Lecture 2a: Igneous classification, midocean ridges

1
Lecture 2a Igneous classification, mid-ocean
ridges

• Questions
• How many igneous rock names should you learn, and
  what do they mean?
• What is a mid-ocean ridge and how does it work?
• How does plate spreading produce the oceanic
  crust?
• How do rocks observed at ridges constrain the
  state of the upper mantle?
• Tools
• Field and shipboard geology and analytical
  chemistry
• Thermodynamics
• Fluid dynamics
• Reading Grotzinger et al., chapter 4 Albarède,
  chapter 8

1
2
Igneous Classification

• Igneous rocks can be classified according to
  composition, mineralogy, texture and/or
  locality(!).
• The first distinction is between volcanic and
  plutonic rocks.
• Volcanic rocks are erupted at the Earths surface
  and cool very quickly. There is insufficient time
  to grow large crystals. This leads to formation
  of glass or very fine-grained rocks, or to
  phenocrysts (crystals that grew before eruption)
  in a fine groundmass.
• Plutonic rocks crystallize at some depth, and
  therefore lose heat relatively slowly. Crystals
  have time to grow after nucleation, and the
  resulting rocks generally have individual
  crystals large enough to see unaided.
• Rocks of exactly the same composition and
  mineralogy get different names in their volcanic
  and plutonic forms, because they look different!

2
3
Plutonic vs. Volcanic
3
4
Classification by mineralogy

• The standard classification scheme uses the
  mineralogy of the rock (how much quartz, how much
  plagioclase, etc.)
• There is one important twistfor volcanic rocks
  you usually cannot measure the actual minerals
  present (or it may be a glass and there are no
  minerals present).
• In this case, instead of the actual minerals, you
  classify based on normative mineralogy
• The norm is a calculation based on the bulk
  composition of a volcanic rock, for what minerals
  would be present if it were fully crystallized.
• The standard norm calculation is called the CIPW
  norm, after Cross, Iddings, Pirsson, and
  Washington (1902).

4
5
Classification by mineralogy

• Mineral content (actual or normative) of the rock
  by volume is divided into Quartz (Q), Alkali
  feldspar (A), Plagioclase (P), Feldspathoids like
  nepheline and leucite (F), and Mafic minerals
  like amphibole, biotite, pyroxenes, and olivine
  (M).
• For rocks with M lt 90, the Streckeisen
  double-triangle is used. It shows names defined
  by Q-A-P-F recalculated to 100.
• Many of these names are really obscure dont try
  to learn all of them.

5
6
Classification by mineralogy

• For rocks with mostly mafic minerals, a different
  scheme is used. The proportion of olivine,
  orthopyroxene, clinopyroxene, plagioclase, and
  hornblende locate a rock using the appropriate
  Streckeisen ternary diagram.

6
7
Classification by composition

• There are several classifications, of individual
  rocks or rock suites.
• By silica percentage
• SiO2 Designation Dark Minerals Designation Exam
  ple rocks
• gt66 Acid lt40 Felsic Granite, rhyolite
• 52-66 Intermediate 40-70 Intermediate Diorite,
  andesite
• 45-52 Basic 70-90 Mafic Gabbro, basalt
• lt45 Ultrabasic gt90 Ultramafic Dunite, komatiite
• By alumina saturation (this controls which dark
  minerals show up)
• Chemistry Designation Distinctive Minerals
• Al2O3gtNa2OK2OCaO Peraluminous Muscovite,
  biotite, topaz,
• corundum, garnet, tourmaline
• Na2OK2OCaOgtAl2O3 Metaluminous Melilite,
  biotite, pyroxene
• Al2O3 gt Na2OK2O hornblende, epidote
• Al2O3 Na2OK2O Subaluminous Olivine,
  pyroxenes
• Al2O3 lt Na2O K2O Peralkaline Sodic pyroxenes
  amphiboles

7
8
Classification by composition

• By Alkali-Lime index for a suite of rocks, CaO
  and Na2OK2O are plotted against SiO2. Generally,
  CaO decreases with increasing SiO2 while Na2OK2O
  increases. Suites are classified by the SiO2
  where the intersection occurs

Rock Suite Alkali-Lime Index Illustrative rock
series Calcic gt61 SiO2 Mid-ocean ridge
basalts Calc-alkaline 56-61 Continental margin
arc series Alkali-calcic 51-56 Some
intraoceanic island arcs Alkaline lt51 Intrapla
te continental melts
8
9
Ocean crust geology

