Pulsars

A pulsar is what a neutron star looks like when it spins fast and carries a strong, misaligned magnetic field: a lighthouse in the sky. The magnetic axis is tilted relative to the rotation axis, so the beamed radiation funnelled out along the magnetic poles sweeps a cone across the heavens once per turn. Each time that cone crosses the line of sight, a radio telescope on Earth registers a pulse, and the interval between pulses is nothing more exotic than the star's rotation period $P$ . The object doing the rotating is a Neutron Star — roughly $1.4\,M_\odot$ packed into a $10\,\mathrm{km}$ ball, held up against gravity by neutron degeneracy pressure and the stiff repulsive core of the nuclear force encoded in the Equation of State of Dense Matter. Two observations nailed the interpretation almost immediately after Jocelyn Bell Burnell's 1967 discovery. First, the periods were far too short for anything but a neutron star: a white dwarf spun to a fraction of a second would fly apart, whereas a $10\,\mathrm{km}$ star can turn hundreds of times a second and hold together. Second, the Crab pulsar sat inside a young supernova remnant, tying pulsars to the core collapse of massive stars, and its steady spin-down revealed the power source — the star is a colossal flywheel radiating away its rotational kinetic energy $\tfrac12 I\Omega^2$ . The organizing model is the oblique magnetic rotator. A rotating inclined dipole radiates electromagnetic energy at the spin frequency; equate that luminosity to the rate of rotational-energy loss and the two numbers you can actually measure — the period $P$ and its derivative $\dot P$ — decode the surface magnetic field, the star's age, and its energy budget. This article develops that magnetic-braking machinery; the charge-filled magnetosphere that carries the real torque is owned by Pulsar Magnetospheres and Spin-Down, and the recycled, ultra-magnetized, and glitching sub-populations by their own articles.

Key takeaways

A pulsar is a rapidly rotating, strongly magnetized neutron star whose beamed emission sweeps past the observer once per rotation, producing a train of extraordinarily regular pulses; the pulse period is the spin period $P$ .
Periods run from $\approx 1.4\,\mathrm{ms}$ (recycled millisecond pulsars) to several seconds, and the best millisecond pulsars keep time to rival atomic clocks.
Rotation powers the emission: pulsars spin down at a measured rate $\dot P>0$ , and the inferred surface dipole fields span $\sim 10^{8}$ – $10^{15}\,\mathrm{G}$ .
The magnetic-dipole braking model turns the observable pair $(P,\dot P)$ into the field strength $B_p\approx3.2\times10^{19}\sqrt{P\dot P}\,\mathrm{G}$ , the characteristic age $\tau_c=P/2\dot P$ , and the spin-down luminosity $L_{\text{sd}}=4\pi^2 I\dot P/P^3$ , and predicts a braking index $n=3$ .
Discovered by Jocelyn Bell Burnell in 1967, pulsars confirmed the neutron stars proposed by Baade and Zwicky and became precision laboratories for gravity, dense matter, and the interstellar plasma.
Key idea: a pulsar is a rotating oblique magnetic dipole — its beamed emission traces the rotation period, and the observable pair $(P,\dot P)$ decodes the magnetic field, age, and rotational energy budget through the magnetic braking law.

The core insight

A pulsar is a rotating oblique magnetic dipole: its beamed emission traces the rotation period, and the measurable pair $(P,\dot P)$ decodes the surface magnetic field, the characteristic age, and the rotational energy budget through the magnetic braking law.

Where this lives today

Pulsars have become instruments. Pulsar timing arrays — NANOGrav, the European and Parkes arrays, and their global consortium — monitor dozens of millisecond pulsars for years, treating each as a clock at a fixed point in the Galaxy; a passing nanohertz gravitational wave from merging supermassive black holes would leave a correlated wobble in their arrival times, and in 2023 the arrays reported the first evidence of exactly this stochastic background. The Hulse–Taylor binary already delivered the first, indirect, detection of gravitational waves: its orbit decays at precisely the rate general relativity predicts for quadrupole radiation, to better than a fraction of a percent. Precise neutron-star masses from binary pulsars such as PSR J0740+6620 ( $\approx2.08\,M_\odot$ ) prune the dense-matter equation of state by demanding it stay stiff enough to support such masses. Fast radio bursts — millisecond flashes at cosmological distances — are now known in at least some cases to come from magnetars, tying the extreme tail of the pulsar population to one of the newest puzzles in astrophysics. And FAST, MeerKAT, and the coming Square Kilometre Array are enlarging the known population from a few thousand toward the tens of thousands, sharpening every one of these uses.

