Skip to main content
GIOPHYSICS
Menu
← University Physics IV

University Physics IV · Limits of Classical Physics · 1.1

The Classical Synthesis & Its Limits

Before you meet a single quantum, learn to predict where you will need one. This lesson sets out the three-part account classical physics had assembled by 1900, then builds the dimensionless ratios v/c, hν/kT and λ/d that tell you, from numbers you already have, which part of it is about to fail.

Open the lab ↓Practice & worked examples ↓
01

Build the model

Connect the measurement to the mechanism.

By 1900 physics held three books that between them claimed everything: Newton's laws with universal gravitation for how matter moves, Maxwell's four equations for how fields propagate, and the Boltzmann–Gibbs machinery for what to do when there are too many particles to track. The synthesis is genuinely closed — hand it every position, momentum, charge and current at one instant and it returns the future, with equipartition supplying ½kT per quadratic degree of freedom whenever only an ensemble can be described. Its cost is three hidden assumptions: that speeds are small compared with c, that energy is a continuous variable divisible without limit, and that a particle has a trajectory, so position and momentum are simultaneously definite.

Each fails in a regime you can identify in advance by forming a dimensionless ratio — v/c for the first, hν/kT for the second, and λ/d for the third, where λ = h/p is the de Broglie wavelength and d is the length the system actually cares about. Each ratio is built from quantities you can already measure, so the boundary can be drawn before anyone knows what lies beyond it. That is all this topic claims.

Naming where a theory stops is not the same as replacing it, and correspondence runs the other way: any successor is obliged to reproduce the classical answers wherever these ratios are small.

Simple definition
The classical synthesis is the closed predictive account formed by Newtonian dynamics, Maxwell's electrodynamics and statistical mechanics; its domain of validity is the region in which the ratios v/c, hν/kT and λ/d are all small.
Example
Room air at 300 K: a nitrogen molecule has v/c = 1.7 × 10⁻⁶, its rotational levels are spaced at hν/kT ≈ 0.02, and its 28 pm de Broglie wavelength is 0.008 of the 3.4 nm mean spacing. All three tiny — so the ideal-gas law works.
The synthesis in three linesF = dp/dt · ∇×B = μ₀J + μ₀ε₀ ∂E/∂t · S = k ln W

Give it the state now and it returns every later observable — that closure is what makes the boundary worth locating.

Newton for matter, Maxwell for fields, Boltzmann for the counting; the Lorentz force joins the first two.

Relativistic ratio ββ = v/c · γ = (1 − β²)(−1/2)

½mv² is 0.75% low at β = 0.1 and 19% low at β = 0.5 — the price of Galilean kinematics.

v in m s⁻¹; c = 2.998 × 10⁸ m s⁻¹; β and γ are dimensionless.

Thermal ratio for a modex = hν/kT · ⟨E⟩ = hν/(ex − 1)

x ≪ 1 gives ⟨E⟩ → kT and equipartition; x ≫ 1 leaves ⟨E⟩/kT = x/(ex − 1) ≈ x e⁻ˣ, which is 1.4 × 10⁻⁴ at x = 11.3.

h = 6.626 × 10⁻³⁴ J s; k = 1.381 × 10⁻²³ J K⁻¹; ν in Hz and T in K.

Crossover frequency at temperature Tν* = kT/h · λ* = hc/kT

Modes well below ν* behave classically. Visible light at 300 K sits at x = 87, which is why nothing in the room glows.

At 300 K, ν* = 6.25 THz and λ* = 48 μm; at 6000 K, λ* = 2.4 μm.

de Broglie length against system sizeΛ = λ/d · λ = h/p · thermal form h/√(3mkT)

Λ = 0.008 for air at 300 K, Λ = 27 for electrons in copper. One gas is ideal, the other degenerate.

d is the length the system cares about: a spacing n(−1/3), a slit width, a well. Use p = γmv once β is not small.

Action measured in units of hS = ∮ p dq ≫ h ⇔ n ≫ 1

The single criterion behind all three ratios, and the form the correspondence principle takes.

S in J s; h = 6.626 × 10⁻³⁴ J s is the quantum of action.

