Density and destiny

Article metrics

Will the Universe perpetually expand, or eventually collapse? Is space flat, or curved? On page 51 of this issue, Perlmutter et al.1 report observations of the explosion of a star so distant that its demise preceded the birth of our planet — observations that begin to answer these grand questions, and promise a final answer soon.

If we inhabit a nice, simple Universe, the destiny of the Universe is determined by its density. The present expansion is being decelerated by gravity, at a rate governed only by the average density of matter, ΩM. If ΩM < 1, space is negatively curved (like a saddle), and will expand for ever. ΩM > 1 means positive curvature (like a sphere) and eventual collapse. ΩM = 1 is supreme simplicity: space is flat and expansion slows asymptotically towards zero.

But empty space may have an energy density of its own, measured by the cosmological constant2, ΩΛ. Introducing this complicates matters, but there are reasons for doing so. Until recently, the main one was the ‘cosmic age problem’ — the expansion time of the Universe appears to be shorter than the ages of the oldest stars. The bigger ΩM, the faster galaxies were separating in the past and the greater the problem (Fig. 1). But a positive ΩΛ acts like a long-range repulsionwhich drives acceleration and increases our estimates of the expansion time. The age problem has eased in the past few years, however, as estimates of the Hubble constant (the present expansion rate) and the ages of the oldest stars have come down.

Figure 1: The furthest supernova.


The first two images, from the Cerro Tololo Interamerican Observatory 4-metre telescope, show a galaxy halfway across the Universe transformed by the apppearance of a type-1a supernova (resolved more clearly in a follow-up observation using the Hubble Space Telescope). Being of well-known brightness, these supernovae help to determine the deceleration, and so the eventual fate, of the Universe.

A more roundabout reason for introducing the cosmological constant is the theory of ‘inflation’. According to the cosmic inflation hypothesis, the nascent Universe underwent such a phenomenal stretching that it should be nearly flat. But most observers, when they add the luminous matter to the dark matter inferred from galaxy motions, find ΩM 0.2. This can be reconciled with inflation if ΩM + ΩΛ = 1, but in that case ΩΛ will rise towards one in the future as ΩM plummets towards zero, and runaway acceleration will lead to another episode of inflation billions of years from now. Why should we just happen to be here at a time when ΩM and ΩΛ are about the same size?

When doing the ΩM sum, however, it is hard to be sure that all matter is being counted; and ΩΛ is even more elusive. A good way to measure both would be to observe the past history of the expansion, by looking deep into space, deep into the past. This boils down to plotting a Hubble diagram (apparent brightness against redshift) for a sample of ‘standard candles’ — things that have nearly the same true brightness, so relative brightness indicates relative distance. This is where the brightest kind of supernova comes in.

Type Ia supernovae3 (SNe Ia) are thought to be such standard candles. They start as white dwarf stars accreting matter from binary companion stars; eventually they approach a critical mass at which nuclear carbon burning begins. But under the electron-degenerate conditions in a white dwarf, this burning happens very rapidly — and the star explodes. These explosions all have about the same brightness because the objects involved are nearly identical: all made of carbon and oxygen, and all of the critical mass.

Indeed, the peak brightnesses of normal SNe Ia are so similar that treating them as standard candles gives relative distances to within 15 per cent. Correlations between peak brightness and distance-independent observables such as colour or the rate of fading can be used to standardize the candles further, and get distances accurate to 10 per cent. The absolute level of the SN Ia peak brightness is needed for measuring the Hubble constant4, but not for ΩM and ΩΛ. The challenge there is to discover and accurately measure the apparent brightnesses of many very remote (high-redshift) SNe Ia.

Perlmutter's team has met this challenge by developing a search strategy that virtually guarantees the discovery of batches of high-redshift (z > 0.3) SNe Ia during an observing run at a large ground-based telescope5. The guarantee means that coordinated observations by other telescopes can be arranged in advance.

In this issue, Perlmutteret al.1 present observations of SN1997ap, at the very high redshift z = 0.83. (The Universe has expanded since then by a factor of 1.83.) The brightness against time of SN1997ap was measured from the ground and from the Hubble Space Telescope, and the Keck telescope contributed a good spectrum to prove that it was a normal type Ia. The data do not support ΩM = 1. A rival pack of high-redshift-supernova hunters6 has also tentatively concluded that ΩM is low. So if space is flat, there must be a cosmological constant.

The prospects for more precisely measuring ΩM and ΩΛ are bright. Both groups already have data on dozens of remote SNe Ia, including a number near z = 1; these will provide the leverage needed to determine ΩM and ΩΛ separately7. Astronomers will be wondering whether these analyses could be underestimating ΩM. The inferred value would come out higher if remote SNe Ia were observed to be brighter than they are, relative to low-redshift supernovae. Could the high-redshift SNe Ia of the younger Universe be intrinsically dimmer than the nearer ones? (Not likely, if they have normal spectra.) Could there be wavelength-neutral intergalactic obscuration? Could the low-redshift sample be seriously biased by observational selection effects?

If the answers to these and other such questions prove to be no, it will become clear that expansion is here to stay. We should keep in mind, though, that if the cosmological principle (large-scale homogeneity) isn't valid, we could simply be exploring the nature and future of just one bubble in a cosmic sea of champagne.


  1. 1

    Perlmutter, al. Nature 391, 51–54 (1998).

  2. 2

    Carroll, S. M., Press, W. H. & Turner, E. L. Annu. Rev. Astron. Astrophys. 30, 499–542 (1992).

  3. 3

    Ruiz-Lapuente, P., Canal, R. & Isern, J. (eds) Thermonuclear Supernovae (Kluwer, Dordrecht, 1997).

  4. 4

    Saha, al. Astrophys. J. 486, 1–20 (1997).

  5. 5

    Perlmutter, al. Astrophys. J. 483, 565–581 (1997).

  6. 6

    Garnavich, P. et al. Astrophys. J. (in the press).

  7. 7

    Goobar, A. & Perlmutter, S. Astrophys. J. 450, 14–18 (1995).

Download references

Author information

Rights and permissions

Reprints and Permissions

About this article

Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.