One tiny experiment, one enormous mystery: electrons act like waves until you look — wave–particle duality and de Broglie's matter waves, explained from zero.
Richard Feynman called it the experiment that contains “the only mystery” of quantum mechanics.
Beginner Module 03 · Duality ~35 minPrerequisites: Module 1 (energy comes in packets, E = hf) and Module 2 (light is both wave and particle). If you're comfortable with those two ideas and can do algebra with scientific notation, you have everything you need.
“We choose to examine a phenomenon which is impossible, absolutely impossible, to explain in any classical way, and which has in it the heart of quantum mechanics. In reality, it contains the only mystery.” — on the double-slit experiment, The Feynman Lectures on Physics
That is quite a claim from one of the greatest physicists of the twentieth century: that a single tabletop experiment — a source, a wall with two narrow openings, and a screen — holds the entire strangeness of quantum mechanics. By the end of this module you'll see exactly what he meant.
Before we get to the mystery, we need one everyday idea: waves can add up, and waves can cancel out. Picture a still pond. You drop a stone in — circular ripples spread outward. Now drop two stones a short distance apart. Each stone makes its own set of ripples, and the two sets spread out and pass right through each other. Where they overlap, something interesting happens.
A wave is a pattern of crests (high points) and troughs (low points). Where the ripples from the two stones meet, the water simply adds the two waves together:
Whether two waves reinforce or cancel at some point depends only on the difference in distance they traveled from their two sources to that point. Along some directions the path difference is a whole number of wavelengths, so crests keep arriving together: permanent reinforcement. Along other directions the paths differ by half a wavelength, so a crest always arrives with a trough: permanent cancellation. The result is a fixed, striped pattern of “lively” and “calm” lanes fanning out across the pond. You can genuinely see this in a bathtub or a ripple tank.
In 1801, the English scientist Thomas Young did exactly the two-stones trick, but with light. He let light pass through two very narrow, closely spaced slits in an opaque barrier and looked at a screen behind them. If light were a stream of tiny classical particles, you'd expect two bright patches, one behind each slit. That's not what he saw.
He saw alternating bright and dark stripes — called interference fringes — spread across the screen. Bright stripes where light from the two slits arrived crest-with-crest (constructive), dark stripes where it arrived crest-with-trough (destructive). There were even bright stripes at positions between the slits' shadows, and dark stripes at spots that were brightly lit when only one slit was open. Opening a second slit made some places on the screen darker. Only waves can do that.
An interference pattern — alternating stripes of “lots” and “none” — is the unmistakable signature of a wave. If two openings produce stripes, whatever came through them behaved as a wave that passed through both openings at once and recombined on the other side. Keep this fingerprint in mind; we're about to go looking for it in a very unexpected place.
Young's stripes settled a century-old debate: light is a wave. Then, as you saw in Module 2, the photoelectric effect un-settled it — light also arrives in particle-like packets called photons. Hold both facts in your head. Now let's run the two-slit experiment properly, three times, with three very different kinds of ammunition.
The setup is always the same: a source that fires things, a barrier with two narrow slits, and a detecting screen behind it that records where things land. The only thing we change is what we fire. Feynman's classic presentation runs it three ways — and the third run is where physics falls off a cliff.
The double-slit setup: the source's wave passes through both slits, the two emerging waves overlap, and alternating bright/dark interference fringes appear on the screen.
Imagine a slightly wobbly machine gun firing bullets at an armored wall with two slits, with a sandbank behind it catching whatever gets through. Each bullet is an indivisible lump: it goes through either the left slit or the right slit, never both, and lands at one spot. After thousands of rounds, the sand shows two piles, one roughly behind each slit, blurring together in the middle. No stripes. The count with both slits open is just the sum of the counts with each slit open alone. Utterly unsurprising — this is what “particle” means.
