Pulsar Magnetospheres and Spin-Down

A pulsar is what a rotating, strongly magnetized neutron star looks like from a few kiloparsecs away: a clock that ticks with the star's spin, discovered and timed through the coherent radio beam it sweeps across the sky. Underneath the lighthouse image the physics is pure electrodynamics. Take an object of radius $R\sim10$ km and moment of inertia $I\sim10^{45}\,\mathrm{g\,cm^2}$ , give it a dipole field $B\sim10^{8}$ – $10^{15}$ G, and spin it with a period $P$ of milliseconds to seconds. The rotating field cannot coexist with empty space — it induces fields strong enough to tear charge off the surface — and the resulting torque bleeds rotational energy at a prodigious rate, most of it into a relativistic wind and only a whisper into the radio pulses we detect. Two ideas do almost all the work. The first is the rotating magnetic dipole: a field tilted by $\alpha$ from the spin axis has a component that sweeps at the spin frequency, and an accelerating magnetic moment radiates by the Larmor formula, yielding the spin-down law $\dot E\propto B^2R^6\Omega^4$ and the sharp prediction $n\equiv\Omega\ddot\Omega/\dot\Omega^2=3$ . The second is the Goldreich–Julian magnetosphere: vacuum is impossible, so the star sits inside a corotating, charge-separated plasma bounded by the light cylinder $R_{LC}=c/\Omega$ , where rigid corotation would demand the speed of light. Real pulsars live between these idealizations, and the measured braking indices ( $n<3$ for every young pulsar) are the fingerprint of that in-between physics. This article assembles the machine: the fields, the torque, the timing observables, and the $P$ – $\dot P$ diagram that organizes the population.

Key takeaways

A pulsar is a spinning neutron star whose magnetic dipole is misaligned with the spin axis; the time-varying moment radiates, torques the star, and produces the slowdown seen as an increasing pulse period.
The vacuum dipole model gives $\dot E\propto B^2R^6\Omega^4$ and predicts $n=3$ ; every measured young pulsar shows $n<3$ , because a plasma-filled magnetosphere and its wind carry the torque.
A rotating conductor induces $E_\parallel\sim10^{12}\,\mathrm{V\,cm^{-1}}$ that rips charge off the star, filling space with Goldreich–Julian plasma of density $\rho_{GJ}=-\boldsymbol\Omega\cdot\mathbf B/2\pi c$ .
The light cylinder $R_{LC}=cP/2\pi$ separates the closed corotating zone from open lines threading the polar caps, where pair cascades feed the coherent radio beam.
Timing $P$ and $\dot P$ gives $\tau_c=P/2\dot P$ and $B\simeq3.2\times10^{19}\,(P\dot P)^{1/2}$ G; the loss $\dot E=-I\Omega\dot\Omega$ powers the wind, and the Crab's $4.6\times10^{38}\,\mathrm{erg\,s^{-1}}$ lights its nebula almost exactly.
Key idea: A rotating misaligned magnetic dipole cannot sit in vacuum — it builds a corotating charge-separated (Goldreich–Julian) magnetosphere out to the light cylinder, and the electromagnetic torque drains rotational energy at $\dot E=-I\Omega\dot\Omega\propto B^2R^6\Omega^4$ , observed as the steady lengthening of the pulse period.

The core insight

A rotating misaligned magnetic dipole cannot sit in vacuum: it induces electric fields that build a corotating charge-separated (Goldreich–Julian) magnetosphere out to the light cylinder, and the electromagnetic torque on the star drains rotational energy at the rate $\dot E=-I\Omega\dot\Omega\propto B^2R^6\Omega^4$ , observable as the secular increase of the pulse period.

Where this lives today

Pulsar spin-down is no longer just a slowing clock — it is a precision instrument. Millisecond pulsars are so stable that arrays of them (NANOGrav and the International Pulsar Timing Array) act as a galaxy-sized gravitational-wave detector; the 2023 nanohertz background rests entirely on modelling each pulsar's spin-down and watching for correlated residuals. The torque physics is now computed rather than guessed: particle-in-cell simulations follow the polar-gap pair cascades and reproduce the transition from near-zone dipole to striped wind, resolving the aligned-rotator paradox. Magnetars ( $B\gtrsim10^{14}$ G) power giant flares and at least one confirmed fast radio burst (SGR 1935+2154), tying magnetospheric physics to transient astronomy. And each clean $\ddot P$ below three sharpens the tension between the textbook vacuum dipole and the plasma-filled reality.

