3.1.3 Wave-Particle Duality#
Prompts
How is the optical path length an ‘action’ for light? What is the analogous quantity in classical mechanics?
State the de Broglie relations. What do they imply about matter?
What does the equation \(\psi = A \mathrm{e}^{\mathrm{i}S/\hbar}\) mean physically?
How does the path integral \(\sum \mathrm{e}^{\mathrm{i}S/\hbar}\) reduce to classical mechanics when \(S \gg \hbar\)?
Lecture Notes#
Overview#
This lesson unifies two classical frameworks: geometric optics (ray picture) and classical mechanics (particle picture). Both obey the same variational principle—Fermat’s principle for light, Hamilton’s principle for particles. The bridge between them is de Broglie’s insight: matter has waves, and the phase of the matter wave is the classical action.
Optical Path as Action#
To avoid symbol overload, we use distinct notation:
optical path length: \(L = \int n\,\mathrm{d}s\)
mechanical action: \(S = \int L_{\mathrm{mech}}\,\mathrm{d}t\) (Lagrangian: \(L_{\mathrm{mech}}\))
Then the variational principles are directly parallel:
Geometric Optics |
Classical Mechanics |
|---|---|
Light ray |
Particle trajectory |
Fermat’s principle: \(\delta L = 0\) |
Hamilton’s principle: \(\delta S = 0\) |
Optical path \(L = \int n\,\mathrm{d}s\) |
Action \(S = \int L_{\mathrm{mech}}\,\mathrm{d}t\) |
Ray optics (short wavelength) |
Classical limit (\(\hbar \to 0\)) |
The physical picture is the same: the observed path is selected by a stationary variational quantity.
De Broglie: Matter Waves#
In 1924, de Broglie proposed that matter, like light, has wave-particle duality.
De Broglie Relations
which associate:
particle properties: momentum \(\boldsymbol{p}\) and energy \(E\), with
wave properties: wave vector \(\boldsymbol{k}\) and angular frequency \(\omega\).
This unifies particle and wave descriptions. Experimental confirmation (Davisson & Germer, 1927): electrons diffracted by crystals exhibit interference patterns with \(\lambda = h/p\), confirming de Broglie’s hypothesis.
Action = Phase: The Punchline#
In wave optics, the physically observed ray is selected by stationary phase. From §3.1.2,
Just as Fermat’s principle (§3.1.1) governs geometric optics, here we see that the same variational structure appears in mechanics: classical particle trajectories satisfy \(\delta S = 0\). This parallel is no accident; it reveals a deep unity.
For monochromatic light, \(k\) is constant, so
Hence \(\delta\Phi=0 \Leftrightarrow \delta L=0\): the stationary-phase condition is exactly Fermat’s principle. Because this is a variational selection rule, a quantity proportional to \(\Phi\) plays the same role for light as action does in mechanics.
Now compare with mechanics, where the path is selected by
So optics and mechanics have the same structure if we identify an action-like quantity for light with
The proportionality constant is fixed below using matter-wave kinematics. For a free matter wave,
using de Broglie \(\boldsymbol{p}=\hbar\boldsymbol{k}\) and \(E=\hbar\omega\), with classical principal function increment
Therefore,
so the proportionality constant is \(\hbar\), and the universal phase-action relation is \(S=\hbar\Phi\) (up to an additive constant in phase).
So the quantum wavefunction takes the form
This is the unification: action controls phase, and stationary action emerges as the stationary-phase limit.
What is Action?#
Intuitively, action is the scalar functional assigned to a candidate history of a system. Classically, physical histories satisfy a stationary-action condition. Quantum mechanically, each history contributes with phase \(\Phi=S/\hbar\), so action becomes the unifying principle connecting classical dynamics and quantum interference.