Now put the same barrier in a ripple tank and send water waves at it. Each slit becomes a new source of ripples (the two stones from the previous section, in effect), the two ripple sets overlap, and the detector behind the barrier measures interference stripes in the wave intensity: strong–weak–strong–weak across the whole width. The energy arrives continuously — smoothly, in any amount, not in lumps. This is what “wave” means.
Now the main event. Fire electrons — certified particles, with a definite mass and charge, the things that make up electric current — at the two slits. And to remove any possibility of electrons bumping into each other and faking a pattern, turn the source down so far that only one electron is in flight at a time. Each electron leaves the source, crosses the apparatus alone, and hits the screen before the next one is even fired.
Watch the screen. Each electron arrives as a single, tiny, localized dot. One electron, one dot, at one spot. That's particle behavior — so far, it looks like the bullets. But keep watching as the dots accumulate: a hundred, a thousand, ten thousand… The dots are not building two piles. They are building stripes. Dense bands of dots separated by lanes where almost no electron ever lands — the exact interference pattern the water waves made. Individually the electrons arrive like particles; collectively they paint the fingerprint of a wave.
Each electron traveled alone — there was nothing else in the apparatus for it to interfere with. Yet the places it can land are dictated by an interference pattern that depends on both slits being open. Close one slit and the stripes vanish. Somehow each single electron “knows” about both slits — as if something wave-like associated with it passes through both and interferes with itself, steering where the electron may land. And remember the killer detail from Young's experiment: there are spots on the screen an electron can reach with one slit open that become forbidden when you open a second one. Giving the electron an extra way through makes some destinations impossible. No picture of little bullets can explain that.
| What we fire | How it arrives | Pattern with both slits open | Verdict |
|---|---|---|---|
| Bullets | One lump at one spot | Two piles, no stripes | Particle |
| Water waves | Continuous, spread out | Interference stripes | Wave |
| Electrons | One dot at one spot (particle-like!) | Stripes emerge from the dots (wave-like!) | Both — duality |
For decades this was taught as an idealized “gedanken” experiment, but it has been done for real, beautifully. In 1989, Akira Tonomura's team at Hitachi fired electrons one at a time through an electron biprism (a two-path setup equivalent to two slits) and filmed the screen: first a few random-looking dots, then, as tens of thousands accumulated, crisp interference fringes. The footage is one of the most famous movies in physics. The same single-particle interference has been demonstrated with photons arriving one by one, with whole atoms, and even with large molecules — C60 “buckyballs” of 60 carbon atoms (Vienna, 1999) and since then molecules of thousands of atoms. The mystery is not a quirk of electrons; it is how everything behaves when the experiment is delicate enough to reveal it.
At this point every honest person asks the obvious question: fine, but which slit did the electron actually go through? Let's just look. Put a small detector next to the slits — say, a light source that scatters a photon off each electron as it passes, revealing its position — so that for every electron we record: left slit or right slit.
The detector works perfectly. Every single electron is caught going through exactly one slit — never both, never half-and-half. Fifty percent left, fifty percent right, just like bullets. Great, mystery solved?
Look at the screen. The stripes are gone. With the which-slit detector switched on, the electrons pile up in two bullet-like heaps, one behind each slit. The interference pattern has vanished. Switch the detector off — give up knowing the path — and the stripes come back.
If the experiment makes it possible to know which path the electron took, the electron behaves like a particle and there is no interference. If the path information is not recorded anywhere — not by you, not by any device, not by any stray atom — the electron behaves like a wave exploring both paths, and interference appears. You can have a known path or stripes. Never both.
Why does looking change anything? Because in physics, “measurement” is a physical interaction, not an act of noticing. To learn which slit the electron used, something must interact with it — a photon must bounce off it, a wire must feel its charge, an atom must be nudged — and that interaction leaves a physical record correlated with the electron's path. It is this interaction-and-record that destroys the delicate wave behavior, whether or not any human ever reads the result. A detector left running in a sealed, empty lab kills the stripes just as dead.
