Wave-particle duality and the double-slit

Intuition [Beginner]

Classical physics sorts phenomena into two bins: waves (spread out, diffract, interfere) and particles (localized, travel in straight lines, collide). Quantum mechanics says the sorting is wrong. Everything is both.

The double-slit experiment proves it. Fire electrons one at a time at a barrier with two narrow slits. Each electron arrives at a single point on the detector — particle-like. After many electrons, the dots build up into an interference pattern — wave-like. Each electron interferes with itself.

Place a detector at the slits to find out which one the electron uses, and the interference vanishes. Observation changes the result.

Visual [Beginner]

Picture a source on the left firing electrons, one at a time, toward a wall with two parallel slits. Behind the wall sits a detector screen.

Stage 1. A few electrons arrive. Each makes a single dot. The dots look random.

Stage 2. Hundreds of electrons. The dots cluster into faint bands — bright stripes where many dots land, dark stripes where almost none do.

Stage 3. Thousands of electrons. The pattern is a classic two-slit interference pattern, identical to what light waves produce.

Bright bands are where "electron waves" from the two slits arrive in phase — constructive interference. Dark bands are where they arrive out of phase — destructive interference. Each electron contributes one dot; the distribution of dots is governed by a wave.

Worked example [Beginner]

The de Broglie relation assigns a wavelength to any moving object:

$$\lambda =\frac{h}{p}$$

where $h=6.626\times 10^{-34}$ J$\cdot$s is Planck's constant and $p$ is momentum. A baseball ($0.145$ kg at $40$ m/s) has $\lambda \approx 10^{-34}$ m — undetectable. An electron is different.

Electron mass $m=9.11\times 10^{-31}$ kg, speed $v=10^6$ m/s:

$$\lambda =\frac{6.626\times 10^{-34}}{9.11\times 10^{-31}\times 10^6}=\frac{6.626\times 10^{-34}}{9.11\times 10^{-25}}=7.27\times 10^{-10}\text{ m}\approx 0.73\text{ nm}$$

This is comparable to the spacing between atoms in a crystal — which is why Davisson and Germer (1927) observed electron diffraction from a nickel crystal. The electron's wavelength matches the atomic lattice, producing the same kind of diffraction pattern X-rays produce.

Check your understanding [Beginner]

Formal definition [Intermediate+]

The double-slit experiment is the entry point to three foundational ideas: the de Broglie relation, the superposition principle, and the Born rule. We develop each in turn, then state the Heisenberg uncertainty principle.

The de Broglie relation. In 1924, de Broglie postulated that a particle of momentum $p$ has an associated wavelength

$$\boxed{\lambda =\frac{h}{p}}$$

where $h$ is Planck's constant. In terms of the wave number $k=2\pi /\lambda$ and the reduced Planck constant $\hslash =h/(2\pi )$, this reads $p=\hslash k$. For a free non-relativistic particle of mass $m$ and velocity $v$, $p=mv$ and $\lambda =h/(mv)$.

The photoelectric effect (Einstein, 1905). Light incident on a metal surface ejects electrons. Classically, increasing the light intensity should increase the ejected electrons' kinetic energy. Instead, only the frequency of the light matters: below a threshold frequency ${\nu }_0$, no electrons are ejected regardless of intensity. Above ${\nu }_0$, the maximum kinetic energy of ejected electrons is

$$K_{\max }=h\nu -\varphi$$

where $\varphi$ is the work function of the metal. Einstein's interpretation: light comes in quanta (photons) of energy $E=h\nu$. Each photon transfers its entire energy to a single electron. This is the particle picture of light — the complement to the wave picture of diffraction and interference.

The Compton effect (1923). X-rays scattered off free electrons undergo a wavelength shift

$$\Delta \lambda =\frac{h}{m_{e}c}(1-\cos \theta )$$

where $\theta$ is the scattering angle and $m_e$ is the electron mass. This formula follows from treating the photon as a particle with energy $E=h\nu$ and momentum $p=h/\lambda$, then applying conservation of energy and momentum to the photon-electron collision. The Compton effect is direct confirmation that photons carry momentum — the particle side of wave-particle duality for light.

