Planetary Nebulae

A planetary nebula is the most photogenic bookkeeping error in the sky: a glowing shell of gas that looks like a birth but is really a death notice. When a star of $0.8$ – $8\,M_\odot$ exhausts its fuel on the asymptotic giant branch (AGB), it does not explode — it undresses. A slow, dense wind ( $\sim10$ km s $^{-1}$ , mass loss up to $10^{-4}\,M_\odot\,\mathrm{yr^{-1}}$ ) peels the envelope away, and when the residual envelope drops below $\sim10^{-2}\,M_\odot$ the naked carbon–oxygen core is exposed. That core contracts and heats at nearly constant luminosity, sliding horizontally across the Hertzsprung–Russell diagram until its surface passes $\sim3\times10^4$ K and floods the departing gas with ultraviolet light. Then the nebula switches on. Photons past the Lyman limit ( $\lambda<912$ Å) tear electrons off hydrogen and heavier atoms; the freed electrons recombine and cascade into the recombination lines of H and He, while electron impacts pump the metastable levels of ions like O $^{2+}$ and N $^+$ , whose slow radiative decay produces the forbidden lines — [O III] $\lambda\lambda4959,5007$ , [N II], [O II] — that give planetary nebulae their eerie green glow and their diagnostic power. The spectacle is fleeting: expanding at $20$ – $40$ km s $^{-1}$ , the shell disperses in $\sim10^4$ yr, the same interval the central star takes to fade toward the white-dwarf cooling track. We catch planetary nebulae only because those two clocks run together for one luminous moment — during which they seed the interstellar medium with carbon, nitrogen, and s-process metals and serve as standard candles out to $\sim20$ Mpc.

Key takeaways

A planetary nebula (PN) is the ionized envelope a star of $0.8\lesssim M/M_\odot\lesssim8$ sheds at the end of the AGB, lit by the ultraviolet of its emerging hot core — nothing to do with planets, a misnomer from William Herschel's small telescopes.
The gas fluoresces: UV photons from the central star ( $T_{\mathrm{eff}}\sim3\times10^4$ – $2\times10^5$ K) photoionize it, and it re-radiates as recombination lines (H $\alpha$ ) and forbidden lines (the [O III] doublet, 4959 and 5007 Å).
A Strömgren balance fixes the size of the ionized zone: every ionizing photon is paid back by one recombination, giving $R_{\mathrm{S}}\sim0.15$ pc for a typical core.
The forbidden lines are bright because nebulae are near-vacuums: below the critical density $n_{\mathrm{crit}}\sim10^5$ – $10^6$ cm $^{-3}$ , metastable levels decay by photon emission instead of being collisionally quenched.
PNe are short-lived ( $t\sim R/v_{\mathrm{exp}}\sim10^4$ yr) and enrich the interstellar medium with $^4$ He, $^{12}$ C, $^{14}$ N, and s-process elements dredged up in AGB thermal pulses.
The [O III] $\lambda5007$ luminosity function has a nearly universal bright cutoff at $M^{*}\approx-4.5$ , making PNe standard candles across the local universe.
Key idea: A planetary nebula is not an evolutionary stage of the gas itself but a temporary fluorescent screen — an ionization-bounded (or density-bounded) shell of ejected envelope material, lit up by the ultraviolet flood from the newly exposed hot stellar core en route to becoming a white dwarf.

Where this lives today

Planetary nebulae remain a working frontier on three fronts. In imaging, JWST's infrared cameras have resolved the dusty inner structure of nearby objects such as the Southern Ring (NGC 3132) and the Ring Nebula, revealing central binary companions and molecular gas that optical images miss — direct tests of the shaping question. That question is the field's central open problem: most planetary nebulae are not round, yet single-star winds are nearly spherical, so most seem to demand a binary companion, a common-envelope interaction, or a magnetic field to sculpt the bipolar forms. Integral-field spectrographs like MUSE on the VLT now map temperature, density, and abundance pixel by pixel, exposing the "abundance discrepancy" between recombination and forbidden lines that hints at a cold, metal-rich second gas component. And on the cosmic-distance side, the [O III] $\lambda5007$ PNLF, calibrated against Gaia parallaxes of Galactic central stars, remains an independent rung of the distance ladder — one of the few methods reaching $\sim20$ Mpc from a single narrow-band image.

