← quantum · contents

§ Phase iv · Spin · Chapter 01

Stern–Gerlach

In the winter of 1922, in an unheated laboratory in Frankfurt, two physicists pointed a beam of silver atoms at a wedge-shaped magnet and developed a glass plate. The smudge they found, two clean spots where classical physics had predicted a vertical smear, was the first direct sighting of a property no one had yet named. The property would later be called spin.

The story starts with a bet. In the autumn of 1921, the young theoretical physicist Otto Stern was sitting in a Frankfurt cafe, irritated. The old quantum theory of Bohr and Sommerfeld claimed that an atom in a magnetic field could only point its angular momentum vector along certain discrete directions, a doctrine the Germans called Richtungsquantelung, or space quantization. To Stern, raised on classical mechanics, the idea was preposterous. An atom, freely tumbling through empty space, suddenly choosing to align itself with a magnetic field that had not yet been switched on? Nonsense. “I am still skeptical that quantum theory is correct,” he wrote to a friend that year. He would prove the Bohr–Sommerfeld people wrong, and he would do it experimentally, using a beam of atoms in a vacuum.

He had the tool to do it. While in Frankfurt he had refined a method he called the Molekularstrahlmethode, the molecular beam technique, in which atoms boiled out of a hot oven through a narrow slit, free of collisions and free of fields. A beam of bare atoms is a physicist’s dream: every atom is a tiny gyroscope, naked and unperturbed, ready to interact with whatever you point at it. Stern had already used such beams to verify the Maxwell–Boltzmann distribution of molecular speeds. Now he proposed to point one at an inhomogeneous magnet, and watch what happened.

The classical prediction was straightforward, and Stern wrote it down clearly. A silver atom is electrically neutral but has a tiny magnetic moment, like a microscopic bar magnet. In a uniform magnetic field the two poles of the atom feel equal and opposite forces and the atom drifts in a straight line. In a non-uniform field, where one pole sits in a stronger field than the other, the forces no longer cancel, and the atom is deflected. The amount of the deflection depends on how the magnetic moment is oriented relative to the field gradient. If, as classical mechanics required, the orientations are random and continuous, the beam should arrive at the detector as a vertical smear, a band wider at the middle and tapering at the ends. If, as Sommerfeld and Bohr insisted, the orientations are quantized, the smear should split into discrete bands. Stern intended to show that the smear was a smear.

To run the experiment he needed an experimentalist, and at Frankfurt he found one. Walther Gerlach, six years younger than Stern, was a quiet, careful man with strong hands and an instinct for vacuum systems. The two were a study in contrasts: Stern theoretical, slightly arrogant, lover of dinners and cigars; Gerlach methodical, taciturn, lover of brass fittings and polished optics. They began work in the early summer of 1921.

The experiment itself was a nightmare of engineering. The active flight path, from furnace to collector plate, was about 20 centimeters long, set inside a much larger glass vacuum vessel that had to be pumped down to a millionth of an atmosphere. At one end, an electrically heated furnace boiled metallic silver at 1000 degrees Celsius, a temperature only modestly forgiving to the brass and the pump oil. A vapor of silver atoms streamed out through a slit the width of a human hair, then through a second slit about three centimeters downstream, producing a thin, collimated beam. The beam then crossed a five-centimeter gap between the two iron pole pieces of a powerful electromagnet, custom-shaped so that one pole was a sharp wedge and the other a grooved channel. That asymmetry was the crucial trick: the wedge produced an enormous field gradient, on the order of ten tesla per centimeter, far stronger than the field itself in absolute terms. After the magnet, the silver beam coasted another ten centimeters or so to a polished glass collector plate, where the atoms condensed in a faint metallic film. Total flight time per atom: a few hundred microseconds.

Funding for this work was nearly nonexistent. Germany in 1921 was a country flattened by the war, gripped by hyperinflation, where a loaf of bread might cost a billion marks by year’s end. Stern was paying for filaments and pump oil out of his own postdoc stipend. The story goes that he wrote to Einstein for help, and Einstein, then in Berlin and already a celebrity, raised a few thousand dollars from an American banker named Henry Goldman of Goldman Sachs, a sum that kept the lab running for another year. Without that loan, the experiment might never have been completed.

The first attempt, in the autumn of 1921, gave no visible result. The field was too weak, the alignment was off, the vacuum was leaky. Gerlach spent the winter rebuilding. By February 7, 1922, the apparatus was running cleanly. Gerlach pulled the plate, breathed on it, and saw it: not the smooth Gaussian smear that classical mechanics predicted, but two distinct lips of silver, one above where the beam would have gone with the field off, one below. In the middle, where most of the silver should have landed in any classical picture, the plate was nearly bare. The atoms had taken sides.

The next morning Gerlach mailed a postcard to Niels Bohr in Copenhagen. The card carried a photograph of the developed glass plate, the unmistakable two-lobed silver deposit, and a single hand-written sentence above it. “Anbei die experimentelle Bestätigung der Richtungsquantelung.” Attached, the experimental confirmation of space quantization. He signed it “with respectful greetings, your devoted Walther Gerlach.” Bohr framed it and hung it in his office.

