# A featureless transmission spectrum for the Neptune-mass exoplanet GJ 436b

## Abstract

GJ 436b is a warm—approximately 800 kelvin—exoplanet that periodically eclipses its low-mass (half the mass of the Sun) host star, and is one of the few Neptune-mass planets that is amenable to detailed characterization. Previous observations1,2,3 have indicated that its atmosphere has a ratio of methane to carbon monoxide that is 105 times smaller than predicted by models for hydrogen-dominated atmospheres at these temperatures4,5. A recent study proposed that this unusual chemistry could be explained if the planet’s atmosphere is significantly enhanced in elements heavier than hydrogen and helium6. Here we report observations of GJ 436b’s atmosphere obtained during transit. The data indicate that the planet’s transmission spectrum is featureless, ruling out cloud-free, hydrogen-dominated atmosphere models with an extremely high significance of 48σ. The measured spectrum is consistent with either a layer of high cloud located at a pressure level of approximately one millibar or with a relatively hydrogen-poor (three per cent hydrogen and helium mass fraction) atmospheric composition7,8,9.

## Access options

from$8.99 All prices are NET prices. ## References 1. 1 Stevenson, K. B. et al. Possible thermochemical disequilibrium in the atmosphere of the exoplanet GJ 436b. Nature 464, 1161–1164 (2010) 2. 2 Stevenson, K. B. et al. Two nearby sub-Earth-sized exoplanet candidates in the GJ 436b system. Astrophys. J. 755, 9 (2012) 3. 3 Knutson, H. A. et al. A Spitzer transmission spectrum for the exoplanet GJ 436b, evidence for stellar variability, and constraints on dayside flux variations. Astrophys. J. 735, 27 (2011) 4. 4 Madhusudhan, N. & Seager, S. High metallicity and non-equilibrium chemistry in the dayside atmosphere of hot-Neptune GJ 436b. Astrophys. J. 729, 41 (2011) 5. 5 Line, M. R. et al. Thermochemical and photochemical kinetics in cooler hydrogen-dominated extrasolar planets: a methane-poor GJ 436b? Astrophys. J. 738, 32 (2011) 6. 6 Moses, J. I. et al. Compositional diversity in the atmospheres of hot Neptunes, with application to GJ 436b. Astrophys. J. 777, 34 (2013) 7. 7 Figueira, P. et al. Bulk composition of the transiting hot Neptune around GJ 436. Astron. Astrophys. 493, 671–676 (2009) 8. 8 Nettelmann, N. et al. Interior structure models of GJ 436b. Astron. Astrophys. 523, A26 (2010) 9. 9 Miller, N. & Fortney, J. J. The heavy element masses of extrasolar giant planets, revealed. Astrophys. J. 736, L29 (2011) 10. 10 McCullough, P. M. & MacKenty, J. Considerations for using Spatial Scans with WFC3. Instrument Science Report WFC3 2012–8 (Space Telescope Science Institute, 2012) 11. 11 Deming, D. et al. Infrared transmission spectroscopy of the exoplanets HD 209458b and XO-1b using the Wide Field Camera-3 on the Hubble Space Telescope. Astrophys. J. 774, 95 (2013) 12. 12 Berta, Z. K. et al. The flat transmission spectrum of the super-Earth GJ 1214b from Wide Field Camera 3 on the Hubble Space Telescope . Astrophys. J. 747, 35 (2012) 13. 13 Mandel, K. & Agol, E. Analytic light curves for planetary transit searches. Astrophys. J. 580, L171–L175 (2002) 14. 14 Gillon, M. et al. Improved parameters for the transiting hot Jupiters WASP-4b and WASP-5b. Astron. Astrophys. 496, 259–267 (2009) 15. 15 Carter, J. A. & Winn, J. N. Parameter estimation from time-series data with correlated errors: a wavelet-based method and its application to transit light curves. Astrophys. J. 704, 51–67 (2009) 16. 16 Sing, D. K. et al. Hubble Space Telescope transmission spectroscopy of the exoplanet HD 189733b: high-altitude atmospheric haze in the optical and near-ultraviolet with STIS. Mon. Not. R. Astron. Soc. 416, 1443–1455 (2011) 17. 17 Pont, F. et al. The prevalence of dust on the exoplanet HD 189733b from Hubble and Spitzer observations. Mon. Not. R. Astron. Soc. 432, 2917–2944 (2013) 18. 