The physical picture

Picture a lighthouse. A beam fixed to a rotating lamp appears as a flash each time it swings toward you; you never see the beam itself, only the flash. A pulsar's "beam" is radiation channelled along the open magnetic field lines above a magnetic pole, and because the magnetic axis is tilted by an angle $\alpha$ to the spin axis, that beam traces out a hollow cone in the sky once per rotation. The pulse you record is the geometry, not a physical throb of the star. The first figure fixes the geometry and puts a real number on the light cylinder — the surface, at radius $R_{\text{lc}}=cP/2\pi$ , where rigid co-rotation with the star would demand the speed of light.

Figure

The oblique rotator: the magnetic axis is tilted by α from the spin axis Ω, so the emission cones sweep a circle on the sky once per period P. Field lines that close inside the light cylinder R_lc = cP/2π co-rotate; for the Crab (P = 0.033 s) that radius is R_lc ≈ 1600 km, about 160 stellar radii

The star is a flywheel storing an enormous $\tfrac12 I\Omega^2$ , and the tilted field is the brake: a rotating dipole radiates, dragging on the star, so the flywheel slowly winds down. Measure how fast it slows — the period derivative $\dot P$ — and you measure the strength of the brake, hence the magnetic field. Plotting every known pulsar in the plane of period versus period-derivative produces the single most useful diagram in the field, the $P$ – $\dot P$ diagram, which does for pulsars what the Hertzsprung–Russell diagram does for stars: it sorts the whole population into families and overlays contours of constant field and constant age.

Figure

The P–Ṗ diagram, pulsar astronomy's Hertzsprung–Russell plot. Young energetic pulsars like the Crab (P = 0.033 s, Ṗ = 4.2×10⁻¹³) sit top-left; the bulk of normal pulsars fill the middle; recycled millisecond pulsars occupy the bottom-left; magnetars the top-right. Diagonals of constant field B = 3.2×10¹⁹√(PṖ) and constant age τ_c = P/2Ṗ turn the plane into a map

The core equations

The starting identity is almost too simple to write, but it is the whole reason pulsars are useful: the observed pulse period equals the neutron-star rotation period. Nothing about a radial oscillation, nothing about an orbit — just rotation. Everything downstream is bookkeeping on this fact.

P_{\text{pulse}} = P_{\text{rot}} = \frac{2\pi}{\Omega} \qquad (1)

where: $P_{\text{pulse}}$ — measured interval between pulses; $P_{\text{rot}}$ — neutron-star rotation period; $\Omega$ — angular spin frequency. Because the flywheel is huge and the braking torque tiny, $P$ is one of the most stable quantities in nature — for a millisecond pulsar it drifts by less than a part in $10^{15}$ . The engine is rotational kinetic energy. Differentiate $\tfrac12 I\Omega^2$ and rewrite in the observables $P$ and $\dot P$ to get the spin-down luminosity — the power the star is losing as it slows.

L_{\text{sd}} = -\frac{\mathrm{d}}{\mathrm{d}t}\left(\tfrac{1}{2} I \Omega^2\right) = -I\,\Omega\,\dot\Omega = 4\pi^2 I\,\frac{\dot P}{P^3} \qquad (2)

where: $L_{\text{sd}}$ — spin-down power; $I\approx 10^{45}\,\mathrm{g\,cm^2}$ — canonical neutron-star moment of inertia; $\dot P>0$ — dimensionless period derivative. Reading it: a fast-spinning, quickly-slowing pulsar (small $P$ , large $\dot P$ ) is a powerful engine, which is why the young Crab outshines the entire rest of the pulsar population in $L_{\text{sd}}$ . What removes that energy, in the simplest model, is magnetic-dipole radiation. An inclined dipole moment of magnitude $|\mathbf m|=\tfrac12 B_p R^3$ has a second time derivative $\ddot{\mathbf m}\neq0$ as the star turns, and the Larmor formula gives the radiated power.

