1
Lithium Isotope Geochemistry applications to
high-temperature processes

• Dr. Ian Buick
• Research School of Earth Science
• The Australian National University
• 4th Feb 2008

2
Overview

• 2 isotopes 6Li (7.5 ) and 7Li (92.5)
• Data as ?7Li ( ) 1000((7Li/6LiUnk -
  7Li/6LiL-SVEC)-1), where 7Li/6LiL-SVEC 12.0192
  (also see ?6Li used)
• L-SVEC manufactured Li2CO3 standard (NST
  SRM-8545)
• Large mass difference between 6Li and 7Li (16)
  large fractionations due to (low T)
  geological processes large range of
  ?7Li in earth materials (gt 60 )

3

• from Tomascak RIMG v55 (2004)

4
Uses for Lithium isotopes

• Li-isotopes may provide information about
• Subduction-related addition and crustal recycling
  at active continental margins
• Fluid transfer processes in subduction zones
• Material cycling between mantle and oceanic crust
• Isotopic fractionation during metamorphism,
  melting and fractional crystallisation

5
Lithium isotope cycle MOR
6

• Lithium isotope cycle Subduction Zones

Elliot et al. (2004) EPSL
7
Methods of ?7Li data collection

• Bulk rock (TIMS or MC-ICP-MS) v. Ionprobe (SIMS)
• Potentially extreme Li-isotopic fractionations
  during wet chemistry need for 100
  extraction of Li
• TIMS significant 7Li/6Li mass fractionation on
  surface of filament. Problems overcome by
  std-bracketed MC-ICP-MS analysis
• SIMS Required due to common occurrence mineral
  inclusions (metamorphic rocks) documented
  extreme ?7Li heterogeneity on a 10s-100s
  ?m-scale in some minerals

8
Li abundances (ppm)

• Li distribution coefficients are commonly poorly
  known as a function of P-T, mineral composition

9
SIMS analytical set-up

• Analyses undertaken on Cameca 3f (Nancy) 4f
  (Edinburgh) 3f 6f (Arizona) 1270 (Misasa,
  Japan Edinburgh)
• Edinburgh 4f method (Kasemann et al., 2005)
• secondary 6Li 7Li produced by a 5-40nA,
  15kV, 16O primary beam focused to 15-40?m spot
  size.
• energy window of 52 eV, at a mass resolution of
  1200 (to resolve closest interference, 6Li1H).
• counted by peak jumping on an ETP electron
  multiplier.
• internal uncertainty of lt1 (1??mean)
  7Li/6Li ratio measured over 50-120 cycles
  (depending on Li ppm), each cycle consisting of
  5-s and 2-s integrations of 6Li and 7Li
  10-50 minutes/analysis.

10
SIMS Standards

• Glass stds widely available
• BCR-2G (USGS) ?7Li 4.08 1.0 , 92 ppm Li
• GSD-1g (NIST) ?7Li 31.140.8 , 30-40 ppm
  Li
• both from Kasemann et al. (2005, Anal. Chem.)
• MPI-DING glasses ( 5 stds with ?7Li 2.0 to
  17.1 Jochum et al. 2006 G3)
• Minerals typically in house developed using
  TIMS/MC-ICP-MS cross calibration, not widely
  available
• Ol (Arizona, Nancy, Misasa)
• Cpx (Misasa, Nancy, Japan)
• Crd (Edinburgh)

11
Matrix Effects

• Instr. mass fractionation (?inst
  7Li/6Litrue/7Li/6Limeas gt1) typically corrected
  for by use of glass stds
• Currently lack of consensus and/or knowledge
  of extent of matrix effects
• Decritre et al. (2002 G3) no matrix effects
  on ?inst for range of low-SiO2 minerals (Ol,
  Opx, Cpx, Amph, Bt etc) using low -SiO2
  (basaltic) glass std
• Kasemann et al.(2005) 4-5 matrix effect for
  high-Si glass using basaltic glass std.
  Unpublished large matrix effect for Crd as a fn
  of XMg.
• Bell et al. (2007) 26 matrix effect in Ol as
  a function of XMg (0.94-0.74)

12
Processes that may potentially change ?7Li values

• Rayleigh-type fractionation e.g. metamorphic
  devolatilisation, fractional crystallisation,
  fluid exsolution 6Li partitions into solids
  phases over fluids or melts.
• Kinetic diffusion driven by mineral-scale
  concentration gradients a potentially major
  problem for mantle samples and mantle-derived
  magmas possibly also during crustal
  metamorphism and magmatism..