The action takes different forms across physics:
System |
Action |
Meaning |
|---|---|---|
Light ray |
\(\hbar k\int n\,\mathrm{d}s\) |
Optical path length |
Non-relativistic particle |
\(\int \left(\frac{1}{2}m\dot{\boldsymbol{x}}^2 - V\right)\mathrm{d}t\) |
Kinetic minus potential |
Relativistic particle |
\(-mc\int \mathrm{d}s\) |
Worldline length / proper time |
Electromagnetic field |
\(-\frac{1}{4}\int F_{\mu\nu}F^{\mu\nu}\,\mathrm{d}^4x\) |
Total gauge curvature |
Spacetime (gravity) |
\(\frac{c^3}{16\pi G}\int R \sqrt{-g}\,\mathrm{d}^4x\) |
Total spacetime curvature |
Relativistic string |
\(-T\int \sqrt{-\gamma}\,\mathrm{d}^2\sigma\) |
Worldsheet area |
Spin |
\(s\int (1-\cos\theta)\,\mathrm{d}\varphi\) |
Geometric (Berry) phase |
Derivation: Relativistic to Non-Relativistic Action
Start from the relativistic point-particle action written directly with proper time:
In a weak, static gravitational potential \(\Phi_g(\boldsymbol{x})\) with \(|\Phi_g|/c^2 \ll 1\), the leading-order proper-time relation is
So
where we kept only first order in \(\Phi_g/c^2\) and \(v^2/c^2\) (dropping products like \(\Phi_g v^2/c^4\)).
Therefore
Hence the low-speed, weak-gravity Lagrangian is
Up to the additive constant \(-mc^2\), this is exactly the usual non-relativistic form \(L=T-V\) with \(V=m\Phi_g\).
Quantization: From Paths to Probability Amplitudes#
Classical mechanics: the observable trajectory satisfies \(\delta S = 0\) (Hamilton’s principle).
Quantum mechanics: the probability amplitude for a particle to travel from \((\boldsymbol{x}_i, t_i)\) to \((\boldsymbol{x}_f, t_f)\) is the propagator \(K(\boldsymbol{x}_f,t_f;\boldsymbol{x}_i,t_i)\), obtained by summing contributions from all possible paths, each weighted by the phase \(\mathrm{e}^{\mathrm{i}S/\hbar}\):
Path Integral Principle
where the integral sums over every conceivable path, weighted by its quantum phase.
Classical limit (\(S \gg \hbar\)): The phase \(\mathrm{e}^{\mathrm{i}S/\hbar}\) oscillates rapidly. Neighboring paths have wildly different phases and cancel by destructive interference. Only the stationary-phase path—where \(\delta S = 0\)—survives. This is the classical trajectory.
Quantum regime (\(S \sim \hbar\)): All paths have comparable phases and interfere constructively. Diffraction, tunneling, and superposition emerge naturally.
This is Huygens’ principle for matter: every point is a source of “matter wavelets” that propagate and interfere. Quantization is the principle of superposing all these wavelets with the correct quantum phase.
Discussion: do non-classical paths matter?
The path integral says we sum \(\mathrm{e}^{\mathrm{i}S/\hbar}\) over all paths, including wildly non-classical ones. In what sense do these “crazy” paths contribute to observable physics? If a particle tunnels through a barrier, which paths dominate the sum?
Poll: Phase and classical action
The remarkable parallel between optics and quantum mechanics arises because the optical phase accumulated along a path (\(\propto L / \lambda\), where \(L\) is optical path length) plays the role of the quantum mechanical phase (\(\propto S / \hbar\), where \(S\) is the classical action). Which observation most directly supports this connection?
(A) Both light and matter have the same speed in vacuum.
(B) When light wavelength matches the de Broglie wavelength, Fermat’s principle and the path integral principle give the same predictions.
(C) Quantum particles are actually tiny waves, so they obey Huygens’ principle.
(D) The refractive index \(n\) in optics is analogous to the potential energy \(V\) in quantum mechanics.
Summary#
Optics-mechanics duality: Geometric optics and classical mechanics obey identical variational principles. Matter is the “geometric optics” limit of quantum mechanics.
