Building the Planets - UChicago GeoScigeosci.uchicago.edu/~tdavison/comptonlectures/Slides...Image courtesy of NASA/Hubble Space Telescope Comet Team 11 larger than Earth, 318 heavier

Constructing the Solar System: A Smashing Success

Building the Planets

Thomas M. Davison

Department of the Geophysical Sciences

Compton Lecture SeriesAutumn 2012

T. M. Davison Constructing the Solar System Compton Lectures – Autumn 2012 1

Compton Lecture Series Schedule

1 10/06/12 A Star is Born

2 10/13/12 Making Planetesimals: The building blocks of planets

3 10/20/12 Guest Lecturer: Mac Cathles

4 10/27/12 Asteroids and Meteorites:10/27/12 Our eyes in the early Solar System

5 11/03/12 Building the Planets

6 11/10/12 When Asteroids Collide

7 11/17/12 Making Things Hot: The thermal effects of collisions

11/24/12 No lecture: Thanksgiving weekend

8 12/01/12 Constructing the Moon

12/08/12 No lecture: Physics with a Bang!

9 12/15/12 Impact Earth: Chicxulub and other terrestrial impacts


Compton Lecture Series Schedule

1 10/06/12 A Star is Born

2 10/13/12 Making Planetesimals: The building blocks of planets

3 10/20/12 Guest Lecturer: Mac Cathles

4 10/27/12 Asteroids and Meteorites:10/27/12 Our eyes in the early Solar System

5 11/03/12 Building the Planets

6 11/10/12 When Asteroids Collide

7 11/17/12 Making Things Hot: The thermal effects of collisions

11/24/12 No lecture: Thanksgiving weekend

8 12/01/12 Constructing the Moon

12/08/12 No lecture: Physics with a Bang!

9 12/15/12 Impact Earth: Chicxulub and other terrestrial impacts


Today’s lecture

1 Formation of the terrestrial planets

2 Why are the outer planets so different from the terrestrialplanets?

3 Recent advances such as the Nice Model and theGrand Tack theory

Image courtesy of NASA


From the first two lectures

Several conditions our model of SolarSystem formation needs to meet

1 The near-circular orbits of theplanets

2 The planets orbiting in the sameplane as the Sun’s equator

3 The planets all orbiting in thesame direction as the Sun rotates

Ñ Nebular disk model

4 The inner planets are rocky anddense; the outer planets aregaseous and large

Ñ Today’s lecture

Images courtesy of NASA


Part 1:Building the terrestrial planets



Mercury


Small radius(� 0.38 � the Earth)

High density(second highest of allthe planets)

70% metal, 30%rock: Large, metalliccore

Possible explanationfor small size andlarge core: Giantimpact

Dominant surfacefeature: Impactcraters


Mercury


Small radius(� 0.38 � the Earth)

High density(second highest of allthe planets)

70% metal, 30%rock: Large, metalliccore

Possible explanationfor small size andlarge core: Giantimpact

Dominant surfacefeature: Impactcraters


Venus


Similar size to Earth(0.95 � Earth’sradius)

Covered in thick,dense cloud

Beneath the cloud,there is a geologicallyactive surface

Volcanoes, impactcraters

All craters ¡ 3km

Clouds preventsmaller impactorsreaching surface�1000 craters


Venus


Similar size to Earth(0.95 � Earth’sradius)

Covered in thick,dense cloud

Beneath the cloud,there is a geologicallyactive surface

Volcanoes, impactcraters

All craters ¡ 3km

Clouds preventsmaller impactorsreaching surface�1000 craters


Earth


Most dense of all theplanets

Very active surface

Plate tectonicsVolcanismErosionOceans

Fewer impact cratersthan other innerplanets / moons

Only 182confirmed craters


Impacts on Earth

Image courtesy of Planetary and Space Science Centre, University of New Brunswick/NASA/Google


Mars


Smaller than Venusand Earth

Lower density thanEarth

Evidence of volcanismand water on surface

Surface split in two:Northern Lowlandsand SouthernHighlands

Possibly result of agiant impact


Mars

Image courtesy of NASA/JPL-Caltech

Smaller than Venusand Earth

Lower density thanEarth

Evidence of volcanismand water on surface

Surface split in two:Northern Lowlandsand SouthernHighlands

Possibly result of agiant impact


How did the terrestrial planets grow?