The physical picture

Picture a bar magnet frozen into a spinning, perfectly conducting sphere, its axis tilted from the spin axis. In the lab frame the dipole-moment vector wheels around once per rotation, like a lighthouse beam. An accelerating magnetic moment radiates, carrying off angular momentum so the star slows — and only the perpendicular component $m\sin\alpha$ accelerates, so an aligned dipole ( $\alpha=0$ ) radiates nothing in vacuum while an orthogonal one ( $\alpha=90^\circ$ ) radiates the most. Because the star is an excellent conductor, its interior obeys $\mathbf E+(\mathbf v\times\mathbf B)/c=0$ . This unipolar-generator condition forces a field-aligned electric component just outside the surface of order $(\Omega R/c)B\sim10^{12}\,\mathrm{V\,cm^{-1}}$ — far too strong for gravity to hold charges down — so the star sprays plasma outward until the exterior lines corotate rigidly. That corotation cannot extend past the light cylinder $R_{LC}=c/\Omega$ , where a corotating point would move at $c$ : inside it, lines close and trap plasma; outside, they open and let plasma stream to infinity, so the period sets the magnetosphere's size.

Figure

The oblique rotator: the magnetic axis μ is tilted by α from the spin axis Ω. Lines closing inside the light cylinder R_LC = cP/2π corotate rigidly; lines crossing it open and channel plasma out through the polar caps. For the Crab (P = 33 ms) the light cylinder sits at R_LC ≈ 1580 km, about 160 stellar radii

Only a tiny fraction of the spin-down power emerges as radio waves — the efficiency is $\sim10^{-5}$ ; the bulk goes into the particle wind. Yet the pulses are what we see, because they are coherent: bunches of pair plasma radiate in phase, boosting the brightness temperature to $10^{25}$ – $10^{30}$ K, far above any thermal source. Plotting each pulsar by $P$ and $\dot P$ sorts the population into families — ordinary pulsars, recycled millisecond pulsars, and magnetars. The second figure is that map, the periodic table of the pulsar zoo.

Figure

The P–Ṗ diagram, the periodic table of pulsars: ordinary pulsars cluster near P ∼ 0.5 s and B ∼ 10¹² G, recycled millisecond pulsars fill the low-P low-Ṗ corner, magnetars the high-field top right. Blue diagonals are constant surface field B = 3.2×10¹⁹ √(P Ṗ) G; green diagonals are constant age τ_c = P/2Ṗ

The core equations

The magnetosphere has one intrinsic length scale, fixed by the spin alone: the radius at which a corotating point would reach the speed of light. All the geometry refers back to it.

R_{LC}=\frac{c}{\Omega}=\frac{cP}{2\pi} \qquad (1)

where: $R_{LC}$ — light-cylinder radius; $\Omega=2\pi/P$ — spin angular frequency; $P$ — pulse period. Reading: a $P=1$ s pulsar has $R_{LC}\approx4.8\times10^4$ km, a $P=1.6$ ms pulsar only $\approx76$ km — the faster the spin, the tighter the magnetosphere. Vacuum is not an option. A conductor spinning in its own dipole field develops a corotation electric field whose divergence demands a real space charge — the Goldreich–Julian density the magnetosphere must supply to screen $E_\parallel$ .

\rho_{GJ}=-\frac{\boldsymbol\Omega\cdot\mathbf B}{2\pi c}\left(1-\frac{\Omega^2 r^2\sin^2\theta}{c^2}\right)^{-1} \qquad (2)

where: $\rho_{GJ}$ — corotation charge density; $\boldsymbol\Omega\cdot\mathbf B$ — field projection on the spin axis; $r,\theta$ — spherical coordinates. Reading: near the star the bracket is unity; $\rho_{GJ}$ flips sign across the null surface $\boldsymbol\Omega\cdot\mathbf B=0$ , setting up the global current circuit. The number density is $n_{GJ}\approx7\times10^{-2}\,B/P\ \mathrm{cm^{-3}}$ . The torque is measured, not assumed. Differentiating the rigid-body rotational energy turns the observed period drift into a spin-down power.