Pop-science loves to say “the electron knows it's being watched” — hinting that a conscious mind changes reality. That is a misreading. The stripes vanish because of the physical interaction that records the path, and they vanish identically whether the detector's output goes to a scientist's eye, a hard drive, or nowhere at all. No eyes, no minds, no mysticism required. The genuinely hard (and genuinely open) question is what exactly counts as a completed measurement and why recorded outcomes are definite — that's the “measurement problem,” and we'll meet it properly in Module 6: Superposition & Measurement. But “consciousness causes collapse” is not part of working quantum mechanics.
So here is the scoreboard after three runs and a twist: electrons are always detected as whole particles, they travel in a way only waves can (both slits at once, interfering with themselves), and any attempt to pin down the path converts them back to bullet behavior. Nature will show you the wave story or the particle story, but never both in the same experiment. This package of facts is wave–particle duality — and in 1924, a French PhD student turned it into an equation.
By 1924 physics had grudgingly accepted half of the duality: light, the textbook wave, also behaves like particles (photons — Module 2). A young French doctoral student, Louis de Broglie (pronounced roughly “de Broy”), asked the symmetric question that in hindsight seems obvious and at the time seemed outrageous: if waves can be particles, why can't particles be waves? Maybe electrons — maybe all matter — have a wave associated with them too.
He even proposed the formula, by carrying over the relationship that already worked for photons. Every moving object has a wavelength — now called its de Broglie wavelength — given by:
$$\lambda = \frac{h}{p} = \frac{h}{mv}$$
Read the formula like a beginner, because that's all it takes: momentum is in the denominator, so the more momentum something has, the shorter its wavelength. Small, slow, light things (like electrons) get relatively long wavelengths; big, fast, heavy things (like baseballs) get absurdly short ones. And because \(h\) is so mind-bogglingly small, every matter wavelength is tiny — the only question is how tiny, and that question decides everything, as the worked examples below will show.
De Broglie put this in his 1924 PhD thesis. His examiners were so unsure what to make of it that they sent it to Einstein, who replied that the young man had “lifted a corner of the great veil.” Within three years the idea was confirmed head-on: in 1927, Davisson and Germer in the USA (and independently G.P. Thomson in Scotland) fired electrons at crystals and observed diffraction patterns — wave interference — at exactly the wavelength de Broglie's formula predicted. De Broglie received the 1929 Nobel Prize in Physics, the first person to win it for a PhD thesis. (A neat historical footnote: G.P. Thomson won a Nobel for showing the electron is a wave; his father J.J. Thomson had won one for showing it is a particle. Both were right.)
The electron's wave is not the electron smeared out like jam, and it's not a ripple in any material. The modern understanding — which we'll sharpen in Module 6 — is that it is a wave of probability: where the wave is strong, the electron is likely to be found; where interference makes the wave cancel, the electron is never found. The wave interferes; the detection is always one whole particle at one spot. That single sentence is the resolution of everything you saw on the screen.
Time to actually use \(\lambda = h/(mv)\). Constants we'll need: Planck's constant \(h = 6.626 \times 10^{-34}\ \text{J·s}\), electron mass \(m_e = 9.11 \times 10^{-31}\ \text{kg}\). Nothing beyond multiplication, division, and powers of ten.
A typical electron in a lab beam moves at about \(v = 1.0 \times 10^{6}\ \text{m/s}\). What is its de Broglie wavelength?
Momentum first:
$$p = mv = (9.11 \times 10^{-31}\ \text{kg}) \times (1.0 \times 10^{6}\ \text{m/s}) = 9.11 \times 10^{-25}\ \text{kg·m/s}$$
Then the wavelength:
$$\lambda = \frac{h}{p} = \frac{6.626 \times 10^{-34}}{9.11 \times 10^{-25}} \approx 7.3 \times 10^{-10}\ \text{m} \approx 0.73\ \text{nm}$$
Why this number matters: atoms are a few tenths of a nanometer across, and the spacing between atoms in a crystal is about \(0.2\)–\(0.5\ \text{nm}\). The electron's wavelength is the same size as atomic-scale structures. Interference only shows up when the slits (or crystal spacings) are comparable to the wavelength — which is exactly why electrons diffract off crystals and produce stripes in nanoscale double-slit setups. The wave is small, but the “slits” nature provides are small too.
