D. Turrini1, M. Barbieri2, F. Altieri1, D. Grassi1, A. Adriani1, G. Piccioni1

1 Institute for Space Astrophysics and Planetology INAF-IAPS

2 Center of Studies and Activities for Space (CISAS) “G. Colombo”

The Role of Impacts during the Formation and Evolution of Giant

Planets: Implications for the Atmospheric Composition

“The Search for Exoplanets in Italy” - INAF Headquarters

Atmospheric Composition of ExoplanetsAtmospheric Composition of Exoplanets

To date, we have information on the atmospheric composition of 13 planets outside our Solar System (source: Extrasolar Planets Encyclopaedia, www.exoplanet.eu).

Planet Detected Molecules

1RXS1609 b CO H2O H

2

GJ 436 b CO CO2

H2O H

HD 189733 b CO CH4

H2O H Na

HD 209458 b H2O H TiO VO Na Mg

HD 80606 b K

WASP-12 b CO CH4

H2O H

2TiO VO HCN

WASP-17 b Na

XO-2 b K Na

GJ 504 b CH4

HD 179949 b CO H2O

tau Boo b H2O

51 Peg b CO H2O

HAT-P-1 b H2O K

Slide 2/22

Elemental abundances (relative to hydrogen) in Jupiter's atmosphere, compared to solar abundances. Figure from Coradini et al. 2011 (updated from Lunine et al. 2004).

The Galileo mission revealed that the Jovian atmosphere is characterized by a factor 3 enhancement of C, N, S and Ar, Kr and Xe (Owen et al. 1999). Jupiter's bulk composition in enriched (3%-13%) in high-Z elements respect to solar (2%) composition (Lunine et al. 2004).

The Case of Giant Planets in the Solar The Case of Giant Planets in the Solar SystemSystem

Also the other giant planets of the Solar System are enriched in high-Z elements, but the enrichment factor varies from planet to planet and possibly from element to element.

For, e.g., C/H we have: Saturn: 10.40.4 (Fouchet et al. 2009); Uranus & Neptune: 4520 (Guillot &

Gautier 2007).

Slide 3/22

A nice and simple idea, which unfortunately does not work: the erosion of the core, an inefficient mixing inside the envelope or the capture and dissolution of planetesimals in the atmosphere can produce (alone or in conjunction) the same outcomes from both scenarios (see e.g. Helled et al. 2011). Moreover...

For a long time it was thought that enrichment could be used to assess whether giant planets formed by core accretion or gravitational instability (see e.g. Coradini et al. 2010 for a discussion). The basic idea was the following:

Formation by gravitational instability

Formation by gravitational instability

Planet is characterized by solar bulk composition

Planet is characterized by solar bulk composition

Formation by nucleated instabilityFormation by nucleated instability

Planet is characterized by over-abundances in high-Z elements

Planet is characterized by over-abundances in high-Z elements

Atmospheric Composition and Formation Atmospheric Composition and Formation ProcessProcess

Slide 4/22

Late and Secular Accretion ProcessesLate and Secular Accretion Processes

T0

T0

NowNow

Capture planetesimals from distant regions due to resonances (e.g. Primordial Heavy Bombardment –

Turrini, Nelson & Barbieri 2014)

Capture planetesimals from distant regions due to resonances (e.g. Primordial Heavy Bombardment –

Turrini, Nelson & Barbieri 2014) = 1 Ma = 1 Ma

Secularly capture planetesimals while terrestrial planets are forming (e.g. Guillot & Gladman 2000)

Secularly capture planetesimals while terrestrial planets are forming (e.g. Guillot & Gladman 2000) = 1-100 Ma = 1-100 Ma

Capture planetesimals during late migration events (e.g. Matter et al. 2009)

Capture planetesimals during late migration events (e.g. Matter et al. 2009) = 0.1-1 Ga = 0.1-1 Ga

Secularly capture planetesimals in stationary planetary systems (e.g. comets hitting Jupiter)

Secularly capture planetesimals in stationary planetary systems (e.g. comets hitting Jupiter) > 1 Ga > 1 Ga

Slide 5/22

… to further complicate things, we know from the Solar System that giant planets can:

Intensity of Late Accretion ProcessesIntensity of Late Accretion Processes

T0

T0

NowNow

Max Max

MinMin

TIME EFFECTS

Slide 6/22

Earlier events have larger effects than later events (the largest, the earliest) and are therefore those affecting the most the composition of giant exoplanets

Later events, however, cover a larger timespan and are therefore more likely to produce transient effects that can affect our observations.

Let's see some examples of these events from the case of the Solar System...

Primordial Bombardments and Late Primordial Bombardments and Late AccretionAccretion

Safronov (1969) originally proposed that the formation of Jupiter should scatter planetesimals from its formation region.

The formation of Jupiter also causes the appearance of mean motion resonances in the inner and outer Solar System (Weidenschilling et al. 2001, Turrini et al. 2011, 2012).

