Uncertainty

Not a limit on measurement, and not about clumsy instruments knocking things. A fact about waves — which turns out to be the reason atoms have a size and matter does not collapse.

1927 – today18 min

By the end you should be able to

State the uncertainty relation and explain it as a property of waves rather than of measurement
Say what is wrong with the microscope story Heisenberg himself told
Derive the size of a hydrogen atom from uncertainty alone

In 1927 Heisenberg published the relation that carries his name: Δx · Δp ≥ ħ/2 The spread in position times the spread in momentum has a floor. You cannot have both sharp. And in the same paper he offered an explanation of why — which is wrong, which Bohr told him was wrong before publication, and which has been in textbooks ever since.

A line of typeset mathematics from Heisenberg’s 1927 paper, giving the commutation rule for position and momentum. — The commutation rule, as it appears in Heisenberg’s 1927 paper. The relation follows from *this* — from the algebra of the operators — and not from any story about photons knocking electrons about. That is precisely what Bohr made him concede in the note added in proof, at the end of the same paper.

The explanation is the gamma-ray microscope. To see where an electron is, you must bounce something off it — there is no other way to look. To locate it precisely you need short-wavelength light, because you cannot resolve detail finer than the wavelength you are using. But short wavelength means high momentum — that is de Broglie — so the photon you bounce off the electron delivers a substantial kick. The more precisely you locate it, the harder you hit it, and the less you know about its momentum afterwards. It is a vivid argument. It feels like an explanation. And it gets the physics backwards.

Bohr read the draft and objected immediately and forcefully. The two argued for weeks — Bohr had a well-earned reputation for wearing people down, and by several accounts Heisenberg was reduced to tears at one point. Bohr's objection was that the microscope story describes a disturbance: a physical kick delivered by an act of measurement. And that is not what the mathematics says. Heisenberg was made to add a note in proof, acknowledging Bohr's criticism and conceding that the relation follows from the formalism rather than from the microscope. The note is in the published paper. Almost nobody reads it. The microscope has been taught for a century.

You might think

Uncertainty arises because measuring a particle inevitably disturbs it — our instruments are too clumsy to catch both quantities at once.

Actually

This says the electron has a definite position and a definite momentum, and we are prevented from learning both. That is a claim about ignorance, and it is not what the theory says. The theory says a state with definite momentum does not have a position — not unknown, absent; there is no value there to be ignorant of. Three things show the disturbance story is the wrong account. It is a property of the state, computable from ψ with no measurement anywhere in the calculation. It is not confined to quantum mechanics — the identical relation holds for sound and radio waves, where nobody invokes disturbance. And measurement disturbance is a real but separate effect, with its own relation, first properly formulated by Ozawa in 2003 and measured in 2012; the two can be distinguished experimentally, and they are different.

Squeeze the position spread and watch the momentum content broaden. Nothing here is being measured.

Loading the wavefunction…

A Gaussian packet and its momentum content, side by side. Narrow in one means broad in the other, with the product pinned at exactly ħ/2 — the smallest any state can achieve. This is computed from the shape of ψ alone; no measurement occurs anywhere in it.

This is not a quantum fact

The relation between the width of a wave and the range of wavelengths it contains is ordinary mathematics, known long before quantum theory and true of every wave there is. A pure tone must last a long time. A very short click has no definite pitch — it contains all pitches at once. Halve its duration and you double the spread of frequencies in it. Any audio engineer knows this; it is why you cannot have a filter that is both perfectly sharp in frequency and instantaneous in time. Radio has the same constraint. A transmission confined to a narrow frequency band cannot be brief; a brief pulse necessarily splatters across the spectrum. Regulators write rules about it. None of that is mysterious and nobody explains it by saying the receiver disturbs the signal. Quantum mechanics adds exactly one thing: de Broglie's relation, p = h/λ. Wavelength is momentum. Substitute that into a fact about waves that everyone already knew, and out comes the uncertainty principle. The strangeness is not in the relation. It is in the claim that a particle is described by a wave in the first place.

Squeeze the atom and watch the confinement cost explode. Then find the minimum.

