Infrared observations of exoplanetary atmospheres in the past have typically been interpreted using models that assumed solar elemental abundances. With chemical composition fixed, attempts have been made to classify hot Jupiter atmospheres on the basis of stellar irradiation. However, recent data have revealed deviations from predictions based on such classifications, and chemical compositions retrieved from some datasets have also indicated non-solar abundances. In this work, I developed a new two-dimensional classification scheme for giant exoplanetary atmospheres with irradiation (or temperature) and chemistry (represented by the C/O ratio) as the two dimensions. We define four classes in this 2-D space (O1, O2, C1, and C2) with distinct chemical, thermal and spectral properties, and apply it to six hot Jupiters. I also discuss how observations using existing and forthcoming facilities can constrain C/O ratios in exoplanet atmospheres. For details, please see the research article here and an essay for non-experts here


4. First Inference of a Carbon-rich Planetary Atmosphere

Selected Recent Results

(See Publications for full list of recent results)

  1. The interior and atmosphere of the habitable-zone exoplanet K2-18b

    N. Madhusudhan, M. Nixon, L. Welbanks, A. Piette, R. Booth 2020, ApJL, 891L, 7M

  1. Mass-Metallicity Trends in Exoplanetary Atmospheres

    L. Welbanks, N. Madhusudhan, N. Allard, et al. 2019, ApJL

  1. Detections of HCN in hot Jupiter atmospheres

    G. Hawker, N. Madhusudhan, S. Cabot, S. Gandhi 2018, ApJL, 863, 11

    S. Cabot, N. Madhusudhan, G. Hawker, S. Gandhi 2019, MNRAS, 482, 4422

  1. Measurements of H2O in ten giant exoplanets

    A. Pinhas, N. Madhusudhan, S. Gandhi, R. Macdonald 2019, MNRAS, 482, 1485

  1. New self-consistent models and retrieval methods for exoplanetary atmospheres

    (Gandhi & Madhusudhan 2017 & 2018, Pinhas & Madhusudhan 2017,2018, MacDonald & Madhusudhan 2017a,b)

  1. Detection of TiO in an exoplanetary atmosphere

    (E. Sedaghati, H. Boffin, R. MacDonald, S. Gandhi, N. Madhusudhan, et al., 2017, Nature, 549, 238)

  1. First thermal map of a super-Earth (B-O. Demory, M. Gillon, J. de Wit, N. Madhusudhan, et al.,

    Nature, 2016, 532, 207)

  1. First spectroscopic inference of a thermal inversion in a hot Jupiter (K. Haynes, A. M. Mandell,

    N. Madhusudhan, et al., 2015, ApJ, 806, 2, 146)

  1. Chemical constraints on hot Jupiter migration (Madhusudhan et al. 2014, ApJ, 794, L12)

  1. High-precision measurements of water in exoplanetary atmospheres (Madhusudhan et al. 2014, ApJ, 791, L9)

  1. First inference of a possible carbon-rich super-Earth (Madhusudhan et al. 2012, ApJ, 759, L40)

  1.   A two-dimensional classification scheme for exoplanetary atmospheres (Madhusudhan 2012, ApJ, 758, 36)

  1.   First constraint on C/O ratio of a giant planetary atmosphere (Madhusudhan et al. 2011, Nature, 469, 64)

  1.   First detection of non-equilibrium chemistry and high metallicity in an exoplanetary atmosphere

     (Stevenson et al. 2010, Nature, 464, 1161; Madhusudhan & Seager, 2011, ApJ, 729, 41)

  1.   First statistical retrieval method for exoplanetary atmospheres (Madhusudhan & Seager, 2009, ApJ, 707, 24)

  1.   Atmospheric characterization of numerous exoplanets, using Hubble, Spitzer, Kepler, and

     ground-based telescopes (please see list of publications).

The carbon-to-oxyg
en (C/O) ratio is one of the most critical parameters governing chemical conditions in all stages of planetary evolution. A C/O ratio less than 1.0, i.e. an oxygen-rich composition, is familiar in the solar system. Consequently, chemical models of extrasolar planets in the past have also typically assumed oxygen-rich compositions. However, I recently led the first measurement of C/O in an exoplanetary atmosphere, of hot Jupiter WASP-12b, and found it to be equal to or greater than 1 (i.e. carbon-rich) (Madhusudhan et al. 2011a), while the host star is known to possess a C/O of 0.4. Based on observations obtained using the Spitzer space telescope and ground-based telescopes, I found that the dayside atmosphere of WASP-12b is substantially depleted in water vapor and enhanced in methane, by at least hundred times each, consistent with expectations for a hot carbon-rich atmosphere. This discovery opens a new territory in planetary chemistry with implications in several areas, from planetary formation, interiors, and atmospheres to questions on astrobiology. For details, please see the research article here, and/or press articles at: NASA news, MIT news, TIME, BBC,, etc. 

