Type I X-ray Bursts

A Type I X-ray burst is a thermonuclear bomb, a few metres thick, going off on the surface of a neutron star. In a low-mass X-ray binary the compact star pulls a steady stream of hydrogen and helium off its companion; the gas spirals in, spreads across the surface, and is crushed under a gravity of order $g\sim10^{14}\,\mathrm{cm\,s^{-2}}$ . Buried beneath fresh fuel, the layer becomes electron-degenerate, and the trap is sprung: degenerate pressure barely depends on temperature, so when nuclear reactions release heat the layer cannot expand and cool the way an ordinary gas would. The temperature climbs, the fiercely temperature-sensitive reactions accelerate, and within about a second the shell detonates — a runaway that briefly outshines the steady accretion glow tenfold. It is the same instability that drives the helium flash in red giants and the nova on a white dwarf, here at the most extreme surface gravity in the ordinary universe. The observable is a sharp X-ray flash rising in $\sim1\,\mathrm{s}$ and decaying over $\sim10$ – $100\,\mathrm{s}$ , its spectrum a cooling blackbody softening from $kT\sim2$ – $3\,\mathrm{keV}$ . The energetics are dominated by gravity, not by the nuclear physics that lights the match: accreting a nucleon liberates $\sim200\,\mathrm{MeV}$ while burning it yields only a few MeV, so the persistent emission between bursts carries $\alpha\sim40$ – $100$ times the burst fluence. The brightest bursts reach the Eddington limit, lift their own photosphere, and become standard candles for weighing neutron stars and probing the dense-matter equation of state.

Key takeaways

Type I X-ray bursts are thermonuclear flashes on an accreting neutron star, triggered when a degenerate layer of accumulated hydrogen and helium ignites unstably.
They show a factor-of-ten X-ray spike rising in $\sim1\,\mathrm{s}$ and decaying over $\sim10$ – $100\,\mathrm{s}$ ; recurrence runs from hours to days.
Ignition is unstable because degenerate electron pressure is nearly temperature-independent: nuclear heating raises $T$ without the expansion and cooling that stabilises an ideal gas.
The burst-to-persistent ratio $\alpha\equiv\int L_{\mathrm{persist}}\,dt/E_{\mathrm{burst}}\sim40$ – $100$ equals $E_{\mathrm{grav}}/E_{\mathrm{nuc}}\approx200\,\mathrm{MeV}/(\text{few MeV})$ per nucleon.
Photospheric-radius-expansion (PRE) bursts reach $L_{\mathrm{Edd}}\approx(3.5$ – $3.8)\times10^{38}\,\mathrm{erg\,s^{-1}}$ for helium and serve as standard candles for distance and mass–radius.
The nuclear engine runs through the triple-alpha reaction, the hot-CNO cycle, and the rapid-proton-capture (rp) process, building proton-rich nuclei up to $A\sim100$ .
Key idea: Because degenerate electron pressure is nearly temperature-independent, nuclear heating of the accreted shell raises its temperature without the expansion and cooling that stabilises an ideal gas — so ignition runs away thermally, producing a Type I X-ray burst.

Where this lives today

Type I bursts are now a working tool of neutron-star physics. NASA's NICER instrument times burst photons to the microsecond, sharpening PRE mass–radius measurements that feed the dense-matter equation of state — the same physics tested by neutron-star mergers such as GW170817. Burst oscillations at a few hundred hertz, first found by Strohmayer, read out the spin frequencies of accreting neutron stars and tie the bursters to the recycled millisecond-pulsar population. The rp-process is a target for radioactive-beam facilities like FRIB, and the rare superbursts — hours-long carbon flashes releasing $10^{42}\,\mathrm{erg}$ — probe the conductivity and heating of the crust. Large catalogues (RXTE, and now the MINBAR compilation of thousands of events) let population studies test the ignition theory against statistics.

The physical picture

Picture accreted gas settling onto the neutron star like snow, each fall burying and compressing what lies beneath. A few metres down the electrons are degenerate: their pressure comes from the Pauli exclusion principle, not thermal motion, so the buried gas behaves like a rigid solid that will not expand when heated. That single fact is the whole story. Once burning starts the heat cannot lift and cool the overlying column, so it raises the temperature, which accelerates the reactions, which release more heat. The layer runs away in about a second, and the star flashes.