18 Benneke, B. & Seager, S. Atmospheric retrieval for super-Earth atmospheres: uniquely constraining the atmospheric composition with transmission spectroscopy. Astrophys. J. 753, 100 (2012) 19. 19 Kreidberg, L. et al. Clouds in the atmosphere of the super-Earth exoplanet GJ 1214b. Nature http://dx.doi.org/10.1038/nature12888 (this issue) 20. 20 Morley, C. V. et al. Quantitatively assessing the role of clouds in the transmission spectrum of GJ 1214b. Astrophys. J. 775, 33 (2013) 21. 21 Adams, E. R., Seager, S. & Elkins-Tanton, L. Ocean planet or thick atmosphere: on the mass-radius relationship for solid exoplanets with massive atmospheres. Astrophys. J. 673, 1160–1164 (2008) 22. 22 Baraffe, I., Chabrier, G. & Barman, T. Structure and evolution of super-Earth to super-Jupiter exoplanets. I. Heavy element enrichment in the interior. Astron. Astrophys. 482, 315–332 (2008) 23. 23 Rogers, L. A. & Seager, S. A framework for quantifying the degeneracies of exoplanet interior compositions. Astrophys. J. 712, 974–991 (2010) 24. 24 Ehrenreich, D., Lecavelier des Etangs, A. & Delfosse, X. HST/STIS Lyman-α observations of the quiet M dwarf GJ 436. Astron. Astrophys. 529, A80 (2011) 25. 25 Heng, K. & Kopparla, P. On the stability of super-Earth atmospheres. Astrophys. J. 754, 60 (2012) 26. 26 Owen, J. E. & Jackson, A. P. Planetary evaporation by UV and X-ray radiation: basic hydrodynamics. Mon. Not. R. Astron. Soc. 425, 2931–2947 (2012) 27. 27 Sanz-Forcada, J. et al. Estimation of the XUV radiation onto close planets and their evaporation. Astron. Astrophys. 532, A6 (2011) 28. 28 Ehrenreich, D. & Désert, J.-M. Mass-loss rates for transiting exoplanets. Astron. Astrophys. 529, A136 (2011) 29. 29 Fortney, J. J. et al. A framework for characterizing the atmospheres of low-mass low-density transiting planets. Astrophys. J. 775, 80 (2013) 30. 30 Benneke, B. & Seager, S. How to distinguish between cloudy mini-Neptunes and water/volatile-dominated super-Earths. Astrophys. J. 778, 153 (2013) 31. 31 Eastman, J., Siverd, R. & Gaudi, B. S. Achieving better than 1 minute accuracy in the heliocentric and barycentric Julian dates. Proc. Astron. Soc. Pacific 122, 935–946 (2010) 32. 32 Knutson, H. A. et al. Using stellar limb-darkening to refine the properties of HD 209458b. Astrophys. J. 655, 564 (2007) 33. 33 Swain, M. R., Vasisht, G. & Tinetti, G. The presence of methane in the atmosphere of an extrasolar planet. Nature 452, 329–331 (2008) 34. 34 Beaulieu, J. P. et al. Methane in the atmosphere of the transiting hot Neptune GJ 436b? Astrophys. J. 731, 16 (2011) 35. 35 Hauschildt, P. et al. The NEXTGEN Model Atmosphere Grid. II. Spherically symmetric model atmospheres for giant stars with effective temperatures between 3000-6800 K. Astrophys. J. 525, 871–880 (1999) 36. 36 Claret, A. A new non-linear limb-darkening law for LTE stellar atmosphere models. Astron. Astrophys. 363, 1081–1090 (2000) 37. 37 von Braun, K. et al. The GJ 436 system: directly determined astrophysical parameters of an M dwarf and implications for the transiting hot Neptune. Astrophys. J. 753, 171 (2012) 38. 38 Gregory, P. Bayesian Logical Data Analysis for the Physical Sciences 167 (Cambridge Univ. Press, 2005) 39. 39 Bean, J. L. et al. A Hubble Space Telescope transit light curve for GJ 436b. Astron. Astrophys. 486, 1039–1046 (2008) 40. 40 Coughlin, J. L. et al. New observations and a possible detection of parameter variations in the transits of Gliese 436b. Astrophys. J. 689, L149–L152 (2008) 41. 41 Alonso, R. et al. Limits to the planet candidate GJ 436c. Astron. Astrophys. 487, L5–L8 (2008) 42. 42 Pont, F. et al. Transit infrared spectroscopy of the hot Neptune around GJ 436 with the Hubble Space Telescope. Mon. Not. R. Astron. Soc. 393, L6–L10 (2009) 43. 43 Shporer, A. et al. Photometric follow-up observations of the transiting Neptune-mass planet GJ 436b. Astrophys. J. 694, 1559–1565 (2009) 44. 