Astronomy › Section V · The Big Bang › Topic V.1

The age of the universe

If galaxies are moving apart today, they were closer together in the past. Running the expansion backwards gives the age of the universe, about 13.8 billion years, and the oldest stars give an independent check.

A universe that cannot stand still

Gravity pulls every galaxy towards every other, so a universe of galaxies at rest would start to fall together. It can avoid collapse only by expanding, much as a ball thrown upwards keeps rising only if it was launched fast enough. In 1917 Albert Einstein applied his general theory of relativity to the whole universe and found the same thing: it could not be static. Believing the universe was unchanging, he added a term, the cosmological constant, to balance gravity. About a decade later, observations of galaxies showed that the universe is in fact expanding, and Einstein is said to have regretted the change. The constant has since returned in a new role.

Running the film backwards

Hubble's law, described in the topic "The expanding universe", says a galaxy's recession speed is proportional to its distance:

v = H_0 d

where \(H_0\) is the Hubble constant. Imagine a film of the expansion played in reverse. The galaxies close in on one another, and at some moment in the past everything we see was packed together at enormous density. That moment marks the start of the expansion, the Big Bang. To date it, astronomers ask how long the galaxies have taken to reach their present separations, just as you could work out when a car set off from its distance and speed.

The Hubble time

Suppose every galaxy has moved at a constant speed since the start. A galaxy now at distance \(d\), moving at speed \(v\), set off a time \(T_0 = \frac{d}{v}\) ago. Substituting \(v = H_0 d\), the distance cancels:

T_0 = \frac{d}{H_0 d} = \frac{1}{H_0}

The same time comes out for every galaxy, near or far, which is exactly what a common starting point requires. This estimate, \(\frac{1}{H_0}\), is called the Hubble time. Because \(H_0\) is given in km/s per megaparsec, finding the time in years needs a unit conversion: 1 Mpc is about \(3.086 \times 10^{19}\) km.

Worked example: the Hubble time for H0 = 70

Find \(\frac{1}{H_0}\) for \(H_0 = 70\) km/s per Mpc. 1. Write \(H_0\) as a speed per distance: \(H_0 = \frac{70\ \text{km/s}}{3.086 \times 10^{19}\ \text{km}}\). 2. The kilometres cancel: \(H_0 = 2.27 \times 10^{-18}\) per second. 3. Invert: \(\frac{1}{H_0} = \frac{3.086 \times 10^{19}}{70} = 4.41 \times 10^{17}\) s. 4. Convert, using 1 year = \(3.156 \times 10^{7}\) s: \(\frac{4.41 \times 10^{17}}{3.156 \times 10^{7}} = 1.40 \times 10^{10}\) years, about 14.0 billion years. 5. Check: a larger \(H_0\) means faster expansion and less time; \(H_0 = 73\) gives 13.4 billion years, and \(H_0 = 67\) gives 14.6 billion years.

The Hubble tension

The Hubble constant is about 70 km/s per Mpc, but its measurements disagree. Measurements based on the cosmic microwave background, combined with the standard model of the universe, give about 67 to 68. Measurements from the distance ladder, using Cepheid variable stars to calibrate type Ia supernovae in nearby galaxies, give about 73. The difference is larger than the stated uncertainties of either method, and other techniques fall in between or on either side. This disagreement, the Hubble tension, is unresolved as of 2026. It may come from hidden measurement errors, or it may hint at physics missing from the standard model. Either way, it shifts the Hubble time by only about 9%.

Slowing down, then speeding up

The Hubble time assumes a constant expansion rate. Gravity should slow the expansion, so the rate would have been faster in the past, and the true age would be less than \(\frac{1}{H_0}\), just as a car that started fast and slowed down took less time than its final speed suggests. In 1998 two teams, one led by Saul Perlmutter and the other including Brian Schmidt and Adam Riess, found the opposite trend. Distant type Ia supernovae looked fainter than expected, meaning they were farther away than a constant or slowing expansion allowed. The expansion has been speeding up. The three shared the 2011 Nobel Prize in Physics.