Superposition and probability amplitudes. In the double-slit experiment, label the two paths through the slits as path 1 and path 2. Quantum mechanics assigns to each path a probability amplitude — a complex number ${\psi }_1$ or ${\psi }_2$. The amplitude for the electron to arrive at a point $x$ on the screen is the sum

$$\psi (x)={\psi }_1(x)+{\psi }_2(x).$$

This is the principle of superposition: amplitudes add, not probabilities. The probability of detecting the electron at $x$ is given by the Born rule:

$$P(x)=|\psi (x){|}^2=|{\psi }_1(x)+{\psi }_2(x){|}^2.$$

The cross terms ${\psi }_1^{\ast }(x){\psi }_2(x)+{\psi }_2^{\ast }(x){\psi }_1(x)$ are what produce the interference pattern. If you detect which slit the electron uses, you force the system into a definite path — only ${\psi }_1$ or ${\psi }_2$ contributes — and the cross terms vanish. The probability becomes $P(x)=|{\psi }_1(x){|}^2+|{\psi }_2(x){|}^2$, which produces two lumps with no interference.

The Heisenberg uncertainty principle. A particle cannot simultaneously have a definite position and a definite momentum. The spreads (standard deviations) $\Delta x$ and $\Delta p$ in repeated measurements satisfy

$$\boxed{\Delta x\cdot \Delta p\ge \frac{\hslash }{2}.}$$

An analogous relation holds for energy and time: $\Delta E\cdot \Delta t\ge \hslash /2$. The uncertainty principle is not a statement about measurement precision — it is a property of the state itself. A particle in a state with well-defined position (small $\Delta x$) necessarily has a large spread in momentum (large $\Delta p$), and vice versa.

Counterexamples to common slips

• The uncertainty principle does not say "measuring position disturbs momentum." That is a consequence, not the principle. The principle says the state does not possess simultaneously sharp values of $x$ and $p$. The disturbance story is a heuristic, not the content.
• The de Broglie wavelength $\lambda =h/p$ applies to the free-particle momentum eigenstate, which has perfectly defined momentum ($\Delta p=0$) and is completely delocalised ($\Delta x=\infty$). A realistic electron in a double-slit experiment is a wave packet — a superposition of many momentum eigenstates — localised enough to pass through the slits but spread enough in momentum to produce interference.
• The photon has zero rest mass but carries momentum $p=h/\lambda$. The formula $p=mv$ does not apply to photons; the correct relativistic relation is $E^2=(pc{)}^2+(m{c}^2{)}^2$, which for $m=0$ gives $E=pc=h\nu$.
• "Wave-particle duality" is a historical label, not a physical principle. The correct statement is that quantum objects are described by a state in a Hilbert space, and different measurements reveal wave-like or particle-like aspects. The duality language is a scaffold, not the building.

Key theorem with proof [Intermediate+]

Theorem (Heisenberg uncertainty from the de Broglie relation and Fourier analysis). Let $\psi (x)$ be a normalised one-dimensional wave function describing a particle. Define the spreads

$$\Delta x:=\sqrt{⟨x^2⟩-⟨x{⟩}^2},\Delta p:=\sqrt{⟨p^2⟩-⟨p{⟩}^2},$$

where $⟨\cdot ⟩$ denotes the expectation value in the state $\psi$. Then

$$\Delta x\cdot \Delta p\ge \frac{\hslash }{2}.$$

Proof (via the Fourier transform). Write the wave function in position space as $\psi (x)$ and in momentum space via the Fourier transform

$$\tilde{\psi }(k)=\frac{1}{\sqrt{2\pi }}{\int }_{-\infty }^{\infty }\psi (x)e^{-ikx}dx.$$

Parseval's theorem gives $\int |\psi (x){|}^{2}dx=\int |\tilde{\psi }(k){|}^{2}dk=1$. The de Broglie relation $p=\hslash k$ identifies the momentum-space wave function as $\varphi (p)=\tilde{\psi }(p/\hslash )/\sqrt{\hslash }$, so that $⟨p^n⟩=\int p^n|\varphi (p){|}^{2}dp$.

The Cauchy-Schwarz inequality applied to the functions $x\psi (x)$ and $d\psi /dx$ yields

$$\left(\int |x\psi {|}^{2}dx\right)\left(\int {\left|\frac{d\psi }{dx}\right|}^{2}dx\right)\ge {\left|\int x{\psi }^{\ast }\frac{d\psi }{dx}dx\right|}^2.$$

The right-hand side is bounded below by $(1/2{)}^2$ via integration by parts with the boundary term vanishing (normalisability):

$$\mathrm{Re}\int x{\psi }^{\ast }\frac{d\psi }{dx}dx=\frac{1}{2}\int x\frac{d}{dx}|\psi {|}^{2}dx=-\frac{1}{2}\int |\psi {|}^{2}dx=-\frac{1}{2},$$

so $\left|\int x{\psi }^{\ast }(d\psi /dx)dx\right|\ge 1/2$.

