Giant Impacts During Planet Formation:

1
Giant Impacts During Planet Formation

• Parallel Tree Code Simulations Using Smooth
  Particle Hydrodynamics

2
Why Giant Impacts?

• Evidence
• 98 degree obliquity of Uranus
• Random element of planetary rotation
• Moon, Mercury, Pluto-Charon
• Core Accretion vs. Gas Instability Model
• Metal abundance vs. planet occurrence rate
• Mass function of planets

3
Impact Effects

• Shortening Formation Timescales
• Observations Median disk lifetime 3 Myr, Jupiter
  0-3 MEarth Core (Saturns 9-22 MEarth)
• Theory Jupiter needs more than 10 Myr to form
  with 3 MEarth Core (and Saturn takes 10 Myr to
  form with 17.5 MEarth core)
• Effects of migration

4
Shortening Formation Timescales How?

• Cut off heating from planetesimal accretion
• Jupiter Start with large core - rapid growth -
  impact removes core mass directly or through
  mixing
• Saturn Impact adds mass - grow core more
  quickly, remove later planetesimal heating (9.8
  - 3.3 Myrs)

5
Uranus and Neptune

• Similar core masses accounted for by core
  accretion model
• Why similar envelope masses?
• Why thermal emission differences?

6
Goals

• Core Mass/Formation Time
• Uranus and Neptune Envelope Mass, Thermal
  Emission
• Satellite Formation

7
Method

• SPH with parallel tree algorithm, self-gravity,
  Runge-Kutta-Fehlberg integrator of 2nd order
• 262,144 particles
• Radial cutoff at 1010 cm
• Tillotson, Saumon-Chabrier, ideal gas EOS
• Continued for 25,000-35,000 seconds (time to
  impact 7-12 free-fall times)

8
Method Comparisons

• Korycansky 1990
• 1-D grid code
• impact onto Uranus
• produces obliquity envelope dispersal
• Slattery 1992
• SPH, 8,000 particles (1 small particle mass of
  entire Uranian satellite system)
• Uranus impact - disks up to 0.3 MEarth
• no gaseous envelope
• Radiative Transfer
• no reasonable algorithm available when
  simulations were run
• simulation time

9
Numerical Tests of Code

• Mass conservation good (
• Angular momentum conservation within 1043 gm
  cm2/sec (
• Energy not conserved well (viscosity effects)

10
Numerical Tests of Code

• Particle Number
• Earth vs. Mars pressure/density wave propagates
  at sound speed

11
Results

• Simulations
• 6 MEarth impactor, 26 km/sec, 5 silicate ice,
  600 particles
• Impact angles 0/0, 90/15, inf/24 degrees w/r/t
  core/envelope
• Jup Model A 30.6 MEarth core 6.24 (5) MEarth
  envelope, dust opacity standard model, 3 (of
  3.1) Myr evolution time
• Jup Model B 14.6 MEarth core 5.0 (3) MEarth
  envelope, dust opacity 0.02 standard model,
  1.75 (of 2.2) Myr

12
Simulation A1

• Head-on collision with massive proto-Jupiter
  (Model A)
• 8 of envelope mass escapes
• Core and impactor remain

13
(No Transcript)
14
(No Transcript)
15
(No Transcript)
16
(No Transcript)
17
Simulation A1half

• Same as A1 but half the particles
• 8 of envelope mass escapes (same as A1)
• Core and impactor remain
• Numerical resolution is sufficient

18
Simulation A2

• Glancing (90 degree impact angle w/ core)
  collision with massive proto-Jupiter (Model A)
• 4 of envelope mass escapes
• Majority of impactor 0.14 MEarth of core
  escapes
• Additional 0.02 MEarth from core 0.06 from
  impactor likely to mix into envelope

19
(No Transcript)
20
Simulation A3

• Glancing (24 degree impact angle w/ envelope)
  collision with massive proto-Jupiter (Model A)
• 2 of envelope mass escapes
• Impactor escapes, core remains

