Galaxies and the Interstellar Medium

Take every galaxy in a reasonable atlas, lay them out on a table, and a pattern jumps out. Some are featureless ellipsoids — glowing eggs, no arms, no structure. Others are flat disks threaded with spiral arms and dust lanes. A surprising number have a straight bar of stars crossing the nucleus before the arms take off. And then there are the train wrecks — galaxies too messy to file.

Edwin Hubble noticed this in the 1920s and published a classification in 1926 that still runs the show. Draw a tuning fork. On the handle: ellipticals, from E0 (near-spherical) through E7 (cigar-shaped). At the branch point: S0 lenticulars, disk galaxies without spiral arms. On the top prong: normal spirals Sa–Sc, from tight-wound with big bulges to loose-armed with small bulges. On the bottom prong: barred spirals SBa–SBc. Trailing off the end: irregulars, everything else.

Hubble thought ellipticals evolved into spirals. He was wrong about that. But the morphological classes turned out to mean something deeper. Ellipticals are gas-poor, dust-poor, dominated by old red stars, with random stellar orbits — they look yellow-orange. Spirals are gas-rich, dust-laden, still forming stars in their arms, orbits organized around a disk — they look blue where the new stars are. The Milky Way is an SBbc: a barred spiral with a modest bulge and well-organized arms. M31 Andromeda, our nearest big neighbour, is Sb. M87 in Virgo is a giant E1, weighing in at maybe a trillion solar masses of mostly old stars.

You can see the morphological sequence at the eyepiece in a single night. M31 is the textbook Sb — the outer arms fade to tantalizing wisps in a 6-inch from dark skies. M51 is a fierce Sc with dust lanes and a companion. M87 is an E1 that looks, honestly, like a fuzzy ball — which is exactly what an old elliptical should look like. M82 in Ursa Major is an irregular starburst, tortured by a close pass with M81 300 million years ago and now gushing superwinds of ionized gas from its centre.

In 1977, Brent Tully and Richard Fisher noticed something quietly miraculous: a spiral galaxy's total luminosity and its rotation speed are locked in a tight power-law relationship. Spin faster, shine brighter — with almost no scatter. Roughly, L ∝ v_rot⁴. Double the rotation speed and the galaxy is sixteen times more luminous.

This is a gift. Distances in extragalactic astronomy are brutal to measure — parallax stops working past a kiloparsec, and even Cepheid variables stop working past 30 Mpc or so. But rotation speed is a kinematic measurement. You don't need to know the distance. You observe the 21-cm line from the galaxy's hydrogen — redshifted on the receding limb, blueshifted on the approaching limb — and measure the line's full width. That width, corrected for the galaxy's inclination, gives you v_rot directly.

The physics behind Tully-Fisher is still not fully wrapped up, but the simple version goes like this. A galaxy's rotation speed is set by the total mass inside a given radius (including dark matter). Its luminosity is set by the mass of its stars. If the ratio of stellar mass to total mass is roughly constant across spirals — which it appears to be — then luminosity and rotation speed are locked together. Why the slope should be exactly 4 rather than 3 or 5 touches on deep questions about dark matter profiles. That it works is enough to use.

Astronomers call every element heavier than helium a "metal". Neon is a metal. Oxygen is a metal. So is iron. This is jargon that embarrasses chemists and has lived for a hundred years anyway.

A star's metallicity — usually written [Fe/H], the logarithmic iron-to-hydrogen ratio relative to the Sun — is one of the three numbers that describe it (the others are mass and age). [Fe/H] = 0 is solar composition. [Fe/H] = -1 means ten times less iron than the Sun. [Fe/H] = -3 means a thousand times less.

Metallicity is a time machine. The very first stars — Population III, hypothetical and not directly observed — had zero metals because nothing heavier than lithium existed when they formed. Population II stars, which we do see in globular clusters and the galactic halo, have [Fe/H] around -1 to -2 — ancient, formed when the Milky Way was young. Population I stars like the Sun and everything in the spiral arms have near-solar metallicity. They formed from gas already seasoned by billions of years of supernovae and stellar winds.

