Disks and the Origins of Planetary Systems
Sanford Davis, Lead Co-Investigator
Co-Investigators: Jack Lissauer, Greg Laughlin, David Hollenbach, Uma Gorti,
Denis Richard, Kevin Zahnle
OBJECTIVE 2: Develop simulations that predict, in a probabilistic sense, the diversity of planetary systems that emerge during accretion of planets from proto-planetary disks, with a particular focus on understanding the resultant juxtaposition of the chemical raw materials, energy, and physicochemical environments that are required to sustain prebiotic evolution and the emergence of complexity.
Objectives, Expected Significance,
and Extending the State of Knowledge
In this investigation we address Objective 2; we are concerned with the increasing complexity of planetary bodies as they form in habitable zones of protoplanetary disks. The planet formation process begins with fragmentation of large molecular clouds. Such collapsing rotating fragments maintain angular momentum by flattening at the poles and extending themselves into a protoplanetary disk. This disk is in many ways an astrochemical "primeval soup," in which cosmically abundant elements are assembled into increasingly complex prebiotic species. An important issue that we will address is the competition between thinning out of the disk (disk dispersal) and the process of planet formation which requires a disk lifetime of about 10 million years. During this period embedded dust grains are extremely important, for they act as both reaction and condensation sites for the capture of important prebiotic compounds such as water and methane. These condensates, enhanced by these prebiotics, combine through collisions into small agglomerations and then into planetesimals. Gravitational attraction between the planetesimals leads to the accretion of Moon-sized objects. Depending on the origin of these embryos, different mixes of prebiotic molecules and volatile species are delivered to the terrestrial planets. The heat generated during accretion induces stratification into planetary cores, mantles, and primitive atmospheres. If the newly formed planet is a suitable distance from its star to support liquid water at the surface, it is in the so called "habitable zone." Such planets are prime candidates for biological evolution.
Early planetary sciences research considered in-place condensation as a key element of the planet formation process. Condensates in a chemically equilibrated solar nebula were described by Grossman (1972) and Lewis (1972). The models predicted a specific condensation sequence radially stratified by temperature and pressure. This theme of radially-stratified condensation was refined and expanded in the following decades. This research is examined in a number of review articles (Arrhenius 1978, Fegley and Prinn 1989, Prinn 1993). This early work mostly concerns the hot inner solar nebula where terrestrial planets form. In other reviews Lunine, Owen, and Brown (2000) considered chemistry of the outer solar nebula (defined as the region beyond the water-ice line at about 5 AU) and emphasize the role of water in the evolution of the outer planets. In recent work (Salyk et al 2008), the presence of water in terrestrial planet forming zones has definitely been detected. The question we ask is how prebiotic compounds formed and processed globally in the disk undergo condensation and are transported to habitable zones in both solar and extra-solar protoplanetary disk systems.
The chemical/physical processes in protoplanetary nebulae now point to a more complex picture with significant material interchange both radially and vertically in the disk. This interchange is driven by dynamic and thermal effects and includes turbulent mixing, grain settling, the drift of grains and solid objects with respect to the gas and species diffusion. In particular, the ramifications of UV and X ray radiation effects on the upper layers of the disk and the subsequent transfer of the radiated (and chemically transformed) products to the habitable zones are considered. Once the gas in the disk is dispersed, condensates form embryos and planetesimals which ultimately produce planets which may be able to support viable atmospheres. Specifically, we consider the following key processes in the development of an environment suitable for supporting habitable planets:
Investigations 2.1 and 2.2
Chemistry, disk dynamics, and disk dispersal
Planetary systems form in evolving protoplanetary disks, and studying the nature of these disks along with the associated chemistry is essential to understanding how planets are assembled. We have sophisticated thermochemical models of the gas and dust in disks, which incorporate many detailed processes that affect the heating and cooling of the gas and dust, and self-consistently solve for the disk thermal and density structure with disk chemistry (Gorti and Hollenbach, 2004, 2008). We plan to couple the earlier thermochemical steady-state models (Gorti and Hollenbach 2004, 2008) that focus on the disk surface layers where much of the stellar energy is deposited to detailed disk models which focus on the midplane chemistry and ice formation (Davis 2007) in a new, time-dependent, chemical framework. There is evidence from the meteoritic record of significant radial and vertical mixing in the early protosolar disk, and as mixing timescales can be comparable to chemical timescales, time-dependent models are necessary. The integrated models will combine thermal balance calculations that incorporate the various photoionization and photodissociation processes with detailed chemistry that includes ice formation and radial and vertical transport in the disk. Our freeze-out model will include processes such as photodesorption, thermal, X-ray and cosmic ray desorption to remove ices from cold dust grains. We will also include the condensation of gas phase species onto grains to form ice mantles and grain surface chemistry to transform molecules on grain surfaces; a process that can lead to ice composed of, for example, water, methanol, carbon dioxide, and methane.