\dot E=-I\Omega\dot\Omega=\frac{4\pi^2 I\dot P}{P^3} \qquad (3)

where: $\dot E$ — spin-down luminosity; $I\approx10^{45}\,\mathrm{g\,cm^2}$ — moment of inertia; $\dot P$ — period derivative (s/s). Reading: since $\dot P>0$ and $\dot\Omega<0$ , $\dot E<0$ ; the $\dot P/P^3$ scaling means young fast pulsars dominate the Galaxy's particle supply. The origin of that torque, in the simplest model, is magnetic-dipole radiation: only $m\sin\alpha$ accelerates, and the Larmor formula turns its second derivative into radiated power.

\dot E_{\mathrm{dip}}=\frac{2}{3c^3}\,|\ddot{\mathbf m}|^2=\frac{B^2R^6\Omega^4\sin^2\alpha}{6c^3} \qquad (4)

where: $\mathbf m$ — dipole moment, $|\mathbf m|=\tfrac12 BR^3$ (polar field); $B$ — polar field; $R$ — radius; $\alpha$ — spin–magnetic angle. Reading: the loss climbs as $\Omega^4$ and $B^2$ , and vanishes for an aligned vacuum rotator ( $\alpha=0$ ). If the torque is a power law in $\Omega$ , its exponent is one measurable number. Timing $P,\dot P,\ddot P$ isolates it free of any unknown constant.

\dot\Omega=-K\,\Omega^{\,n},\qquad n\equiv\frac{\Omega\ddot\Omega}{\dot\Omega^2} \qquad (5)

where: $K$ — torque constant; $n$ — braking index. Reading: vacuum-dipole braking gives $n=3$ ; real young pulsars measure $n<3$ , exposing evolving geometry, field growth, or wind torque. Setting the observed torque equal to the orthogonal vacuum-dipole loss inverts the timing into a field estimate — the formula behind every point on the $P$ – $\dot P$ diagram.

B\simeq3.2\times10^{19}\,(P\dot P)^{1/2}\ \mathrm{G} \qquad (6)

where: $B$ — characteristic surface field; $P$ — period (s); $\dot P$ — (s/s). Reading: uses $I=10^{45}$ , $R=10^6$ cm and the equatorial convention $|\mathbf m|=BR^3$ ; the polar field is twice this. An order-of-magnitude proxy, not a measurement. Integrating the $n=3$ braking law from a fast birth spin gives an age estimate needing no field or distance.

\tau_c\equiv\frac{P}{2\dot P}=-\frac{\Omega}{2\dot\Omega} \qquad (7)

where: $\tau_c$ — characteristic (spin-down) age. Reading: equals the true age only if $P_0\ll P$ and $n=3$ held throughout; for the Crab it gives 1245 yr against a true age near 970 yr. Along the open field lines the unscreened potential is enormous — it powers the pair cascade and the coherent beam.

\Delta\Phi_{pc}\approx\frac{\Omega^2 B R^3}{2c^2}\approx6.6\times10^{12}\,\frac{B_{12}}{P^2}\ \mathrm{V} \qquad (8)

where: $\Delta\Phi_{pc}$ — maximum open-field potential; $B_{12}=B/10^{12}\,\mathrm G$ ; $P$ — period (s). Reading: $\sim10^{13}$ V for an ordinary pulsar, $\sim10^{16}$ V for the Crab — enough to drive primaries to $\gamma\sim10^7$ and ignite the pair avalanche.

Derivation AThe rotating vacuum dipole and the braking index n = 3

The goal is to derive, from the Larmor formula alone, the spin-down luminosity of a magnetic dipole tilted by $\alpha$ and spun at $\Omega$ , and to show that this single model makes a falsifiable prediction — a braking index of exactly three — testable against timing data with no free parameters. We split the dipole into constant axial and rotating perpendicular parts, and convert the radiated power into a torque law whose exponent is the braking index.

