NEWS FEATURE NEWS FEATURE News Feature: How to light a cosmic candle

Astronomers are still struggling to identify the companions that help white dwarf stars self-destruct in violent supernovae explosions.

Nadia Drake Science Writer elements transformed the billowing mate- rial into a blinding beacon of light. In its second death, the star blazed with the Billions of years ago, an ill-fated star not so That was the first time the star died. How- brilliance of three billion suns. different from our Sun began to age and ever, a spectacular resurrection was at hand. Photons expelled by this terminal stellar balloon outward. As it grew, the star’sglow Material stolen from a nearby stellar spasm zoomed across space. For 21 million darkened to a deep, foreboding red, and its companion ignited the dwarf’s carbon and years, they traveled from their home in the outer layers sloughed off into space. Eventu- oxygen layers and triggered a ferocious Pinwheel Galaxy through dust and clouds — — ally, the star’s nuclear furnace blinked out. All thermonuclear explosion that flung smol- until quite improbably they collided that remained was a dense, lifeless core about dering material outward at about 10% of with a telescope perched atop a mountain thesizeofEarth:awhitedwarfstar. the speed of light. Decaying radioactive in southern California. It was August 24, 2011, and the Palomar Transient Factory had just seen the nearest, freshest type 1a supernova that had yet been detected (1, 2). In the days and weeks that followed, that twice-dead star—now called supernova 2011fe—was the most studied ob- ject in the sky. Ordinary in every way, it was just what astronomers needed to learn more about how white dwarf stars explode. Type 1a supernovae have been observed for millennia, but in recent decades, their predictable brightness has made then invalu- able as cosmic distance markers. In the late 1990s, Nobel Prize-winning observations based on these stellar explosions revealed that the Universe is expanding at an accelerating rate, propelled by a poorly understood phenomenon dubbed dark en- ergy (3, 4). However, the processes that produce type 1a supernovae are still funda- mentally mysterious. “That’s a scary thing,” says astrophysicist Brad Tucker of Austra- lian National University and the University of California, Berkeley. “These are very powerful tools in cosmology, but we really don’t know what’s going on with them.” he says. Before 2011fe exploded, astronomers did not have much observational evidence that exploding white dwarfs were responsible for the supernovae. Just as the discovery an- swered one question, it also added tinder to a debate that burns brightly today: What In 1572, the Danish astronomer Tycho Brahe observed and studied the explosion of kind of starry companion is donating mate- a star that became known as Tycho’s supernova. More than four centuries later, Chandra’s rial to the doomed dwarf? imageofthesupernovaremnantshowsanexpanding bubble of multimillion-degree “People are now really embracing the idea debris (green and red) inside a more rapidly moving shell of extremely high energy electrons there may be more than one way to make (filamentary blue). Image courtesy of NASA/CXC/Rutgers/J. Warren and J. Hughes, et al. a type 1a supernova,” says Peter Nugent, an