The physical picture

Picture the AGB star as a bloated onion held loosely at its surface, where pulsations and radiation pressure on freshly formed dust grains lift matter off faster than gravity can retrieve it. Layer after layer floats away as a slow, dense wind, until the envelope is gone and a tiny, ferociously hot ember — the degenerate carbon–oxygen core — shines its ultraviolet straight into the retreating gas, like a fluorescent tube switched on inside an expanding smoke ring. The ionization defines a sharp boundary, the ionization front, where the last ionizing photons are used up: inside it the gas glows, outside it stays neutral and dark. The first figure is the anatomy of that boundary and carries the number that sets its size.

Figure

Anatomy of a planetary nebula. A hot central star (T_eff ~ 10⁵ K) drives a fast tenuous wind that sweeps the earlier slow AGB wind into a dense shell, which its ultraviolet photons ionize out to the Strömgren radius R_S ≈ 0.15 pc for Q₀ = 10⁴⁷ s⁻¹ and n = 10³ cm⁻³. Inside the ionization front the gas glows, outside it stays neutral.

The colours tell a story about density. In a laboratory discharge the green [O III] transitions essentially never happen: atoms are collisionally knocked out of the metastable level long before the slow radiative decay can fire. In the near-vacuum of a nebula, seconds to minutes pass between collisions, so the forbidden decay wins and the nebula shines in lines no earthly gas can reproduce. The morphology encodes the dynamics of dying: a fast tenuous wind ( $v\sim1000$ km s $^{-1}$ ) plows into the slow AGB wind, and equatorial density enhancements — often a binary's fingerprint — funnel the outflow into the bipolar butterfly and ant shapes. But the reason astronomers love planetary nebulae is a single, remarkably stable number: the bright end of their luminosity function, shown next.

Figure

The O III λ5007 planetary-nebula luminosity function. The number of nebulae per magnitude follows N(M) ∝ e^0.307M (1 − e^3(M−M)), which truncates abruptly at the bright cutoff M ≈ −4.5 mag. That cutoff is nearly independent of the parent stellar population, so the brightest few nebulae in any galaxy read off its distance out to ~20 Mpc.

Timing is everything. The nebula expands away in a few thousand years while the central star races across the H–R diagram in the same interval, fading as its hydrogen-burning shell extinguishes. We see a planetary nebula only because these two independent clocks — dispersal of the shell and fading of the ionizing source — are synchronized to $\sim10^4$ yr. Miss the window and there is either an un-ionized wind (too early) or a bare white dwarf drifting through cold, invisible gas (too late).

The core equations

The gate that opens the whole phenomenon is a photon energy. Hydrogen binds its electron with $13.6$ eV, so only photons above that threshold — equivalently, below the Lyman-limit wavelength — can ionize it and start the cascade that lights the nebula.

h\nu \geq 13.6\ \mathrm{eV} \quad\Longleftrightarrow\quad \lambda \leq 912\ \text{\AA} \qquad (1)

where: $h\nu$ — photon energy; $13.6\ \mathrm{eV}$ — ionization energy of ground-state hydrogen; $912\ \text{\AA}$ — the corresponding threshold wavelength, the Lyman limit. Reading: a core below $\sim3\times10^4$ K emits almost no photons here, so its shell stays neutral and dark until the core is hot enough. In steady state the ionized region grows until recombinations exactly consume the ionizing photons. Balancing the supply $Q_0$ against recombinations throughout a uniform sphere fixes its radius — the Strömgren radius — as a cube-root of stellar output over density squared.