Loading the wavefunction…

Confine an electron to a region r and it must carry momentum of order ħ/r, so squeezing costs kinetic energy rising as 1/r². The Coulomb attraction only deepens as 1/r. The competition has a minimum — and that minimum is the atom.

The experiment

The liquid that will not freeze

Heike Kamerlingh Onnes and successors; explanation from Simon and London in the 1930s · Liquefied 1908; the anomaly established through the 1920s–30s · Leiden, and subsequently everywhere with a cryostat

The question: Uncertainty forbids a confined particle from being perfectly at rest, since zero momentum is a perfectly definite momentum. Is that residual motion real enough to have macroscopic consequences?
The apparatus: Helium cooled towards absolute zero at ordinary pressure. Every other substance known solidifies when cooled far enough — thermal motion ceases and the atoms lock into a lattice. Helium is the test case because it is very light and very weakly bound, so its zero-point motion is unusually large compared with the forces holding it together.

Theory predicted

Classically, cooling removes kinetic energy; at absolute zero there is none left and any substance must be solid. Helium should freeze like everything else.

They measured

It does not. Helium remains liquid all the way to absolute zero at atmospheric pressure, and can only be solidified by applying about 25 atmospheres. The residual motion that keeps it liquid is not thermal — it persists when there is no thermal energy left to remove.

How sure could they be? Unambiguous rather than precise: the substance is either solid or it is not. The quantitative test is the pressure required to solidify it, which agrees with calculations of the zero-point energy for helium’s mass and interatomic potential — and with the fact that the heavier isotope ³He behaves differently from ⁴He, as its larger zero-point motion requires.

Why it mattered

Zero-point motion is not a bookkeeping device. It is the reason a bottle of liquid helium stays liquid at temperatures where nothing else could, and it is a direct macroscopic consequence of the uncertainty relation — a confined particle cannot be at rest, because being at rest would mean having an exactly defined momentum.

And why spectral lines have a width

There is a second version of the relation, between energy and time: ΔE · Δt ≥ ħ/2. It resolves something left dangling in Act 5. A state that only exists for a limited time cannot have a perfectly sharp energy — and an excited atom decays, typically within a few nanoseconds. So every spectral line has an irreducible width. For hydrogen's Lyman-α line, with an excited-state lifetime of about 1.6 ns: - natural linewidth: about 100 MHz - against a line frequency of 2,466,000,000 MHz - a fractional width of 4 parts in 10⁸ The lines of Act 5 are sharp — extraordinarily sharp, which is why spectroscopy works at all — but they are not infinitely sharp, and no improvement in instruments will make them so. A short-lived state has a fuzzy energy, necessarily. This is also useful rather than merely limiting: measure a linewidth and you have measured a lifetime, for states far too short-lived to time directly. Much of what is known about unstable particles comes from exactly this.

1807Fourier’s analysis establishes the width–bandwidth relation for waves. No physics of particles involved.
1927Heisenberg publishes the uncertainty relation, with the microscope argument.
1927Bohr objects; Heisenberg adds a note in proof conceding the point.
1929Robertson derives the general relation from the commutator, for any pair of observables.
1930sZero-point motion explains why helium does not freeze.
2003Ozawa formulates the separate measurement–disturbance relation.
2012Experiments confirm that measurement disturbance and the uncertainty relation are distinct, and can be violated independently.

Where this leaves us

Uncertainty is not an apology for the state of our instruments. It is a structural feature of a world described by amplitudes — and it is load-bearing. It sets the size of atoms, it keeps helium liquid at absolute zero, it gives spectral lines their width, and it is the reason matter does not collapse into its own nuclei. But it leaves one thing hanging, and we have now stepped around it in three consecutive lessons. The theory says an electron approaching two slits has no definite path. It also says that if you put a detector at one slit, you find the electron at exactly one slit, every time, whole. So what happens when you look? Schrödinger's equation is smooth, continuous and perfectly deterministic: give it ψ now and it gives you ψ later, with no probabilities anywhere. Then a measurement occurs, and something else happens — something the equation does not describe and does not contain. That gap has a name, it has been there since 1926, and it has not been closed.