The evolving nebula will induce a time dependent change in all the terms of the species concentration equation which must also be taken into account. The proposed method will be compared with non-evolving, Lagrangian convection calculations (Aikawa et al. 1997, 1999); evolving Lagrangian computations (Ilgner et al. 2004); and evolving Eulerian computations (Wehrstedt and Gail 2003). Thus, the chemical and physical components of the protoplanetary disk will be highly interdependent. We also plan to investigate processes by which material may be transported over large radial and vertical distances using a variety of turbulence models (Richard and Davis, 2005) along with relevant advective and diffusive processes where time dependent chemistry is included along with disk evolution. The rapid radial transport of ices in meter-sized objects, balanced by vaporization, can enhance elemental abundances of volatile species rich in oxygen and carbon at the "snow line," a process that we will investigate.
The planet formation process in disks is likely concurrent with disk dispersal via photoevaporation (e.g., Lissauer and Stevenson 2007). The rate at which the disk is destroyed and the evolving surface density distribution (which determines the amount of gas and small dust available in a planet-formation zone) determine the likelihood of forming planetary systems and the architecture (e.g., migration) of these systems. Photoevaporative mass loss rates depend on the stellar UV field, and hence mass, leading to the possibility that planetary systems and habitable planets may be more common around stars of certain mass range. Gorti and Hollenbach will use their disk models to study the photoevaporation of disks around stars of different masses and determine the conditions during disk evolution that are conducive to habitable planet formation.
This work is a major extension of reported research by Hollenbach and Gorti on disk chemistry, thermal balance and dispersal mechanisms; Richard on disk dynamics and transport processes and Davis on nebular dynamics, chemical evolution in disks, and the distribution of ices in solar nebula models. Here we will comprehensively address ice formation/composition in the context of terrestrial planet formation and the delivery of prebiotics to these regions.
Terrestrial planet evolution
Scenarios for the accumulation of planets within our Solar System often begin with minimum mass solar nebula models and this is the starting point for this phase of the work. In such models, the surface density profile of the protoplanetary disk is derived from the observed planetary abundances of refractory materials, smoothed out into a disk, and augmented with volatiles to solar composition. Typical minimum mass profiles derived in this manner have surface density dropping off with distance from the sun as r-3/2 (Weidenshilling 1977, Hayashi 1981). Flatter surface density models are derived in other minimum mass solar nebula models (Davis 2005). Such flat profiles provide better fits to most evolutionary models of protoplanetary disks, and they also have surface densities at 5-20 AU that are more consistent with accretion models of giant planet cores (Lissauer 1987, Lissauer et al. 1995, Pollack et al. 1996, Goldreich et al. 2004, Tsiganis et al. 2005) and the outward migration of Neptune (Malhotra 1995). Our viscosity models are anticipated to have more complex radius-dependent decay rates. We will also adopt the surface density models as described in (1) as initial conditions for computation of the evolution of terrestrial planets.
We will use Newtonian and Monte Carlo simulations to generate a wide variety of synthetic planetary systems. A range of mass distributions and initial configurations will be used in these simulations. Time histories from these simulations will enable us to determine the provenience from which the planets are formed and, in turn, specific volatiles that may be present in these regions of the pre-planetary disk.
Atmosphere of a habitable planet
Proper modeling of a planets atmosphere is needed to determine its climate, and therefore habitability. Extrasolar planets orbiting around cool (M type) stars provide an example where 3D global atmospheric circulation enables a habitable zone to exist whereas 1D models predict it would not (Joshi et al., 1997).
Most habitable zone studies thus far have been carried out nearly exclusively using 1D globally-averaged radiative-convective climate models (e.g., Kasting et al., 1993). We will use existing climate modeling tools to study the atmosphere of typical planets from either the simulations of Lissauer, the planet-finding algorithms of Laughlin, or the results of specialized missions such as Kepler. Hollingsworth is leading this work.
