Breakdown of Classical Physics

In April 1900, Lord Kelvin — the most decorated physicist alive — told the Royal Institution that the "beauty and clearness" of physical theory was marred by only two clouds: the puzzle of the ether, and the failure of a statistical theorem called equipartition to explain how bodies store and radiate heat. Everything else, the community broadly felt, was bookkeeping. Newtonian mechanics ruled matter, Maxwell's equations ruled light, and Boltzmann's statistical mechanics welded the two to thermodynamics. Within thirty years, Kelvin's second cloud had not merely rained — it had washed the whole edifice away. This article is the case file. Four clean, reproducible experiments refused to fit: a warm cavity that classical theory said should blaze with infinite ultraviolet energy; metal surfaces that eject electrons according to light's color, not its brightness; atoms that by Maxwell's own rules should collapse in sixteen picoseconds yet are patently eternal, emitting sharp lines instead of a death wail; and solids whose heat capacity walks off the classical value as they cool. These are not small discrepancies: two are infinitely wrong, one is off by ten million in a tabletop furnace, and one contradicts the stability of everything you have ever touched. You already own all the classical tools these failures broke — equipartition, the Larmor formula, the wave theory of light. Here you will point them at the data and watch them fail quantitatively, which is the only honest motivation for the strange postulates ahead. One constant, $h$ , will turn out to sit behind all four cracks.

Key takeaways

By 1900, physics rested on Newton, Maxwell, and Boltzmann; a handful of experiments broke all three pillars — qualitatively, not by small corrections.
Blackbody radiation: equipartition gives every cavity mode $k_BT$ , the mode count grows as $\nu^2$ , so the predicted energy density diverges — the ultraviolet catastrophe.
Photoelectric effect: electron energies track the light's frequency, not its intensity, with a sharp threshold — impossible for a classical wave that stores energy in amplitude.
Atomic stability: a classically orbiting electron radiates (Larmor), spirals into the nucleus in $\sim 1.6\times10^{-11}\ \mathrm{s}$ , and emits a continuum — real atoms are stable and emit discrete Rydberg lines.
Specific heats: solids obey Dulong–Petit's $3R$ at high temperature but fall toward zero as they cool, violating equipartition again.
Key idea: nature has a quantum of action, $h \approx 6.626\times10^{-34}\ \mathrm{J\,s}$ — energy is exchanged in lumps $E = h\nu$ , and classical physics is the limit where the action dwarfs $h$ .

The core insight

Nature contains a fundamental quantum of action, Planck's constant $h \approx 6.626\times10^{-34}\ \mathrm{J\,s}$ : energy is exchanged between matter and radiation in discrete amounts $E = h\nu$ , not continuously. Classical physics is not wrong so much as it is the limiting case in which every relevant action is enormous compared to $h$ .

Where this lives today

The four anomalies never retired; they became instrumentation. The cosmic microwave background is the most perfect Planck curve ever measured; the photoelectric effect runs every camera sensor; spectral lines let astronomers read galactic compositions, and the cesium line defines the SI second. Most tellingly, since 2019, $h$ is no longer measured at all: it is fixed by definition at $6.62607015\times10^{-34}\ \mathrm{J\,s}$ , and the kilogram is realized from it via the Kibble balance — the constant born in a curve-fitting emergency now anchors the mass of everything. The deeper legacy is the diagnostic habit. Kelvin's framing — locate the clouds, take them seriously — is how physicists speak of dark matter, dark energy, and quantum gravity today. And the action ladder is now an engineering frontier: micromechanical drums and even LIGO's mirrors have been cooled toward their quantum ground states, deliberately dragging macroscopic objects down the ladder to $S \sim \hbar$ , where the old physics ends.