Further collisions between the largest planetesimals(the oligarchs) led to growth of larger bodies

Gravitational focusing allowed largest bodies to grow efficiently

Computer simulations of a colliding population ofplanetesimals allow us to test this hypothesis

These simulations are called N-body simulations


N-body simulations

We test these models by looking at the results to:

See if the outcome results in � 4 rocky inner planetsSee if the planets are at the right distances from the Sun

Mer

cury

Venus

Earth

Mar

s

0.4 AU 0.7 AU 1.0 AU 1.5 AU

It is assumed that the Gas Giants had already formed beforethis stage. We will examine the formation of those planetslater in this lecture


Orbital eccentricity and inclination

Eccentricity

A planet’s orbital eccentricity, e is ameasure of how elliptical its orbit is

e = 0 means the orbit is circular

0 e 1 for an elliptical orbit

Inclination

The orbital inclination, i is the anglebetween the plane of the orbit of theplanet and the ecliptic — which isthe plane containing Earth’s orbitalpath

Sun

Planet

a

ae

e=0.0 e=0.5 e=0.8

i


Simulating terrestrial planet formationN-body simulation

In the lecture I showed a movie of an N-Bodysimulation created by David O’Brien. Thatsimulation can be viewed online here:http://www.psrd.hawaii.edu/WebImg/OBrien_movie_cjs_simulation.gif

Legend

Jupiter Embryos/Planets Planetesimals

Within 10’s ofmillions ofyears, severalplanets form

Stable orbits

Terrestrialplanet region0.5 – 2 AU


http://www.psrd.hawaii.edu/WebImg/OBrien_movie_cjs_simulation.gif

Impacts were violent on the early Earth

Image courtesy of Don Davis

Soon after its formation, thesurface of the Earth wouldhave been molten

The Hadean period

Heat from radioactivematerials and impacts

Impacts much more violent,and more common, duringthe period of planetformation

Many more impactorsHigh eccentricity ofplanetesimals meanthigher velocities


Building the planets

Terrestrialplanet formationis relatively wellunderstood

Why are theterrestrialplanets so

different fromthe gas giants?



Part 2:Building the outer planets



Jupiter

Image courtesy of NASA/JPL/Space Science Institute

11 � larger than Earth,318 � heavier

Much lower density (4 timeslower than Earth)

Possibly contains a rockycore around the size of Earth

Atmosphere of hydrogen andhelium

Impacts into the atmosphereof Jupiter are visible asbrown spots


Jupiter

Impact of cometShoemaker-Levy 9 in 1994

Image courtesy of NASA/Hubble Space Telescope CometTeam

11 � larger than Earth,318 � heavier

Much lower density (4 timeslower than Earth)

Possibly contains a rockycore around the size of Earth

Atmosphere of hydrogen andhelium

Impacts into the atmosphereof Jupiter are visible asbrown spots


Saturn

Image courtesy of NASA and The Hubble Heritage Team(STScI/AURA); R.G. French (Wellesley College), J.

Cuzzi (NASA/Ames), L. Dones (SwRI), and J. Lissauer(NASA/Ames)

9.5 � larger than Earth;95 � heavier

Very low density (half thatof Jupiter)

Large atmosphere ofhydrogen and helium

Probably has a small rockycore, similar to Jupiter

Like all the gas giants,Saturn has rings of dust andice


Uranus

Image courtesy of NASA/JPL/Voyager mission

Smaller than Jupiter andSaturn;4 � larger radius than Earth

Similar density to Jupiter

Along with Neptune, knownas an ice giant

Contains more ices in theatmosphere than Jupiterand Saturn

Rotates on its side: probablythe result of a large impactevent


Neptune


Similar size, density andcomposition to Uranus

Contains ices in itsatmosphere, like Uranus


Planets comparison


Small, rocky

Size range 0.4 – 1.0 � Earth

Thin (or no) atmosphere

Density range 4 – 5.5 g/cm3


Large, gaseous

Size range 3.9 – 11.2 � Earth

Very large atmosphere

Density range 0.7 – 1.6 g/cm3

Why are the gas giants so different from the inner planets?