The specific tasks mentioned above form an extremely complex chain of events and encompass much of the Astrobiology discipline. We plan to examine only those processes that easily couple to ongoing NASA missions, are relevant to Astrobiology, and help explain the biotic path from molecular clouds to habitable planets. Outcomes of this work will be applied to mission derived data as applicable and may affect the design of future missions. Hollenbach and Gorti have been awarded a Spitzer grant to exploit the Spitzer data archive and compare their disk models with infrared observations of protoplanetary disks. In addition, they will propose similar observations and analysis to the Herschel and SOFIA missions. They have a long history of applying theoretical models to mission data. Lissauer will apply the results of the planetary growth calculations that he will perform to analyze and interpret extrasolar planets discovered during the ongoing COROT and Kepler missions respectively.
Technical Approach and Methodology
In this section we consider aspects of the path from primitive carbon molecules in the interstellar medium to the evolution of terrestrial planets. These molecules, which include the PAHs discussed in the research section on Cosmic Distribution of Chemical Complexity, are processed by the disk chemistry into a number of important prebiotic species. The disk environment encourages the formation of a number of ices which capture and transport many interesting molecules such as water, methane, and other hydrocarbons. These distributions are determined by classical gas-phase chemistry, grain processes, and photoevaporation. In addition to disk chemistry, we include radial and vertical transport along with the condensation/vaporization of ices. We ultimately determine the spatial distributions of chemical abundances and surface density for both the gaseous and solid components. The resulting disk surface density is used as input for the planet formation process.
We consider planetary growth, using the disk surface density as an initial condition, during which primitive bodies grow and can move closer or farther from the parent star and condense into terrestrial planets. These dynamic simulations are used to model the late stages of planetary accretion. These models show a diverse range of outcomes indicating a complex relation between initial conditions and the final planets. The terrestrial planets accrete material from a range of radii, the molecular abundances of which are output from our chemical/dynamical models. These constituents ultimately determine the volatile inventory of each planet.
Once terrestrial planets are identified, and initial volatile inventories established, we investigate a selection of earth-like planets to assess the presence of an atmosphere and associated climatic processes. We will use modifications of existing global climate models as well as simpler versions that would be appropriate to extrasolar planets. These climate models, in turn, will be used to model atmospheric chemistry processes in these systems.
Chemistry and Disk Dynamics
Complex chemical networks are now routinely considered for application to molecular clouds and protoplanetary disks (Aikawa et. al. 1997, Ilgner et al. 2004). We extend our previous work to emphasize those processes that feed into evolution of terrestrial planets. Here, for the first time, we plan to include models that deal with interactions between the (observable) disk surface and the mass-loaded midplane. We will investigate in a coordinated manner the mutual effects of differential disk heating, prebiotic species transport, dust/particle settling, and surface chemistry effects. Important outcomes of this work are to predict initial conditions for terrestrial planet formation and to relate observational data from the optically transparent region of the disk to conditions in the (unobservable) midplane where such planetary systems actually form.
The surface layers of the disk are the most interesting area for the following reasons: (1) It is the site of disk IR emission which is a major observable; (2) It is the region where disk photoevaporation originates; (3) It is a region of high rates of photodissociation and photoionization and transport of species from this layer to the midplane and can have important chemical effects. For example, X rays can ionize helium, and the helium can react with CO to destroy it. The resulting O atoms end up as water ice. This process effectively converts CO to water ice and leaves carbon to form hydrocarbons. We will also include mechanisms involving grain surface chemistry involving, for example, water and methane ice formation and the adsorption/desorption of ice mantles as well as the transport of these solid species into the disk midplane.
Vertical and radial transport of matter is an essential component of the time-dependent solution. We will consider vertical mixing due to turbulent diffusion in a viscously evolving disk using the alpha and/or beta viscosity models. At a given spatial location, this will be solved coupled with the chemistry continuity equation for the net production rate of every species. Dust settling is also considered by using a settling algorithm (Dullemond and Dominik 2005), as dust particles drift with respect to the gas in the presence of turbulence. We will also address the important issue of grain-surface chemistry and will use models that include differential layering of prebiotic species onto particles as they traverse different regions of the disk (Charnley et al. 2001).
