Stellar Populations I, II, and III

Every galaxy carries its history written in starlight, and the code is deceptively simple: age and chemistry. When Walter Baade resolved the Andromeda Galaxy into individual stars in 1944, he saw two crowds — bright blue stars strewn through the disk and spiral arms, and a redder, smoother throng filling the bulge and the globular clusters. He named them Population I and Population II. What looked at first like a difference in the brightest visible stars turned out to be a far deeper division: a division in when the stars were born and from what gas. The Big Bang made essentially only hydrogen, helium, and a trace of lithium; every heavier element — everything astronomers loosely call a "metal" — was forged inside stars and flung back into space by winds and supernovae. So each stellar generation enriches the gas that forms the next, and metallicity runs forward in time like a clock that can only tick up. That single fact organizes the whole subject. The oldest stars formed from nearly pristine gas and are the most metal-poor; the youngest formed from heavily recycled gas and are metal-rich. Population II is old, metal-poor, and lives in a hot spheroidal halo on plunging orbits; Population I is young, metal-rich, and rides the cold rotating disk; and the hypothetical Population III — the first stars, born with exactly zero metals — begins the chain. Read a star's metallicity and you read its rough age; read the turnoff of a whole cluster and you date it to better than a billion years — which is why globular clusters, at 12–13 Gyr, help bound the age of the Universe itself.

Key takeaways

Stellar populations classify stars by metallicity Z (the mass fraction in elements heavier than helium) and age, tracing the chemical enrichment history of a galaxy.
Population I is young and metal-rich ( $Z\sim0.02$ , near solar), living in disks and spiral arms; Population II is old and metal-poor ( $Z\sim10^{-3}$ to $10^{-4}$ ), in the halo, globular clusters, and bulge.
Metallicity is measured spectroscopically as $[\mathrm{Fe/H}]=\log_{10}(N_\mathrm{Fe}/N_\mathrm{H})_\star-\log_{10}(N_\mathrm{Fe}/N_\mathrm{H})_\odot$ ; Pop I has $[\mathrm{Fe/H}]\gtrsim-0.5$ , Pop II has $[\mathrm{Fe/H}]\lesssim-1$ .
Population III are hypothetical first stars formed from primordial H/He gas ( $Z=0$ ); lacking metal cooling they were likely very massive, short-lived, and made the first heavy elements.
The main-sequence turnoff is a sensitive clock: since $t_\mathrm{MS}\propto M^{-2.5}$ , an older cluster has a fainter, redder turnoff — a turnoff at $0.9\,M_\odot$ gives $\approx13$ Gyr.
Populations correlate with kinematics: Pop I is dynamically cold ( $\sigma\sim20\ \mathrm{km\,s^{-1}}$ ) on circular disk orbits, Pop II is dynamically hot ( $\sigma\sim100\text{–}150\ \mathrm{km\,s^{-1}}$ ) in a spheroid.
Key idea: metallicity and age are the two axes of stellar classification — because stars enrich the gas from which later stars form, low metallicity is a proxy for great age, and the sequence Pop III → Pop II → Pop I traces cosmic chemical enrichment while the main-sequence turnoff dates each population.

The core insight

Metallicity and age are the two axes of stellar classification: because stars enrich the gas from which later stars form, low metallicity is a proxy for great age. The sequence Population III → II → I traces cosmic chemical enrichment, while the main-sequence turnoff in a colour–magnitude diagram serves as the clock that dates each population.

Where this lives today

The population concept is now the working language of "galactic archaeology," and its sharpest tools are enormous stellar surveys. Gaia has measured positions, motions, and parallaxes for over a billion stars, while spectroscopic surveys such as APOGEE, GALAH, and LAMOST add metallicities and abundance patterns — $[\mathrm{Fe/H}]$ , $[\alpha/\mathrm{Fe}]$ , and dozens of elements — for millions more. Together they let astronomers "chemically tag" stars to their birth environments and reconstruct the Milky Way's assembly: the discovery of the Gaia–Enceladus merger, a dwarf galaxy swallowed some 10 Gyr ago, came from a coherent population of accreted stars with distinctive kinematics and low $[\alpha/\mathrm{Fe}]$ . At the extreme metal-poor frontier, dedicated searches hunt stars with $[\mathrm{Fe/H}]<-4$ (a few now known below $-6$ ) as living fossils that may be the chemical descendants of Population III. And at cosmic dawn, JWST is now probing galaxies within the first few hundred million years, testing whether truly metal-free Population III stars can be caught in the act — the first light that closes the chain the Sun sits at the end of.