Figure

Cross-section of the accreting envelope on a neutron star: fuel piles up under g ≈ 2.4×10¹⁴ cm s⁻² until a degenerate ignition layer at column y ≈ 10⁸ g cm⁻² and T ≈ 3×10⁸ K runs away, burning through the triple-alpha and rp-process into ashes with mass number up to A ≈ 100

The energy budget is the second surprise, and it points the other way. Falling onto a neutron star is far more violent than burning: a nucleon reaching the surface gives up $\sim200\,\mathrm{MeV}$ , whereas fusing it releases at most a few MeV. So the steady accretion light carries tens to a hundred times more energy than the flashes that punctuate it — the ratio $\alpha$ read straight off the light curve. The runaway is finally self-limiting: once the layer heats past $T\sim10^9\,\mathrm{K}$ degeneracy lifts, thermal pressure takes over, the shell expands and cools, and burning shuts off — seen as the cooling blackbody tail.

Figure

A Type I burst light curve: after ~1 s the luminosity spikes from the persistent level to the Eddington value L ≈ 3.8×10³⁸ erg s⁻¹, drives photospheric radius expansion at the peak, then decays as a cooling blackbody over ~10–100 s with kT falling from ~2–3 keV — the persistent-to-burst ratio is α ≈ 40–100

The core equations

Everything begins with the surface gravity, which sets the pressure at any column depth and hence where fuel ignites. The field is strong enough that GR matters at the $30\%$ level, so we keep the redshift factor.

g = \frac{GM}{R^2}\left(1-\frac{2GM}{Rc^2}\right)^{-1/2} \qquad (1)

where: $g$ — proper surface gravitational acceleration; $M,R$ — neutron-star mass and radius; $G,c$ — gravitational constant and speed of light. For canonical $M=1.4\,M_\odot$ , $R=10\,\mathrm{km}$ the Newtonian part is $1.9\times10^{14}\,\mathrm{cm\,s^{-2}}$ and the GR factor $\approx1.30$ lifts it to $g\approx2.4\times10^{14}\,\mathrm{cm\,s^{-2}}$ . The energy released when a nucleon falls to that surface dominates the problem, and is purely gravitational.

E_{\mathrm{grav}} = \frac{GMm_u}{R} \approx 193\ \mathrm{MeV\ per\ nucleon} \qquad (2)

where: $m_u$ — atomic mass unit; $E_{\mathrm{grav}}$ — binding energy released per accreted nucleon (about $0.2\,m_u c^2$ ; the relativistic value is $\approx200\,\mathrm{MeV}$ ), which powers the persistent emission. The nuclear energy powering the burst is far smaller and depends on what burns — hydrogen releases several times more per nucleon than helium, already tightly bound.

E_{\mathrm{nuc}} \approx 6.0\ \mathrm{MeV/nucleon\ (H\!\to He)},\qquad 1.6\ \mathrm{MeV/nucleon\ (He\!\to Fe\text{-}group)} \qquad (3)

where: $E_{\mathrm{nuc}}$ — nuclear energy liberated per nucleon. Hydrogen fusion gives $\approx26.7\,\mathrm{MeV}$ per helium nucleus ( $\approx6\,\mathrm{MeV/nucleon}$ after neutrino losses); helium burning to iron adds only the $\approx1.7\,\mathrm{MeV/nucleon}$ separating $^4$ He from the iron group. Composite fuel of hydrogen fraction $X$ yields $E_{\mathrm{nuc}}\approx1.6+4.0\,X$ . The ratio of these two energies, measured directly from the light curve as $\alpha$ , is the cleanest sign that gravity and nuclear physics compete on the same fuel.

\alpha \equiv \frac{\int L_{\mathrm{persist}}\,dt}{E_{\mathrm{burst}}} \approx \frac{E_{\mathrm{grav}}}{E_{\mathrm{nuc}}} \sim 40\text{–}100 \qquad (4)

where: $L_{\mathrm{persist}}$ — persistent accretion luminosity; $E_{\mathrm{burst}}$ — time-integrated burst fluence; $\alpha$ — their ratio. Helium-rich fuel ( $E_{\mathrm{nuc}}\approx1.6$ ) gives $\alpha\approx120$ ; hydrogen-rich ( $E_{\mathrm{nuc}}\approx5$ – $6$ ) gives $\alpha\approx35$ – $40$ ; the observed spread maps the fuel composition. Why ignition is unstable rather than steady lives in the equation of state: in the degenerate layer the pressure is set by electron momentum through the Pauli principle, not temperature.