21
Simulation B1

• Head-on collision with less massive proto-Jupiter
  (Model B)
• 32 of envelope mass escapes to orbit or to
  nebula
• 0.2 MEarth of core escapes, impactor remains
• Additional 0.16 MEarth from core likely to mix
  into envelope

22
(No Transcript)
23
Simulation B2

• Glancing collision (90 degree impact angle w/
  core) with less massive proto-Jupiter (Model B)
• 9 of envelope mass escapes to orbit or to nebula
• Impactor leaves behind only 0.01 MEarth

24
(No Transcript)
25
Simulation B3

• Glancing collision (24 degree impact angle w/
  envelope) with less massive proto-Jupiter (Model
  B)
• 2 of envelope mass escapes
• Core remains, impactor departs intact

26
Results Summary
27
Results Summary
28
Results Timescale/Core Mass

• Case A1 Shorten formation time by increasing
  core mass
• Case A2 Shorten formation time by increasing
  core mass some mass liberated from core, but
  more is added to core from impactor
• Case A3 Little effect on core mass
• Case B1 Shorten formation time by increasing
  core mass some mass liberated from core, but
  more is added to core from impactor
• Case B2 Little effect on core mass
• Case B3 Little effect on core mass
• A case with intermediate impact angle between B1
  and B2 is likely to result in more core mass
  being lost
• Upper bound of 0.3 MEarth on core mass mixed with
  deposited energy through convection

29
Results Uranus/Neptune Similarities

• Case B1 produces envelope mass similar to
  Uranus/Neptune
• Case B2 produces core mass similar to U/N and
  envelope mass nearly as small
• Angular momentum values in reasonable agreement
  between Case B2 and Uranus today

30
Results Uranus/Neptune Thermal Emission
Differences

• Case B1 - compositional gradient at 89 of
  radius (rhoo 2.7 vs. 0.95 g/cm3)
• Compositional gradients at 50-60 radius require
  impact at much earlier times

31
Results Satellite Formation, Envelope Enrichment

• Solid satellite disks not formed in any of the 6
  cases
• Case B1 results in an upper bound of 0.37 MEarth
  of material mixed into envelope ( up to 0.29
  mixed due to convection)

32
Summary

• Mass and angular momentum well-conserved, energy
  subject to artificial viscosity errors
• Simulations well-resolved numerically
• Simulated giant impacts at 2 evolutionary stages
  with 3 impact angles

33
Summary contd

• Giant impacts disrupt planetary cores, decreasing
  their mass by as much as 0.66 MEarth (but there
  is a net increase in core mass in all cases due
  to added mass of impactor)
• Core mass loss decreases as a function of initial
  proto-planetary mass and increases with impact
  angle, up to angle unclear)
• No case resulted in a net loss of core mass, but
  an impact intermediate between B1 and B2 may do
  so
• Mass added to core by impactor increases as a
  function of initial proto-planet mass and
  decreases as a function of impact angle
• Up to 0.16 MEarth of core is accelerated to half
  the escape speed, but has insufficient angular
  momentum to create a disk. Impact angle between
  B1 and B2 will provide more non-radial velocity.

34
Summary contd

• Giant impacts can leave planet with only 2
  MEarth of envelope, similar to Uranus and Neptune
• Envelope mass removed goes up with decreasing
  impact angle and decreasing core mass
• Impacts in our models created compositional
  differences down to 89 of planetary radius
• No significant disks were generated

35
Future Directions

• Do simulation similar to Case B2 with slightly
  smaller impact angle
• Likely to result in greater mass loss
• Improve formulation of artificial viscosity
• Suite of simulations to determine likelihood of
  envelope mass decrease to 2 MEarth, and to test
  Canup et al 2001 satellite formulae
• Impact simulations at earlier times to determine
  likelihood of compositional gradient at 50-60
  radius
• Use more recent (Hubickyj et al 2005) planet
  models, add migration effects, make choice of
  impactor characteristics more sophisticated, use
  impactors with rock/ice layers

36
Conclusion

• Giant impacts can affect planetary evolution
• Thanks to Peter (!), Doug and Steve for their
  support!