sediment melts at sub-arc depths: an experimental study

Sediment Melts at Sub-arc Depthsan Experimental Study

JOlaquo RG HERMANN1 AND CARLJ SPANDLER2

1RESEARCH SCHOOL OF EARTH SCIENCES THE AUSTRALIAN NATIONAL UNIVERSITY CANBERRA ACT 0200

AUSTRALIA2INSTITUTE OF GEOLOGICAL SCIENCES UNIVERSITY OF BERN CH-3012 SWITZERLAND

RECEIVED APRIL 30 2007 ACCEPTED OCTOBER 16 2007ADVANCE ACCESS PUBLICATION DECEMBER 4 2007

The phase and melting relations in subducted pelites have been

investigated experimentally at conditions relevant for slabs at

sub-arc depths (Tfrac14 600^10508C Pfrac14 25^45 GPa) The fluid-

present experiments produced a dominant paragenesis consisting

of garnet^phengite^clinopyroxene^coesite^kyanite that coexists

with a fluid phase at run conditions Garnet contains detectable

amounts of Na2O (up to 05 wt) P2O5 (up to 056 wt)

and TiO2 (up to 09 wt) in all experiments Phengite is stable

up to 10008C at 45 GPa and is characterized by high TiO2

contents of up to 2 wt The solidus has been determined at

7008C 25 GPa and is situated between 700 and 7508C at

35 GPa At 8008C 45 GPa glass was present in the experiments

indicating that at such conditions a hydrous melt is stable In

contrast at 7008C 35 and 45 GPa a solute-rich non-quenchable

aqueous fluid was presentThis indicates that the solidus is steeply

sloping in P^T space Fluid-present (vapour undersaturated) par-

tial melting of the pelites occurs according to a generalized reaction

phengitethorn omphacitethorn coesitethorn fluid frac14meltthorngarnet The H2O

content of the produced melt decreases with increasing temperature

The K2O content of the melt is buffered by phengite and increases

with increasing temperature from 25 to 10 wt whereas Na2O

decreases from 7 to 23 wt Hence the melt compositions

change from trondhjemitic to granitic with increasing temperature

The K2OH2O increases strongly as a function of temperature

and nature of the fluid phase It is 00004^0002 in the aqueous

fluid and then increases gradually from about 01at 750^8008C to

about 1at 10008C in the hydrous melt This provides evidence that

hydrous melts are needed for efficient extraction of K and other

large ion lithophile elements from subducted sediments Primitive

subduction-related magmas typically have K2OH2O of 01^04

indicating that hydrous melts rather than aqueous fluids are respon-

sible for large ion lithophile element transfer in subduction zones

and that top-slab temperatures at sub-arc depths are likely to be

700^9008C

KEY WORDS experimental petrology pelite subduction UHP meta-

morphism fluid LILE

I NTRODUCTIONSubducted sediments play a key role in the recycling ofincompatible elements in subduction zones The compari-son of primitive arc lavas with mid-ocean ridge basalts(MORB) provides evidence that arc lavas are enriched inH2O large-ion lithophile elements (LILE) and to a lesserextent light rare earth elements (LREE) (Tatsumi ampEggins1995)The enrichment of these components is inter-preted to be related to the addition of a lsquoslab componentrsquoto the locus of partial melting in the mantle wedge(Hawkesworth et al 1993) Detailed chemical analyses ofoceanic crustal material have revealed that subducted sedi-ments are the dominant host of LILE in the slab (Plank ampLangmuir 1998 Tenthorey amp Hermann 2004) The funda-mental role of sediments in the recycling of LILE has alsobeen demonstrated by the matching of trace element ratiosof sediment input and arc output in several subductionzones (eg Plank amp Langmuir 1993 Peate et al 1997)Therefore the systematic study of subducted sediments iscrucial for understanding the LILE transfer The pressurerange where this transfer occurs is constrained bythe depth (70^170 km) of the subducted slab below thevolcanic arc (Tatsumi 1986 Syracuse amp Abers 2006)Hence it is important to study phase and melting relations

Corresponding author Telephone thornthorn61 2 6125 8842E-mail joerghermannanueduau

The Author 2007 Published by Oxford University Press Allrights reserved For Permissions please e-mail journalspermissionsoxfordjournalsorg

JOURNALOFPETROLOGY VOLUME 49 NUMBER 4 PAGES 717^740 2008 doi101093petrologyegm073 by guest on O

ctober 30 2015httppetrologyoxfordjournalsorg

Dow

nloaded from

in subducted sediments that experienced pressures in therange of 25^5GPaOne approach to study such phase relations is through

investigation of ultrahigh-pressure (UHP) metamorphicrocks (Hermann et al 2006) Subducted continental crustoccasionally contains relics of coesite and diamond pro-viding evidence that these rocks were metamorphosedat 100^150 km depth (Chopin 1984 Sobolev amp Shatsky1990) Metapelites metagranites and metacarbonates insuch terrains can be used as natural laboratories to studyphase and melting relations in subducted sediments(Philippot et al 1995 Stolaquo ckhert et al 2001 Hermannet al 2006) Bulk-rock and mineral trace element analysesof such rocks have shown that phengite plays a key role as arepository for LILE (Sorenson et al 1997 Spandler et al2003) However one major problem with the interpretationof such data is that deeply subducted rocks cool relativelyslowly during exhumation and hence it is not clear whetherthe original chemical signature of phases is preservedMoreover in these rocks the record of the fluid phase isat best only fragmentarily preserved and so little infor-mation on the composition and nature of the fluid phasepresent at UHP conditions can be obtainedHigh-pressure experiments on sedimentary bulk compo-

sitions provide a second way of investigating elementrecycling in subduction zones Experimental studies areimportant for understanding mineral relations in sub-ducted sediments and for directly constraining the compo-sition and nature of the fluid phase Experimental studieshave confirmed that phengite is the stable K-phase insubducted sediments at sub-arc depths (Domanik ampHolloway 1996 Schmidt 1996 Ono 1998 Hermann2002b Auzanneau et al 2006) Consequently an under-standing of phengite melting relations is crucial forunderstanding the recycling of LILE One of the problemswith experimental studies is that piston cylinder techni-ques are used at pressures 35GPa whereas multi-anvilexperiments are used at pressures 35GPa (Schmidtet al 2004) This might introduce some systematic errorsexactly in the critical pressure range where the subduc-tion-zone fluids leave the slabThere are contrasting views on the nature and composi-

tion of the fluid phase that transports the key elementsfrom the slab to the mantle wedge Geochemical studiesof arc magmas have attempted to constrain the natureof the fluid phase on the basis of element ratios Twoend-members have been proposed corresponding to a sedi-ment fluid component and a sediment melt component(Class et al 2000 Elliott 2003) However it must bestressed that the actual phase relations and bufferingphases in the slab are often ignored A second approach toconstrain the nature and composition of the fluid phaseinvolves detailed solubility studies of minerals in fluidsand calculation of fluid compositions in equilibrium with

the determined high-pressure phases (Manning 19982004) Such an approach has provided excellent results formajor element concentrations in the aqueous fluid Toestablish whether an aqueous fluid or a hydrous melt isleaving the subducted slab it is necessary to combineexpected top-slab temperatures with experimental studieson the solidus of sedimentary rocks (Nichols et al 1994Hermann et al 2006) To date this approach has beeninconclusive because there exists a large variation in pre-dicted top-slab temperatures according to differentmodels as well as discrepancies in the position of the wetsolidus for typical subducted sediments A further compli-cation arises as a consequence of the potential presenceof the second critical endpoint (Kessel et al 2005b) In thepelitic system the termination of the wet solidus at thesecond critical endpoint might be reached at pressure andtemperature conditions relevant for subducted sedimentsat sub-arc depth and hence transitional supercriticalfluids could be the transporting agent (Manning 2004Kessel et al 2005a Hermann et al 2006)In this paper we present the results of an experimental

study on the phase and melting relations in a pelite compo-sition We used piston cylinder experiments for the entiredataset covering a P^T range of 25^45GPa and600^10508C appropriate for sub-arc conditions This elim-inates any potential systematic errors involved in usingtwo different experimental techniques We systematicallyinvestigated the stability and melting behaviour of phen-gite to constrain the recycling of K2O from subducted sedi-ments Fluid-present melting is explored and specialemphasis is placed on constraining the position of the wetsolidus We show that the compositions of aqueous fluidsand hydrous melts vary as a function of temperatureand we use this information to postulate that sedimentmelts provide the simplest way to explain the K2O andH2O systematics of primitive arc lavas We also reporthow the common UHP phases garnet omphacite andphengite change their composition as a function of pres-sure and temperature and how this information can beused as a guide to evaluate how well UHP rocks preserveoriginal phase compositions

EXPER IMENTAL METHODSStarting materialA starting material has been synthesized with a major-element composition similar to average global subductingsediment (GLOSS Plank amp Langmuir1998) and the aver-age upper continental crust (Fig 1 Taylor amp McLennan1985) The starting material was produced using a lsquosol^gelrsquomethod to eliminate problems of sluggish reaction ofrefractory minerals during experiments Most majorelements and some trace elements (Spandler et al 2007b)were combined as nitrate solutions and mixed withtetraethyl orthosilicate [Si(C2H5O)4] and slowly dried

JOURNAL OF PETROLOGY VOLUME 49 NUMBER 4 APRIL 2008

718

by guest on October 30 2015

httppetrologyoxfordjournalsorgD

ownloaded from

to a gel From this basic gel two starting materials wereprepared that differ only in their H2O content For theanhydrous starting material the gel was combined withappropriate amounts of Al2O3 TiO2 and MnO powdersfinely milled and then fused at 15008C and quenched toglass Multiple analyses of the glass for major elements(by energy-dispersive spectrometry EDS) and trace ele-ments (by laser ablation inductively coupled plasma massspectrometry LA-ICP-MS) indicate homogeneity withinanalytical precision Iron was then added to the glass inthe form of synthetic fayalite The fayalite and glass mix-ture was ground for 30min to form a fine powder For thehydrous starting material no Al2O3 was added prior tomelting Instead Al(OH)3 in a quantity sufficient to pro-duce the same amount of Al2O3 as in the anhydrous start-ing material was added together with fayalite at the endThis second starting material contains 68wt H2OBlending of the hydrous and anhydrous starting materialspermitted us to obtain precise but variable H2O contentsin the experiments which provide the basis for determina-tion of melt H2O contents using mass balance The watercontent of high-pressure melts increases with decreasingtemperature (Hermann 2003) To maintain 15^25 ofhydrous melt in the experiments which facilitates accurateanalysis it is therefore necessary to increase the amountof H2O in the starting material with decreasing tempera-ture (Tables 1 and 2) The final hydrous and anhydrousstarting material compositions are presented inTable 2

Experimental techniqueArc-welded gold capsules of 23mm outer diametercontaining the starting material were placed in anend-loaded 127 cm piston-cylinder apparatus using alsquolow-frictionrsquo and lsquodryrsquo furnace assembly assembled fromTeflon foil a salt outer sleeve Pyrex glass a graphite

heater and MgO inserts Some experiments were also runwithout the Pyrex glass to test if the glass introduces somefriction at low-T high-P conditions No systematic differ-ence was observed indicating that friction was fullyreleased during the long duration of the experimentsIn some experiments (labelled lsquoarsquo inTable 1) two gold cap-sules with different compositions (one carbonate-bearing)were run in one assembly Temperature was controlledusing type-B thermocouples (Pt94Rh6Pt70Rh30) and isaccurate to 108C Pressure is converted from load and isaccurate to 01GPa For the 45GPa experiments theassembly was initially held at 7008C 35GPa for 2 h toallow the vessels and assembly to compact In the 45GPaexperiments piston travel was monitored throughout theduration of the experiments Minimal piston travel after30^40 h indicates that friction has been released The useof the lsquodryrsquo furnace assembly gold capsules and the rela-tively low temperatures of the runs produces conditionsclose to the Ni^NiO transition (W O Hibberson personalcommunication) Hydrogen loss from the glass or alloyingof Fe with the gold capsule would potentially yield moreoxidized mineral assemblages containing acmite andandradite components in pyroxene and garnet respec-tively Normalization of the experimental run products(see the Electronic Appendix available for downloadingfrom httpwwwpetrologyoxfordjournalsorg) did notprovide evidence for any significant acmite and andraditecomponent which suggests that no significant hydrogenloss occurred

Diamond traps

For experiments with diamond traps we used 35mmdiameter gold capsules that were filled first with 18mg of40^60 mm diamonds as a fluid trap followed by 65mg ofthe hydrous starting material containing 68wt H2O

Al2O3

K2O MgO+FeO

SiO210

Na2O K2O

EPSM this studyGLOSSUpper continental crustPeliteGreywackeTonaliteRed Clay

Fig 1 Triangular projection of the experimental pelite starting material (EPSM) used in this study compared with global oceanic subductedsediment (GLOSS Plank amp Langmuir 1998) and upper continental crust (Taylor amp McLennan 1985) as well as starting materials from otherexperimental studies [pelite from Schmidt et al (2004) greywacke from Auzanneau et al (2006) tonalite from Patinlsaquo o Douce (2005) red clayfromJohnson amp Plank (1999)]

HERMANN AND SPANDLER MELTING OF SUBDUCTED PELITES

719



ownloaded from

The length of the fully loaded capsule was about 7mmthis was placed in the hotspot of a 30mm long MgO^graphite^Pyrex^NaCl assembly with the diamond traplocated close to the thermocouple Following each experi-ment the gold capsule was pierced and dried to precipitatethe remaining solutes in the fluid In the experimentsof Spandler et al (2007b) performed with an identical start-ing material the fluid was analysed for major and traceelements after the experiment The fluid contained503wt Na2O and 5005wt K2O It is expectedthat the vapour coexisting with such a fluid containedconsiderably less of these elements Hence the major andtrace element content of the escaping vapour is at leastan order of magnitude lower than that of the fluid phaseat the P andTconditions of the experiments and thereforeis considered to be negligibleThe capsule was then openedat the end containing the diamond trap and the diamondaggregate was extracted The diamond trap was eitherpowdery (Fig 2a) or cemented together (Fig 2b)

The diamonds and accompanying quench material weresubdivided into three fractions taken at different distancesfrom the residue (Tenthorey amp Hermann 2004) Fraction 1was sampled from the uppermost section of the diamondtrap fraction 2 from the intermediate part and fraction 3from the part of the diamond trap that was in contact withthe pelite residue A small sample of each fraction wasmounted on a glass slide and covered by immersion oilsOptical inspection of these samples with a polarized lightmicroscope allowed distinction between isotropic amor-phous quench material and anisotropic hydrothermalminerals that formed by solution precipitation reactionsduring the relatively long experimental runs Fraction 3always contained significant amounts of minerals whereasfraction 2 was dominated by amorphous quench and con-tained less than 5 vol of minerals Only in the experi-ment at 25GPa 6008C was the overall amount ofminerals higher in all fractions Interestingly fraction 1generally contained more precipitated minerals than

Table 1 Experimental conditions and run products

Run no P (GPa) T (8C) hours wt H2O Phases

C-2098D 25 600 168 68 Am Phe Qtz Rt All Ap

C-1866a 25 700 183 66 Am Phe Qtz Rt All Ap Gl

C-2099D 25 700 168 68 Am Phe Qtz Gl Rt All Ap

C-2246 25 750 168 68 Grt Am Qtz Phe Bt Gl Rt All Zir Ap

C-1846a 25 800 166 68 Grt Am Qtz Phe Bt Gl Ky Rt All Zir Ap

C-1928a 25 900 190 20 Grt Cpx Qtz Kfs Bt Gl Rt Ky Mon Zir

D-48 30 850 284 34 Grt Cpx Coe Phe Gl Ap Rt Mon

C-2082D 35 600 144 68 Ctd Cpx Coe Phe Q Ap Rt All graph

C-1857a 35 700 163 66 Grt Cpx Coe Phe Q Ky Rt All Ap

C-2083D 35 700 144 68 Grt Cpx Coe Phe Q Ap Rt All Zir

C-2247 35 750 168 68 Grt Cpx Coe Phe Gl Ap Rt All Mon Zir

C-1848a 35 800 168 68 Grt Cpx Coe Phe Gl Ky Ap Rt Mon Zir

C-1578a 35 900 164 61 Grt Cpx Coe Gl Phe Ky Ap Rt Mon Zir

C-1699a 35 900 260 31 Grt Cpx Coe Phe Gl Ky Ap Rt Mon

C-1867a 35 950 170 20 Grt Cpx Coe Phe Gl Ky Rt Ap Mon

C-2097D 45 700 168 68 Grt Cpx Coe Phe Q Ky Ap Rt All

C-2590 45 750 144 68 Grt Cpx Coe Phe Q Rt All Ap

C-2596 45 800 120 68 Grt Cpx Coe Phe Q Ap Rt Mon Zir


C-1614a 45 900 75 68 Grt Cpx Coe Phe Gl Ky Rt Mon Zir

C-1563a 45 1000 138 28 Grt Cpx Coe Gl Phe Ky Rt Mon Zir

C-1868a 45 1050 120 20 Grt Cpx Coe Gl Ky Rt Ap

Phases in bold have a modal abundance of45 phases in regular font have a modal abundance between 5 and 1 andphases in italics are accessory phases Experiments with suffix lsquoarsquo were run in assemblies containing two capsules suffixlsquoDrsquo indicates experiments with a diamond trap Am amphibole Phe phengite Qtz quartz Coe coesite Grt garnetBt biotite Gl glass graph graphite Kfs K-feldspar Ctd chloritoid Cpx clinopyroxene Ky kyanite Rt rutile Allallanite Ap apatite Mon monazite Q quench material from fluid Zir zircon