L_{\text{dip}} = \frac{2}{3c^3}\,|\ddot{\mathbf{m}}|^2 = \frac{B_p^2 R^6 \Omega^4 \sin^2\alpha}{6 c^3} \qquad (3)

where: $B_p$ — polar surface field; $R$ — stellar radius; $\alpha$ — angle between magnetic and spin axes; $c$ — speed of light. The $\sin^2\alpha$ is the crux: an aligned rotator ( $\alpha=0$ ) has a static moment vector, radiates nothing, and would never spin down. Equating $(2)$ and $(3)$ eliminates the unmeasured luminosity and delivers the workhorse formula of the field — the surface magnetic field inferred straight from timing.

B_p = \left(\frac{3 c^3 I}{2\pi^2 R^6 \sin^2\alpha}\, P\dot P\right)^{1/2} \approx 3.2\times 10^{19}\,\sqrt{P\dot P}\ \mathrm{G} \qquad (4)

where: $P$ in seconds and $\dot P$ dimensionless for the numerical form, which fixes $I=10^{45}\,\mathrm{g\,cm^2}$ , $R=10\,\mathrm{km}$ , $\sin\alpha=1$ and quotes the field at the surface. The scaling $B_p\propto\sqrt{P\dot P}$ is the reason the $P$ – $\dot P$ diagram doubles as a magnetic-field map. The braking law generalizes $(3)$ : write the torque as a single power of the spin frequency and define a dimensionless index measuring that power directly from three timing derivatives.

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

where: $K$ — braking constant set by $B$ , $I$ , and geometry; $n$ — braking index, equal to $3$ for a pure vacuum dipole; $\ddot\Omega$ — measurable for a handful of young pulsars. Observed values (Crab $n=2.51$ ) fall below $3$ , a direct signal that the vacuum-dipole cartoon is incomplete. Integrating the braking law gives the age you can estimate without knowing the birth date, and the boundary of the co-rotating magnetosphere follows from demanding sub-luminal co-rotation.

\tau_c \equiv \frac{P}{2\dot P} = \frac{\Omega}{2|\dot\Omega|}, \qquad R_{\text{lc}} = \frac{c}{\Omega} = \frac{cP}{2\pi} \qquad (6)

where: $\tau_c$ — characteristic (spin-down) age, an upper limit on the true age when $n=3$ and the birth spin was fast; $R_{\text{lc}}$ — light-cylinder radius, the scale of the active magnetosphere and the wind that carries the torque. Finally, the signal itself is a diagnostic of the intervening plasma: a broadband pulse arrives later at low frequency because free electrons disperse it, and the delay measures the electron column.

\mathrm{DM} = \int_0^d n_e\,\mathrm{d}l, \qquad \Delta t = \frac{e^2}{2\pi m_e c}\,\mathrm{DM}\left(\frac{1}{\nu_1^2}-\frac{1}{\nu_2^2}\right) \qquad (7)

where: $\mathrm{DM}$ — dispersion measure, the free-electron column to the pulsar; $n_e$ — electron density along the sightline; $\nu_1,\nu_2$ — two observing frequencies; $\Delta t$ — differential arrival delay. Combined with a Galactic electron model, DM yields distances; the Faraday rotation of the same signal maps the Galactic magnetic field.

Derivation ASpin-down luminosity and the inferred surface magnetic field

The goal is to turn two numbers a radio telescope actually measures — the period $P$ and its slow increase $\dot P$ — into the surface magnetic field of a neutron star whose surface we can never resolve. We do it by insisting on energy conservation: the rotational kinetic energy the star loses as it slows must be carried away by the radiation of its rotating magnetic dipole. Equating the two expressions eliminates the unmeasurable luminosity and leaves the field in terms of $P$ , $\dot P$ , and known constants, which is the single most-used inference in pulsar astronomy.

Step 1. Start from the rotational energy $E_{\text{rot}}=\tfrac12 I\Omega^2$ and differentiate at fixed $I$ , so the spin-down power is $L_{\text{sd}}=-\dot E_{\text{rot}}=-I\Omega\dot\Omega$ . Why: the neutron star is a rigid flywheel; the only reservoir large enough and steady enough to feed the emission for millennia is its rotation, so the loss rate of $\tfrac12 I\Omega^2$ is the power budget.

Step 2. Convert to observables with $\Omega=2\pi/P$ and $\dot\Omega=-2\pi\dot P/P^2$ , giving $L_{\text{sd}}=4\pi^2 I\,\dot P/P^3$ ; the two minus signs cancel because $\dot P>0$ . Why: $\Omega$ and $\dot\Omega$ are not observed directly — the raw data are pulse arrival times, which yield $P$ and $\dot P$ , so the working formula must be written in those.