The physical picture

Think of a galaxy as a set of geological strata. The deepest, oldest layer is the metal-poor halo, laid down when the gas was nearly pristine; on top of it, infall and repeated enrichment build the metal-rich disk. Reading a star's metallicity is like reading its depth in that record — and it works only because enrichment runs one way in time. A star can only be assembled from gas that earlier stars enriched, so a pristine ( $Z=0$ ) star must be first-generation: metallicity is a clock whose hands can never run backward. Crucially, populations remember the dynamical state of the galaxy at the moment of their birth. Gas radiates away its energy and settles into a thin, cold, rotating disk, so stars born late inherit orderly near-circular orbits. Stars born early, before that dissipation had finished, froze in the hot, pressure-supported, quasi-spherical velocity field of the collapsing protogalaxy. That is why chemistry and kinematics track each other so tightly — and the first figure shows the two crowds side by side.

Figure

The Milky Way seen edge-on: cold Population I stars with iron abundance Fe/H ≳ −0.5 ride the thin disk on near-circular orbits with dispersion σ ≈ 20 km/s, while old metal-poor Population II stars and globular clusters with Fe/H ≲ −1 fill a hot spheroidal halo with σ ≈ 100–150 km/s

To turn "how long ago a star formed" into an actual date we need a clock, and the colour–magnitude diagram of a coeval cluster provides it. At any moment exactly the stars whose main-sequence lifetime equals the cluster's age are peeling away from the main sequence and swelling into giants. Because the lifetime falls steeply with mass, $t_\mathrm{MS}\propto M^{-2.5}$ , the mass at this "turnoff" shrinks measurably as the cluster ages, and with it the turnoff luminosity and colour. The second figure makes that clock explicit, with the numbers a globular cluster actually returns.

Figure

The main-sequence turnoff as a clock in a globular-cluster colour–magnitude diagram: stars above the turnoff have already burned out, so the surviving turnoff mass fixes the age through t ≈ 10 Gyr × (M/M⊙)^−2.5, and a turnoff at 0.90 M⊙ gives ≈ 13 Gyr

The core equations

Metallicity is quoted logarithmically, relative to the Sun, so a single number says how a star's iron content compares to solar. The bracket notation is a base-ten logarithm of the abundance ratio, so each unit is a factor of 10.

[\mathrm{Fe/H}] \equiv \log_{10}\!\left(\frac{N_\mathrm{Fe}}{N_\mathrm{H}}\right)_\star - \log_{10}\!\left(\frac{N_\mathrm{Fe}}{N_\mathrm{H}}\right)_\odot \qquad (1)

where: $[\mathrm{Fe/H}]$ — logarithmic iron abundance relative to the Sun, so $0$ is solar and $-1$ is one-tenth solar; $N_\mathrm{Fe},N_\mathrm{H}$ — number densities of iron and hydrogen; $\odot$ — solar reference. Reading: a halo star at $[\mathrm{Fe/H}]=-2$ has $1/100$ the solar iron-to-hydrogen ratio. The theorists' bookkeeping variable is not iron but the total mass fraction in metals. Because every gram of stellar material is hydrogen, helium, or something heavier, the three mass fractions sum to one.