R_{\mathrm{S}} = \left(\frac{3\,Q_0}{4\pi\,n^2\,\alpha_B}\right)^{1/3} \qquad (2)

where: $R_{\mathrm{S}}$ — radius of the ionized zone; $Q_0$ — rate of hydrogen-ionizing photons (s $^{-1}$ ); $n$ — hydrogen (electron $\approx$ proton) density; $\alpha_B$ — Case B recombination coefficient. Reading: denser gas recombines faster and shrinks the sphere, a hotter core enlarges it, and whether $R_{\mathrm{S}}$ falls inside or outside the shell decides ionization- versus density-bounded. The recombination coefficient above is not a free parameter but a well-measured atomic constant. "Case B" means recombinations directly to the ground state are discounted, because the Lyman-continuum photon they emit is immediately reabsorbed nearby.

\alpha_B(10^4\,\mathrm{K}) \approx 2.6\times10^{-13}\ \mathrm{cm^3\,s^{-1}} \qquad (3)

where: $\alpha_B$ — total recombination coefficient to the excited levels of hydrogen at electron temperature $T_e=10^4$ K. Reading: plug this into (2) with $Q_0\sim10^{47}$ s $^{-1}$ and $n\sim10^3$ cm $^{-3}$ and the ionized sphere comes out at $\sim0.15$ pc — exactly the size of a real young planetary nebula. The shell's age is read straight off its kinematics. The Doppler splitting of the lines gives the expansion speed, the angular radius scaled by distance gives the physical radius, and their ratio is the elapsed time since ejection.

t_{\mathrm{dyn}} = \frac{R}{v_{\mathrm{exp}}} = \frac{\theta\, d}{v_{\mathrm{exp}}} \qquad (4)

where: $R$ — physical shell radius; $v_{\mathrm{exp}}$ — expansion velocity ( $20$ – $40$ km s $^{-1}$ ); $\theta$ — angular radius; $d$ — distance. Reading: for $R\sim0.3$ pc and $v_{\mathrm{exp}}\sim25$ km s $^{-1}$ this gives $\sim10^4$ yr — the visible lifetime, and also the central star's crossing time. The forbidden lines that dominate the visible spectrum survive only where the gas is thin enough. The decisive quantity is the critical density, where collisional de-excitation of a metastable level just balances its radiative decay.

n_{\mathrm{crit}} = \frac{A_{ji}}{q_{ji}} \qquad (5)

where: $A_{ji}$ — spontaneous radiative decay rate of the metastable level (s $^{-1}$ , tiny for a forbidden transition); $q_{ji}$ — collisional de-excitation rate coefficient (cm $^3$ s $^{-1}$ ). Reading: for the [O III] $^1D_2$ level $n_{\mathrm{crit}}\sim5\times10^5$ cm $^{-3}$ , and since nebulae sit at $n_e\sim10^2$ – $10^4\ll n_{\mathrm{crit}}$ , almost every excitation ends in a photon. Those same forbidden lines are quantitative thermometers. The auroral line at $\lambda4363$ needs more energetic electrons to populate its upper level than the nebular lines at $\lambda\lambda4959,5007$ , so their intensity ratio is a steep, calibratable function of the electron temperature.

\frac{I(\lambda4959)+I(\lambda5007)}{I(\lambda4363)} = f(T_e) \qquad (6)

where: $I(\lambda)$ — emission-line intensity; $\lambda\lambda4959,5007$ — nebular lines from the $^1D_2$ level; $\lambda4363$ — auroral line from the higher $^1S_0$ level; $T_e$ — electron temperature ( $\sim10^4$ K). Reading: measure the ratio, invert $f$ , and read off the plasma temperature — the first step toward every nebular abundance. Finally, the property that made planetary nebulae a cosmological tool: the empirical luminosity function in the [O III] $\lambda5007$ line, which rises from a sharp bright cutoff and has a nearly universal shape.

N(M) \propto e^{0.307\,M}\left[\,1 - e^{3(M^{*}-M)}\,\right],\qquad M^{*}\approx -4.5 \qquad (7)

where: $N(M)$ — number of nebulae per unit absolute magnitude; $M$ — absolute magnitude in [O III] $\lambda5007$ ; $M^{*}$ — bright-end cutoff, the standard-candle anchor. Reading: because $M^{*}$ barely changes between galaxies, matching the observed cutoff to $M^{*}=-4.5$ gives a distance good to $\sim10\%$ out to $\sim20$ Mpc.

