Skip to content

Module 04

Tunnelling and scattering

Watch a wave packet hit a wall it cannot climb, and find part of it on the other side.

What you will see

  • The classical ball rebounds along its own path while the packet splits into reflected and transmitted parts
  • Inside the barrier |ψ|² decays exponentially — it is not zero
  • Doubling the barrier width costs several orders of magnitude in transmission

Assumed background

  • Stationary states and exponential solutions in the forbidden region (module 03)
  • A wave packet as a superposition of momenta

Module 03 ended with an exponential tail: in the classically forbidden region the wavefunction is not zero, merely decaying.

Now make the wall thin. The tail has not finished decaying by the time it reaches the far side, so an outgoing wave appears beyond the wall. The particle has gone through.

Tunnelling: a wave packet hits a wall it cannot climb

A classical ball with E < V₀ always bounces back. Watch the wave packet hit the barrier instead: part reflects, part decays exponentially inside the wall, and part genuinely turns up on the far side.

Loading 3D scene…

Barrier
2.40
1.60
Incident wave packet
1.60
4.5

A wider σ means a sharper momentum, and a transmission closer to the plane-wave formula

E < V₀: classically forbidden

Evolution
t = 0.0
Speed
×1.0

The vertical scale is pinned to the initial packet, so the transmitted blob really is that small. Magnifying only helps you see its shape.

Transmission T(E)
0E = 4.8

The blue line is the current energy, the grey line the barrier height. Note that T does not equal 1 once E > V₀: it oscillates towards 1, because the waves reflected at the two faces interfere (resonant transmission).

Numerical: reflected R / transmitted T0.000 / 0.000
Analytic (plane wave) T = 6.04e-2κ = √(2(V₀−E)) = 1.265Remaining norm 1.0000
Te2κa,κ=2m(V0E)T \approx \ee^{-2\kappa a},\qquad \kappa = \frac{\sqrt{2m(V_0-E)}}{\hbar}
  • ψ (complex tube — the twist is the phase)
  • |ψ|²
  • barrier V(x)
  • Classical particle

What to look for

  • With the default settings E < V₀. Press play and keep your eyes on the right of the barrier: a small blob of probability density emerges from the wall and keeps travelling. Its area is the transmission T.
  • Pause the moment the packet is pressed against the barrier and switch to the barrier close-up. Inside the wall |ψ|² is not zero but decays as e^{−2κx}. The wider the wall, the less gets through — that is where T ≈ e^{−2κa} comes from.
  • Take the width from 1.6 to 3.2 — merely doubling it — and see how many orders of magnitude T loses. That exponential sensitivity is why a scanning tunnelling microscope can resolve a single atomic step.
  • Push E just above V₀. Classically everything should pass, yet some of the packet still bounces back. "Quantum reflection" is the other face of wave behaviour.
  • Switch on the classical particle: the ball rebounds along its own path and never appears on the right. Run the two side by side once and the difference speaks for itself.

"The particle borrows energy from the wall and climbs over"

Nothing is borrowed. In a stationary state the energy is definite, and the wavefunction is simply non-zero in the forbidden region — a decaying real exponential, not a climb. Treating "borrowed energy" as a real mechanism makes calculations go wrong immediately.

"Inside the barrier the particle has imaginary velocity"

There is no well-defined classical velocity inside the barrier. The wave number k = iκ is purely imaginary, which means decay, not propagation; talking about "velocity" there has already left the domain of the theory.

"Transmission falls off linearly with width"

It falls exponentially. Double the width and T is squared. That distinction is the basis of every application of tunnelling — STM, α decay, stellar fusion.

Think it through

  1. A larger σ (sharper momentum) brings the numerical T closer to the plane-wave formula; a small σ pulls them apart. Why? (Hint: a packet is a superposition of many k, each with its own T.)
  2. Half-lives for α decay run from 10⁻⁷ seconds to 10¹⁷ years — more than thirty orders of magnitude — while the energies differ by only a few MeV. Explain that ferocious sensitivity with T ≈ e^{−2κa}.
  3. Is T exactly zero for an infinitely high barrier? Push V₀ to its maximum, then reconsider where the boundary condition of the infinite square well comes from.

This scene really is solving the equation

The earlier scenes had either closed-form solutions or eigenvalue problems. This one is different: it integrates the time-dependent Schrödinger equation

itψ=[22mx2+V(x)]ψ\ii\hbar\,\partial_t\psi = \left[-\frac{\hbar^2}{2m}\partial_x^2 + V(x)\right]\psi

frame by frame in your browser, using the split-operator method: the evolution operator is split symmetrically,

eiH^dt/eiV^dt/2  eiT^dt/  eiV^dt/2+O(dt3),\ee^{-\ii\hat H \dd t/\hbar} \approx \ee^{-\ii\hat V \dd t/2\hbar}\;\ee^{-\ii\hat T \dd t/\hbar}\;\ee^{-\ii\hat V \dd t/2\hbar} + O(\dd t^3),

with the kinetic part diagonal in momentum space, so an FFT there and back does the work. The scheme is exactly unitary: run it for ten thousand steps and the norm does not drift. That is what the “remaining norm” readout is watching (the absorbing layer at the edges does show up honestly as a decrease).

The algorithm is documented in appendix C.

Transmission: exponentially sensitive

For a plane wave on a rectangular barrier with E<V0E<V_0 the exact result is

T=[1+V02sinh2(κa)4E(V0E)]1,κ=2m(V0E).T = \left[1 + \frac{V_0^2\sinh^2(\kappa a)}{4E(V_0-E)}\right]^{-1}, \qquad \kappa = \frac{\sqrt{2m(V_0-E)}}{\hbar}.

When κa1\kappa a \gg 1, sinh12eκa\sinh \approx \tfrac12\ee^{\kappa a} and

T16E(V0E)V02e2κa.T \approx \frac{16E(V_0-E)}{V_0^2}\,\ee^{-2\kappa a}.

Exponential in the width and in V0E\sqrt{V_0-E}. That single fact underlies three apparently unrelated things:

PhenomenonMechanism
An STM resolving single atoms0.1 nm of tip height changes the tunnelling current by an order of magnitude
α-decay half-lives spanning 30 orders of magnitudeA few MeV of energy, exponentially amplified
The Sun burning at allProtons only get through the Coulomb barrier by tunnelling

Above the barrier: quantum reflection and resonances

Push the energy above the barrier height. Classically everything passes, yet part of the packet still comes back — because an abrupt change in potential reflects waves, exactly as light partially reflects entering glass.

Keep raising the energy and TT does not climb monotonically to 1: it oscillates towards it, hitting exactly 1 whenever k2a=nπk_2 a = n\pi. That is destructive interference between the waves reflected at the two faces — resonant transmission (the Ramsauer–Townsend effect is its three-dimensional cousin).

The T(E)T(E) curve in the control panel draws all of this, with the current working point marked.

Think it through

  1. Take the barrier width from 1.6 to 3.2 and read off how much TT falls. Predict the ratio from e2κa\ee^{-2\kappa a} first, then compare.
  2. As the barrier height goes to infinity, T0T\to 0. How does that connect to the infinite square well’s boundary condition that ψ vanishes at the wall?
  3. Reflection and transmission should sum to one, yet the readout often falls short. Look at the middle segment of the bar labelled “still near the barrier” and explain the deficit.

Go deeper · matching textbook sections

The 3D scenes build the picture; the full derivations and exercises live in the textbook.

Having finished this module