NEWSFOCUS

ASTRONOMY Tanvir, starting the observation some 21 minutes after the burst. “We observed the target for 20 minutes, and then the wind got Scrambling to Read the Meaning too much for us,” he says. Thousands of kilometers away, sitting in an Of the Sky’s Most Ancient Flare auditorium listening to talks about the future of U.K. astronomy, Tanvir kept glancing at his This spring, a seemingly routine burst triggered a worldwide race to catch laptop in anticipation of the first images. He a glimpse of the early was also in touch with scientists operating ’s 8-meter North tele- On 23 April, astrophysicist Nial Tanvir had networks make possible worldwide, round- scope, also on Mauna Kea, which had begun just dropped his daughter off at school when the-clock collaborations, and pressures for taking observations in the optical band within an automated text message appeared on his cooperation and competition often come into minutes of the burst. cell phone. NASA’s Swift satellite had just simultaneous play. “This is extreme astron- The burst appeared as a fuzzy spot in near- detected a gamma ray burst (GRB)—a brief omy,” says Don Lamb, a theoretical astrophysi- images taken by UKIRT but didn’t flash of high-energy radiation emitted by a cist at the University of Chicago in Illinois. show up in optical images taken by Gemini. To collapsing to form a . “Oh boy, Tanvir and others, this was a clue that it had here we go again,” Tanvir, a researcher at the All-night dash occurred at a high : Visible from University of Leicester in the United King- Shortly after Tanvir received the alert, he was the burst had shifted to the near-infrared, and dom, recalls thinking. conversing via e-mail with his colleagues at light—which should have shifted to Tanvir’s mix of excitement and ennui is the Joint Astronomy Centre (JAC) in Hilo, visible —had been absorbed by familiar to GRB , who intergalactic , causing the receive two or three burst alerts a week. burst to vanish in the optical band. They must respond quickly by pointing By noon, Tanvir says, “we knew we ground-based at the burst to were looking at something very inter- observe its afterglow, which fades away esting.” He and Andrew Levan, an on September 23, 2009 in a matter of hours or days. Most GRBs at the University of War- turn out to be from the nearby universe wick in the United Kingdom, grabbed and do not yield new insights. A coveted lunch from the buffet table and found a few, however, went off billions of years quiet spot outside the auditorium to dis- ago and provide a window on the enig- cuss the data. “We weren’t going back matic early universe. When Tanvir read to listen to any more talks,” Tanvir says. the alert, he knew his team was in for a First they announced the UKIRT frenzy of observation and analysis that observations on the Ground Control www.sciencemag.org was more likely to produce a humdrum Network (GCN), a NASA e-mail list result than an interesting one. that had circulated the automated As it turned out, GRB 090423 was burst alert. The GCN system serves as the most distant burst astronomers had a bulletin board for the GRB commu- ever seen. In two papers on the arXiv Far out. A 13-billion-year-old nity, enabling astronomers to coordi- preprint server (arxiv.org), Tanvir’s explosion (center) riveted nate follow-up observations. It also group and a competing team led by astronomers. functions as a logbook and timeline of Italian astronomer Ruben Salvaterra discovery. In flagging the burst for Downloaded from report that the burst’s redshift of 8.2 other astronomers, Tanvir and his col- means that it went off a mere 625 million , who had already started planning leagues had planted their own flag, too. years after the , when the universe observations using the United Kingdom Infra- Across the Atlantic, Harvard University’s was less than 5% of its current age. The pho- Red (UKIRT) near the summit of Edo Berger and Pennsylvania State Univer- tons it spewed into space traveled for more Mauna Kea. The good news was that night had sity’s Derek Fox and Antonino Cucchiara were than 13 billion years before reaching . only just begun in Hawaii—it was 10 p.m.— crunching new data taken by an infrared imager Not only did the burst shatter the previous and the burst had gone off over the Pacific mounted on Gemini’s North telescope starting record for the farthest object seen—a galaxy at Ocean, which meant observations were possi- 75 minutes after the burst. The objective was to redshift 6.96, discovered in 2006—but it also ble. The bad news was that it was a windy confirm that the lack of an optical signal was proved that the universe came alive with night with gusts of up to 80 kilometers an indeed due to high redshift rather than to dust within a few hundred million years of the big hour blowing across the mountain. Opening blocking visible light emanating from the burst. bang. “It brings us close to that magical point of the telescope’s dome would subject the About 4 hours after Tanvir’s circular on GCN, first light,” says Volker Bromm, an astrophysi- instrument to wobbling that would make it Berger and his colleagues posted the first esti- cist at the University of Texas, Austin. “We difficult to get any useful images. mate of how distant the burst was: a redshift of don’t have to get much farther to catch the ear- “We were in fact closed for the night when 9. Later that evening, they would revise the esti- liest stars.” The discovery of GRB 090423 and we got the alert,” says Tom Kerr, one of the mate to somewhere between 7 and 9. the ensuing race to publish observations also observers at JAC. But he and his colleagues More-detailed information and a defini- offer a glimpse into 21st-century astronomy— decided to take a chance. “Against our better tive measurement of redshift could come

a high-stakes pursuit in which communications judgment, we will try it,” Kerr e-mailed only from spectroscopy. For Tanvir and his IMMLER CREDITS: NASA/SWIFT/STEFAN