01

One closed account, assembled by 1900

Three theories, each complete in its own domain and joined at the edges. Newton's laws with universal gravitation take a state — every position and momentum now — and return every later state by integrating F = dp/dt. Maxwell's four equations do the same for fields, and hand back a wave speed c = 1/√(ε₀μ₀) = 3.00 × 10⁸ m s⁻¹ assembled from two bench constants. Statistical mechanics closes the gap: when tracking 10²³ trajectories is hopeless, S = k ln W and the equipartition theorem's ½kT per quadratic degree of freedom deliver the thermodynamics anyway. The Lorentz force F = q(E + v×B) stitches the first two together. Call this the classical synthesis. It is not a slogan about determinism; it is a working prediction machine, and for planets, bridges, engines, radio and dilute gases it still returns the right number to more figures than most experiments can check.

02

v/c: the ratio that breaks the kinematics

The synthesis assumes velocities add the way Galileo said, which is what would let a light wave have a different speed in a moving frame. Maxwell's equations disagree: they contain c and no frame for it to be measured against. The dimensionless β = v/c decides which of the two matters here. Expand the relativistic kinetic energy: (γ − 1)mc² = ½mv²(1 + ¾β² + …), so the classical value is low by 0.75% at β = 0.1 and by 19% at β = 0.5. Concretely, an electron pushed through 100 V reaches β = 0.0198, and there the classical formula overstates β by only 0.015%; pushed through 100 kV, the classical √(2eV/m) predicts β = 0.626 where the true value is 0.548 — 14% too fast, and above 256 kV it returns speeds greater than c. Keep β under about 0.1 and Newtonian dynamics survives; past that, the corrections are the physics.

03

hν/kT: the ratio that breaks equipartition

Equipartition assumes energy is continuous, so a mode can accept any amount however small and therefore averages kT in equilibrium whatever its frequency. Compare hν, the smallest lump the mode can actually take, with kT, the energy on offer: x = hν/kT. At 300 K, kT/h = 6.25 THz, so any mode above roughly 6 THz — any wavelength shorter than 48 μm — is starved. Nitrogen makes the point inside one molecule. Its rotational levels sit at x ≈ 0.02 and behave classically; its 2359 cm⁻¹ vibration sits at x = 11.3, and its mean energy hν/(ex − 1) is 1.4 × 10⁻⁴ of kT, one part in seven thousand of what equipartition promises. So the measured molar heat capacity of N₂ at room temperature is 5R/2 = 20.8 J K⁻¹ mol⁻¹, not the 7R/2 = 29.1 J K⁻¹ mol⁻¹ that counting degrees of freedom predicts. Heat it to 3000 K, x drops to 1.13, and the vibration wakes up.

04

λ/d: the ratio that breaks the trajectory

The third assumption is that a particle has a definite path, so position and momentum can both be sharp at once. De Broglie's λ = h/p sets the scale on which that fails, and what matters is λ compared with d, the length the system actually cares about — a slit, a well, a mean spacing. A nitrogen molecule in room air carries λ = 28 pm against a mean spacing of 3.4 nm, so Λ = 0.008 and the ideal-gas law survives. Conduction electrons in copper are the counter-example. With n = 8.5 × 10²⁸ m⁻³ the spacing is 0.23 nm, while the thermal wavelength h/√(3mkT) for an electron at 300 K is 6.2 nm, giving Λ = 27. That 6.2 nm is itself a classical estimate, and the fact that it dwarfs the spacing is exactly the signal that the estimate had no right to be made: the electrons overlap, classical state counting is simply wrong, and the electronic heat capacity measured at 300 K is about 1% of the 3k/2 per electron that Drude's classical gas assumed.

05

Action in units of h, and correspondence

The three ratios are three faces of one comparison: the system's characteristic action against Planck's constant, h = 6.63 × 10⁻³⁴ J s. For a bound orbit that action is S = ∮ p dq. The old quantum condition of Bohr and Sommerfeld set it to an integer multiple of h, and the semiclassical estimate S ≈ (n + ½)h refines the same idea; neither is a law of quantum mechanics, but both say the classical regime is S/h ≫ 1 — equivalently, large quantum numbers. That is what makes the boundary one-way. Relativity must return Newton as β → 0, and Planck's mean mode energy hν/(ex − 1) must return kT as x → 0; expand the exponential, ex − 1 ≈ x, and it does. Any successor theory is obliged to reproduce the classical answers wherever these ratios are small, because that is exactly where the classical answers are already known to be right to several figures. The correspondence principle is not politeness; it is a hard constraint on what a new theory is permitted to say.