Dark energy and the 13.8-billion-year age

Whatever drives the acceleration is called dark energy; it is not the same as dark matter, and its nature is unknown. One candidate is Einstein's cosmological constant, an energy of empty space itself, but simple estimates of that energy from quantum theory come out larger than the observed value by a factor often quoted as about \(10^{120}\). For its first several billion years the universe decelerated, because matter was densely packed; after that, as matter thinned out, dark energy took over. The early slowing and later speeding up roughly cancel, so the age from the full model, about 13.8 billion years with an uncertainty of well under a billion years, is close to \(\frac{1}{H_0}\).

Checking with the oldest stars

The universe must be older than anything in it. The oldest stars in the Milky Way are in globular clusters, and models of how stars evolve, together with improved distances to the clusters from satellites, date them to about 12 to 13 billion years. The cooling of the oldest white dwarfs gives similar ages. A third method is radioactive dating of a star. Uranium-238 has a half-life of 4.47 billion years. By comparing how much uranium remains in a very old star with how much is normally made alongside stable elements, astronomers can estimate its age without needing its distance or a model of its interior. All three methods give ages a little less than 13.8 billion years.

Worked example: a uranium clock

Suppose a very old star shows about 14% of the uranium-238 it would have had when it formed. How old is it? The half-life is 4.47 billion years. 1. After \(n\) half-lives, the fraction left is \(\left(\frac{1}{2}\right)^{n}\). 2. Try \(n = 3\): \(\left(\frac{1}{2}\right)^{3} = 0.125\), a little less than 0.14, so \(n\) is just under 3. 3. Try \(n = 2.8\): \(\left(\frac{1}{2}\right)^{2.8} = 0.144\), very close. 4. Age: \(2.8 \times 4.47 = 12.5\) billion years. 5. Check: that is less than 13.8 billion years, so the star formed about 1.3 billion years after the Big Bang. Real measurements of this kind carry uncertainties of a few billion years.

Common mistakes to avoid

First, forgetting to convert megaparsecs to kilometres before inverting \(H_0\); the units must cancel to leave seconds. Second, thinking a larger Hubble constant means an older universe; it means a younger one. Third, assuming the Hubble time is exactly the age; it ignores deceleration and acceleration, which happen nearly to cancel. Fourth, confusing dark energy with dark matter. Fifth, stating the Hubble constant as settled; measurements range from about 67 to 73. Sixth, thinking the oldest stars disagree with the age; they are slightly younger, as they must be.

Key takeaways

Gravity means a universe of galaxies must expand or contract; it is expanding. Running Hubble's law backwards gives the Hubble time, T₀ = 1/H₀, about 14 billion years for H₀ = 70 km/s per Mpc. Measurements of H₀ range from about 67 to 73, the unresolved Hubble tension. Gravity slowed the early expansion; since 1998 we know dark energy is now accelerating it. The two effects nearly cancel, giving an age of about 13.8 billion years. Globular clusters, white dwarfs and uranium dating give stellar ages of about 12 to 13 billion years, consistent with that age.

Flashcards

Flashcard

Cosmological constant

Term Einstein added in 1917 to keep the universe static

Now a candidate for dark energy.

Test your understanding

Question 1 of 5

Taking 1 Mpc = 3.086 × 10¹⁹ km and 1 year = 3.156 × 10⁷ s, what is the Hubble time for H₀ = 80 km/s per Mpc?

Question 2 of 5

If gravity alone had acted, steadily slowing the expansion, how would the true age compare with the Hubble time 1/H₀?

Question 3 of 5

What did the two supernova teams find in 1998?

Question 4 of 5

Which statement best describes the Hubble tension?

Question 5 of 5

Globular-cluster stars are dated to about 12 to 13 billion years. Why does this matter for cosmology?

← Previous topicNext topic →