5. Carbon-rich Giant Planets: Atmospheres and Formation Conditions

6. First Discovery of Non-equilibrium Chemistry and High Metallicity in an

    Exoplanetary Atmosphere

7. First Statistical Retrieval Method for Exoplanetary Atmospheres

In planetary atmospheres, processes such as macroscopic mixing and/or photochemistry are expected to drive molecules out of chemical equilibrium. Given the composition of a planetary atmosphere, one can, in principle, derive the parameters of the non-equilibrium processes at play. Recently, my modeling and data interpretation led to the first discovery of non-equilibrium chemistry in the atmosphere of an exoplanet, hot Neptune GJ 436b, using infrared observations  made with the Spitzer space telescope in six photometric channels. I retrieved constraints of unexpectedly low methane and large amounts of carbon monoxide (CO), both in severe disequilibrium. I used the CO abundance to infer, for the first time, vigorous vertical mixing and high metallicity in the dayside atmosphere of an exo-Neptune. More details on this work can be found in two papers (Stevenson et al. 2010; Madhusudhan & Seager 2011), and in some press articles: MIT News and NASA News.

A key difficulty in interpreting atmospheric observations of exoplanets in the past has been the limited amount of data for any planet, often less than the number of model parameters. However, given the long execution times of atmospheric models in the past, it had not been practical to run more than a few models to explain the data, leaving large regions of parameter space unexplored. I addressed this problem in Madhusudhan & Seager (2009) by developing a new modeling approach, and the first statistical retrieval technique, for exoplanetary atmospheres. The model consists of an efficient line-by-line radiative transfer scheme, with parametric composition and temperature structure, includes major opacity sources, and assumes LTE, hydrostatic equilibrium, and global energy balance. The model allows efficient exploration of phase space (computing up to ten million models) to retrieve statistical limits on the atmospheric compositions and temperature profiles for a given data set. In later work (Madhusudhan & Seager 2010, 2011), I optimized the method by using a Bayesian MCMC sampler. Since 2010, this method has led to the characterization of over a dozen transiting exoplanet atmospheres, in over a dozen journal papers, using data from hundreds of hours of collective observing time on HST, Spitzer, and major ground-based facilities.
The recent inference of a carbon-rich atmosphere in the extrasolar giant planet WASP-12b has motivated the exotic new class of carbon-rich planets (CRPs). CRPs are planets in which carbon is more abundant than oxygen, implying a carbon-to-oxygen (C/O) ratio greater than one, contrary to what is believed in the solar system (in the sun, C/O = 0.54). In a recent study (Madhusudhan et al. 2011b) we reported  the  first detailed study of possible formation mechanisms and atmospheric properties of Jupiter-sized gas giant CRPs. We find that formation of a giant CRP requires substantially low quantities of oxygen in the region of the primordial disk around the star where the planet was formed. Considering WASP-12b as a CRP orbiting a star that has more oxygen than carbon, such as our Sun, we estimate that the proto-planetary  disk must have had about 40% lesser oxygen at the formation location than one would expect based on the stellar composition. With such a large depletion in oxygen, the planetesimals, which are the building blocks of planets, are largely devoid of H2O-ice and are instead dominated by ices of CO, CO2, NH3, and CH4, as shown in the figure alongside. This result opens a new dimension for scenarios of planet formation, which have traditionally assumed oxygen-rich composition, in which water-ice has been considered to be the primary component of planetesimals. Such oxygen-poor formation environments also critically influence the habitability of terrestrial exoplanets by reducing the water content delivered to them via the planetesimals or minor bodies. The paper can be accessed here, and some press articles can be read on TIME Magazine and the PAW.

Condensate composition in C-rich conditions

Research Interests

Atmospheres, interiors, and formation mechanisms of extrasolar planets 

Radiative transfer, planetary chemistry, and atmospheric retrieval methods for exoplanets

Internal structure modeling of super-Earths, mass-radius relations, and geophysical processes

Constraints on exoplanetary formation mechanisms using their atmospheric chemical abundances

Optimal planning of observations to characterize exoplanetary atmospheres

8. A New Grid of Cloudy Atmosphere Models for Directly Imaged Giant Planets

    Application to the HR 8799 planets

In a recent study (Madhusudhan, Burrows, & Currie 2011), we generated an extensive new suite of massive giant planet atmosphere models and used it to obtain fits to photometric data for the planets HR 8799b, c, and d. We considered a wide range of cloudy and cloud-free models. The cloudy models incorporate different geometrical and optical thicknesses, modal particle sizes, and metallicities. For each planet and set of cloud parameters, we explored the space of gravity and effective temperature in search of the best-fitting models. Our new models yield statistically significant fits to the data, and conclusively confirm that the HR 8799 planets have much thicker clouds than those required to explain data for typical L and T dwarfs.  Both models with 1) physically thick forsterite clouds and a 60-micron modal particle size and 2) clouds made of 1 micron-sized pure iron droplets and 1% supersaturation fit the data. Based on the spectral fits, we report constraints on the masses and ages of HR 8799b, HR 8799c, and HR 8799d. We also use our model grid to predict the near-to-mid IR colors of soon-to-be imaged planets. Our models predict that planet-mass objects follow a locus in some near-to-mid IR color-magnitude diagrams that is clearly separable from the standard L/T dwarf locus for field brown dwarfs. The paper can be accessed here.