1494 18 SEPTEMBER 2009 VOL 325 SCIENCE www.sciencemag.org Published by AAAS NEWSFOCUS

collaborators, that meant a second night of chances of publishing in a high-profile jour- Burst watchers. Left to right: The Very Large Tele- observations, this time using the European nal by combining observations. Strategic scope, United Kingdom Infra-Red Telescope, Tele- Southern Observatory’s 8.2-meter Very alliances start to form even while the observa- scopio Nazionale Galileo, and Gemini Observatory. Large Telescope (VLT) in Chile. tions are being made, says John Nousek, an Time was of the essence not only for get- astrophysicist at Pennsylvania State Univer- the papers were, how unique the data sets ting the data but also for analyzing it to solidify sity, University Park, and one of 45 co-authors were,” says Gehrels, who eventually joined the claim to discovery. And so, after debating on the Salvaterra paper. Because GRBs are the Tanvir paper while assigning some of his whether to sleep at all, Tanvir went to bed, set- such a community event for astronomers, he colleagues to Salvaterra’s team. After weeks ting the alarm for 1:30 a.m. when the night says, there’s a very public race to publish. “Of of discussion, the two sides agreed to keep would be starting in Chile. When he woke up, course, ego plays a role,” he says. their papers separate but attempt to publish he tiptoed over to his spare bedroom and For Tanvir and Salvaterra, the race to pub- them simultaneously. On 9 June, the two logged on to his computer. lish began as soon as the race to analyze papers appeared together on astro-ph, the There was troubling news: A technical observations had ended. Tanvir quickly began astrophysical section of the arXiv preprint problem at VLT had caused a delay. “For writing a draft, forming an alliance with the server, with a note indicating that each had some time, it felt like that might scup- different groups he been submitted to Nature. on September 23, 2009 per the situation,” Tanvir says. Then, The astro-ph papers reach similar 17.5 hours after the burst, the tele- conclusions about the burst’s redshift scope began taking spectra using the and significance. The Salvaterra paper Infrared Spectrometer And Array Cam- makes the additional inference that the era (ISAAC). Meanwhile, researchers star that collapsed to produce the burst at the Max Planck Institute (MPI) for contained elements heavier than Astrophysics in Garching, Germany, hydrogen and helium. That suggests it had posted a more definitive value of formed well after the first stars in the www.sciencemag.org redshift—about 8.0—based on images universe—made entirely of hydrogen taken 15 hours after the burst by a 2.2- and helium—had had time to synthe- meter telescope owned by MPI and ESO size heavier elements and spew them at the La Silla Paranal Observatory in into the galactic medium. Chile. The Italian group—led by Ruben Although Tanvir agrees that the star Salvaterra, Guido Chincarini, and others could not have been a first-generation at the Italian Institute of Astrophysics in Friendly rivals. Salvaterra (left) and Tanvir agreed object, he says the spectroscopic data Merate—was racing Tanvir’s group to finish to publish their papers together. are too sketchy to support any conclusions Downloaded from the spectral analysis. Using the 3.6-meter Tele- about its composition. Sandra Savaglio, a scopio Nazionale Galileo on La Palma in the had been communicating with, including the researcher at the MPI for Extraterrestrial Canary Islands, the researchers had begun tak- researchers from Garching. Three days after Physics in Garching, Germany, is skeptical ing near-infrared spectra of the burst’s after- the burst, he got an e-mail from Chincarini as well. “You can’t say anything about the glow 3 hours before VLT. suggesting that the two groups work together. [composition] based on these When Tanvir started analyzing the VLT According to Tanvir, Chincarini spelled spectra,” she says. data, dawn was breaking over Cambridge. He out one condition: An Italian astronomer The next time Swift detects a high-redshift saw a GCN circular from the Italian group, would have to be made first author of the burst, astronomers hope, telescopes on the posted shortly after 4 a.m. It was the first spec- paper. Chincarini’s implicit argument, Tanvir ground will respond even faster, capturing troscopic result about the burst, reporting a says, was that “their spectrum was obtained a data that will help paint a more vivid picture redshift of 7.6. At 8:20 a.m., a bleary-eyed Tan- few hours before us.” But “we could argue of the early universe. “There were no success- vir posted the results of his group’s spectral that we got the correct measurements first,” ful attempts to take spectra on the first night, analysis showing a redshift of about 8.2, which he says. Tanvir declined the offer. which is a real pity,” says Lamb. But he’s turned out to be more accurate. Neil Gehrels, an astrophysicist at NASA’s optimistic that nature has more opportunities Goddard Space Flight Center in Greenbelt, in store; and that not too far in the future, Splitting the difference Maryland, and principal investigator for light from an even more ancient collapsing Even though every group that collects data on Swift, attempted to resolve the differences and star will quicken the pulse of observers and a burst could choose to publish independently, have one paper. “We had e-mails and phone gladden the hearts of theorists.

CREDITS: (TOP, LEFT TO RIGHT): EUROPEAN SOUTHERN OBSERVATORY; JOINT ASTRONOMY CENTRE; PATRICK SOGORB; GEMINI OBSERVATORY/AURA; (BOTTOM, LEFT TO RIGHT): MAX ALEXANDER SOGORB; GEMINI OBSERVATORY/AURA; R. SALVATERRA; COURTESY JOINT ASTRONOMY CENTRE; PATRICK LEFT TO RIGHT): EUROPEAN SOUTHERN OBSERVATORY; CREDITS: (TOP, astronomers know they can increase their calls to discuss who was first, how far along –YUDHIJIT BHATTACHARJEE

www.sciencemag.org SCIENCE VOL 325 18 SEPTEMBER 2009 1495 Published by AAAS