A baseball has mass \(145\ \text{g} = 0.145\ \text{kg}\) and a decent pitcher throws it at \(v = 40\ \text{m/s}\) (about 144 km/h). Same formula:
$$p = mv = 0.145 \times 40 = 5.8\ \text{kg·m/s}$$
$$\lambda = \frac{6.626 \times 10^{-34}}{5.8} \approx 1.1 \times 10^{-34}\ \text{m}$$
How small is that? A proton — the nucleus of a hydrogen atom — is about \(1.6 \times 10^{-15}\ \text{m}\) across. The baseball's wavelength is roughly \(10^{19}\) times (ten billion billion times) smaller than a proton. To see interference you'd need slits separated by about a wavelength, and no slit, no crystal, no structure of any kind in the universe is remotely that fine. The baseball has a de Broglie wavelength in principle, but it is so preposterously short that its wave behavior can never, even theoretically, be revealed.
Quantum mechanics doesn't “switch off” for big objects — the same \(\lambda = h/p\) applies to electrons, dust grains, baseballs, and you. But because \(h \sim 10^{-34}\) sits in the numerator, any everyday momentum crushes the wavelength to invisibility. Balls fly in clean arcs and never make stripes not because they obey different laws, but because their wavelengths are absurdly, unobservably small.
Neutrons slowed down to room-temperature speeds (“thermal” neutrons) move at roughly \(v = 2200\ \text{m/s}\), and a neutron's mass is \(1.675 \times 10^{-27}\ \text{kg}\). Then:
$$\lambda = \frac{6.626 \times 10^{-34}}{(1.675 \times 10^{-27}) \times 2200} = \frac{6.626 \times 10^{-34}}{3.69 \times 10^{-24}} \approx 1.8 \times 10^{-10}\ \text{m} \approx 0.18\ \text{nm}$$
Again almost exactly the spacing between atoms in a crystal — a lucky coincidence of nature that makes thermal neutrons a superb probe of materials. Notice the pattern across all three examples: it's the ratio of \(h\) to the momentum that decides whether the quantum world shows itself.
Matter waves are not just a philosophical shock — they are working technology you can find in labs (and factories) all over the world.
Any microscope has a hard limit: it cannot resolve details much smaller than the wavelength of whatever it uses to look. Visible light has wavelengths of \(400\)–\(700\ \text{nm}\), so an ordinary light microscope can see bacteria (micrometers) but a virus (tens of nanometers) is hopelessly below its limit — and an atom is thousands of times smaller still.
Now recall Example 1: a modest lab electron already has \(\lambda \approx 0.73\ \text{nm}\) — hundreds of times shorter than visible light — and accelerating electrons harder shrinks the wavelength further, down to picometers. An electron microscope exploits exactly this: it “illuminates” the sample with electron matter waves and focuses them with magnetic lenses. The payoff is enormous: electron microscopes image viruses in crisp detail (this is how new viruses are photographed), and the best modern instruments resolve individual atoms. Every such image is a direct, everyday confirmation of \(\lambda = h/p\).
Because electron and thermal-neutron wavelengths match the spacing between atoms in a solid (Examples 1 and 3), a crystal acts like a natural multi-slit barrier for them. Fire a beam at a material and the atoms' regular rows produce an interference pattern — and from the geometry of that pattern you can work backwards to the exact arrangement of the atoms. This is precisely Davisson and Germer's 1927 experiment, industrialized: electron diffraction and neutron diffraction are now standard tools for mapping crystal structures, checking semiconductor wafers, and studying new materials. Chemists, geologists, and chip manufacturers routinely rely on the fact that particles interfere like waves.
Three problems: one prediction, one calculation, one concept. Try each honestly before opening the solution — the struggle is where the learning happens.