P_{\mathrm{deg}} = K\left(\frac{\rho}{\mu_e m_u}\right)^{5/3},\qquad \left(\frac{\partial\ln P}{\partial\ln T}\right)_{\!\rho}\approx 0,\qquad K=\frac{(3\pi^2)^{2/3}\hbar^2}{5m_e} \qquad (5)

where: $P_{\mathrm{deg}}$ — non-relativistic degenerate electron pressure; $\rho$ — density; $\mu_e$ — mean molecular weight per electron; $K$ — degeneracy constant. The bracketed near-zero derivative is the crux: heating at fixed pressure barely changes the density, so the layer cannot expand and cool. The other half of the instability is that burning is ferociously temperature-sensitive: the triple-alpha reaction proceeds through the Hoyle resonance, carrying an activation energy in its exponent.

\varepsilon_{3\alpha}\propto \rho^2\,T_9^{-3}\,\exp\!\left(-\frac{4.40}{T_9}\right)\ \Rightarrow\ \nu\equiv\frac{\partial\ln\varepsilon_{3\alpha}}{\partial\ln T}=\frac{4.40}{T_9}-3\approx 40\ \ (T_9=0.1) \qquad (6)

where: $\varepsilon_{3\alpha}$ — triple-alpha energy-generation rate; $T_9\equiv T/10^9\,\mathrm{K}$ ; $\nu$ — logarithmic temperature sensitivity. The constant $4.40$ is $E_a/k_B$ in units of $10^9\,\mathrm{K}$ , with $E_a\approx0.38\,\mathrm{MeV}$ the Hoyle-resonance energy above threshold. Near ignition $\nu\approx40$ , falling to $\sim20$ by $T_9\approx0.2$ . At its peak a burst can reach the Eddington limit, where radiation pressure on the electrons balances gravity and the photosphere lifts off — a nearly universal luminosity, so PRE bursts are standard candles.

L_{\mathrm{Edd}} = \frac{4\pi GMc}{\kappa_{\mathrm{es}}}\left(1-\frac{2GM}{Rc^2}\right)^{-1/2},\qquad \kappa_{\mathrm{es}}=0.2\,(1+X)\ \mathrm{cm^2\,g^{-1}} \qquad (7)

where: $L_{\mathrm{Edd}}$ — surface Eddington luminosity; $\kappa_{\mathrm{es}}=\sigma_T(1+X)/2m_u$ — electron-scattering opacity; $X$ — hydrogen mass fraction. For $M=1.4\,M_\odot$ the Newtonian value is $\approx3.5\times10^{38}\,\mathrm{erg\,s^{-1}}$ (helium, $X=0$ ) and $2.1\times10^{38}$ (solar); GR and redshift roughly offset, so measured touchdown luminosities cluster near $(3.5$ – $3.8)\times10^{38}\,\mathrm{erg\,s^{-1}}$ for helium. Finally, the burst cadence is set by how long fuel takes to reach the ignition column, while the cooling tail is a shrinking blackbody whose apparent area gives the radius.

y_{\mathrm{ign}}=\frac{P_{\mathrm{ign}}}{g}\sim10^{8}\ \mathrm{g\,cm^{-2}}\ (\mathrm{He}),\quad t_{\mathrm{rec}}=\frac{y_{\mathrm{ign}}}{\dot m}\,(1+z),\quad F_{\mathrm{obs}}=\frac{R_\infty^2\,\sigma_{\mathrm{SB}}\,T_{\mathrm{bb}}^4}{d^2} \qquad (8)

where: $y_{\mathrm{ign}}$ — ignition column depth (pressure over gravity); $\dot m$ — local accretion rate per unit area; $t_{\mathrm{rec}}$ — recurrence time; $(1+z)$ — redshift; $R_\infty$ — apparent (redshifted) blackbody radius; $T_{\mathrm{bb}}$ — colour temperature; $d$ — distance. Together these turn a light curve into a measurement of composition and distance.

Derivation AWhy degenerate ignition runs away

The goal is to prove, from thermodynamics alone, that a thin burning shell is stable when it can expand and cool but unstable when degeneracy pins its density — reducing the whole question to comparing two logarithmic derivatives, one for how fast burning grows with temperature and one for cooling. We hold the pressure fixed, perturb the temperature, and ask whether the perturbation grows; the answer separates an ideal gas, which self-corrects, from a degenerate gas, which detonates.