Step 1. Write the moment as $\mathbf m=m(\cos\alpha\,\hat{\mathbf z}+\sin\alpha[\cos\Omega t\,\hat{\mathbf x}+\sin\Omega t\,\hat{\mathbf y}])$ . The axial piece is constant; the perpendicular piece $m\sin\alpha$ sweeps the equatorial plane at the spin frequency. Why: radiation is sourced only by the time-varying part.

Step 2. Differentiate twice: $\ddot{\mathbf m}=-\Omega^2 m\sin\alpha[\cos\Omega t\,\hat{\mathbf x}+\sin\Omega t\,\hat{\mathbf y}]$ , so $|\ddot{\mathbf m}|^2=\Omega^4 m^2\sin^2\alpha$ , constant in time. Why: the total loss is steady — the pulses are a beaming effect, not a modulation of the power.

Step 3. Apply the magnetic-dipole Larmor formula $\dot E=\tfrac{2}{3c^3}|\ddot{\mathbf m}|^2=\tfrac{2}{3c^3}\Omega^4 m^2\sin^2\alpha$ . Why: the magnetostatic analogue of the electric Larmor result (Jackson §9.3) — the whole content of the vacuum model.

Step 4. Insert $m=\tfrac12 BR^3$ (polar-field convention): $\dot E=\dfrac{B^2R^6\Omega^4\sin^2\alpha}{6c^3}$ . Why: this ties the moment to the observable field and radius and exhibits the defining $B^2R^6\Omega^4$ scaling.

Step 5. Equate radiated power to rotational-energy loss, $I\Omega\dot\Omega=-\dot E\propto\Omega^4$ , so $\dot\Omega\propto\Omega^3$ — a torque of exponent 3. Why: energy conservation turns the radiation formula into an equation of motion.

Step 6. With $\dot\Omega=-K\Omega^3$ , differentiate and eliminate $K$ : $n=\Omega\ddot\Omega/\dot\Omega^2=3$ . Why: any vacuum-dipole pulsar must show $n=3$ , so a measured $n\ne3$ proves the model incomplete — which is what every young pulsar tells us.

Go deeper · Derivation B · 1 minGoldreich and Julian — why a pulsar cannot live in vacuum

Derivation BGoldreich and Julian — why a pulsar cannot live in vacuum

The goal is to show, from the physics of a rotating conductor, that empty space around a pulsar is unstable — the induced field-parallel electric component is so much stronger than gravity that charge is inevitably pulled off the surface — and to compute the equilibrium space charge, the Goldreich–Julian density, that the corotating magnetosphere must carry. The route runs from the force-free interior condition through the surface potential and the force comparison to Gauss's law.

Step 1. Inside the perfect conductor the Lorentz force on a charge vanishes, giving $\mathbf E_{in}=-(\boldsymbol\Omega\times\mathbf r)\times\mathbf B/c$ . Why: this defines the corotating equilibrium and fixes the interior field in terms of $\Omega$ and $B$ .

Step 2. Integrating from pole to equator gives $\Delta\Phi\sim\Omega B R^2/c$ . For $B=10^{12}$ G, $P=1$ s, $R=10^6$ cm this is $\sim10^{16}$ V, and $E_\parallel\sim(\Omega R/c)B\sim10^{12}\,\mathrm{V\,cm^{-1}}$ . Why: the induced potential dwarfs any binding or work-function energy.

Step 3. Compare forces on a surface electron: $e|E_\parallel|$ exceeds gravity $m_e g_{\mathrm{surf}}$ by $\sim10^9$ , even in the star's $\sim10^{11}g_\oplus$ surface gravity. Why: gravity loses by nine orders of magnitude, so charges are lifted off and no true vacuum persists.

Step 4. The ejected plasma redistributes until it too corotates force-free: $\mathbf E+(\boldsymbol\Omega\times\mathbf r)\times\mathbf B/c=0$ throughout the closed zone. Why: the interior condition now governs the exterior, extending rigid corotation to the light cylinder.

