Solving the mystery of the Sun's low lithium abundance - and its relation to extrasolar planets

For more than half a century astronomers have wondered why the lithium abundance in our Sun is much lower than that observed in meteorites. Since the birth of the solar system, 4.5 billion years ago, the chemical composition of these celestial objects has been essentially unaltered. The low levels of lithium observed in the Sun, as well as in other distant stars around which planets have been detected, have been suggested as signature of the presence of planets. A team led by scientists from the Max Planck Institute for Astrophysics has recently published the most comprehensive study of lithium abundances in stars very similar to our Sun, which reveals that age is the key factor not the presence of planets as recently claimed by a rivalling research group.

Fig. 1: Evolution of lithium abundance with age in solar twin stars. Field solar twins are shown by open squares, while filled triangles represent solar twins in open clusters. The Sun is shown as the red circle. Non-standard stellar models, which include rotation and internal gravity waves, (solid lines) can explain both the decrease in lithium with increasing age and the scatter seen in the sample; the three different curves are the results for a selection of (assumed) initial rotational velocities of the stars (indicated by the numbers next to the curves).

Fig. 2: Evolution of lithium abundance with age for solar twins and metal-rich solar analogues. The solar twin sample is represented by open squares, those which are known to host planets by filled squares. Metal-rich solar analogues with and without detected giant planets are shown with filled and open triangles, respectively. The metal-rich solar analogues seem to follow a lithium versus age trend as well, similar to the solar twins, independently of hosting planets or not, but their lithium content is somewhat smaller than in solar twins at a given age, at least in the 3 to 6 Gyr age range. This difference is expected from the higher metal content, as predicted by stellar models.

The Sun has a very low surface lithium content compared to the initial value in the proto-solar system as inferred from measurements of the lithium abundance in meteorites, which were formed at about the same time as the Sun - 4.5 billion years ago. Lithium is a fragile element that can be easily destroyed in the interiors of stars where temperatures are very high. Since the external layers of stars such as the Sun are convective, surface material can be transported inside the star, reaching regions possibly hot enough for lithium burning. This would mean that the material is depleted in lithium when it returns to the surface; this mechanism therefore could explain the observed low solar lithium abundance. However, theoretical calculations based on standard stellar models have been unable to quantitatively account for the high level of lithium depletion observed on the Sun's surface.

Does the observed low lithium abundance in the Sun make our star peculiar and therefore unreliable to test general models of how stars work? To answer this important question, astronomers at the MPA (Patrick Baumann, Ivan Ramirez, and Martin Asplund), in collaboration with scientists from the University of Porto (Jorge Melendez, now at the University of Sao Paulo) and the European Southern Observatory (Karin Lind), have recently completed a study of solar twin stars, objects whose temperature, radius, and mass are very similar to the Sun.

A careful selection of candidates from a pool of about 100,000 stars was first made using the star's measured colours, brightnesses, and distances. About 100 of these stars were then observed with a variety of telescopes, including the 6.5 metre telescope at the Las Campanas observatory in Chile and the 2.7 metre telescope at the McDonald observatory in the USA. Additional data were gathered from the public archives of ESO and Keck observatories. The unprecedented high quality of the data obtained for these objects allowed the scientists to measure lithium abundances with high precision as well as fundamental properties of the stars such as mass, age, and overall metal content.

One of the most important findings of this research is that the lithium abundances in solar twins (stars of mass and metallicity essentially identical to solar) decrease monotonically with the age of the star. Data for solar twin stars in open clusters with well-known ages fit the observed trend perfectly. Importantly, the Sun does not appear abnormally low in lithium abundance (see Fig. 1).

In standard solar models the Sun does not rotate, and many important ingredients dealing with diverse transport mechanisms in the solar interior are ignored. Thus, it is perhaps not surprising that standard solar models predict only a minor destruction of lithium. Nevertheless, in more realistic stellar models lithium can be destroyed over time. Thus, the trend exhibited by the observed lithium content versus age can be explained by more realistic representations of how the interiors of stars work. In particular, models including rotation and so-called internal gravity waves can successfully reproduce the observed lithium-versus-age trend.

Recently, it has been claimed by a rivalling research group that stars with planets are in general more depleted in lithium than those not hosting planets. However, the age effect described above was not properly accounted for by this other group.

The MPA team also analysed a sample of metal-rich solar analogues, stars with properties very similar to the Sun but an overall metal content that is about 80% larger. This subset of stars was chosen because stars known to host planets, in general, tend to be more metal-rich than those where no planets have been detected yet. Comparing the lithium-versus-age relation for metal-rich solar analogues with that for stars with planets detected shows no significant differences (see Fig. 2).

The metal-rich stars seem to have a low lithium abundance compared to the combined solar twin and solar analogue sample. However, this is due to a metallicity effect, since metal-rich solar analogues have, on average, a lower lithium abundance than solar twins, regardless of whether the star hosts a planet or not. This metallicity effect is predicted by stellar models, which feature a deeper convective envelope in more metal-rich stars. In their previous study, the rivalling team was led to the incorrect conclusion of more lithium depletion in planet-hosts due to this metallicity effect and the fact that planet host stars are on average more metal-rich than stars not known to host planets.

The new study showed that a low lithium abundance in stars can be naturally explained as an age effect, which is metallicity-dependent and does not imply the presence of planets. Both solar analogues and solar twins follow a strong lithium-age trend, suggesting a secular mixing process in the star, which is well-reproduced by several non-standard stellar models. The more metal-rich stars (independently of hosting a planet or not) are somewhat more depleted in lithium. On the other hand, this research shows that there is no connection between a low level of lithium in solar-type stars (including our Sun) and the presence of planets.


Patrick Baumann, Ivan Ramirez, and Martin Asplund


Further Readings

Baumann, P., Ramirez, I., Melendez, J., Asplund, M., Lind, K. "Lithium depletion in solar-like stars: no planet connection", 2010, Astronomy and Astrophysics, in press, linkPfeilExtern.gifarXiv:1008.0575