Z = 1 - X - Y \qquad (2)

where: $Z$ — mass fraction in elements heavier than helium ("metals"); $X$ — hydrogen mass fraction; $Y$ — helium mass fraction. Reading: with $X_\odot\approx0.7381$ and $Y_\odot\approx0.2485$ , the solar metal fraction is $Z_\odot\approx0.0134$ . A single metallicity number hides a second, independent quantity: the ratio of different metals, which records how fast the enrichment happened. The α-elements (O, Mg, Si, Ca, Ti) are made almost entirely by prompt core-collapse supernovae, whereas much of the iron comes from delayed Type Ia supernovae.

[\alpha/\mathrm{Fe}] = \log_{10}\!\left(\frac{N_\alpha}{N_\mathrm{Fe}}\right)_\star - \log_{10}\!\left(\frac{N_\alpha}{N_\mathrm{Fe}}\right)_\odot \qquad (3)

where: $N_\alpha$ — combined abundance of the α-elements; $[\alpha/\mathrm{Fe}]$ — a tracer of the enrichment timescale, enhanced ( $>0$ ) when core-collapse supernovae dominate over Type Ia. Reading: high $[\alpha/\mathrm{Fe}]$ means fast enrichment, before Type Ia iron could catch up. The clock that dates a population comes from stellar structure. A main-sequence star lives as long as its fuel lasts divided by the rate it burns it; combining fuel $\propto M$ with the mass–luminosity law $L\propto M^{3.5}$ gives a lifetime that plunges with mass.

t_\mathrm{MS} \approx 10\ \mathrm{Gyr}\left(\frac{M}{M_\odot}\right)\!\left(\frac{L_\odot}{L}\right) \approx 10\ \mathrm{Gyr}\left(\frac{M}{M_\odot}\right)^{-2.5} \qquad (4)

where: $t_\mathrm{MS}$ — core-hydrogen-burning lifetime; $M,L$ — stellar mass and luminosity. Reading: a $10\,M_\odot$ star lives only $\sim3$ Myr, flagging recent star formation, while a $0.9\,M_\odot$ star lives $\sim13$ Gyr. Where do the metals come from? The simplest chemical-evolution model treats a galaxy as a closed box of gas that turns into stars, each generation returning fresh metals with a fixed yield. Integrating that balance gives metallicity as a logarithm of the remaining gas fraction.

Z(t) = -y\,\ln\!\left(\frac{g(t)}{g_0}\right) = y\,\ln\!\left(\frac{1}{\mu}\right) \qquad (5)

where: $y$ — nucleosynthetic yield, the mass of new metals produced per unit mass locked into long-lived stars; $g,g_0$ — current and initial gas mass; $\mu=g/(g+s)$ — the gas fraction. Reading: as gas runs out ( $\mu\to0$ ) the closed box predicts unbounded metallicity — real galaxies escape by accreting fresh gas. Finally, the population divide shows up in motion as sharply as in chemistry: the random-velocity "temperature" of the cold disk is several times smaller than that of the hot halo.

\sigma_\mathrm{thin} \sim 20\ \mathrm{km\,s^{-1}} \ \ll\ \sigma_\mathrm{halo} \sim 100\text{–}150\ \mathrm{km\,s^{-1}} \qquad (6)

where: $\sigma$ — stellar velocity dispersion, the spread of random motions; thin and halo denote the thin-disk Population I and spheroidal Population II components. Reading: the disk is rotationally supported and quiet, the halo pressure-supported and hot.

Derivation AThe main-sequence lifetime scales as mass to the minus five-halves

The goal is to show, from nothing more than "fuel divided by burn rate," why a star's time on the main sequence collapses so steeply with mass that a factor-of-two heavier star lives roughly six times less long. This scaling is the engine behind every age we quote for a stellar population: it makes the main-sequence turnoff a usable clock, because the turnoff mass is the mass whose lifetime equals the cluster's age, and the steepness of the relation makes that mass a sharp pointer rather than a blur.

Step 1. A star shines by fusing hydrogen in its core, and only a fixed fraction (roughly the inner $10\%$ of its mass) is ever available as fuel, so the total energy reservoir is $E_\mathrm{avail}\propto M c^2 \propto M$ . The lifetime is this reservoir divided by the rate it is spent, the luminosity $L$ : $t_\mathrm{MS}\propto E_\mathrm{avail}/L \propto M/L$ . Why: fuel scales with mass but the "burn rate" is set independently by how brightly the star shines, so the two must be combined rather than assumed equal.