What had they actually shown? In their own minds, in February 1922, Stern and Gerlach believed they had confirmed Bohr–Sommerfeld space quantization. The Bohr–Sommerfeld picture predicted that an atom with orbital angular momentum quantum number l = 1 (which the ground state of silver was wrongly believed to have at the time) should give 2l + 1 = 3 spots, but with the central undeflected spot suppressed by a kinematic factor, leaving only the two outer ones visible. So the result fit the old theory, sort of, with some fudging. The newspapers in Germany ran the story as a triumph for the new physics, and Stern was instantly famous within the small world of European theoretical physicists. The bet against quantum theory had failed; he had proved himself wrong with his own apparatus.

The real explanation took three more years to surface, and it was nothing the Frankfurt team had imagined. In 1925 two young Dutch physicists in Leiden, George Uhlenbeck and Samuel Goudsmit, proposed that the electron itself carries an intrinsic angular momentum of magnitude h-bar over two, a quantum of spinning that has nothing to do with orbital motion around a nucleus. The orbital angular momentum of silver’s outermost electron is in fact zero (it sits in an s orbital), so the magnetic moment the Frankfurt apparatus was measuring did not come from any orbit at all. It came from the new property. Pauli, who had at first ridiculed the spin proposal and would soon write down the matrix algebra that made it work, eventually accepted it. The two spots on Gerlach’s plate, it turned out, were the two values of electron spin along the z-axis. Spin up. Spin down. Nothing in between.

The reason the two-spot result is so deeply strange is worth slowing down on. In a classical picture, before any measurement, a silver atom heading toward the magnet has some magnetic moment vector pointing in some direction. That direction could be anywhere on a sphere: straight up, straight down, sideways, at 37 degrees, at 89.9 degrees. Inside the magnet, the projection of that moment along z determines how hard the atom is deflected. If the moment points straight up, the atom is deflected the maximum amount upward. If it points sideways, no deflection at all. If it points at 60 degrees, a moderate deflection. A whole oven full of atoms, tumbling in every direction, should give a smooth distribution of deflection amounts. You should see a streak.

You do not see a streak. You see two spots. Every single atom that enters the apparatus, no matter what direction its moment was pointing as it left the oven, arrives at one of exactly two locations on the detector. This is not because the oven only emits two kinds of atoms (it does not, the emission is thermal and isotropic). It is because the very act of measuring the z-component of the magnetic moment forces the answer to be either +h-bar/2 or -h-bar/2. There is no third option. There is no in-between. Whatever the silver atom was doing before it met the magnet, after the magnet it is doing exactly one of two things, and a coin gets flipped to decide which.

F = -∇U = ∇(μ · B)

F_z ≈ μ_z · (∂B_z/∂z)

F_z = +μ_B · (∂B_z/∂z)    (spin up, atom kicked upward)
F_z = -μ_B · (∂B_z/∂z)    (spin down, atom kicked downward)

The aftermath of the experiment is its own quiet drama. Stern left Germany in 1933, fleeing the Nazi regime; he ended up at the Carnegie Institute of Technology in Pittsburgh, where he continued to refine molecular beam techniques and eventually measured the magnetic moment of the proton, finding it nearly three times the value predicted by Dirac’s relativistic equation. (The mismatch hinted that the proton, unlike the electron, had structure. That structure would later be called quarks.) The Nobel Committee awarded Stern the 1943 Physics Prize for his molecular beam method and proton-moment measurement. He could not collect it in Stockholm because of the war; the medal was sent to him in America a year later. Gerlach, who stayed in Germany, spent the war as the head of the German Uranverein, the abortive nuclear-weapons program. He was interned at Farm Hall in England after the war along with Heisenberg and the rest, returned to academic life in Munich, and lived until 1979. The Nobel passed him over. There is no Gerlach prize. The cigar smoke that developed the plate, and the patient hands that built the apparatus, did not earn him a trip to Stockholm.

The experiment itself became, in the decades that followed, the workhorse of quantum measurement. Feynman, in his Lectures, builds the entire formalism of quantum mechanics around chains of idealized Stern–Gerlach magnets, each one preparing or measuring a spin. Modern atomic clocks, the standard of time itself, descend directly from Rabi’s molecular-beam magnetic resonance, which is a Stern–Gerlach magnet with an oscillating field added in the middle. The hydrogen maser that defines our cesium-based second is, at its heart, the Frankfurt apparatus reinvented for the atomic age. Every GPS satellite carries a Stern–Gerlach descendent.

And the two spots? They remain, in 2026, the cleanest single demonstration we have that the world is not classical. The plate is not a metaphor. The silver is real. The split is real. Half a centimeter of separation, after a brass tube the length of your forearm, is enough to prove that the universe does not permit you to know the orientation of an atomic magnet in advance, only to discover it by asking. The asking changes the answer. The answer comes in two flavors. Welcome to spin.