A steamy proposal for Martian clays

Martian clays present a conundrum: the models proposed to explain their formation require conditions that are not predicted by computational climate simulations. Experiments now suggest an alternative scenario.

Clay minerals are found in abundance across the surface of Mars, and many models have been suggested to explain their formation1. These models often invoke the presence of substantial surface water and a warm climate during Mars’s first 500 million to 700 million years. However, computational climate models for early Mars struggle to reproduce such conditions2. On page 88, Cannon et al.3 present a possible solution to this problem: a model for clay formation during the end stages of Mars’s formation that does not require prolonged warm and wet conditions.

During their formation, many planets go through a stage known as a magma ocean, which results from substantial or complete melting of their interiors. Both Earth and Mars went through at least one such phase4. The melting of their silicate mantles led to outgassing of volatile components that were originally incorporated in the solid material. This produced atmospheres containing one or more oceans’ worth of water vapour and other volatile compounds such as carbon dioxide4. As the planets cooled, processes such as water condensation and the escape of gases to space reduced the size of the atmosphere.

Massive early atmospheres are extremely hot and dense, and generate high pressures at the planet’s surface. Under these extreme conditions, water and CO2 exist as supercritical fluids — neither gas nor liquid. Cannon and colleagues performed experiments to show that supercritical water and CO2 can react extremely rapidly with minerals typically found in early planetary crusts to make hydrated silicates — clays. 

Radiometric dating5 of Martian meteorites suggests that crystallization of the magma ocean on Mars occurred within 20 million to 25 million years of the beginning of the Solar System, and that crustal formation took at most another 15 million years. The steam in the planet’s atmosphere would have condensed to form a hot ocean on similar timescales. Cannon and colleagues’ findings suggest that the supercritical atmosphere would have reacted with the crust to form clays during this time (Fig. 1a).

Figure 1 | A model for the formation of primordial clays on Mars.a, For about the first 20 million years of its existence, Mars went through a ‘magma ocean’ phase, during which all, or most, of the planet’s interior was molten. The atmosphere was a supercritical fluid (a phase that is neither gas nor liquid) containing water and carbon dioxide. Cannon et al.3 show that minerals in Mars’s crust could have reacted with steam in the atmosphere (white arrows) to form a layer of clays. b, The magma ocean subsequently solidified, and the atmosphere changed to a gaseous state. During the next few hundred million years, volcanic activity would have generated a fresh layer of material on top of the primordial clays, and asteroid impacts would have churned up the upper crustal layers. c, The authors’ computational simulations show that this chain of events would have resulted in the observed patterns of partially exposed clay.

Magma oceans exist for relatively short periods in the context of geological timescales, and so the authors’ mechanism for clay formation on Mars halts long before most other proposed mechanisms would have even begun. Numerous models have been proposed for the formation of Martian clays during the later Noachian period (4.1 billion to 3.7 billion years ago), including: the alteration of subsurface material through reactions with groundwater6; alteration of crustal material through reactions with water at high temperatures, driven by asteroid impacts7; and surface weathering8. Cannon and colleagues’ proposal does not exclude the possibility of later clay formation, but it does limit the amount of clays that could have formed during the Noachian. Notably, surface-weathering models of clay formation require warm, wet conditions throughout most of the Noachian period — which might have been conducive to life. But the authors’ model is consistent with cold, dry Noachian conditions, which would have been unfavourable for life.

For primordial clays still to be present on Mars today, they must have survived the substantial reworking of the Martian crust that occurred as a result of widespread volcanism, disruptions by asteroid impacts and burial by impact ejecta (Fig. 1b). Cannon et al. performed computational simulations of the physical evolution of the primordial clay layer during this crustal reworking. The simulations’ predictions of the clay content of the Martian regolith (the layer of loose materials, such as dust and broken rock, that overlays the bedrock) and of large clay exposures at the surface are consistent with present-day observations of Mars (Fig. 1c).

According to the simulations, the primordial clay layer remains as a mostly continuous layer buried at depths of 15–25 kilometres under volcanic and impact ejecta, but exposed near impact craters. The simulations also show that the dearth of clays in the northern Martian hemisphere can be explained by the disruption of the primordial layer by the Borealis impact — the collision of a single, large body with Mars that is thought to have occurred in this region. In the southern hemisphere, the buried clay layer might correspond to a low-density crustal layer that has been identified by studies of the gravity and topography of Mars9.

Models of magma-ocean evolution on Earth have sometimes included crustal hydration10, but, unlike for Mars, there is no geological record for Earth that goes back more than 3.8 billion years. The primordial clays on Mars therefore provide a unique window into this hot, early stage of planet formation. For example, they will have compositions that reflect the atmospheric composition before it was altered by the loss of gases to space. By contrast, Noachian clays formed under very different conditions, and will therefore be compositionally distinct. More experimental and modelling work is needed to determine the chemical signatures of the different formation mechanisms. Measurements made by robot missions on Mars, such as NASA’s Curiosity rover and the future Mars 2020 rover, might help to constrain these models.

Some of the assumptions of the primordial-clay scenario will also need to be tested further. Cannon and co-workers’ model assumes that the crustal porosity is initially high, allowing instantaneous alteration of the entire crustal thickness by supercritical fluid. However, clay minerals have larger volumes per unit mass than do the unaltered crustal minerals, and so the formation of clay in the upper crust will cause an expansion that might lower the porosity in this region. This could hinder clay formation at lower levels by sealing off the passages through which supercritical fluid travels to interact with the lower crust. 

Finally, the clays formed in Cannon and colleagues’ experiments have a different mineral structure from that of the vast majority of clays detected on Mars by remote sensing. A second stage of alteration might therefore need to be invoked to produce the structures observed on the red planet11. Further experiments must be performed in the laboratory to identify exactly which clay phases are produced, as a key step towards identifying the primordial clays on the surface of Mars.  

doi: 10.1038/d41586-017-07661-3
Nature Briefing

Sign up for the daily Nature Briefing email newsletter

Stay up to date with what matters in science and why, handpicked from Nature and other publications worldwide.

Sign Up


  1. 1.

    Ehlmann, B. L. et al. Space Sci. Rev. 174, 329–364 (2013).

  2. 2.

    Wordsworth, R. et al. Icarus 222, 1–19 (2013).

  3. 3.

    Cannon, K. M., Parman, S. W. & Mustard, J. F. Nature 552, 88–91 (2017).

  4. 4.

    Elkins-Tanton, L. T. Earth Planet. Sci. Lett. 271, 181–191 (2008).

  5. 5.

    Kruijer, T. S. et al. Earth Planet. Sci. Lett. 474, 345–354 (2017).

  6. 6.

    Ehlmann, B. L. et al. Nature 479, 53–60 (2011).

  7. 7.

    Tornabene, L. L. et al. J. Geophys. Res. Planets 118, 994–1012 (2013).

  8. 8.

    Carter, J., Loizeau, D., Mangold, N., Poulet, F. & Bibring, J.-P. Icarus 248, 373–382 (2015).

  9. 9.

    Baratoux, D. et al. J. Geophys. Res. Planets 119, 1707–1727 (2014).

  10. 10.

    Sleep, N. H., Zahnle, K. & Neuhoff, P. S. Proc. Natl Acad. Sci. USA 98, 3666–3672 (2001).

  11. 11.

    Catalano, J. G. J. Geophys. Res. Planets 118, 2124–2136 (2013).

Download references