Graham's Law of Effusion

What Graham's Law tells you

Poke a hole in a container of gas that is smaller than the average distance a molecule travels between collisions — a hole a few tens of nanometers across — and the gas leaks out one molecule at a time. That process is effusion. Graham's Law makes a clean, quantitative prediction about how fast it happens:

    Rate of effusion  ∝  1 / √M          (M = molar mass)

    For two gases at the same T and P:

        Rate₁       √M₂        √(molar mass of gas 2)
        ─────  =  ─────  =  ──────────────────────────
        Rate₂       √M₁        √(molar mass of gas 1)

The rule is compact but the payload is surprising: a molecule 4× heavier effuses only half as fast, and a molecule 100× heavier effuses only one-tenth as fast. The mass dependence is weak — it goes as the square root, not linearly — which is exactly why isotope separation by effusion is so agonizingly slow and why it's such a good exam question. Get the two masses under the radical the right way up (heavier gas on top when the lighter gas is faster) and every effusion problem collapses to one division and one square root.

The kinetic-theory mechanism, step by step

Graham's Law is not an empirical curve fit that happens to work; it falls straight out of the kinetic molecular theory of gases. Here is the chain of reasoning, each link doing one job:

1. Equal energy, not equal speed. Temperature is average translational kinetic energy. At a shared temperature, every gas — helium, oxygen, sulfur hexafluoride — has the same mean ½mv². Write it per molecule: ½m⟨v²⟩ = (3/2)k_BT.
2. Solve for speed. Rearranging, ⟨v²⟩ = 3k_BT/m, and the root-mean-square speed is v_rms = √(3k_BT/m) = √(3RT/M). The heavy molecule, carrying the same energy, is forced to move slower — its speed drops as 1/√M.
3. Counting hits on the hole. The rate at which molecules strike a patch of wall (the flux) equals ¼·n·v̄, where n is the number density and v̄ is the mean speed. Every molecule that hits the open hole passes through. So the effusion rate is proportional to v̄.
4. Speed carries the mass dependence. Because v̄ (like v_rms) scales as 1/√M, the flux — and therefore the effusion rate — inherits that same 1/√M scaling. Number density n and temperature are held equal for both gases, so they cancel in the ratio, leaving pure molar mass.

The single physical idea underneath all four steps: light molecules are fast molecules, and fast molecules find the exit more often. Nothing about the hole's chemistry matters — only how frequently a molecule happens to be heading at it, which is set by speed, which is set by mass.

    ½ m ⟨v²⟩ = (3/2) k_B T        (equipartition — same energy for all gases)
         │
         ▼
    v_rms = √(3RT / M)            (speed ∝ 1/√M)
         │
         ▼
    flux = ¼ n v̄                 (hits on the hole ∝ mean speed)
         │
         ▼
    Rate of effusion ∝ v̄ ∝ 1/√M   (Graham's Law)

Conditions and where the "exact" law lives

Graham's Law is exact only inside a specific regime. Know the fine print or your predictions drift:

• The hole must be small — molecular-effusion regime. The aperture diameter has to be much smaller than the mean free path λ (the average gap between collisions, ~70 nm for air at 1 atm and 25 °C). Then molecules leave individually, in the free-molecular (Knudsen) regime, and the flux formula ¼n v̄ applies. A hole bigger than λ produces hydrodynamic outflow (a jet), where the gas moves as a bulk fluid and the 1/√M law breaks down.
• Same temperature and pressure for both gases. The 1/√M ratio only isolates molar mass because n (from P and T) and T are matched. Compare gases at different pressures and you must fold the density ratio back in.
• Ideal-gas, non-interacting behavior. The molecules are treated as point masses with no attractions. Near condensation, or for very polar or associating gases, real behavior deviates — but for common gases far from their boiling points the law is remarkably good.
• Molar mass is the whole story. Hole shape, wall material, and molecular size (as long as it's ≪ λ) drop out. That's what makes the law so clean: two isotopologues with identical chemistry differ only in mass, and effusion still separates them.