De Broglie relations: \(\lambda = h/p\) and \(\omega = E/\hbar\) unify waves and particles. Matter-wave phase is \(\Phi = S/\hbar\), so \(\psi \sim \mathrm{e}^{\mathrm{i}S/\hbar}\).
Path integral quantization: Sum over all paths with phase weight \(\mathrm{e}^{\mathrm{i}S/\hbar}\). The classical path dominates when \(S \gg \hbar\) (stationary phase); quantum interference ruins other paths.
Actionequals phase: The exponent of the quantum amplitude is the classical action divided by \(\hbar\)—a deep unity connecting classical and quantum mechanics.
See Also
3.1.1 Geometric Optics: The particle theory of light and Fermat’s principle
3.1.2 Physical Optics: Huygens’ principle, interference, and why stationary phase gives geometric optics
3.2 Propagator: Making the path integral \(\int \mathcal{D}[x]\,\mathrm{e}^{\mathrm{i}S/\hbar}\) mathematically precise; deriving the Schrödinger equation from it
3.3 Stationary Phase: The semiclassical limit—WKB approximation and Bohr-Sommerfeld quantization
3.4 Imaginary Time: Wick rotation \(t \to -\mathrm{i}\tau\) connects the path integral to statistical mechanics
Homework#
1. Free particle action. A free particle has action \(S = \boldsymbol{p}\cdot\boldsymbol{r} - Et\) with \(E = p^2/(2m)\). Write down \(\psi = A\,\mathrm{e}^{\mathrm{i}S/\hbar}\) and identify the wavelength and frequency in terms of \(p\) and \(E\). Verify the de Broglie relations.
2. De Broglie wavelength. An electron is accelerated through a potential difference of \(V = 100\) V. Compute its de Broglie wavelength. Compare this to the lattice spacing of a typical crystal (\(\sim 0.3\) nm). Why does this make electron diffraction possible?
3. Optics-mechanics translation. In a medium with spatially varying refractive index \(n(\boldsymbol{x})\), a light ray bends according to Snell’s law: \(n(\boldsymbol{x})\sin\theta(\boldsymbol{x}) = \text{const}\), where \(\theta\) is the angle to the normal. Using the optics-mechanics analogy from the lecture, translate this into a statement about a classical particle in a potential \(V(\boldsymbol{x})\). What is the mechanical analogue of the refractive index?
4. Dimensions of action. The quantum wavefunction is \(\psi = A\,\mathrm{e}^{\mathrm{i}S/\hbar}\).
(a) What are the dimensions of \(S\)? What are the dimensions of \(S/\hbar\)? Why must the exponent be dimensionless?
(b) A classical baseball (\(m = 0.15\) kg) is thrown at \(v = 30\) m/s across a distance \(d = 10\) m. Estimate its action \(S \sim mvd\) and compute \(S/\hbar\). What does the enormous size of this ratio imply for whether quantum interference is observable?
(c) An electron (\(m = 9.1 \times 10^{-31}\) kg) moves at \(v = 10^6\) m/s across \(d = 1\) nm. Compute \(S/\hbar\). Why is quantum behavior now important?
5. Path cancellation. In the path integral, we sum \(\mathrm{e}^{\mathrm{i}S[\text{path}]/\hbar}\) over all paths.
(a) Consider two neighboring paths whose actions differ by \(\Delta S\). What is the phase difference between their contributions? Under what condition do they interfere constructively?
(b) When \(S \gg \hbar\), argue that a small geometric deformation of the path changes \(S\) by much more than \(\hbar\), so neighboring paths have wildly different phases and cancel. Why does the classical path (\(\delta S = 0\)) survive?
(c) When \(S \sim \hbar\), explain why many paths contribute and the particle no longer follows a single trajectory.
6. Relativistic action limit. For a relativistic particle, \(S = -mc^2\int\mathrm{d}\tau\). Show that in the non-relativistic limit (\(v \ll c\)), the action reduces to \(S \approx \int(\tfrac{1}{2}mv^2 - mc^2)\,\mathrm{d}t\), recovering the familiar Lagrangian up to a constant.