www.pnas.org/cgi/doi/10.1073/pnas.1413121111 PNAS | August 19, 2014 | vol. 111 | no. 33 | 11909–11911 Downloaded by guest on September 28, 2021 Asolitarywhitedwarfisinnodangerof exceeding this mass threshold, but a dwarf in a binary star system is. When two gravita- tionally bound stars circle one another, the dwarf can steal material from its companion, growing and growing until it explodes. In a normal galaxy, this kind of kleptomaniacal relationship ends catastrophically every 200 years or so. A decade ago, many astronomers favored a model in which the dwarf’s companion was large and gassy, such as a red giant star. “Everybody was settled with that. That’swhat we taught in our introductory astrophysics classes,” Foley says. However, some type 1a supernovae seem to have taken about 10 billion years to grow up and die—much too long to be explained by a shorter-lived red giant binary system. The discrepancy sug- gests other partners must be involved. “There’s quite strong evidence that many, if not most, type 1a supernovae come from a different system, with two white dwarfs,” says astronomer Alexei Filippenko of the Supernova 2011fe was discovered just hours after it exploded in the Big Dipper. Studies University of California, Berkeley. If they’re by the Nearby Supernova Factory of its spectrum as it evolved over time have produced close enough, two dwarfs will slowly spiral in a benchmark atlas of data by which to measure all future type Ia supernovas. Image toward one another as they emit gravitational courtesy of B. J. Fulton (Las Cumbres Observatory Global Telescope Network, Goleta, CA). waves—ripples in the fabric of space-time that draw energy from the stars. As they come together, either the more massive dwarf astrophysicist at Lawrence Berkeley National Quite by chance, PIRATE had looked at the steals material from its companion until it Laboratory in California. Pinwheel Galaxy just 4 hours after 2011fe explodes and obliterates both of them or the went off. The image showed no trace of a two dwarfs collide and are annihilated. The Supernova Next Door supernova. The only possibility was that the Every night, the Palomar Transient Factory The double-degenerate scenario provoked exploding star was extremely dense and very skepticism within the field for years. Early takeshundredsofimagesofthesky.Software — ’ small less than 2% of the Sun sdiameter models of their death spiral could not explain subtracts those images from one another, and (5). It was the strongest evidence yet that type flags anything that changes between shots— how the dwarfs transferred mass or came 1a supernovae erupt from white dwarfs. together quickly enough to explode. Plus, the what astronomers call an astrophysical tran- ’ sient. These objects include variable stars, The Unusual Suspect explosion physics weren t quite right: scien- tists could not make the dwarfs in their gamma ray bursts, and supernovae. Two and Carbon-oxygen white dwarfs are dense, models shine brightly enough or explain the a half years ago, however, Nugent was sifting containing roughly a Sun’s mass of material many layers of chemical elements emerging through the images himself because the sys- squeezed into an Earth-size object. In these over the course of the explosion. tem’s software had crashed the night before. stars, the inward crush of gravity is coun- However, new observations and better That’s when he discovered 2011fe. teracted by electron degeneracy pressure, a models are now changing astronomers’ The explosion was remarkably close and, quantum mechanical property that con- minds. “Almost everything has been flip- although it was just 11 hours old, the bal- strains how tightly packed electrons can be ped on its head,” says Nugent. looning debris cloud was already big enough (this is the reason white dwarfs are called to fill the orbit of Jupiter. It was noon in degenerate stars). Degeneracy pressure usu- Missing Companions Berkeley, so he asked a colleague to aim ally prevents a catastrophic collapse, leaving Oneofthemosttellingflawsinthecasefor a telescope in the Canary Islands at the the stable white dwarf to slowly fade away large companion stars is that astronomers growing spot of light. That observation over billions of years. have not seen much evidence of them around revealed the classic spectrum of a normal, However, if the star crosses a crucial type 1a supernovae. young 1a supernova, containing silicon, cal- threshold around 1.4 solar masses, it becomes When a dwarf detonates, the explosion cium, a little bit of iron—and no hydrogen. massiveenoughtobothoverwhelmde- should tear some of the hydrogen gas from Many of the major space observatories and generacy pressure and begin fusing carbon a nondegenerate companion star and fling it telescope arrays on the ground soon joined nuclei, causing a runaway reaction that outward. “And yet, we don’t see any evidence in, eager to gather as much data as possible ends in a type 1a supernova. “Once that for that gas,” Filippenko says. Indeed, one of about the earliest stages of the star’sfinal process starts, for the most part, we think it the hallmarks of a 1a explosion is a lack of performance. However, it was a tiny tele- just continues. It’ll burn through the entire hydrogen gas flying outward at high speed. scope in Mallorca, known by the acronym star,” says astronomer Ryan Foley of the If an explosion is caught early enough, like PIRATE, that garnered a crucial observation. University of Illinois, Urbana–Champaign. 2011fe, models suggest that astronomers