06

Reading the ratios before you choose a model

The procedure is short. Write down the system's speed v, the frequency ν (or level spacing ΔE/h) of the mode you care about, the temperature T, and the length d that confines it. Form β = v/c, x = hν/kT and Λ = λ/d. If all three sit well below 1 — say under 0.1 — use classical physics and expect errors of a few percent or less. If one exceeds 1, that ratio names which part of the account has failed and therefore which repair to reach for: Lorentz kinematics, quantized modes, or a wave description of the particle. Two can fail together and often do; inside a hot plasma β and x are both large. And a small ratio is a licence, not a guarantee — the classical model can still fail for reasons no ratio here tracks, such as the indistinguishability of identical particles or exponential sensitivity to initial conditions.

02

Change one variable at a time

Make the relationship visible.

Interactive model
0.005 c
3.00
300 K

Start at the defaults, d = 1 μm and 300 K: only the hν/kT bar clears the solid line, so it is the cavity radiation, not the electron, that classical physics gets wrong. Now take d down to 1 nm and v down to 0.001c to push λ/d over as well, then run v up to 0.99c and watch v/c climb past the dashed line while λ/d drops back.

Interactive physics modelThree ratios for one system: an electron of speed v inside a cavity of size d at temperature T, whose lowest mode is taken as ν = c/d. Each bar is log₁₀ of one ratio, measured from the solid line at 1, and the de Broglie length uses the relativistic λ = h/(γmv) so the third bar stays honest as v approaches c. Here log₁₀(v/c) = −2.30, log₁₀(hν/kT) = 1.68 and log₁₀(λ/d) = −3.31. Below the dashed 0.1 line the classical answer is good to a few percent; above the solid 1 line it is not even approximately right. v/c is the one ratio that can never exceed 1, so for kinematics the dashed line is the threshold that matters.one electron, one cavity, one temperaturev = 0.005 c d = 1000 nm T = 300 Kabove dashed: percent errorsabove solid: classical fails10⁴10.110⁻⁴v / chν / kTλ / d

log₁₀ (v/c)-2.30

log₁₀ (hν/kT)1.68

log₁₀ (λ/d)-3.31

LARGEST log₁₀ RATIO1.68

Live interpretationlog₁₀ (v/c): −2.30. log₁₀ (hν/kT): 1.68. log₁₀ (λ/d): −3.31. LARGEST log₁₀ RATIO: 1.68

03

Catch the common trap

Explain before calculating.

Copper at 300 K holds 8.5 × 10²⁸ conduction electrons per cubic metre, so their mean spacing is 0.23 nm, while an electron's thermal de Broglie wavelength at 300 K is 6.2 nm. Electrons at the Fermi surface move at 1.6 × 10⁶ m s⁻¹. Which reading of the ratios is correct?

Choose an answer to test the model.

04

Practice & worked examples

Reason from the model, then test the result.

EasyAn electron starts from rest and is accelerated through a potential difference V. Use the classical relation to find v/c for V = 100 V and for V = 100 kV, and in each case decide whether classical dynamics is good enough. Take the electron rest energy mc² = 511 keV.
  1. Classically eV = ½mv², so v = √(2eV/m) and β = v/c.
  2. V = 100 V: v = √(2 × 1.602 × 10⁻¹⁹ × 100 / 9.109 × 10⁻³¹) = 5.93 × 10⁶ m s⁻¹, so β = 5.93 × 10⁶ / 2.998 × 10⁸ = 0.019784.
  3. Check it against the exact result: γ = 1 + eV/mc² = 1 + 100/511000 = 1.00019569, so β = √(1 − γ⁻²) = 0.019781. The classical value is high by 0.015%, so keep it.
  4. V = 100 kV: the same classical formula gives v = 1.88 × 10⁸ m s⁻¹, β = 0.626. But eV/mc² = 0.1957 is no longer small: γ = 1.1957 and the exact β = √(1 − γ⁻²) = 0.548.
  5. The classical answer overstates the speed by 0.626/0.548 = 1.14, i.e. by 14%, and above V = mc²/2 = 256 kV it returns β greater than 1. Once β passes about 0.1, use relativistic kinematics.

Answerβ = 0.0198 at 100 V, where the classical value is high by only 0.015%. At 100 kV the classical β = 0.626 against a true 0.548 — 14% too fast, so v/c has taken this electron out of the classical domain.