Water is one the least volatile species that acts as an important oxygen reservoir. A recent paper (Davis 2007) invokes simple physical modeling to predict relative abundances of both ice and vapor in the colder regions of a typical disk. This work included a new formulation of the thermal adsorption/desorption process as incorporated in the chemical network. An example of water ice abundances for an evolving nebula is shown in Figure 1. At early times the disk is dense and hot and extensive regions of water vapor are evident in the viscously heated midplane and the inner disk close to the illuminating star (non shaded region in the figure). What is noteworthy is that there are two branches that meet at a cusp. This induces a kind of overhanging cloud that persists for longer times. Cuzzi and Zahnle (2004) show that large ice gradients induce vapor phase enhancement in the hotter region closer to the central star. They surmise that this effect can contribute to the planet-building process. This model will be augmented by advection, diffusion and sedimentation that will redistribute the water ice.
Major transport processes in accretion disks are turbulent diffusion and grain settling. Another first-order effect is large scale advection by meridional circulation. Turbulent diffusion is generally described by a turbulent viscosity prescription such as the classic α model. Meridional circulation can be taken into account only with models with two or more dimensions; in the classic one-zone approximation (Shakura and Sunyaev, 1973), all quantities, including the radial drift, are vertically averaged and vertical velocities vanish. In disk models with a vertical structure, a more complex velocity field for the gas phase can be derived from the equations of conservation of mass and momentum.
The computed velocity field in Figure 2 shows an example of the meridional circulation field that we will consider. It is very different from unidirectional constant accretion; the more complex velocity field indicates outflow from the midplane of the disk and the upper atmosphere, a dominant inflow at mid-altitudes. Integration along the vertical shows a net inward mass flux.
Keller and Gail (2004) have shown, using a similar model, but computed from power laws instead of a self-consistent two-dimensional model—that the effect of such a meridional circulation on the transport of a tracer can be significant. Takeuchi and Lin (2002), deriving a comparable radial flow, also showed the effect on dust transport and gas to dust ratio, but did not consider the vertical gas flow. Our approach will enable a more consistent description of the impact of this meridional circulation.
Photoevaporation and disk dispersal
We plan to use the above time-dependent thermo-chemical disk models to study the evolution of the surface density distribution in the disk as it is being simultaneously photoevaporated by high energy radiation from the central star. We will identify important cooling and heating processes in the upper regions of the disk, to determine the density and temperature distributions.
As the gas flows outward from the disk surface towards the interstellar medium, it may cool, reducing the mass loss rate, or may get heated to even higher temperatures and be further accelerated. Flow dynamics are necessary to resolve this complication and to obtain accurate photoevaporation rates. As the full problem is computationally intensive, flow hydrodynamics will be computed using some of the techniques described above using simplified heating/cooling terms and chemistry.
We will use chemical/dynamical model to identify important cooling and heating processes in the upper regions of the disk. We will determine photoevaporation rates in the habitable zones of disks (obtained from our thermal structure models) and see how these affect the formation of terrestrial planets in the habitable zone. Rapid rates of photoevaporation will carry small sub-micron sized dust particles with the gas flow before the dust begins to coagulate to larger objects, whereas slower rates may allow planetesimal build-up in disks. We will also determine how photoevaporation affects the formation of gas giants. This is an important issue since they affect the formation and habitability of the smaller terrestrial planets. Finally, we will determine how photoevaporation may affect formation of asteroids since the volatiles in these bodies may be reservoirs for terrestrial planets in the habitable zone.
Determining disk lifetimes and constraining planet formation timescales will be a direct result of our study. The determination of the surface density distribution of the disk as a function of time supplemented by comparisons of our model results with observational data (e.g., from Spitzer, SOFIA and Herschel) will help understand the early environments and conditions under which planets form. We will also infer disk lifetimes as a function of stellar mass. Longer lived disks may also encourage the formation of planets with circularized orbits which may be more conducive to habitation.
Evolution of terrestrial planets
The disk modeling/evaporation efforts predict spatial distribution of matter in the habitable zone of a protoplanetary disk at the start of the planet formation epoch. The next stages of planetary evolution are local in the sense that they do not involve mixing over large distances. We will skip these intermediate steps and move to consider the chaotic later stages of terrestrial planet growth, during which nascent bodies can move closer to or farther from the parent star. Newtonian N-body simulations are used to model the late stages of planetary accretion (Agnor et al. 1999, Chambers 2001, Raymond et al. 2004, 2007, O'Brien et al. 2006). These simulations show a diverse range of outcomes as a consequence of deterministic chaos (Figure 3), so multiple simulations with very similar initial parameters must be run in order to obtain a statistically robust distribution of results.
