Partial melting of amphibolites and the genesis of Archaean TTG (and some geodynamical implications) Jean-François Moyen and Gary Stevens Stellenbosch.

Partial melting of amphibolites and the

genesis of Archaean TTG(and some geodynamical implications)

Jean-François Moyenand

Gary Stevens

Stellenbosch University, South Africa

TTG are...• Orthogneisses• Tonalites, Trondhjemites & Granodiorites

(Na-rich series)• Fractionnated REE, etc.• Largely homogeneous throughout the

Archaean• Originated by partial melting of amphibolites

(hydrated basalts), in garnet stability field

Trace elements features of Archaean TTGs

Nb-Ta anomaly

Sr contents

Y & HREEdepletion

Les « gneiss gris »

Minéralogie

Eléments majeurs

REE

Conditions for making TTGs

Experimental melts

In Garnet stability field (Gt in residue)

Melting of hydrous basalt

KD

Gt/melt= 10 - 20

(other minerals ≤ 1)

Yb

Geodynamic site ?

Thick (oceanic or continental) crust(e.g. Oceanic plateau)

Subduction

Intermediate cases:• Shallow subduction

(± underplating)• Stacked oceanic crust

Gt-in

Gt-in

Gt-in

Gt-in

Gt-in

Partial melting of amphibolites

15-20 « modern » studies(1990-2000)

+ Phase diagrams (1970-80)

114 exp. fluid present or saturated

209 exp. « dehydration melting »

Goal of the study

• Review and compilation of published data on experimental melting

• Elaboration of a global model for amphibolite melting

• Implications for trace element contents

• Geological/geodynamical consequences

Review and compilation of published work

• Starting materials

• Solidus position & melt productivity

• Mineral stability fields

(Moyen & Stevens, subm. to AGU monographs)

Starting materials

Fluids and melting

• Fluid-saturated (free fluid phase)

• Fluid-present (yielded by breakdown of hydrous minerals in the near sub-solidus), limited availability

• Fluid-absent (dehydration melting)

• Dry

Fluid saturated

Dehydration melting

Fluid-present

Experimental solidus position

Melt productivity: dehydration melting

Melt productivity: water saturated

(+ Qz)

Melt productivity: fluid-present

(- Qz)

Mineral stability limits

Control on amphibole stability

Control on plagioclase stability

Mineralogical models

CaO 11.0 10.0 9.0

Na2O 2.2 2.8 3.3

K2O 0.1 0.5 1.0

TiO2 1.2 2.1 0.8

Amp. Comp.Ti-rich

High Mg#

Si poor

Int.

Ti-poor

Low Mg#

Si rich

KoB ThB AB

Quartz 0 1 10

Plagioclase 25 40 54

Amphibole 75 59 36

Amp:Plag 3:1 3:2 2:3

Mineralogical models

KoBKoB ThBKoB ThB AB

Composition of experimental melts

Very unlikely for amphibolite melting!

Na2O contents in experimental melts

K2O

Major elemen

ts

A linear model, of the form

C/C0 = a F + b

Modelled melts

Model vs. TTGs

Preliminary conclusions (1)

• K2O content depends on the source. Only relatively K-poor sources (< 0.7 %) make TTGs … but really depleted sources won’t.

• This means that K-rich amphibolites can indeed melt into granites (Sisson et al., 2005)

• With appropriate sources, tonalites & trondjhemites occur for F = 20-40 % (900-1100 °C)

Model for trace element

Cl = C0

F + D (1 - F)

Experimental data

D = Kdi. Xi

Arbitrary

Litterature

KoBKoB ThBKoB ThB AB

Trace elements contents of the 3 sources

Melt proportions

KoB ThB AB

Mineral proportions: amphibole and plagioclase

KoB ThB AB

Mineral proportions: garnet

KoB ThB AB

KD

Gt/melt= 10 - 20Yb

Mineral proportions: rutile

KoB ThB AB

KD

Rt/melt= 25 - 150Nb

KD

Rt/melt= 50 - 200Ta

REE contents in (modelled) melts

KoB ThB AB

REE contents in (modelled) melts

KoB ThB AB

REE contents: La/Yb

KoB ThB AB

Y contents

KoB ThB AB

Sr contents and the role of residual plagioclase

(Martin & Moyen, 2001, Geology 30 p 319-322; after Zamora, 2000)

Sr/Y

KoB ThB AB

Nb/Ta

KoB ThB AB

Effect of pressure

TTG composition as a depth indicator

Nb-Ta anomalyand Nb/Ta

Sr contents

Y & HREEdepletion

TTG composition as a depth indicator (cont.)

