From the outlook of a planet that resides next to a quiet, relatively predictable star, the circumstances that lead to dramatic stellar explosions elsewhere in the universe can sound somewhat improbable. Some such blasts, known as type Ia supernovae, occur when a small, dense star known as a white dwarf—roughly the diameter of Earth, but hundreds of thousands of times more massive—grows too large by siphoning material off a neighboring star, igniting a thermonuclear explosion. Other cataclysms, known as type II supernovae, occur when much heftier stars, some of them dozens of times as massive as the sun, implode under their own weight.

Luckily those circumstances arise infrequently enough to spare humankind the fallout of a nearby supernova. But the universe is a big place, and locally rare events such as type Ia and type II supernovae happen in relatively large numbers across the vast expanse of space. Now a sky survey has turned up a much rarer kind of supernova, one that defies the standard explanations for how such blasts work.

"They just don't look like normal supernovae," says Robert Quimby, a postdoctoral researcher at the California Institute of Technology, of the newfound phenomena. "That's the simplest way to put it." Quimby is part of the Palomar Transient Factory (PTF) project, which uses the 1.2-meter Oschin Telescope at Palomar Observatory in California to locate explosions in the universe, some of which are so distant that they occurred several billion years ago, but light from their detonation is only now reaching Earth. The PTF team described the new class of supernovae in a study published online June 8 in Nature. (Scientific American is part of Nature Publishing Group.)

Four new PTF supernovae, along with two events identified in the past several years that defied classification, all share the same unexplained traits: They are extraordinarily bright, and a spectral breakdown of their emitted light shows no trace of common supernova components such as hydrogen, iron and calcium. "If you look at thousands of supernova spectra, as I do, these immediately jump out to you as being peculiar," Quimby says. "They don't have the normal kinds of wiggles that you'd expect to see."

The extreme brightness of the new class of supernovae, some 10 times that of a typical type Ia supernova from an exploding white dwarf, rank them among the most luminous supernovae known. That luminosity enabled Quimby and his colleagues to spot a handful of the new supernovae among the 1,000-plus supernovae of all kinds that have been found by PTF, even though core-collapse supernovae appear to be 10,000 times more common.

But just what produces the brightness of the new class remains unknown. The way the supernovae fade from their peak brightness over time is inconsistent with the decay of radioactive elements, which is what powers the glow of a type Ia supernova. And in core-collapse cataclysms such as type II supernovae, heavy elements such as iron appear in the spectra, usually accompanied by hydrogen from the expanding supernova blast encountering ambient gas in the circumstellar medium.

One possible origin for the superluminous blasts is a very massive star, roughly 100 times the mass of the sun, that ejects a dense shell of hydrogen-depleted material. If it then undergoes core collapse to initiate a supernova, the supernova-driven ejecta would collide with the existing shell to glow brightly. Astronomers have found a precedent of a hydrogen-poor supernova preceded by an eruptive event, says Roger Chevalier, a professor of astronomy at the University of Virginia who did not contribute to the new research. But the scale of that eruption was far too small to explain the luminosity of the PTF group's supernovae.

Alternately, a supernova could have left behind a magnetar, a highly magnetized form of the dense stellar remnants known as neutron stars. The rapid spin of a magnetar could provide an internal power source to light up the supernova ejecta. But that scenario is wanting for observational backup as well; all known magnetars spin far too slowly to account for the glow of the superluminous blasts. "You want it to be formed with a spin rate of one to three milliseconds, and we don't have any evidence to show that magnetars form with those kinds of spin rates," Chevalier says. "So in principle at least you can produce the high luminosity in that way, but again there's a lot we don't understand."

Quimby and his colleagues are continuing to look for new events and to track fading supernovae over time to see how they evolve. They have even marshaled the Hubble Space Telescope to gather their ultraviolet spectra. "By building that whole sequence and incorporating the UV data, we can get a better handle on the physical origins of these things," Quimby says. But for the moment neither mechanism for the newfound supernova class is entirely convincing, Chevalier notes. "They both have their pluses and minuses, and I wouldn't say the community has come to an agreement about what is going on here," he says.