Experiments where no Pyrex sleeve was used


720



ownloaded from

fraction 2 indicating that the fluid in the diamond trapwas circulating through the pelite residue to some extentThis observation also shows that special care needs to betaken to obtain meaningful data from diamond trapsThe three fractions of the diamond trap were thenmounted on sticky carbon tape and the residue was sec-tioned longitudinally and mounted in epoxy Both the sec-tioned residue and the diamond trap material were thenexamined in detail with a JEOL 6400 scanning electronmicroscope (SEM) equipped with EDS The main miner-als precipitated in the trap were quartz and phengitealthough minor amounts of omphacite were also observedAll analytical results presented in this paper derive fromfraction 2 which contained nearly pure amorphousquench

Analytical techniquesThe phase relations were obtained from detailed studiesof backscattered electron (BSE) images Mineral composi-tions were mainly determined using a JEOL 6400 SEM

(Electron Microscopy Unit ANU) with EDS operatingat 15 kVand a focused beam of 1nA At such low beam cur-rent no alkali loss occurs in phengite (Hermann amp Green2001) Also because of the high resolution mixed analysescan be avoided Garnets were also analysed by wavelength-dispersive spectrometry (WDS) for major elements andin particular for accurate determination of the minorelements P Na Ti and Mn TheWDS analyses were con-ducted using a Cameca SX-100 electron microprobe at theRSES ANU operating at 15 kV and a beam current of20 nA The comparison of major element concentrationsbetween the SEM EDS and the electron microprobeWDS analyses showed excellent agreement Major-elementmelt compositions were also analysed by EDS by scanningglass pools over an area of 5 mm10 mm The scanningand the low beam current prevent destruction of the vola-tile-rich glass and permit accurate determination of thealkali elements (Hermann amp Green 2001)Material from the diamond trap mounted on carbon

tape was analysed for major and trace elements by

Table 2 Composition of the anhydrous and hydrous starting material (EPSM) and of experimentally produced glasses

normalized to 100 anhydrous

Exp EPSM EPSM 2246 1846 1928 D29 2247 1848 1578 1699 1867 1872 1614 1563 1868

P (GPa) 25 25 25 3 35 35 35 35 35 45 45 45 45

T (8C) 750 800 900 900 750 800 900 900 950 800 900 1000 1050

SiO2 6883 6415 7343 7459 7132 7272 7487 7377 7357 7318 7224 7272 7231 7066 6872

TiO2 067 063 021 017 025 031 027 031 035 044 043 046 055 060 103

Al2O3 1470 1370 1447 1490 1726 1532 1356 1378 1443 1485 1502 1232 1363 1521 1491

FeO 467 435 089 078 065 081 062 088 088 077 112 093 094 104 129

MnO 011 010 5005 5005 5005 5005 5005 5005 5005 5005 006 5005 008 5005 5005

MgO 251 234 027 050 052 055 027 043 040 029 023 031 061 035 040

CaO 245 228 117 159 094 125 110 142 119 106 084 186 162 079 081

Na2O 262 244 693 416 377 509 623 596 376 344 294 687 365 316 228

K2O 294 274 261 339 509 406 308 329 519 576 701 445 637 793 1005

P2O5 032 030 5020 5020 034 5020 5020 022 025 026 010 5020 029 026 049

Sum 998 9301 9997 10008 10014 10010 9999 10007 10001 10006 10000 9991 10005 10000 9998

Meas 8528 8476 9381 8679 7903 8088 8600 8561 8618 7032 8328 9096 8889

melt 35 60 40 25 23 30 55 20 26 21 40 35 30

H2O wt 68 16 10 5 10 25 20 11 11 6 29 16 8 6

K2OH2O 014 030 097 037 009 013 042 047 110 011 033 091 157

XMg 035 053 059 055 043 046 045 040 027 037 053 037 036

ASI 089 112 129 102 087 086 103 107 108 063 086 100 093

Ab 75 57 49 60 70 67 48 44 37 63 42 36 24

An 7 12 7 8 7 9 8 7 6 9 10 5 5

Or 18 31 44 32 23 24 44 48 58 27 48 59 71

The measured total (Meas) is given The melt fraction ( melt) in each experiment and the water content of the melt hasbeen estimated on the basis of mass balance XMg refers to the Mg(Mgthorn Fe) atomic ratio and ASI is the aluminasaturation index Al(NathornKthorn 2Ca) Percentages of albite (Ab) anorthite (An) and orthoclase (Or) components as used inFig 7 are also given


721



ownloaded from

LA-ICP-MS at the Research School of Earth Sciences(ANU) using a pulsed 193 nm ArF Excimer laser with100mJ output energy at a repetition rate of 5Hz (Egginset al 1998) coupled to an Agilent 7500 quadrupole ICP-MS system Laser sampling was performed in a HeÂratmosphere Data acquisition was performed by peak hop-ping in pulse counting mode acquiring intensity datafor each element during each mass spectrometer sweepThe sample stage was driven by a step motor allowingcontinuous analysis along scans through the mountedmaterial Four scans per fraction were run using a spotsize of 112 mm diameter with 25 sweeps of background fol-lowed by 80^100 sweeps of data collection A syntheticglass (NIST 612) was used as external standard referencevalues were taken from Pearce et al (1997) Experimentswere also doped with Cs in the hope of using this element

as an internal standard as previously done in ultramafic(Tenthorey amp Hermann 2004) and K-free mafic systems(Kessel et al 2005b) However the residual phengite inthe experiments retained the great majority of the Cs andthus this technique did not provide satisfactory resultsBecause no reliable internal standard is available for theaqueous fluids absolute solubilities could not be directlydetermined However element ratios can be accuratelydetermined with this method

RESULTSPhase relationsPhase relations were examined in detail by SEM (Fig 2)and results are summarized in Table 1 and Fig 3 In runswhere glass is present experimental mineral phases are

(a) (b) (c)

(d) (e) (f)

(g) (h) (i)

Fig 2 Backscattered electron and secondary electron image (g) of experimental run products For abbreviations seeTable 1 F quench productsfrom aqueous fluid D diamond M melt (hydrous glass) Scale bar in all images represents 10 mm (a) Powdery quench material in the diamondtraps (b) Diamonds from the trap are glued together by a hydrous melt indicating appearance of melt at 7008C 25GPa (c) Hydrous meltcoexisting with hydrous phases garnet and accessory minerals (d) Bubble-rich hydrous melt at 7508C 35GPa coexisting with small crystals ofclinopyroxene rutile monazite and zircon (e) Large melt pools and idiomorphic garnet at 8008C 35GPa (f) Phengite is still stable at 9508C35GPa and is in textural equilibrium with the main assemblage garnet^kyanite^coesite^garnet^melt (g) Idiomorphic garnet and fine-grainedphengite at 7008C 45GPa no glass (h) The first appearance of bubble-rich hydrous glass at 8008C 45GPa with the typical high-pressurepelite assemblage (i) Phengite is still present at 10008C 45GPa zircon and apatite have not melted out


722



ownloaded from

well crystallized and euhedral The mineral grain sizesare in the range 20^100 mm for garnet and some coesite10^20 mm for other major and minor phases and510 mmfor accessory phases In experiments where hydrousquench was present the phases had significantly smallergrain sizes and were more difficult to analyse

Solid phases

At 25GPa phase relations are complex with the phaseassemblage changing over small temperature intervalsThe assemblage is dominated by hydrous phases At6008C amphibole phengite and quartz are the majorphases and they coexist up to at least 8008C Biotite andgarnet appear at 7508C (Fig 2c) At 9008C K-feldsparand clinopyroxene appear at the expense of phengite andamphibole respectively whereas biotite is still stable(Fig 3) At the higher investigated pressures phase rela-tions are simpler The assemblage garnet clinopyroxenecoesite phengite and minor kyanite is stable over nearlythe entire range of investigated P^T conditions (Fig 2f hand i) Only at 7008C 35GPa is chloritoid present insteadof garnet Phengite is stable up to 9508C at 35GPa(Fig 2f) and breaks down between 10008C and 10508C at45GPa (Fig 2i) Kyanite is a minor phase in almost allexperimentsRutile is the stable accessory Ti phase over the entire

investigated P^T range Apatite is also found in all experi-mental charges Allanite is present at 25GPa up to 8008C(Fig 2c) and at 35 and 45GPa up to 7508C At higher

temperatures monazite is stable up to 10008C Only at thehighest conditions of 45GPa 10508C was monazite notfound Zircon grains are very small and difficult to detectbut zircon has been observed in nearly all experimentsIt is definitely present up to 9008C at 25 and 35GPa andup to 10008C 45GPa (Fig 2i)

Fluid phases

The water added to the experiments always exceeded thecapacity of the hydrous phases to store H2O resultingin the presence of a fluid phase in all experiments At thehigher investigated temperatures this fluid phase quenchedto a hydrous glass which indicates that a hydrous meltwas present The added water content was always lowerthan that required to obtain vapour-saturated melting forthe given pressure range (Huang amp Wyllie 1981) and thusthe experiments constrain fluid-present vapour undersatu-rated melting in subducted sediments No mineral quenchphases have been observed within experimentally pro-duced glassAt 25GPa clean bubble-free glass was observed down to

temperatures of 7508C (Fig 2c) In experiment C-1866 at7008C 25GPa no glass pools were observed Howeverthe experimental phases were well cemented together aswas generally observed in charges where glass is presentIn a repeat experiment a diamond trap was added to thestarting material (C-2099D) to facilitate separation offluid from solid phases After opening the capsule the dia-mond trap could be extracted as a coherent blockInspection of the diamond trap under the SEM revealedthat the diamonds were cemented together by a hydrousglass (Fig 2b) In contrast at 6008C 25GPa the diamondtrap after the experiment was powdery and appeared verydifferent under the SEM (Fig 2a) A very fine hydrousquench powder which is not stable under the electronbeam was found These observations indicate that at6008C a solute-rich aqueous fluid was present whereas at7008C there was a hydrous meltAt 35GPa hydrous glass is present in experiments

down to temperatures of 7508C (Fig 2d) At 900 and9508C the glass is bubble free whereas at 800 and 7508Cbubbles are abundant in the glass (Fig 2e) At 7008C(C-1857) fine-grained minerals were observed the grainswere loose and significant porosity was present in thecharge indicative of an aqueous fluid during the runTo further test this hypothesis a diamond trap experimentwas run at the same conditions (C-2083D) The diamondtrap showed the same characteristics as observed at6008C 25GPa indicating the presence of an aqueousfluid At 6008C 35GPa the diamond trap again containedfine-grained amorphous hydrous quench materialAt 45GPa a hydrous glass was observed from 10508C

down to 8008C The glass in the 1050 and 10008C runs isbubble free A small amount of bubbles appears in the9008C experiment and a significant amount of bubbles

35

45

P (

GP

a)

600 700 800 900 1000T (degC)

25

Grt Cpx CoePhe Gl

Grt CpxCoe GlGrt Cpx Coe

Phe F

Ctd Cpx CoePhe F

Am QtzPhe F

Am Qtz BtGrt Phe Gl Grt Cpx Bt

Qtz Kfsp Gl

non

quen

chab

le fl

uid

hydr

ous

glas

s

CoeQtz

Phe

ngite

Garnet

Amphibole

Biotite

Fig 3 P^Tstability fields of different parageneses of the main miner-als in subducted pelites It should be noted that water contentsvary between experiments (Table 1) The approximate stability limitsof key phases are indicated Triangles refer to runs where hydrousquench has been found Glass and phengite are present in experimentsshown by squares Phengite has disappeared in experiments shownas circles At sub-arc pressures the main solid assemblage in pelitesconsists of phengite^clinopyroxene^garnet^coesite and minor kyanite


723



ownloaded from

is present in the 8008C experiment (Fig 2h) This suggeststhat some aqueous fluid exsolved during quenching from amelt that contained high water contents In a secondexperiment at 8008C (C-2596) no glass pools were foundThis might be related to the fact that the first experimentwas run with two capsules which produced a slight hot-spot in which the melt pooled At 7508C the run productswere fine-grained and the charge was not well cementedA Si-rich hydrous quench was observed in pores betweenmineral phases The experiment at 7008C contained adiamond trap and the residual minerals were generallyfine grained and not well cemented (Fig 2g) Examinationof the trap after the experiment revealed the presence of apowdery amorphous quench material similar to runs2099D and C-2083D From all these observations we sug-gest that the transition from quenchable hydrous glass tonon-quenchable solute-rich fluid occurs at 8008C at45GPa

Phase compositionsThe average composition of all analysed phases in theexperiments is given in the Electronic Appendix (httpwwwpetrologyoxfordjournalsorg)

Garnet

Large idomorphic garnet is generally slightly zonedThe cores are characterized by high TiO2 contents (often1wt) and by slightly higher MnO and CaO contentswhereas the MgO content is generally lower than inthe rimThe following discussion is based on rim composi-tions which are interpreted to be in textural equilibriumwith the other phasesThemost prominent feature in garnet

is the increase of pyrope content with increasing tempera-ture (Fig 4a) Grossular contents decrease slightly withincreasing pressure The cation sum of garnet whennormalized with all iron as ferrous iron is within errorof 80 indicating that there are insignificant amounts offerric iron There is no evidence for supersilicic garnet[Si43 cations per formula unit (pfu)] in any of the experi-ments An interesting feature of the garnet compositions isthe relatively high amounts of minor elements such asTiO2P2O5 and Na2O elements that are buffered in theexperiments by coexisting rutile apatite and omphaciterespectively In all analysed garnets theTiO2 content variesbetween 04 and 09wt Generally the TiO2

content increases with increasing temperature P2O5 con-tents range from 02 to 056wt and increase with increas-ing pressure Na2O was detected in all garnets and rangesfrom 01to 05wt Na2O increases with pressure and tem-perature Figure 4b displays the variation of Na vs P cations(pfu) contouredwith the experimental conditions At 700^8008C Na and P (pfu) values plot on a 11line indicatingthat at these conditions these elements enter garnet via acoupled substitution NaPM2thorn

1Si1 At higher temperaturesNa (pfu) is much higher than P (pfu) and thus Namust be incorporated by other substitutions such asNaTifrac12oM2thorn

1Al1 Some of theTi might also be accommo-datedby the coupled substitutionTifrac12oAlfrac12TAlfrac12o1Si1

Clinopyroxene

Clinopyroxene generally forms small grains that aredifficult to analyse It often exhibits zoning in FeO withhigher contents in the core than in the rim Because of thesmall grain size it has not always been possible to analyse

Gro

Alm Py

60 408010

10

20

30

10

30

30

50

50

700 800900

1000

0

002

004

006

008

01

Na

(pfu

)

0 001 002 003 004 005P (pfu)

253545

700ndash800

800ndash900

900ndash1000

Garnet

(a) (b)

Fig 4 (a) Garnet compositions plotted in molecular proportions of grossular (Gro) almandine (Alm) and pyrope (Py) show a strong increasein pyrope content with increasing temperature (numbers in8C) There is a slight decrease in grossular (Gro) content with increasing pressure(œ 25 and 3GPa s 35GPa 45GPa) (b) Variation of Na and P (in cations per formula unit) in garnet from all experiments At lowtemperature (700^8008C) the data plot on a 11line At higher temperatures there is more Na than P Error bars refer to 1s standard deviation ofmultiple analyses


724



ownloaded from

the rim only The average analyses presented in theElectronic Appendix are based preferentially on rimcompositions Clinopyroxene is omphacitic in compositionin all experiments The jadeite component stronglyincreases with increasing pressure and reaches a maximumof 74 at 7008C 45GPa (Fig 5a) At 35GPa the jadeitecontent is nearly constant over the entire investigatedtemperature range In contrast at 45GPa the jadeite con-tent decreases significantly from 700 to 10508C (Fig 5a)TiO2 content ranges from 01 to 06wt and generallyincreases with increasing temperature and decreases withincreasing pressure In some omphacite grains K2O wasdetected when analysed by WDS however the ratherlarge variation indicated that the analyses were affectedby contamination from neighbouring phengite In factwhen omphacite was analysed by SEM using a finer beamand a lower beam current K2O was always below thedetection limit of 02wt All omphacite analyses exhi-bit Al4Na (pfu) Therefore no ferric iron is requiredto charge balance Na The excess Al which is notbound in a jadeite (NaAlSi2O6) or a Ca-Tschermakrsquos(CaAl2SiO6) component provides evidence for the pre-sence of a Ca-eskolaite (Ca05œ05AlSi2O6) componentin the omphacite Because Ca-eskolaite contains a vacancyit is expected that the sum of the cations should decreasewith increasing Ca-eskolaite component Figure 5b dis-plays the correlation between excess Al and the sum ofcations in relation to the theoretically expected line Thegood agreement provides evidence that the experimentallyproduced omphacite contained up to 15 eskaloite compo-nent The Ca-eskaloite component increases mainly asfunction of increasing temperature and does not changesystematically with pressure This is in good agreement

with a recent comprehensive experimental study ofKonzett et al (2007) who reported that the Ca-eskolaitecomponent in the assemblage clinopyroxene^garnet^coesite^rutile increases with increasing temperature butthat no significant increase with pressure can be observed