Step 2. Main-sequence stars obey a steep mass–luminosity relation, $L\propto M^{3.5}$ , because a more massive star must support a larger weight and therefore runs a hotter, far more luminous core. Substituting gives $t_\mathrm{MS}\propto M\cdot M^{-3.5}=M^{-2.5}$ . Why: the luminosity climbs much faster than the fuel supply, so heavier stars squander their larger reservoirs disproportionately quickly.

Step 3. Calibrate on the Sun. The Sun will spend about $10$ Gyr on the main sequence, so $t_\mathrm{MS}\approx10\ \mathrm{Gyr}\,(M/M_\odot)^{-2.5}$ . A $2\,M_\odot$ star then lives $10\times2^{-2.5}\approx1.8$ Gyr and a $0.9\,M_\odot$ star $\approx13$ Gyr. Why: fixing the single free constant to a known star turns a proportionality into a predictive formula, and the numbers explain why massive stars are absent from old populations.

Go deeper · Derivation B · 2 minDating a coeval cluster from its main-sequence turnoff

Derivation BDating a coeval cluster from its main-sequence turnoff

The goal is to convert "older clusters have fainter turnoffs" into a quantitative age, by identifying the one turnoff mass in a coeval population whose main-sequence lifetime exactly equals the cluster's age and inverting the lifetime law to read that age off. This is the single most important dating technique in stellar populations: it is how globular clusters are placed at 12–13 Gyr, how the disk is shown to be younger than the halo, and how stellar astrophysics delivers an independent lower bound on the age of the Universe.

Step 1. A star cluster is essentially coeval — all its members formed at the same time $t=0$ from one cloud — so every star shares a single age $t$ . Any star more massive than some threshold has $t_\mathrm{MS}(M)<t$ and has already exhausted its core hydrogen, evolving off onto the subgiant and giant branches. Why: the common birth time is what makes the turnoff a single well-defined point rather than a smear, so one mass marks the boundary between survivors and evolved stars.

Step 2. That threshold is the turnoff mass $M_\mathrm{TO}$ , defined by $t_\mathrm{MS}(M_\mathrm{TO})=t$ : the most massive star still burning core hydrogen. Using the lifetime law, $t=10\ \mathrm{Gyr}\,(M_\mathrm{TO}/M_\odot)^{-2.5}$ . Why: setting lifetime equal to age picks out the exact mass at the bend in the colour–magnitude diagram, tying an observable feature to a physical clock.

Step 3. Invert to get either quantity from the other: $M_\mathrm{TO}=M_\odot\,(t/10\ \mathrm{Gyr})^{-1/2.5}$ . For $t=12$ Gyr this gives $M_\mathrm{TO}\approx0.93\,M_\odot$ ; conversely a measured $M_\mathrm{TO}\approx0.90\,M_\odot$ returns $t\approx13$ Gyr. Why: the same relation runs both ways, so a turnoff mass read from a diagram becomes an age, and a target age predicts where to look for the turnoff.

Step 4. In practice one fits theoretical isochrones — curves of constant age at fixed metallicity — to the observed colour–magnitude diagram, because the turnoff colour and luminosity depend on $Z$ as well as age. The result: globular clusters give $M_\mathrm{TO}\sim0.8$ – $0.9\,M_\odot$ , ages $\sim12$ – $13$ Gyr, older than the disk and comfortably below the $13.8$ Gyr cosmic age. Why: real colours shift with metallicity, so a self-consistent fit rather than a single-point read yields a defensible age and its cross-check against cosmology.