In the momentum representation, $⟨p^2⟩={\hslash }^2\int k^2|\tilde{\psi }(k){|}^{2}dk={\hslash }^2\int |d\psi /dx{|}^{2}dx$ (the last equality is Parseval applied to the Fourier derivative theorem). Combining:

$$⟨x^2⟩\cdot \frac{⟨p^2⟩}{{\hslash }^2}\ge \frac{1}{4}.$$

Shifting the origin so that $⟨x⟩=0$ and $⟨p⟩=0$ (which does not change the variances) gives $\Delta x^2\cdot \Delta p^2\ge {\hslash }^2/4$, hence $\Delta x\cdot \Delta p\ge \hslash /2$. ∎

Remark. The equality $\Delta x\cdot \Delta p=\hslash /2$ holds if and only if $\psi (x)$ is a Gaussian — the minimum-uncertainty wave packet. Squeezing one side of the product below $\hslash /2$ is impossible in any state.

Bridge. The uncertainty principle builds toward 12.01.02 pending where the non-commuting spin operators $S_x,S_z$ exhibit the same variance trade-off in the finite-dimensional setting of spin-1/2, and appears again in 12.02.01 pending where the canonical commutation relation $[\hat{x},\hat{p}]=i\hslash$ becomes an axiom of the Hilbert-space formalism. The foundational reason is that Fourier duality between position and momentum space is the mathematical statement of the physical fact that a localised particle requires a broad superposition of momentum eigenstates; this is exactly the content of the de Broglie relation $p=\hslash k$. The bridge is between the wave picture (Fourier transform, diffraction) and the operator picture (non-commuting Hermitian operators, spectral theory) — the two languages are equivalent, and the uncertainty principle is their shared cornerstone.

Exercises [Intermediate+]

Advanced results [Master]

Theorem 1 (Englert-Greenberger-Yasin duality relation). For any two-path interferometer — including the double-slit — define the fringe visibility $V=(I_{\max }-I_{\min })/(I_{\max }+I_{\min })$ and the which-path distinguishability $D=|p_1-p_2|/(p_1+p_2)$ where $p_j$ is the probability that path $j$ is correctly identified. Then

$$D^2+V^2\le 1.$$

Equality holds if and only if the two path amplitudes have equal magnitude. The relation quantifies the trade-off: perfect which-path information ($D=1$) forces $V=0$ (no interference), and perfect interference ($V=1$) forces $D=0$ (no which-path information).

The duality relation ^{[Englert 1996]} is the quantitative statement behind the qualitative "wave-particle duality" slogan. It subsumes the double-slit observation that placing a detector at the slits destroys the interference pattern: the detector increases $D$, which forces $V$ down. The proof appears in the Full proof set below. A tighter inequality $D^2+V^2\le 1$, including the effect of the quantum state's purity on the bound, was given by Bimonte and Musto ^{[Bimonte-Musto 2003]} and generalises to multipath interferometers. The relation also connects to complementarity in the sense of Bohr: the wave aspect (high $V$) and the particle aspect (high $D$) are mutually exclusive, but the exclusion is quantitative, not absolute — partial which-path information allows partial interference.

Theorem 2 (Mott, 1929). A spherical outgoing wavefunction describing alpha-particle emission from a radioactive nucleus produces straight-line tracks in a cloud chamber. The apparent particle-like trajectory is a consequence of the spherical wave interacting with a series of atoms, each acting as a position measurement. The joint probability amplitude for ionising atoms at positions ${\mathbf{r}}_1,{\mathbf{r}}_2,\dots ,{\mathbf{r}}_n$ concentrates on configurations where the points are colinear with the source — not because the alpha particle "is" a classical particle, but because the spherical wave's correlations enforce colinearity in the many-ionisation limit.

Mott's analysis ^{[Mott 1929]} is the founding example of environment-induced superselection: a spatially extended wavefunction, coupled to an environment with many degrees of freedom, produces outcomes that look classical — straight tracks, localised spots, single trajectories — without any collapse postulate. The same mechanism underlies the electron dots in the double-slit experiment: each electron's wavefunction is spread across the slits, but the detector screen ionises at a single point, and the pattern of many such points reveals the wave. Mott's key insight was that the question "why do alpha-particle tracks look straight?" is ill-posed if asked about a single ionisation event — the track is a correlation between many events, and the wavefunction's structure enforces the correlation. The single-particle detection events at the double-slit screen are the same phenomenon in a simpler geometry.

Theorem 3 (Aharonov-Bohm, 1959). A charged particle travelling around a region of nonzero magnetic vector potential $\mathbf{A}$ — even when the magnetic field $\mathbf{B}=\nabla \times \mathbf{A}$ is identically zero along the particle's path — acquires a phase shift $\Delta \varphi =(q/\hslash )\oint \mathbf{A}\cdot d\mathbf{l}=(q/\hslash ){\Phi }_B$ where ${\Phi }_B$ is the magnetic flux enclosed. This phase shift is observable as a displacement of interference fringes in a double-slit configuration where the two paths enclose the flux.