How did they form?


Planets comparison


Building the gas giants: Core accretion theory

Core of the planet growsrapidly via two-bodycollisions

Same process as forplanetesimals and theinner planets

Gas attracted into envelopesurrounding core

Core continues to grow asmore planetesimals collidewith it

38

formation lead to planetary properties (such as the ec-centricity, and the mass of Mars compared to the otherterrestrial planets) that di!er somewhat from those ob-served in the Solar System. Thus, although there is gen-eral confidence that the basic physics of terrestrial planetformation is understood, it is clear that current modelsdo not include all of the ingredients needed to accuratelymatch Solar System constraints (Raymond et al., 2009).

C. Gas giant formation

Two theoretical models vie to explain the formationof gas giant planets. The core accretion model (Boden-heimer & Pollack, 1986; Mizuno, 1980), developed in itsmost refined form by Pollack et al. (1996), postulates thatthe envelopes of gas giants are accreted subsequent to theformation of a large core, which is itself assembled in amanner analogous to terrestrial planet formation. Coreaccretion is the dominant theory for massive planet for-mation. The gravitational instability model, on the otherhand, is based on the idea that a massive protoplane-tary disk might collapse directly to form massive planets(Cameron, 1978; Kuiper, 1951). Boss (1997) is the mostrecent advocate of this idea, which has come under re-newed theoretical scrutiny with the discovery of manyextrasolar planets with masses much larger than that ofJupiter.

In this Section, we review the physics of these theoriesin turn. We also discuss the observational constraints onthe di!erent theories, which include inferences as to thecore masses of the gas giants in the Solar System, the hostmetallicity / planet frequency correlation for extrasolarplanetary systems, and — indirectly — comparison ofthe theoretically derived time scales with observationsof protoplanetary disk lifetimes. This is a critical issue,since gas giants must form prior to the dispersal of the gasdisk. Any successful model of massive planet formationmust grow such bodies within at most 5-10 Myr (Haisch,Lada & Lada, 2001).

1. Core accretion model

The main stages in the formation of a gas giant viacore accretion are illustrated in Figure 26. A core ofrock and / or ice forms via the same mechanisms that wehave previously outlined for terrestrial planet formation.Initially, there is either no atmosphere at all (becausethe potential is too shallow to hold on to a bound at-mosphere), or any gas is dynamically insignificant. How-ever, as the core grows, eventually it becomes massiveenough to hold on to a significant envelope. At first,the envelope is able to maintain hydrostatic equilibrium.The core continues to grow via accretion of planetesi-mals, and the gravitational potential energy liberated asthese planetesimals rain down on the core provides themain source of luminosity. This growth continues until

incr

easi

ng ti

me,

pla

net m

ass

planetesimals

gas

FIG. 26 Illustration of the main stages of the core accretionmodel for giant planet formation.

the core reaches a critical mass. Once the critical massis reached, the envelope can no longer be maintained inhydrostatic equilibrium. The envelope contracts on itsown Kelvin-Helmholtz time scale, and a phase of rapidgas accretion occurs. This process continues until (a) theplanet becomes massive enough to open up a gap in theprotoplanetary disk, thereby slowing down the rate of gassupply, or (b) the gas disk itself is dispersed.

The novel aspect of the core accretion model is theexistence of a critical core mass. Mizuno (1980) used nu-merical models to demonstrate the existence of a maxi-mum core mass, and showed that it depends only weaklyon the local properties of the gas within the protoplane-tary disk. A clear exposition of this type of calculationis given in, for example, Papaloizou & Terquem (1999).Here, following Stevenson (1982), we show that a toymodel in which energy transport is due solely to radia-tive di!usion displays the key property of a critical coremass.

Consider a core of mass Mcore and radius Rcore, sur-rounded by a gaseous envelope of mass Menv. The totalmass of the planet,

Mt = Mcore + Menv. (191)

The envelope extends from Rcore to some outer radius

Image courtesy of P. Armitage


How did the cores form so quickly?

