A fundamental quest of modern astronomy is to locate the earliest galaxies and study how they influenced the intergalactic medium a few hundred million years after the Big Bang1,2,3. The abundance of star-forming galaxies is known to decline4,5 from redshifts of about 6 to 10, but a key question is the extent of star formation at even earlier times, corresponding to the period when the first galaxies might have emerged. Here we report spectroscopic observations of MACS1149-JD16, a gravitationally lensed galaxy observed when the Universe was less than four per cent of its present age. We detect an emission line of doubly ionized oxygen at a redshift of 9.1096 ± 0.0006, with an uncertainty of one standard deviation. This precisely determined redshift indicates that the red rest-frame optical colour arises from a dominant stellar component that formed about 250 million years after the Big Bang, corresponding to a redshift of about 15. Our results indicate that it may be possible to detect such early episodes of star formation in similar galaxies with future telescopes.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
We thank K. Umetsu for a discussion of gravitational lensing models; K. Nakanishi, F. Egusa and K. Saigo for discussions about handling ALMA data; S. Kikuchihara for supporting MOSFIRE observations; and H. Yajima and A. Zitrin for discussions. We acknowledge support from: NAOJ ALMA Scientific Research Grant number 2016-01 A (T.H. and A.K.I.); European Research Council Advanced Grant FP7/669253 (N.L. and R.S.E.) and 339177 (C.E.R.); KAKENHI grants 26287034 and 17H01114 (K.M. and A.K.I.), 17H06130 (Y. Tamura), 17H04831 (Y.M.), 16H01085 (T.O.), 16H02166 (Y. Taniguchi), 15K17616 (B.H.), 17K14252 (H.U.), JP17H01111 (I.S.), 16J03329 (Y.H.) and 15H02064 (M.O.); the grant CONICYT-Chile Basal-CATA PFB-06/2007, FONDECYT Regular 1141218 (F.E.B.); NAOJ Visiting Fellow Program (N.H.H.). ALMA is a partnership of ESO (representing its member states), NSF (USA) and NINS (Japan), together with NRC (Canada), NSC and ASIAA (Taiwan) and KASI (Republic of Korea), in cooperation with the Republic of Chile. The Joint ALMA Observatory is operated by ESO, AUI/NRAO and NAOJ. This work incorporates observations made with ESO Telescopes at the La Silla Paranal Observatory. This work is also partly based on observations made with the Spitzer Space Telescope, which is operated by the Jet Propulsion Laboratory, California Institute of Technology under a contract with NASA, as well as observations obtained with the NASA/ESA Hubble Space Telescope at the Space Telescope Science Institute (STScI), which is operated by the Association of Universities for Research in Astronomy, Inc. under NASA contract NAS 5-26555.
Extended data figures and tables
a, ALMA dust contours of MACS1149-JD1 overlaid on the HST F160W image. Contours are drawn at ± 2σ, where σ = 17.7 μJy per beam. Negative contours are shown by the dashed line. The ellipse at the lower left corner indicates the synthesized beam size of ALMA. b, Dust continuum of a dusty galaxy at z = 0.99 in our ALMA field of view, overlaid on the HST F160W image. Contours are drawn at −2σ, 2σ, 4σ, 6σ, 8σ and 9.5σ, where σ = 16.0 μJy per beam.
a, Orientation of the X-shooter slit (white dashed-line rectangle), demonstrating the successful acquisition of MACS1149-JD1 data by alignment of the slit to follow the lensed elongation, as well as the inclusion of a bright foreground galaxy. b, Two-dimensional X-shooter spectra of MACS1149-JD1, with the position of Lyα marked with a green arrow and the two negative counterparts shown by red arrows. Sky lines are highlighted by blue rectangles. c, Extracted one-dimensional spectra in a 0.8″ aperture. Lyα is shaded in yellow, 2σ is shown in grey, and the sky lines are marked by blue rectangles.
The flux densities Fν of [O iii] and Lyα emissions are shown with a resolution of ∼42 km s−1 and 15 km s−1, respectively. The values are normalized by the peak flux densities. The zero-velocity point corresponds to the [O iii] redshift, z = 9.1096 (red dashed line), and the Lyα offset is ∼450 km s−1 (blue dashed line). Grey rectangles show regions contaminated by night-sky emission. The data at around −100 km s−1 to 0 km s−1 were removed from the analysis because the night-sky emission was too strong. The black solid lines indicate the 1σ noise level for these velocity resolutions.
a, b, Best-fit SEDs obtained with a single stellar component, assuming an exponentially declining (a) and a constant (b) SFH. c, d, Best-fit SEDs obtained with two stellar components, assuming a constant SFH. The star formation duration of the old component is τ = 10 Myr (c) and 200 Myr (d). The reduced χ2 value, , and the best-fit stellar age for each model is shown in the upper left corner. The meanings of the symbols are the same as those in Fig. 2.
The red and blue rectangles show the old and young stellar components with constant SFRs, respectively. The old component stops its star formation activity after a fixed duration, τ. The black vertical dashed line indicates the observation at z = 9.1 (Universe age of ~550 Myr). Each component is described using age and SFR parameters. For simplicity, both components have a common dust attenuation, Av, and metallicity, Z.
a–c, (a), IRAC colour (b) and dust emission (c), plotted against stellar age. All model grids are shown with grey dots, and the best-fit models at given stellar ages are indicated with red circles. In b, the black horizontal dashed line indicates the observed value and the yellow shaded region its 1σ uncertainty. In c, the black horizontal dashed line refers to the 2σ upper limit. d–f, Best-fit SEDs at ages of 1 Myr, 10 Myr and 100 Myr indicated by the blue squares in a. The insets show the flux density of the dust continuum (in μJy) and the [O iii] 88-μm flux (in 10−18 erg cm−2 s−1). In d we demonstrate that a strong nebular continuum plus [O ii] 3,727-Å emission counteracts intense Hβ plus [O iii] 4,959-Å emission, producing an ‘inverse Balmer break’ for very young metal-poor cases.
Extended Data Fig. 7 Evolution of the ultraviolet luminosity of MACS1149-JD1 as a function of redshift.
For each redshift bin (Δz = 1), we extrapolated the magnitude by assuming a constant SFR over the redshift interval (blue curve). We over-plotted the sensitivity of the NIRCam filters (pink and grey) covering the 1,500-Å rest frame (10σ in ~20 min) and the NIRSpec sensitivity (dashed black line) at the same wavelength (10σ in 3 h).