The Aharonov-Bohm effect ^{[Aharonov-Bohm 1959]} demonstrates that the electromagnetic potential, not the field, is the fundamental object in quantum mechanics. The particle never enters the region of nonzero $\mathbf{B}$; the phase shift is entirely due to the gauge potential. This is incompatible with any local-action classical picture and confirms that the probability amplitude (including its phase) is physically real, not merely a calculational device. The effect also illustrates that the superposition principle is sensitive to the topology of the configuration space, not just local geometry.

Theorem 4 (Feynman path integral, 1948). The probability amplitude for a particle to travel from $(x_a,t_a)$ to $(x_b,t_b)$ is a sum over all possible paths $\gamma$ connecting the two events, weighted by the action evaluated along each path:

$$⟨x_b,t_b|x_a,t_a⟩=\int \mathcal{D}\gamma \exp \left(\frac{i}{\hslash }S[\gamma ]\right).$$

In the classical limit $\hslash \to 0$, the stationary-phase approximation localises the integral to paths near the classical trajectory — the principle of least action emerges from the wave description. For the double-slit, the sum over paths reduces to two dominant contributions (through slit 1 and slit 2), recovering the superposition $\psi ={\psi }_1+{\psi }_2$.

The path-integral formulation ^{[Feynman 1948]} identifies the double-slit experiment as the universal template for quantum mechanics: every process is an interference of amplitudes over all histories. This builds toward 12.02.01 pending where the Hilbert-space formalism encodes the same superposition in abstract vector language, and appears again in 12.10.01 pending (forthcoming) where the path integral becomes the generating functional of quantum field theory.

Theorem 5 (Kochen-Specker, 1967). In a Hilbert space of dimension $\ge 3$, there is no function $v:\mathcal{O}\to \mathbb{R}$ from the set of quantum observables to the reals satisfying: (i) $v(A)\in \mathrm{Spec}(A)$ for every observable $A$; (ii) $v(f(A))=f(v(A))$ for every Borel function $f$. Non-contextual hidden-variable theories are impossible in dimension 3 and above.

The original proof ^{[Kochen-Specker 1967]} constructed 117 directions on the sphere in ${\mathbb{R}}^3$ that overdetermine any 0-1 assignment respecting the orthogonality constraints of spin-1 observables. Simpler proofs exist: Peres (1991) gives a 33-direction proof in dimension 3; Cabello (1996) gives an 18-vector proof in dimension 4.

The consequence is contextuality: the measured value of an observable can depend on which other observables are measured simultaneously. The double-slit experiment shows contextuality in its simplest form — the position distribution at the screen depends on whether a which-path measurement is performed — and Kochen-Specker proves this dependence is irreducible.

The connection to Bell's theorem: Kochen-Specker rules out non-contextual hidden variables within a single system. Bell's theorem ^{[Bell 1964]} rules out local hidden variables across spatially separated systems. Both results constrain the space of possible classical-underneath explanations, and both are independent — neither implies the other.

Theorem 6 (Quantum eraser — Kim et al., 1999). In a double-slit experiment with entangled photon pairs (signal photon through the slits, idler photon carrying which-path information), erasing the which-path information on the idler — by measuring it in a basis complementary to the which-path basis — restores interference in the conditional distribution of the signal photon. The unconditional (unconditioned) signal distribution shows no interference: it is the sum of two complementary conditional patterns.

The quantum eraser ^{[Kim et al. 1999]} makes the role of information explicit: the relevant variable is whether the two paths are, in principle, distinguishable from the state of the rest of the world, not whether a force is exerted on the particle. The delayed-choice variant, in which the erasure decision is made after the signal photon has already been detected, demonstrates that the erasure need not be temporally prior to the detection. The "recovered" interference appears only in a conditional subset of the data — you must post-select the signal-detector results based on the idler measurement outcome.

Wheeler's delayed-choice experiment (proposed 1978, realised by Jacques et al. 2007) pushes this further: the decision of whether to insert a second beamsplitter (recovering interference) or remove it (revealing which-path information) is made after the photon has entered the interferometer. The result respects the delayed choice: with the beamsplitter in place, interference fringes appear; without it, which-path information is recovered. The photon's behaviour is not predetermined by any property it carries into the apparatus — it is co-determined by the entire experimental arrangement, including the final choice of measurement. Wheeler's summary: "no phenomenon is a phenomenon until it is an observed phenomenon." This is contextuality in its temporally most dramatic form.