The core must grow rapidly while gas is still present in the diskHow did they form more quickly than the inner planets?Further from the Sun:

More mass in an annular diskColder temperatures: ices condense

All that extra mass means more collisions and therefore fastergrowth

Snowline


Building the gas giants: Core accretion theory

Core forms quickly

Takes � 0.5 Myr

Hydrostatic growth

Slow accretion of gasfrom diskFeeding zone growswith planetTakes � 7 Myr

Runaway growth

When the mass of thegas becomes largerthan the mass of thecore, runaway growthstarts

0 1 2 3 4 5 6 7 8Time [Myr]

0

20

40

60

80

100

Mas

s[M

C]

Core

Planet

Coreformation Runaway

growth

Adapted from Rice & Armitage 2003


Part 3:Recent advances



Some open questions

1 Uranus and Neptune could notform in the current locations fastenough

Gas not dense enough

2 The giant planets all have e, i ¡ 0

The gas should have dampenedthe orbits down to circular andcoplanar

3 There is not enough mass in theKuiper belt to form Pluto and theother dwarf planets

Mass of Kuiper belt is around themass of the Moon

Images courtesy of NASA/JPL


Some open questions







Sun

Planet

a

ae

e=0.0 e=0.5 e=0.8

i


Some open questions









The answer may lie in orbital resonances

Remember from last weekthe role that orbitalresonance can have on theasteroids

In the Nice model, the sameapplies for the Giant Planets


















































The Nice ModelTop-down view of the Solar System

Image courtesy of Wikimedia Commons

In the lecture I showed a movie of theNice Model. A version of that moviecan be viewed on YouTube, here:http://www.youtube.com/watch?v=6LzQfR-T5_A

Big changes happen whenJuptier and Saturn reachan orbital resonance

Uranus and Neptunemove outwards (andswitch position!)

The Kuiper Beltstarted with highermass that wasscattered by theplanets’ changingorbits

The resonance movesthe planets intohigher e, i orbits


http://www.youtube.com/watch?v=6LzQfR-T5_A

Orbital evolution in the Nice Model

Image adapted from Gomes et al. (2005), Nature


Nice Model outcomes

In summary, the Nice modeloffers one possible mechanismto match some observations ofthe Solar System:

Uranus and Neptune formedcloser to the Sun than theyare nowThis overcomes the problemthat at their currentlocations, there would nothave been enough material toform planets of that sizeAllows them to reach theirpresent large sizes and thenmigrate outwards

Image adapted from Gomes et al. (2005) Nature


Nice Model outcomes


The resonance of Jupiter andSaturn throws all the giantplanets into higher e, i orbitsExplains why they are notcloser to circular orbits



Nice Model outcomes


The dwarf planets formed inthe Kuiper Belt when therewere still many planetesimalspresentMost of that mass was lost inthe resonance event,explaining the smaller masstoday



What does this mean for the rest of the Solar System?

Mare Imbrium


Lots of objects from the Kuiperbelt thrown inwards towards theinner Solar System

Evidence for this on the Moon!

e.g. The Lunar basins such asMare Imbrium

We’ll cover this in the lectureabout the Moon


Mars is too small!

One more open question: Why is Mars is so small?

Greater orbital distance; should be larger than the Earth

However, it is actually only one tenth of the mass of EarthWhy?



Mars stopped growing early

timescale of Mars can be compared with those of planetesimals, whoseformation history is known from measurement of meteorites. Theparent bodies of iron meteorites were formed at the same time or soonafter CAIs10. Chondrites, which formed .3 Myr after CAIs25, containpristine materials such as CAIs and presolar grains that indicate they

accreted from pristine nebular dust, rather than from collisional debris.Detailed thermochronology also shows that chondrites’ parent bodieswere probably around 100 km in size21. An important finding of thiswork, therefore, is that large planetesimals (for example, 10–100-kmbodies) were still being formed in the protoplanetary disk during Mars’accretion.