The ultimate distributions of terrestrial planet systems depend upon giant planet configurations as well as the surface density profile of the disk within the terrestrial planet growth zone. Thus, in order to do the 'inverse problem' of mapping the observed distribution of planetary systems into initial disk profiles, we will need to first generate a 'library' consisting of an ensemble of many realizations for each of a variety of initial disk profiles and giant planet configurations. We already have such a library, consisting of more than 30 realizations, for a disk similar to the minimum mass solar nebula and giant planets like Jupiter and Saturn around a single solar mass star, and much smaller collections of results for a similar disk around a single star without giant planets and within various binary star configurations (Quintana et al. 2002, 2007, Quintana and Lissauer 2006). The codes that we have used for these simulations (Chambers 1999, Chambers et al. 2002) are robust and well tested. We have also developed an ensemble of statistics that quantify various characteristics of the planetary systems output from our integrations (Quintana et al. 2002).
Our first runs will model growth closer to stars than the standard region followed by terrestrial planet formation simulations. These runs will examine a poorly-characterized region of phase space, and one that is of particular importance to COROT and Kepler, because the detection statistics of both of these space missions will be strongly biased in favor of close-in planets. Although the time steps for these runs will need to be short (because orbital periods are short), the run time is less. The shorter run time is a consequence of bodies nearer the star occupying a greater fraction of their Hill Spheres, implying that accretion requires fewer orbits to occur (for a theoretical analysis, see Greenzweig and Lissauer 1990, 1992; for an example concerning planetary accretion around stars of differing mass see Lissauer 2007; for the more extreme case of accretion of Earth's Moon, involving bodies that almost fill their Hill Spheres, see Stevenson 1987, Ida et al. 1997, Canup 2004).
For stars near the bottom of the main sequence, a downward extrapolation of the minimum-mass solar nebula (MMSN) might not be warranted. Sub-mm flux measurements of low-mass primary systems suggest no clear correlation between stellar-mass and disk mass for low-mass stars (Andrews and Williams, 2005). We have recently completed a suite of accretion simulations (Montgomery and Laughlin 2008) that study the formation of potentially habitable terrestrial planets orbiting low-mass red dwarfs. In these models, the presence of a Neptune-mass "failed core" beyond the disk's ice line meters the inward flow of gas through the gap that it has opened in the protoplanetary disk. This leads to a reduced gas-density in the interior orbital regions and significantly reduces the deleterious effect of Type I migration. Terrestrial planets then form in a manner that is more akin to the formation of Jovian satellites (e.g. Canup and Ward 2006), and achieve a total planet-to-star mass ratio of order 2x10-4, as is observed for the regular satellite systems of Jupiter, Saturn and Uranus (see Figure 4). We are eager to widen our library of planetary systems by including the results of formation scenarios for low mass stars that depart from simple downward extrapolation of the MMSN.
Once we have compiled a 'library' of artificial planetary systems for a given disk configuration, we will use this library to determine distributions of transit observations from COROT and Kepler for each of the resulting systems. Each system will be 'viewed' from a large number of randomly chosen points on the celestial sphere. These simulations will yield a distribution of observed transiting planets and of systems with multiple transiting planets; the characteristics of these systems will be available for comparison with COROT and Kepler findings.
Kepler is capable of finding true Earth analogs, planets the size of Earth that orbit 1 AU from a Sun-like star; COROT is limited to larger planets orbiting closer to their stars. A prime reason why this parameter range is of interest is that Earth is the only known planet suitable for life. In addition to size and location, a key factor that determines habitability by familiar forms of life is the presence of moderate quantities of volatile compounds. Kepler will not be able to assess the abundance of such compounds, and because of the distance to Kepler targets, follow-up observations are unlikely to either. However, our simulations of the late stages of planetary growth will tell us the material that each planet is composed of and where it came from. Our work on protoplanetary disks, in turn, will predict the composition of solid material in the disk as a function of radial location. Thus, by combining these models with Kepler observations we will be able to estimate the volatile content of observed terrestrial planets in the habitable zone.