Theorem 7 (Born rule from Gleason's theorem). Let $\mathcal{H}$ be a Hilbert space of dimension $\ge 3$. Any measure $\mu$ on the lattice of projection operators that satisfies $\mu ({\sum }_i{P}_i)={\sum }_i\mu (P_i)$ for any set of mutually orthogonal projectors $\{P_i\}$ must take the form $\mu (P)=\mathrm{tr}(\rho P)$ for some density operator $\rho$. The Born rule $P(x)=|\psi (x){|}^2=\mathrm{tr}(|x⟩⟨x|\cdot |\psi ⟩⟨\psi |)$ is the special case $\rho =|\psi ⟩⟨\psi |$.

Gleason's theorem (1957) derives the Born rule from the structure of Hilbert space alone, without invoking probability as a separate postulate. The dimension-$\ge 3$ condition is essential: in dimension 2, non-Born measures exist (which is why Kochen-Specker also requires dimension $\ge 3$). Gleason's result is the strongest known argument that the probability interpretation of quantum mechanics is forced by the Hilbert-space structure rather than being an additional physical axiom. The implication for the double-slit: the rule $P(x)=|\psi (x){|}^2$ is not an independent hypothesis about nature but a mathematical consequence of representing states as vectors in a Hilbert space of dimension $\ge 3$ and demanding consistent probability assignments across all measurements.

Synthesis. The foundational reason that wave-particle duality is not a paradox but a structural feature of quantum mechanics is that the probability amplitude $\psi$ is a complex-valued object whose squared modulus gives real probabilities and whose phase produces interference. Putting these together with the Englert-Greenberger duality relation, the Aharonov-Bohm effect, and Mott's analysis, the central insight emerges: what is "wave-like" (interference, phase sensitivity) and what is "particle-like" (discrete detection events, which-path information) are two aspects of the same probability-amplitude structure, and the quantitative trade-off between them is exact. The bridge is the path integral, which identifies the superposition principle — the sum $\psi ={\psi }_1+{\psi }_2$ underlying the double-slit — as the universal rule for quantum amplitudes. This is exactly the structure that 12.02.01 pending re-formulates in the language of Hilbert spaces and operators, and the pattern generalises from the double-slit to every process in quantum field theory via the functional integral.

Full proof set [Master]

Proposition 1 (Englert-Greenberger duality). In a two-path interferometer with path amplitudes ${\psi }_j=\sqrt{p_j}{e}^{i{\varphi }_j}$ ($j=1,2$; $p_1+p_2=1$), the fringe visibility $V$ and which-path distinguishability $D$ satisfy $D^2+V^2\le 1$.

Proof. The probability at the screen, parameterised by position variable $\delta$ (proportional to the phase difference $kd\sin \theta$ at the detector), is

$$P(\delta )=|\sqrt{p_1}{e}^{i{\varphi }_1}+\sqrt{p_2}{e}^{i{\varphi }_2}{|}^2=p_1+p_2+2\sqrt{p_1{p}_2}\cos (\Delta \varphi +\delta )=1+2\sqrt{p_1{p}_2}\cos (\Delta \varphi +\delta )$$

where $\Delta \varphi ={\varphi }_1-{\varphi }_2$. The maximum and minimum are $I_{\max }=1+2\sqrt{p_1{p}_2}$ and $I_{\min }=1-2\sqrt{p_1{p}_2}$, giving visibility $V=2\sqrt{p_1{p}_2}$.

The optimal which-path measurement assigns the particle to path 1 with probability $p_1$ and to path 2 with probability $p_2$. The distinguishability is $D=|p_1-p_2|$.

Compute: $D^2+V^2=(p_1-p_2{)}^2+4p_1{p}_2=(p_1+p_2{)}^2=1$. Equality holds for all values of $p_1,p_2$ satisfying $p_1+p_2=1$ — that is, for all pure two-path states. A partial which-path measurement (e.g., a detector with finite efficiency) introduces mixedness and makes the inequality strict: $D^2+V^2<1$. $\square$

Proposition 2 (Mott colinearity). The joint probability for an outgoing spherical wave of momentum $\hslash k$ to ionise atoms at positions ${\mathbf{r}}_1$ and ${\mathbf{r}}_2$ in a cloud chamber concentrates on colinear configurations ${\mathbf{r}}_1\parallel {\mathbf{r}}_2$.