Metallicity bleeds into almost every stellar property. Low-metallicity stars are hotter and more luminous for their mass, because metals provide the opacity that slows radiation leaking out of a stellar interior. Strip the metals, and photons escape more easily, the star adjusts by becoming brighter and bluer. Metallicity also sets how transparent a protoplanetary disk is, how efficiently a stellar wind carries mass away, how likely a star is to form planets at all (high-metal stars host more giant planets), and ultimately which kind of supernova it produces.

When you observe M31, the yellowish glow of the bulge is Population II — old, metal-poor stars. The bluish glow of the arms is Population I — young, metal-rich stars. You are literally looking at two different chemical generations in the same telescope view.

Take a star 25 times the Sun's mass. Let it burn through its hydrogen core in a few million years. Normally it would swell into a red supergiant — but if it's massive enough, its radiation pressure is so ferocious that the outer envelope doesn't just swell; it gets blown off. Layer by layer, the star strips itself down to its inner helium-burning core. What's left is a Wolf-Rayet star.

Wolf-Rayets were catalogued in 1867 by Charles Wolf and Georges Rayet at the Paris Observatory, who noticed three stars in Cygnus whose spectra were unlike anything else — dominated by broad emission lines of helium, carbon, nitrogen, and oxygen, with almost no normal absorption. Those lines come from a tenuous but blindingly hot wind — surface temperatures around 50,000–200,000 K, driving mass outflows at 1,000–3,000 km/s.

A Wolf-Rayet sheds a solar mass of material every 100,000 years. Compare that to the Sun, which loses one solar mass every trillion years. These stars are on fire, and they are running out of fuel fast. Every Wolf-Rayet you see today will go core-collapse supernova — probably a Type Ib or Ic — in less than a million years. They are the primary pathway by which the universe makes its carbon, nitrogen, and oxygen and delivers those elements back to the interstellar medium.

In northern skies, you can find the WR signature visually if you know where to look. The Crescent Nebula (NGC 6888) in Cygnus is a bubble blown into the interstellar medium by a Wolf-Rayet star, WR 136, whose winds have swept up and shocked surrounding gas into a ring-shaped nebula. It's 5,000 light-years away, faintly visible in a good OIII filter even in modest apertures, and you're looking at active stellar butchery.

A HII region — pronounced "H-two" — is a cloud of ionized hydrogen. The Roman numeral is a spectroscopist's convention: HI is neutral hydrogen, HII is ionized hydrogen (one proton, no electron). What strips the electron is ultraviolet light from nearby massive stars.

Only O- and B-stars — the most massive, hottest young stars — produce enough UV above 13.6 eV (the ionization energy of hydrogen) to strip electrons off gas out to distances of tens of light-years. The ionized plasma glows by recombination: a free electron finds a proton, drops into the ground state in steps, and each step spits out a photon. The brightest of those is Hα at 656.3 nm — the deep-red line that makes every Hubble and amateur narrowband image of these regions look the same colour.

The Orion Nebula (M42) is the nearest HII region to Earth, just 1,300 light-years away, ionized by the four massive stars of the Trapezium — of which θ¹ Orionis C, an O6 type, provides most of the UV. M8 Lagoon, M16 Eagle, M17 Omega, NGC 7000 North America — every pink-red blob in long-exposure photos of the Milky Way is an HII region. In the Large Magellanic Cloud, the Tarantula Nebula (NGC 2070) is ionized by an entire cluster of Wolf-Rayets and O-stars; if it were as close as M42, it would cast shadows at night.

HII regions are where stars are being born — the hot young stars ionizing them formed out of the same cold molecular cloud a few million years earlier, and deeper in the cloud, thousands more stars are still collapsing out of the gas. The ionization and the star formation are coupled; UV from the new stars eventually blows apart the cloud that made them, ending the party.

The space between stars isn't empty. Per cubic centimetre, you'll find roughly one hydrogen atom, plus traces of heavier elements, plus tiny grains of silicate and carbon dust — soot, basically, measured in nanometres. There's almost nothing there. But multiply almost-nothing by a few thousand light-years and you have a lot of stuff.

Dust scatters and absorbs starlight, and — crucially — it does so more strongly at short wavelengths than at long wavelengths. Blue light bounces off a grain more efficiently than red light does. The net effect is that a star behind a dust cloud looks both dimmer and redder than it should. Astronomers call this interstellar extinction, and the reddening is how we detect it.