The physical picture

Think of energy as money that changes hands only in coins, the coin for a process of frequency $\nu$ costing $h\nu$ . For slow processes, the coin is so tiny that transactions look continuous — pour coffee, swing a pendulum, orbit a planet, and classical physics is exact to any measurable precision. For fast processes — ultraviolet light, the stiff vibrations of a light atom in a crystal — the coin becomes expensive, and a mode whose coin costs more than the thermal budget $k_BT$ almost never gets paid: it is frozen out. That one sentence, made quantitative, will cure two of the four anomalies on its own. Each anomaly attacks a different pillar: blackbody radiation breaks statistical mechanics (equipartition on field modes); the photoelectric effect breaks the wave theory of light (energy in amplitude); atomic collapse and line spectra break electrodynamics (accelerating charges must radiate); the specific-heat falloff breaks equipartition again, inside matter. Four pillars, four experiments — one constant, $h$ , repairs them all. That convergence is the strongest hint in the history of physics that one new theory, not four patches, was required.

Figure

The four anomalies — each panel is one broken pillar of classical physics

One diagnostic tells you in advance which physics applies: compare the characteristic action $S$ of a process — energy × time, or momentum × distance — against $h$ . Earth's orbit carries about $2.7\times10^{40}\ \mathrm{J\,s}$ , some $10^{74}\,\hbar$ ; discreteness is unimaginably invisible and Newton is exact. A one-gram pendulum bob at ten centimetres per second still sits near $10^{28}\,\hbar$ . But an electron bound in an atom has angular momentum of order $\hbar$ itself — and there classical reasoning does not lose accuracy, it produces nonsense: infinite energies, collapsing matter.

Figure

The action ladder — where classical physics is safe and where it lies

Anomaly 1 · The blackbody catastrophe

Heat a cavity — an oven with a pinhole — and the radiation leaking out has a spectrum that depends only on temperature, not on the oven's material, size, or shape. Kirchhoff proved that universality in 1859 and challenged theorists to find the function. Classical physics has exactly two ingredients to offer, and both seem beyond reproach. First, counting: the electromagnetic field in a cavity is a sum of standing-wave modes, and the number of modes per unit volume per unit frequency interval is a purely geometric quantity,

\rho_{\text{modes}}(\nu) = \frac{8\pi\nu^2}{c^3} \qquad (1)

where: $\rho_{\text{modes}}$ — modes per unit volume per unit frequency; $c$ — speed of light; the factor includes both polarizations. Read the structure: the count grows without bound as $\nu^2$ — there are always more ways to fit short waves than long ones. This counting survives into quantum theory untouched; it is the innocent half of the disaster. Second, filling: the equipartition theorem of classical statistical mechanics awards every quadratic degree of freedom an average energy $\tfrac12 k_BT$ , so each field mode — one kinetic-like and one potential-like quadrature — carries $k_BT$ regardless of its frequency. Multiply the two ingredients and integrate:

u_{\mathrm{RJ}}(\nu, T) = \frac{8\pi\nu^2}{c^3}\,k_BT, \qquad \int_0^\infty u_{\mathrm{RJ}}\,d\nu = \infty \qquad (2)

where: $u_{\mathrm{RJ}}$ — the Rayleigh–Jeans spectral energy density; $k_B = 1.381\times10^{-23}\ \mathrm{J/K}$ . The law fits the data beautifully at low frequency — and then climbs as $\nu^2$ forever, promising that every oven, and indeed your own body, should radiate an infinite blaze of ultraviolet and X-rays. Ehrenfest later named it the ultraviolet catastrophe.

Derivation AThe Rayleigh–Jeans law and the ultraviolet catastrophe

We want the classical blackbody prediction, built honestly from Maxwell plus equipartition, so the failure cannot be blamed on sloppy modelling. The route: count the standing-wave modes of a cubical cavity, convert the count to a density per unit frequency, then award each mode its equipartition energy $k_BT$ . Watch which ingredient causes the divergence — the filling rule, not the counting — because that is precisely the part Planck will replace. The mode count is purely geometric and is recalled, unchanged, in the Planck-law article that follows.

Step 1. Fit standing waves in a box. For a cube of side $L$ with reflecting walls, each field component must vanish at the boundaries, so the allowed wavevectors are $k_i = n_i\pi/L$ with $n_i = 1, 2, 3, \dots$ Each triple $(n_x, n_y, n_z)$ is one spatial mode. Why: the guitar-string condition, once per axis — boundary conditions discretize the wavevector, classically just as quantum-mechanically.

