Skip to content
Planet 29

The erosion of planetary atmospheres

Principal Investigators

  • Dr. E. W. Günther
    Thüringer Landessternwarte Tautenburg , Tautenburg


In our Solar System there are basically two species of planets: gas, or ice giants which have masses larger than 15 MEarth and densities between 0.7 to 1.6 gcm-3, and rocky planets with masses of one Earth-mass or less, and densities between 3.7 to 5.5 gcm-3. It was thus expected that extrasolar planets would have similar properties, but observations with the Kepler and CoRoT satellites showed that extrasolar planets are far more diverse. Planets in the mass-range between 1 and 15 Earth-masses can have extended Hydrogen atmospheres but can also be rocky. The rocky planets are often called super-Earths, and low-mass planets with extended Hydrogen atmospheres mini-Neptunes. Statistical studies using the data obtained with the Kepler satellite show that mini-Neptunes/super-Earth are the most abundant species of planets. For our understanding of planet formation and evolution it is thus essential to find out why some planets in this mass-regime have Hydrogen atmospheres and others not. However, we should keep in mind that the so-called min-Neptunes are unlike the gas-giants in our solar-system. Theoretical studies indicate that they are most likely rocky planets that have Hydrogen-atmospheres that contain only one or two percent of the mass of the planet.

There are three possible scenarios that can explain why some low-mass planets have extended Hydrogen atmospheres, and others not:

  1. The gas-poor formation scenario: In this scenario rocky planets do not accrete a substantial Hydrogen atmosphere. On possibility for that is that the inner disk is cleared from Hydrogen gas before the inner planets form.
  2. Atmospheric losses due to X-ray and EUV-radiation (XUV) of the star: In this scenario, it is assumed that most planets have initially a Hydrogen atmosphere. The XUV- radiation of the host star then erodes the planetary atmosphere. The main erosion phase for planets of solar-like stars are the first 100 Myrs.
  3. Atmospheric losses driven by the energy from formation: In this scenario, planets also have initially a Hydrogen atmosphere. The difference is that atmospheric mass-loss are powered by the cooling luminosity of a planet’s core and impacts. This type of erosion takes longer, 500 Myrs to 3 Gyrs.

Possibly all three mechanism play a role. The first scenario is studied by the university in Munich within the SPP 1992.

In this project, we study the second scenario. Because X-ray radiation from young stars plays an important role in the first as well as the scenarios, we exchange information and data. The basic idea of the second scenario is that planets that receive a lot of XUV-radiation from the host star lose their Hydrogen atmospheres, whereas planets that receive only a small amount keep it. Because the amount XUV-ration of young stars is orders of magnitudes larger than of old stars, the main erosion phase is when the stars are young and active. Because 20-50% of the XUV-radiation of young stars is due to flares, such events are likely to play an important role in this process. Flares are events in which a large amount of energy is released within a relatively short interval of time. They involve the heating of plasma, mass ejection, and particle acceleration that generates high-energy particles. Coronal-Mass-Ejections are large releases of plasma and magnetic field from the solar corona, and perhaps also play a role for the erosion of planetary atmospheres.


David Wöckel


M-stars stay for a longer time in the activity phase than solar-like stars. Since the activity is related to the rotation of the stars, we consider the time when more than 50% of the stars have rotation periods shorter than 3 days as the main erosion phase. This phase lasts for about 100 Myrs for solar-like stars, and 500 Myrs for M-stars. 

Another reason why planets of M-stars are particularly interesting is that potentially habitable planets orbit relatively close to the star and thus receive a lot of XUV-radiation. This so-called habitable zone has a distance of 0.04 to 0.3 AU even for planets of M-stars. 

Furthermore, 98.5% of the known planets with R<3REarth that orbit stars with M<0.5 MSun orbit closer than Mercury in our solar-system, 87.6% of them orbit even within 0.1 AU. Thus, almost all low-mass planets of M-stars that we know about orbit close to their host stars. Thus, the erosion of planetary atmospheres is particularly relevant for planets of M-stars, and young M-stars are the best targets to study these processes. 

Particularly interesting are the M-stars in the Upper Scorpius OB association, because Upper Sco has an age of 5-10 Myrs, and because the Kepler satellite has observed these stars for 1860 hours. Furthermore, planets have also been discovered in this region. K2-33 is a low mass planet in this region which currently has an extended Hydrogen atmosphere. Depending on its mass and the amount of XUV-radiation that it receives it may lose its atmosphere to become a rocky planet, or it may keep it.

In the first step of our work, we re-evaluate the status of other bona-fide M-stars in this region and identify 56 members using high-resolution spectra obtained with UVES on the VLT and AAT spectra. The AAT spectra were provided by Prof. Preibisch from the university of Munich as part of the SPP collaboration.

In the second step, we analysed the light curves obtained with the Kepler satellite. We find that 90% of the M-stars in Upper Sco have rotation periods of less than 10 days and 59% less than three days. Most M-star are thus in the high activity phase. 

The most active M-star in this region is 2MASS J16111534-1757214. Analysing the K2-light curves, we find that 2MASS J16111534-1757214 has, on average, one flare with E≥1035erg every 620 hours, and one with E≥1034erg every 52 hours. For comparison, the largest flare on the sun emitted only about 2 1030erg (Carrington flare from 1859). Flares have a power-law distribution of their energies. A power-law index β<-1 means that the total energy released in small events is larger than that of large events. A power law-index β>-1 means that the energy released in large events is larger. Using the maximum likelihood estimation, we derive β=−0.65±0.15 for 2MASS J16111534-1757214 and β=−0.52±0.13 for all other M-stars in this region. This means that rare, large events dominate the energy output. The XUV-flux thus is underestimated, if the contribution from large flares is ignored.

In a related study of the active, but older M-star, AD Leo and EV Lac, we derived β-values that are smaller than -1. Perhaps, the relative contribution of large events changes with the age of the star. 

As part of part of the SPP 1992-collaboration, send the results of our study Prof. Kuiper from the university of Tübingen who used it for their models of planetary erosion. The result of their modelling-efforts is that a Hydrogen atmosphere of a 5 MEarth-planet orbiting an M-star at 0.1 AU would be completely eroded by the XUV-radiation of flares. 

Invited Guests

Priscilla Muheki

To top