The total extinction in visual light is written A_V, in magnitudes. The selective extinction — how much redder the star appears — is written E(B-V), the difference between how the star is dimmed in the blue B band and the visual V band. Their ratio is a fundamental constant of the interstellar medium:

R_V = A_V / E(B-V) ≈ 3.1

R_V ≈ 3.1 is astonishingly universal. Sight lines through the Milky Way disk, through the LMC, through external galaxies — they almost all come out around 3.1. The implication is that dust grains and their size distributions are more or less the same everywhere the interstellar medium has had time to process them. For amateur observers, the practical upshot is that when you read that a star is 30% dimmer than its intrinsic brightness, that same star is about 10% redder than it should be — and both effects are tied together by the same 3.1.

Dust also matters morphologically. Those black gashes you see cutting across the disk of M104 Sombrero or NGC 891 aren't empty holes in the galaxy. They are equatorial dust lanes silhouetted against the billions of stars behind them. Every spiral galaxy has them; we mostly notice them when we see the galaxy edge-on.

Neutral hydrogen is mostly invisible. A cold hydrogen atom sitting quietly in interstellar space doesn't produce Hα, doesn't show up in optical surveys, doesn't block starlight nearly as effectively as dust. For decades, astronomers had to assume there was gas out there and squint at indirect evidence.

Then in 1944, a young Dutch astronomer named Henk van de Hulst, working in occupied Holland, predicted that hydrogen does produce one very specific photon: when the spin of the electron in a ground-state atom flips from being aligned with the proton's spin to being anti-aligned, it releases a 5.9 microelectronvolt photon. Wavelength: 21.106 centimetres. Frequency: 1,420 megahertz. In the microwave band, cutting clean through every dust cloud in the galaxy.

In 1951, three independent groups — at Harvard, in the Netherlands, and in Australia — detected the line. It cracked Milky Way astronomy wide open. Before 1951, nobody knew whether our galaxy had spiral arms. Within a decade, radio maps of 21-cm emission had traced the arms straight through the dust-blocked galactic centre — because radio waves don't care about dust. Rotation curves measured from 21-cm Doppler shifts later became the first unambiguous evidence for dark matter: the outer disks of galaxies rotate too fast for the visible mass to hold them together.

Today, 21-cm observations are the backbone of extragalactic kinematic studies. That rotation curve you need for Tully-Fisher? Measured at 21 cm. Surveys like HIPASS and ALFALFA have catalogued tens of thousands of galaxies purely from their neutral hydrogen signatures — including many "dark" dwarf galaxies with very few stars but plenty of gas. Radio astronomy's humblest line is astrophysics's workhorse.

Zoom out once more. Galaxies themselves cluster. A single cluster can contain a thousand galaxies embedded in a vast cloud of hot, X-ray-emitting gas at 10–100 million Kelvin — the intracluster medium. How do we detect that gas if it's not bright enough to see directly?

One elegant method: watch the cosmic microwave background shine through it. The CMB is the 2.7-Kelvin glow of the universe 380,000 years after the Big Bang, filling every direction on the sky. When CMB photons pass through the hot electrons of a galaxy cluster's intracluster medium, some photons get kicked up to higher energy by inverse Compton scattering. This shifts the local CMB spectrum slightly — deficit below the peak, excess above — producing a small but unmistakable dip at microwave wavelengths in the direction of the cluster.

Rashid Sunyaev and Yakov Zel'dovich predicted this in 1969–1972. The Sunyaev-Zel'dovich effect has three beautiful properties. First, it is redshift-independent — the fractional brightness change is the same whether the cluster is at z = 0.1 or z = 2, because the CMB shines through all of them. Second, it is additive rather than integrated along line of sight — the signal depends on the line-integral of electron pressure, not on luminosity dimming with distance. Third, it is a direct thermometer of the intracluster gas.

Surveys like the South Pole Telescope and the Atacama Cosmology Telescope now routinely find galaxy clusters by their SZ signatures alone, including clusters too distant for their stars to be seen. In a sense, the SZ effect lets the ancient light of the Big Bang become a backlight, silhouetting every massive structure in the universe on its way to us.

That is the arc of this article. From the geometry of a single galaxy's shape, through the kinematics of its rotation, to the chemistry of its stars, the physics of its forbidden-line gas, the optics of its dust, and finally the shadow it casts against the oldest light we know — every tool is a different way of asking the same question: what is out there, and how did it get that way?