Derivation AStrömgren radius from ionization balance

The goal is to turn the intuitive statement "the ionized zone grows until it runs out of ionizing photons" into the quantitative Strömgren radius, showing exactly which quantities set the size of a planetary nebula. We assume a uniform, pure-hydrogen sphere of density $n$ around a star emitting $Q_0$ ionizing photons per second, demand a photoionization–recombination steady state, and integrate over the sphere. The result is the workhorse length scale of every ionized nebula, and it reproduces the observed $\sim0.1$ pc size to within a factor of order unity.

Step 1. Write the global balance: in equilibrium the star's ionizing output equals the total recombinations per second inside the ionized volume, so $Q_0 = \int_0^{R_{\mathrm{S}}} n_e n_p\,\alpha_B\,4\pi r^2\,\mathrm{d}r$ . Why: the region settles where ion creation (photons in) matches destruction (recombinations) — the definition of the ionization front.

Step 2. Use Case B and pure hydrogen: recombinations directly to $n=1$ are excluded because their photon is reabsorbed on the spot, and for fully ionized hydrogen $n_e=n_p=n$ . Why: Case B counts only photons that escape the local gas against $Q_0$ ; dropping helium and metals shifts the answer by just tens of percent.

Step 3. With constant density the integrand is uniform, so the integral collapses to a volume, $Q_0 = \tfrac{4}{3}\pi R_{\mathrm{S}}^3\,n^2\,\alpha_B$ . Why: uniform density is the cleanest first model — every part of the sphere recombines at the same rate.

Step 4. Solve for the radius, $R_{\mathrm{S}} = \left(3Q_0/4\pi n^2\alpha_B\right)^{1/3}$ , and evaluate with $Q_0\sim10^{47}$ s $^{-1}$ , $n\sim10^3$ cm $^{-3}$ , $\alpha_B\sim2.6\times10^{-13}$ cm $^3$ s $^{-1}$ to get $R_{\mathrm{S}}\approx4.5\times10^{17}$ cm $\approx0.15$ pc. Why: the cube root makes the size insensitive to the inputs — a factor-10 error in $Q_0$ moves $R_{\mathrm{S}}$ by only $\sim2$ , so the $\sim0.1$ pc scale is robust. *Units check: $[Q_0/(n^2\alpha_B)] = \mathrm{s^{-1}}/(\mathrm{cm^{-6}\cdot cm^3\,s^{-1}}) = \mathrm{cm^3}$ , and cube-rooting gives cm, as required; the numerical result is $R_{\mathrm{S}}=0.146$ pc.*

Go deeper · Derivation B · 1 minDynamical age of an expanding shell

Derivation BDynamical age of an expanding shell

The goal is to show that a planetary nebula wears its own age on its sleeve, recoverable from two observable quantities plus a distance, and to make explicit why that age lands near $10^4$ yr — the number that explains why PNe are rare snapshots rather than long-lived objects. We convert the Doppler expansion velocity and the angular size into a physical radius, assume constant-speed coasting since ejection, and divide. The estimate is an upper bound for a decelerating flow, yet it agrees with the independent central-star evolutionary timescale.

Step 1. Get the physical radius from geometry, $R = \theta\,d$ , using the small-angle relation between the angular radius $\theta$ (in radians), the distance $d$ , and the true radius $R$ . Why: the nebula is resolved on the sky, so its apparent size plus a distance is a direct ruler — no modelling of the emission needed.

Step 2. Assume constant-velocity coasting, $R \approx v_{\mathrm{exp}}\,t$ , so the radius grows linearly with time. Why: once launched, the envelope feels negligible gravity from the low-mass core over $10^4$ yr, so free expansion is an excellent approximation; slight acceleration by the fast wind makes this an upper bound on the age.

