Electromagnetism
The Betatron: How a Changing Magnetic Field Whips Electrons to Millions of Volts
In July 1940, in a basement at the University of Illinois, Donald Kerst switched on a 60 Hz electromagnet the size of a washing machine and watched a faint glow appear on a phosphor screen. Electrons, injected at a few hundred volts, had spiraled in a fixed circle and emerged at 2.3 million electron-volts — accelerated not by any electrode, not by any voltage anyone could measure with a meter, but by the pure electric field that a changing magnetic flux conjures out of empty space. There were no accelerating gaps. The electron never touched a high-voltage terminal. It was pushed entirely by Faraday's law.
The betatron is the most literal machine ever built to embody ∮E·dl = −dΦ/dt. It is a transformer whose secondary winding is a single beam of electrons, and its central trick — the reason the beam neither flies outward nor collapses inward as it gains energy — is a beautifully simple geometric rule: the field at the orbit must always equal exactly half the average field enclosed by it.
- Governing law∮E·dl = −dΦ/dt (Faraday)
- Key conditionB_orbit = ½⟨B⟩ (2:1 rule)
- First machineKerst, 1940 — 2.3 MeV e⁻
- Drive frequency50–200 Hz AC magnet
- Typical energy1–300 MeV electrons
- LimitSynchrotron radiation loss ∝ E⁴/R²
Interactive visualization
Press play, or step through manually. The visualization is yours to drive — try it before reading on.
Watch the 60-second explainer
A condensed visual walkthrough — narrated, captioned, under a minute.
The core idea: an electric field with no charges in sight
Faraday's law says a changing magnetic flux threads a closed loop with a circulating electric field: ∮E·dl = −dΦ/dt, where Φ = ∫B·dA is the flux through the loop. In a betatron the loop is the electron's circular orbit of radius R, so the line integral is just E multiplied by the circumference: E·(2πR) = −dΦ/dt. The tangential electric field at the orbit is therefore
- E_θ = −(1/2πR)·dΦ/dt
This E_θ points around the ring and does exactly what an accelerating cavity would — it exerts a force −eE_θ on the electron along its direction of motion. Because the field is induced, not electrostatic, there is no source charge, no terminal, no gap. The relativistic momentum equation is dp/dt = −eE_θ, and substituting gives
- dp/dt = (e/2πR)·dΦ/dt
Integrating over the rising quarter-cycle of the AC magnet, the electron's momentum grows in lockstep with the enclosed flux: p = eΦ/(2πR), plus its small injection momentum. Ramp the flux from zero to a few tenths of a weber in a couple of milliseconds and the momentum — hence the energy — climbs to megaelectron-volts. The betatron is, in the cleanest possible sense, a device that converts flux into momentum.
The 2:1 rule: why the orbit doesn't drift in or out
Here is the problem the machine must solve. The same magnet that accelerates the electron (via the flux inside the orbit) also has to bend the electron in a circle (via the field at the orbit). A relativistic particle of momentum p moving in a magnetic field B_R follows a circle of radius R given by the magnetic-rigidity relation p = e·B_R·R. As acceleration proceeds, p grows — so to keep R fixed, B_R must grow in exact proportion to p.
Compare the two momentum expressions. Acceleration demands p = eΦ/(2πR), and guidance demands p = eB_R·R. Setting them equal:
- eB_R·R = eΦ/(2πR)
- B_R = Φ/(2πR²)
Now define the average field enclosed by the orbit as ⟨B⟩ = Φ/(πR²). Then the equation above reads
- B_R = ½⟨B⟩
That is the celebrated betatron condition, or Widerøe's 2:1 flux rule: the magnetic field guiding the electron at its orbit must be exactly half the mean field threading the enclosed area, and this must hold at every instant as the AC current ramps. The magnet's pole pieces are shaped — a dense flux plug in the middle, a weaker field at the rim — so that this ratio is baked into the geometry and automatically maintained for all time. Get it right and R never changes: the electron rides a single circle while its energy multiplies ten-thousand-fold.
Betatron oscillations: the stability that keeps the beam alive
A fixed orbit is not enough; the beam also needs to be stable against small kicks. Nudge an electron slightly outward and it must feel a net restoring force back toward R, both radially and vertically. This is set by how fast the guide field falls off with radius, captured by the field index n = −(R/B_R)·(∂B_R/∂R).
- Vertical (axial) focusing requires the field lines to bow outward, which happens when n > 0.
- Radial (horizontal) focusing requires n < 1.
Both together give the stability window 0 < n < 1, and practical machines run near n ≈ 0.5 to 0.75. Within this band a displaced electron doesn't wander off — it executes small transverse oscillations about the ideal orbit at frequencies ω_r = ω₀√(1−n) radially and ω_z = ω₀√n vertically, where ω₀ is the orbital angular frequency. These are the betatron oscillations, and the term is so fundamental that it survived into every synchrotron and storage ring built since; physicists still speak of a beam's "betatron tune" even in machines with no betatron in them. The oscillations are the reason a millimeter-scale injected bunch stays confined through the roughly 250,000 revolutions of a single acceleration cycle.
A worked cycle: from a few hundred volts to a few MeV
Picture Kerst's machine running on the rising quarter of a 60 Hz sine, so the flux ramps over about T/4 ≈ 4.2 ms. Electrons are injected from a hot filament at only ~few-hundred-eV energy, so they are already moving at a few percent of c. In each turn they gain a tiny sliver of energy — the induced EMF around one loop is
- EMF = dΦ/dt ≈ ΔΦ/Δt
If the enclosed flux swings by ΔΦ ≈ 0.4 Wb over Δt ≈ 4 ms, the average induced EMF is about 100 volts per turn. That sounds feeble — but the electron circulates the ~5 m-circumference orbit at nearly the speed of light, roughly 6 × 10⁷ revolutions per second, so it collects those 100 eV kicks a quarter-million times before the flux peaks. The total energy is EMF × (number of turns) ≈ 100 V × 250,000 ≈ 2.5 × 10⁷ eV = 25 MeV for a large machine, and a couple of MeV for Kerst's first small one. The genius is that time, not voltage, does the work: a modest EMF applied over a huge path length replaces the impossibly large single voltage a DC machine would need.