Proof sketch. The ionisation amplitude for the $j$-th atom at ${\mathbf{r}}_j$ is proportional to $e^{ik|{\mathbf{r}}_j|}/|{\mathbf{r}}_j|$. The joint amplitude for ionising atoms at ${\mathbf{r}}_1$ and ${\mathbf{r}}_2$ involves the product of successive spherical waves: the outgoing spherical wave from the source hits ${\mathbf{r}}_1$, then the scattered spherical wave from ${\mathbf{r}}_1$ hits ${\mathbf{r}}_2$. The probability amplitude involves the angular factor $\sin (k|{\mathbf{r}}_1-{\mathbf{r}}_2|)/|{\mathbf{r}}_1-{\mathbf{r}}_2|$ which, for $k$ corresponding to MeV-scale alpha particles and atomic-scale scattering centres, oscillates rapidly. The integral over the spherical wavefront averages to zero unless ${\mathbf{r}}_1$, ${\mathbf{r}}_2$, and the source are nearly colinear — the stationary-phase condition. The many-ionisation limit amplifies this: three or more non-colinear ionisation points contribute vanishing amplitude. $\square$

Proposition 3 (Aharonov-Bohm phase). In a region where $\mathbf{B}=0$ but $\mathbf{A}\ne 0$, the probability amplitude for a charged particle to travel along a path $\gamma$ acquires the additional phase $\varphi =(q/\hslash ){\int }_{\gamma }\mathbf{A}\cdot d\mathbf{l}$. Two paths enclosing flux ${\Phi }_B$ accumulate a relative phase $\Delta \varphi =(q/\hslash ){\Phi }_B$.

Proof. The Schrödinger equation for a charged particle in an electromagnetic potential is obtained by the minimal coupling substitution $\mathbf{p}\to \mathbf{p}-q\mathbf{A}$, $E\to E-q\Phi$ in the free-particle Hamiltonian. For a free particle in a region where $\mathbf{B}=\nabla \times \mathbf{A}=0$, there exists a local gauge in which $\mathbf{A}=\nabla \chi$ for some scalar function $\chi$. The solution is $\psi ={\psi }_0\exp (iq\chi /\hslash )$ where ${\psi }_0$ is the free-particle solution. Along a path from ${\mathbf{r}}_a$ to ${\mathbf{r}}_b$, the accumulated phase is $(q/\hslash ){\int }_{{\mathbf{r}}_a}^{{\mathbf{r}}_b}\mathbf{A}\cdot d\mathbf{l}=(q/\hslash )(\chi ({\mathbf{r}}_b)-\chi ({\mathbf{r}}_a))$.

For two paths ${\gamma }_1,{\gamma }_2$ forming a closed loop $C={\gamma }_1-{\gamma }_2$, the relative phase is $(q/\hslash ){\oint }_C\mathbf{A}\cdot d\mathbf{l}=(q/\hslash ){\int }_S\mathbf{B}\cdot d\mathbf{S}=(q/\hslash ){\Phi }_B$ by Stokes' theorem, where $S$ is any surface bounded by $C$. This phase is gauge-invariant and observable: the intensity at the detector is $I\propto 1+\cos (\Delta \varphi )=1+\cos (q{\Phi }_B/\hslash )$, which shifts as ${\Phi }_B$ varies. $\square$

Interpretation previews [Master]

The double-slit experiment is the common reference point for all major interpretations of quantum mechanics. Each explains why the interference pattern appears and what happens during measurement, using a different ontological picture.

Copenhagen interpretation (Bohr, Heisenberg). The wave function is not a physical object — it is a tool for calculating probabilities. Asking "which slit did the electron go through?" is a category error when no detector is present: the question has no answer because the observable was not measured. Measurement brings a property into being. This is the operational default of most working physicists and requires no additional mathematical structure, but leaves the measurement process itself unexplained (see 20.03.01 pending).

de Broglie-Bohm pilot-wave theory (de Broglie 1927, Bohm 1952). The electron is a particle with a definite trajectory, guided by a real physical wave (the pilot wave) that passes through both slits. The wave produces interference; the particle rides the interference landscape. The trajectory is deterministic — given the initial position and the wave function, the future path is fixed. Measurement is a special case of the guiding equation. In the double-slit, the pilot wave $\psi (x)$ passes through both slits, interferes with itself, and creates a landscape of high and low $|\psi {|}^2$; the electron particle, guided by the local current $\mathbf{j}=|\psi {|}^2\nabla (\arg \psi )$, is channelled into the bright fringes. After many electrons, the particle distribution matches $|\psi {|}^2$ — recovering the Born rule as a statistical consequence, not a postulate. The cost: the theory is non-local (the pilot wave depends instantaneously on the configuration of all particles in the universe) and the trajectories are inaccessible without disrupting the wave.

