Title: Moving beyond the Earth: What use is mineral physics to planetary scientists
1Moving beyond the EarthWhat use is mineral
physics to planetary scientists?
Francis Nimmo (U. C. Santa Cruz)
2Talk Outline
- What do we care about?
- What do we know?
- Earth, solar system, extra-solar planets
- What would we like to know (and why)?
- Static properties
- EOS
- Melt behaviour
- Dynamic properties
- Rheology
- Dissipation
- Conductivity
- Partitioning
3What do planetary scientists care about?
- Present-day interior structure
- Formation
- Evolution
What do mineral physicists care about?
- Measuring things (preferably under extreme
circumstances) - Here are some justifications for doing so . . .
4Solar nebula and planets . . .
- Nebular material can be divided into gas
(mainly H/He), ice (CH4,H2O,NH3 etc.) and
rock (including metals) - In our solar system, the proportions of
gas/ice/rock are roughly 100/1/0.1 - Planets will contain variable mixtures of these
- The compounds which actually condense will
depend on the local nebular conditions
(temperature) - E.g. volatile species will only be stable
beyond a snow line (distance depends on stellar
luminosity) - But planets can (and do) migrate subsequent to
their formation! (e.g. hot Jupiters)
5Classes of planetary bodies
Ice H,He 15 Me 800 GPa 8000 K
Mainly H,He 300 Me 7000 GPa 20,000 K
Rockice 0.1 Me 10 GPa 1500 K
6What do we know?
7What are we going to know?
- Jupiter/Saturn internal structure (JUNO,Cassini)
- Extra-solar planet atmospheric compositions
- Extra-solar planet flattening?! (MoI)
- Earth-like planets mass/radii (COROT, Kepler)
- Mars seismology (dont hold your breath)
81. Static properties
- Equations of state
- Hydrogen Helium
- Everything else
- (Silicate) Melting
9Hydrogen EOS
- Why do we care?
- Fundamental to deducing structure of gas giants
- A 5 error in the EOS for hydrogen translates
into a factor of six uncertainty in the abundance
of ices - Different EOS lead to different conclusions!
Laser (high compression)
Pulse-shock (low comp.)
Guillot, Ann. Rev. 2005
Podalak and Hubbard 1998
10Hydrogen - Experiments
DAC
Hubbard et al. Ann. Rev. 2002
11He EOS
- He makes up 20 by mass of giant planets
- He EOS only measured to 50 GPa (less than 5 of
depth within Jupiter) - Extent to which He and H are miscible is
important (energy balance) - Ne only 0.1 x solar in Jupiter envelope is this
because it dissolves in He?
12H/He - Summary
- H EOS/compressibility
- Size of Jupiters core, envelope composition
- H molecular/metallic transition
- Convective barrier, chemistry, temperature
- H/He miscibility
- Internal structure, energy budget
- He EOS and noble gas solubility
- Experiments only up to 50 GPa
13H/He -A Caution!
HD149026b (1.25 g/cc)
1g/cc
1g/cc
20 g/cc
7 g/cc
GJ436b (2.02 g/cc)
3 g/cc
10 g/cc
Mixing ratios can be more important than EOS
accuracy
Gillon et al. 2007
14EOS Everything else
- Super-Earths e.g. GJ876d (7.5 Me), Gl581c (5
Me), OGLE-2005-BLG-390Lb (5.5 Me), more to come! - Need for EOS data up to several TPa (Valencia et
al. 2007) - Incompressible oxides e.g. Gd3Ga5O12 (Mashimo et
al. 2006) - Carbon-rich planets (?) (Kuchner and Seager,
submitted)
Super Earths (P few TPa)
Fortney et al. 2005 parameterization
15(Silicate) melt behaviour
- Why do we care?
- Mass transfer (chemistry, differentiation)
- Heat transfer (e.g. Io)
- Rheology
- Many other reasons!
- What things to measure?
- Liquidus
- Density
16Liquidus/Density
- Deep mantle liquidus controls whether magma ocean
solidifies from top or bottom important! - Melt-solid density contrast controls whether
magma can move upwards or not affects e.g. CMB
heat flux
Mosenfelder et al. JGR in press
17Summary Static properties
- Equations of state
- Hydrogen metallic transition He miscibility
- Helium high pressure EOS, noble gas solubility
- Super-Earths imply pressures up to few TPa
- (Silicate) Melting
- Solidification from top or bottom?
- Density compared with solid
182. Dynamic Properties
- Rheology
- Dissipation
- Conductivity
- Partitioning
19Rheology (viscosity)
- Why does it matter?