Step 5. Take the divergence: $\rho_{GJ}=\nabla\cdot\mathbf E/4\pi=-\boldsymbol\Omega\cdot\mathbf B/2\pi c$ . Near the pole $\rho_{GJ}<0$ ; the sign reverses across the null surface $\boldsymbol\Omega\cdot\mathbf B=0$ , separating opposite-charge zones and driving the global current. Why: its sign structure seeds the polar-cap and outer-gap accelerators. *The implied number density is $n_{GJ}=|\rho_{GJ}|/e\approx7\times10^{-2}\,B/P\ \mathrm{cm^{-3}}$ ( $B$ in G, $P$ in s).*

Go deeper · Derivation C · 1 minReading the field and the age off the clock

Derivation CReading the field and the age off the clock

The goal is to invert the two timed quantities $P$ and $\dot P$ into the numbers plotted on the $P$ – $\dot P$ diagram — the surface field and the characteristic age — and, along the way, to confront the factor-of-two convention that makes the famous prefactor either $3.2\times10^{19}$ or $6.4\times10^{19}$ . We equate the measured torque to the orthogonal vacuum-dipole loss, solve for $B$ , and integrate the braking law for the age.

Step 1. Set the observed torque equal to the maximal ( $\alpha=90^\circ$ ) vacuum-dipole loss: $I\Omega|\dot\Omega|=B_p^2R^6\Omega^4/6c^3$ , with $B_p$ the polar field. Why: lacking any handle on $\alpha$ , the estimate assumes orthogonal geometry, so $B$ is a proxy, not a fit.

Step 2. Solve: $B_p^2=6c^3 I|\dot\Omega|/(R^6\Omega^3)$ ; with $\Omega=2\pi/P$ , $|\dot\Omega|=2\pi\dot P/P^2$ one has $|\dot\Omega|/\Omega^3=P\dot P/4\pi^2$ , so $B_p^2=\dfrac{3c^3 I}{2\pi^2 R^6}\,P\dot P$ . Why: the field collapses onto $P\dot P$ times constants — why a diagram line is a line of constant field.

Step 3. Insert $I=10^{45}\,\mathrm{g\,cm^2}$ , $R=10^6$ cm, $c=3\times10^{10}\,\mathrm{cm\,s^{-1}}$ : the prefactor $3c^3I/2\pi^2R^6=4.1\times10^{39}$ , so $B_p=6.4\times10^{19}(P\dot P)^{1/2}$ G. Why: the polar-field value in the $m=\tfrac12 B_pR^3$ convention — twice the catalogue value.

Step 4. The tabulated $B=3.2\times10^{19}(P\dot P)^{1/2}$ G is the same physics in the equatorial convention $m=BR^3$ , where $B=B_p/2$ ; both describe the identical star. Why: the factor of two is only a choice of which field one quotes.

Step 5. For the age, integrate $\dot\Omega=-K\Omega^3$ : $t=\tfrac{1}{2K}(\Omega^{-2}-\Omega_0^{-2})$ , and for $\Omega_0\gg\Omega$ this is $t\to P/2\dot P\equiv\tau_c$ . Why: the birth spin drops out only when it was much faster than today's, so $\tau_c$ overshoots the Crab's true age by about 30 percent. *Both prefactors — $6.4\times10^{19}$ (polar) and $3.2\times10^{19}$ (equatorial) — follow from $B^2\propto(3c^3I/2\pi^2R^6)P\dot P$ ; applying $\tau_c=P/2\dot P$ to the Crab gives 1245 yr.*

Limiting cases

The real torque lives between two idealizations — the radiating vacuum dipole and the force-free wind — and which one dominates controls every diagnostic.