Step 2. Count modes up to frequency $\nu$ . Modes below $\nu$ fill one octant ( $n_i > 0$ ) of a sphere of radius $n = 2L\nu/c$ in $n$ -space, times two polarizations: $N(\nu) = 2 \cdot \tfrac18 \cdot \tfrac43\pi\,(2L\nu/c)^3 = \tfrac{8\pi\nu^3}{3c^3}L^3$ . Why: counting lattice points by volume is legitimate — at optical frequencies, the $n_i$ are enormous.

Step 3. Differentiate to a density. Per unit volume and frequency, $\rho_{\text{modes}}(\nu) = \tfrac{1}{L^3}\tfrac{dN}{d\nu} = \tfrac{8\pi\nu^2}{c^3}$ — equation (1), the cavity size dropping out. Why: universality appears already here; only the filling of each mode can carry the temperature.

Step 4. Fill each mode classically. Each mode is a harmonic oscillator; equipartition gives it $\langle E\rangle = 2 \times \tfrac12 k_BT = k_BT$ regardless of frequency, so $u_{\mathrm{RJ}} = (8\pi\nu^2/c^3)\,k_BT$ . Why: the fatal step — classical statistics cannot charge more for faster oscillation, so the growing mode count goes unpunished.

Step 5. Integrate and face the catastrophe. $\int_0^\infty \nu^2\,d\nu$ diverges: infinite energy density at any temperature, the excess piling up at high frequency. Why: classical physics owns no cutoff to stop it — the failure is structural, not numerical.

How bad is it in practice? In a pottery kiln at $1500\ \mathrm{K}$ , green light has $x = hc/\lambda k_BT \approx 19.2$ , and Rayleigh–Jeans over-predicts the energy density by $(e^x - 1)/x \approx 1.1\times10^{7}$ — ten million, at a wavelength you can see. Yet at $\lambda = 1\ \mathrm{cm}$ , $x \approx 10^{-3}$ and the classical law is essentially exact. The theory is broken only where the coin $h\nu$ becomes expensive — the clue everything hinges on. The repair is the business of the next article, Blackbody Radiation and the Planck Quantum.

Anomaly 2 · Light that pays in lumps

Shine ultraviolet light on a clean metal, and electrons pop out — the photoelectric effect, noticed by Hertz in 1887 while confirming Maxwell's wave theory. That theory (Maxwell's) makes three confident predictions: brighter light should eject faster electrons; any frequency should work at sufficient intensity; and feeble light should require minutes of soak time while an electron accumulates its escape energy. All three are wrong. The electron energies track frequency alone,

K_{\max} = h\nu - W, \qquad \nu_0 = \frac{W}{h} \qquad (3)

where: $K_{\max}$ — maximum kinetic energy of an ejected electron; $W$ — the work function, the escape cost from the metal; $\nu_0$ — a sharp threshold below which no intensity ejects anything. Read it as a receipt: energy arrives in packets of $h\nu$ , one packet per electron; intensity merely counts packets. Emission is instantaneous even in near-darkness, and the same slope $h$ appears for every metal. The full machinery — stopping potentials, Millikan's decade of hostile precision, the failed soak-time estimate — is developed in Photoelectric Effect and the Photon; here it stands as the second crack: a wave that stores energy in amplitude cannot produce a frequency threshold.

Anomaly 3 · Atoms should not exist

By 1911 Rutherford's scattering experiments had put a tiny positive nucleus at the atom's centre with electrons around it. Classically, an electron cannot hover — it must orbit, and an orbiting charge accelerates. Maxwell's equations are ruthless about accelerating charges: they radiate, with power given by the Larmor formula. The orbit therefore decays. This is not a slow leak you could ignore.

Go deeper · Derivation B · 1 minA classical atom dies in sixteen picoseconds

Derivation BA classical atom dies in sixteen picoseconds

We want the lifetime Maxwell's equations assign to a hydrogen atom — the sharpest way to state the stability crisis, because the answer is a definite, absurdly short number. The strategy: put the electron on a circular orbit of radius $r$ , let the Larmor formula drain its mechanical energy, convert that drain into a shrinkage rate $dr/dt$ , and integrate from the atomic radius down to zero. Keep an eye on the sign structure — the total energy of a bound orbit is negative and gets more negative as $r$ shrinks, which is why radiating energy means falling inward, faster and faster.

