melt inclusions & volatiles in silicic magmatic systems · what can we infer about volatiles in...
TRANSCRIPT
Melt Inclusions & Volatiles in Silicic Magmatic Systems
Photo by D. Harlow, USGS
Outline
• Generation of intermediate & silicic magma
• Melt inclusions – host crystals & inclusion textures
• Volatile concentrations in silicic magmas
• Vapor saturation & magma chamber configurations
• Processes that cause volatile variations
• Inferring vapor compositions
• Sulfur & chlorine
How are intermediate and silicic magmas formed?
• Continental crust has mafic lower crust and more evolved granite-dominated upper crust
• Mantle-derived magmas in subduction zone settings are basalt to high-Mg andesite in composition
• Processes for forming more evolved magma:
• Differentiation of primary mafic magmas by crystallization
• Partial melting of older crustal rocks
• Partial melting of earlier formed cumulates/plutons in the lower crust
KOS data
Blundy MSH figures
Importance of Mafic Magma• Crustal magmatic systems are fundamentally basaltic – heat & mass
• Basaltic magma transfers volatiles from mantle to crust
• Mantle volatile sources include upper mantle & subduction-recycled components
Hildreth (1981)
Conceptual representation of a deep crustal hot zone(Annen et al., 2006)
• Hydrous basaltic melts intruded into thelower crust as sills
• Heat & H2O from the crystallizing basaltpromote partial melting of the lower crust
• Mixing of residual & crustal melts
• Ascent of H2O-rich melts into the uppercrust
• Degassing & crystallization at shallow depthlead to large increases in viscosity & stallingof magma to form magma chambers
Strong crystal fractionation signature in rhyolites
Halliday et al. (1991)
• Data suggest a genetic link between crystal-poor & crystal-rich rhyolites &I-type granitoids
• But how are viscous silicic melts physically separated from crystals?
Crystal mush model – Bachmann & Bergantz (2004)
Upward migration of low-density, residual melt from a crystallizing boundary layer is a popular, but problematic, hypothesis for shallow magmatic differentiation
• Difficult to explain old, complexly zoned phenocrysts• Large silicic magma bodies have sill-like aspect ratios• Re-entrainment of magma in cap may be relatively rapid
An alternative model for formation of intermediate & silicic magma
Bachmann &Bergantz (2004)
What can we learn from melt inclusions about formation, crystallization & storage of silicic magmas?
Montserrat – photo by B. Voight
Quartz-hosted melt inclusions, Bishop Tuff
Cathodoluminescence images
Wafers for FTIR & microbeam analysis
Rhyolitic melt inclusions in quartz phenocrysts
Secondary electron images of bipyramidalquartz from the 1912 eruption of Katmai showing semi-skeletal growth form & trappingof inclusions (Lowenstern, 1995)
QuickTime™ and a decompressor
are needed to see this picture.
Rhyolitic melt inclusions – vapor bubbles & quench crystals
Wallace et al. (2003)
• Melt inclusions in high-silica rhyolities are typically bubble free (a)• With slower cooling, inclusions develop a darker color (c), bubbles (b), & fine crystals (g, k, l)• In some cases, bubbles nucleate on daughter crystals (g, h, l)• Bubbles can also be caused by cracking (decrepitation) of the host
Rhyolitic melt inclusions in plagioclase
QuickTime™ and a decompressor
are needed to see this picture.
Transmitted light photomicrographsCrater Lake (Bacon et al., 1992)
• Melt inclusions in plagioclase areoften poorly sealed
• Vapor bubbles are common
SEM images of Mount St. Helensplagioclase (courtesy of J. Blundy)
Volatiles in melt inclusions from subduction zone rhyolites
Wallace (2005)
• Agreement between melt inclusion vapor saturation pressures & total pressure constrained by experimental phase equilibria
• Volcanic SO2 and CO2 emissions
• CO2 vs. trace elements in melt inclusions
• Fluid inclusions in phenocrysts in volcanic rocks
Independent Evidence for Vapor Saturation
Pasteris et al. (1994)
Quartz phenocryst, PinatuboPlagioclase phenocryst, Guagua
Pichincha, Ecuador
Photo courtesy of Jake Lowenstern, USGS
0
200
400
600
800
1000
1200
0 1 2 3 4 5 6 7 8
H2O (wt%)
YellowstoneBishop Tuff Pine
GroveTuff
Pinatubo
0.5 kb
1 kb
2 kb3 kb
N. Chile
Toba
KatmaiMazama
Data sources: Chesner & Newman (1989); Bacon et al. (1992); Lowenstern (1994); Wallace & Gerlach(1994); Gansecki (1998); Wallace et al. (1999); Schmitt (2001); Wallace (unpubl.)
