Pressure from microscopic collisions

In brief

Stand under a hailstorm and you feel a steady push against your umbrella, not because hail is a continuous pushing fluid but because thousands of separate stones strike it every second, each delivering a tiny jolt. Blow up a balloon and the same thing is happening on the inside surface of the rubber, except the hailstones are gas molecules, moving far faster and striking far more often, so many collisions per second that the jolts blur into what a pressure gauge reads as a smooth, constant push. Pressure, in other words, is not a separate ingredient added to matter. It is the statistical residue of an enormous number of individual collisions, and once that residue is derived from first principles, it becomes possible to predict exactly how pressure must change with depth, temperature, or the size of the container doing the confining.

The full treatment

First look: one molecule, one wall

Imagine a single gas molecule of mass m trapped in a cube-shaped box with sides of length L, moving straight toward one wall with speed v in that direction, perpendicular to the wall. When it strikes the wall, the collision is elastic: the wall is far more massive than the molecule and does not move, so the molecule simply bounces straight back with the same speed, its velocity reversed. Its momentum along that direction was m times v before the collision and negative m times v after, so the wall received a momentum kick of two times m times v, delivered in the brief instant of contact.

That molecule does not hit the same wall again immediately. It travels the length of the box, bounces off the opposite wall, and travels back, covering a round trip distance of two times L before striking the original wall again. The time between successive hits on that one wall is therefore two L divided by v.

Building the idea: force as momentum delivered per unit time

Force, by Newton's second law in its impulse form, is the rate at which momentum is delivered. Averaged over many round trips, this single molecule delivers a momentum kick of two m v every two L divided by v seconds, which works out to an average force of m times v squared, divided by L. This is the average push from just one molecule on just one wall.

A real gas has an enormous number of molecules, call it N, all doing this independently and simultaneously, each with its own speed in the direction perpendicular to the wall. Summing their individual average forces gives a total force on that wall equal to m divided by L, multiplied by the sum of every molecule's squared velocity component perpendicular to the wall, which can be written as m times N times the average of that squared velocity component, all divided by L.

Pressure is force per unit area, and the wall in question has area L squared, so dividing the total force by L squared gives pressure equal to m times N times the average squared velocity component, divided by L cubed, which is the volume of the box. Writing n for the number density N divided by volume V, pressure equals n times m times the average of the squared velocity component perpendicular to that particular wall.

From one direction to three, and to a usable formula

Molecules move in three dimensions, and there is nothing special about the direction chosen for the wall above; on average, motion is equally vigorous in all three perpendicular directions, so the average of the squared velocity component in any one direction equals one third of the average of the total squared speed, since the total squared speed splits into the sum of the squared components along the three perpendicular axes. Substituting that in gives the working result: pressure equals one third times the number density times the mass times the average of the squared speed, or P equals one third n m times average v squared. Multiplying both sides by the volume gives P V equals one third N m times average v squared, which can be rewritten as P V equals two thirds N times one half m times average v squared, that last piece, one half m times average v squared, being nothing other than the average kinetic energy of a single molecule's motion. Pressure times volume, in other words, is exactly two thirds of the total kinetic energy of every molecule bouncing around inside.

Extending the picture: depth, temperature, and confinement

This same collision picture, extended slightly, predicts how pressure responds to changing conditions. Confinement: squeezing the same number of molecules into a smaller volume raises the number density n directly, so more collisions strike each patch of wall every second, and pressure rises, exactly what is felt compressing a bicycle pump with the outlet blocked. Temperature: a hotter gas has faster molecules, a larger average squared speed, so both the force per collision and the rate of collisions increase, and pressure rises at fixed volume, exactly what a can of compressed gas shows if warmed. Depth in a fluid follows a related but distinct argument: pressure at depth in a liquid increases because each layer must support the weight of every layer of fluid above it, a static-equilibrium argument rather than a collision-rate argument, yet the underlying cause is the same molecular pushing, now transmitted downward through the fluid's own weight rather than driven by faster motion. In a liquid, molecules are packed close enough that they are also weakly attracting and briefly repelling their neighbors at every instant rather than flying freely between rare collisions, but the same basic fact holds: pressure is what a surface feels from the momentum constantly being delivered to it by the matter in contact with it.

Lineage

The idea that a gas's pressure comes from countless molecular impacts rather than from some static repulsive fluid was proposed by Daniel Bernoulli in 1738, well before atoms were experimentally confirmed, as part of an early kinetic account of gases. It was set aside for over a century while chemistry favored other models, then revived and made rigorous in the 1850s and 1860s by Rudolf Clausius, who introduced the idea of molecular mean free path, and by James Clerk Maxwell, who derived the distribution of molecular speeds in a gas at a given temperature. Ludwig Boltzmann completed the statistical framework in the following decades, and the picture was placed beyond reasonable doubt when Jean Perrin's early twentieth century experiments on Brownian motion confirmed the real, countable existence of the molecules the theory required all along.

The strongest case for it

The derivation above requires nothing but Newton's laws applied to elastic collisions and a definition of pressure as force per area, yet it correctly predicts, without any adjustable fitting constant, how pressure scales with density, with temperature, and with molecular mass. It explains why lighter gases at the same temperature and pressure move faster on average, why pumping more gas into a fixed container raises pressure, and why heating a sealed container is dangerous. It also generalizes cleanly: the same momentum-transfer logic underlies the pressure of light on a surface, the pressure exerted by a beam of particles, and the recoil force in a rocket exhaust, all cases where pressure is fundamentally about momentum delivered per unit time and area rather than about any single named force.

The strongest case against it

The derivation assumes molecules that are effectively point particles exerting force on each other only during instantaneous, perfectly elastic collisions, an assumption that is excellent for dilute gases but breaks down as a gas is compressed toward the density of a liquid, where molecules spend most of their time within reach of a neighbor's attractive or repulsive force rather than flying freely between brief collisions. It also assumes the number of molecules is large enough that individual jolts blur into a smooth average; for a handful of molecules in a nanoscale container, pressure fluctuates measurably from instant to instant, and the smooth picture no longer applies. A common misconception is imagining pressure as some fluid-like substance pushing outward on its own; there is no pushing agent beyond the molecules themselves, and a vacuum exerts no pressure precisely because there is nothing left to deliver momentum. Another misconception is expecting the depth argument and the collision-rate argument to be the same calculation; they answer different questions, one about a static column of weight, the other about a dynamic exchange of momentum, and conflating them leads to sign errors in either case.

Where it stands now

The kinetic derivation of pressure from molecular collisions is settled, foundational physics, confirmed to high precision across dilute gases of every composition and extended successfully, with appropriate corrections for molecular size and intermolecular attraction, to real gases and liquids. It underlies the ideal gas law, the design of pressure vessels, weather and climate modeling, and the physics of stars, where pressure from particle collisions, and in extreme cases from light itself, holds a star's own gravity at bay.

Test yourself

A sealed rigid metal cylinder holds a fixed amount of gas at room temperature and a measured pressure. Predict, with a one sentence justification grounded in the collision picture above rather than a memorized rule, what happens to the pressure in each of these three separate changes: the cylinder is placed in direct sunlight and warms significantly while its volume stays fixed; a valve is opened and half the gas is allowed to escape into a much larger evacuated tank, after which the valve is closed; and the same original amount of gas is instead compressed into a cylinder of half the original volume at the original temperature. For each case, say whether the number of collisions per second, the average force per collision, or both, are responsible for the change you predict.