Zwicky and his colleague Walter Baade speculated that the explosive energy comes from gravity. Their idea was that a normal star implodes until its core reaches the density of an atomic nucleus. Like a crystal vase falling onto a concrete floor, the collapsing material releases enough gravitational potential energy to blow the rest of the star apart. An alternative emerged in 1960, when Fred Hoyle of the University of Cambridge and Willy Fowler of Caltech conceived of the explosions as giant nuclear bombs. When a sunlike star exhausts its hydrogen fuel and then its helium, it turns to its carbon and oxygen. Not only can the fusion of these elements release a titanic pulse of energy, it produces radioactive nickel 56, whose gradual decay would account for the months-long after-glow of the initial explosion.
Both these ideas have proved to be right. Of the supernovae that show no signs of hydrogen in their spectra (designated type I), most (type Ia) appear to be thermonuclear explosions, and the rest (types Ib and Ic) result from the collapse of stars that had shed their outer hydrogen layers. Supernovae whose spectra include hydrogen (type II) are thought to arise from collapse as well. Both mechanisms reduce an entire star to a shell of gaseous debris, and gravitational collapse events also leave behind a hyperdense neutron star or, in extreme cases, a black hole. Observations, notably of supernova 1987A (a type II event), have substantiated this basic theoretical picture [see "The Great Supernova of 1987," by Stan Woosley and Tom Weaver; Scientific American, August 1989].
Even so, explaining supernovae is still a major challenge for astrophysicists. Computer simulations have had trouble reproducing the explosions, let alone their detailed properties. It is reassuringly hard to get stars to explode. They regulate themselves, remaining very stable for millions or billions of years. Even dead or dying stars have mechanisms causing them to peter out rather than blowing up. Figuring out how these mechanisms are overcome has taken multidimensional simulations that push computers to, and beyond, their limits. Only very recently has the situation improved.
Blowing Up Is Hard to Do
Ironically, the stars that are thought to blow up as type Ia supernovae are usually paragons of stability--namely, white dwarf stars. A white dwarf is the inert remnant of what used to be a sunlike star. If left unmolested, it stays more or less in the state it was born, gradually cooling down and fading out. But Hoyle and Fowler argued that if a white dwarf tightly orbits another star, it may accrete matter from its companion, grow in mass and become ever more compressed at its center, until it reaches densities and temperatures sufficiently high to explosively fuse carbon and oxygen.
The thermonuclear reactions should behave rather like an ordinary fire. A front of burning should sweep through the star, leaving behind a pile of nuclear ash (mainly nickel). At any moment, the fusion reactions would occur in a tiny volume, most likely at the surface of ash-filled bubbles floating in the deep interior of the white dwarf. Because of their lower density, the bubbles would be buoyant and try to rise toward the surface of the star--much like steam bubbles in a pot of boiling water.