ESA Science & Technology - INFO 13-1997: A vista of new knowledge from ESA's Hipparcos astronomy mission

Hipparcos is a milestone in the history of astronomy. In 1985 the American physicist Freeman J. Dyson hailed Hipparcos as the first major new development in space science to come from outside the United States.

The spacecraft operated in orbit 1989-93, measuring the angles between stars in the sky. Over a further three years, computing teams across Europe generated a consistent, high-precision plot of 118,000 stars in the Hipparcos Catalogue and somewhat less accurate (but still unprecedented) data on a million stars in the Tycho Catalogue.

The distances, motions, pairings and variability of stars are now known far more accurately than ever before. Hipparcos will make an impact on every branch of astronomy, from the Solar System to the history of the Universe, and especially on theories of stars and their evolution. For almost a year, astronomers most closely associated with the mission have had an early view of the completed catalogues and in Venice they will summarize their initial results. The Hipparcos data will be published in June, as an extraordinary contribution from Europe to astronomy all around the world.

The success of Hipparcos has created problems for the organizers of Venice symposium. Altogether 190 scientific papers were offered for presentation by various groups of astronomers. With three mornings and three afternoons available for the main scientific sessions, 67 oral presentations are accommodated, by restricting speakers to 10-15 minutes each. For the rest, there will a generous display of results in the form of posters. Thus Hipparcos will be celebrated by a vista of new knowledge.

The Stars are Looking Younger

Already Hipparcos seems to cure a headache concerning the ages of stars. As recently as last year, astronomers were perplexed by a contradiction between their estimates of the age of the Universe, and stars that seemed to be older. An early Hipparcos result announced in February 1997 (ESA Information Note 04/97) concerned the winking stars called Cepheids, used to measure cosmic distances. Corrections to the Cepheid distances made the Universe bigger and increased its age to 10-13 billion years, even according to the shorter of two rival scales from the Hubble Space Telescope. At the same time the Hipparcos Cepheid scale drastically reduced, from 14.6 billion years to about 11 billion years, the ages of the most ancient stars, occurring in globular clusters of stars that orbit independently around the centre of the Milky Way Galaxy.

Astronomers have been anxious to know if Hipparcos results on other kinds of stars would support these novel and reassuring conclusions from the Cepheids. A broad consensus seems to be emerging, although there may well be adjustments to make when more results are in and the discussion is carried further.

Martin Barstow and his colleagues at Leicester University in England have examined Hipparcos data on white dwarfs, which are the embers of burnt-out stars. At the time of their formation, white dwarfs glow at temperatures of hundreds of thousands of degrees C, and then spend billions of years cooling down, before they disappear from view. Good Hipparcos data on the distances of white dwarfs makes the theoretical picture of their evolution much more precise.

For example V471 Tauri is a double star, in which a white dwarf and an ordinary star, similar in size to the Sun, orbit around one another.

Both objects emit X-rays -- the sun-like star from its atmosphere and the white dwarf from its surface, which glows at 32000 degrees C. The Hipparcos distance of V471 Tauri (153 light-years) prompts calculations showing that the white dwarf is only 30 per cent wider than the Earth.

Yet it is 230,000 times more massive, and its gravity is a million times stronger than at the Earth's surface. From the diameters of this and other white dwarfs a cooling rate appears, and an age for the oldest and coolest white dwarfs which is less than 11 billion years.

Although this age is about the same as that suggested by the Cepheid distance scale, visible white dwarfs are not the oldest stars in the Galaxy. Those examined by Barstow and his colleagues exist in the disk that makes the band of the Milky Way across the sky. Astronomers believe that some stars on independent orbits, among the so-called halo stars and the globular clusters, are older than the disk.

From Hipparcos data on 30 elderly halo stars, Neill Reid of the California Institute of Technology finds that they are 10-15 per cent more distant that previously supposed, and therefore about 20 per cent more luminous. The more luminous a star is, the shorter is its life. Reid therefore cuts the ages of the oldest stars from more than 14 billion years to 11-13 billion years.

An Italian team offers a rather precise age of 12.1 billion years for the oldest stars. There is still a margin of error on the figure, but the oldest stars should be younger than 13.3 billion years. Astronomers from the observatories of Padua, Bologna, Rome and Turin studied Hipparcos data for 100 old stars and carefully matched them to other stars within nine globular clusters. Although all globular clusters may not be of exactly the same age, the astronomers identified as true Methuselahs of the Galaxy: the Oosterhoof II clusters (M92, M68 and M30); the Blue Horizontal Branch globulars (M13, NGC288 and M30); and 47 Tucanae.

These are just a sample of results on stellar ages that are expected to be presented in Venice and discussed with fervour by the assembled astronomers. But Gisella Clementini of the Bologna Observatory and her collaborators are now convinced that, at the present level of accuracy of globular cluster ages, no contradiction remains between the ages of stars and the age of the Universe as predicted by standard inflationary models.