Step 3. Equate this to the magnetic-dipole luminosity $L_{\text{dip}}=B_p^2 R^6\Omega^4\sin^2\alpha/(6c^3)$ and substitute $\Omega^4=(2\pi/P)^4$ . Why: the vacuum inclined-dipole radiation is the simplest torque that both drains rotational energy and depends on the field, so setting output equal to loss closes the system with $B_p$ as the one unknown.

Step 4. Solve for the field: the powers of $\pi$ and $P$ combine so that only the product $P\dot P$ survives, giving $B_p^2=\dfrac{3c^3 I}{2\pi^2 R^6\sin^2\alpha}\,P\dot P$ , and plugging $I=10^{45}\,\mathrm{g\,cm^2}$ , $R=10\,\mathrm{km}$ , $\sin\alpha=1$ yields $B_p\approx3.2\times10^{19}\sqrt{P\dot P}\,\mathrm{G}$ . Why: the collapse to $P\dot P$ is what makes the field observable at all — you never need $\dot\Omega$ separately, only the product, and the numerical coefficient bakes in the canonical neutron-star structure. *As a check: setting $L_{\text{sd}}=L_{\text{dip}}$ does indeed reduce to $B_p\propto(P\dot P)^{1/2}$ ; with the canonical constants the coefficient is $3.2\times10^{19}\,\mathrm{G}$ , and inserting the Crab values $P=0.0334\,\mathrm{s}$ , $\dot P=4.21\times10^{-13}$ returns $B_p\approx3.8\times10^{12}\,\mathrm{G}$ , the standard catalogue field.*

Go deeper · Derivation B · 2 minThe characteristic age and the braking index

Derivation BThe characteristic age and the braking index

The goal is to estimate how old a pulsar is from timing alone, without any recorded birth, and to build a dimensionless test of the braking model that a real pulsar can pass or fail. Integrating the braking law under the vacuum-dipole assumption produces the characteristic age $\tau_c=P/2\dot P$ , a quantity computable for every pulsar; forming the second frequency derivative gives the braking index $n$ , whose measured departure from the ideal value $3$ tells us precisely how the simple model breaks down.

Step 1. Adopt the braking law $\dot\Omega=-K\Omega^n$ with constant $K$ , and specialize to the vacuum dipole where the torque $\propto\Omega^3$ , so $n=3$ . Why: dividing the dipole luminosity by $\Omega$ gives a torque $I\dot\Omega=-(B_p^2R^6\sin^2\alpha/6c^3)\Omega^3$ , i.e. exactly a single cube of the frequency with a constant coefficient — the defining assumption of the model.

Step 2. Integrate $\dot\Omega=-K\Omega^3$ from the birth frequency $\Omega_0$ to the present $\Omega$ , obtaining the true age $t=\dfrac{1}{2K}\left(\Omega^{-2}-\Omega_0^{-2}\right)$ . Why: the age is the time to spin down from birth to now, and for a power-law torque that integral is elementary; the initial term carries all the dependence on unknown birth conditions.

Step 3. If the star was born spinning much faster than today, $\Omega_0\gg\Omega$ , drop the initial term to get $t\to\dfrac{\Omega}{2|\dot\Omega|}=\dfrac{P}{2\dot P}\equiv\tau_c$ . Why: the birth spin is almost never known, so we discard it — which is exactly why $\tau_c$ is only an upper limit: if the star was born spinning near its present rate, the true age is smaller.

Step 4. Define the braking index $n\equiv\Omega\ddot\Omega/\dot\Omega^2$ and evaluate it for the vacuum dipole: from $\dot\Omega=-K\Omega^3$ we get $\ddot\Omega=3K^2\Omega^5$ , hence $n=3$ exactly. Why: $n$ is measurable from three timing derivatives with no knowledge of $K$ , so it is a clean falsifiable prediction — and the Crab's $n=2.51$ proves real torques (field growth, magnetospheric currents, changing inclination) go beyond the vacuum model. *As a check: $\tau_c=P/2\dot P$ does follow from integrating $\dot\Omega\propto\Omega^3$ with $\Omega_0\to\infty$ ; the index algebra gives $n=3$ identically; and the measured Crab index $n=2.51\pm0.01$ confirms the predicted deviation from the vacuum-dipole value.*

Go deeper · Derivation C · 1 minThe breakup limit — why milliseconds, and why not white dwarfs

Derivation CThe breakup limit — why milliseconds, and why not white dwarfs

The goal is to explain the hard floor on the spin period and, with it, why the 1967 discovery forced the neutron-star interpretation rather than a rotating white dwarf. A self-gravitating body cannot spin so fast that a parcel at its equator is flung off; the limit is set by balancing the required centripetal acceleration against surface gravity. Working this out gives a minimum period near half a millisecond for a neutron star and several seconds for a white dwarf, and the observed sub-second pulses fall cleanly on the neutron-star side.

