Astrophysics George Field Harvard-Smithsonian Center for Astrophysics, Cambridge, Massachusetts 02138 During the 100 years that astrophysics has been recognized as a separate discipline, there has been progress in interpreting observations of stars and galaxies using the principles of modern physics. Here we review some of the highlights, including the evolution of stars driven by nuclear reactions in their interiors, the emission of detectable neutrinos by the sun and by a supernova, the rapid and extremely regular rotation of neutron stars, and the accretion of matter onto black holes. A comparison of the observed Universe with the predictions of general relativity is also given. [S0034-6861(99)04602-4] I. INTRODUCTION stars is of the order of 107 K, (b) that the earth is Giga- years (Gy) old, and (c) that the sun and stars are mostly Astrophysics interprets astronomical observations of hydrogen. Bethe’s cycle works on hydrogen at about stars and galaxies in terms of physical models. During 107 K, and the luminosity of the sun can be balanced for this century new classes of objects were discovered by 10 Gy by burning only 10% of it. astronomers as novel instruments became available, The stage had been set by Hertzsprung (1911) and challenging theoretical interpretation. Russell (1914), who had found that, in a diagram in Until the 1940s, astronomical data came entirely from which the luminosity of a star is plotted against its sur- optical ground-based telescopes. Photographic images face temperature, most stars are found along a ‘‘main enabled one to study the morphology of nebulae and sequence’’ in which the hotter stars are brighter and the galaxies, filters permitted the colors of stars and hence cooler are fainter. A sprinkling of stars are giants, which their surface temperatures to be estimated, and spectro- greatly outshine their main-sequence counterparts, or graphs recorded atomic spectral lines. After World War white dwarfs, which though hot, are faint. Eddington II, physicists developed radio astronomy, discovering (1924) had found that the masses of main-sequence stars relativistic particles from objects like neutron stars and correlate well with their luminosities, as he had pre- black holes. NASA enabled astronomers to put instru- dicted theoretically, provided the central temperatures ments into earth orbit, gathering information from the were all about the same. Bethe’s proposal fitted that re- ultraviolet, x-ray, infrared, and gamma-ray regions of quirement, because the fact that only the Maxwell- Boltzmann tail of the nuclear reactants penetrates the the spectrum. Coulomb barrier makes the reaction rate extremely sen- As the century opened, astrophysicists were applying sitive to temperature. But Bethe’s discovery did not ex- classical physics to the orbits and internal structure of plain the giants or the white dwarfs. stars. The development of atomic physics enabled them Clues to this problem came with the application of to interpret stellar spectra in terms of their chemical photoelectric photometry to the study of clusters of stars composition, temperature, and pressure. Bethe (1939) like the Pleiades, which were apparently all formed at demonstrated that the energy source of the sun and stars the same time. In such clusters there are no luminous— is fusion of hydrogen into helium. This discovery led hence massive—main-sequence stars, while giants are astrophysicists to study how stars evolve when their common. In 1952 Sandage and Schwarzschild showed nuclear fuel is exhausted and hence contributed to an that main-sequence stars increase in luminosity as he- understanding of supernova explosions and their role in lium accumulates in the core, while hydrogen burns in a creating the heavy elements. Study of the interstellar shell. The core gradually contracts, heating as it does so; medium is allowing us to understand how stars form in in response, the envelope expands by large factors, ex- our Galaxy, one of the billions in the expanding uni- plaining giant stars. Although more massive stars have verse. Today the chemical elements created in super- more fuel, it is consumed far faster because luminosity nova explosions are recycled into new generations of increases steeply with mass, thus explaining how massive stars. A question for the future is how the galaxies stars can become giants, while less massive ones are still formed in the first place. on the main sequence. The age of a cluster can be computed from the point II. STELLAR ENERGY AND EVOLUTION at which stars leave the main sequence. Sandage found that ages of clusters range from a few million to a few A key development in astrophysics was Bethe’s pro- billion years. In particular, globular star clusters— posal that the carbon cycle of nuclear reactions powers groups of 105 stars distributed in a compact region—all the stars. H fuses with 12C