Many-worlds interpretation (Everett 1957). The wave function always evolves unitarily — no collapse. When a measurement occurs, the total system (particle + detector + observer) enters a superposition of branches: one branch where the particle went through slit 1 and the detector registered slit 1, another where it went through slit 2 and the detector registered slit 2. Each branch is equally real. The "loss" of interference when a detector is present is the decoherence of these branches — they become mutually inaccessible due to entanglement with the macroscopic environment. The cost: the interpretation must explain why we experience a single outcome (the "preferred basis problem") and why the Born rule probabilities emerge from deterministic unitary evolution.

These interpretations make identical empirical predictions for the double-slit experiment. They differ on the ontology — what exists — and on the dynamics of measurement. No experiment known to date distinguishes between them. Bell's theorem rules out a large class of local hidden-variable theories but does not settle the interpretation question.

Connection to Hilbert-space formalism [Master]

The double-slit experiment, treated informally in this unit, is the physical motivation for the abstract Hilbert-space formalism developed in 12.02.01 pending. The probability amplitudes ${\psi }_1,{\psi }_2$ that interfere on the screen are the components of a state vector in a Hilbert space. The Born rule $P=|\psi {|}^2$ is the norm squared of a projection. The superposition $\psi ={\psi }_1+{\psi }_2$ is vector addition. The vanishing of interference under measurement is the projection postulate.

The de Broglie relation $p=\hslash k$ and the Planck-Einstein relation $E=\hslash \omega$ together motivate the identification of momentum and energy as operators acting on a wave function: $\hat{p}=-i\hslash d/dx$ and $\hat{E}=i\hslash d/dt$. The Heisenberg uncertainty principle $\Delta x\cdot \Delta p\ge \hslash /2$ is then a consequence of the non-commutativity $[\hat{x},\hat{p}]=i\hslash$ via the Robertson-Schrödinger relation (Exercise 9). The canonical commutation relation is the load-bearing piece: once $[\hat{x},\hat{p}]=i\hslash$ is on the table, the uncertainty principle, the energy quantisation of bound states, and the entire operator-algebraic framework follow.

The bridge to 12.01.02 pending is the realisation that not all quantum observables arise from position and momentum. Spin has no classical position-space representation — it requires a finite-dimensional Hilbert space (${\mathbb{C}}^2$) and operators (the Pauli matrices) that have no $\hat{x},\hat{p}$ decomposition. Stern-Gerlach is to spin what the double-slit is to spatial degrees of freedom: the smallest experiment that forces the Hilbert-space picture on you.

Connections [Master]

• Stern-Gerlach and spin-1/2 12.01.02 pending. The double-slit is to spatial wavefunctions what Stern-Gerlach is to spin: the smallest experiment that reveals quantum behaviour. Both demonstrate superposition, measurement projection, and the non-commutativity of incompatible observables — position versus momentum at the slits, $S_z$ versus $S_x$ in sequential Stern-Gerlach. The formalism of 12.01.02 pending (Hilbert space ${\mathbb{C}}^2$, Born rule, projection postulate) is the finite-dimensional instance of the general framework this unit motivates for continuous observables.

• Hilbert-space formalism 12.02.01 pending. Provides the abstract mathematical framework that the double-slit experiment motivates physically: state vectors, inner products, Hermitian operators, spectral decomposition, and the projection postulate. The superposition principle observed in the interference pattern becomes vector addition in Hilbert space; the Born rule becomes the norm-squared of the projection; the collapse under measurement becomes the projection postulate. The generalisation from the two-path double-slit to arbitrary superpositions in an infinite-dimensional Hilbert space is the content of 12.02.01 pending.

• Measurement problem in QM 20.03.01 pending. The double-slit experiment is the canonical instance of the measurement problem: the wavefunction passes through both slits (coherent superposition), but the detector records a single point (apparent collapse). Every interpretation of QM must account for this transition; none resolves it to universal satisfaction. The philosophy unit on the measurement problem takes the double-slit as its primary case study.

• Hydrogen atom quantum chemistry 14.04.01 pending. The de Broglie wavelength introduced here is the physical input for the chem-side hydrogen atom: the electron in a hydrogen atom behaves as a standing wave with wavelength $\lambda =h/p$, and the Bohr radius emerges from requiring an integer number of wavelengths to fit around the orbit. The chem unit references the de Broglie relation when explaining the discrete energy levels of atomic electrons.

Historical & philosophical context [Master]

Planck introduced energy quantisation in 1900 to resolve the ultraviolet catastrophe in blackbody radiation ^{[Planck 1900]}; the birth of quantum theory. Einstein explained the photoelectric effect in 1905, proposing light quanta carrying energy $E=h\nu$ ^{[Einstein 1905]}, for which he received the Nobel Prize in 1921. Bohr proposed quantised electron orbits in 1913 to explain the hydrogen spectrum ^{[Bohr 1913]}. Compton scattering, observed in 1923, confirmed that photons carry momentum $p=h/\lambda$ ^{[Compton 1923]}.