Phengite

Phengite grains are relatively large and easy to analyseWe preferred to analyse phengite by EDS using a lowbeam current and where possible a rastered beamUnder such analytical conditions no K loss during theanalysis occurred Phengite displays a systematic changein the celadonite component [KAlMgSi4O10(OH)2] withchanging conditions The Si content (pfu) increases withincreasing pressure and decreasing temperature (Fig 6a)An interesting feature is the systematic change in TiO2

content The highest TiO2 content of 2 wt has been mea-sured at 10008C 45GPa Overall TiO2 increases stronglywith increasing temperature and decreases slightly withincreasing pressure (Fig 6b) Na2O was detected in allphengites and the paragonite component reaches 19at the lowest pressure investigated At higher pressurethe paragonite component is typically between 4 and 8All phengites contain MgthornFe4Si 3 indicating thatthere is a biotite component present as previouslyobserved by Green amp Hellman (1982) This biotite compo-nent is highest when phengite coexists with biotite (ie at25GPa) and at this pressure it increases with increasingtemperature to reach more than 30 at 8008C 25GPaIncorporation of Ti presumably on the octahedral site ofphengite could either be charge balanced by substitutionof Al for Si on the tetrahedral site or by a coupled substi-tution M2thornTiAl2 on the octahedral site The first

04

05

06

07

08

Na

(pfu

)

500 600 700 800 900 1000 1100

T (degC)

45

35

25

3

0

005

01

015

02

Exc

ess

Al (

pfu

)

392 394 396 398 4high T

low T

(a) (b)

Clinopyroxene

Fig 5 (a) Na per formula unit in clinopyroxene as a function of temperature and pressure (b) Excess Al (Al neither bound in a jadeite norCa-tschermakite component) vs cation sum indicates the presence of a Ca-eskaloite component in clinopyroxene Error bars refer to 1s standarddeviation of multiple analyses Symbols as in Fig 4


725



ownloaded from

substitution produces phengite with a higher Al contentwhereas the second substitution produces phengite withan Al content that is lower than would be expected forthe celadonite and biotite substitution alone We observean Al deficit in all Ti-rich phengites which indicatesthat Ti is dominantly incorporated by the M2thornTiAl2

substitution

Other minerals

Only three runs at 25GPa contained biotite The TiO2

content of biotite increases from 2wt at 7508C to4wt at 9008C whereas the FeO content significantlydecreases over the same temperature interval All biotitescontain about 05wt Na2O and are characterizedby high Si content (pfu) of 30 Four runs at 25GPacontained sodic^calcic amphibole With increasing tem-perature from 600 to 8008C the Na2O SiO2 and FeO con-tents decrease whereas theTiO2 K2O and Al2O3 contentsincrease The NaM4 (when normalized to 23 oxygens)decreases from 098 pfu (6008C) to 056 pfu (8008C)the NaA is approximately constant at 033 pfu whereasthe Al[T] increases from 1 to 15 pfu Only one run at9008C 25GPa contained K-feldspar with the compositionOr58Ab39An3 The K-feldspar also contains detectableamounts of FeO (025wt)

Hydrous glass (quenched melt compositions)

Table 2 presents the glass analyses normalized to 100on an anhydrous basis whereas the Electronic Appendixcontains the originally measured element concentrationdata All normalized melt compositions are broadly gra-nitic with SiO2 contents between 69 and 75wt andAl2O3 contents from 12 to 17wt K2O and Na2O vary

strongly as a function of temperature Na2O decreasesfrom about 7wt to 23wt from 750 to 10508C K2O isbuffered in the melt by coexisting phengite up to tempera-tures of 10008C at 45GPa and increases from 26wt at7508C to 10wt at 10508C (Fig 7a) Consequentlythe K2ONa2O ratio varies over about an order of magni-tude (Fig 7b) Melt compositions have low anorthitecomponents and trend from trondhjemitic to granitic com-positions with increasing temperature (Fig 7c) Low-temperature melts are typically peralkaline (Table 2)whereas only one melt composition determined at 9008C25GPa is clearly peraluminous The low totals of theglass provide evidence that there is a significant amountof H2O dissolved in the melt in agreement with the occur-rence of bubbles in the glass Mass-balance calculationsusing the known starting material composition and thecompositions of all phases in the runs (ElectronicAppendix) allowed determination of the amount of meltin each experiment (Table 2) From the known amount ofwater in the starting material of each experiment (Table 1)and the amount of OH stored in hydrous phases (deter-mined by mass balance) the H2O content of the meltcould also be calculated The principal source of errorin this approach derives from the amount of melt deter-mined from mass balance and this error can be significantespecially at very low melt fractions Our experimentswere designed to produce relatively high amounts (420)of melt in each experiment which results in estimateduncertainties in the H2O content of the melt of 10^15relative at the lower melt fractions and of 5^10 relativeat the higher melt fractions Figure 7d shows the H2Ocontent of the melt as a function of pressure and tempera-ture The resulting H2O isopleths are characterized by

31

32

33

34

35

36

500 600 700 800 900 1000 1100

45

35

25

(a) (b)

Si (

pfu

)

T (degC)500 600 700 800 900 1000 1100

T (degC)

059

051

063

066

061

042

096

072

042

145

089

054

068

187

176

177

107

191

199

25

35

45

P (

GP

a)

Phe

out

05 075 10 15 20TiO2

Phengite

Fig 6 (a) Si content of phengite as a function of pressure and temperature Error bars refer to 1s standard deviation of multiple analyses (b)Contours of TiO2 content (wt) in phengite as a function of P^TgridThe high content of TiO2 close to the phengite breakdown curve shouldbe noted


726



ownloaded from

2

4

6

8

10

SiO

2A

l 2O

3

500 600 700 800 900 1000 1100

T (degC)

500 600 700 800 900 1000 1100

T (degC)

800degC 45 GPa

0

05

1

15

Na 2

OA

l 2O

3

aqueous fluid

hydrous melt

aqueous fluid

hydrous melt

01

1

10

K2O

Na 2

O

0

25

5

75

10

125

K2O

500 600 700 800 900 1000 1100

T (degC)

500 600 700 800 900 1000 1100

T (degC)

An

Ab Or

tonalite

trondhjemite

granodiorite

granite

(95)

(92)

20

(89)

(85)

16

25

10

20

29

5

10

1111

16

6

8 6

25

35

45

P (

GP

a)

600 700 800 900 1000 1100

T (degC)

51015202530

(a) (b)

(c)(d)

(f)(e)

25 3 GPa35 GPa45 GPadiamond traps

aqueous fluid

hydrous melt

wet

sol

idus

H2O

Fig 7 (a) The K2O content of melt is buffered by phengite in most experiments and increases systematically with increasing temperaturePressure has no significant influence on the K2O of the melt (b) The ratio of K2O to Na2O increases over an order of magnitude from 600 to10508C The significant difference between hydrous melts and aqueous fluids should be noted (c) The aqueous fluid and the low-temperaturemelts are albite-rich whereas the high-temperature melts are orthoclase-rich All fluids are poor in anorthite component Classification of silicicmelts after Barker (1979) (d) Estimated water contents of hydrous melts in the experimental runs as a function of P andTand schematic con-tours of wt H2O in the melt The numbers in parentheses represent rough estimates of the wt H2O of the aqueous fluid The wet solidus isdefined up to 35GPa Aqueous fluid and hydrous melt display distinctly different SiO2Al2O3 (e) and Na2OAl2O3 ratios (f)


727



ownloaded from

a modest positive slope of 88Ckbar The highest amountof 30wt H2O was determined for the experiment at8008C 45GPa Fluids with higher amount of H2O werenot quenchable

Diamond traps

To constrain the composition of the non-quenchable fluidphase some experiments were run with diamond trapsFraction 2 of the diamond traps contained more than95 precipitates formed from the fluid during quenchingTherefore this fraction was analysed by LA-ICP-MS toprovide major element ratios in the quench materialThe total sum of the major elements was then normalizedto 100 (Table 3) Experiment 2099 provides a good testfor the applied method Detailed textural investigationsshowed that this trap contains hydrous melt rather thanan aqueous fluid (Fig 2b) The normalized values displayall the characteristics of hydrous melts similar to the com-positions reported in Table 2 and the data obtainedplot along the trends established from where the meltpools have been directly measured The quench materialsin the other four low-temperature experiments have a

significantly different major element composition from thehydrous melts and are interpreted to represent quenchfrom aqueous fluidsWhereas the SiO2Al2O3 value variesin a very limited range (4^6) in the hydrous melt theaqueous fluids are characterized by much higher values(7^34 Fig 7e) The aqueous fluid is also characterizedby very high Na2O and low K2O contents resulting inK2ONa2O values that are much lower than those of thehydrous melts but still consistent with the general tem-perature trend of this ratio (Fig 7b) The relatively highNa2O in conjunction with low Al2O3 makes the aqueousfluids strongly peralkaline Therefore Na2OAl2O3 isanother effective ratio for distinguishing between aqueousfluid and hydrous melt compositions (Fig 7f) Similar ele-ment ratios were reported for hydrous fluids trappedas synthetic fluid inclusions equilibrated at 22GPa and600^6508C by Spandler et al (2007b) those results furthersupport the validity of using these chemical criteria for dis-tinguishing aqueous fluids from hydrous melt

Element partitioningThe compositions of solid and fluid phases (ElectronicAppendix) permit determination of the partitioning ofmajor and minor elements among the phases presentOmphacite always contains higher Na than coexistinghydrous melts at all conditions investigated The DCpx=Melt

Naincreases with pressure as well as with temperature from15 at 7508C 35GPa to 45 at 10508C 45GPa (Fig 8a)Interestingly inthetwoexperimentswhereamphibole ispre-sent DAm=Melt

Na is below unityThe general affinity for Na2Oincorporation is omphacite4melt4amphibole4phengite4biotite4garnet Because Na in garnet increases and Na inphengite decreases with increasing pressure and tempera-ture DPhe=Grt

Na is close to unity at 10008C 45GPa Ti is aminor element that can be detected in all major phasesThegeneral affinity for TiO2 incorporation is biotite4phengite4amphibole4garnet4melt omphacite The FeandMgpartitioningof coexistingminerals isbest illustratedusing XMgfrac14Mg(MgthornFe) which decreases in the follow-ingway biotite4phengite4omphacite4amphibole4melt4garnet Amphibole andomphacitehave notbeen observedin the same run and the relative position is inferred fromphengiteômphacite^melt and from phengiteâmphibole^melt partitioning An interesting observation concernsthe XMg in coexisting garnet and melt Generally the XMg

in the melt is higher than in garnet However at 9508C35GPa and at 10008C10508C at 45GPa XMg of garnet ishigher than in the melt Such an inversion in the XMg isin excellent agreement with previous studies Rubatto ampHermann (2007) found in an experimental study at 2GPathat garnet has a lower XMg than the melt up to 9508Cat 9508C it is identical and at 10008C garnet has a higherXMg than the melt Moreover previous experimentsby Green (1977) and Ellis (1986) showed that melts haveahigherXMg thangarnetbelow10008Cat 30GPa

Table 3 Composition of quench material in fraction 2 of the

diamond traps determined by LA-ICP-MS

Exp 2099 2098 2082 2083 2097

P (GPa) 25 25 35 35 45

T (8C) 700 600 600 700 700

SiO2 7247 7538 6902 7505 8847

TiO2 024 032 042 027 022

Al2O3 1478 1021 880 774 261

FeO 183 187 293 309 027

MnO 004 008 014 026 001

MgO 060 247 300 180 393

CaO 184 097 168 183 061

Na2O 578 786 1255 848 320

K2O 209 083 146 147 069

Sum 100 100 100 100 100

H2O 20 95 92 89 85

K2OH2O 0084 00004 00013 00018 00012

XMg 037 070 065 051 096

ASI 098 065 035 041 037

Alb 71 88 87 81 80

An 12 6 6 10 8

Or 17 6 7 9 11

Analyses are normalized to 100 anhydrous The watercontent of experiment 2099 has been determined by massbalance Other water contents are rough estimates basedon published total solubilities in similar systems (see text fordiscussion) Other entities are as defined in Table 2


728



ownloaded from

It has long been known that the KD of the Fe^Mgexchange in minerals is very sensitive to temperature(Ellis amp Green 1979 Green amp Hellman 1982) Figure 8bshows the variation of the Fe^Mg partitioning betweengarnet and phengite as a function of inverse temperatureat 45GPa Data from lower pressure experiments havelarger errors so for simplicity these data are not plottedin the figure Three important observations can be madefrom examination of the plot (1) there is a clear trend inln KD with inverse temperature (2) this trend is in verygood agreement with the formulated thermometer ofGreen amp Hellman (1982) (3) the propagated errors of thevariability in analyses are rather large Although there isalso a general trend in garnet^clinopyroxene partitioningwith inverse temperature the large error in determiningthe FeO content in clinopyroxene because of zoning pre-vents a sensible test of garnet^clinopyroxene thermometerswith the existing dataset

DISCUSSIONThe results of this experimental study provide informa-tion on two main aspects of research on the behaviourof deeply subducted sediments The first part of the discus-sion addresses issues related to the use of ultrahigh-pressure(UHP) rocks as a natural laboratory for subductionprocesses The second part focuses on the phase andmelting relations in subducted sediments and how thenature and composition of the fluid phase determinesthe way these elements are recycled in the subductionfactory

UHP indicators in mineral compositionsIt is not always easy to distinguish between minerals thatformed at UHP conditions and minerals that formedor equilibrated during retrograde evolution Most UHProcks occur within subducted continental crust Hencedetailed phase compositions in a pelitic starting materialare useful to evaluate whether or not there are clear com-positional signatures that can indicate a UHP origin ofcrustal rocks Generally UHP conditions are deducedfrom the presence of coesite andor diamond In this sec-tion we compare the major and minor element composi-tion of the main phases garnet and phengite in theexperiments with data from UHP rocks as well as fromdeep-seated xenoliths to provide some constraints on possi-ble UHP indicators in rock-forming minerals

Garnet

The experimental garnets are characterized by high NaTi and P At temperatures of 700^8008C the exchangeNaPM2thorn

1Si1 seems dominant for the incorporation of Pand NaThis exchange has been experimentally confirmedand a complete solid solution series has been documentedat very high pressures (15^17GPa) and high temperatures(Brunet et al 2006) In our experiments at temperatureshigher than 8008C a second substitution linked to Ti suchas NaTifrac12oM2thorn

1Al1 is proposed to accommodate Na inexcess of P This substitution was first documented experi-mentally by Ringwood amp Major (1971) at 5GPa Thoseworkers also produced experimental garnets at 35GPaand 12008C with significant amounts of Na and TiOur results are in agreement with the conclusion of

05

1

15

2

25

3

Ln

KD

(G

rt-P

he)

75 8 85 9 95 10 1051T (degK) x 10000

45 GPa1

2

3

4

5

Na 2

O C

px

Mel

t

700 800 900 1000 1100T (˚C)

25 3 GPa35 GPa45 GPa

T (degC)

7008009001000

T (degC)

GH82

(a) (b) (c)

Garnet-Phengite thermometry

Garnet-Phengite-Clinopyroxene barometry

minus30

minus20

minus10

0

10

20

30

(Pca

lc -

Pex

p)

Pex

p x

100

600 700 800 900 1000 1100

Fig 8 (a) Na2O (wt) partitioning between clinopyroxene and melt as a function of pressure and temperature It should be noted thatclinopyroxene always contains higher Na2O than the coexisting melt (b) Fe^Mg partitioning between garnet and phengite at 45GPaKDfrac14 (MgpheFephe)(MggrtFegrt) The continuous line represents the formulated thermometer of Green amp Hellman (1982) for a peliticcomposition with high XMg Error bars in (a) and (b) refer to propagated errors on the ratio from 1s standard deviation derived from multipleanalyses of the minerals (c) Comparison between run pressures with calculated pressures at the given run temperature using the garnet^phen-gite^clinopyroxene barometer formulated by Krogh-Ravna amp Terry (2004) The grey band refers to the estimated calibration error of thebarometer There is a very good match between calculated and nominal pressures in the temperature range 800^9008C At lower temperaturesthe calculated pressures are lower and at higher temperatures the calculated pressures are higher than the nominal run pressures