Go deeper · Derivation C · 2 minThe closed-box model and the G-dwarf problem

Derivation CThe closed-box model and the G-dwarf problem

The goal is to derive the metallicity distribution that a self-enriching galaxy would produce if it were a sealed box of gas, then confront that prediction with the metallicities of long-lived dwarfs in the solar neighbourhood — a confrontation the simple model loses badly, and whose failure is one of the cleanest pieces of evidence that the Galactic disk grows by continuous infall rather than evolving in isolation. Long-lived G dwarfs are ideal witnesses because they preserve, essentially unchanged, the metallicity of the gas from which they formed.

Step 1. Assume no inflow or outflow, instantaneous recycling, and a constant yield $y$ . Conservation of mass gives $\mathrm{d}g=-\mathrm{d}s$ (gas turns into stars), and the metal budget evolves as $\mathrm{d}(Zg)=y\,\mathrm{d}s-Z\,\mathrm{d}s$ : fresh metals are produced at rate $y$ per unit star formation, while metal-bearing gas is locked away at the ambient $Z$ . Why: the closed-box assumptions strip the problem to a single self-contained reservoir, so every metal atom is either made in situ or removed with the gas it sits in.

Step 2. Expanding $\mathrm{d}(Zg)=Z\,\mathrm{d}g+g\,\mathrm{d}Z$ and cancelling the $Z\,\mathrm{d}s$ terms leaves $g\,\mathrm{d}Z=-y\,\mathrm{d}g$ . Integrating from pristine gas ( $Z=0,\ g=g_0$ ) gives $Z=-y\,\ln(g/g_0)=y\,\ln(1/\mu)$ . Why: the logarithm is the signature of a box enriching itself as it depletes — metallicity climbs without bound as the gas fraction $\mu\to0$ , already a warning that no real accreting galaxy behaves this way.

Step 3. The mass of stars formed while metallicity was below $Z$ equals the gas consumed by then, $s(<Z)=g_0-g=g_0\,(1-e^{-Z/y})$ . Normalising by the final population gives the cumulative metallicity distribution $\dfrac{s(<Z)}{s_\mathrm{tot}}=\dfrac{1-e^{-Z/y}}{1-e^{-Z_1/y}}$ , where $Z_1$ is the present metallicity. Why: inverting $Z=-y\ln(g/g_0)$ to $g/g_0=e^{-Z/y}$ turns "how much gas is left" into "how many stars are below a given metallicity," which is exactly the observable G-dwarf histogram.

Step 4. With $Z_1\approx y$ the model predicts a large fraction of stars at very low metallicity — of order a third to a half below quarter-solar. Solar-neighbourhood G dwarfs instead show a sharp deficit of metal-poor stars: the G-dwarf problem. The resolution is that the disk is not a closed box; continuous infall of near-primordial gas dilutes and regulates the metallicity near an equilibrium value, erasing the predicted metal-poor tail. Why: the mismatch is not a small correction but a qualitative failure, so it points to a missing ingredient — accretion — rather than a tweak of the yield.

Limiting cases

The population framework spans an enormous range of metallicity and age; the useful limiting regimes are:

Case	Result	Meaning
Population III (primordial)	$Z=0,\quad[\mathrm{Fe/H}]=-\infty$	No metals, so cooling relies only on H₂ and HD; the high Jeans mass favours very massive ( $\sim10$ – $100\,M_\odot$ ), short-lived stars. None yet observed directly.
Extreme Population II (halo)	$[\mathrm{Fe/H}]\lesssim-3,\quad t\sim13\ \mathrm{Gyr}$	Ultra metal-poor halo stars; enhanced $[\alpha/\mathrm{Fe}]$ signals rapid early enrichment by core-collapse supernovae before Type Ia iron arrived.
Population I (solar disk)	$Z\approx Z_\odot=0.0134,\quad[\mathrm{Fe/H}]\approx0$	Young thin-disk stars including the Sun; near-circular orbits, low dispersion, concentrated in the spiral arms and plane.
Massive-star lifetime limit	$t_\mathrm{MS}\propto M/L\propto M^{-2.5}$	High-mass stars ( $M\gg M_\odot$ ) live only Myr, so their presence marks recent star formation — a Population I signature.
Gas-exhausted limit	$\mu\to0,\quad Z\to\infty$ (closed box)	Unphysical divergence of the simple model; real galaxies saturate near $Z\approx y$ .

