Stellar Absorption Spectra — How to Read the Dark Lines in Starlight

In 1814 a Bavarian lens grinder named Joseph von Fraunhofer pointed a precision prism at the Sun. He expected a smooth rainbow. He saw, instead, a rainbow chopped by hundreds of razor-thin dark stripes. He labelled the most prominent ones A through K — Fraunhofer lines — and noted that the same lines appeared, slightly shifted, when he looked at Sirius and Castor. He had no idea what they were. He died fifteen years later still not knowing.

Half a century later Gustav Kirchhoff and Robert Bunsen lit the answer on a Heidelberg bench. A hot gas, they showed, emits bright lines at specific wavelengths that are unique to each element — sodium glows yellow at 589 nm, no matter whether it's in a flame or a supernova. A cooler gas placed between you and a hotter continuous source absorbs those same wavelengths. The Fraunhofer lines are shadows cast by specific atoms in the solar atmosphere, sitting above the blazing photosphere below.

Every star is a Bunsen burner turned inside out. The deep photosphere emits a thermal rainbow. A cooler gas layer just above it subtracts the wavelengths its atoms happen to like. What's left — continuum minus dips — is the spectrum we record.

By 1900, Annie Jump Cannon at Harvard had classified the spectra of more than a quarter of a million stars, sorting them into the sequence we still use: O, B, A, F, G, K, M. Hot O stars showed helium lines and weak hydrogen. Cool M stars showed sprawling dark bands where no atomic line should be. A stars like Vega had the most brutal hydrogen Balmer lines of anyone.

The obvious conclusion, at the time, was that the stars had wildly different chemistry — some made of helium, some of hydrogen, some of metals, some of whatever made those strange molecular bands. Stellar spectra were a catalogue of alien elemental soups.

Then in 1925, a 25-year-old doctoral student named Cecilia Payne showed that almost all of it was wrong. Using the brand-new quantum physics of ionisation, she demonstrated that stellar spectra change from O to M because the outer layers get cooler. At 30,000 K, hydrogen is too hot — every atom is ionised, stripped of its electron, and ionised hydrogen can't absorb Balmer lines. At 3,000 K, hydrogen is too cold — the electron sits placidly in the ground state and ignores visible photons. Only at the in-between temperature of an A star, around 9,500 K, do hydrogen electrons live exactly on the second energy level, ready to leap when a red photon passes by.

Every star in the galaxy, Payne concluded, is made of essentially the same stuff — three-quarters hydrogen, a quarter helium, a trace of everything else. The spectra look different because the thermometers read different temperatures, not because the ingredients are different.

Here's what each spectral class looks like in Nightbase's simulator, and the star on your next clear night where you can verify it.

B stars — hot enough for helium

At 25,000 K, Spica (B1 IV) is hot enough to ionise helium. A B-star spectrum is sparse: a few sharp He I lines (447, 471, 501 nm), a handful of faint Balmer lines, almost nothing else. The continuum blazes blue because the Planck curve peaks in the ultraviolet.

Rigel (B8 Ia), cooler at 12,100 K, sits at the B/A boundary. He I has started to fade; Balmer lines are growing fast.

A stars — the Balmer showcase

At ~9,500 K, hydrogen is at its sweet spot. Vega (A1 V, 9,600 K) and Sirius (A1 V, 9,984 K) display the most violent Balmer absorption of any class: Hα at 656 nm carves a canyon out of the red; Hβ (486 nm), Hγ (434 nm), Hδ (410 nm) march in ever-closer spacing toward the ultraviolet, tracing the quantised energy levels of Bohr's hydrogen atom exactly.

Open Vega's detail page and scroll to the Stellar Absorption Spectrum panel. The four great hydrogen gashes are unmistakable. Compare it to Sirius on Sirius's page — nearly identical. That's why Sirius and Vega were once used as flux standards: their spectra are simple, reproducible, and dominated by one well-understood element.

F and G stars — the Sun's neighbourhood

Cool off to 6,500 K and Balmer weakens. What takes over are calcium II H and K at 393 and 397 nm — the two deepest lines in any Sun-like spectrum — plus a forest of iron lines and the sodium D doublet at 589 nm. This is Procyon (F5 IV-V, 6,516 K), and it is our own Sun (G2 V, 5,778 K). If you'd like to see a G-star spectrum tonight, look at Capella (G1 III + K0 III) — the brightest G-type star in our sky.

K stars — calcium takes over

Arcturus (K0 III, 4,291 K) is the canonical K-giant spectrum: Ca II H and K are now bottomless, Na D is obvious, hydrogen is a faint footnote. Aldebaran (K5 III, 3,881 K) pushes further down — faint TiO bands begin to appear in the red.