13
Fractional Crystallisation

• eqm Li-isotope partitioning (?7LiMin-Min or
  ?7LiMin-Melt) a fn of T and Li coordination
• v. small little shifts in ?7LiWR in
  differentiation series small ?7Li Min-Min
  or -Melt at T of mafic magmatism

14
Fractional Crystallisation

• Li in 4-fold co-ordn in high-Si melts,but Li in
  minerals may be 8-fold (Grt) 6-fold (e.g.Crd,
  Micas, Px), 4-fold (Sta). 6Li favours larger
  co-ordn number

• significant ?7Li(Min-Melt) and
• ???????????7LiMelt
• inc. for extreme fraction at low T
• little change in
• ?7LiMelt due to fluid exsolution

from Teng et al. (2006, Am. Mineral)
15
Metamorphic devolatilisation

• some contradictory results
• Alpine eclogites may have very low ?7Li
  compared to MORB ie ?7Li(Omph) as low as -11
  (Zack et al., 2003 EPSL). Used as a basic input
  into recent subduction zone Li cycling models
• Zack et al. model recently challenged by
  coupled ?7Li - metamorphic modelling-mass balance
  approach (Marshall et al., 2007, EPSL)
  devolatilisation cannot account for such low
  ?7Li(WR). Therefore what is the cause?

16
Metamorphic devolatilisation

• studies of metamorphism in contact aureoles and
  regional LP/HT metamorphism imply little or no
  change in ?7Li(WR) from Chl Zone to beginning
  of anatexis e.g Teng et al. (2007, Chem Geol.)

from Teng et al (2007, Chem. Geol.)
17
Kinetically-driven fractionation
Li-doped
RB5 6 mins _at_1350C
RB4 1 hr _at_1350C

• Exp diffusion couples at 1350-1450 C for
  0.1-15hours, 12-13 kbar
• Diffusivity Li 102-103 x gt all other major and TE
• Diffusivity 6Li gt7Li by 2-3

6 mins _at_1350C
from Richter et al (2003, GCA)
18
Isotope modelling

• model assumes 3 relative difference in
  diffusivity of 6Li v. 7Li ?7Li initially
  homogeneous 10x Li concentration gradient.
• trough of low ?7Li migrates towards the low-Li
  phase as initial Li-concentration step function
  is removed due to diffusion

19
Kinetically-driven fractionation

• Diffusion due to gradient between low-Li
  phenocrysts (Cpx/Ol) and high-Li matrix
  Li-isotope fractionation.
• Developed in lavas with recrystallised matrix v.
  glassy matrix cooling history ?7Li
  of mantle minerals (and Melt Incs) may not
  give source information

after Beck et al. (2006), GCA
20
Kinetically-driven fractionation potential
processes

• changing D(Li)Min-Min or D(Li)Min-Melt during
  cooling
• dissolution of xenocrysts in melt
• infiltration of out of equilibrium melts magma
  mixing
• fluid exsolution into country rocks adjacent to
  plutons
• homogenisation of Li-growth zoning during
  prograde metamorphism

21
Is kinetic fractionation always a problem ?

• may depend on the T range and minerals of
  interest
• there are few experimental studies of Li
  diffusivity for different materials (Ol (?) lt Cpx
  ltltPlg Glass)
• can look at empirical studies of natural Li
  diffusion couples at Tgt 700 C.
• Expect LiCrd, Bt gtgt LiPl, Kfs LiGrt,Opx
  gt LiPl,Kfs

22
Granulite-facies metased, S. India

• highly restitic Li-poor, metapelitic lenses
  (Crd-Grt-Sil-Hrc-Bt) within relatively Li-rich
  leucogranite.
• Feldspar thermometry in leucogranite T 800C

23
Granulite-facies metapelite, S. India

• Preliminary data indicate Grt (also CrdBt) are
  little or unaffected by Li diffusion on a
  100?m-mm scale at T 750-900C.

Ian Buick, unpublished data
24
?7Li Mineral-Fluid factors

• Limited experimental data for staurolite, Li
  white mica and Li-clinopyroxene
• Different behavior Sta due to tetrahedral Li
  co-ordn.
• Potential use for estimating ?7Lifluid
  liberated during devolatilisation e.g in
  subduction settings

25
?7Li in Crd LP/HT melting

• Crd is product of 3-3.5kbar, 680-810C melting
  of metapelites at Mt. Stafford, central Oz.

26

• grain-scale uniformity in ?7Li(Crd) and values
  are low.
• very little change in ?7Li(Crd) with inc. grade
  (?T 100C) most Crd is produced at 675-700C
  (3.5 kbar) from BtSil dehydration melting in
  which Bt is rapidly exhausted

Ian Buick, unpublished data
27

• Restitic B-rich granulite-facies metapelites
  have much higher ?7Li(Crd) than in B-poor
  restitic metapelites at the same grade

• preservation of distinctly different ?7Li(Crd)
  in different bulk comps at T800C suggests
  potential for isotopic tracing

Ian Buick, unpublished data
28
?7Li in pegmatites

• Core to Rim dec in Li, HREE, Y, Ti
  magmatic zoning
• high ?7Li(Grt) consistent with advanced
  fractionation
• little variation in ?7Li(Grt) Li
  gradient doesnt drive diffusive fractionation
  in Grt, at least at 650C

Ian Buick, unpublished data