The equations, and where each constant comes from

The full toolkit for effusion problems, with the standard gas constant R = 8.314 J·mol⁻¹·K⁻¹:

    Root-mean-square speed:   v_rms = √(3RT / M)
    Mean (average) speed:     v̄     = √(8RT / πM)
    Most probable speed:      v_p    = √(2RT / M)

    Graham's Law (rate ratio):     Rate₁ / Rate₂ = √(M₂ / M₁)

    Equivalent time-to-effuse form (heavier is slower, takes longer):
        t₂ / t₁ = Rate₁ / Rate₂ = √(M₂ / M₁)

    Molar-mass ratio for isotopes (single-stage separation factor):
        α = √(M_heavy / M_light)

All three speeds (v_rms, v̄, v_p) carry the same 1/√M dependence — they differ only by numerical prefactors (√3 ≈ 1.732, √(8/π) ≈ 1.596, √2 ≈ 1.414). That's why it doesn't matter which "average speed" you invoke: the mass scaling is identical, and it's the mass scaling that Graham's Law captures.

Worked example: an unknown gas and a stopwatch

A classic lab problem. A fixed volume of oxygen (O₂, M = 32.00 g/mol) takes 45 s to effuse through a pinhole. The same volume of an unknown gas, under identical T and P, takes 90 s. Identify the gas.

    Longer time  →  slower rate.  Rate is 1/time.

        Rate_O₂ / Rate_unk  =  t_unk / t_O₂  =  90 / 45  =  2

    Graham's Law:

        Rate_O₂ / Rate_unk  =  √(M_unk / M_O₂)
                        2   =  √(M_unk / 32.00)
                        4   =  M_unk / 32.00
                  M_unk     =  4 × 32.00  =  128 g/mol

    A molar mass of 128 g/mol  →  sulfur dioxide? (SO₂ = 64) no.
    128 g/mol matches HI (127.9) — hydrogen iodide.

The gas effuses half as fast, so it is four times as heavy (because the mass sits under a square root). The answer, 128 g/mol, points to hydrogen iodide. Two things to internalize: slower means heavier, and the factor gets squared going from rate ratio to mass ratio (a 2× rate difference is a 4× mass difference, a 3× rate difference is a 9× mass difference).

A second quick one — pure numbers, no unknown. How much faster does helium (4.00) effuse than argon (39.95)?

    Rate_He / Rate_Ar = √(39.95 / 4.00) = √9.99 ≈ 3.16

    Helium empties about 3.2× faster — which is exactly why a helium
    party balloon goes limp overnight while an air-filled one stays firm.

Effusion vs Diffusion (and where Graham's Law fits)

	Effusion	Diffusion
What happens	Gas escapes through a hole into vacuum/low pressure	Gas spreads through another gas (or air)
Hole vs mean free path	Hole ≪ mean free path (molecules leave singly)	No hole — bulk mixing throughout the space
Dominant physics	Wall-collision flux, ¼n v̄	Molecule-molecule collisions, random walk
Molecule collisions matter?	No — free-molecular regime	Yes — collisions set the step length
Graham's 1/√M law	Exact	Approximate (also depends on cross-section)
He vs O₂ rate ratio	√(32/4) = 2.83, cleanly observed	Roughly √(M) but blurred by collisions
Timescale for a room	Instant through the aperture	Minutes to hours across a room (still air)
Textbook use	Molar-mass determination, isotope separation	Smell traveling across a room, gas mixing

The common textbook demo — cotton balls of ammonia (NH₃, M = 17) and hydrochloric acid (HCl, M = 36.5) at opposite ends of a glass tube, meeting nearer the HCl end to form a white NH₄Cl ring — is technically a diffusion experiment, so it only approximately obeys Graham's Law. The observed ratio (about 1.3, vs the predicted √(36.5/17) ≈ 1.47) is close but not exact, precisely because molecule-molecule collisions slow the advancing fronts. If you want the law to nail it, you need genuine effusion through a pinhole.