- Heat transfer
- Mixing/stirring rates (chemistry)
- Dissipation (see later)
- What would we like to know?
- Deep earth
- Ices
Convection inside Enceladus (image courtesy James
Roberts)
20What would we like to know?
- Deep Earth
- Perovskite
- Post-perovskite . . .
- Influence of water . . .
- Ices (outer solar system sats.)
- Ice I diffusion creep!
- Higher pressure ice rheology not well known
Ice II
Kubo et al. Science 2006
Ice I
5mm
Forte and Mitrovica Nature 2001
21Dissipation
- Deforming real materials results in dissipation
- Tidal dissipation very important to planets
- How do we define dissipation?
e
Gribb and Cooper 1998
Apples vs. oranges?
22Dissipation measurements
0 -1 -2 -3 -4 -5 -6
-7 -8 -9
Increasing dissipation
Maxwell model
Andrade model (a0.3)
Andrade model (a0.2)
Apples vs. oranges?
23Conductivity
- High pressure ice conductivities important for
Neptune, Uranus mag.fields (Cavazzoni et al.
1999, Lee et al. 2006) - Fe conductivity uncertain by a factor of 2
(Matassov 1977, Bi et al. 2002) - Affects strength of magnetic field
- Affects age of inner core (Nimmo et al. 2004)
24Partitioning
- Vital for using geochemical observations to
constrain physical processes. Examples - Re/Os/Pt and age of inner core (?) (Brandon et
al.) - He/U/Th and mantle layering (Parman)
- Siderophile elements and core formation (various)
- Experimentally challenging e.g. high temperature
gradients can drive diffusion - Affected by many factors e.g. oxygen fugacity,
silicate polymerization
25Summary
- Available observational constraints much poorer
than for Earth, but . . . - Parameter space much wider!
- Higher P,T
- Different and exotic compositions (hydrogen,
noble gases, carbides etc.) - N gtgt 1
- Major growth areas (e.g. extra-solar planets)
- Funding possibilities? (e.g. NASA PIDDP)
26Conclusion a shopping list
- H molecular-metallic transition and He
miscibility - He EOS (gt 50 GPa)
- High P silicate melting
- Q (at 1 hr periods better theoretical
understanding) - High P silicate/ice rheology
- Fe/high P ice conductivity
- High P partition behaviour
27Questions?
28Dihedral angle
- Controls melt separation and movement
- Important for core formation, magma transport
- Results depend on O content of liquid Fe (P,T
dependent) - Inefficient Fe separation in lower mantle?
- Hard experiments very large T gradients
Terasaki et al. 2005, 5-20 GPa
29Extra-solar planets
- Hot Jupiters have more heating (radiative,
tidal) - Larger core masses? (close-in means less easy to
scatter planetesimals)
30How do we calculate Q?
- For solid bodies, we assume a viscoelastic
rheology - Such a body has a rigidity m, a viscosity h and a
characteristic relaxation (Maxwell) timescale
tmh/m - The body behaves elastically at timescales ltlttm
and in a viscous fashion at timescales gtgt tm
- Dissipation is maximized when timescale tm
Tobie et al. JGR 2003
31 Interior Structure of GJ 876d
20,000
7.5 ME
DENSITY (kg/m3)
12,000
Valencia, Sasselov, OConnell (2006)
4,000
2,000
6,000
10,000
RADIUS (km)
32Partitioning
Can siderophile element abundances be explained
by high P,T partition coefficients?
Walter et al. 2000
Kegler et al. 2005
33Compressibility Density
- As mass increases, radius also increases
- But beyond a certain mass, radius decreases as
mass increases. - This is because the increasing pressure
compresses the deeper material enough that the
overall density increases faster than the mass - The observed masses and radii are consistent with
a mixture of mainly HHe (J,S) or H/Heice (U,N)
radius
Constant density
mass
34Basic Parameters
- Data from Lodders and Fegley 1998. Surface
temperature Ts and radius R are measured at 1 bar
level. Magnetic moment is given in 10-4 Tesla x
R3.