729



ownloaded from

Ringwood amp Major (1971) that in the presence of TiNa can enter garnet at relatively modest pressuresPrevious experiments in the pressure range of 20^35GPain natural compositions suggested that Na P and Ti areincorporated into garnet Thompson (1975) conductedanhydrous melting experiments in basaltic compositionsand found up to 04wt Na2O 06wt P2O5 and up to3wt of TiO2 in garnet Auzanneau et al (2006) investi-gated phase relations in a hydrous metagreywacke andreported garnet compositions with significant amounts ofTiO2 (1wt) and 01^03wt of Na2O in runs at22^24GPa 800^9008C In experiments on a naturaltonalite composition Patinlsaquo o Douce (2005) found garnetswith Na2O 01wt and TiO241wt even at 15GPa9508CGarnet grains in UHP rocks occasionally display

exsolution of Ti- Na- and P-rich phases indicating thatthere was a P- andTi-rich garnet precursorYe et al (2000)found such garnets in the Sulu UHP belts of China andMposkos amp Kostopoulos (2001) described similar garnetfrom the Rhodope UHP belt in Greece By integratingthe inclusions in garnet Ye et al (2000) obtained 034wtNa2O 028wt P2O5 and 107wt TiO2 and Mposkosamp Kostopoulos (2001) obtained similar values of 06wtNa2O 033wt P2O5 and 103wt TiO2 These valueswere used by both groups to postulate that the rocksoriginated from more than 200 km depth However theconcentrations of these elements are very similar to whatwe find in garnet at 35^45GPa (100^150 km equivalentdepth) indicating that such compositions are not unequi-vocal proof of an ultra-deep origin of subducted crustAccording to our experiments and literature data

garnets with highTi Na and P should be the rule and notthe exception in UHP terrains However it is interestingto note that most reported garnet analyses from UHProcks display rather low concentrations of these minorelements For example garnet from diamond-bearingmeta-granitic to meta-pelitic rocks from the Kokchetavmassif and the Erzgebirge generally has TiO25025wtand Na2O501wt (Zhang et al 1997 Massonne 2003Massonne amp Nasdala 2003) Only in some rare garnetsfrom eclogites did Zhang et al (1997) find some elevatedTiO2 contents of 034wt and Na2O of 014wtHowever all these values are significantly lower thanwhat is expected based on our experiments given the dia-mond-facies metamorphic conditions of these rocks(Tfrac14 950^10008C P45GPa) This inconsistencybetween garnet compositions from experiments and nat-ural UHP garnet compositions provides evidence thatelement re-equilibration occurred during exhumationSuch re-equilibration for P and Ti has been documentedin UHP mafic rocks from the Northern Dabie Complex(Malaspina et al 2005) It is expected that garnet grainsin crustal xenoliths are less affected by such retrograde

modification Although available literature data arerestricted to mafic UHP xenoliths in which the bufferingof P andTi is not always clear there is convincing evidencethat garnet from such rocks contains much higher NaP and Ti than is generally observed in UHP rocks of simi-lar composition and metamorphic grade For exampleSobolev amp Lavrentrsquoev (1971) observed that Na2O contentsof 01^022wt are common in eclogitic garnet fromdiamond-bearing rocks whereas 001^002wt Na2O isgenerally present in garnet from UHP eclogite-faciesrocks Bishop et al (1978) reported garnet compositionsfrom eclogite xenoliths with up to 01wt P2O5 014wtNa2O and 08 wt TiO2In summary there is clear evidence that garnet formed

in subducted metapelitic rocks contains significantamounts of Na2O P2O5 and TiO2 The much lowercontent of these elements in garnet from UHP felsic rocksindicates that in most cases garnet is not able to retain itsoriginal UHP chemical signature in such slowly coolingrocks Garnet is one of the most important phases forthermobarometry Therefore the application of thermoba-rometers using garnet in relatively slowly cooled UHProcks should be done with care as in some cases these ther-mobarometers may be unable to provide meaningful con-straints on peak metamorphic conditions

Phengite

The Si content of phengite increases strongly with increas-ing pressure but depends also on the buffering mineralparagenesis as has been demonstrated in severalearlier studies (Massonne amp Schreyer 1987 Massonne ampSzpurzka 1997 Hermann 2002b) The Si content of phen-gite must therefore be used in conjunction with the compo-sitions of coexisting phases to estimate pressure The mostinteresting observation on experimental phengite is thesystematic change in TiO2 content when buffered withrutile and quartz^coesite a feature that has not beendescribed before The TiO2 content increases stronglywith increasing temperature and decreases moderatelywith increasing pressure (Fig 6b) and provides a usefulsemi-quantitative tool to evaluate whether or not phengitecompositions record peak conditions or equilibratedduring exhumation For example Meyre et al (1999)reported phengite containing 05^06wt TiO2 fromhigh-pressure pelites that experienced peak conditions ofT 6508C P 25GPa This is in very good agreementwith the experimental values of 059 wt TiO2 (6008C25GPa) and 063wt TiO2 (7008C 25GPa) Phengitein metapelites of the Dora^Maira massif (peak conditionsT 7308C P 4GPa) have 04wt TiO2 (Hermann2003) in good agreement with the measured 042wtTiO2 in the experiment at 7508C 45GPa Phengite-bearing eclogites of the Erzgebirge equilibrated atsignificantly higher temperatures of 8508C with lessprecise pressure estimates of 35GPa (Schmalaquo dicke amp


730



ownloaded from

Mulaquo ller 2000) These phengites contain 11^15wt TiO2which again is consistent with our experimental results(Fig 6b) Even higher TiO2 contents in phengite of 175^204wt have been described from UHP metapelites ofthe Greenland Caledonides for which peak metamorphicconditions of 9708C and 36GPa have been proposed(Lang amp Gilotti 2007) This is in very good agreementwith theTiO2 content of 19wt determined in an experi-ment at 9508C 35GPa Because phengite breaks down atsignificantly lower temperatures of 700^7508C at crustalpressure (Vielzeuf amp Holloway 1988) than at UHP condi-tions (Fig 3) phengite with such very high TiO2 contentsmight be used as an indicator for UHP conditions Forphengites that do not contain a major biotite component(520) TiO2 contents of42wt seem to be stable onlyat pressures43GPa (ie at UHP conditions) In fact suchphengite has been observed as inclusions in zircon from thediamond-bearing gneisses of the Kokchetav massif(27wt TiO2 Hermann et al 2001) and as inclusionsin garnet from the Erzgebirge (33wt TiO2 Massonneamp Nasdala 2003) These first results are promisinghowever clearly more experimental data are needed tofully understand and quantify the TiO2 incorporation inphengite as a function of pressure temperature bulk-rockcomposition and buffer paragenesis

Implications for garnet^clinopyroxene^phengite barometryAlthough the composition of single minerals as discussedabove can give some insights into the conditions of poten-tial UHP metamorphism the most complete approachis to obtain pressure information by combining thecomposition of garnet clinopyroxene and phengite Thecompositions of these three phases are related throughthe fluid-absent pressure-sensitive equilibrium

6 diopsidecpx thorn 3 muscovitephe

frac14 2 grossulargrt thorn 1 pyropegrt thorn 3 celadonitepheeth1THORN

The large stability field of garnet^phengite^clinopyroxenein a great variety of subducted crustal rocks makesthis assemblage ideal for determination of pressure inUHP terrains (Carswell et al 1997 Krogh Ravna ampTerry 2004) In the following discussion we compare thecalculated pressure at the nominal run temperature fromphase compositions in experiments using the recentlyformulated garnet^phengite^clinopyroxene barometer ofKrogh Ravna amp Terry (2004) with the nominal pressureof the experimental runs (Fig 8c) This barometerhas been calibrated empirically and has so far not beentested by experiments The calibration uncertainty at35^45GPa is 10 as indicated by the grey band inFig 8c For most of the runs the pressure obtained fromthe barometer is within this grey band indicating thatit can be successfully used to identify UHP rocks

However there is a clear trend in the data at 35and 45GPa showing that the barometer is probablyunderestimating pressures at the low-temperature end(ie 7008C) and overestimating the pressures at thehigh-temperature end (ieT 9508C)

Pelite phase relations at sub-arc depthsThe experiments of this study on a pelitic starting materialdocument an important change in phase relations fromfore-arc to sub-arc depth There are complex phase rela-tions in subducted sediments that involve a variety ofhydrous phases at P527GPa (Patinlsaquo o Douce ampMcCarthy 1998 Hermann 2002b Poli amp Schmidt 2002Auzanneau et al 2006) In our experiment this is mani-fested by the presence of amphibole phengite and biotiteat 25GPa Amphibole is stable only in runs at 25GPaand below 9008C as has also been found in experimentalstudies on mafic hydrous systems (eg Liu et al 1996 Poliamp Schmidt 2002) In contrast the lack of garnet at 600and 7008C 25GPa and at 6008C 35GPa is not consistentwith natural phase assemblages or thermodynamic grids ofhigh-pressure low-temperature metapelites (Yardley 1989Powell amp Holland 1990) or other experimental studiesExperiments on pelitic compositions at 2^3GPa byNichols et al (1994) produced garnet as part of the solidassemblage at temperatures down to 6008C Spandleret al (2007b) using the same starting compositions as usedhere formed garnet in runs down to 6008C at 22GPaIt should be noted however that in the experimentsof Spandler et al (2007b) garnet growth at subsolidus runconditions was achieved only by adding pyrope seeds tocircumvent the sluggish nucleation of new garnet grainsTherefore we expect that the lack of garnet in the lowesttemperature experiments reported here is an artefact ofsluggish nucleation of garnet at these conditionsWith increasing pressure from 25GPa amphibole trans-

forms to clinopyroxene and biotite reacts to form phengiteThe transition from biotite-bearing to phengite-bearingparageneses has been investigated in detail by Auzanneauet al (2006) in a greywacke bulk composition They foundthat biotite disappears at 23^24GPa at 800^9008CAt higher pressure the paragenesis garnetthorn clinopyroxenethorn phengitethorn coesite kyanite is stable over a large P^Trange (Fig 9) Interestingly this paragenesis is stable notonly in metapelites but also in greywacke and K-bearingMORB (Schmidt et al 2004) This confirms that at sub-arc depths phengite will be the stable K phase insubducted crustWithin this dominant sub-arc paragenesisa fluid-present melting reaction can be postulated based onthe observed proportions and compositions of phasesEspecially useful are those experiments in which the P^Tconditions are constant but the added water contentchanges At 9008C 35GPa we performed two experimentswith 68 and 31wt H2O added Based on mass balanceusing the compositions given in the Electronic Appendix


731



ownloaded from

a significant decrease in the abundance of clino-pyroxene phengite and to a lesser extent coesitehas been observed in the experiment with the higherwater content whereas there is a strong increase in theamount of melt and a smaller increase in the amount ofgarnet These observations suggest the following meltingreaction

Cpxthorn Phethorn Coethorn Fluid GrtthornMelt eth2THORN

Such a melting reaction has been proposed previously(Schmidt 1996 Schmidt et al 2004) but is different fromthat of Auzanneau et al (2006) who suggested that garnetis on the reactant side and not on the product side whichcreates considerable problems in balancing Fe and MgPhengite is stable to surprisingly high temperatures at

sub-arc depths At 35GPa it was still stable at 9508Cwhereas at 45GPa phengite disappeared between 1000and 10508C Our experiments provide information on theminimum stability of phengite because we did not performfluid-absent melting experiments In fact it has beenobserved in a natural tonalite composition (no fluidadded 1wt H2O) that phengite was stable up to10108C at 3GPa and up to 10508C at 32GPa (Patinlsaquo oDouce 2005)

The new data in combination with literature dataprovide evidence for the key role of phengite as a K (andother LILE) carrier at sub-arc depths Other traceelements are likely to be controlled by accessory phasessuch as apatite zircon rutile allanite and monazitewhich are stable to high temperatures and pressures(Table 1) The influence of accessory phases on the traceelement contents of the liberated liquids will be presentedelsewhere The new data support previous suggestions thatsignificant liberation of a range of trace elements is mostprobably related to fluid-present metamorphic conditionsin the subducted sediments (Domanik amp Holloway 1996Hermann amp Green 2001)

The position of the wet solidusExperimental determination of the position of the wetsolidus at low pressures (P 525GPa) in felsic systemsis relatively straightforward as a clear transition fromglass-absent to glass-present runs can be observed (Huangamp Wyllie 1981) For example Spandler et al (2007b) deter-mined that the wet solidus for the same chemical systemas investigated here is at 6758C at 22GPa Howeverat higher pressures experimental determination of thesolidus becomes more complex as the amount of dissolved

2

4

600 800 1000

T (degC)

600 800 1000

T (degC)

N94

S04

A06

J99th

is s

tudy

biotite out

phen

gite

out

1

3

5

hydr

ous

glas

s

2

4

P (

GP

a)

P (

GP

a)

N94

S04

this

stu

dy

1

3

551015202530

8

N94

hydrous melt

(a) (b)

H2O

Fig 9 (a) The positions of the wet solidus in metapelite and metagreywacke compositions in P^T space Bold line this study N94 Nicholset al (1994) J99 Johnson amp Plank (1999) S04 Schmidt et al (2004) A06 Auzanneau et al (2006) The grey band indicates the pressure range ofthe subducted sediments at sub-arc depths Also shown is the phengite-out curve determined in this study (bold line)The continuation to lowerpressures is based on the experiments of Vieulzeuf amp Holloway (1988) The approximate position of the disappearance of biotite is based on thisstudy and the results of Auzanneau et al (2006) (b) Water content of melts (in wt H2O) based on this study compared with results fromNichols et al (1994) The fields of aqueous fluid and hydrous melts are shown The solidus and the second critical endpoint () for pelitesproposed by Schmidt et al (2004) are also shown


732



ownloaded from

H2O in the hydrous melts at the solidus increases withincreasing pressure (Huang amp Wyllie 1981) In this experi-mental set we found that the melt produced is quenchableto a glass only if the H2O content is530wt (Fig 6a)Thus we are confident that we were able to track the wetsolidus up to pressures of 35GPa where it is situatedbetween 700 and 7508C This is in very good agreementwith the study of Nichols et al (1994) who determinedthe wet solidus in subducted red clay sediments up to pres-sures of 3GPa (Fig 9) However the wet solidus deter-mined is very different from that in other studies Forexample Johnson amp Plank (1999) reported a wet solidusin the red clay system that is more than 1008C higherthan our results This might be related to a difference inthe starting material because the red clay they used hassignificantly lower SiO2 and higher FeOthornMgO contentsthan the starting material used in the present study(Fig 1) or to sluggish reaction in the lower temperatureruns of their experiments because of the use of an alreadycrystalline starting material The locus of fluid-presentmelting for a metagreywacke composition (Auzanneauet al 2006) which is very similar in composition to ourstarting material (Fig1) is also at considerably higher tem-peratures (1008C) at 35GPa than the results of ourstudyAs pointed out above at higher pressures we are only

able to determine the transition from glass to non-quench-able fluid From theoretical considerations of meltingtopologies it is evident that the wet solidus must have asteeper slope than the H2O isopleths in the hydrous melt(Hack et al 2007) This means that for the investigatedsystem the wet solidus should have a steeper slope than808CGPa (Fig 9) Interestingly the proposed wet solidiof Schmidt et al (2004) and Auzanneau et al (2006) atP435GPa are very similar to the 25wt H2O isoplethdetermined in this study (Fig 9) It is important to notethat these two studies were not well designed to tracethe wet solidus The amount of excess water in theseexperiments is very small and ranges from 04wt in themetagreywacke to 09wt in the metapelite Consideringthat hydrous melts contain at least 25^30 of water atthe wet solidus at P435GPa (Fig 8b) this will result inextremely low melt fractions of 1^3 Such low melt frac-tions are very hard to detect and virtually impossibleto trace by mass balance In contrast in this study wehave applied a different approach by increasing the watercontent at decreasing temperatures to keep the amount ofmelt produced high at about 20 Such melt proportionsare much easier to detect and provide larger pools foraccurate analyses of the melt composition

Composition of the fluid phaseThe first-order control on the composition of the fluidphase is given by the H2O content Where the fluid phaseis able to quench to a glass we classify it as a hydrous

melt (Hermann et al 2006) For the studied pelitecomposition the transition from quenchable glass to non-quenchable fluid occurs at 30wt H2O With increasingtemperature the H2O content in the hydrous meltdecreases in a systematic way (Fig 9b) The liquidus deter-mined by Nichols et al (1994) for a red clay compositionwith 8wt H2O is situated between the 5 and 10wtH2O isopleths and thus is consistent with the isoplethsdetermined in this study At the low-temperature side ofthe solidus the fluid phase is dominated by H2O and thusrepresents an aqueous fluid (Fig 9) In our study we wereunable fully to quantify the amount of solute in the dia-mond traps However a comprehensive compilation ofmajor element solubilities by Manning (2004) indicatesthat such aqueous fluids are surprisingly dilute and arelikely to contain 55wt of total dissolved solutes atP520GPa and T 6008C At 22GPa and 600^6508CSpandler et al (2007b) constrained the solubilities ofelements with synthetic fluid inclusions using a startingmaterial identical to that in this study They found thattotal dissolved solutes range from 4 to 55wt Thisvalue is likely to be representative also for our run at6008C 25GPa (Table 3) Solubilities tend to increase withincreasing pressure and temperature (Manning 2004)and guided by the general topology of the solubility iso-pleths (Hack et al 2007) we have estimated that the totaldissolved solutes are 8 and 11wt at 35GPa at 600 and7008C respectively This seems reasonable consideringthat Kessel et al (2005b) reported total solute contents of12wt at 40GPa 7008C in the MORBthornH2O systemThis system produced a mineral assemblage (apart fromphengite) similar to that in the pelite system and consider-ing that the main solutes are SiO2 Al2O3 and Na2O(Table 3) total solute contents are likely to be comparableFor the run at 45GPa 7008C we estimated a total amountof solutes of 15 based on the findings of Kessel et al(2005b) in the MORBthornH2O system those workersreported aqueous fluids with total dissolved solutes of7^17wt in the range 40^50GPa and 700^8008CAs solute contents increase along the solidus for the

aqueous fluid and the H2O increases in the hydrous meltthe two compositions merge together until they are thesame at the second critical endpoint where the solidusterminates The solubility isopleths obtained in this studyindicate that at 35GPa the second critical point has notbeen reached for the pelitic bulk composition However itmust be noted that the design of our and the Schmidt et al(2004) experiments is not suitable to detect any closure ofthe miscibility gap between aqueous fluid and hydrousmelt The position of the second critical endpoint cannotbe determined with such experiments because the meltsare no longer quenchable along the solidus above 35GPaIt is necessary to conduct experiments at high and variablewater contents to locate the second critical endpoint


733



ownloaded from

As long as solubilities in the fluid phase are consideredthe actual position of the second critical endpoint is notof paramount importance because there is a strongincrease in solubilities over a temperature interval of1008C (as observed for example between 7008C and8008C 45GPa) whether or not the wet solidus is still pre-sent (eg Manning 2004 Hermann et al 2006)The large melt pools obtained in this study permit

accurate analysis of the experimental melts as a functionof increasing temperature (Fig 7) The melts are essentiallygranitic in composition The most significant observedchange within the granite field is the change of Na2O andK2O with temperature Close to the wet solidus the meltsare trondhjemitic with Na2O4K2O At higher tempera-tures the melts are granitic sensu stricto with K2O4Na2OIt is interesting to note that a very similar trend hasbeen observed at lower pressures by Patinlsaquo o-Douce ampMcCarthy (1998) Fluid-present melting in a muscoviteschist at 700^7508C 06^10GPa produced trondhjemiticmelt compositions whereas melting at higher temperaturesproduced granitic melt compositions This similarityin melt composition evolution with temperature athigh and low pressures is not expected as the bufferingparagenesis changes significantly from quartz^plagioclase^muscovite^biotite^garnet (low pressure) tocoesite^clinopyroxene^phengite^garnet (high pressure)