• Recall the typical sequence of rocks observed in
  ophiolite exposures and in drilling the ocean
  crust

• Deep-sea marine sediments
• Massive sulfide deposits
• Pillow basalts
• Sheeted basaltic dikes
• Layered gabbro
• Serpentinized peridotites

• This sequence is consistent with the seismic
  velocity profile of oceanic crust

9
10
Ocean Crust Geology
Modern and ancient pillow basalts, photographed
by submersible and by field geologist,
respectively.
10
11
Ocean Crust Geology
Modern and ancient sheeted dike complexes
observed by seismology and by field geologist,
respectively.
11
12
Ocean Crust Geology
Modern and ancient layered gabbro (oceanic layer
3) complexes observed by ocean drilling and by
field geologist, respectively.
12
13
Ocean Crust Geology
Modern and ancient harzburgite/dunite uppermost
mantle assemblages An abyssal peridotite and the
Muscat massif in Oman
13
14
Mid-ocean Ridges

• Now lets discuss the origin of this rock
  sequence in the detailed mechanisms and
  variations that occur along ridges.
• Fluid dynamics plate-driven flow, internal
  buoyancy
• Physically we think of a ridge as a viscous fluid
  asthenosphere overlain by a lithosphere whose
  thickness is very small at the ridge axis, and
  which is being pulled laterally apart along a
  linear rift by externally-imposed forces. This is
  approximately a two-dimensional flow.
• Flow in the mantle is slow enough that inertia
  can be neglected (i.e. the Reynolds number is
  tiny) hence the 2-D stream function y (a scalar
  function whose contours are everywhere parallel
  to the flow vector and whose spacing is
  proportional to the velocity) for incompressible
  flow satisfies the biharmonic equation

14
15
Mid-ocean ridge dynamics

• In a simple uniform viscosity case, the steady
  solution to this corner flow is sort of given by
  Batchelor (1967 he actually gives the solution
  for different boundary conditions)

15
16
Mid-ocean ridge dynamics

• Things to note about this flow field
• It fills all of half-space, out to infinite depth
  and infinite lateral extent.
• The pressure goes to 8 at the corner!
• Flow under the ridge axis is near vertical, flow
  far out to the side is nearly horizontal, but
  actually there is a positive upward vertical
  component to flow everywhere.
• There are real complexities superposed on this
  simple solution that should be noted right away
• The lithosphere thickens away from the ridge axis
  with the square root of time, so the upper
  boundary of the asthenospheric flow is not
  horizontal. Streamlines get gradually
  incorporated into the lithosphere and the depth
  of material points becomes fixed.

16
17
Mid-ocean ridge dynamics More complexities

• The viscosity is NOT constant. Viscosity
  variations are due to
• Temperature to first order this is the
  difference between lithosphere and asthenosphere,
  but it is not really a sharp cutoff.
• Strain Rate the viscosity may be strain or
  stress-dependent, so that areas flowing fast are
  weaker and tend to concentrate strain ever more.
• Melting the presence of melt above the solidus
  may weaken the partially molten region
• Water the 100 ppm dissolved H2O in mantle
  olivine lowers the viscosity by a factor of 500.
  This H2O is removed to the melt phase when
  melting begins, so melting may actually increase
  the viscosity!
• For your information, the viscosity of mantle
  flow in and near the melting region is somewhere
  in the range 1018 to 1023 Pa-s

17
18
Mid-ocean ridge dynamics More complexities

• The drag from the upper boundary condition is not
  the only driving force for flow. There is also
  internal buoyancy. Density variations are due to
• The presence of retained melt in the partially
  molten region. Silicate liquids are substantially
  less dense than mantle minerals at these
  pressures. The more melt is retained, the more
  buoyant the rock.
• The change in composition of the residue with
  progressive melt extraction. Because the melt
  concentrates Fe relative to Mg, and because Al is
  incompatible and forms the dense mineral garnet,
  the solid residue actually becomes less dense
  even as low density melt is being extracted!
• This is not non-physical there is no law of
  conservation of volume
• The flow is not incompressible both melt
  production and melt extraction cause changes in
  the volume of the fluid (really a multiphase
  medium).

18
19
Mid-ocean ridge dynamics More complexities

• Here are some models of realistic flow fields
  incorporating all these complexities, showing
  relative importance of plate drag and internal
  buoyancy as functions of spreading rate and
  mantle viscosity.
• The figures show streamlines and the shading
  shows temperature.

