Astronomy › Section P · Stars from adolescence to old age › Topic P.3

Checking stellar evolution against clusters

Computer models make bold predictions about how stars change over millions and billions of years. Star clusters let astronomers check them: plot a cluster's stars on an H-R diagram, and the pattern should match the models for a single age.

Turning theory into a test

Models of stellar evolution predict how a star of any mass changes with time. To test them, astronomers compute where stars of every mass should sit on the H-R diagram at one particular age, then draw a curve through those points. This curve, a snapshot of a whole population at one moment, is called an isochrone, from Greek words meaning "equal time". As "Star clusters" explained, the members of a cluster share an age, a composition and a distance. So a real cluster's H-R diagram should match an isochrone for one age. If no single age fits, something in the theory or the data is wrong.

A cluster only a few million years old

In a very young cluster, theory predicts a split. The massive stars, which form fastest, should already be on the main sequence, shining by hydrogen fusion. The low-mass stars, which take tens of millions of years to form, should still sit to the upper right of the main sequence, larger and cooler than they will be, still contracting. That is exactly what is seen. NGC 2264, the Christmas Tree Cluster about 2,500 light-years away, is a few million years old and still wrapped in the gas it formed from; its H-R diagram follows the computed pattern closely. The cluster in the Orion Nebula looks much the same.

The main-sequence turnoff

As a cluster ages, its most massive stars use up their core hydrogen first and leave the main sequence for the giant region, as described in "From the main sequence to red giants". Stars of lower and lower mass follow in turn. At any moment the main sequence is therefore cut off at the top: above a certain point there are no main-sequence stars left. That point, where stars are just peeling away, is the main-sequence turnoff. The stars at the turnoff have main-sequence lifetimes equal to the cluster's age. As time passes, the turnoff slides steadily down the main sequence towards cooler, fainter, lower-mass stars, like the level of a draining tank.

Worked example: an age from the turnoff

In a cluster, the stars just leaving the main sequence have about 2 solar masses and luminosities of about 16 \(L_\odot\), where \(L_\odot \approx 3.83 \times 10^{26}\) W is the solar luminosity. How old is the cluster? 1. Turnoff stars have lived exactly one main-sequence lifetime. 2. Use the fuel-over-spending estimate from the Sun's 10-billion-year lifetime:

t \approx 10\ \text{billion years} \times \frac{M/M_\odot}{L/L_\odot}

3. Substitute: \(t \approx 10 \times \frac{2}{16} = 1.25\) billion years. 4. Check: detailed models give about 1.1 to 1.2 billion years for 2 solar masses, so the estimate is close. 5. Interpret: the cluster is roughly a billion years old, far younger than the Sun but far older than the Christmas Tree Cluster.

Gaps and giants

In a cluster about 100 to 200 million years old, such as M41, several stars have reached the red-giant region, and between them and the main sequence there is a gap with almost no stars. The gap does not mean stars avoid those temperatures. It means stars cross that part of the diagram so fast, compared with how long they spend on the main sequence or as giants, that we rarely catch one in transit. This is another successful prediction: the models say the crossing is quick, and real clusters show few stars caught in the act. In older clusters, the stages described in "Further evolution of stars" appear as further groups of stars.

Old clusters and globular clusters

After a few billion years, only stars of about one and a half solar masses or less remain on the main sequence. The open cluster M67, roughly 4 billion years old, has a turnoff just above and to the left of the Sun's own position on the diagram, at stars of about 1.2 to 1.3 solar masses. Globular clusters are older still. Their main sequences stop near the Sun's luminosity, around 0.8 to 0.9 solar masses, and above that lies only a well-populated sequence of giants. No luminous blue main-sequence stars remain at all. Fitting isochrones gives globular-cluster ages of about 11 to 13 billion years, making them among the oldest objects known and, as a group, far older than the open clusters.

Worked example: comparing two clusters

Cluster A has its turnoff at stars of 3 solar masses; cluster B at stars of 1.2 solar masses. Roughly how many times older is B? 1. Main-sequence lifetime scales roughly as \(t \propto M^{-2.5}\), as shown in "From the main sequence to red giants". 2. The age ratio is

\frac{t_B}{t_A} = \left(\frac{M_A}{M_B}\right)^{2.5} = \left(\frac{3}{1.2}\right)^{2.5}

3. Evaluate: \(2.5^{2.5} = 9.9\), so B is about 10 times older. 4. Check the direction: B's turnoff is lower, at less massive stars, so B must be older. 5. Caution: the power law is only approximate, so real ages come from fitting detailed isochrones.

Checks, puzzles and fixes

Testing works because it can fail. In the mid-1990s, globular-cluster ages of about 15 billion years seemed older than the universe, an obvious contradiction. Better distances to the clusters, from the Hipparcos satellite, and improved models brought the ages down to about 12 billion years, comfortably below today's figure for the universe of about 13.8 billion years. Odd stars are tested too. Blue stragglers sit above a cluster's turnoff and look too young, but they are explained by binary stars: one star gaining mass from a companion, or two stars merging. Fits also need the right composition, since stars poor in heavy elements are hotter and bluer at a given mass, and the right distance.

What the fits tell us

Hundreds of clusters, from a million years old to more than 10 billion, can each be fitted by an isochrone of a single age, using the same physics. That agreement is strong evidence that the models of stellar structure and evolution are basically right. It also turns clusters into clocks for the galaxy. The youngest clusters and associations, at one to a few million years, mark star formation happening now; open clusters span up to a few billion years; and globular clusters record the galaxy's earliest era. Since the ESA Gaia mission's data releases from 2018 onwards, cluster H-R diagrams have become far sharper, tightening these tests further.

Common mistakes to avoid

First, a lower turnoff means an older cluster, not a younger one. Second, the stars below the turnoff are not younger than the others; all members are about the same age, but low-mass stars evolve more slowly. Third, a gap in a cluster's H-R diagram does not mean stars skip those temperatures; they cross them quickly. Fourth, blue stragglers do not disprove the theory; binary interactions explain them. Fifth, the \(M^{-2.5}\) rule gives only rough ages, especially for low-mass stars.

Key takeaways

An isochrone shows where stars of all masses sit at one age; a cluster's H-R diagram should match one. Very young clusters show massive stars on the main sequence and low-mass stars still contracting above it. As clusters age, the main-sequence turnoff moves down, and turnoff stars have lifetimes equal to the cluster's age, roughly \(10\) billion years \(\times \frac{M}{L}\) in solar units. Gaps mark fast stages. Globular clusters turn off near 0.8 to 0.9 solar masses and are about 11 to 13 billion years old, younger than the 13.8-billion-year universe. Blue stragglers and old puzzles have been explained.

Flashcards

Flashcard

Isochrone

Curve on the H-R diagram showing stars of all masses at one age

Means "equal time".

Test your understanding

Question 1 of 5

In a young cluster, the low-mass stars lie to the upper right of the main sequence. What does this show?

Question 2 of 5

A cluster's turnoff stars have 2.5 solar masses and luminosities of about 50 L☉. Using t ≈ 10 billion years × M/L, how old is the cluster?

Question 3 of 5

Cluster X has its turnoff at 4 solar masses and cluster Y at 1 solar mass. Using t ∝ M⁻²·⁵, about how many times older is Y than X?

Question 4 of 5

Why does a cluster's H-R diagram often show a gap with few stars between the main sequence and the red giants?

Question 5 of 5

Which observation would most clearly mark a cluster as a very old globular cluster rather than a young one?

← Previous topicNext topic →