Step 1. Energy of a circular orbit. With $\kappa \equiv e^2/4\pi\varepsilon_0$ , the Coulomb force provides the centripetal acceleration, $\kappa/r^2 = mv^2/r$ , giving kinetic energy $\tfrac12 mv^2 = \kappa/2r$ and total energy $E(r) = -\kappa/2r$ . Why: the virial pattern — kinetic is half of potential in magnitude — makes the total negative: bound.

Step 2. Radiated power. The centripetal acceleration is $a = \kappa/mr^2$ , and Larmor gives $P = \dfrac{e^2 a^2}{6\pi\varepsilon_0 c^3} = \dfrac{2}{3}\dfrac{\kappa^3}{c^3 m^2 r^4}$ . Why: power scales as $a^2 \propto 1/r^4$ — the tighter the orbit, the more savage the radiation; the fall accelerates.

Step 3. Convert to a shrinkage rate. Energy balance $dE/dt = -P$ with $dE/dr = \kappa/2r^2$ gives $\dfrac{dr}{dt} = -\dfrac{4}{3}\dfrac{r_e^2 c}{r^2}$ , where $r_e = \kappa/mc^2 = 2.82\times10^{-15}\ \mathrm{m}$ is the classical electron radius. Why: bundling constants into $r_e$ exposes the scales — light speed and a nuclear-sized length set the rate.

Step 4. Integrate the fall. $r^2\,dr = -\tfrac43 r_e^2 c\,dt$ from $r = a_0 = 5.29\times10^{-11}\ \mathrm{m}$ to $r = 0$ :

t_{\text{collapse}} = \frac{a_0^3}{4\,r_e^2\,c} \approx 1.6\times10^{-11}\ \mathrm{s} \qquad (4)

Why: sixteen picoseconds — every atom in your body should have imploded in a hundredth of a nanosecond, radiating a continuous swan-song sweeping up the spectrum as the orbit tightened. *A more precise evaluation using CODATA constants gives $t_{\text{collapse}} \approx 1.556\times10^{-11}\ \mathrm{s}$ ; the initial radiated power at $r = a_0$ is $\approx 4.7\times10^{-8}\ \mathrm{W}$ .*

Reality is doubly insolent: atoms last forever, and they emit discrete lines, not a spiral-song. Hydrogen's visible lines had even been captured in a formula by a Swiss schoolteacher fitting numbers — Balmer's series, generalized by Rydberg to

\frac{1}{\lambda} = R\left(\frac{1}{n_1^2} - \frac{1}{n_2^2}\right), \qquad R = 1.097\times10^{7}\ \mathrm{m^{-1}} \qquad (5)

where: $n_1 < n_2$ — positive integers; $R$ — the Rydberg constant. Integers — in a spectroscopy formula! For $n_1 = 2$ it delivers the visible Balmer lines ( $656\ \mathrm{nm}$ , $486\ \mathrm{nm}$ , …, limit $365\ \mathrm{nm}$ ). Whatever replaces the death spiral must explain the stability and these integers. (Bohr's 1913 planetary patch reproduced $R$ and the ladder $-13.6\ \mathrm{eV}/n^2$ ; its story belongs to the Bohr-model article of the optics track — here it matters only as evidence that quantizing action in units of $h$ fit the lock.)

Anomaly 4 · Heat capacity walks away

The rule of Dulong and Petit (1819) says one mole of almost any elemental solid needs the same heat to warm it by one kelvin: about $25\ \mathrm{J\,mol^{-1}K^{-1}}$ . Classical statistical mechanics explains this with ease — and that ease becomes the fourth embarrassment.

Go deeper · Derivation C · 1 minDulong–Petit from equipartition and how it fails

Derivation CDulong–Petit from equipartition and how it fails

We want the classical molar heat capacity of a crystal — both why the Dulong–Petit law works at room temperature and why its low-temperature failure is fatal to equipartition. The model: each of the $N_A$ atoms in a mole vibrates about its lattice site in three dimensions — three oscillators per atom. Equipartition then fixes the energy budget with no free parameters, which is what makes the cold data so damning: there is nothing to adjust.

