what makes a lettuce crisp and a stem stand. Roots drink the same way, as long as the soil water is less concentrated than the root cells.

Salt the soil, or overdo the fertiliser, and water flows the wrong way: out of the roots. The plant wilts even in wet soil. A limp carrot or wilted coriander does the reverse: drop it in cold water and it firms up again as its cells refill.

Diffusion is fast over tiny distances and hopeless over big ones. The time to wander a distance x grows as x²: t ≈ x² ÷ 2D. Oxygen crosses a cell, 10 µm, in about 25 milliseconds. But 1 cm takes around 7 hours, and a metre about 8 years. That is why anything bigger than a flatworm needs a heart and blood vessels: the blood carries oxygen the long way, and diffusion only does the last few micrometres (see HeartClear and LungsClear).

Real solutions aren't ideal. π = i M R T assumes the dissolved particles ignore each other, like an ideal gas. In real salt water the Na⁺ and Cl⁻ attract, so they act like slightly fewer particles: about 8% less pressure at 0.5 M. In very strong solutions the opposite happens: salt and sugar molecules grab water around them, and π climbs above the ideal line. Chemists fix this with an osmotic coefficient φ: π = i φ M R T. Membranes aren't perfect either: urea slips through cell membranes, so it barely counts for osmosis there.

Myth: "water flows to where there is more water." It is the other way round: water flows towards the solute, to where there is less water. And it isn't really about counting water molecules at all. 0.3 M salt water still has 99.5% as much water per litre as pure water, yet it pulls with 14 bar. What matters is the water potential, ψ = P − π. Push on the salty side with a piston and you can stop or even reverse the flow without changing any concentration. That is exactly what RO does.