Because of the large stability field of phengite most experi-mental melts are buffered in K2O In fact there is asystematic increase in K2O with increasing temperatureas is expected for an element that has a constant composi-tion in the buffering phase (Fig 7a) In conjunction withthe decrease in H2O with increasing temperature thereis a very pronounced and systematic change of K2OH2Oas a function of temperature (Fig 10a) Figure 10a is usedto show how efficiently K2O can be removed from sub-ducted sediments as a function of temperature nature ofthe fluid phase and available H2O At low temperaturesan aqueous fluid is stable and hence a very high fluid^rock ratio is needed to strip substantial K2O out of sub-ducted sediments In contrast at higher temperaturesmuch less H2O is required to efficiently remove K2O witha hydrous melt Figure 10 can also be used to predictin experiments with a given K2OH2O at which tempera-ture it is expected that the solid K phase is completelydissolved in the produced fluid The temperature rangeof phengite dissolution used as an argument for thepresence of supercritical fluids by Schmidt et al (2004)can be readily explained in terms of this diagramA K2OH2O 12 in their starting material is reachedby the hydrous melt at 1000^11008C in agreementwith the disappearance of phengite observed by Schmidtet al (2004)

K2O

H2O

600 800 1000

2535

45

aqueous fluid

hydrous meltprimitive arc

lavas

top slab temperatures

2

4

400 600 800 1000

T (degC)

P (

GP

a)

1) 2) 3) 4) 5)

N94

this

stu

dy

phen

gite

out

1

3

5

hydr

ous

glas

s

T (degC)

scenario 1 scenario 2 scenario 3

hydrous melt

aqueous fluid


fluid absentmica melting

0001

001

01

1

500 700 900

(a) (b)

Fig 10 (a) Variation of K2OH2O in aqueous fluids and hydrous melts produced in the experiments compared with the range of K2OH2O inprimitive arc lavas and the K2OH2O that would result from fluid-absent melting of phengite- or phlogopite-bearing rocks (b) P^Tdiagramillustrating the three fundamentally different scenarios for element transfer in subduction zones when thermal models for top-slab temperaturesare compared with the position of the wet solidus and the stability of phengite In scenario 1 elements are transferred by an aqueous fluid inscenario 2 by fluid-present melting in scenario 3 by fluid-absent melting Top-slab temperatures from numerical and analogue models 1 278angle of subduction 10 cmyear convergence rate 50Myr old oceanic crust 2 278 angle of subduction 10 cmyear convergence rate 20Myr oldoceanic crust 3 278 angle of subduction 3 cmyear convergence rate 20Myr old oceanic crust [numerical models 1^3 are from Peacock et al(1994)] 4 scaled analogue model 498 dip 35 cmyear convergence rate (Kincaid amp Griffiths 2004) 5 numerical model with enhanced couplingwith the mantle wedge 608 dip 45 cmyear convergence rate 10Myr old oceanic crust (van Keken et al 2002)


734



ownloaded from

We have already highlighted the difficulties of constrain-ing the nature of the fluid phase based on textural observa-tions when the amount of fluid phase is low andor thefluid phase does not quench to a glass However in thisstudy we have observed some compositional characteristicsthat might help researchers in future studies to constraintthe nature of the fluid phase Element ratios are particu-larly useful because determination of absolute concentra-tions is often difficult in the hydrous quench from aqueousfluids We found that SiO2Al2O3 and Na2OAl2O3

abruptly increase from hydrous melts to aqueous fluids(Fig 7e and f) This is in excellent agreement with thecalculated solubilities in aqueous fluids in equilibriumwith eclogite showing that Si and Na are the main cationsand Al is a minor component (Manning1998) The Na2OAl2O3 of aqueous fluids in equilibrium with a high-pressure metapelite has also been determined in fluidinclusions trapped in experiments at 22GPa 600^6508Cand is close to unity (Spandler et al 2007b) in agreementwith our data (Fig 7f) Green amp Adam (2003) analysed thequench material precipitated from aqueous fluids that hadbeen equilibrated with a basaltic eclogite assemblage at650^7008C 30GPa The quench material was character-ized by high SiO2Al2O3 of 12^20 and high Na2OAl2O3

of 07^08 providing further evidence for the distinct char-acteristics of these ratios in aqueous fluidsA surprising feature is the very low K2O content of the

aqueous fluids which is between 004 and 016wt whenthe H2O content of the fluid is taken into account(Table 3) Such low K2O contents have also been observedin the fluids present in the K-bearing basalt experimentsof Green amp Adam (2003) Although their starting materialcontained only 03wt K2O and they applied a highH2Ostarting material ratio of unity all experiments pro-duced residual phengiteThis provides strong evidence thatthe solubility of K2O in the aqueous fluid must be smallerthan 03wt In fact if the K2O content of the fluid isnormalized to a total solute content of 5 wt as suggestedby Green amp Adam (2003) then it reaches values of007^011wt these values are in excellent agreementwith the low concentrations found in our study

Element recycling andtop-slab temperaturesOne of the most remarkable characteristics of primitivearc lavas compared with MORB is the distinct enrichmentof H2O and LILE (Tatsumi amp Eggins 1995) A range ofisotopic and geochemical criteria provide strong evidencethat these components are derived from the subductedslab (Hawkesworth et al 1993 Plank amp Langmuir 19931998) Although the H2O may be sourced from a numberof lithologies in the slab (Poli amp Schmidt 2002) severalstudies have confirmed that subducted sediments play akey role in the LILE budget of subduction zones (Plank ampLangmuir 1993 1998 Peate et al 1997) In particular

pelites are regarded as the most important componentof the subducted sediment for hosting most lithophiletrace elements (Plank amp Langmuir 1998) It is thereforeof prime interest to understand the nature and compositionof the fluid phase that is formed at sub-arc depths in sub-ducted pelites To evaluate the process of LILE recycling itis necessary to compare the position of the wet metapelitesolidus (and the extension of the transition of hydrous meltto non-quenchable fluid) to thermal models of top-slabtemperatures Figure 10b shows a variety of thermalmodels with respect to the wet pelite solidus the fieldwhere hydrous melts are stable and the stability of phen-gite Thermal modelling has shown that temperaturesat the top of the subducting slab are mainly controlled bythe age of the incoming plate (the older the cooler) andthe speed of subduction (the faster the cooler (Peacocket al 1994) Recent thermal and mechanical modelling hasalso highlighted the importance of coupling between theslab and mantle wedge which may dramatically influencetemperatures on top of the slab (van Keken et al 2002Kincaid amp Griffiths 2004)The variety of controlling para-meters results in vastly different model conditions at sub-arc depths that range from very cold (400^6008C) to veryhot (950^10008C) This highlights the need to obtain inde-pendent constraints on top-slab temperaturesTo characterize recycling of elements from the subducted

sediments to themantlewedge we canconsider three funda-mentally different scenarios In the first scenario the top-slabtemperature iswithintheaqueous fluid fieldat relativelylow temperatures at sub-arcdepthsThis is thecase forallbutthe hottest proposed geotherms of the models given byPeacock et al (1994) Under such conditions effective extrac-tion of K2O and other LILE will be difficult Our experi-ments have shown that the K2O content and moreimportantly the K2OH2O ratio of such fluids are very low(Fig 10a) This means that LILE can be removed by fluidsonly in zones where there is a very high fluid^rock ratioAlso the solubility of LREE which are often enriched inarc lavas is extremely low (Hermann 2002a Spandler et al2007b) In the second scenario top-slab temperatures arehigh enough (47008C) for fluid-present melting at sub-arcdepths but no fluid-absent melting of phengite occursKincaidampGriffiths (2004) reported results from scaled ana-logue models that produced top-slab temperatures that fallwithin this scenario K2O and LILE will be efficientlyremoved from the sediments depending on the availabilityof an externally derived fluid that triggers melting It hasbeen suggested that the breakdown of hydrous phases inmafic and ultramafic rocks in the deeper and cooler parts ofthe slab might provide the fluids necessary for melting thesediments on top of the slab (Hermann amp Green 2001)Alternatively hybrid chlorite- and talc-rich rock types thatform as a result of interaction between the subducted slaband the mantle wedge might liberate significant amounts of


735



ownloaded from

fluid at 700^8008C at sub-arc depths and thus will triggermelting of the sediments (Spandler et al 2007a) In the thirdscenario very high top-slab temperatures of 950^10008Cmight lead to fluid-absent melting of phengite in subductedsediments Such high top-slab temperatures could occurwhere there is enhanced coupling with a hot mantle wedge(van Keken et al 2002) or where asthenospheric inflowoccurs in response to slab rollback (Kincaid amp Griffiths2004) In such a case nearly complete transfer of K2O andLILE from the subducted sediments to the mantle wedgeoccurs at sub-arc depths independent of whether or notthere is additional fluid fluxingOne interesting aspect of fluid-present melting in sedi-

ments is that the amount of K2O in the melt (bufferedby phengite) increases systematically with increasingtemperature It was suggested previously that in manyisland-arc systems there exists a relationship between theK2O content of primitive lavas and the depth to the top ofthe subducting slab beneath the volcanoccedilthe so-calledlsquoK^hrsquo or lsquopotash^depthrsquo relationship (Dickinson 1975)Consequently it has been postulated that pressure is theprimary control on K2O release from the slab Our studyhowever suggests that temperature and not pressure isthe controlling parameter Because the temperature atthe top of the subducting slab increases with increasingdepth the observed lsquopotash^depthrsquo relationship still holdsWe suggest that the systematic change in the K2OH2O

ratio in the slab fluid as a function of temperature and toa lesser extent pressure provides a useful tool to distin-guish between the three scenarios outlined above K2O isbuffered by residual phengite and is a good proxy forthe behaviour of the LILE which are the most importantelements in the subduction component of arc lavas Therationale behind normalizing to the water content is thatthe amount of water added to the mantle wedge will be aprincipal factor in determining how much basaltic meltis produced during wet melting of the mantle wedge peri-dotite which in turn will have an influence on the finalconcentration of LILE in the resultant arc magmasThe K2OH2O of the fluid phase leaving the subductedsediments at the top of the slab is extremely low (0002)in scenario 1 (Fig 10a) As soon as hydrous melts are pre-sent (scenario 2) the K2OH2O increases drastically to01^1 for the temperature range of fluid-present meltingFor fluid-absent phengite melting at sub-arc depths as out-lined in scenario 3 the K2OH2O is bracketed between 1given by the high temperature of melting and 23 givenby the stoichometry of phengite (105wt K2O 45wtH2O)These constraints can now be compared with the K2O

H2O ratio of primitive arc lavas Primitive lavas from arange of volcanic arcs have between 02 and 08wtK2O (Kelemen amp Hanghoj 2003) and between 2 and4wt H2O (Sobolev amp Chaussidon 1996 Stern 2002)

giving K2OH2O ratios of between 005 and 04 Anexcellent dataset has been recently published on melt inclu-sions in olivine from the Kamtchatka arc where K2Oand H2O have been measured in the same inclusions(Portnyagin et al 2007) The observed K2OH2O valuesrange from 005 to 042 with the majority of data cluster-ing at 025 The depleted mantle source of arc magmas isexpected to contain of the order of 0007wt K2O and001wt H2O (Salters amp Stracke 2004) and so is unlikelyto contribute significantly to the budget of these elementsin arc magmas Therefore we can reasonably assume thatthe K2OH2O ratio of primitive arc magmas is primarilyderived from slab contributionsTo make a link between the K2OH2O of arc magmas

and the K2OH2O of the slab-derived sediment flux wehave to consider potential alteration of this ratio in themantle wedge according to the three different scenariosFor models predicting cool top-slab temperatures (sce-nario 1) aqueous fluids leaving the slab will interact withthe mantle wedge and form a variety of hydrous phasessuch as antigorite amphibole talc chlorite and phlogopite(Poli amp Schmidt 2002) Talc chlorite and phlogopitein particular are stable up to considerable pressuresand temperatures and the formation of these phases willalter the H2O and K2O content of the residual fluidHowever overall it is expected that the solubilities inthe peridotite system are lower than in the pelite systemand that the residual fluids are likely to be H2O-richaqueous fluids with very low K2OH2O An interestingalternative model is that the fluid which transportsthe slab signature to the locus of partial melting in themantle wedge derives from the breakdown of the hydrouschlorite^talc assemblage which will occur at 8008C(Spandler et al 2007a) We can estimate the proportion ofhydrous minerals from the interaction of peridotite withaqueous fluids derived from the subducted sedimentsFor a fluid composition such as that determined at 6008C35GPa (Table 3) we obtain an estimated talcchloritephlogopite ratio of 30151 If upon dehydration of such anassemblage at 8008C all the K2O hosted in phlogopiteenters the fluid phase the resulting K2OH2O ratiowould be 003 This is a maximum value as potentiallysome phlogopite will remain This calculated ratio isconsiderably higher than in the original aqueous fluidleaving the slab but is still significantly lower thanthe K2OH2O ratio in arc lavas (Fig 10) Therefore itis very difficult to explain the K2OH2O of arc lavas withscenario 1 Clearly more experiments are needed to con-strain the solubility of K2O in aqueous fluids in suchsystemsIn the second scenario a hydrous melt is produced from

the subducted sediments This melt will most likely enterthe mantle wedge at pressures above the amphibolestability field (Fumagalli amp Poli 2005) and at temperatures


736



ownloaded from

that exceed the stability of antigorite talc and chlorite(Poli amp Schmidt 2002) Hence the only mineral phasecapable of altering the K2OH2O is phlogopite Sekine ampWyllie (1982) have shown that interaction of graniticmelts with peridotite will generally produce phlogopiteThe extent of phlogopite formation depends on the tem-perature and especially on whether the melts are incontact with olivine or are shielded by the formation ofnew orthopyroxene Malaspina et al (2006) have investi-gated garnet-orthopyroxenites that have been interpretedas the products of interaction of peridotites with hydroussilicate melts at sub-arc pressures They have observedthat a residual fluid was trapped in the garnet that wasenriched in incompatible elements These garnet-ortho-pyroxenites are surrounded by layers of phlogopitewhich probably formed during the interaction process(N Malaspina personal communication 2007) Phlogopitehas K2OH2O of 2 given by the stoichiometry Figure 10shows that hydrous melts released from the slab attemperatures of 700^9008C have a K2OH2O of 01^1which is lower than the K2OH2O of phlogopite Thusif phlogopite is formed in the mantle wedge peridotitethis will lower the K2OH2O of the residual fluidthat will enter the locus of wet melting The observedK2OH2O of 005^04 of primitive arc lavas thusconstrains the likely minimum K2OH2O in the slab-derived fluid phaseTherefore estimated top-slab tempera-tures based on the K2OH2O of arc lavas would representminimum values and would depend on the amountof interaction of the slab-derived fluid with the mantlewedge In the case of channel flow such interaction wouldbe minimal and the slab signature could be transportednearly entirely to the locus of partial melting In sucha case K2OH2O temperature estimates based on thiswork would be accurate On the other hand if the fluid istransported via porous flow extensive interactionand phlogopite formation is expected K2OH2O wouldthen be determined by the solubility of phlogopite inaqueous fluids in the peridotite system No experimentalresults are available yet to quantify this process but itis expected that K2OH2O would increase systematicallywith increasing temperatureFluid-absent melting of phengite at high temperatures

(scenario 3) produces melts that have virtually the sameK2OH2O as phlogopite Therefore in such a casethe K2OH2O is not significantly changed by the potentialformation of phlogopite in the mantle wedge It is alsoworth noting that fluid-absent melting of a phlogopite-peridotite would produce melts with K2OH2O of 2These values are significantly higher than what is observedin primitive arc lavas and consequently we do not considerthis scenario to be likelyThe combination of our experimental results on the

systematics of K2OH2O variation in the fluid phase

released from subducted sediments with considerations ofthe interaction of these fluids with the overlying mantlewedge suggests that the easiest way to explain the observedK2OH2O in arc lavas is if fluid-present melting (scenario2) at top-slab temperatures of 750^9008C occurs in thesubducted sediments (Fig 10) Although this study clarifiesthe nature and composition of the fluid phase presentin subducted sediments clearly more experiments areneeded to constrain the interaction of these fluids with themantle wedge for a complete understanding of LILE recy-cling in subduction zones

CONCLUSIONS(1) Phengite is the most important host for K (and otherLILE) in the subducted crust This study shows that phen-gite is stable to high temperatures at sub-arc depths (9508Cat 35GPa 10008C at 45GPa) and hence fluid-absentphengite breakdown is not a viable process to release Kfrom the subducted slab in most subduction zones(2) The pelite solidus is steeply sloping and is located at

