How this instrument works
The Knudsen number compares two lengths: how far, on average, a gas molecule travels before it collides with another molecule — the mean free path — against the smallest dimension of whatever geometry the gas is moving through, whether that is a microchannel, a dust particle, or the leading edge of a spacecraft. Kn = λ ⁄ L is nothing more than that ratio. When λ is tiny next to L, a molecule collides with its neighbours thousands of times before it ever reaches a wall, and the gas behaves as the smooth, continuous fluid that the Navier-Stokes equations describe. When λ grows to rival or exceed L, molecules cross the gap and strike a boundary far more often than they strike each other, and the whole idea of a continuous fluid stops making sense.
The formula carries no separate constant because none is needed — it is a direct ratio, not an empirical fit. Martin Knudsen introduced it in 1909 while measuring gas flow through narrow tubes at low pressure for vacuum research, work that later gave 'Knudsen flow' and the Knudsen effusion cell their name in thin-film deposition. Engineers split the ratio into four working regimes: below 0.01 is continuum flow, where ordinary fluid dynamics and no-slip walls hold; 0.01 to 0.1 is slip flow, where the fluid picture survives but the boundary condition at a wall needs correcting; 0.1 to 10 is transition flow, genuinely hard territory; and above 10 is free-molecular flow, where kinetic theory and direct simulation replace continuum equations entirely.
Mean free path is not fixed to a substance — it depends on pressure and temperature through kinetic theory, roughly λ = kT ⁄ (√2 π d² p), where d is the molecule's collision diameter. Lower the pressure and molecules travel farther between collisions, so λ grows and Kn climbs even though nothing about the geometry changed at all. That is why the same 1 mm channel can sit safely in continuum flow at sea level and drift toward slip flow at the fringes of the atmosphere: shrinking a channel is only one way to raise the Knudsen number, and vacuum engineers reach the same regime shift by lowering pressure instead.
- Enter Mean free path — the average distance a gas molecule travels before its next collision; air at sea level sits near 68 nm.
- Enter Characteristic length — the smallest relevant scale of the geometry: a channel width, a particle diameter, or a nose radius.
- Read Knudsen number, the dimensionless ratio of the two lengths you just entered.
- Compare the result against 0.01, 0.1, and 10 to place the flow in the continuum, slip, transition, or free-molecular regime.
Worked example — air in a 1 mm microchannel
Air's mean free path at sea level is about 68 nm — enter 68 in the Mean free path field with its unit set to nm. Set Characteristic length to 1 mm, a typical microchannel bore. Converted to consistent units the calculation is Kn = 6.8 × 10⁻⁸ m ⁄ 1.0 × 10⁻³ m = 6.8 × 10⁻⁵, exactly what the Knudsen number field reads back.
6.8 × 10⁻⁵ sits three orders of magnitude below the 0.01 continuum threshold, so this channel is safely in continuum flow — the Navier-Stokes equations and ordinary no-slip walls apply without correction. Shrink Characteristic length to 1 nm instead, keeping the same air, and Kn climbs to 68: deep in free-molecular territory, where molecules strike the channel wall far more often than they strike one another and continuum fluid dynamics gives the wrong answer entirely.
Questions
What do the Knudsen number regimes actually mean?
They mark how well continuum fluid dynamics still describes the gas. Below 0.01 is continuum flow, where the Navier-Stokes equations and no-slip walls hold exactly. From 0.01 to 0.1 is slip flow, where the equations survive but the wall boundary condition needs a correction. From 0.1 to 10 is transition flow, genuinely hard to model. Above 10 is free-molecular flow, where individual molecule-wall collisions dominate and kinetic theory replaces continuum equations outright.
Why does mean free path change with pressure?
Because mean free path measures the average gap a molecule crosses before hitting another one, and that gap depends on how crowded the gas is. Kinetic theory gives λ = kT ⁄ (√2 π d² p): halve the pressure p and λ doubles, since each molecule has twice the empty space to travel through before a collision. This is why the same geometry can be firmly in continuum flow at sea level and rarefied at high altitude or inside a vacuum chamber.
Does the Knudsen number apply to liquids?
No. Mean free path is a kinetic-theory concept describing the wide spacing between gas molecules between collisions; in a liquid, molecules are packed shoulder to shoulder and the idea of a free path between collisions loses its meaning. The Knudsen number is a gas-dynamics tool — it governs when microfluidic gas flows, vacuum systems, aerosol particles, and high-altitude vehicles depart from ordinary continuum behaviour, not liquid flow.
Is a small characteristic length always the reason Kn is large?
No, and that is the most common misreading of this number. Kn is a ratio, so a large value can come from a small L, as in a MEMS microchannel, or from a large λ at low pressure, as inside a vacuum chamber or the thin upper atmosphere a re-entry vehicle passes through. Two systems built at wildly different physical scales can share the same Knudsen number and the same flow physics, provided that ratio matches.
Who actually uses the Knudsen number?
MEMS and microfluidics engineers check it before assuming standard fluid equations apply inside a channel a few microns wide. Vacuum-system and thin-film deposition engineers use it to size Knudsen effusion cells and predict flow through narrow tubing at low pressure. Aerospace engineers track it during spacecraft re-entry, where thinning atmosphere pushes flow from continuum toward free-molecular as altitude climbs. Aerosol scientists use it to compute the Cunningham slip correction for small particle drag.
What is the difference between the Knudsen number and the Reynolds number?
Both are dimensionless ratios describing a flow, but they answer different questions. Reynolds number compares inertial and viscous forces within a continuum fluid and predicts laminar versus turbulent behaviour. Knudsen number asks whether the continuum assumption itself is valid in the first place, by comparing molecular mean free path to geometry. A flow can carry any Reynolds number and still fail the continuum test if its Knudsen number is too high.