Case	Result	Meaning
Aligned vacuum rotator ( $\alpha=0$ )	$\dot E_{\mathrm{vac}}=0$	$\ddot{\mathbf m}=0$ , so vacuum radiation vanishes — yet a plasma-filled aligned rotator still spins down via its wind, the sharpest proof that vacuum intuition misses the dominant channel.
Orthogonal vacuum rotator ( $\alpha=90^\circ$ )	$\dot E=\dfrac{B^2R^6\Omega^4}{6c^3}$	Maximal vacuum-dipole loss; the geometry assumed by the standard field estimate.
Force-free magnetosphere	$\dot E\approx\dfrac{\mu^2\Omega^4}{c^3}(1+\sin^2\alpha)$	Spitkovsky (2006): even the aligned rotator loses $\dot E\approx\mu^2\Omega^4/c^3\ne0$ , with $\mu=BR^3$ ; the wind, not near-zone radiation, carries the torque.
Millisecond pulsar	$R_{LC}=cP/2\pi\sim$ tens of km	The light cylinder shrinks to a few stellar radii, a compact relativistic magnetosphere; recycled by accretion to $B\sim10^{8}$ – $10^{9}$ G.
Braking-index diagnostic	$n\equiv\dfrac{\Omega\ddot\Omega}{\dot\Omega^2}$	Pure dipole gives $n=3$ ; Crab $\approx2.5$ , Vela $\approx1.4$ , B1509−58 $\approx2.8$ all fall short, signalling wind torque, geometric evolution, or field growth.

Common trapThe surface field on the P–Ṗ diagram is a measurement

It is not. The quantity $B=3.2\times10^{19}(P\dot P)^{1/2}$ G stacks three idealizations: a perpendicular ( $\alpha=90^\circ$ ) rotator, a pure vacuum dipole, and the fiducial $I=10^{45}\,\mathrm{g\,cm^2}$ , $R=10^6$ cm. Change the inclination, admit the wind, or vary the equation of state and it shifts by factors of a few — and the polar-versus-equatorial convention alone moves it by two. The same holds for $\tau_c=P/2\dot P$ : it equals the true age only if the star was born spinning much faster than now and $n=3$ held throughout. Both are coordinates that organize the population, not calibrated readings of any one star.

Worked examples

1 · The Crab pulsar, fully worked

The Crab pulsar has $P=33$ ms, $\dot P=4.2\times10^{-13}$ s/s. Find its spin-down luminosity, characteristic age, surface field, light-cylinder radius, and Goldreich–Julian density. Take $I=10^{45}\,\mathrm{g\,cm^2}$ , $R=10^6$ cm.

Step 1. Spin-down luminosity, equation (3): $\dot E=4\pi^2(10^{45})(4.2\times10^{-13})/(0.033)^3=\mathbf{4.6\times10^{38}\ erg\,s^{-1}}$ — matching the Crab Nebula almost exactly, the classic proof that spin-down powers it.

Step 2. Characteristic age, equation (7): $\tau_c=P/2\dot P=0.033/(2\times4.2\times10^{-13})=\mathbf{1245\ yr}$ . SN 1054 makes the true age about 970 yr, so $\tau_c$ overshoots by a third — a birth period $P_0\approx19$ ms not negligible against today's 33 ms.

Step 3. Surface field: the polar value $B_p=6.4\times10^{19}(0.033\times4.2\times10^{-13})^{1/2}=\mathbf{7.5\times10^{12}\ G}$ , and the equatorial value is half, $B=3.8\times10^{12}$ G — strong but ordinary.

Step 4. Light cylinder, equation (1): $R_{LC}=cP/2\pi=(3\times10^{10})(0.033)/2\pi=\mathbf{1.6\times10^8\ cm}=1580\ \mathrm{km}$ , about 160 stellar radii.

Step 5. Goldreich–Julian density $n_{GJ}\approx7\times10^{-2}(7.5\times10^{12})/0.033=\mathbf{1.6\times10^{13}\ cm^{-3}}$ ; the maximum potential, equation (8), is $\Delta\Phi_{pc}\approx6.6\times10^{12}(7.5)/(0.033)^2\approx\mathbf{4.6\times10^{16}\ V}$ — far above the pair-cascade threshold.

2 · A millisecond pulsar for contrast

PSR B1937+21, the first millisecond pulsar, has $P=1.558$ ms, $\dot P=1.05\times10^{-19}$ . Find its field, age, light-cylinder radius, and spin-down power, and compare with the Crab.