7008C at 25GPa and between 700 and 7508C at 35GPaAt higher pressures we could only determine the transitionfrom hydrous glass to non-quenchable fluid at 8008C45GPa Hence the solute content of the fluid phasereleased from subducted sediments will strongly increasein the temperature interval from 700 to 8008C at sub-arcdepths(3) The K2O and H2O content of the fluid phase is

buffered by the phengite-bearing residue and changesdrastically as a function of the nature of the fluid phaseand the slab temperature Aqueous fluids which are stableat temperatures lower than the wet solidus are dilute andcontain K2O 51wt and H2O 480wt In contrasthydrous melts display increasing K2O and decreasingH2O contents with typical K2OH2O of 01^1 Thusefficient extraction of LILE from the subducted sedimentsrequires wet melting(4) The comparison of K2OH2O produced in hydrous

experiments with primitive arc basalts provides evidencethat the most important process for element recyclingfrom the slab to the mantle wedge occurs via fluid-presentmelting of the sediments and suggests that top-slabtemperatures must be in the range of 700^9008C Oncethe interaction of the fluids with the mantle wedge isunderstood the K2OH2O of primitive arc basalts canbe used to obtain independent constraints on top-slabtemperatures

ACKNOWLEDGEMENTSWe would like to thank D Scott andW O Hibberson forassistance with the experiments A Norris and F Brink fortheir help with the electron microprobe and SEM analysesrespectively and C Allen and M Shelly for their help


737



ownloaded from

with the ICP-MS analyses David Green encouragedthis work and provided helpful comments We thankMWilson for editorial handling and S Poli and an anon-ymous reviewer for constructive comments which helpedto improve the paper This work has been financially sup-ported by the Australian Research Council

SUPPLEMENTARY DATASupplementary data for this paper are available at Journalof Petrology online

REFERENCESAuzanneau E Vielzeuf D amp Schmidt MW (2006) Experimental

evidence of decompression melting during exhumation ofsubducted continental crust Contributions to Mineralogy and Petrology

152 125^148Barker F (1979) Trondhjemite definition environment and

hypothesis of origin In Barker F (ed) Trondhjemites Dacites and

Related Rocks Developments in Petrology Amsterdam Elsevierpp 1^12

Bishop F C Smith J V amp Dawson J B (1978) Na K P and Ti ingarnet pyroxene and olivine from peridotite and eclogite xenolithsfrom African kimberlites Lithos 11 155^173

Brunet F Bonneau V amp Irifune T (2006) Complete solid-solutionbetween Na3Al2(PO4)3 and Mg3Al2(SiO4)3 garnets at highpressure American Mineralogist 91 211^215

Carswell D A OrsquoBrien P J Wilson R N amp Zhai M (1997)Thermobarometry of phengite-bearing eclogites in the DabieMountains of central China Journal of Metamorphic Geology 15239^252

Chopin C (1984) Coesite and pure pyrope in highgrade blueschists ofthe western Alps A first record and some consequencesContributions to Mineralogy and Petrology 86 107^118

Class C Miller D M Goldstein S L amp Langmuir C H (2000)Distinguishing melt and fluid subduction and componentsin Umnak volcanics Aleutian Arc Geochemistry Geophysics

Geosystems 1 paper number 1999GC000010DickinsonW R (1975) Potash^depth (K^H) relations in continental

margin and intra-oceanic magmatic arcs Geology 3 53^56Domanik K J amp Holloway J R (1996) The stability of phengitic

muscovite and associated phases from 55 to 11GPa implicationsfor deeply subducted sediments Geochimica et Cosmochimica Acta 604133^4150

Eggins S M Rudnick R L amp McDonough W F (1998) Thecomposition of peridotites and their minerals a laser ablationICP-MS study Earth and Planetary Science Letters 154 53^71

Elliott T (2003) Tracers of the slab In Eiler J (ed) Inside the

Subduction Factory Geophysical Monograph American GeophysicalUnion 183 23^45

Ellis D (1986) Garnet^liquid Fe2thorn^Mg equilibria and implicationsfor the beginning of melting in the crust and subduction zonesAmericanJournal of Science 286 765^791

Ellis D J amp Green D H (1979) An experimental study of the effectof Ca upon garnet^clinopyroxene Fe^Mg exchange equilibriaContributions to Mineralogy and Petrology 71 13^22

Fumagalli P amp Poli S (2005) Experimentally determined phaserelations in hydrous peridotites to 65GPa and their consequenceson the dynamics of subduction zones Journal of Petrology 46555^578

Green T H (1977) Garnet in silicic liquids and its possible use as aP^T indicator Contributions to Mineralogy and Petrology 65 59^67

GreenTH amp Adam J (2003) Experimentally-determined trace ele-ment characteristics of aqueous fluid from partially dehydratedmafic oceanic crust at 30 GPa 650-7008C European Journal of

Mineralogy 15 815^830Green T H amp Hellman P L (1982) Fe^Mg partitioning between

coexisting garnet and phengite at high pressure and comments ona garnet^phengite geothermometer Lithos 15 253^266

Hack A C Hermann J amp Mavrogenes J A (2007) Mineralsolubility and hydrous melting relations in the deep Earth analysisof some binary A^H2O system pressure^temperature^compositiontopologies AmericanJournal of Science 307 833^855

Hawkesworth C J Gallagher K Herget J M amp McDermott F(1993) Mantle and slab contributions in arc magmas Annual

Review of Earth and Planetary Sciences 21 175^204Hermann J (2002a) Allanite thorium and light rare earth element

carrier in subducted crust Chemical Geology 192 289^306Hermann J (2002b) Experimental constraints on phase relations in

subducted continental crust Contributions to Mineralogy and Petrology

143 219^235Hermann J (2003) Experimental evidence for diamond-facies

metamorphism in the Dora^Maira massif Lithos 70 163^182Hermann J amp Green D H (2001) Experimental constraints on high

pressure melting in subducted crust Earth and Planetary Science

Letters 188 149^186Hermann J Rubatto D Korsakov A amp Shatsky V S (2001)

Multiple zircon growth during fast exhumation of diamondiferousdeeply subducted continental crust (Kokchetav massifKazakhstan) Contributions to Mineralogy and Petrology 141 66^82

Hermann J Spandler C Hack A amp Korsakov A V (2006)Aqueous fluids and hydrous melts in high-pressure and ultra-highpressure rocks Implications for element transfer in subductionzones Lithos 92 399^417

Huang W L amp Wyllie P J (1981) Phase relationships of S-typegranite with H2O to 35 kbar Muscovite granite fromHarney Peak South Dakota Journal of Geophysical Research 8610515^10529

Johnson M C amp Plank T (1999) Dehydration and meltingexperiments constrain the fate of subducted sedimentsGeochemistry Geophysics Geosystems 1 paper number 1999GC000014

Kelemen P B amp Hanghoj K (2003) One view of the geochemistryof subduction-related magmatic arcs with emphasis on primitiveandesite and lower crust In Rudnick R L (ed) The Crust

Treatise on Geochemistry Amsterdam Elsevier pp 593^659Kessel R Schmidt M W Ulmer P amp Pettke T (2005a) Trace

element signature of subduction-zone fluids melts and supercriticalliquids at 120^180 km depth Nature 437 724^727

Kessel R Ulmer P Pettke T Schmidt MW amp Thompson A B(2005b) The water^basalt system at 4 to 6GPa Phase relationsand second critical endpoint in a K-free eclogite at 700 to 14008CEarth and Planetary Science Letters 237 873^892

Kincaid C amp Griffiths R W (2004) Variability in flow andtemperature within mantle subduction zones Geochemistry

Geophysics Geosystems 5 paper number Q06002Konzett J Frost D J Proyer A amp Ulmer P (2007) The Ca-Eskola

component in eclogitic clinopyroxene as a function of pressuretemperature and bulk composition an experimental study to15GPa with possible implications for the formation of orientedSiO2-inclusions in omphacite Contributions to Mineralogy and

Petrology (in press) doi 101007s00410-007-0238-0Krogh Ravna E J amp Terry M P (2004) Geothermobarometry of

UHP and HP eclogites and schistsccedilan evaluation of equilibria


738



ownloaded from

among garnet^clinopyroxene^kyanite^phengite^coesitequartzJournal of Metamorphic Geology 22 579^592

Lang H M amp Gilotti J A (2007) Partial melting of metapelites atultrahigh-pressure condition Greenland Caledonides Journal of

Metamorphic Geology 25 129^147Liu J Bohlen S R amp ErnstW G (1996) Stability of hydrous phases

in subducting oceanic crust Earth and Planetary Science Letters 143161^171

Malaspina N Hermann J Scambelluri M amp Compagnoni R(2005) Multistage metasomatism in ultrahigh-pressure mafic rocksfrom the North Dabie Complex (China) Lithos 90 19^42

Malaspina N Hermann J Scambelluri M amp Compagnoni R(2006) Polyphase inclusions in garnet-orthopyroxenite (DabieShan China) as monitors for metasomatism and fluid relatedtrace element transfer in subduction zone peridotite Earth and

Planetary Science Letters 249 173^187Manning C E (1998) Fluid composition at the blueschistêclogite

transition in the model system Na2O^MgOÂl2O3^SiO2^H2O^HCl Schweizerische Mineralogische und Petrographische Mitteilungen 78225^242

Manning C E (2004) The chemistry of subduction-zone fluids Earthand Planetary Science Letters 223 1^16

Massonne H-J (2003) A comparison of the evolution of diamondifer-ous quartz-rich rocks from the Saxonian Erzgebirge and theKokchetav Massif are so-called diamondiferous gneisses magmaticrocks Earth and Planetary Science Letters 216 347^364

Massonne H-J amp Nasdala L (2003) Characterization of an earlymetamorphic stage through inclusions in zircon of a diamondifer-ous quartzofeldspathic rock from the Erzgebirge GermanyAmerican Mineralogist 88 883^889

Massonne H-J amp SchreyerW (1987) Phengite geobarometry basedon the limiting assemblage with K-feldspar phlogopite and quartzContributions to Mineralogy and Petrology 96 212^224

Massonne H-J amp Szpurzka Z (1997) Thermodynamic propertiesof white micas on the basis of high-pressure experiments in thesystems K2O^MgOÂl2O3^SiO2^H2O and K2O^FeOÂl2O3^SiO2^H2O Lithos 41 229^250

Meyre C De Capitani C Zack T amp Frey M (1999) Petrology ofhigh-pressure metapelites from the Adula Nappe (Central AlpsSwitzerland) Journal of Petrology 40 199^213

Mposkos E D amp Kostopoulos D K (2001) Diamond former coesiteand supersilic garnet in metasedimentary rocks from the GreekRhodope a new ultrahigh-pressure metamorphic province estab-lished Earth and Planetary Science Letters 192 497^506

Nichols G T Wyllie P J amp Stern C R (1994) Subduction zonemelting of pelagic sediments constrained by melting experimentsNature 371 785^788

Ono S (1998) Stability limits of hydrous minerals in sedimentand mid-ocean ridge basalt compositions implications for watertransport in subduction zones Journal of Geophysical Research 10318253^18267

Patinlsaquo o Douce A E (2005) Vapor-absent melting of tonalite at 15^32kbar Journal of Petrology 46 275^290

Patinlsaquo o Douce A E amp McCarthy T C (1998) Melting of crustalrocks during continental collision and subduction In Hacker BR amp Liou J G (eds) When Continents Collide Geodynamics and

Geochemistry of Ultrahigh-Pressure Rocks Dordrecht KluwerAcademic pp 27^55

Peacock S M Rushmer T amp Thompson A B (1994) Partial melt-ing of subducted oceanic crust Earth and Planetary Science Letters 121227^244

Pearce N J G Perkins W T Westgate J A Gorton M PJackson S E Neal C R amp Chenery S P (1997) A compilation

of new and published major and trace element data for NISTSRM 610 and NIST SRM 612 glass reference materialsGeostandards News 21 115^144

Peate D W Pearce J A Hawkesworth C J Colley HEdwards C M H amp Hirose K (1997) Geochemical variationsinVanuatu arc lavas the role of subducted material and a variablemantle wedge composition Journal of Petrology 38 1331^1358

Philippot P Chevallier P Chopin C amp Debussy J (1995) Fluidcomposition and evolution in coesite-bearing rocks (Dora^Mairamassif Western Alps) Implications for element recycling duringsubduction Contributions to Mineralogy and Petrology 121 29^44

PlankT amp Langmuir C H (1993) Tracing trace elements from sedi-ment input to volcanic output at subduction zones Nature 362739^743

Plank T amp Langmuir C H (1998) The chemical composition ofsubducting sediments and its consequences for the crust andmantle Chemical Geology 145 325^394

Poli S amp Schmidt M W (2002) Petrology of subducted slabsAnnual Review of Earth and Planetary Sciences 30 207^235

Portnyagin M Hoernle K Plechov P Mironov N ampKhubunaya S (2007) Constraints on mantle melting andcomposition and nature of slab components in volcanic arcs fromvolatiles (H2O S Cl F) and trace elements in melt inclusionsfrom the Kamchatka Arc Earth and Planetary Science Letters 25553^69

Powell R amp Holland T (1990) Calculated mineral equilibria in thepelite system KFMASH (K2O^FeO^MgOÂl2O3^SiO2^H2O)American Mineralogist 75 367^380

Ringwood A E amp Major A (1971) Synthesis of majorite and otherhigh-pressure garnets and perovskites Earth and Planetary Science

Letters 12 411^441Rubatto D amp Hermann J (2007) Experimental zirconmelt and

zircongarnet trace element partitioning and implications for thegeochronology of crustal rocks Chemical Geology 241 38^61

Salters V J M amp Stracke A (2004) Composition of the depletedmantle Geochemistry Geophysics Geosystems 5 paper number Q05004

Schmalaquo dicke E amp Mulaquo ller W F (2000) Unusual exsolution phenom-ena in omphacite and partial replacement of phengite by phlogopi-tethornkyanite in an eclogite from the Erzgebirge Contributions to

Mineralogy and Petrology 139 629^642Schmidt M W (1996) Experimental constraints on recycling of

potassium from subducted oceanic crust Science 272 1927^1930Schmidt M W Vielzeuf D amp Auzanneau E (2004) Melting and

dissolution of subducting crust at high pressures the key role ofwhite mica Earth and Planetary Science Letters 228 65^84

Sekine T amp Wyllie P J (1982) Phase relationships in the systemKAlSiO4^Mg2SiO4^SiO2^H2O as a model for hybridizationbetween hydrous siliceous melts and peridotite Contributions to

Mineralogy and Petrology 79 368^374Sobolev A V amp Chaussidon M (1996) H2O concentrations in pri-

mary melts from supra-subduction zones and mid-ocean ridgesImplications for H2O storage and recycling in the mantle Earthand Planetary Science Letters 137 45^55

Sobolev N V amp Lavrentrsquoev J G (1971) Isomorphic sodium admix-ture in garnets formed at high pressures Contributions to Mineralogy

and Petrology 31 1^12Sobolev N V amp Shatsky V S (1990) Diamond inclusions in garnets

from metamorphic rocks A new environment for diamond forma-tion Nature 343 742^746

Sorenson S S Grossman J N amp Perfit M R (1997) Phengite-hosted LILE enrichment in eclogite and related rocks implicationsfor fluid-mediated mass transfer in subduction zones and arcmagma genesis Journal of Petrology 38 3^34


739



ownloaded from

Spandler C J Hermann J Arculus R J amp Mavrogenes J A(2003) Redistribution of trace elements during prograde meta-morphism from lawsonite blueschist to eclogite facies implicationsfor deep subduction-zone processes Contributions to Mineralogy and

Petrology 146 205^222Spandler C Hermann J Faure K Mavrogenes J A amp

Arculus R J (2007a) The importance of talc and chlorite lsquohybridrsquorocks for volatile recycling through subduction zonesevidence from the high-pressure subduction mecurren lange of NewCaledonia Contributions to Mineralogy and Petrology (in press)doi 101007s00410-007-0236-2

Spandler C Mavrogenes J A amp Hermann J (2007b) Experimentalconstraints on element mobility from subducted sediments usinghigh-P synthetic fluidmelt inclusions Chemical Geology 239 228^249

Stern R J (2002) Subduction zones Reviews of Geophysics 40 doi2001RG000108

Stolaquo ckhert B Duyster J Trepman C amp Massone H-J (2001)Microdiamond daughter crystals precipitated from supercriticalCOHthorn silicate fluids included in garnet Erzgebirge GermanyGeology 29 391^394

Syracuse E M amp Abers G A (2006) Global compilation ofvariations in slab depth beneath arc volcanoes and implicationsGeochemistry Geophysics Geosystems 7 paper number Q05017

TatsumiY (1986) Formation of the volcanic front in subduction zonesGeophysical Research Letters 17 717^720

Tatsumi Y amp Eggins S M (1995) Subduction Zone MagmatismCambridge Blackwell

Taylor S R amp McLennan S M (1985) The Continental Crust its

Composition and Evolution Oxford BlackwellTenthorey E amp Hermann J (2004) Composition of fluids during

serpentinite breakdown in subduction zones Evidence for limitedboron mobility Geology 32 865^868

Thompson R N (1975) Is upper mantle phosphorus contained insodic garnet Earth and Planetary Science Letters 26 417^424

van Keken P E Kiefer B amp Peacock S M (2002) High-resolutionmodels of subduction zones Implications for mineral dehydrationreactions and the transport of water into the deep mantleGeochemistry Geophysics Geosystems 3 doi1010292001GC000256

Vielzeuf D amp Holloway J R (1988) Experimental determination ofthe fluid-absent melting relations in the pelitic system Contributionsto Mineralogy and Petrology 98 257^276

Yardley BW D (1989) An Introduction to Metamorphic Petrology HarlowLongman

Ye K Cong B amp Ye D (2000) The possible subduction ofcontinental material to depths greater than 200 km Nature 407734^736

Zhang RY Liou J G ErnstW G Coleman R G Sobolev NVamp ShatskyV S (1997) Metamorphic evolution of diamond-bearingand associated rocks from the Kokchetav Massif NorthernKazakhstan Journal of Metamorphic Geology 15 479^496