Common trapLow metallicity always means old age, and the populations are three sharp boxes

Both halves are traps. First, "metal" to an astronomer means anything heavier than helium — carbon, nitrogen, oxygen, neon all count — not the chemist's metallic elements, so a nitrogen-rich, iron-poor star is still "metal-poor." Second, the populations are not discrete bins but a continuum: real galaxies show a smooth spread in $[\mathrm{Fe/H}]$ and age, with thick-disk and "intermediate Population II" stars filling the gap, and the ratio $[\alpha/\mathrm{Fe}]$ carries independent information — the enrichment timescale — that a single metallicity cannot. Third, low metallicity is only a statistical proxy for age: accreted dwarf galaxies can deposit genuinely young metal-poor stars, and a few metal-rich stars form early in dense, rapidly enriched regions. Use metallicity, kinematics, α-enhancement, and the turnoff together — never one alone.

Worked examples

1 · Reading a halo globular cluster (fully worked)

A globular cluster has $[\mathrm{Fe/H}]=-1.5$ and a main-sequence turnoff at $M_\mathrm{TO}\approx0.90\,M_\odot$ . Find its iron abundance relative to the Sun, estimate its age, and classify it.

Step 1. Convert the metallicity. By equation (1), $[\mathrm{Fe/H}]=-1.5$ means $N_\mathrm{Fe}/N_\mathrm{H}=10^{-1.5}=0.0316$ of the solar ratio — about 1/32 of the solar iron abundance, deeply metal-poor and characteristic of the halo.

Step 2. Date it from the turnoff. Using equation (4), $t\approx10\ \mathrm{Gyr}\times(0.90)^{-2.5}=10\times1.301=\mathbf{13.0\ Gyr}$ — everything heavier than $0.9\,M_\odot$ has already died.

Step 3. Classify and cross-check. A $13$ Gyr age with $[\mathrm{Fe/H}]=-1.5$ is a textbook old, metal-poor Population II system — older than the thin disk, yet below the cosmic age $13.8$ Gyr, so it does not overconstrain cosmology. With an expected enhanced $[\alpha/\mathrm{Fe}]$ , the picture is of gas enriched rapidly and early, then frozen into stars before the disk assembled. Result: 1/32 solar iron, age ≈ 13 Gyr, Population II.

2 · The G-dwarf problem in numbers

Take the solar-neighbourhood disk to have present metallicity $Z_1=Z_\odot=0.0134$ and a current gas fraction $\mu\approx0.15$ . Using the closed-box model, estimate the effective yield, then the fraction of long-lived stars predicted to be more metal-poor than quarter-solar, and compare with observation.

Step 1. Fit the yield from present-day values. From equation (5), $Z_1=y\,\ln(1/\mu)$ , so $y=Z_1/\ln(1/\mu)=0.0134/\ln(1/0.15)=0.0134/1.897=\mathbf{7.1\times10^{-3}}$ — the single free parameter of the model.

Step 2. Predict the metal-poor fraction. Quarter-solar is $Z=0.25\,Z_\odot=3.35\times10^{-3}$ . The cumulative distribution from Derivation 3 gives

\frac{s(<Z)}{s_\mathrm{tot}}=\frac{1-e^{-Z/y}}{1-e^{-Z_1/y}}=\frac{1-e^{-0.474}}{1-e^{-1.897}}=\frac{0.378}{0.850}=\mathbf{0.44}.

The closed box predicts about 44% of surviving stars more metal-poor than quarter-solar ( $[\mathrm{Fe/H}]\lesssim-0.6$ ).

Step 3. Compare. Observed solar-neighbourhood G dwarfs show only a few percent that metal-poor — a dramatic deficit relative to the predicted $\sim44\%$ . This is the G-dwarf problem: the disk cannot be a closed box. Continuous infall of near-primordial gas pins the metallicity near equilibrium from early on, so most stars form already moderately enriched and the metal-poor tail never accumulates. Result: model predicts ≈ 44% metal-poor, observation gives a few percent — the disk is an open, accreting system.