You run the double-slit experiment twice. Run A: you pour fine grains of sand toward the two slits. Run B: you send single photons, one at a time, toward two extremely narrow, closely spaced slits. For each run, predict the pattern that accumulates on the far screen, and explain why they differ.
Run A (sand): two piles, one behind each slit, blending in the middle — the bullet pattern, with no stripes. Each grain is a macroscopic object; its de Broglie wavelength (\(\lambda = h/p\)) is fantastically smaller than any slit, so its wave nature is utterly unobservable and it simply goes through one slit or the other. Run B (single photons): each photon lands as one dot at one spot, but as thousands of dots accumulate they build interference stripes — bright bands separated by lanes where photons never land. A photon's wavelength is comparable to the slit spacing, so its wave passes through both slits and interferes with itself; the wave sets the probabilities, the detection is always one whole photon. The difference between the runs is purely the ratio of wavelength to slit scale — not a different set of laws.
An electron (\(m_e = 9.11 \times 10^{-31}\ \text{kg}\)) moves at \(v = 2.0 \times 10^{6}\ \text{m/s}\) — twice the speed of the one in Worked Example 1. Compute its de Broglie wavelength with \(h = 6.626 \times 10^{-34}\ \text{J·s}\). Before calculating, predict: should the answer be longer or shorter than the \(0.73\ \text{nm}\) we found earlier?
Prediction: momentum is in the denominator, so doubling the speed should halve the wavelength — expect about \(0.36\ \text{nm}\). Calculation:
$$p = mv = (9.11 \times 10^{-31}) \times (2.0 \times 10^{6}) = 1.82 \times 10^{-24}\ \text{kg·m/s}$$
$$\lambda = \frac{h}{p} = \frac{6.626 \times 10^{-34}}{1.82 \times 10^{-24}} \approx 3.6 \times 10^{-10}\ \text{m} \approx 0.36\ \text{nm}$$
Exactly half of Worked Example 1's answer, as predicted — still comfortably atom-sized, so this electron would also diffract beautifully off a crystal.
A classmate says: “I get it — the electron splits into two halves, one half goes through each slit, and the halves recombine at the screen.” What experimental facts contradict this picture, and what is the better way to describe what goes through both slits?
What's wrong: nobody has ever detected half an electron. Every detection — at the screen, or at a which-slit detector — finds exactly one whole electron at one spot, with its full charge and full mass. If the electron literally split, a detector at the slits should sometimes catch half an electron at each slit simultaneously; instead it always catches the entire electron at one slit or the other. The better picture: what passes through both slits is not a fragment of the electron but its wave — a wave of probability. The wave takes both paths and interferes with itself, and the resulting pattern of strong and cancelled regions dictates where the (always whole, always single) electron is likely or forbidden to land. “Wave when traveling, particle when detected” is the honest beginner summary — and making the word probability precise is exactly the job of Module 6 on superposition and measurement.
How waves interfere and why stripes are a wave's fingerprint; what bullets, water waves, and single electrons each do at two slits — and why the electron result contains, in Feynman's words, the only mystery; why recording the path destroys interference (physical interaction, not conscious watching); de Broglie's matter-wave formula \(\lambda = h/p\) and how to use it; and why electron wavelengths reveal the quantum world while a baseball's wavelength — \(10^{19}\) times smaller than a proton — keeps the everyday world looking classical.
And here's the cliffhanger: matter waves are about to solve a puzzle we've been quietly ignoring since Module 1. According to classical physics, an electron orbiting a nucleus should radiate away its energy and spiral into the nucleus in about a hundred-millionth of a second — every atom in the universe should collapse almost instantly. They obviously don't. Why not? Because the electron is a wave, and a wave wrapped around a nucleus can only settle into certain snug, self-consistent patterns — discrete energy levels. That is Bohr's quantum atom, it explains why every element glows with its own barcode of colors, and it's the whole story of Module 4: Atoms, Energy Levels & Spectra.