MediumNitrogen's vibrational mode has wavenumber 2359 cm⁻¹. Find x = hν/kT at 300 K and at 3000 K, and use it to explain why the molar heat capacity of N₂ at room temperature is 5R/2 = 20.8 J K⁻¹ mol⁻¹ rather than the 7R/2 = 29.1 J K⁻¹ mol⁻¹ that equipartition predicts.
  1. Convert the wavenumber: ν = cν̃ = 2.998 × 10¹⁰ cm s⁻¹ × 2359 cm⁻¹ = 7.07 × 10¹³ Hz, so hν = 6.626 × 10⁻³⁴ × 7.07 × 10¹³ = 4.69 × 10⁻²⁰ J = 0.293 eV.
  2. At 300 K, kT = 1.381 × 10⁻²³ × 300 = 4.14 × 10⁻²¹ J = 0.0259 eV, so x = 4.69 × 10⁻²⁰ / 4.14 × 10⁻²¹ = 11.3.
  3. Equipartition would award this mode a full kT (½kT kinetic, ½kT potential). Its actual mean energy is hν/(ex − 1) = 4.69 × 10⁻²⁰ / (8.19 × 10⁴) = 5.7 × 10⁻²⁵ J, which is 1.4 × 10⁻⁴ of kT.
  4. So the vibration is frozen out. Only three translations and two rotations (x ≈ 0.02, classical) are active: CV, m = (5/2)R = 20.8 J K⁻¹ mol⁻¹, which is what is measured.
  5. At 3000 K the ratio scales as 1/T: x = 11.3 × 300/3000 = 1.13, and hν/(ex − 1) = 4.69 × 10⁻²⁰/2.10 = 2.23 × 10⁻²⁰ J, now 0.54 of kT = 4.14 × 10⁻²⁰ J.
  6. Heat capacity wakes faster than energy: in the harmonic model the mode contributes x²ex/(ex − 1)² = 0.90 R, so CV, m has already climbed to about 3.40R = 28 J K⁻¹ mol⁻¹, most of the way to 7R/2.

Answerx = 11.3 at 300 K, so the vibration holds 1.4 × 10⁻⁴ of kT and is frozen: CV, m = 5R/2 = 20.8 J K⁻¹ mol⁻¹. At 3000 K, x = 1.13, the mode carries 54% of kT, and its 0.90R contribution lifts CV, m to about 28 J K⁻¹ mol⁻¹.

HardHelium gas sits at P = 1.00 atm. Using λ = h/√(3mkT) for the thermal de Broglie wavelength and d = (kT/P)¹⁄³ for the mean spacing, find the temperature at which Λ = λ/d reaches 1, and say what that implies about reaching the quantum regime in a real gas.
  1. At fixed P the two lengths move opposite ways: λ = h(3mkT)(−1/2) shrinks as T(−1/2) while d = (kT/P)¹⁄³ grows as T¹⁄³, so Λ ∝ T(−5/6) and Λ = 1 has exactly one solution.
  2. Set λ = d and clear fractions: hP¹⁄³ = (3m)¹⁄²(kT)¹⁄²(kT)¹⁄³ = (3m)¹⁄²(kT)⁵⁄⁶, hence kT = [hP¹⁄³/√(3m)]⁶⁄⁵.
  3. Numbers for ⁴He: m = 6.65 × 10⁻²⁷ kg, √(3m) = 1.41 × 10⁻¹³, P¹⁄³ = 46.6, so the bracket is 6.626 × 10⁻³⁴ × 46.6 / 1.41 × 10⁻¹³ = 2.19 × 10⁻¹⁹ in SI units.
  4. kT = (2.19 × 10⁻¹⁹)1.2 = 4.06 × 10⁻²³ J, so T = 4.06 × 10⁻²³ / 1.381 × 10⁻²³ = 2.9 K. Check it: at 2.9 K both λ and d come out at 0.74 nm.
  5. But ⁴He at 1 atm liquefies at 4.2 K, above 2.9 K. One classical failure pre-empts the other, and the vapour never reaches Λ = 1.
  6. To get there as a gas you must cut n, not only T. At n = 10¹⁹ m⁻³ the spacing is 0.46 μm and, for ⁸⁷Rb, Λ = 1 arrives near 3 × 10⁻⁷ K — the dilute, sub-microkelvin conditions in which Bose–Einstein condensates are actually made.

AnswerT = 2.9 K, where λ = d = 0.74 nm. Helium at 1 atm condenses at 4.2 K first, so Λ = 1 is unreachable in the vapour; laboratory gases reach it by going dilute (n ≈ 10¹⁹ m⁻³) and cold (≈ 3 × 10⁻⁷ K).