Planetary science

Early start for rocky planets

Article metrics

The chemical composition of stars that host small planets seems to be more varied than that of large planets. This finding may push back the clock for the start of rocky planets and of life around stars other than the Sun. See Letter p.375

On page 375 of this issue, Buchhave et al.1 describe a spectroscopic analysis of the chemical composition of stars hosting planet candidates that have been detected by NASA's Kepler mission. The authors have focused on 152 stars harbouring planet candidates that are a similar size to Neptune or smaller. The key result of this work is that small planets — those with a radius up to four times that of Earth — form under a broader range of environmental conditions than do gas-giant planets. This is an important finding, as it shows that the formation of small planets is a less constrained process than is the formation of large planets. It also implies that small, rocky planets formed at an earlier epoch in the Universe's history than gas giants.

To appreciate this result, flash back to only a few hundred million years after the Big Bang, when the first stars in the Universe had just formed; the Universe today is approximately 14 billion years old. Those stars were composed of only hydrogen and helium, because heavier elements did not yet exist. Metals — defined by astronomers as all elements that are heavier than helium — are manufactured by the fusion of lighter elements in the 'pressure-cooker' cores of stars. When stars explode at the end of their lifetimes, their repositories of metals are cast into space, where they seed molecular clouds to produce new generations of stars that are enriched in heavy elements. Thus powered by the life cycle of stars, the chemical composition of the Universe has evolved from a simple mixture of hydrogen and helium gases to its current state replete with all of the elements in the periodic table.

The chemical evolution of the Universe has profound implications for the formation of planets. Planets are born in 'protoplanetary' disks around nascent stars; both the star and the protoplanetary disk inherit the chemical composition of the parent molecular cloud. A star that consists of only hydrogen and helium will have a disk with that same composition. This makes it unlikely that the first stars in the Universe formed with planetary companions or life; the elements from which planets and our bodies are made did not exist. Even if exotic gas-giant planets composed of hydrogen and helium existed, they could not have harboured pools of liquid water or a veneer of prebiotic chemistry. Therefore, planets that could have acted as platforms for biology did not populate the early Universe.

Fast-forward 13.6 billion years to the present time and the situation is markedly different. Ground-based surveys, a hallmark of the past two decades, have detected hundreds of planets orbiting nearby stars. As the first of these planets were discovered, an interesting correlation was immediately noted: a large fraction of gas-giant planets was being detected around metal-rich stars2,3,4. Conversely, gas-giant planets were rarely detected around stars that had metallicities (metal abundances) lower than that of the Sun5. The general interpretation of this planet–metallicity correlation was that the presence of heavy elements is an environmental condition that increases the density of the protoplanetary disk and therefore the efficiency of planet formation. In the past few years, however, ground-based surveys have also detected super-Earths — planets that have masses from a few to ten times the mass of Earth. Interestingly, the host stars of these low-mass planets do not seem to exhibit the same high metal content as do hosts of gas-giant planets.

NASA's Kepler mission has revolutionized exoplanet science by virtue of its vast number of planet detections — more than 2,000 planet candidates to date. Buchhave and colleagues' analysis is the first comprehensive assessment of host-star metallicity for the super-Earth- and Neptune-sized planets detected by Kepler, and it elevates the trend inferred by ground-based surveys6 to a solid statistical result: small planets can form around stars that have a lower metal content than do the host stars of gas-giant planets. The authors obtained the element composition of Kepler stars by matching observed spectra to reference spectra.

“This result is exciting because it illuminates the history of the environmental requirements for planet formation”

This result is exciting because it illuminates the history of the environmental requirements for planet formation. In the beginning, stars formed in isolation, without planets or life. As the chemical evolution of the Universe proceeded, protoplanetary disks developed with a sufficient inventory of heavy elements to build the cores of planets, whether rocky or gaseous. However, for a gas-giant planet to form, there is a race against time: the core needs to reach a critical mass before the gas-rich material dissipates from the disk, which occurs within about 5 million years7,8. If the core reaches the critical mass before the gas is lost, runaway gas accretion occurs. High metallicity helps the core to win this race. Small, rocky planets can accrete dust and rocky material long after the volatile gas has dissipated from the disk. They then acquire tenuous atmospheres by releasing light, volatile elements as gas or by accreting trace amounts of gas from the disk. This process can continue for 50 million to 100 million years9,10,11.

Unfortunately, it is difficult to use knowledge of metallicity to pinpoint the beginning of the 'age of planet formation', because the chemical enrichment of the Universe is not a monotonic process — some regions inside galaxies may have developed high metallicity rapidly whereas other regions are still metal-poor today. However, knowing that the formation of rocky planets can occur in environments of lower metallicity than those of gas giants implies that there could be some places in the Universe where rocky planets and life got an earlier start than did Earthlings.


  1. 1

    Buchhave, L. A. et al. Nature (2012).

  2. 2

    Gonzalez, G. Mon. Not. R. Astron. Soc. 285, 403–412 (1997).

  3. 3

    Santos, N. C., Israelian, G. & Mayor, M. Astron. Astrophys. 415, 1153–1166 (2004).

  4. 4

    Fischer, D. A. & Valenti, J. Astrophys. J. 622, 1102–1117 (2005).

  5. 5

    Mortier, A. et al. Astron. Astrophys. (in the press); preprint at (2012).

  6. 6

    Mayor, M. et al. preprint at (2011).

  7. 7

    Haisch, K. E. Jr, Lada, E. A. & Lada, C. J. Astrophys. J. 553, L153–L156 (2001).

  8. 8

    Pascucci, I. & Tachibana, S. Protoplanetary Dust: Astrophysical and Cosmochemical Perspectives (eds Apai, D. & Lauretta, D. S.) 263–298 (Cambridge Univ. Press, 2010).

  9. 9

    Lissauer, J. J. Icarus 69, 249–265 (1987).

  10. 10

    Raymond, S. N., Quinn, T. & Lunine, J. I. Icarus 168, 1–17 (2004).

  11. 11

    Raymond, S. N., O'Brien, D. P., Morbidelli, A. & Kaib, N. A. Icarus 203, 644–662 (2009).

Download references

Author information

Correspondence to Debra Fischer.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Fischer, D. Early start for rocky planets. Nature 486, 331–332 (2012) doi:10.1038/nature11199

Download citation

Further reading


    • T. L. Campante
    • , T. Barclay
    • , J. J. Swift
    • , D. Huber
    • , V. Zh. Adibekyan
    • , W. Cochran
    • , C. J. Burke
    • , H. Isaacson
    • , E. V. Quintana
    • , G. R. Davies
    • , V. Silva Aguirre
    • , D. Ragozzine
    • , R. Riddle
    • , C. Baranec
    • , S. Basu
    • , W. J. Chaplin
    • , J. Christensen-Dalsgaard
    • , T. S. Metcalfe
    • , T. R. Bedding
    • , R. Handberg
    • , D. Stello
    • , J. M. Brewer
    • , S. Hekker
    • , C. Karoff
    • , R. Kolbl
    • , N. M. Law
    • , M. Lundkvist
    • , A. Miglio
    • , J. F. Rowe
    • , N. C. Santos
    • , C. Van Laerhoven
    • , T. Arentoft
    • , Y. P. Elsworth
    • , D. A. Fischer
    • , S. D. Kawaler
    • , H. Kjeldsen
    • , M. N. Lund
    • , G. W. Marcy
    • , S. G. Sousa
    • , A. Sozzetti
    •  & T. R. White

    The Astrophysical Journal (2015)


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.