44 Cáceres, C. et al. High cadence near infrared timing observations of extrasolar planets I. GJ 436b and XO-1b. Astron. Astrophys. 507, 481–486 (2009) 45. 45 Ballard, S. et al. A search for additional planets in the NASA EPOXI observations of the exoplanet system GJ 436. Astrophys. J. 716, 1047–1059 (2010) 46. 46 Knutson, H. A., Howard, A. W. & Isaacson, H. A correlation between stellar activity and hot Jupiter emission spectra. Astrophys. J. 720, 1569–1576 (2010) Download references ## Acknowledgements We thank P. McCullough for his assistance in the planning and executing of these observations. We are also grateful to J. Moses, M. Line and N. Nettelmann for conversations on the nature of high-metallicity atmospheres as well as discussions of specific interior and atmosphere models for GJ 436b. D.H. has received support from the European Research Council under the European Community’s Seventh Framework Programme (FP7/2007-2013 grant agreement number 247060). ## Author information ### Affiliations Authors ### Contributions H.A.K. carried out the data analysis for this project with input from D.D. B.B. provided the planetary atmosphere models and accompanying fits, and D.H. supplied the PHOENIX atmosphere models used to calculate the stellar limb-darkening coefficients. ### Corresponding author Correspondence to Heather A. Knutson. ## Ethics declarations ### Competing interests The authors declare no competing financial interests. ## Extended data figures and tables ### Extended Data Figure 1 Representative raw image from the 29 November 2012 ut observation, showing the scanned spectrum. ### Extended Data Figure 2 Raw white-light photometry for the four individual transits. Data are vertically offset for clarity. Transits shown were obtained on the following dates (from top to bottom): 26 October 2012 ut, 29 November 2012 ut, 10 December 2012 ut and 2 January 2013 ut. The raw fluxes are shown as filled black circles, and the best-fit solutions for the instrumental effects and transit light curves are shown as filled red circles. ### Extended Data Figure 3 White-light residuals. Data are vertically offset for clarity. Transit residuals shown were obtained on the following dates (from top to bottom): 26 October 2012 ut, 29 November 2012 ut, 10 December 2012 ut and 2 January 2013 ut. The difference between the white-light fluxes and best-fit model solutions are shown as filled black circles. ### Extended Data Figure 4 Individual transmission spectra for each of the four visits. Transmission spectra are shown as filled circles, with colours indicating the date of the observations: 26 October 2012 ut (dark blue), 29 November 2012 ut (light blue), 10 December 2012 ut (yellow) and 2 January 2013 ut (red). This plot shows the errors in the measured transit depths, but does not include the additional systematic errors from the limb-darkening models. ### Extended Data Figure 5 Observed minus calculated transit times using the new best-fit ephemeris. The solid line denotes observed minus calculated equal to zero. Transit times from this paper are plotted as filled stars, and previously published observations are shown as filled circles, with 1σ uncertainties overplotted for both. The colour of the points denotes the wavelength of the observations (blue for visible, red for infrared). Transits shown include all previously published observations for this planet3,39,40,41,42,43,44,45. The figure is adapted from figure 8 of our previous study3. ### Extended Data Figure 6 Comparison to published transit depths for GJ 436b. Filled black circles show previously published transit depths3,42,43,45,46, with 1σ uncertainties overplotted. The white-light transit depths from our WFC3 observations are plotted as black stars. We show three models for comparison, including a solar-metallicity cloud-free model (red line), a hydrogen-poor 1,900 × solar model (blue line), and a solar metallicity model with optically thick clouds at 1 mbar (grey line). As we discuss in ref. 3, the