Physics of the portrait

Stellar populations reorganized a whole field the moment they were seen clearly — quite literally, by one observer at one telescope during a wartime blackout.

Origins · 1 minWalter Baade and how this was discovered

Walter Baade

1893–1960 · resolving stellar populations in the Andromeda Galaxy

Walter Baade was a German-born observational astronomer whose patient work on the brightest resolvable stars in nearby galaxies reshaped both stellar and extragalactic astronomy. During the Second World War, as an enemy alien barred from war work and with wartime blackouts darkening the skies over Los Angeles, Baade had near-exclusive use of the 100-inch Hooker telescope at Mount Wilson. In 1944 he pushed it to resolve, for the first time, individual stars in the central bulge of the Andromeda Galaxy, and recognized that the bright blue stars of the spiral arms (Population I) and the redder stars of the bulge and globular clusters (Population II) were two physically distinct groups — a distinction we now read as age and metallicity. The consequences were enormous: in 1952 he showed that the two populations follow different Cepheid period–luminosity relations, and correcting the mistake roughly doubled the extragalactic distance scale and the estimated age of the Universe. He also, with Fritz Zwicky, coined the term "supernova." Baade left few papers, but an extraordinary eye and a habit of taking the sky exactly as it was.

History

1944

— Baade resolves Populations I and II in the Andromeda Galaxy

1952

— Baade finds separate Cepheid relations for the two populations, doubling the distance scale

1957

— Burbidge, Burbidge, Fowler and Hoyle detail the stellar nucleosynthesis of the elements

1962

— Eggen, Lynden-Bell and Sandage tie halo-star kinematics to low metallicity in a monolithic collapse

1978

— Searle and Zinn argue the outer halo instead grew by accretion of many fragments

Going further

Texts. Carroll & Ostlie, An Introduction to Modern Astrophysics, 2nd ed., Ch. 24 (the Milky Way) and Ch. 25 (galaxies) · Binney & Tremaine, Galactic Dynamics, 2nd ed., Ch. 1 and 4 · Mo, van den Bosch & White, Galaxy Formation and Evolution, Ch. 10 (chemical evolution) · Ryden, Introduction to Cosmology, 2nd ed., on structure formation and Big Bang nucleosynthesis · Dodelson, Modern Cosmology, 2nd ed., on primordial abundances · Baade, W. 1944, ApJ 100, 137 — the original two-population identification. The solar values $X_\odot\approx0.7381$ , $Y_\odot\approx0.2485$ , and $Z_\odot\approx0.0134$ follow directly from $Z=1-X-Y$ . The closed-box yield $y=Z_\odot/\ln(1/0.15)\approx7.1\times10^{-3}$ gives a predicted $\approx44\%$ metal-poor fraction below quarter-solar. The turnoff clock, calibrated to $t_\odot\approx10$ Gyr via $t_\mathrm{MS}\propto M^{-2.5}$ , returns $M_\mathrm{TO}=0.90\,M_\odot\Rightarrow t\approx13$ Gyr and $t=12\ \mathrm{Gyr}\Rightarrow M_\mathrm{TO}\approx0.93\,M_\odot$ . The conversion $[\mathrm{Fe/H}]=-1.5$ gives $N_\mathrm{Fe}/N_\mathrm{H}\approx1/32$ solar. The mass–luminosity index $2.5$ is the standard main-sequence average and is approximate.

Prerequisites. Hertzsprung–Russell diagram and stellar evolution · Nucleosynthesis and stellar interiors · Structure of the Milky Way Galaxy. Leads to. Galactic Halo and Globular Cluster System · Galactic Bulge, Bar, and Central Regions · Interstellar Medium of Galaxies · Elliptical Galaxies and the de Vaucouleurs Profile.