Step 1. Count the quadratic degrees of freedom. Each atom vibrates along $x$ , $y$ , $z$ ; each direction contributes one kinetic term $\propto p^2$ and one potential term $\propto x^2$ — six quadratic terms per atom. Why: equipartition pays per quadratic term in the energy, so the bookkeeping is just counting squares.

Step 2. Apply equipartition. Each term averages $\tfrac12 k_BT$ , so the molar internal energy is $U = N_A \cdot 6 \cdot \tfrac12 k_BT = 3N_Ak_BT = 3RT$ . Why: frequency never enters — a stiff bond and a soft bond store the same average energy classically. Remember this: it is the point of failure.

Step 3. Differentiate.

C_V = \frac{dU}{dT} = 3R = 24.9\ \mathrm{J\,mol^{-1}K^{-1}} \qquad (6)

Why: a universal constant, independent of material and of temperature — matching Dulong–Petit's observation across dozens of metals at room temperature.

Step 4. Confront the cold data. Diamond already violates the rule at room temperature ( $C_V \approx 6\ \mathrm{J\,mol^{-1}K^{-1}}$ , four times too small), and every solid's $C_V$ slides toward zero as $T \to 0$ . Why: diamond's stiff bonds and light atoms make its vibrational quanta expensive — at $300\ \mathrm{K}$ most of its oscillators cannot afford one coin and sit frozen; soft, heavy lead has cheap quanta and obeys the rule. The threshold is again $h\nu \sim k_BT$ — the blackbody currency, spent inside matter. Einstein made this quantitative in 1907; the Debye refinement lives in the statistical-mechanics track. *Using the CODATA value of $R$ , the Dulong–Petit prediction is $3R = 24.94\ \mathrm{J\,mol^{-1}K^{-1}}$ ; diamond and lead room-temperature heat capacities are consistent with standard tabulated values.*

Limiting cases

Case	Result	Meaning
$h\nu \ll k_BT$	Planck's mean mode energy $\to k_BT$	equipartition and Rayleigh–Jeans recovered — classical physics is the cheap-coin limit
$S \gg h$	quantum levels crowd into a continuum	correspondence principle: classical trajectories emerge for macroscopic actions
$k_BT \gg h\nu_{\text{vib}}$	$C_V \to 3R$	Dulong–Petit restored at high temperature; freeze-out is invisible when coins are cheap
photon number $n \gg 1$	field mode $\to$ classical wave	graininess $h\nu$ is a negligible fraction of the mode energy; Maxwell recovered
$h \to 0$ formally	every anomaly vanishes	all four failures are proportional to the size of the quantum — one constant is behind everything

Common trap“The Bohr model explains the atom”

The old quantum theory — Planck's quantized oscillators, Einstein's heat-capacity fix, Bohr's orbits — is a set of brilliant patches, not a theory, and this article's anomalies are motivation, not doctrine. The Bohr model in particular gets hydrogen's energies right for the wrong reasons: it keeps classical orbits and simply forbids most of them, predicts the wrong angular momentum for the ground state, cannot handle helium or any two-electron system, and offers no account of why action comes in units of $h$ . Treat every semiclassical fix you meet here as a clue pointing at the real explanation — matter waves, the wavefunction, and the Schrödinger equation, which will re-derive the successes and discard the scaffolding. If an exam asks why atoms are stable, "electrons orbit without radiating because Bohr said so" is a historical statement, not a physical one; the honest answer arrives with standing waves and stationary states.

Worked examples

1 · Auditing the catastrophe in a kiln (confidence builder)

A pottery kiln sits at $T = 1500\ \mathrm{K}$ . Compare the Rayleigh–Jeans prediction with the true (Planck) spectral energy density at $\lambda = 500\ \mathrm{nm}$ (green light) and at $\lambda = 1\ \mathrm{cm}$ (microwaves). The over-prediction factor is $(e^x - 1)/x$ with $x = hc/\lambda k_BT$ .

Step 1. Coin-to-budget ratio at $500\ \mathrm{nm}$ : $x = \dfrac{hc}{\lambda k_BT} = \dfrac{(6.626\times10^{-34})(3.00\times10^8)}{(5.00\times10^{-7})(1.381\times10^{-23})(1500)} = 19.2$ — the photon coin costs 19 thermal budgets.

