How this instrument works
Rudolf Clausius introduced this quantity in 1858 to answer an objection to kinetic theory: if gas molecules really move at hundreds of metres per second, why does an opened bottle of perfume take minutes to cross a room? Because no molecule flies straight for long. Clausius swept out a target area πd² and counted hits, treating every other molecule as stationary; James Clerk Maxwell repaired that assumption two years later by averaging over relative velocities, and his correction is the √2 still standing in this denominator.
Substituting number density from an ideal gas law, n = P⁄k_BT, produces this working form: path length climbs with temperature, falls with pressure, and falls hard with collision diameter, which enters squared. Ordinary air near sea level yields tens of nanometres. Pump a chamber down to 1 Pa and molecules cover millimetres between hits; near 10⁻⁴ Pa they cover tens of metres, at which point walls rather than fellow molecules end most flights. Loschmidt exploited exactly that chain in 1865, combining measured free paths with liquid density to produce history's first credible estimate of molecular size.
Two assumptions carry all of that weight: molecules act as hard elastic spheres of one fixed size, and collisions stay binary. Both weaken at high density, where molecular volume itself matters and triple encounters appear, and for ions or electrons, whose long Coulomb reach makes any fixed diameter meaningless. Individual flights also scatter exponentially about λ, so roughly 37% of them run longer than average — read this figure as a mean, never as a limit.
- Enter your gas temperature into Absolute temperature (K) — kelvin, never Celsius; 20 °C becomes 293.15 K.
- Give Pressure as an absolute value, picking Pa, kPa, bar or atm from its unit menu. Gauge readings need one atmosphere added first.
- Set Molecular diameter (m) to a kinetic collision diameter: 3.7e-10 for air, 3.64e-10 for nitrogen, 2.6e-10 for helium.
- Read Mean free path, switching to nm for ambient gases, or to mm and m once you work under vacuum.
- Divide that result by vessel or feature size; anything above about 1 puts you in free-molecular flow rather than fluid flow.
Worked example — room air at sea level
Ordinary room air, 20 °C at sea level. Absolute temperature (K) = 293.15, Pressure = 101325 Pa, Molecular diameter (m) = 3.7e-10, an effective collision size covering nitrogen and oxygen together. Numerator k_B·T works out at 4.0474 × 10⁻²¹ J; denominator √2·π·d²·P works out at 6.1629 × 10⁻¹⁴. Dividing leaves 6.5673 × 10⁻⁸ m, or 65.67 nm.
Sixty-six nanometres runs to about 180 molecular diameters, and to roughly one eighth of green light's wavelength — molecular worlds are crowded, yet mostly empty. Mean speed near 463 m/s turns each flight into about 0.14 nanoseconds, some seven billion collisions every second. Ferocious speed, negligible progress: that is why perfume, and heat, crawl across still air.
Questions
Which molecular diameter should I enter?
Use a kinetic (collision) diameter taken from transport measurements, not a covalent radius or bond length. Common values: 3.64 × 10⁻¹⁰ m for nitrogen, 3.46 × 10⁻¹⁰ m for oxygen, 3.3 × 10⁻¹⁰ m for carbon dioxide, 2.6 × 10⁻¹⁰ m for helium, and 3.7 × 10⁻¹⁰ m as an effective figure for air. Care pays off here: diameter enters squared, so 10% off on width becomes 21% off in your result.
Does a heavier gas have a shorter mean free path?
Not through mass — mass never appears in this formula. Only diameter and number density fix path length, so helium, being small, gets furthest at a given pressure and temperature. Mass does govern mean molecular speed, and therefore mean free time: a helium atom covers its longer path faster than a xenon atom covers its shorter one.
Why is there a √2 in the denominator?
It accounts for motion of both partners in a collision. Clausius held every target molecule still in 1858, which undercounts encounters. Maxwell averaged over a Maxwell-Boltzmann spread of relative velocities in 1860 and found mean relative speed exceeding mean speed by exactly √2, shortening path length by that factor. Drop it and your answer runs 41% too long.
Is this valid for a mixture such as air?
Only approximately. A strict treatment gives each pair of species its own collision diameter, (d₁ + d₂)⁄2, plus its own reduced mass, then weights by partial pressures. For air, feeding one effective diameter near 3.7 × 10⁻¹⁰ m into this instrument reproduces measured values to within a few percent, which is plenty for vacuum work or flow-regime checks.
How does mean free path relate to vacuum quality?
Directly — vacuum ranges are often described by it. Divide path length by characteristic chamber size L to get your Knudsen number. Below 0.01, gas behaves as a continuous fluid; above roughly 10, molecules fly wall to wall without meeting, which is what sputtering and evaporation coating rely on so that arriving atoms are not deflected. Between those limits sit slip and transition flow, where neither picture works cleanly.
Does every molecule really travel this far between hits?
No. Free paths follow an exponential distribution, so short flights are commonest while about 37% (a fraction 1/e) exceed λ, and a few run several times longer. Treat λ as an arithmetic mean, good for transport coefficients and regime estimates, and misleading if read as a typical maximum.