Step 1. Perturb the temperature upward by $\delta T$ at fixed pressure $P$ . The density response is governed by $(\partial\ln\rho/\partial\ln T)_P$ , so the outcome hinges on the equation of state. Why: at fixed pressure it is the density change that decides whether the layer expands and sheds heat or sits still and accumulates it.

Step 2. For an ideal gas, $P=\rho k_B T/\mu m_u$ gives $\rho\propto T^{-1}$ : a temperature rise drops the density and the layer expands. With $\varepsilon\propto\rho^2 T^\nu$ , the falling density claws back part of the heating and radiative losses finish the job. Why: expansion is the thermostat, which is why ordinary stellar burning is stable.

Step 3. For a degenerate gas, equation (5) makes $P=P_{\mathrm{deg}}(\rho)$ independent of $T$ , so at fixed pressure $\rho$ barely changes. The thermostat is disconnected: no expansion, and every joule goes straight into raising $T$ . Why: degeneracy breaks the pressure–temperature link an ideal gas relies on to regulate itself.

Step 4. Compare growth rates. Burning grows with $\nu=\partial\ln\varepsilon/\partial\ln T\approx40$ for triple-alpha, while radiative cooling grows only as $\partial\ln L_{\mathrm{cool}}/\partial\ln T\approx4$ . The instability condition $\nu>\partial\ln L_{\mathrm{cool}}/\partial\ln T$ is met overwhelmingly. Why: the runaway is not marginal — burning beats cooling by an order of magnitude, so once lit the layer cannot stabilise.

Step 5. The runaway self-terminates. As $T$ climbs toward $10^9\,\mathrm{K}$ the thermal energy overtakes the Fermi energy, $k_BT\gtrsim E_F$ ; degeneracy lifts, thermal pressure surges, the layer expands and cools, and burning ceases. Why: the equation of state that forbade expansion now permits it, ending the burst. *The instability criterion $\nu\approx40\gg4$ is met overwhelmingly, and ignition occurs near $T\sim(2$ – $3)\times10^8\,\mathrm{K}$ .*

Go deeper · Derivation B · 1 minThe Eddington limit and photospheric radius expansion

Derivation BThe Eddington limit and photospheric radius expansion

The aim is to derive the luminosity at which radiation pressure exactly balances gravity on the accreted plasma, show why it depends only on the neutron-star mass and hydrogen fraction (and so is nearly universal), and explain how a burst that reaches it lifts its own photosphere to become a standard candle. We balance the radiative force on electrons against gravity on the electron–ion fluid, evaluate it, and follow what happens when the burst tries to exceed it.

Step 1. Radiation of luminosity $L$ at radius $r$ pushes each electron through Thomson scattering with force $L\sigma_T/4\pi r^2 c$ ; the electrons drag the ions along, so the force is shared over the plasma mass per electron. Setting this force per unit mass equal to gravity $GM/r^2$ defines $L_{\mathrm{Edd}}$ , and the $r^{-2}$ cancels. Why: because both forces fall as $1/r^2$ , a single luminosity marks the tipping point everywhere at once.

Step 2. The plasma mass per electron is $2m_u/(1+X)$ for a fully ionised H–He mix, giving $\kappa_{\mathrm{es}}=0.2(1+X)\,\mathrm{cm^2\,g^{-1}}$ and equation (7). For $1.4\,M_\odot$ this yields $L_{\mathrm{Edd}}\approx3.5\times10^{38}\,\mathrm{erg\,s^{-1}}$ (helium), $2.1\times10^{38}$ (solar). Why: the result depends only on $M$ and $X$ , which makes it a calibratable standard candle.

Step 3. When $L$ reaches $L_{\mathrm{Edd}}$ , radiation cannot push harder without unbinding the gas, so the surplus inflates the atmosphere. The photosphere expands to $R_{\mathrm{ph}}>R$ ; with $L\approx L_{\mathrm{Edd}}$ pinned, $L=4\pi R_{\mathrm{ph}}^2\sigma_{\mathrm{SB}}T_{\mathrm{bb}}^4$ forces $T_{\mathrm{bb}}$ down as $R_{\mathrm{ph}}$ rises. Why: radius expansion is the fingerprint of a burst hitting the Eddington ceiling — colour temperature dips while luminosity holds flat.