Step 2. Over-prediction factor: $\dfrac{e^{19.2} - 1}{19.2} \approx \boxed{1.1\times10^{7}}$ — the catastrophe is not a subtlety at infinity; it is a factor of ten million at a visible wavelength.

Step 3. At $\lambda = 1\ \mathrm{cm}$ : $x = 9.6\times10^{-4}$ , $(e^x - 1)/x \approx 1.0005$ . Classical theory is accurate to 0.05% in the microwave.

Step 4. The failure switches on precisely where $h\nu \gtrsim k_BT$ : any repair must leave cheap coins untouched and tax expensive ones into silence — exactly what the Planck factor does in the next article.

2 · The integers hiding in hydrogen light (exam level)

Using the Rydberg formula (5) with $R = 1.097\times10^{7}\ \mathrm{m^{-1}}$ , find the first two Balmer lines and the series limit, and state what a classical spiralling electron would emit instead.

Step 1. Balmer means $n_1 = 2$ . For $n_2 = 3$ : $\dfrac{1}{\lambda} = R\left(\dfrac14 - \dfrac19\right) = R\,\dfrac{5}{36}$ , so $\lambda = \dfrac{36}{5R} = \boxed{656\ \mathrm{nm}}$ — the red H $\alpha$ line.

Step 2. For $n_2 = 4$ : $\dfrac{1}{\lambda} = R\left(\dfrac14 - \dfrac{1}{16}\right) = R\,\dfrac{3}{16}$ , so $\lambda = \dfrac{16}{3R} = \boxed{486\ \mathrm{nm}}$ — the blue-green H $\beta$ .

Step 3. The series limit is $n_2 \to \infty$ : $\lambda_\infty = \dfrac{4}{R} = 365\ \mathrm{nm}$ , just into the ultraviolet. Lines crowd together as they approach it — a fingerprint no classical model produces.

Step 4. Contrast: the Larmor spiral of equation (4) predicts a continuous sweep lasting sixteen picoseconds, then silence. Observation shows eternal atoms emitting an integer-patterned comb. Stability and integers alike must come from the new mechanics — and will, as standing-wave conditions on matter waves (de Broglie Matter Waves).

Origins · 1 minLord Kelvin and how this was discovered

Lord Kelvin

1824–1907 · Glasgow

William Thomson, Baron Kelvin, was Victorian physics incarnate: he laid the transatlantic telegraph cable, formulated the second law of thermodynamics, and had the absolute temperature scale named for him in his lifetime. His April 1900 lecture "Nineteenth-Century Clouds over the Dynamical Theory of Heat and Light" named just two blemishes on an otherwise finished physics: the ether problem, and the equipartition failures this article catalogues. The irony is total — cloud one dissolved into special relativity, cloud two into quantum mechanics, the two revolutions that ended the classical era. Kelvin died in 1907, having seen Planck's constant born but not what it would become. His lecture's lesson: the clouds are never decoration; they are the weather arriving.

History

1859

— Kirchhoff proves the blackbody spectrum is universal and challenges theory to derive it.

1885

— Balmer fits hydrogen's visible lines with an integer formula; Rydberg generalizes it in 1888.

1887

— Hertz, while confirming Maxwell's waves, stumbles on the photoelectric effect.

1897

— J. J. Thomson identifies the electron, making atomic structure a live problem.

1900

— Kelvin's "two clouds" lecture in April; Planck's quantum fix of the blackbody spectrum in December.

1907

— Einstein applies quantization to solids, explaining the collapse of Dulong–Petit at low temperature.

1911

— Rutherford's nucleus sharpens the stability crisis; the first Solvay Conference convenes on "Radiation and the Quanta."