Step 1. Balance forces at the equator: mass-shedding sets in when the centripetal requirement equals the gravitational acceleration, $\Omega_{\max}^2 R=GM/R^2$ . Why: a surface element stays bound only while gravity can supply the centripetal acceleration for co-rotation; when rotation outpaces gravity the equator lifts off, which is the physical meaning of "spinning apart".

Step 2. Solve for the maximum angular frequency, $\Omega_{\max}=\sqrt{GM/R^3}$ — recognizably the Keplerian orbital frequency just grazing the surface. Why: the breakup spin is simply the fastest orbit the surface material could hold, so the mass-shedding limit and the surface orbital period are the same statement.

Step 3. Convert to a minimum period $P_{\min}=2\pi\sqrt{R^3/GM}$ and insert $M=1.4\,M_\odot$ , $R=10\,\mathrm{km}$ to get $P_{\min}\approx0.5\,\mathrm{ms}$ . Why: this Newtonian estimate — full general relativity lowers it modestly — sets the shortest period nature allows, and the fastest known pulsar, PSR J1748−2446ad at $P=1.396\,\mathrm{ms}$ , sits safely above it.

Step 4. Repeat for a white dwarf, $R\sim10^4\,\mathrm{km}$ , which gives $P_{\min}\sim$ several seconds. Why: because $P_{\min}\propto R^{3/2}$ at fixed mass, a thousandfold-larger radius pushes the floor up by $\sim10^{4.5}$ , so a white dwarf physically cannot produce a sub-second pulse — the observed periods demand an object $\sim10\,\mathrm{km}$ across, a neutron star. *As a check: substituting $G=6.674\times10^{-8}$ cgs, $M=2.8\times10^{33}\,\mathrm{g}$ , $R=10^6\,\mathrm{cm}$ gives $P_{\min}\approx4.6\times10^{-4}\,\mathrm{s}\approx0.5\,\mathrm{ms}$ ; the same formula at $R=10^9\,\mathrm{cm}$ gives $P_{\min}\approx15\,\mathrm{s}$ .*

Limiting cases

The pulsar population spans several decades in both period and field, and a few limiting regimes organize the whole $P$ – $\dot P$ plane.

Case	Result	Meaning
Young, high-field (Crab-like)	$P\sim0.03\,\mathrm{s},\ B\sim10^{12}\,\mathrm{G}$	huge $L_{\text{sd}}$ powers a bright pulsar-wind nebula; short $\tau_c$ ; top-left of the diagram
Aligned rotator	$\alpha\to0\Rightarrow L_{\text{dip}}\to0$	a rotating dipole aligned with the spin axis radiates nothing and would not spin down
Millisecond / recycled	$P\sim1.4$ – $10\,\mathrm{ms},\ B\sim10^{8}$ – $10^{9}\,\mathrm{G}$	spun up by binary accretion; tiny $\dot P$ gives clock-grade timing stability
Magnetar	$P\sim2$ – $12\,\mathrm{s},\ B\sim10^{14}$ – $10^{15}\,\mathrm{G}$	field decay, not rotation, powers X-ray and gamma bursts; large $\dot P$
Death line	$B/P^2\lesssim\mathrm{const}$	polar-cap voltage too low to sustain pair production; radio emission ceases

Common trapThe pulse is a rotation, and τc is not the age

Two traps sit at the heart of the subject. First, a pulsar does not pulsate: the period is the rotation period of a rigidly spinning star, not a radial oscillation and not an orbit. The lighthouse geometry, not any throbbing of the star, makes the flashes. Second, the "characteristic age" $\tau_c=P/2\dot P$ is an order-of-magnitude estimate, not a measurement. It assumes a braking index of exactly $3$ , a magnetic field that never changes, and a birth spin much faster than today's — assumptions that fail routinely. The Crab is the cautionary example: its recorded birth in AD 1054 makes it about $970$ years old, yet $\tau_c\approx1260$ years, an overestimate of nearly a third precisely because the Crab was not born spinning arbitrarily fast.