The accretion of planetary embryos from planetesimals began with aperiod of runaway growth, in which large bodies accreted at a muchhigher rate than smaller ones. When the larger bodies (that is, embryos)became massive enough to dynamically stir the planetesimals aroundthem, their growth was slowed down by reduced gravitational focusingfactors, a process known as ‘oligarchic growth’. The embryos eventuallycleared their orbits of smaller objects, limiting their mass to Moon- toMars-sized objects. Computer simulations predict that the growth ofembryos at 1.5 astronomical units (AU) from the Sun should occur on amaximal timescale of a few million years (note that this timescale issensitive to the assumed initial surface density)17,18,26. The growth ofterrestrial planets such as Earth, on the other hand, is thought to haveproceeded by collisions between these embryos on an expected time-scale of several tens of millions of years, a process known as ‘chaoticgrowth’1,5. This timescale on Earth is best constrained by the age of theMoon, which is thought to have formed by a collision that occurred50–150 Myr after CAIs3 between a Mars-size planetary embryo andthe proto-Earth2. It is difficult to reconcile Mars’ accretion timescalebased on the e182W data (t 5 1:8z0:9

{1:0 Myr) with an Earth-like modeof accretion. Our findings, however, are entirely consistent withthe timescale predicted by models of oligarchic growth (Fig. 2). Wetherefore conclude that Mars is most likely to be an embryo thatescaped collisions and merging with other bodies, thus explainingits small mass compared to Earth and Venus.

Rapid accretion of Mars also has important implications for itsmagmatic history. The gravitational energy deposited on the growingembryo from planetesimals during oligarchic growth would not havebeen sufficient to cause large-scale melting27, except for a mode ofrunaway core formation triggered by impacts with .3,500-kmembryos28. In this scheme, the temperature rise would not have beensufficient to induce silicate melting, and molten metal would havesegregated through a solid matrix. The Martian mantle (based onthe composition of SNC meteorites) contains large 182W and 142Ndisotopic heterogeneities that arise from magmatic fractionation ofHf/W and Sm/Nd while 182Hf and 146Sm (half-life 103 Myr) were stillpresent7,8,29,30. The event that caused this chemical fractionation isdated at ,20–60 Myr after CAIs, and may correspond to crystalliza-tion of a magma ocean on Mars. A similar age range is obtained from

0.05

0.10

0.15

0.20

0.25

0.30

0.35

0.40

0.2815 0.2825176HF/177HF

176 H

F/17

7 HF

0.2835 0.2845

Th/H

f

C-chondrites

E-chondrites

O-chondrites

(Th/Hf)CHUR= 0.2144 ± 0.0075

0.28

2785

Typical error bars (95% con!dence)

a

0.2815

0.2820

0.2825

0.2830

0.2835

0.2840

0.2845

1 2 3 4 5 6 7

Petrologic type

Metamorphism Alteration

b

Figure 1 | Determination of the Th/Hf ratio of CHUR. a, Correlationbetween Th/Hf and 176Hf/177Hf (a proxy for Lu/Hf) ratios in chondrites(Table 1; C, carbonaceous; E, enstatite; O, ordinary; meteorite Happy Canyonnot included). The 176Hf/177Hf ratio of CHUR is estimated to be0.282785 6 0.000011 (ref. 14), allowing us to estimate (Th/Hf)CHUR 50.2144 6 0.0075 (the uncertainty is the 95% confidence interval based onregression of the data). At a Th/Hf ratio of 0, we calculate a 176Hf/177Hf ratio of,0.280, which is close to the Solar System initial ratio of 0.27978 (ref. 14). Thereis no correlation between the Lu/Th and Lu/Hf ratios (Table 1). These twoobservations suggest that Th behaves very similarly to Lu but that bothelements can be decoupled from Hf in meteorites. If chondrite parent bodieshad begun with different Th/Hf ratios (for example, owing to evaporation/condensation processes in the nebula), they would not define a singlecorrelation line. b, 176Hf/177Hf ratios of chondrites as a function of petrologictypes (data from Table 1, refs 12, 13 and references therein). The data pointshave been moved horizontally by random values to decrease overlap andimprove readability. The degree of aqueous alteration increases from type 3 to1, while the degree of thermal metamorphism increases from type 3 to 7 (that is,the most pristine samples are of type 3). The dispersion in 176Hf/177Hf ratios ofmetamorphosed chondrites (types 4–7) is much larger compared with othersamples, indicating that redistribution during parent-body metamorphism isthe most likely explanation for the dispersion in Lu/Hf (and Th/Hf) ratiosmeasured in bulk chondrite specimens. The average Th/Hf ratio ofunmetamorphosed chondrites (types 1–3) is 0.2217 6 0.0135, identical withinuncertainties with the value derived from the regression in a.