Going further

Texts. Griffiths & Schroeter, 3rd ed., Ch. 1 + historical afterword · Shankar, 2nd ed., the failure-of-classical-physics chapter · Cohen-Tannoudji, Diu & Laloë, Vol. I, Ch. I · Townsend, Ch. 1 · Eisberg & Resnick, Ch. 1–4. Key numerical results in this article: mode density $8\pi\nu^2/c^3$ from the octant count; kiln over-prediction $(e^x-1)/x \approx 1.1\times10^7$ at $500\ \mathrm{nm}$ , $1500\ \mathrm{K}$ ; collapse time $a_0^3/4r_e^2c \approx 1.6\times10^{-11}\ \mathrm{s}$ ( $r_e = 2.818\times10^{-15}\ \mathrm{m}$ ); Balmer lines $656\ \mathrm{nm}$ and $486\ \mathrm{nm}$ , series limit $365\ \mathrm{nm}$ , from $R = 1.097\times10^7\ \mathrm{m^{-1}}$ ; $3R = 24.94\ \mathrm{J\,mol^{-1}K^{-1}}$ ; Earth-orbit action $\sim 2.5\times10^{74}\,\hbar$ ; and $K_{\max} = h\nu - W$ (Einstein 1905). All constants taken from CODATA values of $h$ , $k_B$ , and $R$ .

Prerequisites. Classical electromagnetism (Maxwell's equations) · thermodynamics and statistical mechanics · Newtonian mechanics and the harmonic oscillator. Leads to. Blackbody Radiation and the Planck Quantum · Photoelectric Effect and the Photon · Wave–Particle Duality · de Broglie Matter Waves.

If classical physics predicts infinite energy in every oven, why don't ovens explode?

Because the prediction is false, not because ovens are clever. The divergence is a statement about the classical theory: apply equipartition to the cavity field and you compute infinite energy density — a proof by absurdity that equipartition cannot govern high-frequency modes. Real cavities obey the Planck law, which taxes expensive quanta into silence. The catastrophe never happened anywhere in nature; its value is forensic, showing classical statistical mechanics fails structurally, by an infinite amount, not by a percent-level correction.

Was classical physics simply wrong all along, and if so why did it work for two centuries?

It is the $S \gg h$ limit of the true theory, and everything humans handled before 1900 lives deep in that limit — a pendulum carries $\sim10^{28}\,\hbar$ , so discreteness sits fifty orders below any measurement. The anomalies appeared only when experiments first probed processes with $S \sim h$ : light emission by single electrons, vibrations of single atoms. That is also why quantum mechanics must and does reduce to classical mechanics in the correspondence limit — the old theory's successes are data the new one has to reproduce.

Why is the same constant h behind all four anomalies rather than four separate fixes?

That is the deepest observation here, and historically the argument that convinced people a revolution — not repairs — was underway. Planck's blackbody fit, Einstein's photoelectric slope, the freeze-out scale, and (via Bohr) the Rydberg constant each yield a measured $h$ , from unrelated experiments — and they agree. Four coincidentally equal patch-constants would be an absurd conspiracy; one quantum of action explains the agreement automatically. When independent phenomena triangulate on one number, that number is structural.

Is the ultraviolet catastrophe really about ultraviolet light?

The name marks where the divergence accumulates, not a special property of UV. The Rayleigh–Jeans density grows as $\nu^2$ , so the error grows without bound toward high frequency; for kiln-temperature bodies it becomes grotesque around the visible-to-ultraviolet, hence Ehrenfest's name. For colder bodies the "catastrophe" region slides toward the infrared. The physics is one sentence: classical statistics cannot charge more for faster modes, and there are infinitely many fast modes.

The collapse calculation uses classical formulas — is it legitimate to trust it?

That is precisely its power: it is an internal consistency test, using only Coulomb's law, circular mechanics, and the Larmor formula — all impeccably classical. If classical physics were the whole story, its own equations would forbid atoms; it refutes itself the moment Rutherford's nucleus is accepted. The calculation describes nothing real (atoms do not spiral); it is a reductio ad absurdum, and its sharpness — sixteen picoseconds, not "eventually" — made the stability crisis impossible to shelve.

Where exactly does the number 16 picoseconds come from — why not much faster or slower?

From three constants arranged as $t = a_0^3/4r_e^2c$ : the atom's size enters cubed because the spiral accelerates as $1/r^2$ , the classical electron radius $r_e \approx 2.8\ \mathrm{fm}$ sets the Larmor drain, and $c$ converts field stiffness into a rate. Feed in the numbers: $1.6\times10^{-11}\ \mathrm{s}$ , nothing tuned. A classical hydrogen atom is not marginally unstable; it is dead before one of your nerve impulses completes.