Step 3. Solve for the age, $t_{\mathrm{dyn}} = R/v_{\mathrm{exp}}$ , and evaluate with $R\sim0.3$ pc $=9.3\times10^{17}$ cm and $v_{\mathrm{exp}}\sim25$ km s $^{-1}=2.5\times10^6$ cm s $^{-1}$ : $t_{\mathrm{dyn}}\sim3.7\times10^{11}$ s $\approx1.2\times10^4$ yr. Why: this coincides with the time the central star takes to heat, ionize, and fade across the H–R diagram, so source and screen are both transient on the same $\sim10^4$ yr — hence the narrow visibility window. *The numerical result is $t_{\mathrm{dyn}}=1.17\times10^4$ yr, which coincides with the post-AGB crossing times of $10^3$ – $10^4$ yr seen in stellar evolution tracks.*

Go deeper · Derivation C · 1 minWhy forbidden lines shine — the critical density

Derivation CWhy forbidden lines shine — the critical density

The goal is to resolve the central paradox of nebular spectroscopy — that "forbidden" transitions, strongly suppressed in any laboratory, are among the brightest features a nebula emits — by deriving the critical density that separates the two regimes. We follow one ion excited into a metastable level, ask whether it de-excites by a photon or a collision, balance the two rates, read off the density where they are equal, and compare it to real nebular densities.

Step 1. A metastable upper level $j$ empties two ways: spontaneous radiative decay at rate $A_{ji}$ , and collisional de-excitation by electrons at rate $n_e\,q_{ji}$ . Why: for a forbidden transition $A_{ji}$ is tiny — only weakly allowed as a magnetic-dipole or electric-quadrupole process — so whether it radiates depends entirely on how it competes with collisions.

Step 2. Define the critical density where the two channels are equal, $n_e\,q_{ji}=A_{ji}$ , giving $n_{\mathrm{crit}}=A_{ji}/q_{ji}$ . Why: this number partitions the physics — above it collisions win and the line is quenched, below it radiative decay wins and every excitation yields a photon.

Step 3. Evaluate for the [O III] $^1D_2$ level: $A_{ji}\sim0.026$ s $^{-1}$ and $q_{ji}\sim5\times10^{-8}$ cm $^3$ s $^{-1}$ at $10^4$ K, so $n_{\mathrm{crit}}\sim5\times10^5$ cm $^{-3}$ . Why: nebulae sit at $n_e\sim10^2$ – $10^4$ cm $^{-3}$ , far below $n_{\mathrm{crit}}$ , so the level almost always radiates — while a lab discharge at $\gg10^{6}$ cm $^{-3}$ quenches it, which is why the lines were once ascribed to a fictitious element, "nebulium". *Numerically, $n_{\mathrm{crit}}=0.026/5.0\times10^{-8}=5.2\times10^5$ cm $^{-3}$ , the standard tabulated value for [O III] and far above typical nebular densities.*

Limiting cases

An observable planetary nebula is a coincidence of several conditions, each with a sharp boundary.

Case	Result	Meaning
Ionization switched on	$T_{\mathrm{eff}}\gtrsim3\times10^4\,\mathrm{K}$	Below this the core emits too few Lyman-continuum photons; the ejected shell stays a neutral, invisible AGB wind rather than a glowing nebula.
Progenitor mass window	$0.8\,M_\odot\lesssim M_{\mathrm{ZAMS}}\lesssim8\,M_\odot$	Below $\sim0.8$ the star has not left the main sequence in a Hubble time; above $\sim8$ it ignites carbon and dies as a core-collapse supernova, not a PN plus white dwarf.
Visibility window	$t_{\mathrm{PN}}\sim10^4\,\mathrm{yr}$	Shell dispersal ( $t\sim R/v_{\mathrm{exp}}$ ) and central-star fading both act on $\sim10^4$ yr; outside this window there is no PN to see.
Forbidden-line regime	$n_e \ll n_{\mathrm{crit}}\sim10^{5\text{--}6}\,\mathrm{cm^{-3}}$	At nebular densities metastable levels radiate before collisions quench them, so the forbidden lines are strong; above $n_{\mathrm{crit}}$ (a lab) they vanish.
Ionization- vs density-bounded	$R_{\mathrm{S}} \lessgtr R_{\mathrm{shell}}$	If $R_{\mathrm{S}}$ is smaller than the shell an outer neutral zone survives (ionization-bounded); if larger, ionizing photons escape (density-bounded), and the nebula leaks Lyman-continuum radiation.