35Compositions (1)
- Well discuss in more detail later, but briefly
- (Surface) compositions based mainly on
spectroscopy - Interior composition relies on a combination of
models and inferences of density structure from
observations - We expect the basic starting materials to be
similar to the composition of the original solar
nebula - Surface atmospheres dominated by H2 or He
(Lodders and Fegley 1998)
36Interior Structures again
- Same approach as for Galilean satellites
- Potential V at a distance r for axisymmetric body
is given by
- So the coefficients J2, J4 etc. can be determined
from spacecraft observations - We can relate J2,J4 . . . to the internal
structure of the planet
37Interior Structure (contd)
C
R
- What we would really like is C/MR2
- If we assume that the planet has no strength
(hydrostatic), we can use theory to infer C from
J2 directly - For some of the Galilean satellites (which ones?)
the hydrostatic assumption may not be OK
A
- Is the hydrostatic assumption likely to be OK for
the giant planets? - J4,J6 . . . give us additional information about
the distribution of mass within the interior
38Results
- Densities are low enough that bulk of planets
must be ices or compressed gases, not silicates
or iron (see later slide) - Values of C/MR2 are significantly smaller than
values for a uniform sphere (0.4) and the
terrestrial planets - So the giant planets must have most of their mass
concentrated towards their centres (is this
reasonable?)
39Pressure
- Hydrostatic approximation
- Mass-density relation
- These two can be combined (how?) to get the
pressure at the centre of a uniform body Pc
- Jupiter Pc7 Mbar, Saturn Pc1.3 Mbar, U/N Pc0.9
Mbar - This expression is only approximate (why?)
(estimated true central pressures are 70 Mbar, 42
Mbar, 7 Mbar) - But it gives us a good idea of the orders of
magnitude involved
40Temperature (1)
- If parcel of gas moves up/down fast enough that
it doesnt exchange energy with surroundings, it
is adiabatic - In this case, the energy required to cause
expansion comes from cooling (and possible
release of latent heat) and vice versa - For an ideal, adiabatic gas we have two key
relationships
Always true
Adiabatic only
Here P is pressure, r is density, R is gas
constant (8.3 J mol-1 K-1), T is temperature, m
is the mass of one mole of the gas, g is a
constant (ratio of specific heats, 3/2)
- We can also define the specific heat capacity of
the gas at constant pressure Cp - Combining this equation with the hydrostatic
assumption, we get
41Temperature (2)
- At 1 bar level on Jupiter, T165 K, g23 ms-2,
Cp3R, m0.002kg (H2), so dT/dz 1.4 K/km
(adiabatic) - We can use the expressions on the previous page
to derive how e.g. the adiabatic temperature
varies with pressure
(Here T0,P0 are reference temp. and pressure, and
c is constant defined on previous slide)
This is an example of adiabatic temperature and
density profiles for the upper portion of
Jupiter, using the same values as above, keeping
g constant and assuming g1.5 Note that density
increases more rapidly than temperature why?
Slope determined by g
42Heavy elements
Guillot 2005
- He subsolar sedimentation?
- Ne depleted dissolves in He?
- Others supersolar delivery by cold bodies
(comets)?
43He miscibility
Hubbard et al. 2002
44Nebular Composition
- Based on solar photosphere and chondrite
compositions, we can come up with a best-guess at
the nebular composition (here relative to 106 Si
atoms)
- Blue are volatile, red are refractory
- Most important refractory elements are Mg, Si,
Fe, S (in the ratio 110.90.45)
Data from Lodders and Fegley, Planetary
Scientists Companion, CUP, 1998 This is for all
elements with relative abundances gt 105 atoms.
45Temperature and Condensation
Nebular conditions can be used to predict what
components of the solar nebula will be present as
gases or solids
Mid-plane
Photosphere
Earth
Saturn
Condensation behaviour of most abundant elements
of solar nebula e.g. C is stable as CO above
1000K, CH4 above 60K, and then condenses to
CH4.6H2O. From Lissauer and DePater, Planetary
Sciences
Temperature profiles in a young (T Tauri) stellar
nebula, DAlessio et al., A.J. 1998
46Where is everything?
Note logarithmic scales!
Ma
V
E
J
S
U
KB
Me
N
P
1 AU is the mean Sun-Earth distance 150 million
km Nearest star (Proxima Centauri) is 4.2
LY265,000 AU
47Basic data
See e.g. Lodders and Fegley, Planetary
Scientists Companion
48Sequence of events
- 1. Nebular disk formation
- 2. Initial coagulation (10km, 104 yrs)
- 3. Orderly growth (to Moon size, 105 yrs)
- 4. Runaway growth (to Mars size, 106 yrs), gas
loss (?) - 5. Late-stage collisions (107-8 yrs)
49From Guillot, 2004
50Magnetic fields
51Giant Impacts and Temperature
52Siderophile Elements
- Earth deep magma ocean required
- Mars shallow magma ocean (?)
2250K, 270 kbar
Righter AREPS 2003
53Hf-W system
- Core formation indicates an at least partially
molten silicate mantle (Stevenson 1990) - 182Hf decays to 182W, half-life 9 Myrs
Kleine et al. 2002