Comparison of subduction zone & other rhyolites
CO
2(p
pm)
Bishop Tuff & Long Valley Caldera
Cathodoluminescence images ofquartz-hosted melt inclusion
Melt Inclusions from the Bishop Tuff, California
Wallace et al. (1999)
Crystalpoor
Crystal rich
Configuration of pre-caldera magma body, Long Valley Caldera, CA
5 km
10 km
North South
middle
late
unerupted mush
initial vent site
ring fractures formedduring eruption andcaldera collapse
no vertical exaggeration
early
Use of melt inclusions to infer magma body configuration
Wallace et al. (1999)
QuickTime™ and a decompressor
are needed to see this picture.
Wark et al. (2007)
Late Bishop Tuff preserves evidence of pre-eruption intrusion of hot, CO2-rich rhyolitic melt into crystal mush in lower part of magma chamber
Crystal-rich mush Intrusion by hotter rhyoliteDissolution of quartz
Quartz overgrowths traphigh CO2 melt inclusions
What processes cause variations in volatile contents?
Decompression – leads mainly to variations in CO2
QuickTime™ and a decompressor
are needed to see this picture.
Tuff of Pine Grove, Utah (Lowenstern, 1994)
2
3
4
5
6
7
8
0.1 1 10 100 10 3
Ba (ppm)
Late BT
Early BT
Middle BT
Sanidine fractionation Yellowstone
N. Chi le
PineGrove
H2O
(ppm
)
What causes variations in H2O content?
• Variation of H2O with Ba suggests fractional crystallization control onincreasing H2O contents
• Magmatic H2O contents increase during vapor-saturated crystallizationif CO2 is present
0
500
1000
1500
2000
2500
0 1 2 3 4 5 6 7
CO
2 (ppm
)
H2O (wt.%)
3 kb
2 kb
1 kb
vapor saturation isobars
CO
2(p
pm)
Vapor–Saturated Crystallization
Vapor–saturatedcrystallization
Anderson et al. (1989)
Exsolution of vapor during closed-system crystallization
<
<
volume
crystal %
vapor mass
volume
crystal %
vapor mass<
initial finalcrystals
exsolved vapormelt
% crystallization(C
O
/ CO
in
itial
) dis
solv
ed in
mel
t2
2
302010000.0
0.2
0.4
0.6
0.8
1.0
2%
3%
4%
7%
1%
2%
2%3%
6%
1%
5%2%1%0.5%0%
initial exsolved vapor (wt.%)
CO2 exsolution during vapor-saturated crystallization
Melt inclusions from clast CHAL-9
Loss of CO2 to Exsolved Vapor During Closed-System Crystallization
Wallace et al. (1995)
Wallace et al. (1999)
0
200
400
600
800
1000
1 10 100
CO
2 (pp
m)
Ba (ppm)
Melt Inclusions from the Bishop Tuff
Late
Early
Middle
Sanidine fractionation
Vaporexsolution
error
CO
2(p
pm)
Loss of CO2 to Exsolved Vapor During Vapor-Saturated Crystallization
What can we infer about volatiles in magmas that are parental to rhyolites?
• Note that the range of H2O in arc rhyolites is similar to that in arc basalts
• How are the two related?
Effects of fractional crystallization & ascent into upper crust
• Parental basaltic magmas must have relatively low H2O, and/or
• Lower crustal melting must involve relatively H2O-poor sources
Wallace (2005)
Kos Plateau Tuff – quartz crystals recycled from mush(?)
Bachmann et al.(in press)
• Crystal rich (≤40 vol%) rhyolitic tuff; eruptive volume >60 km3
• Geologic evidence suggests magma chamber drawdown of ~1 km during eruption
• Crystals may mostly be derived from mush zone beneath & surrounding chamber
Effects of composition & temperature on S contents of magmas
QuickTime™ and a decompressor
are needed to see this picture.