Atmospheres and climates of habitable planets
Once suitable terrestrial planets are identified in habitable zones, we propose to investigate these planets to assess the presence of an atmosphere. We will apply a small suite of general circulation models that are either simplified (although still fully 3D) or fully complex to investigate habitability (e.g., Hollingsworth et al., 2007). NASA Ames has a long heritage in global circulation modeling for Mars atmosphere using the Ames Mars GCM (e.g., Haberle et al., 1999; Kahre et al., 2006). This 3D climate code, with its newly updated radiative physics package is much more portable and adaptable for use in investigations of atmospheric circulation and planetary climates with realistic radiative heating for a general range of extrasolar, terrestrial-like atmospheres having various chemical compositions (e.g., CO2/H2O atmospheres) and orbital configurations about different Sun-like stellar objects. In addition, the Ames climate model also includes a cloud microphysics package for both CO2 and H2O-ice clouds (Montmessin et al., 2004) that could be generalized to other chemical species and condensates. Together with the new radiative package, cloud multiple scattering in the presence of gaseous absorption in both short (e.g., solar-like) and long (infrared) wavelengths can be accounted for.
The GCM codes are relatively straightforward for a particular set of realistic configurations to examine extrasolar terrestrial-like atmospheres and circulation and to explore the nature of the habitable zone in a fully 3D climate context. Numerical studies of the extrasolar planet's orbital configuration (e.g., obliquity, eccentricity and longitude of periapsis) will be conducted for terrestrial-like planets in order to explore the range of seasonal/annual near-surface temperature extremes. We also plan to investigate climate extremes using more simplified land-surface prescriptions (e.g., an aqua-planet or an aqua-planet with equatorially confined continents) to assess near-surface temperature extremes for different orbital configurations.
We plan to utilize a so-called "simple physics" 3D global circulation model (i.e., with simplified right-hand side forcing terms of the governing fluid equations) as exemplified by Hollingsworth et al. (2007), who investigated dynamical mechanisms for Venus' super-rotation. Such a model is fully 3D and global but removes highly specific atmosphere/climate physical processes and tuning-parameters that are frequently embedded in full general circulation models (GCMs) for terrestrial planetary atmospheres (e.g., Earth and Mars). Such climate modeling tools are well poised to perform mechanistic and idealized studies of the impacts of circulation on the nature of the habitable zone. Certain outcomes from this work are used directly in the atmospheric chemistry effort described below.
Chemistry and Climate of a Prebiotic Atmosphere
OBJECTIVE 3: Develop simulations of atmospheres and climates of planets that form in circumstellar habitable zones, and simulate the chemical consequences in early atmospheres of solar radiation and impacts. Identify the range of climates of potentially habitable planets. Investigate the carbon chemistry of cold dry atmospheres and also impact shock chemistry, as these might affect the chemical inputs to prebiotic evolution.
Objectives, Expected Significance,
and Extending State of Knowledge
We will pursue Objective 3 by investigating the carbon chemistry of cold dry atmospheres as well as the chemical transformations that can occur in the aftermaths of large impacts. Both of these are likely scenarios in early habitable environments and they might have provided reduced chemical species for prebiotic evolution.
After Earth had cooled from the Moon-forming impact, it was left with a ~100 bar CO2 atmosphere and a surface temperature of ~500 K (Zahnle et al 2007). This CO2 atmosphere was then largely removed into the mantle or into carbonate rocks by rock weathering, hydrothermal alteration of the seafloor, and subduction. Once subduction of carbonates began in earnest it might have taken as little as 10 million years to remove 100 bars of CO2 into the mantle (Zahnle et al 2007).
A key aspect of the Hadean era (the pre-4 Ga period of Earth history) is that the Sun was 30% fainter then than it is now (Sackmann et al 1993) and imposed a stringent constraint on early climates (Sagan and Mullen 1972). A Hadean temperate climate would have required either enormous geothermal heat flow or abundant potent greenhouse gases. Geothermal heat was insignificant for climate after 4.4 Ga, save immediately following major impacts. The best candidate greenhouse gases are CO2 and CH4. Water vapor is in fact the most important greenhouse gas, but its atmospheric abundance is a dependent variable. CH4 can contribute substantially if there are reducing agents or catalysts to generate it from CO2 and H2O. Biological processes are the dominant sources of CH4 on Earth today. So CH4 is a good candidate for keeping Earth warm once it teemed with life, but it is not clear that it had a substantial source when Earth was lifeless.