Real-world application: separating uranium isotopes

Graham's Law's most consequential application is gaseous-diffusion uranium enrichment — the method that supplied the fissile material for the first reactors and weapons of the 1940s.

• The gas. Uranium is turned into uranium hexafluoride, UF₆, the only uranium compound that is a gas at convenient temperatures (it sublimes at 56.5 °C). Fluorine is monoisotopic (only ¹⁹F), so the only mass difference between molecules comes from the uranium.
• The masses. ²³⁵UF₆ has M = 235 + 6(19) = 349 g/mol; ²³⁸UF₆ has M = 238 + 6(19) = 352 g/mol. A gap of just 3 out of 352.
• The separation factor. The ideal single-stage factor is α = √(352/349) = √1.00860 ≈ 1.0043 — the light isotopologue effuses only 0.43% faster per pass through a porous nickel barrier.
• The cascade. To climb from 0.72% natural ²³⁵U to 3–5% reactor grade, you multiply that 1.0043 factor over more than a thousand stages, each stage's slightly-enriched stream feeding the next. The Manhattan Project's K-25 plant in Oak Ridge, Tennessee, was a U-shaped building enclosing 44 acres under one roof — at the time the largest building in the world — running thousands of stages and consuming a significant fraction of U.S. electrical output.

Gaseous diffusion has since been retired in favor of gas centrifuges, which exploit the same mass difference far more efficiently (centrifugal separation scales with the mass difference rather than the square-root ratio, so it needs orders of magnitude fewer stages). But the fact that a 0.43% speed advantage could be leveraged — through sheer cascade multiplication — into weapons-grade material is the most dramatic demonstration of Graham's Law ever built.

Limitations and common mistakes

• Applying it to diffusion as if it were exact. The 1/√M law is exact only for effusion. For diffusion, cross-sectional size and collision frequency also matter, so predictions are approximate. Don't expect the ammonia/HCl ring to land exactly at the Graham's-Law position.
• Inverting the radical. The single most common error: writing Rate₁/Rate₂ = √(M₁/M₂) instead of √(M₂/M₁). Sanity-check with the physics — the lighter gas must come out faster. If your formula predicts the heavy gas is faster, flip the ratio.
• Confusing rate ratio with mass ratio. The factors are related by a square. A gas that effuses 3× faster is 9× lighter, not 3× lighter. Forgetting to square (or square-root) the conversion is a frequent slip.
• Using it on a big hole. Once the aperture exceeds the mean free path, you have hydrodynamic flow (a jet), governed by pressure drop and viscosity, not by Graham's Law. Puncturing a tire is not effusion.
• Ignoring gas mixtures. For a mixture effusing together, each component effuses at its own rate, so the escaping stream is enriched in the lighter component and the residue in the heavier — which is the whole basis of isotope separation, but a trap if you assume the composition stays constant.

Who discovered it, and when

The law is named for Thomas Graham (1805–1869), a Scottish physical chemist working in Glasgow and later at University College London. In an 1846 paper (following earlier work on the diffusion of gases dating to 1829–1833), Graham reported that the rates at which different gases passed through a fine plug or a small aperture were inversely proportional to the square roots of their densities. Because a gas's density at fixed T and P is proportional to its molar mass, "inversely proportional to √density" and "inversely proportional to √molar mass" are the same statement.

Graham had no molecular theory to explain why — the kinetic theory of gases that grounds the law in equipartition of energy was not fully developed until the work of Clausius, Maxwell, and Boltzmann in the 1850s–1870s, a decade or more after Graham's measurements. His result was purely experimental, and it stood because the underlying physics turned out to be exactly right. Graham is doubly remembered in chemistry: alongside this law, his study of the slow passage of substances through membranes gave us the word dialysis and the distinction between crystalloids and colloids.