Shinohara (2008)
aa
0
500
1000
1500
2000
2500
3000
3500
0 1 2 3 4 5 6 7 8
H2O (wt.%)
Saturation with H
2O-rich vapor
Bishop Tuff
Pine Grove Tuff
Bandelier (CTR)
2 kbar
Saturation withconcentrated brine Taupo
Cl (
ppm
)
Chlorine in Rhyolitic Melt Inclusions
• Most non-mineralized rhyolites do not have enough Cl to be saturatedwith concentrated brine (hydrosaline melt)
Decompression-driven versus cooling-driven crystallisation
0
1
2
3
4
5
6
7
68 70 72 74 76 78
wt%
H2O
wt% SiO2 (anhyd.)
IV. Syn-eruptivedegassing
III. DecompressionH
2O-saturated
I. IsobaricH
2O-saturated
II. IsobaricH
2O-undersaturated
Inclusion populations record ascent trajectories – time and pressure variations
Blundy & Cashman (2008)
Decompression-driven crystallisation
• Decompression-driven crystallization (H2O loss) is much faster than crystallization driven by cooling
0
1
2
3
4
5
6
7
66 68 70 72 74 76 78 80
1980-81
Apr-10 (plag)Crypto (plag)Crypto (opx)May-18 (plag)May-18 (hbl)May-25 (plag)
Jun-12 (plag)Jul-22 (plag)Aug-7 (plag)Aug-7 (opx)Oct-16 (plag)Oct-16 (hbl)
Dec-27 (plag)Dec-27 (cpx)Jun 1981 (plag)May-18 (gm)1980 (gm)
wt%
H2O
wt% SiO2(n)
rupturedinclusions
Blundy & Cashman (2008)
Modified from Hochstein & Browne (2000)
Silicic magma bodies & hydrothermal systems
QuickTime™ and a decompressor
are needed to see this picture.
Student & Bodnar (2004)
Melt Inclusions from Porphyry Copper Deposits
QuickTime™ and a decompressor
are needed to see this picture.
Key points to remember:
• Melt inclusions in rhyolitic systems preserve a complex record of magmacrystallization, degassing, mixing and storage
• Melt inclusion H2O & CO2 data can be used to infer magma bodyconfiguration and depths of crystallization
• Rhyolitic magmas are typically vapor saturated in the upper crust.Some may be saturated with a hydrosaline melt (brine) phase
• Dissolved sulfur concentrations are low, but there may be considerableS as H2S and SO2 in the coexisting vapor phase
0
500
1000
1500
2000
2500
0 1 2 3 4 5 6 7
CO
2 (ppm
)
H2O (wt.%)
3 kb
2 kb
1 kb
Vapor saturation isobars
Vapor composition isopleths mol% H2O in vapor
10%20%
30%
40%50%
60%70%
80%
90%
Inferring Vapor Composition from Melt Inclusion Data
CO
2(p
pm)
Isobars & isopleths from VolatileCalc (Newman & Lowenstern, 2002)
0.50.40.30.20.10.0300
250
200
150
100early & middle inclusionslocal averagelate inclusions
X in vapor
Pre
ssur
e (M
Pa)
CO2
error
Exsolved Vapor Composition in Pre-Caldera Magma Body
Bishop Tuff
Wallace et al. (1999)
Importance of Mafic Magma
• Crustal magmatic systems are fundamentally basaltic
• Basaltic magma transfers volatiles from mantle to crust
“Degassing of basalt crystallizing in the roots of these systems provides a flux of He, CO2, S, halogens, and other components.”
Hildreth (1981)
“[Stable isotope] data suggest that magmatic fluxes of C and S are dominated by mantle sources”
0
50
100
150
200
250
700 750 800 850 900 950 1000
Temperature (°C)
Fish CanyonPinatubo
Toba
Redoubt
Krakatau
MSH
Ruiz
2 kbar (NNO)FeS saturated
Katmai
BishopTuff
S (p
pm)
Sulfur in Rhyolitic Melt Inclusions
• Many rhyolitic magmas are sulfide saturated• Saturation limit of S increases with temperature & oxygen fugacity