Step 4. As fuel exhausts, the atmosphere recontracts through "touchdown" ( $R_{\mathrm{ph}}\to R$ ), where the flux equals $L_{\mathrm{Edd}}/4\pi d^2$ . Combining this with the cooling-tail apparent area (8), and folding in $f_c$ and redshift, constrains $M$ , $R$ and $d$ . Why: one PRE burst supplies two independent observables — a luminosity anchored to $M$ and an area anchored to $R$ . *The helium Eddington luminosity $L_{\mathrm{Edd}}(\mathrm{He})\approx3.5\times10^{38}\,\mathrm{erg\,s^{-1}}$ , and observed peak PRE luminosities cluster near this value.*

Go deeper · Derivation C · 1 minRecurrence time from the ignition column

Derivation CRecurrence time from the ignition column

The purpose is to turn "fuel piles up until it ignites" into a number for how often bursts repeat, by identifying the critical column depth at which the base becomes thermally unstable and dividing by the rate at which accretion delivers fuel. We build the column in time, set the ignition condition, estimate the recurrence, and apply the redshift separating the star's clock from ours.

Step 1. Fuel arrives at local rate $\dot m$ and accumulates, so the column grows linearly, $y(t)=\dot m\,t$ . The depth variable $y=\int\rho\,dz\approx P/g$ fixes the local pressure regardless of the density profile. Why: pressure, not geometric depth, controls when the base ignites, and $y=P/g$ converts one to the other through $g$ .

Step 2. Ignition occurs when $y$ reaches the critical column at which nuclear heating at the base outruns conductive and radiative losses. For helium $y_{\mathrm{ign}}\sim10^8\,\mathrm{g\,cm^{-2}}$ ; for mixed fuel $\sim10^{7}$ – $10^{8}$ , set by composition and the crustal heat flux. Why: the ignition column, not any fixed mass, sets the trigger.

Step 3. The recurrence time is $t_{\mathrm{rec}}=y_{\mathrm{ign}}/\dot m$ . For $\dot m\sim10^4\,\mathrm{g\,cm^{-2}\,s^{-1}}$ — a few percent of Eddington — this gives $t_{\mathrm{rec}}\sim10^4\,\mathrm{s}\approx2.8\,\mathrm{hours}$ , in the observed range. Why: dividing two measured quantities reproduces the cadence, confirming that accumulation paces the system.

Step 4. Convert the star's clock to ours by the redshift $(1+z)\approx1.30$ : $t_{\mathrm{rec}}^{\mathrm{obs}}=t_{\mathrm{rec}}(1+z)\approx3.6\,\mathrm{hours}$ . Why: the surface runs $30\%$ slow to us, and precise "clocked bursters" only make sense with that factor. *The local recurrence time $t_{\mathrm{rec}}\approx2.8\,\mathrm{h}$ becomes $3.6\,\mathrm{h}$ as observed, matching the clocked burster GS 1826 $-$ 24.*

Limiting cases

The character of a burst — hydrogen-triggered, pure-helium, or absent — is set almost entirely by the accretion rate $\dot M$ in units of $\dot M_{\mathrm{Edd}}$ , with composition following from it.

Case	Result	Meaning
Low $\dot M$ (mixed H/He)	$\dot M\lesssim0.01\,\dot M_{\mathrm{Edd}}$	Hydrogen ignites unstably via hot-CNO; short, weak bursts with H and He both present.
Intermediate $\dot M$ (pure-He)	$0.01\lesssim\dot M/\dot M_{\mathrm{Edd}}\lesssim0.1$	Stable hot-CNO burning exhausts hydrogen first; a pure-helium triple-alpha flash gives the brightest bursts, frequently with PRE.
High $\dot M$ (stable burning)	$\dot M\gtrsim0.1\,\dot M_{\mathrm{Edd}}$	Burning turns thermally stable; bursts weaken and cease, giving way to steady burning.
Superbursts (carbon)	$E_{\mathrm{burst}}\sim10^{42}\,\mathrm{erg}$ , $t_{\mathrm{rec}}\sim\mathrm{yr}$	Rare, hours-long events from unstable carbon ignition in a deep layer of rp-process ashes.