The combined effects of scattering and resonances trigger a phase of remixing of material throughout the planetary system (Safronov 1969, Weidenschilling 1975, Turrini et al. 2011, 2012).

Slide 7/22

Secular Late AccretionSecular Late Accretion

Slide 8/22

After the appearance of the giant planets, the interplay between them and the planetary embryos in the inner Solar System shaped the evolution of the latter (Wetherill 1992; Chambers & Wetherill 2001; Petit et al. 2001; O'Brien et al. 2007).

The mechanism causing the evolution of the inner Solar System is the following:

the planetesimals in the inner Solar System are perturbed by the planetary embryos;

planetesimals get pushed into the orbital resonances in the asteroid belt and ejected;

on a timescale of 100 Ma, about 99% of the mass originally in the belt is lost.

Dynamical clearing of the inner Solar System. Figure from Morbidelli et al. (2012).

Building on the case of extrasolar planets, more dramatic scenarios have been proposed for the evolution of the early Solar System: among these is the Grand Tack scenario (Walsh et al. 2011, Nesvorny et al. 2011).

The Grand Tack scenario have a low probability to produce the Solar System we know (D'Angelo & Marzari 2012), yet the richness of orbital configurations of the extrasolar planets does not allow to rule out we may be a “lucky case”.

Evolution of the Solar System across the Grand Tack scenario (Walsh et al. 2011).

Slide 9/22

Migration and the “Migration and the “Grand TackGrand Tack” Scenario” Scenario

Late Bombardments and Late AccretionLate Bombardments and Late Accretion

Slide 10/22

A “Jumping Jupiters” scenario

proposed to explain the Late Heavy Bombardment is the “Nice Model” (Tsiganis et al. 2005; Gomes et al. 2005; Morbidelli et al. 2005; Morbidelli et al. 2007; Levison et al. 2011).

During the “Jumping Jupiters” evolution, the giant planets accrete solid material but the accreted mass is limited as fewer bodies are around at such late time (Matter et al. 2009).

Evolution of the Solar System across the Nice Model (Gomes et al. 2005).

Secular Delivery and Atmospheric Secular Delivery and Atmospheric ContaminationContamination

Once planetary systems complete the most active and violent phases of their evolution, impacts continue the remixing process that acted across the earliest phases but at a much lower rate.

An example of this process in the Solar System is the impact of comet Shoemaker-Levy 9 on Jupiter in 1994. On average, across the last 19 years the giant planet has been hit by one comet every four-five years.

Results from the Herschel mission (Cavalié et al. 2013) indicate that the spatially-resolved distribution of stratospheric water of Jupiter is a reflection of the impact(s) of SL9.

The contamination by SL9 lasted longer than the average time between impacts → effects of accumulation are in principle possible.

Slide 11/22

Slide 12/22

Two Toy ModelsTwo Toy Models

In the following, we will take advantage of two toy models to illustrate the effects of:

1. Post-formation accretion of a giant planet due to the bombardment triggered by its very birth.

2. Transient atmospheric contamination due to sporadic cometary impacts.

Toy Model 1: JEB/PHB & Late AccretionToy Model 1: JEB/PHB & Late Accretion

We simulated a disk of 2x104 test particles (the planetesimals) under the influence of a fully-formed Jupiter on its present orbit.

To account for the formation of Jupiter's core, we removed the particles between 4.7 AU and 5.7 AU.

The evolution of the system was followed for 2 Ma.

Initial eccentricity of the planetesimals up to 0.1 (Weidenschilling 2008).

Slide 13/22

The bulk of the late accretion of Jupiter is completed in about 3x105 years since the beginning of the simulations, consistently with the duration of the bulk of the JEB (0.3-0.5 Ma, Turrini et al. 2011, 2012).

The total accretion efficiencies are:

No Migration: 4.8%; Fast Migration: 3.3% Slow Migration: 3.8%

Translated into units of M, these become:

No Migration: 1.1 M; Fast Migration: 0.86

M; Slow Migration: 0.75

M;

Slide 14/22

Toy Model 1: JEB/PHB & Late AccretionToy Model 1: JEB/PHB & Late Accretion

Unexpectedly, ~40% of the accreted masses come from the inner region (i.e. volatile-depleted, silicate- and metal-rich) of the planetary system independently on migration.

Slide 15/22

Toy Model 1: JEB/PHB & Late AccretionToy Model 1: JEB/PHB & Late Accretion

The Composition of the PlanetesimalsThe Composition of the Planetesimals

To estimate the mass delivered to Jupiter by the captured planetesimals we assumed that:

Our simulated disk was a Minimum Mass Solar Nebula (Weidenschilling 1977) with a surface density of 2700 g cm-2 at 1 AU and a R-3/2 density profile (Coradini et al. 1981);

In the outer Solar System the total mass of the disk was reduced by 5 Earth masses to account for the gap due to the formation of Jupiter's core.