M stars — the molecule-dominated world

Below ~3,700 K, atoms can keep their electrons, and some atoms can even form molecules. Betelgeuse (M4 Ib, 3,479 K) and Antares (M1.5 Iab-Ib, 3,497 K) are the showpieces. Their spectra are barely recognisable: instead of sharp atomic needles, broad titanium oxide (TiO) bands mow down entire swaths of the red and near-infrared. The star still glows red-hot, but whole wavelength ranges are absorbed away by molecules in its atmosphere.

Open Betelgeuse's page and the simulator makes the jaw drop. The continuum sags in orange channels around 620, 670, and 705 nm where TiO devours photons. Hydrogen is invisible. Metals are lost in the molecular noise.

The Balmer series is where you first feel quantum mechanics in an astronomical spectrum. A hydrogen electron sits on a ladder of discrete energy rungs. Starting from the second rung (the n = 2 level), it can absorb a photon and jump to n = 3 (that's Hα, 656 nm), to n = 4 (Hβ, 486 nm), to n = 5 (Hγ), to n = 6 (Hδ), and so on — each leap requiring a larger, bluer photon. The series crowds ever tighter toward a limit at 365 nm, the Balmer break, where infinity rungs merge into a continuum.

Why does Vega show them so clearly? Because at 9,500 K, most hydrogen atoms have their electron excited to exactly n = 2 (thermally, by collisions with photons and other atoms). A photon passing by can pick any atom up to a higher rung. Hotter, and the electrons are stripped off entirely — ionised hydrogen doesn't absorb at all. Colder, and the electrons fall back to n = 1, where they can only absorb ultraviolet Lyman photons that are hidden from our telescopes by Earth's atmosphere.

A stars are the hydrogen sweet spot. That's why they dominate every introductory spectroscopy chart.

The coolest stars don't look like the warmer ones with a temperature dial turned down. They look like a different kind of object entirely, because they are cool enough to preserve chemical bonds in their photospheres. Once molecules exist, they absorb photons in bands rather than single lines — each molecular vibration or rotation can accept a continuous range of photon energies, and the absorption spreads into wide brushstrokes.

• TiO (titanium oxide) — the signature of M stars. Wide troughs near 705, 670, and 620 nm. Turns M-giant spectra into orange-and-black rainbows with no orange left.
• C₂ (diatomic carbon) and CN — the carbon stars, where the photosphere holds more carbon than oxygen. Carbon steals all the oxygen into CO (which doesn't absorb in the visible), leaving none for TiO. The spectrum flips from TiO-dominated to C₂/CN-dominated. These stars look intensely red because blue photons get eaten by CN bands.
• ZrO (zirconium oxide) — the elegant S stars, intermediate between M and C, where the carbon-to-oxygen ratio is almost exactly one. Zirconium from the star's own nuclear ashes floats up to the surface and claims the leftover oxygen.

An M or C star's spectrum tells you the stellar atmosphere is cool, convective, and churning freshly-fused elements up from the interior. That's why molecular spectroscopy was how astronomers first confirmed that AGB stars dredge nuclear ash to their surfaces — a cornerstone of stellar evolution and the heavy-element enrichment of the galaxy.

Open any star you like on Nightbase — Vega, Betelgeuse, Arcturus, Capella — and expand the Stellar Absorption Spectrum panel under Explore. You'll see:

• A Planck continuum coloured by the star's actual B−V colour index — blue for hot stars, deep red for M giants.
• Labelled absorption dips for every major line the spectral class predicts: Hydrogen Balmer, Ca II H and K, Na D, Mg b, Fe I, He I/He II for hot stars, TiO/C₂/CN/ZrO for cool ones.
• A legend of the six strongest elements for that star.
• Hover the canvas — a tooltip tells you the wavelength, the flux, and the nearest absorption line with its strength as a percentage.

The simulator draws each line as a Gaussian dip whose depth and width depend on the star's spectral subclass. Drag Nightbase's spectral type from A0 to M5 (via the catalogue search) and watch hydrogen die while TiO is born. It is the O-B-A-F-G-K-M sequence on a single canvas — exactly what Payne explained, rendered in 60 frames per second.

Spectroscopy is the thread that runs through almost every article on this site. The Hertzsprung-Russell diagram plots stars by luminosity and spectral class. Nuclear fusion in stars explains why the Sun's core produces a G-type spectrum. The life of stars is basically a trip along the O-M sequence as a star ages. Reading a spectrum is the skeleton key to all of it.