Common trapThe brightest bursts must come from the fastest accretion

It is natural to assume that piling on fuel faster gives bigger bangs, but the opposite is true near the top of the range. As $\dot M$ climbs past $\sim0.1\,\dot M_{\mathrm{Edd}}$ the base is kept hot by compression and the flux from below, so hot-CNO hydrogen burning becomes stable — it simmers steadily instead of accumulating, the instability criterion fails, and the bursts weaken and stop. The most energetic bursts, including the PRE ones, come from the intermediate regime, where hydrogen burns stably but helium accumulates until it flashes as a nearly pure triple-alpha detonation. Two further traps: Type I (thermonuclear) is not Type II (accretion-instability), and the burst blackbody temperature is not the true effective temperature until divided by a colour-correction $f_c\approx1.4$ – $1.9$ .

Worked examples

1 · Surface gravity, accretion energy, and the α ratio (fully worked)

Take the canonical neutron star, $M=1.4\,M_\odot=2.78\times10^{33}\,\mathrm{g}$ , $R=10\,\mathrm{km}=10^6\,\mathrm{cm}$ . Find its surface gravity, the energy released per accreted nucleon, and the expected $\alpha$ for a pure-helium burst.

Step 1. Newtonian surface gravity: $GM/R^2=(6.67\times10^{-8})(2.78\times10^{33})/(10^6)^2=1.86\times10^{14}\,\mathrm{cm\,s^{-2}}$ . With compactness $2GM/Rc^2=0.41$ , the GR factor $(1-0.41)^{-1/2}=1.30$ gives $g=\mathbf{2.4\times10^{14}\ cm\,s^{-2}}$ .

Step 2. Accretion energy per nucleon: $E_{\mathrm{grav}}=GMm_u/R=3.09\times10^{-4}\,\mathrm{erg}$ . With $1\,\mathrm{erg}=6.24\times10^5\,\mathrm{MeV}$ this is $E_{\mathrm{grav}}=\mathbf{193\ MeV/nucleon}$ , about $0.2\,m_u c^2$ .

Step 3. For a pure-helium burst $E_{\mathrm{nuc}}\approx1.6\,\mathrm{MeV/nucleon}$ , so $\alpha=E_{\mathrm{grav}}/E_{\mathrm{nuc}}=193/1.6=\mathbf{121}$ ; a hydrogen-rich burst with $E_{\mathrm{nuc}}\approx5$ gives $\alpha\approx39$ . The observed range $\alpha\approx40$ – $100$ is exactly the span from hydrogen-rich to helium-rich fuel — a readout of what burned.

2 · Recurrence time of a clocked burster

A source accretes at $\dot m=1.0\times10^4\,\mathrm{g\,cm^{-2}\,s^{-1}}$ (a few percent of Eddington) and ignites helium at $y_{\mathrm{ign}}=1.0\times10^8\,\mathrm{g\,cm^{-2}}$ . Predict the observed recurrence time.

Step 1. Local recurrence time: $t_{\mathrm{rec}}=y_{\mathrm{ign}}/\dot m=(1.0\times10^8)/(1.0\times10^4)=1.0\times10^4\,\mathrm{s}=\mathbf{2.8\ hours}$ .

Step 2. Convert to the observer's clock with $(1+z)=1.30$ : $t_{\mathrm{rec}}^{\mathrm{obs}}=1.3\times10^4\,\mathrm{s}=\mathbf{3.6\ hours}$ .

Step 3. This is precisely GS 1826 $-$ 24, the "clocked burster", whose bursts recur every few hours with metronome regularity — the clearest sign that recurrence is set by fuel accumulation, not by anything stochastic. A steadier $\dot m$ makes a steadier clock.

3 · Distance from a PRE touchdown flux

A helium PRE burst reaches a touchdown flux $F_{\mathrm{Edd}}=3.0\times10^{-8}\,\mathrm{erg\,cm^{-2}\,s^{-1}}$ . Taking the standard-candle $L_{\mathrm{Edd}}=3.8\times10^{38}\,\mathrm{erg\,s^{-1}}$ , estimate the distance.

Step 1. From $F_{\mathrm{Edd}}=L_{\mathrm{Edd}}/4\pi d^2$ : $d^2=(3.8\times10^{38})/(3.77\times10^{-7})=1.01\times10^{45}\,\mathrm{cm^2}$ , so $d=3.18\times10^{22}\,\mathrm{cm}$ .