Our disk was divided into 3 regions with different planetesimal composition:

Region A (1-3 AU): planetesimals are assumed similar to ordinary chondrites (~25 wt% Fe, ~19 wt% Si, ~05 wt% C, ~1.8 wt% S);

Region B (3-4 AU): planetesimals are assumed similar to carbonaceous chondrites (~25% Fe, ~14% Si, ~2 wt% C, 1.8 wt% S, ~10 wt% H

2O);

Region C (4-10 AU): planetesimals are assumed similar to comets (half carbonaceous chondrites, half as ~74 wt% H

2O, ~24 wt% C, ~0.7 wt% N, ~1.2 wt% S);

Slide 16/22

Source Regions and Jupiter's Source Regions and Jupiter's EnrichmentEnrichment

The accreted mass in our Solar Nebula is equivalent to 1.1 Earth masses.

The mass accreted from Region A (red area) is 1.7x1027 g.

The mass accreted from Region B (black area) is 8.5x1026 g.

The mass accreted from Region C (blue area) is 4x1027 g.

Slide 17/22

Planetary and Atmospheric Planetary and Atmospheric EnrichmentEnrichment

If the accreted mass of 1.1 Earth masses is distributed:

over the whole planet, the enrichment is still low (16%) and the Jovian high-Z fraction is only 2.4% instead of the lower limit of 3% (Lunine et al. 2004).

over the molecular shell with thickness of 5000 km, the produced enrichment is of about a factor 2;

Slide 18/22

Element/MoleculeEnrichment

(no migration)Enrichment

(fast migration)Enrichment

(slow migration)

Fe 5.9 4.4 4.8

Si 6.44 4.7 5.2

C 1.5 1.4 1.4

N 1.2 1.1 1.1

S 2.5 2.0 2.2

H2O 2.5 2.0 2.2

Atmospheric Mixing RatiosAtmospheric Mixing Ratios

In the considered scenarios, diluting the accreted mass in the molecular shell of a Jupiter-like exoplanet would result in mixing ratios that could be detected (and in principle discriminated between one case and another) by future EChO-like space missions and the JWST.

Slide 19/22

Element/Molecule

No Migration Mixing Ratios

No Migration vs Solar

Abundances

No Migration vs Fast Migration

No Migration vs Slow Migration

Slow Migrationvs Fast Migration

Fe 10-3 6.610-3 10-4 10-4 10-4

Si 10-3 10-4 10-4 10-4 10-4

C 10-3 10-4 10-4 10-5 10-5

N 10-4 10-5 10-5 10-6 10-6

S 10-4 10-4 6.610-5 10-5 10-5

H2O 10-3 10-3 10-3 10-4 10-4

Toy Model 2: Secular DeliveryToy Model 2: Secular Delivery

We considered the planet HD 189733 b as our test case: orbital and physical parameters for star and planet were obtained from the Extrasolar Planets Encyclopaedia.

As impactors, we considered the Sun-grazing comets observed by SOHO: their orbital parameters were obtained from the JPL Small Bodies Database Search Engine.

Sun-grazing comets observed by SOHO have too high orbital inclination: a preliminary estimate gives 1 impact every 200 years.

Assuming an ecliptic population of Sun-grazing comets: a preliminary estimate gives 1 impact every 20 years.

In the latter case, impact rate is comparable to survival of contaminants in the atmosphere of HD 189733 b.

Slide 20/22

Toy Model 2: Secular DeliveryToy Model 2: Secular Delivery

Using water as our tracer (72.8% of the cometary mass, Mumma & Charnley 2011) and assuming the dissolution and mixing of the cometary impactor in an atmospheric shell of thickness ΔR (in km) of a Jupiter-like planet, the resulting mixing ratio ξ can be approximated as:

In principle, the dissolution of a 5 km wide comet (i.e. Shoemaker-Levy 9) into a 300 km thick atmospheric shell (i.e. the Jovian stratosphere) could produce a mixing ratio consistent with the lower end (~10-5) considered by Tinetti et al. (2007) for HD 189733 b.

ξ=2.98×10−6 χ ( D c

1 km )3

( ρatm2×10−5 kg m−3 )( Δ R

100 km )−1

where Dc is the diameter of the comet in km, ρ

atm is the density of the atmospheric layer

expressed in kg m-3 and χ is the fraction of the comet dissolved in the considered shell.

Slide 21/22

Slide 22/22

Wrapping up...Wrapping up...

First, the interaction of a giant planet with the surrounding disk of planetesimals can result in the capture of metals and silicate-based refractory materials (whether these materials can actually stay in the atmosphere for us to see them or not is a completely different question...)

Second, the secular impacts of comets or asteroids on an exoplanet bring transient contaminants in its atmosphere. If the delivery rate is comparable to the removal rate of the contaminants, in principle we can have a prolonged non-equilibrium situation.

…here are the three main points that these toy models want to drive.

Third, current technology allows for enough resolution to differentiate between the different formation and evolution scenarios: future observations will therefore allow probing the past of hundreds of exoplanetary systems.

For these (and more) results, see Turrini, Nelson & Barbieri, Exp. Astron., DOI: 10.1007/s10686-014-9401-6 (also available at http://arxiv.org/abs/1401.5119).