HREEdepletion

Eu anomaly

Preliminary conclusions (2)

• Appropriate depletion in Y, Yb, etc. requires pressures above ca. 15 kbar (rather than 10 kbar = Gt-in)

• Y, Yb, Sr/Y, Nb/Ta etc. are indicators of melting depth

• Low- and high-pressure TTGs with contrasted signatures?

High P TTGs

Low P TTGs

Not really TTGs

Archaean granulites (and intraplate geotherms)

Subduction of younglithosphere

(5 M

a)

(20

Ma)

(50

Ma)

Subduction of oldlithosphere

Tonalites & trondhjemites

(F = 20-40 %)Appropriate trace elts. signature

High Sr, La/Yb, Nb/TaLow Y, Yb

Low Sr, La/Yb, Nb/TaHigh Y, Yb

TTG genesis in P-T space

A regional example

• Barberton, South Africa

• 3.5 to 3.2 greenstone belt and gneisses

Swaziland

R.S.A

.

20 km

Crust accretion around BSB3600-3500 Ma

Steynsdorp pluton

3509 ± 7 Ma

Ngwane gneisses (Swaziland)

3490 ± 3 to 3644 ± 2 Ma

Lower Onverwacht groupca. 3500 Ma

Dwalile Suite greenstone remnantsCa. 3500 Ma ?

Swaziland

R.S.A

.

20 km

Crust accretion around BSB3450 Ma

Stolzburg, Theespruit, etc. plutons

3443 ± 4 to 3460 ± 5 Ma

Tsawela gneisses (Swaziland)

3458 ± 6 to 3437 ± 6 Ma

Upper Onverwacht groupca. 3400 Ma

Swaziland

R.S.A

.

20 km

Crust accretion around BSB3220 Ma

Kaap Valley, Neelshoogte, Badplaas, etc. plutons

ca. 3220 Ma

Usutu granodiorite (Swaziland)

3231 ± 4 to 3216 ± 3 Ma

Fig Tree and Moodies groupsca. 3200 Ma

Dalmein plutonCa 3220 Ma

Geochemistry:3600-3500 Ma

Steynsdorp plutonSteynsdorp pluton

Ngwane gneissesNgwane gneisses

Geochemistry:3450 Ma event

Stolzburg & Theespruit plutonsStolzburg & Theespruit plutons

Tsawela gneissesTsawela gneisses

Geochemistry:3220 Ma event

Kaap Valley, Nelshoogte Kaap Valley, Nelshoogte & Badplaas plutons& Badplaas plutons

TTG evolution around Barberton Greenstone Belt

3.6 – 3.4 Ga

3.4 – 3.2 Ga

Amphibolites with HP relicts

Preliminary conclusions (3)

• TTGs in Barberton record progressively deeper sources

• This is consistent with progressive steepening or onset of subduction, and could witness the progressive accretion of a continental nucleus and its early growth

• At 3.2 Ga (true subduction established), the geothermal gradient recorded in some metamorphic rocks is consistent with the gradient corresponding to TTG genesis

Secular/Geodynamical implications

Progressively cooler gradients ?

Early ArchaeanLate ArchaeanModern

Geodynamical implicationsSteepening/onset of subduction ?

Preliminary conclusions (4)

• Secular chemical evolution of TTGs reflects increasing melting depth and increasing interactions with the mantle

• This is consistent with a subduction origin for TTGs

• Secular cooling of the Earth makes the melting deeper and deeper along the subducted slab, allowing more and more interactions with the mantle

• Alternately, this could witness progressive onset of subduction

Conclusions

• TTGs are diverse, and their chemistry reflects the depth of melting; melting occurred mostly at 15-20 kbar, but can have occurred anywhere between 10-12 and 30 kbar.

• Most TTGs are probably originated in subductions, and interacted with the mantle to some degree

• The changes in TTG compositions can probably be correlated with changes in tectonic styles –either in terms of secular evolution, or in one single area

The Sand River GneissesCa. 3.1 Ga TTG gneisses in Messina area,Limpopo Belt, South Africa(R. White, Melbourne, for scale)