Common trapThe name, the glow, and the word "forbidden" all mislead

Three traps live in the vocabulary. First, planetary nebulae have nothing to do with planets — the name is a visual accident from William Herschel's small telescopes, in which the round greenish disks resembled Uranus. Second, the central star is not what you see glowing: the hot core is a compact ultraviolet point source, and the extended visible light is the shell reprocessing that UV into recombination and forbidden lines. Third, the green [O III] lines are "forbidden" only relative to Earth laboratories; the transition is real, just slow, and at nebular densities ( $n_e\sim10^{3\text{--}4}$ cm $^{-3}$ , far below $n_{\mathrm{crit}}$ ) the metastable level decays radiatively before a collision can quench it. What is impossible in a discharge tube is routine in a cubic parsec of near-vacuum.

Worked examples

1 · The size of the ionized sphere (confidence builder)

A planetary-nebula nucleus emits $Q_0=10^{47}$ hydrogen-ionizing photons per second into surrounding gas of uniform density $n=10^3$ cm $^{-3}$ at $T_e=10^4$ K, where $\alpha_B=2.6\times10^{-13}$ cm $^3$ s $^{-1}$ . Find the Strömgren radius.

Step 1. Assemble the ratio inside the cube root, $\dfrac{3Q_0}{4\pi n^2\alpha_B}$ . The denominator is $4\pi(10^3)^2(2.6\times10^{-13}) = 4\pi\times2.6\times10^{-7} = 3.27\times10^{-6}$ cm $^3$ s $^{-1}$ ·cm $^{-6}$ . Dividing, $3\times10^{47}/3.27\times10^{-6} = 9.2\times10^{52}$ cm $^3$ .

Step 2. Take the cube root: $R_{\mathrm{S}} = (9.2\times10^{52})^{1/3} = 4.5\times10^{17}$ cm. Converting with $1\ \mathrm{pc}=3.086\times10^{18}$ cm gives $R_{\mathrm{S}}\approx0.15$ pc.

Step 3. Sanity check: young planetary nebulae are observed at $0.05$ – $0.3$ pc, so the estimate lands in range. The cube root also makes the size stiff — a tenfold rise in $Q_0$ enlarges $R_{\mathrm{S}}$ by only $10^{1/3}\approx2.15$ — so very different central stars still make comparable nebulae.

2 · Reading the age off the sky

A planetary nebula has an angular radius of $\theta=15''$ and lies at $d=1$ kpc. Its emission lines are Doppler-split by $\pm25$ km s $^{-1}$ . How old is the shell?

Step 1. Convert the angular radius to radians: $\theta = 15'' \times 4.848\times10^{-6} = 7.27\times10^{-5}$ rad. With $d=1\ \mathrm{kpc}=3.086\times10^{21}$ cm, the physical radius is $R=\theta d = 2.24\times10^{17}$ cm $=$ $0.073$ pc.

Step 2. The half-split of the lines is the expansion speed, $v_{\mathrm{exp}}=25\ \mathrm{km\,s^{-1}}=2.5\times10^6$ cm s $^{-1}$ . Divide: $t_{\mathrm{dyn}} = R/v_{\mathrm{exp}} = 9.0\times10^{10}$ s.

Step 3. Convert to years, $t_{\mathrm{dyn}} = 9.0\times10^{10}/3.156\times10^7 =$ $\approx2.8\times10^3$ yr. This is a young, compact nebula — well inside the $\sim10^4$ yr window, with a central star still near the top of its horizontal H–R track.

3 · A planetary nebula as a distance marker

In a distant galaxy the brightest planetary nebulae pile up at an apparent [O III] $\lambda5007$ magnitude of $m^{*}=26.0$ . Taking the universal cutoff $M^{*}=-4.5$ , estimate the galaxy's distance.

Step 1. The distance modulus is the difference between apparent and absolute cutoff magnitudes, $\mu = m^{*}-M^{*} = 26.0-(-4.5) = 30.5$ mag.

