Welcome to my homepage!

I am a postdoctoral researcher in the group of Thomas Henning at the Max Planck Institute of Astronomy (MPIA). In 2022 I will become a staff scientist in Laura Kreidberg's new APEx department at MPIA. My current research focuses on modeling the structure of exoplanet atmospheres, and the synthesis of planetary spectra in low, medium and high resolution, which I compare to data of the world's most sensitive telescopes. For this I use a suite of self-written software. Additionally, I work on planet formation and evolution, and connecting the atmospheric abundances of exoplanets to their formation history. On this website you can find information on my research interests, atmospheric model grids, a description of my codes, a list of my publications, and my CV.

For the petitRADTRANS documentation website, please click here.

Research Grids petitCODE petitRADTRANS Publications CV Contact

Research interests

I am interested in topics ranging from the formation and evolution of exoplanets to their observable spectra. Below you can find a more detailed summary of topics that I am working on.


Classical planet detection techniques such as transit and radial velocity measurements allow us to infer a planet's mass and radius. Studying spectra of exoplanets atmospheres is the only way to gain additional insight into the properties of these objects: what is the planet's atmospheric temperature and composition, and how may its composition link back to its formation history? Prerequisite for such studies are retrieval methods, which construct posterior distributions of the atmospheric parameters, given a planet's spectral observations. My retrieval code petitRADTRANS (Mollière et al. 2019, 2020) has been applied to data of the world's most sensitive telescopes, such as the VLT and HST. The code is open-source and available for download here.

Exoplanet atmospheric modeling

Exoplanet atmospheres can vary strongly for different insolation strengths, host star types, and elemental composition. To study their behavior I am using a self-written software called petitCODE. petitCODE calculates atmospheric structures, assuming radiative-convective and chemical equilibrium. The code allows for a variety of cloud species to be included, and treats absorption and scattering processes. The codes functionality (including scattering) has been succesfully benchmarked (see Baudino et al. 2017). petitCODE is described in Mollière et al. (2015) and Mollière et al. (2017). Using the results for the atmospheric structures, I can calculate spectra at low (λ/Δλ = 10, 50), medium (λ/Δλ = 1,000) and high (λ/Δλ = 106) resolution.

Isotopologues in high-resolution spectra

Inferring the presence of isotopes in exoplanet atmospheres may shed light on their formation and evolution. This is especially true for constraints on their deuterium-to-hydrogen number ratio (D/H ratio); especially for lower mass exoplanets such as terrestrial planets, super-Earths, and ice giants formation scenarios predict that icy-body accretion, or possibly atmospheric evaporation, may have led to high D/H ratios. In Mollière & Snellen (2019) we study the detectability of isotpologues such as HDO, CH3D, and 13CO. We find that detections of such species is possible with high-resolution observations of the upcoming VLT CRIRES+ and ELT METIS instruments.

Connecting planet formation and atmosheric characterization

Understanding how planets form is a great challenge, especially given that observations of forming planets are difficult. Another avenue may be to connect planet formation to the spectral energy distribution of mature exoplanets. I contributed to such a study in Mordasini et al. (2016), where I produced synthetic spectra for planets formed by the Core Accretion paradigm. We found that while the formation location of a planet within the protoplanetary disk may be retrieveable from its spectrum, especially disk chemistry and processes such as pebble accretion need to be better understood.

Properties of transiting and self-luminous planets: from super-Earths to gas giants

In Mollière et al. (2015) we systematically studied the properties of hot Jupiters with petitCODE, and found, e.g., that coolant depletion can cause temperature inversions.

In Mollière et al. (2017) we studied prime exoplanet targets for observations with JWST using petitCODE atmospheric structures, and found that the telescope may shed light on current exoplanet conundra, such as the nature of potentially highly enriched super-Earth atmospheres, or the properties of clouds in hot Jupiters.

My grid for self-luminous exoplanets was used to derive the planetary properties of the directly imaged planet 51 Eridani b, see Samland et al. (2017).

Planet formation and evolution

The observed population of fully-formed exoplanets is often the only probe we have to study process of planet formation itself. A good formation model has to explain the properties of the known exoplanet distribution. It can also help to constrain the masses of directly imaged planets, by predicting their post-formation and evolved luminosities. I have worked on the formation of planets via the Core Accretion paradigm, and how deuterium burning may influence their structure and luminosity during, and after formation (Mollière & Mordasini 2012). Moreover I have contributed to studies investigating the evolution of planets (Mordasini et al. 2017, Linder et al. 2018), as well as to two review papers on planet formation (Mordasini et al. 2015, Baruteau et al. 2016).

Back to top

Atmospheric grids

The atmospheric grids I calculated for various publications can be found here. The file headers and names should be self-explanatory (also check the associated publications for further information). In case something remains unclear please contact me via email.