About a bar of CO2 would have been required to provide enough greenhouse warming to stabilize liquid water at the surface (Zahnle et al 2007), but would necessitate that several bars of CO2 were dissolved as bicarbonate ions in the oceans. This is much more than could be sustained by a seafloor weathering cycle resembling today's cycle (Sleep et al 2001). With no obvious geochemical buffer to maintain this high level, CO2 levels were probably much lower (albeit still higher than today), and therefore climates were much colder and drier. The two dashed curves in Figure 5 relate surface temperature to pCO2 according to Kasting's greenhouse warming models (Kasting 1991) for early (faint Sun) and current Mars. Although these simulations are not directly relevant to the Hadean Earth, an ice-covered Hadean Earth might have comparably cold, and thus CO2 would have been less stable relative to CO.
CO is significant for early habitable environments because it is one of the easiest molecules to synthesize in plausible abiotic atmospheres and it might have participated in prebiotic chemistry. CO is packed with energy and therefore precious for life. Modern methanogens that use CO2 first convert CO2 to CO with one enzyme (CO dehydrogenase, or CODH), and then transfer the CO as a gas to a second enzyme complex where it is used for energy or for cell material. Organisms that can use CO directly consume it very quickly, because with CO an organism can split water. These are ancient metabolisms. The key enzymes are both based on NiFeS cubes, fitting well with a widespread speculation that the first metabolism made use of natural iron sulfides as catalysts (Ragsdale 2004).
The following investigations address key atmospheric processes that probably influenced prebiotic chemical evolution during the Hadean Era on Earth and perhaps also affected early habitable environments on other planets.
Explore the photochemical production of CO as a potentially abundant prebiotic molecule in plausible early atmospheres.
A cold and very dry atmosphere is likely to be CO rich. If an atmosphere is cold and dry, CO can be more photochemically stable than CO2. A key fact about the Hadean is that the Sun was only about 70-75% as luminous as it is now (Sackmann et al 1993). In the absence of potent greenhouse gases, the Hadean Earth should have frozen over and mostly as white as ice. Together, the high albedo of ice and the faint Sun could reduce the average surface temperature of Earth to 200 K. In recent work we have found that a cold, thick CO2 atmospheres on early Mars is unstable with respect to photochemical conversion to CO (Figure 5). The key reason why CO2 becomes unstable is that recombination of CO2 from CO is catalyzed by water vapor photolysis: a very cold atmosphere is very dry, and thus congenial to CO. As a consequence a cold CO2-rich atmosphere can be photochemically converted into a cold CO-rich atmosphere, with the excess oxygen leaving the system as H2O2. A CO-rich early atmosphere may well make a cold young Earth seem a more attractive venue for the first steps toward life.
Explore the consequences of impact shock chemistry in plausible early atmospheres, with particular attention paid to the role played by the impactor's composition in determining the chemical composition of the post-impact atmosphere.
The late bombardment of solar system debris was another factor affecting the Hadean surface environment. By creating mountains of highly reactive mafic and ultramafic debris, big impacts would have enhanced the geochemical sink on CO2, strengthening the case for a generally cold Hadean. But in its immediate aftermath the energy released by a big impact could transform a snowball Earth into a water world. A very big impact would transform the snowball into a sauna or even, briefly, a furnace with surface temperatures exceeding the melting point of rock (Figure 6). It could have taken a long time – hundreds, thousands, or even tens of thousands of years, depending on the greenhouse gases that might have been synthesized by the impact or injected into the atmosphere, and leveraging the relatively low albedo of liquid oceans – for the climate to return to the more stable ice-cold state. There were probably thousands of impacts large enough to trigger impact summers.
Impacts in the Hadean have been discussed as a source of HCN in reduced atmospheres (Fegley et al. 1986, Fegley and Prinn 1989, Chyba and Sagan 1992) and as a source of CO in a CO2 atmosphere (Kasting 1990). The latter in particular stressed the role that metallic iron in the meteoritic material (much of which is vaporized and dispersed in the atmosphere by the impact) would play chemically reducing CO2 to CO.