Step 2. Invert the modulus, $d = 10^{(\mu/5)+1}$ pc $= 10^{(30.5/5)+1} = 10^{7.1}$ pc.

Step 3. Evaluate: $d = 1.26\times10^{7}$ pc $=$ $\approx12.6$ Mpc, well within the $\sim20$ Mpc reach — and it needs only imaging through a narrow [O III] filter, no resolved stars or light curves. This is why the PNLF is a favourite cross-check on Cepheid and surface-brightness-fluctuation distances.

Physics of the portrait

For more than sixty years the brightest lines in every nebular spectrum belonged to no known element. William Huggins recorded a strong green line in 1864 and, since it matched nothing in the lab, astronomers invented a substance to carry it: "nebulium". The resolution was not a new element but a new regime — the sky running an experiment no laboratory could.

Origins · 1 minIra Sprague Bowen and how this was discovered

Ira Sprague Bowen

1898–1973 · nebular spectroscopy and the identification of forbidden lines

Ira Sprague Bowen was an American physicist trained in the precision spectroscopy of Robert Millikan's Caltech, and it was that laboratory rigour, turned on the heavens, that dissolved the sixty-year "nebulium" mystery. In 1927 Bowen showed that the unidentified nebular lines — above all the green pair at 4959 and 5007 Å — were not an exotic element but forbidden transitions of ordinary, doubly ionized oxygen (O III) and singly ionized nitrogen and oxygen: magnetic-dipole and electric-quadrupole transitions from metastable levels that a terrestrial gas can never show, because at laboratory densities collisions de-excite the level long before the slow radiative decay can occur. Only in the extreme rarefaction of a nebula, far below the critical density, do they radiate. The insight converted the emission-line spectrum from a mystery into the field's most powerful diagnostic — the basis of every nebular temperature and abundance measured since. Bowen went on to identify the fluorescence mechanism still named for him, and later, as director of the Mount Wilson and Palomar Observatories, oversaw commissioning of the 200-inch Hale Telescope.

History

1785

— William Herschel coins the term "planetary nebula" for the round, planet-like disks he catalogues

1864

— William Huggins records the first nebular spectrum and attributes its green line to "nebulium"

1927

— Ira Bowen identifies the nebulium lines as forbidden transitions of ionized oxygen and nitrogen

1939

— Bengt Strömgren derives the radius of the ionized region around a hot star

1978

— Kwok, Purton, and FitzGerald propose the interacting-stellar-winds model of nebular shaping

1989

— Ciardullo and Jacoby establish the O III λ5007 luminosity function as an extragalactic standard candle

Going further

Texts. Carroll & Ostlie, An Introduction to Modern Astrophysics, 2nd ed., Ch. 12–13 (post-main-sequence evolution) · Osterbrock & Ferland, Astrophysics of Gaseous Nebulae and Active Galactic Nuclei, 2nd ed. (the definitive treatment of Strömgren spheres, forbidden lines, and nebular diagnostics) · Kwok, The Origin and Evolution of Planetary Nebulae (2000) · Ryden, Introduction to Cosmology / Foundations of Astrophysics, on nucleosynthesis and chemical enrichment. Key values used throughout this article: the Lyman limit $13.6\ \mathrm{eV}\leftrightarrow912$ Å; $\alpha_B(10^4\,\mathrm{K})\approx2.6\times10^{-13}$ cm $^3$ s $^{-1}$ ; the Strömgren radius $R_{\mathrm{S}}\approx0.15$ pc for $Q_0=10^{47}$ s $^{-1}$ , $n=10^3$ cm $^{-3}$ ; the dynamical age $\sim10^4$ yr from $R/v_{\mathrm{exp}}$ ; the critical density $n_{\mathrm{crit}}([O\ III]\ ^1D_2)\sim5\times10^5$ cm $^{-3}\gg n_e$ ; the [O III] wavelengths 4363, 4959, 5007 Å; the PNLF cutoff $M^{*}\approx-4.5$ and the distance-modulus example yielding $d\approx12.6$ Mpc; and the progenitor mass window $0.8$ – $8\,M_\odot$ . All are standard order-of-magnitude figures.