19
20
Melting and melt extraction

• Most of the mantle is solid. When mantle material
  is dragged upwards towards the ridge by the
  spreading-induced corner flow, it is initially
  solid. Why does it melt? How does it melt?
• (see my chapter from the Encyclopedia of
  Volcanoes)
• The temperature path along which a large body of
  mantle upwells is to very good approximation
  adiabatic and reversible and therefore
  isentropic. In problem set 3 we calculate the
  formula for the temperature change along an
  isentrope in the absence of phase changes
• For mantle materials at 1500 K, this slope is 10
  C/GPa.
• The experimental solidus of fertile peridotite
  has a slope of 130 C/GPa
• It follows that unless the mantle is really cold
  (and it is not), the adiabat will intersect the
  solidus at some point upon decompression

20
21
Melting and melt extraction

• Where the intersection of adiabat and solidus
  occurs depends on the potential temperature of
  the mantle (the temperature where the adiabat
  would reach 1 atmosphere if no melting took
  place).
• For ordinary regions, this is close to 1350C
  (1250 to 1500 global range) and the solidus is
  intersected at about 2.5 GPa (1.5 GPa to 6 GPa
  global range).

21
22
Melting and melt extraction

• Once melting begins on an upwelling streamline,
  it will continue until either
• upwelling ceases
• conduction to the surface cools the system
• internal thermochemistry of the residue slows or
  stops melting
• The cartoon version of the melting regime under
  the ridge axis is then drawn on the cartoon
  version of the flow field as some generally
  triangular shape in cross-section

22
23
The mid-ocean ridge Melting Regime

• The bottom of the melting regime is drawn
  horizontal on the assumption that the mantle is
  locally isothermal before melting begins
• The upper edges of the triangle either represent
  where the flow lines become effectively
  horizontal or where cooling from the surface has
  penetrated to the relevant depth
• The extent of melting increases upwards along
  each streamline from the solidus intersection to
  the exit of the melting regime (indicated by
  shading in the figure)

23
24
The mid-ocean ridge Melting Regime

• The erupted melt is going to be some average of
  the melts produced throughout this melting regime
• Several models are possible of how and where the
  melt is extracted and what happens to it during
  transport
• This average melt is primary mid-ocean ridge
  basalt (MORB).

The melting regime is wide (hundreds of
kilometers), but eruptions are focused in a
neovolcanic zone only a few km wide melts have
to be focused somehow to the ridge axis. This
figure is the result of a large project to
seismically image the melting regime on the East
Pacific Rise.
24
25
Melt productivity

• After melting begins, the P-T path and the amount
  of melt production can be inferred from
  conservation of entropy.
• At constant total entropy, the only way for the
  system to find the entropy of fusion is by
  cooling more steeply than the slope of the solid
  adiabat.

25
26
Melt Migration

• There are multiple mechanisms to get the partial
  melt out of the residue and deliver it to the
  crust at the ridge axis.
• Porous flow
• The melt phase quickly forms an interconnected
  network along the grain boundaries of the rock
  (mantle source rocks are made of crystals about
  1mm-1cm in diameter).

• Interconnection occurs because the energy of a
  melt-crystal interface is low enough that the
  melt adopts a high-surface area geometry.
• This geometry implies that the partially molten
  rock is permeable to flow of the melt relative to
  the solid.
• Permeable flow is governed by DArcys Law
• where P is pressure, k is permeability, f is
  porosity or melt fraction, m is the viscosity of
  the melt phase, and v is the velocity of melt
  relative to solid.

26
27
Melt Migration

• The pressure gradients driving porous flow of the
  melt arise from buoyancy of the melt (density of
  mantle 3300 kg m-3, density of basaltic liquids
  2700 kg m-3), as well as from deformation of the
  solid matrix.
• The permeability is normally an increasing
  function of melt-filled porosity and grain size
  (or spacing of melt channels).
• The exact function for the equilibrated mantle
  melt geometry is unknown, but it may be k d2f2
  or d2f3.
• For a grainsize of 1 mm and a melt-filled
  porosity of 1, the permeability is 1011 cm2.
• This translates, for a basaltic melt with
  viscosity 10 Pa s, into migration velocities
  10 mm/yr, not much faster than solid flow rates!
  
• We know melt somehow eventually moves much faster
  than this, so there must be other melt flow
  mechanisms.