What sets the ceiling: synchrotron radiation and the flux budget
Betatrons top out around 100–300 MeV for electrons, and the reason is synchrotron radiation. Any charge on a curved path radiates, and for a relativistic electron the power scales viciously as P ∝ E⁴/(m⁴R²). Every joule radiated away must be resupplied by the induced field, and past a few hundred MeV the losses climb faster than the ramp can pump energy in — the 2:1 condition would need field ramps the iron core cannot deliver. A second ceiling is the flux budget: since p = eΦ/(2πR), reaching higher momentum at fixed R demands proportionally more enclosed flux, and iron saturates near ~1.5–2 T. You cannot simply push more current through the coils. These twin limits are why the betatron was superseded by the synchrotron — which decouples the guide field from acceleration, uses separate RF cavities, and can climb into the GeV and TeV range. Kerst himself built ever-larger betatrons, culminating in a 300 MeV machine at Illinois in 1949, which for years was the record electron energy on Earth.
Where betatrons actually went to work
The betatron was never a curiosity — it became a workhorse for high-energy X-rays. Slam a 20–30 MeV electron beam into a tungsten target and it produces intense bremsstrahlung, a penetrating photon beam far harder than any X-ray tube can make. Two applications dominated:
- Industrial radiography. Compact betatrons could image welds through 30 cm of steel, inspecting pressure vessels, castings, and rocket motors where cobalt-60 sources were too weak. Portable betatrons were even lowered into boreholes for oil-well logging.
- Radiation therapy. From the 1950s through the 1980s, medical betatrons delivered deeply penetrating photon and electron beams to treat tumors, sparing skin because the dose peaks centimeters below the surface. They were eventually displaced by more compact linear accelerators (linacs), which give the same photon energies from a machine that fits on a gantry.
The betatron's deepest legacy, though, is conceptual. The stability analysis Kerst and Robert Serber published in 1941 — the mathematics of betatron oscillations and the field index — is the direct ancestor of the transverse-focusing theory that governs every circular accelerator today, from medical synchrotrons to the LHC.
| Property | Betatron | Cyclotron |
|---|---|---|
| Accelerating field | Induced E from dΦ/dt (no electrodes) | Oscillating E across a gap (dee voltage) |
| Orbit radius | Fixed (constant R by design) | Grows with energy (spiral out) |
| Guiding field | Same magnet also accelerates | Separate static field guides only |
| Particle | Electrons (light, relativistic fast) | Protons, ions (heavy, non-relativistic) |
| Energy ceiling | ~300 MeV (radiation loss) | ~25 MeV protons (relativistic detuning) |
| Duty cycle | Pulsed, one AC quarter-cycle | Continuous or pulsed RF |
Frequently asked questions
Why does the guide field have to be exactly half the average field?
Because the same magnet must do two jobs at once. The flux inside the orbit sets the accelerating EMF (p = eΦ/2πR), while the field at the orbit sets the bending radius (p = eB_R·R). Forcing both to give the same momentum yields B_R = Φ/2πR², and since the average field is ⟨B⟩ = Φ/πR², that ratio is exactly ½. Break the 2:1 rule and the orbit spirals in or out as the electron accelerates.
Where does the energy come from if there are no electrodes?
From the changing magnetic flux itself, via Faraday's law. A time-varying flux induces a circulating electric field (∮E·dl = −dΦ/dt) that has no source charges — it exists purely because the magnetic field is changing. That induced field, wrapped around the electron's circular path, pushes it forward on every one of its ~250,000 orbits. Ultimately the energy is supplied by the AC power feeding the electromagnet's coils.
Why can't a betatron reach GeV energies like a synchrotron?
Two hard limits. First, synchrotron-radiation losses scale as E⁴/R², so beyond a few hundred MeV a relativistic electron radiates energy faster than the induced field can replenish it. Second, the flux budget: higher momentum needs more enclosed flux, but the iron core saturates near 1.5–2 T. Synchrotrons dodge both by separating the guide magnet from RF accelerating cavities, so each can be optimized independently.
What are 'betatron oscillations' and why do modern accelerators still use the term?
They are the small transverse wobbles an electron makes about its ideal orbit, at frequencies ω₀√(1−n) radially and ω₀√n vertically, where n is the field index. Stable focusing requires 0 < n < 1. The concept generalizes to any circular machine: engineers of synchrotrons and storage rings still specify a beam's 'betatron tune' — the number of these oscillations per revolution — even though those machines contain no betatron.
How fast is the electron actually moving in a betatron?
Almost at the speed of light almost immediately. A 2 MeV electron already has a Lorentz factor γ ≈ 5, meaning it travels at about 0.98c. At only a few hundred keV it is already relativistic (rest energy of an electron is just 511 keV), which is why betatrons accelerate light electrons rather than heavy protons — the electron's speed is essentially pinned at c and only its momentum and energy keep climbing.
Who invented the betatron and when?
Donald Kerst built the first working betatron at the University of Illinois in 1940, accelerating electrons to 2.3 MeV. The idea of induction acceleration and the 2:1 flux rule traces to earlier work by Rolf Widerøe and Joseph Slepian, but Kerst — with theorist Robert Serber — cracked the orbit-stability problem that made a practical machine possible. Kerst later reached 300 MeV in 1949, then a world-record electron energy.