These earlier studies greatly understate the importance of impact shocks and the chemical influence of the impacting materials. This is because the earlier studies treated impacts as a steady, low-to-moderate level influence on the chemistry of the atmosphere, rather than as discrete stochastic events. In the Hadean there would have been tens or even hundreds of impacts big enough to heat a standard one bar atmosphere to >1500 K; i.e., on tens to hundreds of occasions the entire atmosphere would have shock heated and chemically reset to something far from its usual state. Thus the true impact of impacts differs profoundly from a perturbed steady state. At most times and most places the impacts do nothing, while at rare times they are everything. In such events the particular composition of the impacting body can count for a lot.
Recently Hashimoto et al (2007) revisited the composition of atmospheres in thermochemical equilibrium with carbonaceous chondritic impactors (Figure 7). They show that CO and H2 are major products in these simulations. At somewhat lower temperatures CH4 is expected to replace CO as a major gas; it is not known if methane formation is kinetically inhibited.
NO production in an O2-N2 atmosphere is well described by a quench temperature on the order of 2000 K. It is not obvious that CO production can be well described by a quench temperature, because the reaction CO + OH ---> CO2 + H is fast at all temperatures. This makes the quench temperature, and therefore the CO/CO2 ratio, sensitive to water. It is also interesting that CH4 is expected to be very abundant at the 500 K surface temperature that is expected for a thick CO2 atmosphere on early Earth, both after the Moon-forming impact and after the occasional big impacts that punctuated the Hadean thereafter. These 500 K atmospheres cool very slowly, because the cooling rate is determined by CO2 removal by geochemical weathering. It is interesting to ask whether warm methane atmospheres are an inevitable consequence of big impacts.
Technical Approach and Methodology
Chemistry of cold, dry atmospheres
The photochemistries of cold, dry Hadean atmospheres are readily addressed using 1D atmospheric photochemistry codes. We have already developed a model for early Mars that appears to be fully self-consistent and debugged. The version to be adapted for this investigation is based on the photochemistry model for early Earth described by Kasting et al (1989). Many of its features are described by Kasting et al (1989) and Kasting (1990). We have successfully used the code to study sulfur photochemistry of Earth's atmosphere during the Archean (Zahnle et al. 2006b, Zahnle et al. 2008), it would be straightforward to adapt this model further to a cold, ice-covered Hadean Earth.
Impact shock chemistry
We will address the kinetically limited cooling histories of plausible impact-modified atmospheres using a modified 1-D photochemical code in a time marching mode. We will consider only the big impacts that heat the whole atmosphere to very high temperatures where thermochemical equilibrium holds. The model needs to be supplemented by a full suite of high temperature reactions from the combustion literature, with care taken to ensure that every reaction is paired with a self-consistent reverse reaction. The chemical system to be considered will be limited to simple species containing C, H, and O. The critical reaction path for methane formation from CO is hydrogenation of formaldehyde (HCHO) to eventually form methanol. The key reaction is known to have high activation energy and its import has been extensively discussed in the context of jovian planets (Yung, 1988).
The chief reason for investigating the time evolution of a 1D model (rather than a single parcel) is that photochemistry is likely to be important in the Hadean when the Sun was a much stronger source of UV than it is today. The vertical structure will be treated as crudely as possible given the uncertainties associated with vertical mixing (without which high altitude photochemistry could not influence the deeper atmosphere). The temperature structure will be treated as an adiabatic lower atmosphere connected to an isothermal stratosphere at the skin temperature, which is the approach used by Kasting (1988) to describe radiative transfer in water-rich atmospheres.
To prepare for numerical experiments to investigate post-impact chemistry, we will adapt reaction networks, modify the model's vertical resolution of the atmosphere, and adopt proper time integrations for experiments. We will then perform experiments for a variety of post-impact atmospheric compositions and conditions.
AMES TEAM RESEARCH
Early Habitable Environments and the Evolution of Complexity
---David J. Des Marais, Principal Investigator
Cosmic Distribution of Chemical Complexity
---Lou Allamandola, Lead Co-Investigator
Disks and the Origins of Planetary Systems
---Sanford Davis, Lead Co-Investigator
Mineralogical Traces of Early Habitable Environments
---Tori Hoehler, Lead Co-Investigator
Origins of Functional Proteins and the Early Evolution of Metabolism
---Andrew Pohorille, Lead Co-Investigator