740



ownloaded from

in subducted sediments that experienced pressures in therange of 25^5GPaOne approach to study such phase relations is through

investigation of ultrahigh-pressure (UHP) metamorphicrocks (Hermann et al 2006) Subducted continental crustoccasionally contains relics of coesite and diamond pro-viding evidence that these rocks were metamorphosedat 100^150 km depth (Chopin 1984 Sobolev amp Shatsky1990) Metapelites metagranites and metacarbonates insuch terrains can be used as natural laboratories to studyphase and melting relations in subducted sediments(Philippot et al 1995 Stolaquo ckhert et al 2001 Hermannet al 2006) Bulk-rock and mineral trace element analysesof such rocks have shown that phengite plays a key role as arepository for LILE (Sorenson et al 1997 Spandler et al2003) However one major problem with the interpretationof such data is that deeply subducted rocks cool relativelyslowly during exhumation and hence it is not clear whetherthe original chemical signature of phases is preservedMoreover in these rocks the record of the fluid phase isat best only fragmentarily preserved and so little infor-mation on the composition and nature of the fluid phasepresent at UHP conditions can be obtainedHigh-pressure experiments on sedimentary bulk compo-

sitions provide a second way of investigating elementrecycling in subduction zones Experimental studies areimportant for understanding mineral relations in sub-ducted sediments and for directly constraining the compo-sition and nature of the fluid phase Experimental studieshave confirmed that phengite is the stable K-phase insubducted sediments at sub-arc depths (Domanik ampHolloway 1996 Schmidt 1996 Ono 1998 Hermann2002b Auzanneau et al 2006) Consequently an under-standing of phengite melting relations is crucial forunderstanding the recycling of LILE One of the problemswith experimental studies is that piston cylinder techni-ques are used at pressures 35GPa whereas multi-anvilexperiments are used at pressures 35GPa (Schmidtet al 2004) This might introduce some systematic errorsexactly in the critical pressure range where the subduc-tion-zone fluids leave the slabThere are contrasting views on the nature and composi-

tion of the fluid phase that transports the key elementsfrom the slab to the mantle wedge Geochemical studiesof arc magmas have attempted to constrain the natureof the fluid phase on the basis of element ratios Twoend-members have been proposed corresponding to a sedi-ment fluid component and a sediment melt component(Class et al 2000 Elliott 2003) However it must bestressed that the actual phase relations and bufferingphases in the slab are often ignored A second approach toconstrain the nature and composition of the fluid phaseinvolves detailed solubility studies of minerals in fluidsand calculation of fluid compositions in equilibrium with

the determined high-pressure phases (Manning 19982004) Such an approach has provided excellent results formajor element concentrations in the aqueous fluid Toestablish whether an aqueous fluid or a hydrous melt isleaving the subducted slab it is necessary to combineexpected top-slab temperatures with experimental studieson the solidus of sedimentary rocks (Nichols et al 1994Hermann et al 2006) To date this approach has beeninconclusive because there exists a large variation in pre-dicted top-slab temperatures according to differentmodels as well as discrepancies in the position of the wetsolidus for typical subducted sediments A further compli-cation arises as a consequence of the potential presenceof the second critical endpoint (Kessel et al 2005b) In thepelitic system the termination of the wet solidus at thesecond critical endpoint might be reached at pressure andtemperature conditions relevant for subducted sedimentsat sub-arc depth and hence transitional supercriticalfluids could be the transporting agent (Manning 2004Kessel et al 2005a Hermann et al 2006)In this paper we present the results of an experimental

study on the phase and melting relations in a pelite compo-sition We used piston cylinder experiments for the entiredataset covering a P^T range of 25^45GPa and600^10508C appropriate for sub-arc conditions This elim-inates any potential systematic errors involved in usingtwo different experimental techniques We systematicallyinvestigated the stability and melting behaviour of phen-gite to constrain the recycling of K2O from subducted sedi-ments Fluid-present melting is explored and specialemphasis is placed on constraining the position of the wetsolidus We show that the compositions of aqueous fluidsand hydrous melts vary as a function of temperatureand we use this information to postulate that sedimentmelts provide the simplest way to explain the K2O andH2O systematics of primitive arc lavas We also reporthow the common UHP phases garnet omphacite andphengite change their composition as a function of pres-sure and temperature and how this information can beused as a guide to evaluate how well UHP rocks preserveoriginal phase compositions

EXPER IMENTAL METHODSStarting materialA starting material has been synthesized with a major-element composition similar to average global subductingsediment (GLOSS Plank amp Langmuir1998) and the aver-age upper continental crust (Fig 1 Taylor amp McLennan1985) The starting material was produced using a lsquosol^gelrsquomethod to eliminate problems of sluggish reaction ofrefractory minerals during experiments Most majorelements and some trace elements (Spandler et al 2007b)were combined as nitrate solutions and mixed withtetraethyl orthosilicate [Si(C2H5O)4] and slowly dried


718



ownloaded from




Diamond traps


Al2O3

K2O MgO+FeO

SiO210

Na2O K2O




719



ownloaded from






























720



ownloaded from







Exp EPSM EPSM 2246 1846 1928 D29 2247 1848 1578 1699 1867 1872 1614 1563 1868

P (GPa) 25 25 25 3 35 35 35 35 35 45 45 45 45

T (8C) 750 800 900 900 750 800 900 900 950 800 900 1000 1050

SiO2 6883 6415 7343 7459 7132 7272 7487 7377 7357 7318 7224 7272 7231 7066 6872

TiO2 067 063 021 017 025 031 027 031 035 044 043 046 055 060 103

Al2O3 1470 1370 1447 1490 1726 1532 1356 1378 1443 1485 1502 1232 1363 1521 1491

FeO 467 435 089 078 065 081 062 088 088 077 112 093 094 104 129

MnO 011 010 5005 5005 5005 5005 5005 5005 5005 5005 006 5005 008 5005 5005

MgO 251 234 027 050 052 055 027 043 040 029 023 031 061 035 040

CaO 245 228 117 159 094 125 110 142 119 106 084 186 162 079 081

Na2O 262 244 693 416 377 509 623 596 376 344 294 687 365 316 228

K2O 294 274 261 339 509 406 308 329 519 576 701 445 637 793 1005

P2O5 032 030 5020 5020 034 5020 5020 022 025 026 010 5020 029 026 049

Sum 998 9301 9997 10008 10014 10010 9999 10007 10001 10006 10000 9991 10005 10000 9998

Meas 8528 8476 9381 8679 7903 8088 8600 8561 8618 7032 8328 9096 8889

melt 35 60 40 25 23 30 55 20 26 21 40 35 30

H2O wt 68 16 10 5 10 25 20 11 11 6 29 16 8 6

K2OH2O 014 030 097 037 009 013 042 047 110 011 033 091 157

XMg 035 053 059 055 043 046 045 040 027 037 053 037 036

ASI 089 112 129 102 087 086 103 107 108 063 086 100 093

Ab 75 57 49 60 70 67 48 44 37 63 42 36 24

An 7 12 7 8 7 9 8 7 6 9 10 5 5

Or 18 31 44 32 23 24 44 48 58 27 48 59 71



721



ownloaded from




(a) (b) (c)

(d) (e) (f)

(g) (h) (i)



722



ownloaded from


Solid phases




Fluid phases





35

45

P (

GP

a)

600 700 800 900 1000T (degC)

25

Grt Cpx CoePhe Gl


Phe F

Ctd Cpx CoePhe F

Am QtzPhe F


Qtz Kfsp Gl

non

quen

chab

le fl

uid

hydr

ous

glas

s

CoeQtz

Phe

ngite

Garnet

Amphibole

Biotite



723



ownloaded from



Garnet






Clinopyroxene


Gro

Alm Py

60 408010

10

20

30

10

30

30

50

50

700 800900

1000

0

002

004

006

008

01

Na

(pfu

)

0 001 002 003 004 005P (pfu)

253545

700ndash800

800ndash900

900ndash1000

Garnet

(a) (b)



724



ownloaded from



Phengite



04

05

06

07

08

Na

(pfu

)

500 600 700 800 900 1000 1100

T (degC)

45

35

25

3

0

005

01

015

02

Exc

ess

Al (

pfu

)

392 394 396 398 4high T

low T

(a) (b)

Clinopyroxene



725



ownloaded from


substitution

Other minerals






31

32

33

34

35

36

500 600 700 800 900 1000 1100

45

35

25

(a) (b)

Si (

pfu

)

T (degC)500 600 700 800 900 1000 1100

T (degC)

059

051

063

066

061

042

096

072

042

145

089

054

068

187

176

177

107

191

199

25

35

45

P (

GP

a)

Phe

out

05 075 10 15 20TiO2

Phengite



726



ownloaded from

2

4

6

8

10

SiO

2A

l 2O

3

500 600 700 800 900 1000 1100

T (degC)

500 600 700 800 900 1000 1100

T (degC)

800degC 45 GPa

0

05

1

15

Na 2

OA

l 2O

3

aqueous fluid

hydrous melt

aqueous fluid

hydrous melt

01

1

10

K2O

Na 2

O

0

25

5

75

10

125

K2O

500 600 700 800 900 1000 1100

T (degC)

500 600 700 800 900 1000 1100

T (degC)

An

Ab Or

tonalite

trondhjemite

granodiorite

granite

(95)

(92)

20

(89)

(85)

16

25

10

20

29

5

10

1111

16

6

8 6

25

35

45

P (

GP

a)

600 700 800 900 1000 1100

T (degC)

51015202530

(a) (b)

(c)(d)

(f)(e)


aqueous fluid

hydrous melt

wet

sol

idus

H2O



727



ownloaded from


Diamond traps










Exp 2099 2098 2082 2083 2097

P (GPa) 25 25 35 35 45

T (8C) 700 600 600 700 700

SiO2 7247 7538 6902 7505 8847

TiO2 024 032 042 027 022

Al2O3 1478 1021 880 774 261

FeO 183 187 293 309 027

MnO 004 008 014 026 001

MgO 060 247 300 180 393

CaO 184 097 168 183 061

Na2O 578 786 1255 848 320

K2O 209 083 146 147 069

Sum 100 100 100 100 100

H2O 20 95 92 89 85

K2OH2O 0084 00004 00013 00018 00012

XMg 037 070 065 051 096

ASI 098 065 035 041 037

Alb 71 88 87 81 80

An 12 6 6 10 8

Or 17 6 7 9 11



728



ownloaded from




Garnet




05

1

15

2

25

3

Ln

KD

(G

rt-P

he)

75 8 85 9 95 10 1051T (degK) x 10000

45 GPa1

2

3

4

5

Na 2

O C

px

Mel

t

700 800 900 1000 1100T (˚C)


T (degC)

7008009001000

T (degC)

GH82

(a) (b) (c)



minus30

minus20

minus10

0

10

20

30

(Pca

lc -

Pex

p)

Pex

p x

100

600 700 800 900 1000 1100



729



ownloaded from






Phengite



730



ownloaded from










731



ownloaded from







2

4

600 800 1000

T (degC)

600 800 1000

T (degC)

N94

S04

A06

J99th

is s

tudy

biotite out

phen

gite

out

1

3

5

hydr

ous

glas

s

2

4

P (

GP

a)

P (

GP

a)

N94

S04

this

stu

dy

1

3

551015202530

8

N94

hydrous melt

(a) (b)

H2O



732



ownloaded from







733



ownloaded from




K2O

H2O

600 800 1000

2535

45

aqueous fluid


lavas


2

4

400 600 800 1000

T (degC)

P (

GP

a)

1) 2) 3) 4) 5)

N94

this

stu

dy

phen

gite

out

1

3

5

hydr

ous

glas

s

T (degC)


hydrous melt

aqueous fluid



0001

001

01

1

500 700 900

(a) (b)



734



ownloaded from









735



ownloaded from









736



ownloaded from











737



ownloaded from

















































738



ownloaded from















































739



ownloaded from




















740



ownloaded from




Diamond traps


Al2O3

K2O MgO+FeO

SiO210

Na2O K2O




719



ownloaded from






























720



ownloaded from







Exp EPSM EPSM 2246 1846 1928 D29 2247 1848 1578 1699 1867 1872 1614 1563 1868

P (GPa) 25 25 25 3 35 35 35 35 35 45 45 45 45

T (8C) 750 800 900 900 750 800 900 900 950 800 900 1000 1050

SiO2 6883 6415 7343 7459 7132 7272 7487 7377 7357 7318 7224 7272 7231 7066 6872

TiO2 067 063 021 017 025 031 027 031 035 044 043 046 055 060 103

Al2O3 1470 1370 1447 1490 1726 1532 1356 1378 1443 1485 1502 1232 1363 1521 1491

FeO 467 435 089 078 065 081 062 088 088 077 112 093 094 104 129

MnO 011 010 5005 5005 5005 5005 5005 5005 5005 5005 006 5005 008 5005 5005

MgO 251 234 027 050 052 055 027 043 040 029 023 031 061 035 040

CaO 245 228 117 159 094 125 110 142 119 106 084 186 162 079 081

Na2O 262 244 693 416 377 509 623 596 376 344 294 687 365 316 228

K2O 294 274 261 339 509 406 308 329 519 576 701 445 637 793 1005

P2O5 032 030 5020 5020 034 5020 5020 022 025 026 010 5020 029 026 049

Sum 998 9301 9997 10008 10014 10010 9999 10007 10001 10006 10000 9991 10005 10000 9998

Meas 8528 8476 9381 8679 7903 8088 8600 8561 8618 7032 8328 9096 8889

melt 35 60 40 25 23 30 55 20 26 21 40 35 30

H2O wt 68 16 10 5 10 25 20 11 11 6 29 16 8 6

K2OH2O 014 030 097 037 009 013 042 047 110 011 033 091 157

XMg 035 053 059 055 043 046 045 040 027 037 053 037 036

ASI 089 112 129 102 087 086 103 107 108 063 086 100 093

Ab 75 57 49 60 70 67 48 44 37 63 42 36 24

An 7 12 7 8 7 9 8 7 6 9 10 5 5

Or 18 31 44 32 23 24 44 48 58 27 48 59 71



721



ownloaded from




(a) (b) (c)

(d) (e) (f)

(g) (h) (i)



722



ownloaded from


Solid phases




Fluid phases





35

45

P (

GP

a)

600 700 800 900 1000T (degC)

25

Grt Cpx CoePhe Gl


Phe F

Ctd Cpx CoePhe F

Am QtzPhe F


Qtz Kfsp Gl

non

quen

chab

le fl

uid

hydr

ous

glas

s

CoeQtz

Phe

ngite

Garnet

Amphibole

Biotite



723



ownloaded from



Garnet






Clinopyroxene


Gro

Alm Py

60 408010

10

20

30

10

30

30

50

50

700 800900

1000

0

002

004

006

008

01

Na

(pfu

)

0 001 002 003 004 005P (pfu)

253545

700ndash800

800ndash900

900ndash1000

Garnet

(a) (b)



724



ownloaded from



Phengite



04

05

06

07

08

Na

(pfu

)

500 600 700 800 900 1000 1100

T (degC)

45

35

25

3

0

005

01

015

02

Exc

ess

Al (

pfu

)

392 394 396 398 4high T

low T

(a) (b)

Clinopyroxene



725



ownloaded from


substitution

Other minerals






31

32

33

34

35

36

500 600 700 800 900 1000 1100

45

35

25

(a) (b)

Si (

pfu

)

T (degC)500 600 700 800 900 1000 1100

T (degC)

059

051

063

066

061

042

096

072

042

145

089

054

068

187

176

177

107

191

199

25

35

45

P (

GP

a)

Phe

out

05 075 10 15 20TiO2

Phengite



726



ownloaded from

2

4

6

8

10

SiO

2A

l 2O

3

500 600 700 800 900 1000 1100

T (degC)

500 600 700 800 900 1000 1100

T (degC)

800degC 45 GPa

0

05

1

15

Na 2

OA

l 2O

3

aqueous fluid

hydrous melt

aqueous fluid

hydrous melt

01

1

10

K2O

Na 2

O

0

25

5

75

10

125

K2O

500 600 700 800 900 1000 1100

T (degC)

500 600 700 800 900 1000 1100

T (degC)

An

Ab Or

tonalite

trondhjemite

granodiorite

granite

(95)

(92)

20

(89)

(85)

16

25

10

20

29

5

10

1111

16

6

8 6

25

35

45

P (

GP

a)

600 700 800 900 1000 1100

T (degC)

51015202530

(a) (b)

(c)(d)

(f)(e)


aqueous fluid

hydrous melt

wet

sol

idus

H2O



727



ownloaded from


Diamond traps










Exp 2099 2098 2082 2083 2097

P (GPa) 25 25 35 35 45

T (8C) 700 600 600 700 700

SiO2 7247 7538 6902 7505 8847

TiO2 024 032 042 027 022

Al2O3 1478 1021 880 774 261

FeO 183 187 293 309 027

MnO 004 008 014 026 001

MgO 060 247 300 180 393

CaO 184 097 168 183 061

Na2O 578 786 1255 848 320

K2O 209 083 146 147 069

Sum 100 100 100 100 100

H2O 20 95 92 89 85

K2OH2O 0084 00004 00013 00018 00012

XMg 037 070 065 051 096

ASI 098 065 035 041 037

Alb 71 88 87 81 80

An 12 6 6 10 8

Or 17 6 7 9 11



728



ownloaded from




Garnet




05

1

15

2

25

3

Ln

KD

(G

rt-P

he)

