The nature of exoplanetary atmospheres is hotly debated. The thermal spectrum of an exoplanet called a hot Jupiter reveals the presence of an analogue of Earth's ozone layer, although its composition is unknown. See Letter p.58
Deciphering the chemical properties of atmospheres using remote sensing is the next frontier in exoplanetary science1. Short of mastering interstellar travel, it is the only conceivable path towards probing whether exoplanets are habitable or inhabited2. Hot Jupiters are a class of exoplanet with a size comparable to that of Jupiter and short orbital periods (of typically several days). Such features make hot Jupiters a good starting point for astronomers to hone their observational and theoretical techniques as they work their way towards examining smaller exoplanets that have cooler atmospheres. On page 58, Evans et al.3 report the detection of water in the thermal spectrum of the hot Jupiter WASP-121b. The discovery suggests that WASP-121b contains an analogue of Earth's ozone layer, causing the exoplanet's atmosphere to feature a temperature inversion — in which temperature increases with altitude.
To understand Evans and colleagues' findings, it helps to use Earth as an analogy (Fig. 1). The lower terrestrial atmosphere is divided into two distinct regions: the troposphere and the stratosphere. The latter, upper layer contains ozone, which, despite being relatively low in abundance, is extremely absorbent to ultraviolet radiation from the Sun. This absorption causes the stratosphere to heat up and exhibit a temperature inversion. By contrast, the temperature decreases with altitude in the troposphere, and is approximately constant in the tropopause and stratopause — the boundaries above the troposphere and stratosphere, respectively. Vigorous convection occurs in the troposphere, compared with the relatively calm stratosphere.
Using the analogy with Earth, astrophysicists initially proposed that hot Jupiters fall into two classes on the basis of their temperature. The hotter objects would have temperature inversions, whereas the cooler ones would not. This proposal was based on the reasoning that titanium oxide (TiO) and vanadium oxide (VO) would act as the extrasolar analogue of ozone in hot-Jupiter atmospheres4. The plausibility of this argument is rooted in both physics and analogy. Both TiO and VO are extremely absorbent to radiation stretching from the visible to the near-infrared range of wavelengths5. Furthermore, both molecules are commonly detected in the spectra of stars less massive than the Sun and of brown dwarfs6 — astronomical objects too massive to be exoplanets and too diminutive to be stars capable of sustaining full-blown nuclear fusion.
These analogies have their limits. 'Stratosphere' is an odd term to use for a hot Jupiter — the planet undergoes intense heating by its star that varies from its equator to its poles, driving vigorous winds that churn the upper atmosphere7. And it remains unproven that brown dwarfs and hot Jupiters have a common heritage (formation mechanism, evolutionary history, and so on). The claimed detections of TiO and/or VO were all made at fairly low spectral resolutions, satisfying the threshold of plausibility to different degrees8,9,10,11, and are subject to an extensive debate in the literature concerning temperature inversions12,13,14,15,16,17,18,19.
Evans and colleagues used the Wide Field Camera 3 (WFC3) on board the Hubble Space Telescope, which has become the standard go-to instrument for detecting water on exoplanets. They obtained a thermal spectrum for WASP-121b that covered a range of wavelengths from 1.1 to 1.6 micrometres. Although the spectral resolution was insufficient to resolve the individual spectral lines of molecules, the authors could trace the shape of molecular bandheads — spectral features caused by large numbers of unresolved lines melding together.
The authors observed bandheads at wavelengths of about 1.2 and 1.4 μm. Of particular interest is the 1.4-μm bandhead that is associated with water and that forms a blunt peak, rather than a trough. Evans et al. interpreted this bandhead as being due to water seen in emission, rather than in absorption. In an atmosphere in which the temperature decreases with altitude, water would be seen in absorption. To be seen in emission requires the existence of a temperature inversion and therefore a strong absorber of stellar radiation that causes the upper atmosphere to be heated. The 1.2-μm bandhead is consistent with the presence of VO, but the definitive detection of this molecule remains elusive. The recorded WFC3 spectrum is blind to the absence or presence of TiO.