27
28
Melt Migration

• How do we know that the melt-filled porosity
  resulting from melting coupled to melt migration
  is 1?
• We cannot guess accurately from the balance of
  melting rate and permeability, since the
  permeability is a poorly known function of melt
  fraction.
• But we have seismology, which does not show the
  extraordinarily low velocities that would be
  expected for large melt fractions.

And we have chemistry of residues (There is a
homework question on this)
28
29
Melt Migration

• During porous flow at low melt fractions, melt
  and solid are expected to remain in chemical
  equilibrium. But primary mid-ocean ridge basalt
  is not in equilibrium with the uppermost oceanic
  mantle. So again, something else must happen.
• Fast porous flow
• Porosity waves The equations for porous flow in
  a viscous two-phase system allow solutions in the
  form of solitary waves that may propagate much
  faster than the DArcy velocity and speed up melt
  transport

(contours of porosity gridded area is lt1 this
figure shows the collision of two magma solitons
moving upwards in a deformable porous medium)
29
30
Melt Migration

• High-porosity zones, reactive transport channels
  focusing the melt flow

This may naturally arise due to the reactive
infiltration instability, which makes flow of a
liquid break up into fingers if it can dissolve
the pyroxenes out of the solid, leaving high melt
fraction and olivine residue.
The evidence for this process in nature is veins
of dunite (pure olivine) in ophiolites if melt
flows through the dunite channels, this may
explain the observation that mid-ocean ridge
basalt is not in equilibrium with abyssal
peridotite (when you find dunites, these are in
equilibrium with MORB).
30
31
Formation of the crust, magma chambers,
differentiation

• The end result of mantle melting and melt
  migration is the delivery of primary magma into
  the crust. Here it begins to cool by conduction
  and hydrothermal convection and hence to
  crystallize.
• Fractional crystallization has two essential
  consequences
• the formation of a gabbroic lower crust from the
  crystallized fraction
• the compositional evolution of the remaining melt
  before eruption.

Pprimary MORB crystallizes olivine, then
olivineplagioclase, then olivine
plagioclaseclinopyroxene. These are the minerals
of the oceanic lower crust and also the
phenocryst phases in erupted MORB. The primary
magma never erupts there are no MORBs that are
in equilibrium with orthopyroxene (a ubiquitous
mantle phase) at any pressure!
31
32
Formation of the crust, magma chambers,
differentiation

• Where does the differentiation take place?
• The phase equilibria indicate low pressure
  (0-2000 bars)
• seismic images show a melt lens 1 km below the
  axis of the East Pacific Rise.
• This suggests crystallization takes places in the
  shallow melt lens, and the gabbroic crust is
  formed by ductile flow of new gabbro down and
  away from the shallow melt lens.

32
33
Primary liquids?

• So if the primary liquid never erupts, what can
  we say about melting conditions in the mantle?
  How do we see back through the filter of
  differentiation?
• One solution is to pick an arbitrary MgO content
  at which to compare data and models. MgO
  systematically and monotonically decreases during
  differentiation, so it is a good index to
  normalize against.

Here is the correction of data to 8 MgO. Note
that variation along the liquid line of descent
is the first principal component of variability
within basalt samples from a given location, but
there remains local variability at 8 MgO, and
when each region is averaged there are systematic
differences between regions
33
34
Global systematics

• The values of regionally-averaged Na8 (i.e., Na2O
  concentration corrected to 8 MgO), Fe8, water
  depth above the ridge axis, and crustal thickness
  show significant global correlations.
• Where Na8 is high, Fe8 is low
• Where Na8 is high, the ridges are deep
• Where Na8 is high, the crust is thin

34
35
Global systematics

• What is the interpretation of the global
  correlations in Na8, Fe8, axial depth, and
  crustal thickness?
• Answer Na8, an incompatible element, is an
  indicator of mean extent of melting. Fe8 is an
  indicator of mean pressure of melting. Axial
  depth is an indicator of mantle temperature,
  extent of melting, and crustal thickness
  combined.
• So the global correlation implies that extent of
  melting and pressure of melting are positively
  correlated, on a global scale.

35
36
Synthesis of global systematics

• The correlation of extent of melting with
  pressure of melting requires that the first-order
  control on variation among ridge segments is
  mantle temperature
• Some places the mantle is cold. Some places it is
  hot. This causes variations in the depth where
  melting begins. If melting continues under the
  axis to the base of the crust everywhere, then
  hot mantle means long melting column, high mean
  pressure of melting, and high mean extent of
  melting, high Fe8, high crustal thickness, low
  Na8, and shallow axial depth. Cold mantle yields
  the opposite.

36