A simulated snapshot of a Type Ia supernova simulation taken very shortly after the moment of detonation.

Credit: DOE NNSA ASC/Alliance Flash Center

AT LEAST ONCE A SECOND, a dim, elderly star somewhere in the cosmos turns into a colossal thermonuclear bomb. Briefly outshining their entire home galaxies, theses explosion, known as a Type Ia supernovae, unleash the equivalent of 1028 megatons of TNT; enough energy to destroy an entire solar system.

Astronomers have marvelled at these cosmic firecrackers for centuries. But, so far, nobody has explained in detail how they explode. Now, theorists are on the verge of understanding - and just in time, because astronomers are observing Type Ia supernovae with a new urgency. In fact, the story these stars have to tell is a matter of cosmic life and death.

Within the two broad categories of supernovae, Type Ia supernovae, which are caused by the explosion of white dwarf stars, are the only type that have a regular, known brightness. This makes them ideal 'standard candles' for measuring distances in the universe - the further they are away, the dimmer they appear from Earth.

When astronomer Robert Kirshner, now at Harvard University, in Boston, first began observing these cataclysmic explosions in 1972, it didn't matter that nobody understood how they happen. A lack of knowledge about the explosion process didn't stop Kirshner and his colleagues, along with another team, in 1998, using Type Ia supernovae to discover that a mysterious entity, later dubbed dark energy, is accelerating the expansion of the universe. But today, ignorance about Type Ia supernovae is no longer bliss, say Kirshner and other astronomers. Researchers now are not only relying on supernovae as distance markers to deduce the presence of dark energy, but also to unveil its character.

Dark energy is a strange substance that cosmologists predict makes up 70% of the substance of the universe and generates a repulsive force that is accelerating its expansion. One of the deepest mysteries in all of physics and astronomy, the nature of dark energy determines the fate of the universe. If its density across the universe increases over time, the cosmos will end in a Big Rip, with every atom torn asunder. If it somehow vanishes, cosmic expansion will continue but at a slower rate. And if its strength remains fixed in time, akin to the cosmological constant that Einstein inserted into his equations of general relativity, every galaxy will someday become its own island universe.

To determine whether dark energy varies or remains the same throughout time, astronomers need to measure its 'equation of state', defined as the ratio of its density to its pressure. And to measure the equation of state at different epochs in the universe, researchers urgently need more detailed information on Type Ia supernovae, says Don Lamb, a professor of astrophysics at the University of Chicago.

Theorists are beginning to crack the riddle of supernova explosions by borrowing some of the techniques - and computer models - applied to a surprisingly down-to-Earth system: combustion in gasoline engines. Thanks to these models, which require the processing power of supercomputers, researchers can now view the full three-dimensional evolution of a stellar explosion instead of a muted, one-dimensional facsimile.

On the computer screen, "it's like watching a fire consume a forest, you just see these flames working through the star, with all this structure to it," says theoretical astrophysicist Daniel Kasen of the University of California, Santa Cruz.

Simulations developed by supernova expert Stan Woosley, also of UC Santa Cruz, along with Kasen, Fritz Röpke of the Max Planck Institute for Astrophysics in Garching, Germany, and others now suggest that supernovae that erupted a few billion years back in time may be different - intrinsically brighter - than those exploding today. The team has begun to identify several other features that may affect the brightness - such as how rapidly a star rotated before it exploded and its abundance of elements heavier than helium - which might confound dark energy measurements if overlooked.

"We're starting to make meaningful comments about how useful these supernovae can be for precision cosmology," Woosley says.

Astronomers rely on Type Ia supernovae to probe the expansion history of the universe because these explosions are almost perfect cosmic mile markers.

Since all Type Ia's appear to have the same starting point - blowing up the same amount of mass - they all should have roughly the same luminosity. After adjusting for variations by applying the 'Phillips relation', which holds that intrinsically brighter supernovae take more time to fade than dimmer ones, researchers can, in principle, read off the wattage of these cosmic light bulbs. Just as the apparent brightness of a 60-watt bulb predictably diminishes with distance, so too should the observed brightness of a supernova.