De Broglie proposed matter waves in 1924 ^{[de Broglie 1924]}: wavelength $\lambda =h/p$ for material particles, receiving the Nobel Prize in 1929. Born proposed the probability interpretation $P=|\psi {|}^2$ in 1926 ^{[Born 1926]}. Davisson and Germer observed electron diffraction from a nickel crystal in 1927 ^{[Davisson-Germer 1927]}, confirming de Broglie wavelengths experimentally. Bohr introduced complementarity the same year: wave and particle aspects are mutually exclusive descriptions of the same phenomenon.

The Bohr-Einstein debates (1927-1930) sharpened the measurement problem. Einstein attempted to show QM is incomplete; Bohr defended complementarity. The EPR paradox (1935) argued QM is incomplete if locality holds ^{[EPR 1935]}, setting the stage for Bell's theorem (1964) ^{[Bell 1964]}, which proved no local hidden-variable theory can reproduce all QM predictions. The Kochen-Specker theorem (1967) ^{[Kochen-Specker 1967]} ruled out non-contextual hidden variables in dimension $\ge 3$. Gleason's theorem (1957, published 1975) derived the Born rule from Hilbert-space structure alone.

Experimental confirmations followed: Freedman and Clauser (1972), Aspect (1982) with fast-switching detectors providing definitive Bell-inequality violations, and Kim et al. (1999) ^{[Kim et al. 1999]} demonstrating the delayed-choice quantum eraser. Aharonov and Bohm (1959) ^{[Aharonov-Bohm 1959]} predicted and Tonomura et al. (1986) observed the electromagnetic phase shift in electron interferometry — confirming that potentials, not fields, are the fundamental quantum-mechanical objects. Tonomura's 1989 double-slit experiment with electrons, in which the build-up of the interference pattern from single-electron detections was recorded in real time, remains the most visually direct demonstration of wave-particle duality. Each electron arrives at a single point; the wave-like pattern emerges only in the statistical accumulation. The experiment closed the last reasonable objection (that the interference might be an artefact of electron-electron interaction) by using electron currents so low that only one electron was in the apparatus at any time.

Bibliography [Master]

• Dirac, P. A. M. The Principles of Quantum Mechanics, 4th ed. Oxford University Press, 1958. Ch. I: The Principle of Superposition.

• Feynman, R. P., Leighton, R. B., and Sands, M. The Feynman Lectures on Physics, Vol. III. Addison-Wesley, 1965. Ch. 1: Quantum Behavior.

• Sakurai, J. J. and Napolitano, J. Modern Quantum Mechanics, 2nd ed. Cambridge University Press, 2011. §1.1.

• Susskind, L. and Friedman, A. Quantum Mechanics: The Theoretical Minimum. Basic Books, 2014. Lecture 1.

• Einstein, A. "Uber einen die Erzeugung und Verwandlung des Lichtes betreffenden heuristischen Gesichtspunkt." Ann. Phys. 322(6), 132–148, 1905.

• de Broglie, L. "Recherches sur la theorie des quanta." PhD thesis, Sorbonne, 1924.

• Davisson, C. and Germer, L. H. "Diffraction of Electrons by a Crystal of Nickel." Phys. Rev. 30(6), 705–740, 1927.

• Born, M. "Zur Quantenmechanik der Stossvorgange." Z. Phys. 37, 863–867, 1926.

• Mott, N. F. "The Wave Mechanics of alpha-Ray Tracks." Proc. Roy. Soc. A 126, 79–84, 1929.

• Aharonov, Y. and Bohm, D. "Significance of Electromagnetic Potentials in the Quantum Theory." Phys. Rev. 115, 485–491, 1959.

• Feynman, R. P. "Space-Time Approach to Non-Relativistic Quantum Mechanics." Rev. Mod. Phys. 20, 367–387, 1948.

• Kochen, S. and Specker, E. P. "The Problem of Hidden Variables in Quantum Mechanics." J. Math. Mech. 17(1), 59–87, 1967.

• Bell, J. S. "On the Einstein Podolsky Rosen Paradox." Physics Physique Fizika 1, 195–200, 1964.

• Englert, B.-G. "Fringe Visibility and Which-Way Information: An Inequality." Phys. Rev. Lett. 77, 2154–2157, 1996.

• Kim, Y.-H., Yu, R., Kulik, S. P., Shih, Y., and Scully, M. O. "A Delayed Choice Quantum Eraser." Phys. Rev. Lett. 84(1), 1–5, 1999.

• Gleason, A. M. "Measures on the Closed Subspaces of a Hilbert Space." J. Math. Mech. 6, 885–893, 1957.