9. Observational Planning and Data Interpretation

10. Thermal Inversions and Geometric Albedos

Besides the above efforts, I have been involved in a number of other studies extensively characterizing exoplanetary properties using standalone models as well as interpreting data from Spitzer, EPOXI, HST, Kepler, and ground-based facilities. Several studies in the recent past have inferred the existence of thermal inversions in some transiting hot Jupiter atmospheres using Spitzer photometry. Given the limited amount of data available, the inference of a thermal inversion depends critically on the chemical composition assumed in the models. In a recent study, Madhusudhan & Seager (2010), we explored the degeneracies between thermal inversions and molecular abundances in four highly irradiated hot Jupiter atmospheres. We demonstrated that some of the datasets which have been interpreted previously as evidence for thermal inversions with solar abundance models can instead be explained without thermal inversions with a wide range of chemical compositions. On another front, using data from Kepler, EPOXI, and Spitzer, my work also led to the first joint constraints on albedos and thermal inversions in several hot Jupiters (Christiansen et al. 2010; Demory et al. 2011; Desert et al. 2011).

A substantial fraction of my research efforts is spent in investigating science that can be pursued with current and forthcoming instruments, and in data interpretation. I am involved in several major programs for characterizing atmospheres of exoplanets using existing space telescopes (Hubble, Spitzer, Kepler) and several ground-based facilities (please see the activities page for a detailed listing). For example, I provided science justification, and am currently involved in the atmospheric modeling and data interpretation, for our large HST program (115 orbits) now underway to observe sixteen transiting exoplanets with the WFC3 instrument to constrain atmospheric thermal inversions and chemistry. I play a similar role in several other programs to study atmospheres of transiting exoplanets (hot Jupiters, hot Neptunes, and a super-Earth) as well as directly imaged exoplanets. Recently, I also played a primary role in a proposal to build a dedicated facility class instrument for SOFIA for atmospheric characterization of transiting exoplanets. I was involved in drafting the science goals, deriving the band pass and precision requirements, designing the spectral channels, and generating the target list. Currently, I am conducting feasibility analyses for several existing and proposal-stage instruments for space-borne and ground-based observations of exoplanetary atmospheres. Fun News: Our WFC3 observation of exoplanet HAT-P-7b is the millionth science observation of the Hubble Space Telescope!

Right: Star map showing the planet-hosting star 55 Cancri in the constellation of Cancer. The star is visible to the naked eye, though better through binoculars. Image credits: Nikku Madhusudhan (created using Sky Map Online).

Left: Illustration of the interior of 55 Cancri e (Artist: Haven Giguere) - an extremely hot (3900 degree F) surface of mostly graphite surrounding a thick layer of diamond, below which is a layer of silicon-based minerals and a molten iron core at the center. See paper and following press articles for details: Yale News, CBS, ABC News, CBC,, NBC News, Reuters, National Geographic, Scientific American, New Scientist, Science Daily, NPR, VoA, MSNBC, Yahoo news, etc.


2. First Inference of a Possible Carbon-rich Super-Earth around a Nearby Star

3. A Two-dimensional Classification Scheme for Exoplanet Atmospheres

In a new study, we report measurements of water vapor in three giant extrasolar planets using the Hubble Space Telescope, including the most-precise H2O measurement ever made in an exoplanet (the scientific paper can be accessed here). Our results show significantly low water abundances in these planets, with major implications for planet formation and the search for water on Earth-like habitable exoplanets in future. Left: Illustration of a 'hot Jupiter' orbiting a sun-like star. A hot Jupiter is a Jupiter-sized gas giant planet orbiting extremely close to its star, with a period of a few days compared to a year on Earth. Such planets can have atmospheric temperatures between about 1300 and 5000 Fahrenheit (or 700-3000 degree Celcius) which are ideal for detecting gaseous water vapor in their atmospheres through spectroscopy. The atmospheric spectrum of a hot Jupiter, HD 209458b, shown in the inset was observed using the Hubble Space Telescope when the planet was passing in front of its parent star as viewed from Earth. The yellow circles with error bars show the observed data. The red curve shows a model matching the data. Image Credits: Haven Giguere, Nikku Madhusudhan

1. Highest-precision measurement of water vapor in an exoplanetary atmosphere

Selected Past Results