Step 2. With $1\,\mathrm{pc}=3.086\times10^{18}\,\mathrm{cm}$ this is $\mathbf{10.3\ kpc}$ — typical for a galactic-bulge or globular-cluster LMXB. The near-universal $L_{\mathrm{Edd}}$ turns a measured flux into a distance.

Physics of the portrait

Type I bursts arrived as a genuine surprise: nobody predicted them, they were caught by satellites built for other purposes, and a race of theory papers in a single year explained them.

Origins · 1 mintype i x-ray bursts and how this was discovered

type i x-ray bursts

thermonuclear flashes on accreting neutron stars, discovered 1975–1976 · dense-matter and accretion astrophysics

The bursts were discovered essentially by accident. In 1975 Grindlay and Heise, analysing data from the Astronomical Netherlands Satellite, saw enormous X-ray flashes from the source 4U 1820 $-$ 30 in the globular cluster NGC 6624; almost simultaneously Belian, Conner and Evans recognised similar flashes in Vela-satellite records. Within months the interpretation was in place: Woosley and Taam (1976), and independently Maraschi and Cavaliere (1977), identified them as unstable thermonuclear shell burning on an accreting neutron star, building on Hansen and Van Horn's 1975 thin-shell analysis. Fujimoto, Hanawa and Miyaji (1981) mapped the ignition regimes against accretion rate, and Wallace and Woosley, then Schatz and collaborators, worked out the rp-process reaching the tin region. What began as an unexplained spike became, within a decade, a precision probe of matter at nuclear density.

History

1975

— the ANS satellite records the first recognised X-ray bursts from 4U 1820−30

1976

— Grindlay and colleagues, and independently the Vela team, announce the new bursters

1976

— Woosley and Taam, and Maraschi and Cavaliere, identify them as unstable thermonuclear shell burning

1981

— Fujimoto, Hanawa and Miyaji map the ignition regimes against accretion rate

1996

— Strohmayer discovers burst oscillations, reading off the neutron-star spin

1998

— Schatz and collaborators trace the rp-process up to the tin region

Going further

Texts. Carroll and Ostlie, An Introduction to Modern Astrophysics, 2nd ed., §18 (neutron stars, accretion and X-ray sources) · Shapiro and Teukolsky, Black Holes, White Dwarfs and Neutron Stars, Ch. 13 · Frank, King and Raine, Accretion Power in Astrophysics, 3rd ed. · Lewin, van Paradijs and Taam, Space Science Reviews 62 (1993) — the classic X-ray burst review · Strohmayer and Bildsten, in Compact Stellar X-ray Sources (2006) — new views of thermonuclear bursts. All numbers use canonical parameters $M=1.4\,M_\odot$ , $R=10\,\mathrm{km}$ . Key values: $g\approx2.4\times10^{14}\,\mathrm{cm\,s^{-2}}$ (1); $E_{\mathrm{grav}}\approx193\,\mathrm{MeV/nucleon}$ (2); nuclear yields $\approx6$ (H) and $\approx1.6$ (He) MeV/nucleon from binding energies; $\alpha=E_{\mathrm{grav}}/E_{\mathrm{nuc}}\approx40$ – $120$ (4); $L_{\mathrm{Edd}}\approx3.5\times10^{38}\,\mathrm{erg\,s^{-1}}$ (He) and $2.1\times10^{38}$ ( $X=0.7$ ) with $\kappa_{\mathrm{es}}=0.2(1+X)$ (7); $\nu=4.40/T_9-3\approx40$ at $T_9=0.1$ , activation energy $0.38\,\mathrm{MeV}$ = the Hoyle resonance (6); $t_{\mathrm{rec}}\approx2.8$ – $3.6\,\mathrm{h}$ (8). Ignition columns and burst-regime boundaries are drawn from Lewin, van Paradijs and Taam (1993) and Strohmayer and Bildsten (2006); the rp-process endpoint from Schatz et al. (1998).

Prerequisites. Neutron Stars · Electron Degeneracy Pressure · Neutron Star Structure and Interior · Equation of State of Dense Matter. Leads to. Millisecond and Recycled Pulsars · Tolman–Oppenheimer–Volkoff Equation.