Publication Description Temperature Range* Download
Linder et al. (2018) Cool clear and cloudy atmospheric grid for self-luminous atmospheres (emission only). This is an extension of the Samland et al. (2017) grid, to low temperatures. Only Na2S and KCl clouds are considered for the cloudy models. 150-1000 K, log(g): 1.5-5, [Fe/H] = -0.4-1.4, fseds = 0.5-3 3.7 GB
Samland et al. (2017) Clear and cloudy atmospheric grid for self-luminous atmospheres (emission only) 500-1700 K (clear)
500-850 K (cloudy)
483 MB
Mollière et al. 2017** Atmospheric grid for prime JWST targets (emission and transmission) 547-2734 K 359 MB
Mollière et al. 2015 Cloud-free hot jupiter (emission only) grid 1000-2500 K 2.23 GB

*Only temperature is listed here, but the grids also vary other parameters (like [Fe/H], settling parameter fsed, C/O, log(g)...

**This is not a regularly spaced grid, but a collection of models for known exoplanets which cover the log(g)-Teq space homogeneously, with varying parameters like [Fe/H], settling parameter fsed, C/O, for all targets.

Back to top


General description

petitCODE is a 1-dimensdional code for calculating the structures as well as the emission and transmission spectra of planet atmospheres in radiative-convective and chemical equilibrium. "petit" in petitCODE stands for "Pressure-Temperature Iterator and Spectral Emission and Transmission Calculator for Planetary Atmospheres".

petitCODE has been used to calculate the atmospheric models for clear and cloudy transiting and self-luminous planets, ranging in mass from super-Earths to planets of multiple Jupiter masses. It has been used for spectral characterization, but also for planet evolutionary calculations.

So far, petitCODE models have been used in Mollière et al. (2015), Mordasini et al. (2016), Mancini et al. (2016a), Mancini et al. (2016b), Mollière et al. (2017), Samland et al. (2017), Mancini et al. (2017), Southworth et al. (2017), Baudino et al. (2017), Tregloan-Reed et al. (2018), Bonnefoy et al. (2018), Müller et al. (2018), Tinetti et al. (2018), Linder et al. (2018), Molaverdikhani et al. (2018), Mollière & Snellen (2019), Mancini et al. (2019), Allard et al. (2019) Mollière et al. (2019), Carone et al. (2020), Molaverdikhani et al. (2020a), Molaverdikhani et al. (2020b), Yan et al. (2020), Stolker et al. (2020), Schlecker et al. (2020), Sarkis et al. (2020), Maire et al. (2020), Edwards et al. (2020), Barth et al. (2020), Dransfield et al. (2020), and Mollière et al. (2020).

Some of petitCODE's main properties are summarized below:

The code is not public at this moment. Contact me via email if you are interested in using petitCODE models for your research.

Code testing

petitCODE is well tested, and benchmarked code (see Baudino et al. 2017). Some examples for tests can be found below.

  1. Comparison between correlated-k and line-by-line calculations:
    In order to test the reliability of the correlated-k implementation, the resulting planetary emission spectra were compared with spectra calculated at high resolution using a line-by-line radiative transport scheme. The correlated-k calculations were carried out at a resolution R=λ/Δλ of 10, 50 and 1000. The line-by-line calculation was carried out at a resolution of 106 and was subsequently rebinned to the three correlated-k resolutions. The error between correlated-k and the line-by-line calculation was in the low single-digit percentage range, usually at ~ 1 %.
    A figure showing such a comparison calculation can be found here (taken from Mollière et al. 2015).
  2. Checking for flux conservation:
    The radiation field arising from the converged solutions of the atmospheric structures fulfill flux conservation. At the top of the atmosphere the upward-moving radiative flux is equal to the imposed insolation and internal flux. In the deeper layers of the atmosphere the upward-directed radiative flux is equal to the imposed internal flux (the stellar flux has been absorbed in layers above). Finally, the radiative flux dwindles in the deepest regions because the atmosphere becomes convectively unstable and the flux is transported by convection. A figure showing an example calculation of the upward-moving flux can be found here (taken from Mollière et al. 2015).
  3. Reproducing analytical solutions:
    When enforcing the appropriate simplyfiying assumptions, petitCODE reproduces analytical solutions for planetary structures, such as the angle-dependent double-gray solutions from Guillot et al. (2010). A figure showing a comparison can be found here. The lines from left to right are pressure-temperature structures for insolation incidence angles of cos(θ) = 0.001, 0.01, 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1.
  4. Scattering implementation:
    When enforcing the appropriate simplyfiying assumptions, petitCODE scattering results perfectly agree with the predictions from Chandrasekhar's H functions.
  5. Benchmarking:
    petitCODE has been succussfully benchmarked (see Baudino et al. 2017).

Back to top


  • Address

    Max-Planck-Institut für Astronomie
    Königstuhl 17
    69117 Heidelberg
  • Email


Jupiter image credits: NASA/JPL-Caltech/SwRI/MSSS/Gerald Eichstäd/Seán Doran

Back to top