Helium is the second-most abundant element in the Universe after hydrogen and is one of the main constituents of gas-giant planets in our Solar System. Early theoretical models predicted helium to be among the most readily detectable species in the atmospheres of exoplanets, especially in extended and escaping atmospheres1. Searches for helium, however, have hitherto been unsuccessful2. Here we report observations of helium on an exoplanet, at a confidence level of 4.5 standard deviations. We measured the near-infrared transmission spectrum of the warm gas giant3 WASP-107b and identified the narrow absorption feature of excited metastable helium at 10,833 angstroms. The amplitude of the feature, in transit depth, is 0.049 ± 0.011 per cent in a bandpass of 98 angstroms, which is more than five times greater than what could be caused by nominal stellar chromospheric activity. This large absorption signal suggests that WASP-107b has an extended atmosphere that is eroding at a total rate of 1010 to 3 × 1011 grams per second (0.1–4 per cent of its total mass per billion years), and may have a comet-like tail of gas shaped by radiation pressure.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
We thank S. Seager, A. Dupree, V. Andretta, M. Giampapa and B. Drummond for discussions. This work is based on observations made with the NASA/ESA HST, obtained at the Space Telescope Science Institute (STScI) operated by AURA, Inc. J.J.S. is supported by an STFC studentship. The research leading to these results has received funding from the European Research Council (ERC) under the European Union’s Seventh Framework Programme (FP7/2007-2013)/ERC grant agreement number 336792. Support for this work was provided by NASA grants under the HST-GO-14916 programme of the STScI. G.W.H. and M.H.W. acknowledge support from Tennessee State University and the State of Tennessee through its Centers of Excellence programme. The MEarth Team (J.I. and D.C.) gratefully acknowledges funding from the David and Lucille Packard Fellowship for Science and Engineering, the US National Science Foundation (NSF) and the John Templeton Foundation. The opinions expressed in this publication are those of the authors and do not necessarily reflect the views of the John Templeton Foundation. This work was carried out in the framework of the National Centre for Competence in Research PlanetS supported by the Swiss National Science Foundation (SNSF). V.B., D.E., A.W. and S.U. acknowledge financial support from the SNSF. D.E. and V.B. acknowledge funding from the ERC under the European Union’s Horizon 2020 research and innovation programme (project FOUR ACES; grant agreement No 724427). B.V.R acknowledges support from an NSF Graduate Research Fellowship (grant DGE-1746060) and the Earth in Other Solar Systems Team, NASA Nexus for Exoplanet System Science. IRAF is distributed by the National Optical Astronomy Observatory, which is operated by the Association of Universities for Research in Astronomy (AURA) under a cooperative agreement with the NSF. J.M.G. acknowledges support from a Leverhulme Trust Research Project Grant.Reviewer information
Nature thanks D. Deming and S. Redfield for their contribution to the peer review of this work.
Extended data figures and tables
Extended Data Fig. 1 G102 white-light curve and broadband spectroscopic light curves covering the wavelength range 0.88–1.14 μm for WASP-107b.
a, Relative flux of the white-light curve with respect to systematics model results (blue points), with the best-fitting transit light curve plotted in black. b, White-light residuals and 1σ errors, after removing the combined transit and systematics components of the best-fitting model. c, Points are spectroscopic light curves divided by systematics model results, and black curves are best-fitting transit models, with vertical offsets applied for clarity. d, Best-fitting spectroscopic model residuals, with vertical offsets applied for clarity. Differently coloured points in c and d are used to highlight separate channels.
Extended Data Fig. 2 Narrowband (four-pixel-wide) spectroscopic light curves covering the wavelength range 1.06–1.12 μm.
a, Points are light curves divided by systematics model results, black curves are best-fitting transit models. b, Best-fitting model residuals, with vertical offsets applied for clarity. Differently coloured points correspond to different channels. The five non-overlapping channels used to measure the 10,833-Å line absorption are highlighted in blue.
We performed a Lomb–Scargle periodogram search and found a best-fitting period of 19.7 ± 0.9 days (dashed red line), with a relative amplitude of about 0.15%. Solid red lines show the results of the best-fitting sinusoidal model.
a, Nightly photometric observations of WASP-107, acquired with Tennessee State University’s C14 AIT at the Fairborn Observatory during the 2017 observing season. The number of observations (Nobs) was 120. ΔR is the relative flux in the Cousins R band. b, The frequency spectrum of the 2017 observations shows low-amplitude variability with a period (P) of 8.675 days (a frequency, f, of 0.115 cycles per day). c, The data phased to the 8.675-day period have a peak-to-peak amplitude of only 0.005 mag. HJD, heliocentric Julian Date; UCT, coordinated universal time; c/d, cycles per day; Tmin, best-fit ephemeris.
a, Measurements for 30 stars of different colour indices with 1σ errors from a previous work60. The two helium lines are expected to form in the same stellar-atmosphere regions and their equivalent widths are clearly correlated. Our 5,876-Å line measurement for WASP-107 (colour index B–V > 0.7), obtained from HARPS spectra, is plotted as a red line and the red-shaded region shows the 1σ error. b, Co-added spectra of WASP-107b around the 5,876-Å line of metastable helium from the HARPS radial-velocity campaign (blue line). Absorption lines are fitted with Gaussian profiles, and the best-fit results are shown as green, yellow and red lines, with the sum of the profiles shown in black. The best-fitting line profile of the 5,876-Å line is shaded in grey.
Extended Data Fig. 6 Effects of an inhomogeneous photosphere on the transmission spectrum of WASP-107b.
Lines show the stellar contamination produced by unocculted spots and faculae. Shaded regions indicate the 1σ uncertainty on the stellar contamination due to the uncertainty on spot- and faculae-covering fractions. a, The region around the 10,830-Å (air wavelength) helium triplet at the resolution of the PHOENIX spectra (resolving power, R = 500,000). (b) The full G102 wavelength range in 15-Å bins.