Ask the tutor: Ask me how a single photospheric-radius-expansion burst pins down a neutron star's mass, radius, and distance at once — or why cranking up the accretion rate switches the bursts off.

What is the difference between Type I and Type II X-ray bursts?

They have completely different engines. A Type I burst is thermonuclear: fuel accumulates on the surface and ignites in an unstable shell flash, so its energy comes from fusion and its spectrum is a cooling blackbody. A Type II burst is accretion-driven: a sudden release of gravitational energy when the accretion flow itself becomes unstable and dumps matter onto the star, seen almost only in the Rapid Burster. The tells are the spectral softening in the decay and the value of $\alpha$ : Type I bursts obey $\alpha\sim40$ – $100$ because nuclear energy is a small fraction of accretion energy, while Type II bursts show no such fingerprint.

Why is the α ratio so large, and what does it tell us?

$\alpha$ compares the energy released between bursts (gravitational, from accretion) with the burst energy (nuclear, from fusion) on the same fuel. It is large — $40$ to $100$ — because falling onto a neutron star is far more violent than burning: a nucleon liberates $\sim200\,\mathrm{MeV}$ reaching the surface but only a few MeV when it fuses. So $\alpha\approx E_{\mathrm{grav}}/E_{\mathrm{nuc}}$ , and its value diagnoses the fuel: helium-rich ( $E_{\mathrm{nuc}}\approx1.6$ ) gives $\alpha\approx120$ , hydrogen-rich ( $E_{\mathrm{nuc}}\approx5$ – $6$ ) gives $\alpha\approx35$ . Measuring $\alpha$ is effectively reading off what burned.

Why do the brightest bursts come from intermediate accretion rates, not the highest?

Because at high rates the fuel never gets to accumulate. When $\dot M$ exceeds about a tenth of Eddington, the base is kept hot enough that hydrogen and eventually helium burn stably — simmering as fast as they arrive, so no unstable reservoir builds and the bursts fade out. At the lowest rates the fuel is hydrogen-rich and burns in weaker mixed flashes. The sweet spot is intermediate: hot-CNO burning quietly consumes the hydrogen, leaving a growing layer of nearly pure helium that ignites in a single energetic triple-alpha flash — the kind that reaches the Eddington limit and shows radius expansion.

What is photospheric radius expansion, and why does it make a standard candle?

When a burst reaches the Eddington luminosity, radiation pressure cannot be balanced by pushing harder on the gas, so the surplus inflates the atmosphere: the photosphere expands, sometimes to many times the stellar radius. Because the luminosity stays pinned at $L_{\mathrm{Edd}}$ , the blackbody temperature drops as the radius grows — an unmistakable signature. It is useful because $L_{\mathrm{Edd}}$ depends only on the mass and hydrogen fraction, not on distance or history, so it is nearly universal. The flux at "touchdown", when the photosphere settles back to the surface, gives the distance, and the cooling-tail area gives the radius.

What is the rp-process, and where does it stop?

The rp-process — rapid proton capture — carries burst nucleosynthesis to heavy elements. After the hot-CNO cycle and the $\alpha p$ -process, protons are captured onto seed nuclei faster than the intervening beta decays, marching the composition up the proton-rich side of the chart to $A\sim100$ . It stalls at the tin–antimony–tellurium region, where a closed cycle of alpha decays (the SnSbTe cycle) blocks further advance. The long tails of some burst light curves are its signature, and the carbon-bearing ashes it leaves seed deeper crustal processes and superbursts.

Where does the "4.4" in the triple-alpha rate come from, and why is the sensitivity about 40 not 500?

The triple-alpha reaction proceeds resonantly through the Hoyle state of carbon-12, $0.38\,\mathrm{MeV}$ above the three-alpha threshold. That resonance energy enters the rate as $\exp(-E_a/k_BT)$ with $E_a\approx0.38\,\mathrm{MeV}$ , which with temperature in units of $10^9\,\mathrm{K}$ becomes $\exp(-4.40/T_9)$ — so " $4.40$ " is $E_a/k_B$ in those units, not an energy in MeV. Differentiating gives $\nu=4.40/T_9-3\approx40$ near $T_9=0.1$ . The common slip is to read the exponent as $4.4\,\mathrm{MeV}/k_BT$ , which would give $\nu\approx500$ ; that is a units error. The true $\nu\approx40$ is still far above the cooling sensitivity of $\sim4$ , all the instability needs.