To claim that a spectral feature is seen in emission rather than absorption requires the use of a reference. Evans and colleagues analysed their WFC3 spectrum using a technique called atmospheric retrieval. In this technique, the abundances of molecules are free parameters in the analysis, which means that chemically implausible abundances are permitted. The authors then used the measured spectra of two brown dwarfs, spanning roughly the same range of wavelengths as the WFC3 spectrum, as references. These brown-dwarf spectra have deep absorption features near 1.4 μm associated with hot water vapour in the objects' atmospheres. A limitation of the authors' work is that atmospheric retrieval does not treat radiation, chemistry and atmospheric motion self-consistently, but this is an ideal that currently eludes all practitioners of the craft.
As the number of claimed detections of TiO and VO in hot Jupiters continues to increase, the presence of these molecules could be tested by other means. The chemistry of hot Jupiters is sensitively controlled by the planets' carbon-to-oxygen ratio20. Besides being water-poor and methane-rich, carbon-rich atmospheres also have abundant carbon monoxide, which sequesters most of the oxygen atoms available, leaving few for TiO and VO to form. Therefore, if TiO and VO are the extrasolar analogues of ozone, then only hot, carbon-poor atmospheres should have temperature inversions. Given a large enough sample of hot Jupiters for which spectra have been measured and chemical properties inferred, this is a falsifiable hypothesis. Furthermore, temperature inversions might counteract disequilibrium chemistry, driving atmospheres towards chemical equilibrium — a hypothesis that could be tested using high-quality spectra analysed with state-of-the-art techniques.Footnote 1
Deming, L. D. & Seager, S. J. Geophys. Res. 122, 53–75 (2017).
Seager, S. Proc. Natl Acad. Sci. USA 111, 12634–12640 (2014).
Evans, T. M. et al. Nature 548, 58–61 (2017).
Fortney, J. J., Lodders, K., Marley, M. S. & Freedman, R. S. Astrophys. J. 678, 1419–1435 (2008).
Sharp, C. M. & Burrows, A. Astrophys. J. Suppl. 168, 140–166 (2007).
Kirkpatrick, J. D. Annu. Rev. Astron. Astrophys. 43, 195–245 (2005).
Showman, A. P. & Guillot, T. Astron. Astrophys. 385, 166–180 (2002).
Désert, J.-M. et al. Astron. Astrophys. 492, 585–592 (2008).
Haynes, K., Mandell, A. M., Madhusudhan, N., Deming, D. & Knutson, H. Astrophys. J. 806, 146 (2015).
Evans, T. M. et al. Astrophys. J. 822, L4 (2016).
Mancini, L. et al. Mon. Not. R. Astron. Soc. 461, 1053–1061 (2016).
Burrows, A., Hubeny, I., Budaj, J., Knutson, H. A. & Charbonneau, D. Astrophys. J. 668, L171–L174 (2007).
Harrington, J., Luszcz, S., Seager, S., Deming, D. & Richardson, L. J. Nature 447, 691–693 (2007).
Knutson, H. A., Charbonneau, D., Allen, L. E., Burrows, A. & Megeath, S. T. Astrophys. J. 673, 526–531 (2008).
Charbonneau, D. et al. Astrophys. J. 686, 1341–1348 (2008).
Madhusudhan, N. & Seager, S. Astrophys. J. 707, 24–39 (2009).
Stevenson, K. B. et al. Astrophys. J. 754, 136 (2012).
Line, M. R., Knutson, H., Wolf, A. S. & Yung, Y. L. Astrophys. J. 783, 70 (2014).
Diamond-Lowe, H., Stevenson, K. B., Bean, J. L., Line, M. R. & Fortney, J. J. Astrophys. J. 796, 66 (2014).
Madhusudhan, N. Astrophys. J. 758, 36 (2012).