Prerequisites. Asymptotic Giant Branch and Thermal Pulses · Post-Main-Sequence Evolution and the Red Giant Branch · Helium Burning and the Triple-Alpha Process · The Hertzsprung–Russell Diagram. Leads to. White Dwarfs and Electron Degeneracy Pressure · Advanced Burning Stages in Massive Stars.

Ask the tutor: Ask me to walk through why a shell only 0.15 pc across can outshine its own star in green light — and how astronomers turn that glow into a distance across 20 million parsecs.

Why are planetary nebulae called that if they have nothing to do with planets?

It is a historical accident of instrument quality. In the 1780s William Herschel was cataloguing faint fuzzy objects, and a subset appeared as small, round, greenish disks — much like the newly discovered planet Uranus seen through the same modest telescope. He grouped them as "planetary nebulae" for their appearance, not their nature. The name stuck even after spectroscopy revealed them to be hot ionized gas ejected by dying stars, unrelated to planets — but every textbook now warns students not to read it literally.

What is the difference between a planetary nebula and a supernova remnant?

Both are expanding shells of glowing gas, but they mark opposite fates. A planetary nebula is the gentle death of a $0.8$ – $8\,M_\odot$ star: the envelope drifts off in a slow wind, is lit by the exposed core, and leaves a white dwarf. A supernova remnant is the debris of a violent death — core collapse above $\sim8\,M_\odot$ , or the thermonuclear disruption of a white dwarf — expanding at thousands of km s $^{-1}$ (versus tens for a PN) and shining by shock heating and synchrotron rather than photoionization, leaving a neutron star or black hole.

If the central star is so hot, why does the nebula look green rather than blue?

The colour you see is the gas, not the star. The central star is extremely hot ( $\sim10^5$ K), radiates mostly in the ultraviolet where the eye cannot follow, and as a compact point contributes little visible light. That ultraviolet is absorbed by the gas and re-emitted at optical wavelengths — above all the [O III] forbidden doublet at 4959 and 5007 Å, in the green. So a nebula converts an invisible UV source into a visible line spectrum tinted green by [O III]; regions rich in H $\alpha$ recombination glow reddish instead.

How do the forbidden lines actually measure temperature and density?

Through line ratios that respond to the emitting plasma. For temperature, compare a line whose upper level is easy to excite (nebular [O III] $\lambda\lambda4959,5007$ ) with one that needs more energetic electrons (auroral $\lambda4363$ ): their intensity ratio is a steep function of $T_e$ , inverting to $\sim10^4$ K. For density, use two lines from levels of nearly equal energy but different critical densities — the [S II] $\lambda6716/\lambda6731$ or [O II] $\lambda3729/\lambda3726$ doublets — whose ratio shifts as collisional de-excitation turns on with rising $n_e$ . Together they pin down the two numbers every abundance needs.

What determines whether a nebula is round, elliptical, or bipolar?

The shaping of planetary nebulae is a live debate, but the leading answer is asymmetry in the mass-loss geometry, usually traced to a companion. A single, non-rotating star sheds a nearly spherical wind, yet most observed PNe are elliptical, bipolar, or multipolar. The favoured mechanism is a dense equatorial torus — from a binary companion, a common-envelope interaction, or magnetic fields — that pinches the later fast wind at the waist so it breaks out along the poles into the butterfly and hourglass forms. The strong statistical link between complex shapes and detected binaries is why many now suspect close binarity is central to making a PN at all.

Will the Sun make a planetary nebula, and what about Earth?

Yes, most likely. The Sun sits in the $0.8$ – $8\,M_\odot$ window, so in about 5 billion years it will ascend the AGB, shed its envelope, and — if its exposed core heats fast enough before the ejecta disperse — briefly light a planetary nebula before settling as a carbon–oxygen white dwarf. A subtlety: low-mass stars cross the H–R diagram slowly, so a very low-mass core might not ionize its ejecta before they thin, giving a faint or absent nebula. Either way, Earth's fate is sealed earlier, in the red-giant phase, when the swollen Sun scorches and may engulf the inner planets.