"Thanks to the stunning Hipparcos data," she says, "there seems to be no longer a conflict between the age of the oldest stars and the age of the Universe. In our opinion the Universe was born perhaps 12.5 to 13 billion years ago. The most ancient globular-cluster and halo stars formed about 12 to 12.5 billion years ago, making the spherical halo of our Galaxy. The flat disk of the Milky Way, where most of luminous stars live, was in existence about 0.5 to 1 billion years later"

Are Redshifts to be Trusted?

Two technical innovations from the Lund Observatory in Sweden will impress the world's astronomers assembling in Venice. They concern direct measurement of the motions of stars towards or away from the Earth, and the imaging of stars from Hipparcos data.

Traditional astrometric measurements, now much improved by Hipparcos, track the very small motions of stars across the sky. Combined with knowledge of a star's distance, its "proper motion" tells astronomers how fast the star moves sideways, in relation to the Sun and the Earth.

To discover the speeds along the line of sight (radial velocities) astronomers have hitherto relied on the stretching or squeezing of the wavelengths of light by the Doppler effect, producing redshifts or blueshifts.

Direct measurements of changing distances were first proposed 80 years ago, but ground-based astrometric techniques were never adequate for the task. For the first time, Hipparcos measured the stars so accurately that alterations in their distances are directly perceptible in many cases, during the four years of the space operations 1989-93.

The Lund astronomers deduce, for example, that a well-known star cluster, the Hyades, is receding from the Sun and the Earth at a speed of 40 kilometres per second. The uncertainty in the measurement is astonishingly small -- just a fraction of a kilometre per second. The Swedish team is now busy comparing Hipparcos results, on the radial velocities of selected star clusters, with very precise measurements of the stars' redshifts and blueshifts in a special programme using the 1.93-metre telescope at the Haute-Provence Observatory in France.

How trustworthy are the redshifts and blueshifts as a guide to stellar motions? This is currently a hot topic in astronomy, because searches for stars possessing planets rely on small variations in their radial velocities, supposedly due to wobbles caused by the massive planets swinging around them. The identification of 51 Pegasi, which caused a sensation two years ago, has recently been challenged on the grounds that changes in the wavelength of light may not be due to the wobble of the star.

Astronomers know other possible causes of wavelength shifts, besides the Doppler effect of a star's bodily motion. One is the Doppler effect of turbulence in the star's outer layers, seen graphically in the nearby Sun. And strong gravity shortens the wavelengths of light, creating a gravitational redshift independently of the star's motion. Changes of surface gravity in oscillating variable stars could create the illusion of a planetary wobble. Distinguishing true motions in the Hipparcos data gives astronomers their best opportunity to refine the use of the redshift, which is one of astronomy's most basic tools.

The First Images from Hipparcos

The telescope and instruments in Hipparcos were not designed to produce images of the stars. Instead, grids and detectors measured angles between the pinpoints of light as the satellite slowly turned, scanning the sky. Nevertheless the Lund team has been able to adapt a technique from radio astronomy, to generate pictures. The first series to be released shows a double star in which the relative positions of the stars change quite plainly over the period of the Hipparcos observations, as they orbit around each other.

The observational and data-processing method called aperture synthesis won a Nobel Prize for the late Martin Ryle of Cambridge, England. He showed how to combine observations from a number of radio telescopes to produce far sharper images than any of them could provide individually.

Each of the observations in an aperture-synthesis set gives a very sketchy view, but they are all of the same object. Only one computable configuration of the object explains all of the observations.

The form of the Hipparcos data is mathematically comparable to signals from a trio of small telescopes spaced 10 centimetres apart. The short wavelength of the visible light used by Hipparcos makes that equivalent, in sharpness of viewing, to a trio of radio telescopes spaced many kilometres apart. The Lund astronomers adapt aperture synthesis to a succession of Hipparcos observations of a double star over a period of about 18 months. The computation uses a standard radio astronomy software package but takes account of the motions of the stars to generate an image that freezes them at one selected moment.

Another set of observations, for a later period, then shows a change in the stars' positions. At the Venice meeting, other applications of the imaging technique will be demonstrated. Lennart Lindegren of the Lund Observatory is leader of the Northern Data Analysis Consortium, one of two parallel multinational groups that undertook the enormous calculations required for the Hipparcos Catalogue. He is pleased by his team's achievements in respect of radial motions and star imagery.

"Astronomers have always lived by their wits," Lindegren comments. "It was never easy to extract information from those small and distant lights in the sky. Planning the Hipparcos mission and its computations required novel techniques. Now we are glad to show our colleagues some new tricks for making the Hipparcos results even more useful.