Worked examples

1 · A normal pulsar: field and age from two numbers (confidence builder)

A pulsar times out at $P=0.5\,\mathrm{s}$ with $\dot P=5\times10^{-15}$ . Estimate its surface magnetic field and characteristic age, and place it in the population.

Step 1. Field from equation $(4)$ : $B_p\approx3.2\times10^{19}\sqrt{P\dot P}=3.2\times10^{19}\sqrt{0.5\times5\times10^{-15}}=3.2\times10^{19}\sqrt{2.5\times10^{-15}}$ . The root is $5.0\times10^{-8}$ , so $B_p\approx\mathbf{1.6\times10^{12}\ G}$ — a textbook normal-pulsar field.

Step 2. Characteristic age from equation $(6)$ : $\tau_c=P/2\dot P=0.5/(2\times5\times10^{-15})=0.5/10^{-14}=5\times10^{13}\,\mathrm{s}$ . Dividing by $3.156\times10^{7}\,\mathrm{s\,yr^{-1}}$ gives $\tau_c\approx\mathbf{1.6\times10^{6}\ yr}$ .

Step 3. Interpretation: a field of $\sim10^{12}\,\mathrm{G}$ and an age of a couple of million years place this squarely in the main island of the $P$ – $\dot P$ diagram, well away from both the recycled millisecond pulsars (weak field, bottom-left) and the magnetars (extreme field, top-right). It is an ordinary middle-aged rotation-powered pulsar. B ≈ 1.6×10¹² G, τc ≈ 1.6×10⁶ yr.

2 · The Crab pulsar, fully diagnosed

The Crab pulsar (PSR B0531+21) has $P=0.0334\,\mathrm{s}$ and $\dot P=4.21\times10^{-13}$ . Take $I=10^{45}\,\mathrm{g\,cm^2}$ , $R=10\,\mathrm{km}$ . Compute its characteristic age, spin-down luminosity, surface field, and light-cylinder radius, and check each against what is independently known.

Step 1. Characteristic age: $\tau_c=P/2\dot P=0.0334/(2\times4.21\times10^{-13})=0.0334/8.42\times10^{-13}\approx3.97\times10^{10}\,\mathrm{s}$ . In years, $\tau_c\approx\mathbf{1260\ yr}$ . The recorded supernova of AD 1054 makes the true age about $970$ years, so $\tau_c$ overestimates by $\sim30\%$ — exactly the sign expected when the birth spin was not $\gg$ the present spin.

Step 2. Spin-down luminosity from equation $(2)$ : $L_{\text{sd}}=4\pi^2 I\dot P/P^3=39.48\times10^{45}\times4.21\times10^{-13}/(0.0334)^3$ . The denominator $(0.0334)^3=3.73\times10^{-5}$ , so $L_{\text{sd}}\approx\mathbf{4.5\times10^{38}\ erg\,s^{-1}}$ . This matches the observed luminosity of the Crab Nebula — the historically decisive proof that the spinning pulsar powers the whole nebula.

Step 3. Surface field from equation $(4)$ : $B_p\approx3.2\times10^{19}\sqrt{0.0334\times4.21\times10^{-13}}=3.2\times10^{19}\sqrt{1.41\times10^{-14}}\approx\mathbf{3.8\times10^{12}\ G}$ . A typical young-pulsar field, a trillion times any laboratory magnet.

Step 4. Light-cylinder radius from equation $(6)$ : $R_{\text{lc}}=cP/2\pi=(3\times10^{10}\times0.0334)/6.283\approx\mathbf{1.6\times10^{8}\ cm}=1600\,\mathrm{km}$ , about $160$ stellar radii — the scale on which the co-rotating magnetosphere hands its torque to the relativistic wind. τc ≈ 1260 yr, L_sd ≈ 4.5×10³⁸ erg s⁻¹, B ≈ 3.8×10¹² G, R_lc ≈ 1600 km.

Physics of the portrait

Pulsars are one of the rare cases where a graduate student, watching a chart recorder for something else entirely, stumbled on a whole new class of object and read the sky correctly against her supervisor's first instinct.