apparent variations in transit depth at different epochs could plausibly be explained by the occultation of active regions on the surface of the star. If correct, this would make broadband photometry collected at different epochs unreliable for the purpose of constraining the planet’s transmission spectrum. p.p.m., parts per million. ### Extended Data Figure 7 Stellar activity versus time. Filled black circles show the measured emission levels in the Ca ii H and K line cores from Keck HIgh Resolution Echelle Spectrometer (HIRES) spectroscopy of GJ 436 (refs 3 and 46); these are parameterized using the SHK index, where larger values indicate increased stellar activity. Vertical lines mark the approximate dates of the six most recent Spitzer transit observations (dashed line), as well as the four HST transits presented in this paper (solid line). ### Extended Data Figure 8 Averaged stellar spectrum versus PHOENIX model atmospheres. Spectra are averaged over each HST visit and then normalized using the sensitivity curve for that visit. These spectra are plotted as dark blue (October), light blue (November), yellow (December), and red (January) lines. For comparison we show two PHOENIX stellar atmosphere models with effective temperatures of 3,500 K and log(g) = 5.0 (black line, where g is surface gravity) and 3,350 K and log(g) = 4.8 (grey line) binned to the same pixel resolution as our data. These data include an additional component of instrumental broadening that smoothes out the sharp spectral features visible in the model spectra. ### Extended Data Figure 9 Joint constraints on cloud-top pressure versus atmospheric metallicity for an oxygen-rich and a carbon-rich atmosphere. a, C/O = 0.3; b, C/O = 1.0. The models shown in Fig. 3 assume a solar C/O ratio of 0.5. The coloured shading indicates the normalized probability density as a function of cloud-top pressure and metallicity derived from a variation of the Bayesian retrieval methods. We vary the amount of metals in the atmosphere (defined as elements heavier than H and He) linearly using the scaling factor shown on the lower x axes. The black contours show the 68%, 95% and 99.7% Bayesian credible regions. ## PowerPoint slides ### PowerPoint slide for Fig. 1 ### PowerPoint slide for Fig. 2 ### PowerPoint slide for Fig. 3 ## Rights and permissions Reprints and Permissions ## About this article ### Cite this article Knutson, H., Benneke, B., Deming, D. et al. A featureless transmission spectrum for the Neptune-mass exoplanet GJ 436b. Nature 505, 66–68 (2014). https://doi.org/10.1038/nature12887 Download citation • Received: • Accepted: • Published: • Issue Date: ## Further reading • ### Chemistry of Temperate Super-Earth and Mini-Neptune Atmospheric Hazes from Laboratory Experiments • Sarah E. Moran • , Sarah M. Hörst • , Véronique Vuitton • , Chao He • , Nikole K. Lewis • , Laurène Flandinet • , Julianne I. Moses • , Nicole North • , François-Régis Orthous-Daunay • , Joshua Sebree • , Cédric Wolters • , Eliza M.-R. Kempton • , Mark S. Marley • , Caroline V. Morley • & Jeff A. Valenti The Planetary Science Journal (2020) • ### Deflating Super-puffs: Impact of Photochemical Hazes on the Observed Mass–Radius Relationship of Low-mass Planets • Peter Gao • & Xi Zhang The Astrophysical Journal (2020) • ### Seeing above the clouds with high-resolution spectroscopy • Siddharth Gandhi • , Matteo Brogi • & Rebecca K Webb Monthly Notices of the Royal Astronomical Society (2020) • ### ExoReL${}^{{\mathfrak{R}}}\$: A Bayesian Inverse Retrieval Framework for Exoplanetary Reflected Light Spectra

• Mario Damiano
•  & Renyu Hu

The Astronomical Journal (2020)

• ### Detectability of Molecular Signatures on TRAPPIST-1e through Transmission Spectroscopy Simulated for Future Space-based Observatories

• Daria Pidhorodetska
• , Thomas J. Fauchez
• , Geronimo L. Villanueva
• , Shawn D. Domagal-Goldman
•  & Ravi K. Kopparapu

The Astrophysical Journal (2020)