0

0.2

0.4

0.6

0.8

1

0 2 4 6 8 10

M(t)

/M!n

al

Time after Solar System birth (Myr)

1

1 AU

3.2 AU

Figure 2 | Accretion timescale of Mars inferred from 182Hf–182Wsystematics. The red curve is calculated from M(t)/Mfinal 5 tanh3(t/t) witht 5 1:8z0:9

{1:0 Myr (see text for details). The yellow band corresponds to the 95%confidence interval on the accretion curve of Mars obtained by propagating alluncertainties on model parameters using a Monte Carlo simulation. The twoblack curves are model simulations of embryo growth at 1 and 3.2 AU (ref. 26)for comparison.

LETTER RESEARCH

2 6 M A Y 2 0 1 1 | V O L 4 7 3 | N A T U R E | 4 9 1

Macmillan Publishers Limited. All rights reserved©2011

Image courtesy of Dauphas and Pourmand(2011) Nature

Recent cosmochemistry studies byresearchers at UChicago

Measured the isotopic compositions ofMartian meteorities

Able to estimate the time that Marsstopped growing

Mars must have stopped growingafter only 2–4 million yearsWell before Earth stopped growing(10’s of millions of years)i.e. Mars was a stranded planetaryembryo


Mars and the Grand Tack


So, we think that Mars wasan embryo that just stoppedgrowing

What could cause it to stop?

Several recent modelsprovide possible reasons

We will look at one of themnow, called the Grand Tackmodel


The Grand Tack Scenario

Jupiter migratesinwards due tocurrents in the gas

Moves from 3.5 AUto 1.5 AU

In the lecture, I showed some movies of theGrand Tack model. You can view thosemovies, and see some more detailed discussionof the model, on Sean Raymond’s website,here:http://www.obs.u-bordeaux1.fr/

e3arths/raymond/movies_grandtack.html


http://www.obs.u-bordeaux1.fr/e3arths/raymond/movies_grandtack.html



Saturn also movesinwards

3:2 resonance withJupiter means bothplanets moveoutwards

Planetesimals(asteroids) arescattered and mixedin the process

In the lecture, I showed some movies of theGrand Tack model. You can view thosemovies, and see some more detailed discussionof the model, on Sean Raymond’s website,here:http://www.obs.u-bordeaux1.fr/

e3arths/raymond/movies_grandtack.html





Asteroid belt ends upas a mix of somebodies that formedoutside Jupiter andsome that formedinside

Fits with predictionfrom last week aboutthe E, S and C classasteroids

Planetesimal diskends at around 1 AU— which helps toexplain the small sizeof Mars

Image courtesy of K. Walsh


Other scenarios could work instead


One example is calledplanetesimal-drivenmigration

Mars pushed outwards bygravitational interations withlots of planetesimals

Moves outwards until itleaves the region mostpopulated by planetesimals

Once there, growth is halted


Summary


Inner planets grow bypairwise collisions

Gas giants grow quickly dueto extra mass in outer diskfrom condensed ices

Cores quickly grow, andaccumulate large gaseousatmospheres

Migration of gas giants canlead to severe consequencesin the inner Solar System

Mars is smallMixed asteroid belt

Next week: Collisions


Thank you

Questions?


Documents

Building the Planets - UChicago GeoScigeosci.uchicago.edu/~tdavison/comptonlectures/Slides...Image courtesy of NASA/Hubble Space Telescope Comet Team 11 larger than Earth, 318 heavier