to produce 13N, then 14N, have the same age, about 10 Gy, suggesting that this is 15O, and 15N. The latter reacts with H to form 12C again, the age of the Galaxy. The article by Turner and Tyson plus 4He. Thus each kilogram of H fuses to form slightly in this volume explains why the age of globular clusters less than a kilogram of He, with the release of 631014 is a key datum in cosmology. joules. Bethe was trying to find an energy source that As more helium accumulates, the core of a star con- would satisfy three conditions: (a) Eddington’s finding tracts and its temperature increases. When it reaches (1926) that the central temperature of main-sequence 108 K, 4He burns to 12C via the triple-a process discov- Reviews of Modern Physics, Vol. 71, No. 2, Centenary 1999 0034-6861/99/71(2)/33(8)/$16.60 ©1999 The American Physical Society S33 S34 George Field: Astrophysics ered by Salpeter (1952); the core shrinks until the den- versing their inward collapse to outward expansion. sity is so high that every cell in phase space is occupied Most of the star is ejected at 20 000 km s21, causing a by two electrons. Further compression forces the elec- flash known to astronomers as a supernova of Type II. tron momenta to increase according to the Pauli prin- This scenario was confirmed in 1987 when Supernova ciple, and, from then on, the gas pressure is dominated 1987A exploded in the Large Magellanic Cloud, allow- by such momenta rather than by thermal motions, a con- ing 19 neutrinos to be detected by underground detec- dition called electron degeneracy. In response, the enve- tors in the U.S. and Japan. If the core is not too massive, lope expands to produce a giant. Then a ‘‘helium flash’’ neutrons have a degeneracy pressure sufficient to halt the collapse, and a neutron star is formed. Analogous to removes the degeneracy of the core, decreasing the stel- 18 23 lar luminosity, and the star falls onto the ‘‘horizontal a white dwarf but far denser, about 10 kg m , it has a branch’’ in the Hertzsprung-Russell diagram, composed radius of about 10 km. The ‘‘bounce’’ of infalling mate- of giant stars of various radii. Formation of a carbon rial as it hits the neutron star may be a major factor in core surrounded by a helium-burning shell is accompa- the ensuing explosion. nied by an excursion to even higher luminosity, produc- Ordinary stars are composed mostly of hydrogen and ing a supergiant star like Betelgeuse. helium, but about 2% by mass is C, N, O, Mg, Si, and If the star has a mass less than eight solar masses, the Fe, with smaller amounts of the other elements. The central temperature remains below the 63108 K neces- latter elements were formed in earlier generations of sary for carbon burning. The carbon is supported by de- stars and ejected in supernova explosions. As the super- generacy pressure, and instabilities of helium shell burn- nova shock wave propagates outward, it disintegrates ing result in the ejection of the stellar envelope, the nuclei ahead of it, and as the material expands and explaining the origin of well-known objects called plan- cools again, nuclear reactions proceed, with the final etary nebulae. The remaining core, being very dense products being determined by how long each parcel of (;109 kg m23), is supported by the pressure of its de- material spends at what density and temperature. Nu- generate electrons. Such a star cools off at constant ra- merical models agree well with observed abundances. dius as it loses energy, explaining white dwarfs. However, there is a serious problem with the above Chandrashekhar (1957) found that the support of description of stellar evolution. In 1964 John Bahcall massive white dwarfs requires such high pressure that proposed that it be tested quantitatively by measuring electron momenta must become relativistic, a condition on earth the neutrinos produced by hydrogen burning in known as relativistic degeneracy. ‘‘Chandra,’’ as he was the core of the sun, and that available models of the called, found that for stars whose mass is nearly 1.5 sun’s interior be used to predict the neutrino flux. Ray- times the mass of the sun (for a helium composition), mond Davis took up the challenge and concluded (Bah- the equation of state of relativistic degenerate gas re- call et al., 1994) that he had detected solar neutrinos, quires that the equilibrium radius go to zero, with no qualitatively confirming the theory, but at only 40% of solutions for larger mass. Though it was not realized at the predicted flux, quantitatively contradicting it. Since the time, existence of this limiting mass was pointing to then several other groups have confirmed his result. A black holes (see Sec. III). new technique, helioseismology, in which small distur- Stars of mass greater than eight solar masses follow a bances observed at the surface of the sun are interpreted different evolutionary path.