Ask the tutor: ask me to date a specific cluster from its turnoff, convert any $[\mathrm{Fe/H}]$ into an abundance and rough age, or unpack how $[\alpha/\mathrm{Fe}]$ reveals a galaxy's star-formation timescale.

What exactly counts as a "metal" in this context?

To an astronomer, a metal is any element heavier than helium — carbon, nitrogen, oxygen, and neon all count, despite being non-metals to a chemist. The reason is practical: the Big Bang made only hydrogen, helium, and a trace of lithium, so everything else was manufactured later inside stars, and one word captures the distinction between primordial material and stellar products. The total mass fraction in metals is denoted $Z$ ; iron ( $\mathrm{Fe}$ ) serves as the observational stand-in because its many spectral lines are easy to measure, giving the metallicity index $[\mathrm{Fe/H}]$ .

Why does low metallicity usually mean old age — and when does it not?

Because enrichment only runs forward: gas gains metals from dying stars but never loses them back to a pristine state, so on average the earlier a star formed, the poorer the gas it formed from. That makes metallicity a rough clock — but only statistically. Accreted dwarf galaxies can drop genuinely young metal-poor stars into the Milky Way, since those small systems enriched slowly; conversely a few metal-rich stars formed very early in dense, rapidly self-enriching regions like the proto-bulge. This is why astronomers combine metallicity with kinematics, α-element ratios, and the cluster turnoff rather than trusting $[\mathrm{Fe/H}]$ alone.

How do we actually measure a globular cluster's age?

By fitting its colour–magnitude diagram. Because all the cluster's stars formed together, they lie along a single isochrone — a curve of constant age at fixed metallicity. The most age-sensitive feature is the main-sequence turnoff, where stars are just now exhausting their core hydrogen; since $t_\mathrm{MS}\propto M^{-2.5}$ , an older cluster has a lower turnoff mass and a fainter, redder turnoff. In practice one measures the metallicity spectroscopically, computes isochrones for that $Z$ over a range of ages, and slides them until the turnoff and subgiant branch match the data. For the oldest Galactic globulars the best-fit age comes out at $12$ – $13$ Gyr.

What makes Population III stars so different from every later generation?

They formed from gas with literally zero metals, and that changes the physics of star formation itself. Cooling a collapsing cloud is what lets it fragment, and the most effective coolants — carbon, oxygen, dust grains — did not yet exist. With only molecular hydrogen and HD to radiate away heat, primordial clouds stayed warmer, so the Jeans mass (the minimum mass that can collapse) was much larger. The result was a top-heavy population of very massive stars, perhaps tens to a hundred solar masses. Such stars live only a few Myr, so none survive today; we look for their fingerprints instead in the abundance patterns of the most extreme metal-poor halo stars, which may be their chemical descendants.

What is the G-dwarf problem, in one paragraph?

The simplest chemical-evolution model treats a galaxy as a closed box of gas that steadily turns into stars, each generation returning fresh metals. It predicts a large population of very metal-poor, long-lived (G-type) dwarfs that formed early when the gas was still poor. But surveys of solar-neighbourhood G dwarfs find far fewer metal-poor ones than predicted: only a few percent below quarter-solar, against a predicted $\sim40\%$ . The deficit — the G-dwarf problem — tells us the disk is not a closed box. It grows by continuous infall of near-primordial gas, which keeps the metallicity near equilibrium from early times, so most stars form already moderately enriched.

What does the alpha-to-iron ratio tell us that metallicity alone cannot?

It tells us how fast the enrichment happened. The α-elements (oxygen, magnesium, silicon, and the like) are made almost entirely by core-collapse supernovae, which explode within a few million years of a burst of star formation. Iron, by contrast, comes largely from Type Ia supernovae, delayed by hundreds of millions to billions of years. So a population that formed rapidly was enriched before Type Ia iron could catch up, leaving a high $[\alpha/\mathrm{Fe}]$ ; a low or solar value signals slow, extended star formation. Two stars can share the same $[\mathrm{Fe/H}]$ yet have formed on utterly different timescales — and $[\alpha/\mathrm{Fe}]$ is what distinguishes them.