Step 1. Field: $B=3.2\times10^{19}(1.558\times10^{-3}\times1.05\times10^{-19})^{1/2}=\mathbf{4.1\times10^{8}\ G}$ — four orders below the Crab, a fossil field buried by accretion.

Step 2. Age: $\tau_c=P/2\dot P=1.558\times10^{-3}/(2\times1.05\times10^{-19})=\mathbf{2.3\times10^{8}\ yr}$ — a thousand times older than the Crab, in the low- $\dot P$ graveyard.

Step 3. Light cylinder: $R_{LC}=cP/2\pi=(3\times10^{10})(1.558\times10^{-3})/2\pi=\mathbf{74\ km}$ — a few stellar radii, a compact magnetosphere.

Step 4. Spin-down power: $\dot E=4\pi^2(10^{45})(1.05\times10^{-19})/(1.558\times10^{-3})^3=\mathbf{1.1\times10^{36}\ erg\,s^{-1}}$ . Despite a field ten thousand times weaker, the $\Omega^4$ leverage keeps the output within a couple hundred of the Crab's — spin, not field alone, sets the budget.

Physics of the portrait

Pulsar magnetospheres are the intellectual property of a single 1969 paper that turned a curiosity into a self-consistent electrodynamic system.

Origins · 1 minPeter Goldreich and how this was discovered

Peter Goldreich

1939– · American theoretical astrophysicist · the pulsar magnetosphere and force-free electrodynamics

Peter Goldreich is one of the most versatile theorists of the last half-century, with fingerprints on planetary rings, the tides of moons, helioseismology, and magnetohydrodynamic turbulence — but among astrophysicists his name is welded first of all to pulsars. When the first pulsar was found in 1967 and identified as a rotating neutron star, the immediate problem was the surrounding fields: could a spinning magnetized star sit in vacuum? In 1969, with his graduate student William H. Julian, Goldreich answered no. Their paper showed the induced surface electric field beats gravity by nine orders of magnitude, so the star must be swaddled in a corotating, charge-separated plasma; they computed its density — $\rho_{GJ}=-\boldsymbol\Omega\cdot\mathbf B/2\pi c$ — and introduced the light cylinder as the boundary of rigid corotation, the concepts on which all later theory is built. The picture was so fertile that its central problems — the global current, the gaps, the transition to a wind — occupied the field for fifty years, settled only numerically in the 2000s. Goldreich spent his career at Caltech and the Institute for Advanced Study, and his honours — the National Medal of Science (1995), the Shaw Prize (2007), the Crafoord Prize (2020) — recognize a rare gift for the one clean calculation at the heart of a messy phenomenon. The pulsar magnetosphere is the archetype: a lighthouse reduced to a divergence and a boundary condition.

History

1934

— Baade and Zwicky propose that neutron stars form in supernova explosions

1967

— Bell and Hewish detect the first pulsar, CP 1919, as a regular radio ticking

1968

— Gold and Pacini identify pulsars as rotating magnetized neutron stars

1969

— Goldreich and Julian derive the charge-separated corotating magnetosphere

2006

— Spitkovsky computes the force-free oblique-rotator spin-down numerically

Going further

Texts. Carroll and Ostlie, An Introduction to Modern Astrophysics, 2nd ed., Ch. 16 · Shapiro and Teukolsky, Black Holes, White Dwarfs and Neutron Stars, Ch. 10 · Lorimer and Kramer, Handbook of Pulsar Astronomy, 2005 · Jackson, Classical Electrodynamics, 3rd ed., §9.3, §11 · Goldreich and Julian 1969, ApJ, 157, 869. Key numerical checks: (1) $\dot E=4\pi^2 I\dot P/P^3\approx4.6\times10^{38}\,\mathrm{erg\,s^{-1}}$ for the Crab; (2) the field prefactor is $6.4\times10^{19}$ (polar) or $3.2\times10^{19}$ (equatorial), both from $B^2=(3c^3I/2\pi^2R^6)P\dot P$ ; (3) $n=\Omega\ddot\Omega/\dot\Omega^2=3$ for $\dot\Omega\propto\Omega^3$ ; (4) $\rho_{GJ}=-\boldsymbol\Omega\cdot\mathbf B/2\pi c$ from $\rho=\nabla\cdot\mathbf E/4\pi$ ; (5) for PSR B1937+21: $B\approx4.1\times10^{8}$ G, $\tau_c\approx2.3\times10^{8}$ yr, $R_{LC}\approx74$ km.

