34
Part One · Quantum mechanics

Tunneling and barriers

Bring a metal needle within a nanometre of a surface, with no contact, and an electric current flows across the empty gap. Pull the needle back by a tenth of a nanometre, about a third of an atom’s width, and the current falls about eightfold. That extreme sensitivity let Gerd Binnig and Heinrich Rohrer, in the early 1980s, build a microscope that sees individual atoms. The current flows by tunnelling: particles crossing regions they do not have the energy to enter.

On the plate, a wave packet meets a barrier higher than its energy. Part of it gets through anyway.

Through the wall

A classical ball rolling at a hill it cannot climb comes back, always. The wave packet on the plate has energy 0.8 and the barrier is 1.0 high. Set its width to 1.0 and .

Inside the barrier the wavefunction does not oscillate; it decays, as e−κxe^{-\kappa x} with κ=2m(V0−E)/ℏ\kappa = \sqrt{2m(V_0 - E)}/\hbar, just as it did inside the walls of the finite well in Chapter 33. If the barrier is thin, the decaying wave still has some amplitude at the far side, and there it becomes a travelling wave again. The readout counts how much probability ends up on the far side. Established

The exact answer

For a rectangular barrier of height V0V_0 and width dd, matching the wave and its slope at both edges gives the transmission probability:

T=[1+V02sinh⁡2(κd)4E(V0−E)]−1≈16E(V0−E)V02 e−2κdT = \left[1 + \frac{V_0^2\sinh^2(\kappa d)}{4E(V_0 - E)}\right]^{-1} \approx \frac{16E(V_0-E)}{V_0^2}\,e^{-2\kappa d}

The approximation holds once κd\kappa d is more than about one. The crucial part is the exponential: every extra 1/κ1/\kappa of thickness cuts transmission by e2≈7.4e^2 \approx 7.4. The plate plots TT against energy for the barrier you chose. Below the top, it rises steeply. Above the top, it is still not one: waves partly reflect even when they could pass classically, except at special energies where the barrier holds a whole number of half-wavelengths and becomes perfectly transparent.

Seeing atoms

In a scanning tunnelling microscope, electrons cross a vacuum gap whose barrier is the metal’s work function, about 4 eV, so κ≈10\kappa \approx 10 per nanometre. Widen the gap to 0.60 nm and watch the current.

Each 0.1 nm costs a factor of about 7.8. A feedback loop moves the tip up and down to hold the current constant as it scans, and the tip’s height traces the surface. Because almost all the current flows through the single atom at the tip’s apex nearest the surface, the resolution is atomic. Binnig and Rohrer shared the 1986 Nobel Prize in Physics for it.

Why some nuclei last for ever

In 1928 George Gamow, and independently Ronald Gurney and Edward Condon, explained alpha decay as tunnelling. An alpha particle inside a nucleus is held by a Coulomb barrier tens of MeV high, and it escapes with only 4 to 9 MeV. Because the escape probability is exponential in the barrier, a modest change in energy makes an enormous change in lifetime.

The plate shows six alpha emitters. Their energies differ by only a factor of two, but their half-lives run from 0.3 microseconds for polonium-212 to fourteen billion years for thorium-232: twenty-four powers of ten. Plotted against 1/E1/\sqrt{E}, the logarithm of half-life is nearly a straight line, the Geiger–Nuttall law found empirically in 1911 and explained by Gamow seventeen years later. This was the first triumph of quantum mechanics inside the nucleus. Established

A whole circuit tunnels

Tunnelling is not only for single particles. In 1984 and 1985, John Clarke, Michel Devoret and John Martinis cooled a superconducting circuit built around a Josephson junction, on a chip about a centimetre across, to a few hundredths of a degree above absolute zero. The circuit’s single collective variable, the phase difference across the junction, sat in a shallow dip of its potential, and it escaped by tunnelling, at the rate quantum theory predicted. They also showed that the circuit’s energy came in discrete levels, like an atom’s.

Billions of electrons, acting as one, behaved as a single quantum particle. That result won the 2025 Nobel Prize in Physics, and the same kind of circuit is the superconducting qubit of Part Three. Established

How long does it take?

This is still argued. Some definitions of tunnelling time give answers that do not grow with barrier thickness, the Hartman effect, which seems to suggest faster-than-light crossing, though no signal can be sent that way. In 2020 Ramón Ramos and colleagues in Toronto used the spin of rubidium atoms as a clock inside a barrier made of light and measured about 0.6 milliseconds spent inside: not zero. Attosecond experiments on electrons tunnelling out of atoms have been read both as nearly instant and as taking a finite time, depending on how the data are analysed.

Part of the difficulty is that there is no time operator in quantum mechanics (Chapter 30), so “how long” has several inequivalent answers. Debated

What you now have

Exponential decay under a barrier, the exact transmission formula and its exponential approximation, the scanning tunnelling microscope, Gamow’s theory of alpha decay, macroscopic tunnelling in circuits, and an open question about time.

· · ·

Three questions to carry with you.

What sets how fast transmission falls with barrier thickness?tap
κ = √(2m(V₀ − E))/ħ; transmission falls roughly as e^(−2κd).
Why do alpha half-lives span 24 orders of magnitude?tap
Escape is by tunnelling, which is exponential in the barrier; small energy differences become huge rate differences.
What did the 2025 Physics Nobel show?tap
That a macroscopic electrical circuit can tunnel and has quantized energy levels, as one quantum object.

And one to work out. An STM current falls by a factor of about 7.8 per 0.1 nm. By what factor does it fall if the gap widens by 0.2 nm?

factor ≈

Sources: G. Gamow, Z. Phys. 51, 204 (1928); R. W. Gurney and E. U. Condon, Nature 122, 439 (1928). G. Binnig, H. Rohrer et al., Phys. Rev. Lett. 49, 57 (1982). J. M. Martinis, M. H. Devoret and J. Clarke, Phys. Rev. Lett. 55, 1543 (1985); Nobel Prize in Physics 2025. R. Ramos et al., Nature 583, 529 (2020). Half-lives from NNDC. Checked by code: κ(4 eV) = 10.25 nm⁻¹, factor 7.76 per 0.1 nm; T(E = 0.8, V₀ = 1, d = 1) = 0.584; log₁₀ half-lives −6.5 (²¹²Po) to 17.6 (²³²Th).

Next · Chapter 35
The harmonic oscillator

A molecule can never stop vibrating, even at absolute zero.

Turn the page