11910 | www.pnas.org/cgi/doi/10.1073/pnas.1413121111 Drake Downloaded by guest on September 28, 2021 should also see what’s called a shock break- million light-years away in the Cigar Galaxy, However, the race to accept multiple NEWS FEATURE out. This is when a large companion star acts are more constraining—buttheydonotrule progenitor systems worries Kerzendorf. as a roadblock for some of the supernova out dimmer nondegenerate stars like red “We have to try to find one model that ejecta; instead of flying off into space, the m-dwarfs, which could easily evade detection. fits all. I’mnotsayingit’s impossible that materialpilesupbehindthecompanion. “M-dwarfs are the most common kind of star there are two progenitor systems,” he Eventually, the material will heat up and in the galaxy,” says J. Craig Wheeler, an as- says. Kerzendorf adds, “But if you open produce a rapid brightening that appears as trophysicist at the University of Texas at that door, then there could be three. And an abnormal bump in the supernova’searly Austin. “How often would white dwarfs and then every single supernova could have light curves. However, no one has found m-dwarfs pair up? The answer is a lot. There its own progenitor system. That might much evidence for shock breakouts. are billions of each of them,” he says. not be the truth.” Astronomers also expected to find beaten- There is at least one strong, recent piece up former companions within the debris of of evidence for the traditional scenario: a Illuminating Dark Energy ancient 1a supernovae. “If we find an actual supernova called PTF 11kx, surrounded by Although scientists do not fully understand companionintheremnant,it’s the best proof complex shells of gas and ejected material how to explode a white dwarf, the dead stars’ we’re ever going to get to finding out what that suggest it blew up with the help of a red role as cosmic milestones is on solid ground. the progenitor system is,” says Wolfgang giant companion (9). Foley has also studied However, a better understanding of how Kerzendorf, an astronomer at the University distant supernova remnants and found that these candles light up will make distance of Toronto, Canada. about 20% have tell-tale outflows of sodium measurements more accurate and help Rocketing through space with peculiar gasthathintatarelativelylargeandgassy scientists figure out whether dark energy speeds and spins and possibly carrying companion star (10). “The simplest expla- has changed over time. To do that, scien- chemical scars from the explosion, these stars nation for that is they come from single de- tists need to find some really old, distant should be odd enough to identify. However, generate systems,” he says. supernovae and be confident that they with one possible exception (6), scientists With a growing number of observations understand the mechanics of the explo- have not seen them. Kerzendorf has scruti- supporting each scenario, many scientists sions. “The concern is that maybe the star nized the remnants of Tycho’s supernova, now strongly suspect that doomed white systems that are exploding in type 1a which exploded in 1572, Kepler’s supernova dwarfs could be dancing with a variety supernovae are different,” says Jha, “but of 1604, and the remains of a supernova that of companions. “There’sreasonablecir- we don’tknowhowthey’re going to exploded in the year 1006. Of those, Tycho’s cumstantial evidence for both channels,” be different.” is the only remnant with a possible com- say Saurab Jha, an astrophysicist at Rut- Still, scientists are confident that they will panion star, but the claim is disputed. gers University. That’s a real surprise, he crack the mystery of these twice-dead can- Conversely, a white dwarf companion adds, because it would mean that different dles. “It’s just a hard slog, the way science would not survive the explosion, so there combinations of ingredients and varied sometimes is,” Wheeler admits, “but there would be no leftover star to find. That’swhat cooking times could produce remarkably are so many people getting so much data, Ashley Pagnotta and Bradley Schaefer, then similar type 1a supernovae. I suspect we’ll figure it out.” both at Louisiana State University in Baton Rouge, concluded in 2012 after studying a supernova remnant called SNR 0509-67.5 in 1 Nugent PE, et al. (2011) Supernova SN 2011fe from an exploding 6 Ruiz-Lapuente P, et al. (2004) The binary progenitor of Tycho the Large Magellanic Cloud (7). carbon-oxygen white dwarf star. Nature 480(7377):344–347. Brahe’s 1572 supernova. Nature 431(7012):1069–1072. 2 Nugent PE, et al. (2011) Young type 1a supernova PTF 11kly in 7 Schaefer BE, Pagnotta A (2012) An absence of ex-companion stars However, absence of evidence is not M101. The Astronomer’s Telegram. Available at http://www. in the type Ia supernova remnant SNR 0509-67.5. Nature 481(7380): evidence of absence, and there has been no astronomerstelegram.org/?read=3581. Accessed July 25, 2014. 164–166. 3 Perlmutter S, et al. (1998) Measurements of Ω and Λ from 42 8 Li W, et al. (2011) Exclusion of a luminous red giant as a direct confirmation of a white dwarf com- high-redshift supernovae. Astrophysical Journal 517(2):565–586. companion star to the progenitor of supernova SN 2011fe. Nature panion. Even archival Hubble images of 4 Riess AG, et al. (1998) Observational evidence from supernovae for 480(7377):348–350. 2011fe could only rule out companions an accelerating universe and a cosmological constant. Astron J 9 Dilday B, et al. (2012) PTF 11kx: A type Ia supernova with a 116(3):1009–1038. symbiotic nova progenitor. Science 337(6097):942–945. dimmer than a Sun-like star (8). Observa- 5 Bloom JS, et al. (2012) A compact degenerate primary-star 10 Foley R, et al. (2012) Linking Type 1a supernova progenitors and tions of 2014J, a recent supernova some 11.5 progenitor of SN 2011fe. Astrophysical Journal Letters 744(2):L17. their resulting explosions. ApJ 752:101.

Drake PNAS | August 19, 2014 | vol. 111 | no. 33 | 11911 Downloaded by guest on September 28, 2021