Prerequisites. Neutron Stars · Neutron Star Structure and Interior · Pulsars. Leads to. Millisecond and Recycled Pulsars · Magnetars · Pulsar Glitches and Neutron Star Superfluidity.

Ask the tutor: Ask me to walk through why a measured braking index below 3 rules out the pure vacuum dipole, or to work any pulsar's field, age, and light cylinder from its $P$ and $\dot P$ .

Why must the magnetosphere fill with plasma instead of staying empty?

A spinning conductor with an interior field induces an electric field along the lines just outside its surface — for pulsar parameters, about $10^{12}$ V per centimetre. The force it exerts on a surface electron outweighs the star's enormous gravity by roughly a billion to one, so charges are lifted off and arrange themselves into a corotating, charge-separated cloud at the Goldreich–Julian density $\rho_{GJ}=-\boldsymbol\Omega\cdot\mathbf B/2\pi c$ . A true vacuum is impossible against that ratio, so within one rotation the star swaddles itself in plasma.

If the vacuum dipole predicts n = 3, why does no pulsar actually show it?

The vacuum dipole assumes the star radiates only near-zone dipole waves into empty space, at fixed inclination and field. Real pulsars are wrapped in a plasma-filled, force-free magnetosphere whose relativistic wind carries a large share of the torque with a shallower dependence on $\Omega$ ; on top of that, $\alpha$ can drift, the field geometry can evolve, and the field may grow or decay. Any of these pushes the braking index below three — Crab near 2.5, Vela near 1.4, the very data that force the model beyond vacuum.

What is the difference between the characteristic age and the true age?

The characteristic age $\tau_c=P/2\dot P$ comes from integrating the $n=3$ braking law backward while assuming the star was born spinning infinitely fast — that is, $P_0\ll P$ and $n=3$ for its whole life. When the birth spin was not negligible, as for the Crab (born with $P_0\approx19$ ms against today's 33 ms), $\tau_c$ overestimates the true age, giving 1245 yr against a real age near 970 yr fixed by SN 1054 — a model-dependent estimate, usually an upper bound, not a calibrated clock.

Why does a millisecond pulsar have such a weak field but such a huge spin-down power?

The two facts have different origins. The weak field — $10^8$ – $10^9$ G, ten thousand times below an ordinary pulsar — is a consequence of recycling: accretion from a binary companion spun the star up and buried the surface field. The large power comes from the spin, because $\dot E\propto B^2\Omega^4$ and the fourth power is a ferocious multiplier: a millisecond pulsar spins hundreds of times faster, and $\Omega^4$ more than compensates for the tiny $B^2$ — which is why PSR B1937+21 still radiates within a couple hundred of the Crab's luminosity.

How can only a hundred-thousandth of the power come out as radio and yet dominate what we see?

Energetically, radio is a rounding error: the efficiency is about $10^{-5}$ , and essentially all the spin-down power goes into the relativistic wind and, for young pulsars, the nebula. Observationally, though, the pulses are overwhelming because they are coherent — bunches of pair plasma radiate in phase, so the emission scales with the square of the number of charges, not linearly. That lifts the brightness temperature to $10^{25}$ – $10^{30}$ K, far beyond anything thermal — a feeble channel that is the easiest to detect across the Galaxy.

Does the light cylinder mean plasma actually moves at the speed of light there?

No — nothing material reaches $c$ . The light cylinder is the radius $R_{LC}=c/\Omega$ at which a point rigidly corotating with the star would have to move at $c$ . Inside it, rigid corotation is possible and the field lines close and trap plasma. At and beyond it, rigid corotation is impossible, so the lines cannot close — they open, and plasma slides along them outward instead of being dragged around, forming the wind. It is a boundary in the pattern of corotation, not a place where matter is forced superluminal.