When astronomers applied this prescription, they found that light from distant supernovae appeared dimmer than it ought to be based on what had been the accepted model of the universe's evolution. That unexpected result led in 1998 to an astonishing conclusion: Rather than slowing down, the cosmos has recently sped up its rate of expansion, putting extra distance between nearby and remote supernovae - and the galaxies in which they originated.

Now, astronomers want to know the inherent brightness of Type Ia supernovae to within a few per cent, rather than the previous error margin of 20%, and how that brightness varies among different populations. Suppose, for example, that supernovae containing a lower abundance of heavy elements (typical of stars earlier in the history of the universe) are on average intrinsically brighter than supernovae exploding today.

The Phillips relation says that the supernovae with fewer metals should remain bright for a longer period of time than others. Indeed, models suggest that such cosmic bulbs would last longer than younger supernovae, but not quite as long as the relation predicts, Woosley and collaborators have found.

This effect cannot be ignored if researchers want to use Type Ia's to measure distances to an accuracy of 1 to 2 %, which will be required to assess whether or not dark energy varies with time, Kasen says.

If Type Ia supernovae vary in brightness according to a random statistical distribution, with some explosions brighter and some dimmer than average, simply observing many more of them will beat down the error in using them as standard light bulbs, Kasen says. But if some Type Ia's, such as distant ones, are systematically different from others, as his team now suggests, a problem emerges. If such properties aren't accounted for, "our errors would be greater than we really believe" in using Type Ia supernovae to measure the expansion of the universe and the nature of dark energy, says Mike Zingale of Stony Brook University in New York.

Most astronomers agree that a Type Ia supernova starts with a white dwarf — an ageing star that crams as much mass as the Sun into a volume no bigger than the Earth. Most white dwarfs are relatively cool and inert. But if the star has a companion, it will siphon mass off the neighbour star until tipping the scales at about 1.4 times the mass of our Sun. At that mass, the white dwarf becomes dense and hot enough to initiate an explosion.

No one really knows the nature of the partner stars, exactly how or where the initial nuclear flame is sparked, or how a relatively slow flame transitions into an inferno that races through the star at supersonic speeds. Because white dwarfs are so dim, astronomers have never even seen one right before it blows up as a supernova.

But based on the information they do have, theorists have developed dazzling, if complex, computer models to mimic and learn about these stellar bombs. Watching flames racing across the screen, it's easy to lose sight of one of the most important properties that researchers are now trying to pin down. A single number, the amount of nickel-56 a radioactive isotope forged at the core of the exploding star by the fusing together of lighter nuclei, determines a Type Ia's luminosity.

Although the explosion itself lasts for only a few seconds, the slow decay of nickel-56, which generates photons that diffuse out of the exploded star's core and heat the outlying shrapnel, causes supernovae to glow brightly and linger in the sky for months. It's Kasen's task, in the UC Santa Cruz group, to determine if models reproduce the observed amounts of nickel-56, how long photons would take to travel through the supernova debris and how bright the simulated supernova would appear to observers on Earth.

Different supercomputer models have to handle different aspects of a mind-boggling array of distance scales — from less than a millimetre to 2,000 km. Also, the approximations embraced by physicists and astronomers for other computational problems do not apply to supernovae, which are highly asymmetrical, involve complex, turbulent flows, and explode under conditions of high density and extreme gravity.

"We've been learning a lot from the people who study combustion," Woosley says. Internal combustion engines exhibit two types of burning that supernovae, at least in theory, also exhibit. A car engine normally operates at a slow burn, with the flame ignited by the compression of gasoline and oxygen travelling at speeds considerably slower than the speed of sound through the fluid. That sluggish burning is known as deflagration. In car engines that knock, the flame travels supersonically, a burning known as detonation.

In exploding stars, models in which a thermonuclear flame travels exclusively at subsonic speeds produce a much dimmer explosion than telescopes have recorded. Also, such models leave too much carbon and oxygen unburned. At the other extreme, simulations in which a flame travels only at supersonic speeds burn the white dwarf's material so rapidly and thoroughly that all the lighter-weight elements are squeezed together. This squeezing forges the heaviest elements a supernova can make in abundance — nickel, cobalt and iron. But that also doesn't match observations, which reveal intermediate-weight elements including magnesium, calcium and silicon in the supernova debris.

