Gamma-ray burst has highest redshift yet seen

On 23 April, NASA’s Swift orbiter detected the most distant and therefore earliest stellar object ever seen. ¹ With a record redshift z of 8.26, the gamma-ray burst recorded by Swift is presumed to manifest the collapse of a rapidly spinning massive star—a collapsar—to a black hole about 625 million years after the Big Bang. In fact, GRB090423 is earlier than anything yet observed, at any electromagnetic wavelength, except the cosmic microwave background.

No redshift as high as 7 has ever been confirmed for a galaxy or quasar, or for any previous GRB. (Supernovae, not in the same league, are rarely seen with z beyond 2.) But typical GRBs are luminous enough to be detectable at least out to z = 20, corresponding to 180 Myr after the Big Bang—if there were any back then. Harvard University theorist Avi Loeb has been arguing for a decade that GRBs beyond z = 7 would be excellent probes of the earliest generations of stars and of the cosmic reionization of intergalactic hydrogen that their UV emission began.

Aside from record distance and the dimming, redshifting, and 1 + z time dilation that go with it, GRB090423 looks like an unremarkable representative of the majority class of latter-day GRBs, which are attributed to collapsars (see Physics Today, August 2005, page 21 ). Thus we learn that such stars already existed at 0.6 Gyr, and we begin to learn something of their formation rate in that early epoch. (The minority class of GRBs, characterized by unusually short burst durations and presumably having nothing to do with collapsars, are not luminous enough to be seen much beyond z = 1.)

In 2006 Loeb and Volker Bromm predicted that the rate of GRB occurrence at z = 8 would already have been 40% of the rate at z = 6, some 300 million years later. ² That’s consistent with the fact that Swift has thus far seen only this one GRB above z = 7 in the five years since its launch, provided one takes account of how easily an observer might miss evidence of high redshift.

Recognizing high redshift

A GRB’s initial burst of gammas and its x-ray afterglow (see figure 1(a)) provide no direct measure of z. Determining redshift requires spectrographic, or at least photometric, recording of the burst’s optical afterglow—usually with telescopes on the ground. Of course, high z moves optical afterglow well into the IR. So when several small robotically alerted ground telescopes failed to find any optical afterglow from the April GRB within minutes of its onset, two large telescopes in Hawaii swung into action and found it at IR wavelengths beyond 1.1 µm (see figure 1(b)).

Within a few hours, a team led by Derek Fox (Pennsylvania State University) had established with photometric data from the 8-m Gemini North telescope that the new GRB’s redshift was about 8, clearly a record!

By then dawn was breaking over Hawaii, so IR spectroscopy had to wait for nightfall at large telescopes in Chile and the Canaries. ³ Together with the earlier photometry, the IR spectra yielded z = 8.26 ± 0.08 (see figure 2). But because a GRB’s afterglow fades rapidly after a few hours, the spectra were too noisy to reveal absorption features that could address two cosmologically important issues: the abundance of “metals” (astronomers’ jargon for anything heavier than helium) in this very early star and its galaxy, and the abundance of neutral atomic hydrogen in the intergalactic medium (IGM) nearby.

Swift’s burst-alert telescope monitors more than 10% of the entire sky at once. So it discovers GRBs almost daily. For about a third of them, however, no optical afterglow is found and therefore no redshift is measured. It’s likely that among those discards, most of them local, are a few high-z GRBs for which no one troubled to look for the IR afterglow. “So there was some serendipity in the recognition of GRB090423’s record red-shift,” says Edo Berger (Harvard) of the Gemini North team. “And a handful of comparably distant GRBs recorded by Swift have almost certainly evaded such recognition.”

Population III stars and reionization

Because Big Bang nucleosynthesis produced almost nothing but hydrogen and helium, it’s assumed that the first generation of stars—perversely called “population III”—began life with no metals. Astronomers have yet to find a population III star. The population III epoch had probably begun by z = 20. “But we don’t know how long it lasted or how much of a role it played in cosmic reionization,” says Loeb. “Nor do we know if population III stars could have produced GRBs.” They were probably massive enough. But to generate an observable GRB, a collapsar must previously have shed its hydrogen envelope, and the stellar wind that strips off the envelope may require some threshold metallicity to get the job done.

There’s no spectral information about GRB090423’s metallicity. But its progenitor was probably not a population III star. Population III GRBs, if they exist at all, are expected to have atypically long durations.

Why is the ionization of intergalactic hydrogen by UV radiation from early stars called reionization? The cosmic microwave background is a snapshot of the cosmos at about z = 1100, its first moment of transparency, 0.4 Myr after the Big Bang. That’s when the opaque primordial plasma first became cool enough for neutral hydrogen to survive. Nowadays the hydrogen in the IGM is almost entirely reionized. It’s known from neutral-hydrogen absorption features in the spectra of high-redshift quasars that the reionization was largely complete by z = 6 (see Physics Today, October 2001, page 17 ). But cosmologists want an observational record of its progress to completion as early galaxies evolved, at least back to z = 12.

Quasars are hyperactive galactic nuclei energized by accretion onto exceptionally massive central black holes. None have been found beyond z = 6.5, probably because the earliest galaxies were small, with correspondingly modest central black holes. In any case, the telltale absorption feature (the so-called Gunn–Peterson trough) in the quasar spectrum saturates at about z = 6, when there was as little as 1% neutral hydrogen left in the IGM. So to study earlier stages of reionization, one has to look elsewhere.

Because GRBs manifest individual stars rather than galaxies, they occur in galaxies large and small. So they let astronomers look back to when there were only small galaxies, and they illuminate more typical intergalactic environments than those around quasars. In the vicinity of a high-z quasar, the IGM is likely to have suffered atypically severe reionization.

Prompt spectroscopy

For the study of cosmic reionization, GRBs, like quasars, serve to backlight the IGM. One seeks to measure how much light has been lost by Lyman-α absorption in neutral atomic hydrogen along the continuum of redshifts on the way here. For probing cosmic reionization, it turns out to be an advantage for GRBs that they, unlike quasars, have no Lyman-α line emission of their own. Their optical afterglow is entirely synchrotron emission from electrons spiraling in shock-driven magnetic fields.

Quasars, for all their limitations, do wait patiently for observers. They shine for millions of years. But a GRB’s afterglow grows fainter with every passing hour. “So we’re obliged to move fast,” says Fox. How is one to promptly identify high-z GRBs and catch their IR afterglow fast enough to yield spectra adequate for studying cosmic reionization—and for measuring the collapsing star’s metallicity?

Happily, the promptness problem is less exacerbated by increasing distance than one might think. That’s because the inverse-square falloff of the afterglow’s brightness with distance is largely compensated by the 1 + z time dilation of its duration.

The main problem has been the shift from optical to IR. Swift, whose name celebrates the speed with which it can turn to and pinpoint a GRB, carries no IR instrument. In the short run, new broad-spectrum IR spectrometers soon to be fitted onto large telescopes in Chile and Hawaii should greatly speed up the identification and spectrography of high-z GRBs.

By the middle of the next decade, the 6.5-m James Webb Space Telescope should be able to acquire afterglow spectra of GRBs even at z = 30, sometimes together with long-exposure images of the host galaxies. There might be some advantage in also having a smaller, dedicated GRB/IR space telescope that could repoint faster than the mammoth, multipurpose JWST in response to a burst alert.

“After 10 years of anticipation, we’ve finally seen a GRB more distant than any galaxy whose redshift we can measure,” says Loeb. “With JWST and such a dedicated satellite in place, we could expect to harvest about one GRB spectrum a month with redshift beyond 8.”