Melt inclusions in basaltic and associated volcanic rocks

1
Melt inclusions in basaltic and associated
volcanic rocks
Adam Kent, Oregon State University
2
Melt inclusions An introduction
Parcels of melt trapped in igneous crystals ?
Fluid inclusions
Occur in basaltic and related rocks wherever they
are found Arcs, OIB, CFB, MORB, LMI,
ETs Silicic and Plutonic Rocks Xenoliths
3
Saal et al. 1998
4
Scope Basaltic and related volcanic rocks
5
Scope Basaltic and related volcanic rocks
Nanos gigantum humeris insidentes Bernard of
Charles, 1159
6
Why study melt inclusions?
Melt inclusions preserve compositions that are
different from those of erupted lavas/tephra
7
9N Mid Atlantic Ridge
Ultra-depleted
Sobolev and Shimizu, 1993
8
Why study melt inclusions?
Melt inclusions preserve compositions that are
different from those of erupted lavas/tephra

• More variable than host and associated lavas
• Bulk rock, Matrix glass ? Averages
• Provide larger data sets per rock
• Preserve low volume or low survivability melts
• Primitive Melts
• Trap volatile elements
• Compare volatile and non volatile behaviour
• Provide melt samples in altered rocks

9
But theres a catch
Melt inclusions are NOT a universal panacea!

• Specific samples phyric rapidly cooled
• More work/time/money per sample
• Mineral separation, mounting and polishing
• Specialized analysis techniques
• Melt inclusions are small!
• Typically (trace element and isotope) analyses
  are less precise
• Isotopic data are limited
• Require significant additional interpretation

10
Melt Inclusion Variations

• Magmatic
• Crystallization
• Assimilation
• Magma mixing
• Source heterogeneity
• Degassing

• Inclusion-specific
• Boundary layers
• Post-entrapment crystallization
• Re-equilibration with host or external melt
• Non representative trapping

i.e. things that drive changes in magma
compositions
i.e. things that are unique to inclusions
11
Inclusion-Specific Processes

• Re-equilibration between inclusion and host
• Portnyagin et al. 2007, Spandler et al. 2007,
  Cottrell et al. 2002, Danyushevsky et al. 2000
  Gaetani and Watson, 2000, 2002
• Preferential trapping of unusual,
  non-representative compositions
• Michael et al. 2002, Danyushevsky et al. 2004,
  Yaxley et al. 2005
• Trapping boundary layers
• Kohut and Nielsen, 2004 Faure and Schiano, 2005,
  Baker et al. 2008. Goldstein and Luth, 2007
• Alteration of inclusions
• Nielsen et al. 1998

12
Analysis
plus speciation
13
How do melt inclusions form?
The widespread occurrence of melt inclusions in
basaltic rocks shows that their formation is a
normal part of the process of crystallization in
igneous rocks
Melt inclusions form in regions of relatively
slow crystal growth
14
How do melt inclusions form?
Modified from V.S. Sobolev and Kostyuk 1975
Roedder, 1979, 1984
15
Do melt inclusion formation processes fractionate
trapped compositions?
Faure and Schiano 2005
16
Baker et al. 2008
17
Not all experimental studies show boundary layer
effects
18

• Most natural suites do not show clear indications
  of boundary layer effects
• Perhaps we sample larger inclusions (only
  significant at lt 30 µm)
• Longer isothermal times in natural samples
• Are boundary layers static?
• Kinetic experiments

19
Evolution of melt inclusions after trapping
Important impact on physical appearance and
chemical compositions
25 µm
25 µm
20
Evolution of melt inclusions after trapping
Wallace, 2005

• Venting/breaching/alteration
• Post-entrapment crystallization
• Diffusive exchange

21
Correction for postentrapment crystallization

• Experimental
• Reheat to (estimated) trapping temperature
• Numerical
• Based on chemical equilibrium
• Olivine KDFeO/MgO 0.33 0.03

Loihi Seamount (Kent et al., 1999)
22
Compatible elements are the least robust after
correction for post-entrapment crystallization
23
Equilibration between Host and Inclusion

• Equilibration more rapid at
• Higher Diffusivity
• Higher Temperatures
• More compatible
• Larger inclusion
• Smaller host

Qin et al. (1992)
24
Fe Loss
Danyushevsky et al. 2000
Yaxley et al. 2005

• Negative correlation between measured FeO and
  Fohost
• Anomalously low FeO wrt liquid line of descent

25
Trace element re-equilibration

• The most robust data sources in melt inclusions
  are slow diffusing and incompatible elements
• Altered only by dilution/concentration
• Ratios unchanged

Are incompatible trace elements affected by
diffusional re-equilibration?
26
Cottrell et al. 2002
REE equilibration with host after 2500 years
27
Cottrell et al 2002
28
(No Transcript)
29
Slater et al., 2001
30
(No Transcript)
31
Trace element re-equilibration
Spandler et al. 2007
32
(No Transcript)
33
Baffin Island olivine-hosted n 103
34
Preserve inter-crystal variations
35
Driving Force?
36
The message from melt inclusions Variability

• In many basaltic systems it is clear that the
  primary control on melt inclusion compositions is
  the variability of melts present within the
  system
• These are sampled by erupted lavas as well, but
  are homogenized
• Implies large scale mixing of smaller melt
  batches is extremely widespread
• Melt inclusions and host lavas related by mixing

37

• Basaltic melt generation and transport systems
  are variable at scales smaller than individual
  eruptive units (factors of 10s)
• Phenocrysts
• Melt inclusions sample this variation
• Some real and apparent homogenization (mixing)
  occurs prior to eruption
• Rates Transport gtgt Re-equilibration

38
Comparison between melt inclusions and host lavas
Baffin Island olivine-hosted
Melt inclusions sample the same population of
melts as host lavas Variability in trace element
composition is driven by the same processes in
inclusions and in lavas
39
Magma
Magma
Melt Inclusion
Kellogg et al. 2002
40
Baffin Island olivine-hosted
41
Borgahraun, Iceland
Maclennan et al. 2003
42
Comparison between host and inclusions provides a
means to assess relationship between inclusions
and magmatic systems
43
9N MAR
Ultra-depleted
Sobolev et al 2000
Sobolev Shimizu 1993

• Anomalous melt inclusions
• Low volume melts?
• Magma chamber or primary?
• Artifacts of trapping?

44
There is no necessary connexion between the
size of an object and the value of a fact,
andthough the objects I have described are
minute the conclusions to be derived from the
facts are great Sorby 1858 Geol. Soc. London.
Quart. Jour. 14 453-500 from Roedder (1979)
Bull. Mineral.