In the early 1990s, Woosley and Alexei Khokhlov, now at the University of Chicago, independently proposed that a hybrid model, in which a supernova begins as a deflagration and transitions to the more rapid detonation, might be the most likely scenario. The original simulation, however, was only in one dimension, limiting its usefulness. In 2003, a team led by Elaine Oran and Vadim Gamezo, both of the Naval Research Laboratory in Washington, D.C., showed that the hybrid model, when extended to three dimensions, did indeed match observations. But the underlying physics that would cause a transition from deflagration to detonation remains unclear.

"It looks promising, but no one is there yet," Woosley says. Determining how fast a thermonuclear flame moves and where it starts is critical, he says. His team's most recent studies show that these properties affect how much nickel-56 will be produced and how bright a supernova can become.
For instance, Röpke now finds that if the flame originates as a slightly off-centre deflagration, just 20 to 80 km from the core, the star doesn't "puff up" as much in response to the slow-moving burning front. Then, when the burning switches to a detonation, the higher density of the exploding star makes it easier to fuse lighter nuclei into heavier ones to produce nickel-56.

Differences in the location of ignition, which may vary from one white dwarf to another and result in a lopsided explosion, "may be the critical factor" for accounting for the diversity of Type Ia supernovae, and why they don't all have exactly the same brightness, Kasen says. Because the central regions of the stars are so turbulent before they explode, "we don't expect ignition to originate in the same way in every supernova." Kasen, Röpke and Woosley report their findings online at the arXiv.org physics web site and in an upcoming edition of the British journal Nature.

Building on previous studies, the team also finds that small deficits in a white dwarf's metal content - in this case a lack of elements heavier than oxygen and carbon - can generate slightly brighter supernovae by generating more nickel-56. In a few cases, the model created some supernovae that were as much as 10 % more luminous than others, Kasen says.

That's important because white dwarfs with few metals tend to hail from remote reaches of the universe, seen as they appeared farther back in time, before stars had a chance to produce an abundance of heavy elements. So Type Ia's from more distant reaches of the universe might be systematically brighter than the nearby explosions. Observing more supernovae won't address these differences; it will only eliminate the statistical ones.

"We're at a point now where we can vary the properties of the white dwarf and get a sense of what the systematic errors [in brightness] might be," Kasen adds.

Although theorists have made progress in simulating the two-step burning process, they're still debating how a slow-burn becomes a detonation. How this happens could have consequences for nickel-56 production, and ultimately how bright a Ia supernova can become.

In Woosley's view, the flame acts as a barrier, keeping apart hot ash and the cold, unburned carbon and oxygen fuel. Late in the explosion, as the white dwarf expands and densities within the star become low enough, turbulent gases rip through the flame and quench it, allowing the hot ash and cold fuel to mix. If the ash and fuel remain well mixed, they can combine into a large volume of material that ignites all at once, triggering a high-speed burning front, or detonation, says Woosley. "We don't completely understand the physics of [the transition] yet, but we understand the density when it would happen."

In another model, developed by Lamb and his colleagues at the University of Chicago's Centre for Astrophysical Thermonuclear Flashes in 2004, a series of ignition points within a white dwarf meld into a hot, burning bubble that rises rapidly, breaks through the surface of the star and spreads quickly across the surface. Waves of ash sloshing around the surface in opposite directions collide at high temperatures, creating a set of outwardly and inwardly directed jets. The inward jets penetrate the star's interior, generating temperatures and densities high enough to initiate a detonation.

In an updated version of the FLASH model, reported last year, researchers demonstrated that the simulation produces a range of nickel-56 abundances that could explain observed variations in supernova brightness, Lamb says. In an upcoming Astrophysical Journal, the researchers show that the detonations can naturally occur in their three-dimensional models. In past versions, the detonation had to be added to the model.

At the Naval Research Laboratory, Oran and Gamezo are exploring how turbulent gases in a white dwarf might generate shock waves that force the transition. They expect to unveil a new simulation in a few months. "We're [all] getting at the physical underpinnings of supernovae," says Kasen. Researchers are hoping that those details will prove to be a giant step forward in unmasking dark energy.

Follow Cosmos on Twitter!
twitter.com/cosmosmagazine