Origins · 1 minJocelyn Bell Burnell and how this was discovered

Jocelyn Bell Burnell

born 1943; discovered pulsars 1967 · radio astronomy of neutron stars

As a doctoral student at Cambridge, Jocelyn Bell (later Bell Burnell) helped build a $4.5$ -acre radio array designed to study the scintillation of quasars, and it was she who then combed through the hundreds of feet of chart paper it produced each day. In late 1967 she noticed a "bit of scruff" — a signal that recurred at the same sidereal time, ruling out terrestrial interference, and that on fast recording resolved into a train of pulses spaced every $1.337$ seconds with astonishing regularity. The period was so precise that the team half-seriously labelled the source LGM-1, for "Little Green Men," before Bell found a second, third, and fourth pulsar in other parts of the sky, killing the alien hypothesis: four independent civilizations would not all be beaming at Cambridge. The clinching interpretation — a rotating, strongly magnetized neutron star, the compact remnant that Baade and Zwicky had proposed in 1934 and that most astronomers had filed as a theorist's curiosity — followed within months from Thomas Gold and Franco Pacini. The 1974 Nobel Prize in Physics went to her supervisor Antony Hewish (with Martin Ryle) and controversially not to Bell Burnell; she has spoken about it with characteristic grace, and in 2018 directed the entire $3\,\mathrm{million}$ of a Breakthrough Prize toward scholarships for under-represented physics students. The object she found now anchors tests of general relativity, maps the Galactic plasma, and hunts for nanohertz gravitational waves.

History

1934

— Baade and Zwicky propose that neutron stars form in supernovae

1967

— Bell Burnell detects the first pulsar, PSR B1919+21, initially dubbed LGM-1

1968

— Gold and Pacini identify pulsars as rotating magnetized neutron stars

1969

— Goldreich and Julian model the charge-filled co-rotating magnetosphere

1974

— Hulse and Taylor discover the binary pulsar PSR B1913+16

1982

— Backer and colleagues find the first millisecond pulsar, PSR B1937+21

Going further

Texts. Carroll and Ostlie, An Introduction to Modern Astrophysics, 2nd ed., ch. 16 (Neutron Stars and Pulsars) · Lorimer and Kramer, Handbook of Pulsar Astronomy (2004) · Shapiro and Teukolsky, Black Holes, White Dwarfs, and Neutron Stars (1983), ch. 10 · Condon and Ransom, Essential Radio Astronomy, ch. 6 · Goldreich and Julian, ApJ 157, 869 (1969). The key numerical benchmarks are: the breakup period $P_{\min}=2\pi\sqrt{R^3/GM}\approx0.5\,\mathrm{ms}$ for $1.4\,M_\odot$ and $10\,\mathrm{km}$ ; for the Crab ( $P=0.0334\,\mathrm{s}$ , $\dot P=4.21\times10^{-13}$ , $I=10^{45}\,\mathrm{g\,cm^2}$ ), $\tau_c\approx1260\,\mathrm{yr}$ , $L_{\text{sd}}\approx4.5\times10^{38}\,\mathrm{erg\,s^{-1}}$ , and $R_{\text{lc}}\approx1600\,\mathrm{km}$ ; the surface-field coefficient $B_p\approx3.2\times10^{19}\sqrt{P\dot P}\,\mathrm{G}$ (with $R=10\,\mathrm{km}$ , $\sin\alpha=1$ ) gives Crab $B_p\approx3.8\times10^{12}\,\mathrm{G}$ ; and the vacuum-dipole braking index $n=3$ follows from $\dot\Omega\propto\Omega^3$ . The 1967 discovery (Bell Burnell and Hewish, PSR B1919+21) and the fastest known pulsar ( $P=1.396\,\mathrm{ms}$ , PSR J1748−2446ad) are historically established.

Prerequisites. Neutron Stars · Equation of State of Dense Matter · Electron Degeneracy Pressure. Leads to. Pulsar Magnetospheres and Spin-Down · Millisecond and Recycled Pulsars · Magnetars · Pulsar Glitches and Neutron-Star Superfluidity.

Ask the tutor: ask me to walk through the Crab's full $(P,\dot P)$ diagnosis, or to show why an aligned rotator refuses to spin down.

Is a pulsar really pulsating, or is it rotating?

It is rotating, not pulsating. The word "pulsar" is a historical accident — the source produces regular pulses, and the name stuck — but the pulses come from a beam swinging past you like a lighthouse, not from the star swelling and shrinking. This distinction was decisive at the discovery: a genuine radial pulsation of a compact star would have a period tied to its mean density, and no known star could pulsate as fast as one second, let alone a millisecond, without a density as high as a neutron star's. Rotation, by contrast, can be arbitrarily fast up to the breakup limit, so the rotating-neutron-star model fit the data where a pulsating white dwarf could not.

How can a magnetic field be read off from just two timing numbers?

Through an energy argument, not a direct measurement. You never see the field; you see the period $P$ and its slow growth $\dot P$ . The model assumes the rotational energy the star is losing is radiated by its spinning magnetic dipole, and the dipole's power depends on the field strength. Equating loss to radiation lets you solve for the field, and the algebra conveniently collapses everything into the single product $P\dot P$ , giving $B_p\approx3.2\times10^{19}\sqrt{P\dot P}\,\mathrm{G}$ . The number is only as good as the model — it assumes a vacuum dipole, a canonical moment of inertia, and $\sin\alpha=1$ — so it is best read as a characteristic field, accurate to a factor of a few, not a precision value.

Why do measured braking indices come out below the predicted 3?

The value $n=3$ is the prediction of a rotating dipole radiating into a vacuum, with a fixed field and a torque that scales exactly as $\Omega^3$ . Real pulsars violate every one of those idealizations a little. The magnetosphere is filled with charged plasma that carries its own currents and torques, not empty vacuum; the effective magnetic moment may grow over time; the inclination angle between the axes can evolve; and a relativistic particle wind adds a torque with a different frequency dependence. Any of these makes the effective torque scale as a power slightly softer than $\Omega^3$ , pulling the measured index below $3$ — the Crab's $n=2.51$ is the classic example, and the deviation is itself a probe of the magnetosphere.

What sets the shortest possible pulsar period?

The breakup, or mass-shedding, limit. If a star spins fast enough that a piece of its equator would need more centripetal acceleration than gravity can supply, that material lifts off and the star cannot spin faster. Setting the two accelerations equal gives $P_{\min}=2\pi\sqrt{R^3/GM}$ , which for a canonical neutron star is about half a millisecond. The fastest pulsar actually known, PSR J1748−2446ad, spins at $1.396\,\mathrm{ms}$ — comfortably above the floor — and the observed cutoff of the millisecond-pulsar population near a few milliseconds is thought to be set not by breakup but by the balance between accretion spin-up and magnetic spin-down during the recycling phase.

What is the light cylinder and why does it matter?

The light cylinder is the imaginary cylinder of radius $R_{\text{lc}}=cP/2\pi$ centred on the spin axis, where a point rigidly co-rotating with the star would have to move at the speed of light. Nothing can co-rotate beyond it. Magnetic field lines that close inside the light cylinder co-rotate with the star; field lines that would cross it cannot close and are dragged open, and it is along these open lines that the relativistic particle wind streams out, carrying away angular momentum and forming the beamed emission near the poles. So the light cylinder sets the size of the "active" magnetosphere and separates the region that spins with the star from the region that launches its wind.

How do pulsars help detect gravitational waves?

In two very different ways. The Hulse–Taylor binary pulsar is itself a source: two neutron stars orbiting each other radiate gravitational waves, losing orbital energy so their orbit shrinks at a rate matching general relativity to a fraction of a percent — the first, indirect, proof that gravitational waves exist. Separately, millisecond pulsars are used as detectors: a pulsar timing array watches dozens of them for years, and a passing nanohertz gravitational wave from distant supermassive black-hole binaries stretches and squeezes the space between us and the pulsars, leaving a distinctive correlated pattern in their pulse arrival times. In 2023 several arrays reported evidence for exactly such a background.

What is the dispersion measure good for besides being a nuisance?

When a broadband pulse travels through the ionized interstellar medium, the free electrons slow the low frequencies more than the high ones, so the pulse arrives later at low frequency by an amount proportional to the electron column, the dispersion measure. You must correct for it to see a sharp pulse at all — hence "nuisance" — but the same delay is a free probe of the Galaxy. Combined with a model of the Galactic electron density, DM gives a rough distance to the pulsar; the DM of a fast radio burst measures the electron column all the way across intergalactic space and has been used to locate the Universe's "missing baryons"; and the Faraday rotation of the polarized signal, alongside DM, maps the Galactic magnetic field.