75 8 85 9 95 10 1051T (degK) x 10000

45 GPa1

2

3

4

5

Na 2

O C

px

Mel

t

700 800 900 1000 1100T (˚C)


T (degC)

7008009001000

T (degC)

GH82

(a) (b) (c)



minus30

minus20

minus10

0

10

20

30

(Pca

lc -

Pex

p)

Pex

p x

100

600 700 800 900 1000 1100



729



ownloaded from






Phengite



730



ownloaded from










731



ownloaded from







2

4

600 800 1000

T (degC)

600 800 1000

T (degC)

N94

S04

A06

J99th

is s

tudy

biotite out

phen

gite

out

1

3

5

hydr

ous

glas

s

2

4

P (

GP

a)

P (

GP

a)

N94

S04

this

stu

dy

1

3

551015202530

8

N94

hydrous melt

(a) (b)

H2O



732



ownloaded from







733



ownloaded from




K2O

H2O

600 800 1000

2535

45

aqueous fluid


lavas


2

4

400 600 800 1000

T (degC)

P (

GP

a)

1) 2) 3) 4) 5)

N94

this

stu

dy

phen

gite

out

1

3

5

hydr

ous

glas

s

T (degC)


hydrous melt

aqueous fluid



0001

001

01

1

500 700 900

(a) (b)



734



ownloaded from









735



ownloaded from









736



ownloaded from











737



ownloaded from

















































738



ownloaded from















































739



ownloaded from




















740



ownloaded from






























720



ownloaded from







Exp EPSM EPSM 2246 1846 1928 D29 2247 1848 1578 1699 1867 1872 1614 1563 1868

P (GPa) 25 25 25 3 35 35 35 35 35 45 45 45 45

T (8C) 750 800 900 900 750 800 900 900 950 800 900 1000 1050

SiO2 6883 6415 7343 7459 7132 7272 7487 7377 7357 7318 7224 7272 7231 7066 6872

TiO2 067 063 021 017 025 031 027 031 035 044 043 046 055 060 103

Al2O3 1470 1370 1447 1490 1726 1532 1356 1378 1443 1485 1502 1232 1363 1521 1491

FeO 467 435 089 078 065 081 062 088 088 077 112 093 094 104 129

MnO 011 010 5005 5005 5005 5005 5005 5005 5005 5005 006 5005 008 5005 5005

MgO 251 234 027 050 052 055 027 043 040 029 023 031 061 035 040

CaO 245 228 117 159 094 125 110 142 119 106 084 186 162 079 081

Na2O 262 244 693 416 377 509 623 596 376 344 294 687 365 316 228

K2O 294 274 261 339 509 406 308 329 519 576 701 445 637 793 1005

P2O5 032 030 5020 5020 034 5020 5020 022 025 026 010 5020 029 026 049

Sum 998 9301 9997 10008 10014 10010 9999 10007 10001 10006 10000 9991 10005 10000 9998

Meas 8528 8476 9381 8679 7903 8088 8600 8561 8618 7032 8328 9096 8889

melt 35 60 40 25 23 30 55 20 26 21 40 35 30

H2O wt 68 16 10 5 10 25 20 11 11 6 29 16 8 6

K2OH2O 014 030 097 037 009 013 042 047 110 011 033 091 157

XMg 035 053 059 055 043 046 045 040 027 037 053 037 036

ASI 089 112 129 102 087 086 103 107 108 063 086 100 093

Ab 75 57 49 60 70 67 48 44 37 63 42 36 24

An 7 12 7 8 7 9 8 7 6 9 10 5 5

Or 18 31 44 32 23 24 44 48 58 27 48 59 71



721



ownloaded from




(a) (b) (c)

(d) (e) (f)

(g) (h) (i)



722



ownloaded from


Solid phases




Fluid phases





35

45

P (

GP

a)

600 700 800 900 1000T (degC)

25

Grt Cpx CoePhe Gl


Phe F

Ctd Cpx CoePhe F

Am QtzPhe F


Qtz Kfsp Gl

non

quen

chab

le fl

uid

hydr

ous

glas

s

CoeQtz

Phe

ngite

Garnet

Amphibole

Biotite



723



ownloaded from



Garnet






Clinopyroxene


Gro

Alm Py

60 408010

10

20

30

10

30

30

50

50

700 800900

1000

0

002

004

006

008

01

Na

(pfu

)

0 001 002 003 004 005P (pfu)

253545

700ndash800

800ndash900

900ndash1000

Garnet

(a) (b)



724



ownloaded from



Phengite



04

05

06

07

08

Na

(pfu

)

500 600 700 800 900 1000 1100

T (degC)

45

35

25

3

0

005

01

015

02

Exc

ess

Al (

pfu

)

392 394 396 398 4high T

low T

(a) (b)

Clinopyroxene



725



ownloaded from


substitution

Other minerals






31

32

33

34

35

36

500 600 700 800 900 1000 1100

45

35

25

(a) (b)

Si (

pfu

)

T (degC)500 600 700 800 900 1000 1100

T (degC)

059

051

063

066

061

042

096

072

042

145

089

054

068

187

176

177

107

191

199

25

35

45

P (

GP

a)

Phe

out

05 075 10 15 20TiO2

Phengite



726



ownloaded from

2

4

6

8

10

SiO

2A

l 2O

3

500 600 700 800 900 1000 1100

T (degC)

500 600 700 800 900 1000 1100

T (degC)

800degC 45 GPa

0

05

1

15

Na 2

OA

l 2O

3

aqueous fluid

hydrous melt

aqueous fluid

hydrous melt

01

1

10

K2O

Na 2

O

0

25

5

75

10

125

K2O

500 600 700 800 900 1000 1100

T (degC)

500 600 700 800 900 1000 1100

T (degC)

An

Ab Or

tonalite

trondhjemite

granodiorite

granite

(95)

(92)

20

(89)

(85)

16

25

10

20

29

5

10

1111

16

6

8 6

25

35

45

P (

GP

a)

600 700 800 900 1000 1100

T (degC)

51015202530

(a) (b)

(c)(d)

(f)(e)


aqueous fluid

hydrous melt

wet

sol

idus

H2O



727



ownloaded from


Diamond traps










Exp 2099 2098 2082 2083 2097

P (GPa) 25 25 35 35 45

T (8C) 700 600 600 700 700

SiO2 7247 7538 6902 7505 8847

TiO2 024 032 042 027 022

Al2O3 1478 1021 880 774 261

FeO 183 187 293 309 027

MnO 004 008 014 026 001

MgO 060 247 300 180 393

CaO 184 097 168 183 061

Na2O 578 786 1255 848 320

K2O 209 083 146 147 069

Sum 100 100 100 100 100

H2O 20 95 92 89 85

K2OH2O 0084 00004 00013 00018 00012

XMg 037 070 065 051 096

ASI 098 065 035 041 037

Alb 71 88 87 81 80

An 12 6 6 10 8

Or 17 6 7 9 11



728



ownloaded from




Garnet




05

1

15

2

25

3

Ln

KD

(G

rt-P

he)

75 8 85 9 95 10 1051T (degK) x 10000

45 GPa1

2

3

4

5

Na 2

O C

px

Mel

t

700 800 900 1000 1100T (˚C)


T (degC)

7008009001000

T (degC)

GH82

(a) (b) (c)



minus30

minus20

minus10

0

10

20

30

(Pca

lc -

Pex

p)

Pex

p x

100

600 700 800 900 1000 1100



729



ownloaded from






Phengite



730



ownloaded from










731



ownloaded from







2

4

600 800 1000

T (degC)

600 800 1000

T (degC)

N94

S04

A06

J99th

is s

tudy

biotite out

phen

gite

out

1

3

5

hydr

ous

glas

s

2

4

P (

GP

a)

P (

GP

a)

N94

S04

this

stu

dy

1

3

551015202530

8

N94

hydrous melt

(a) (b)

H2O



732



ownloaded from







733



ownloaded from




K2O

H2O

600 800 1000

2535

45

aqueous fluid


lavas


2

4

400 600 800 1000

T (degC)

P (

GP

a)

1) 2) 3) 4) 5)

N94

this

stu

dy

phen

gite

out

1

3

5

hydr

ous

glas

s

T (degC)


hydrous melt

aqueous fluid



0001

001

01

1

500 700 900

(a) (b)



734



ownloaded from









735



ownloaded from









736



ownloaded from











737



ownloaded from

















































738



ownloaded from















































739



ownloaded from




















740



ownloaded from







Exp EPSM EPSM 2246 1846 1928 D29 2247 1848 1578 1699 1867 1872 1614 1563 1868

P (GPa) 25 25 25 3 35 35 35 35 35 45 45 45 45

T (8C) 750 800 900 900 750 800 900 900 950 800 900 1000 1050

SiO2 6883 6415 7343 7459 7132 7272 7487 7377 7357 7318 7224 7272 7231 7066 6872

TiO2 067 063 021 017 025 031 027 031 035 044 043 046 055 060 103

Al2O3 1470 1370 1447 1490 1726 1532 1356 1378 1443 1485 1502 1232 1363 1521 1491

FeO 467 435 089 078 065 081 062 088 088 077 112 093 094 104 129

MnO 011 010 5005 5005 5005 5005 5005 5005 5005 5005 006 5005 008 5005 5005

MgO 251 234 027 050 052 055 027 043 040 029 023 031 061 035 040

CaO 245 228 117 159 094 125 110 142 119 106 084 186 162 079 081

Na2O 262 244 693 416 377 509 623 596 376 344 294 687 365 316 228

K2O 294 274 261 339 509 406 308 329 519 576 701 445 637 793 1005

P2O5 032 030 5020 5020 034 5020 5020 022 025 026 010 5020 029 026 049

Sum 998 9301 9997 10008 10014 10010 9999 10007 10001 10006 10000 9991 10005 10000 9998

Meas 8528 8476 9381 8679 7903 8088 8600 8561 8618 7032 8328 9096 8889

melt 35 60 40 25 23 30 55 20 26 21 40 35 30

H2O wt 68 16 10 5 10 25 20 11 11 6 29 16 8 6

K2OH2O 014 030 097 037 009 013 042 047 110 011 033 091 157

XMg 035 053 059 055 043 046 045 040 027 037 053 037 036

ASI 089 112 129 102 087 086 103 107 108 063 086 100 093

Ab 75 57 49 60 70 67 48 44 37 63 42 36 24

An 7 12 7 8 7 9 8 7 6 9 10 5 5

Or 18 31 44 32 23 24 44 48 58 27 48 59 71



721



ownloaded from




(a) (b) (c)

(d) (e) (f)

(g) (h) (i)



722



ownloaded from


Solid phases




Fluid phases





35

45

P (

GP

a)

600 700 800 900 1000T (degC)

25

Grt Cpx CoePhe Gl


Phe F

Ctd Cpx CoePhe F

Am QtzPhe F


Qtz Kfsp Gl

non

quen

chab

le fl

uid

hydr

ous

glas

s

CoeQtz

Phe

ngite

Garnet

Amphibole

Biotite



723



ownloaded from



Garnet






Clinopyroxene


Gro

Alm Py

60 408010

10

20

30

10

30

30

50

50

700 800900

1000

0

002

004

006

008

01

Na

(pfu

)

0 001 002 003 004 005P (pfu)

253545

700ndash800

800ndash900

900ndash1000

Garnet

(a) (b)



724



ownloaded from



Phengite



04

05

06

07

08

Na

(pfu

)

500 600 700 800 900 1000 1100

T (degC)

45

35

25

3

0

005

01

015

02

Exc

ess

Al (

pfu

)

392 394 396 398 4high T

low T

(a) (b)

Clinopyroxene



725



ownloaded from


substitution

Other minerals






31

32

33

34

35

36

500 600 700 800 900 1000 1100

45

35

25

(a) (b)

Si (

pfu

)

T (degC)500 600 700 800 900 1000 1100

T (degC)

059

051

063

066

061

042

096

072

042

145

089

054

068

187

176

177

107

191

199

25

35

45

P (

GP

a)

Phe

out

05 075 10 15 20TiO2

Phengite



726



ownloaded from

2

4

6

8

10

SiO

2A

l 2O

3

500 600 700 800 900 1000 1100

T (degC)

500 600 700 800 900 1000 1100

T (degC)

800degC 45 GPa

0

05

1

15

Na 2

OA

l 2O

3

aqueous fluid

hydrous melt

aqueous fluid

hydrous melt

01

1

10

K2O

Na 2

O

0

25

5

75

10

125

K2O

500 600 700 800 900 1000 1100

T (degC)

500 600 700 800 900 1000 1100

T (degC)

An

Ab Or

tonalite

trondhjemite

granodiorite

granite

(95)

(92)

20

(89)

(85)

16

25

10

20

29

5

10

1111

16

6

8 6

25

35

45

P (

GP

a)

600 700 800 900 1000 1100

T (degC)

51015202530

(a) (b)

(c)(d)

(f)(e)


aqueous fluid

hydrous melt

wet

sol

idus

H2O



727



ownloaded from


Diamond traps










Exp 2099 2098 2082 2083 2097

P (GPa) 25 25 35 35 45

T (8C) 700 600 600 700 700

SiO2 7247 7538 6902 7505 8847

TiO2 024 032 042 027 022

Al2O3 1478 1021 880 774 261

FeO 183 187 293 309 027

MnO 004 008 014 026 001

MgO 060 247 300 180 393

CaO 184 097 168 183 061

Na2O 578 786 1255 848 320

K2O 209 083 146 147 069

Sum 100 100 100 100 100

H2O 20 95 92 89 85

K2OH2O 0084 00004 00013 00018 00012

XMg 037 070 065 051 096

ASI 098 065 035 041 037

Alb 71 88 87 81 80

An 12 6 6 10 8

Or 17 6 7 9 11



728



ownloaded from




Garnet




05

1

15

2

25

3

Ln

KD

(G

rt-P

he)

75 8 85 9 95 10 1051T (degK) x 10000

45 GPa1

2

3

4

5

Na 2

O C

px

Mel

t

700 800 900 1000 1100T (˚C)


T (degC)

7008009001000

T (degC)

GH82

(a) (b) (c)



minus30

minus20

minus10

0

10

20

30

(Pca

lc -

Pex

p)

Pex

p x

100

600 700 800 900 1000 1100



729



ownloaded from






Phengite



730



ownloaded from










731



ownloaded from







2

4

600 800 1000

T (degC)

600 800 1000

T (degC)

N94

S04

A06

J99th

is s

tudy

biotite out

phen

gite

out

1

3

5

hydr

ous

glas

s

2

4

P (

GP

a)

P (

GP

a)

N94

S04

this

stu

dy

1

3

551015202530

8

N94

hydrous melt

(a) (b)

H2O



732



ownloaded from







733



ownloaded from




K2O

H2O

600 800 1000

2535

45

aqueous fluid


lavas


2

4

400 600 800 1000

T (degC)

P (

GP

a)

1) 2) 3) 4) 5)

N94

this

stu

dy

phen

gite

out

1

3

5

hydr

ous

glas

s

T (degC)


hydrous melt

aqueous fluid



0001

001

01

1

500 700 900

(a) (b)



734



ownloaded from









735



ownloaded from









736



ownloaded from











737



ownloaded from

















































738



ownloaded from















































739



ownloaded from




















740



ownloaded from




(a) (b) (c)

(d) (e) (f)

(g) (h) (i)



722



ownloaded from


Solid phases




Fluid phases





35

45

P (

GP

a)

600 700 800 900 1000T (degC)

25

Grt Cpx CoePhe Gl


Phe F

Ctd Cpx CoePhe F

Am QtzPhe F


Qtz Kfsp Gl

non

quen

chab

le fl

uid

hydr

ous

glas

s

CoeQtz

Phe

ngite

Garnet

Amphibole

Biotite



723



ownloaded from



Garnet






Clinopyroxene


Gro

Alm Py

60 408010

10

20

30

10

30

30

50

50

700 800900

1000

0

002

004

006

008

01

Na

(pfu

)

0 001 002 003 004 005P (pfu)

253545

700ndash800

800ndash900

900ndash1000

Garnet

(a) (b)



724



ownloaded from



Phengite



04

05

06

07

08

Na

(pfu

)

500 600 700 800 900 1000 1100

T (degC)

45

35

25

3

0

005

01

015

02

Exc

ess

Al (

pfu

)

392 394 396 398 4high T

low T

(a) (b)

Clinopyroxene



725



ownloaded from


substitution

Other minerals






31

32

33

34

35

36

500 600 700 800 900 1000 1100

45

35

25

(a) (b)

Si (

pfu

)

T (degC)500 600 700 800 900 1000 1100

T (degC)

059

051

063

066

061

042

096

072

042

145

089

054

068

187

176

177

107

191

199

25

35

45

P (

GP

a)

Phe

out

05 075 10 15 20TiO2

Phengite



726



ownloaded from

2

4

6

8

10

SiO

2A

l 2O

3

500 600 700 800 900 1000 1100

T (degC)

500 600 700 800 900 1000 1100

T (degC)

800degC 45 GPa

0

05

1

15

Na 2

OA

l 2O

3

aqueous fluid

hydrous melt

aqueous fluid

hydrous melt

01

1

10

K2O

Na 2

O

0

25

5

75

10

125

K2O

500 600 700 800 900 1000 1100

T (degC)

500 600 700 800 900 1000 1100

T (degC)

An

Ab Or

tonalite

trondhjemite

granodiorite

granite

(95)

(92)

20

(89)

(85)

16

25

10

20

29

5

10

1111

16

6

8 6

25

35

45

P (

GP

a)

600 700 800 900 1000 1100

T (degC)

51015202530

(a) (b)

(c)(d)

(f)(e)


aqueous fluid

hydrous melt

wet

sol

idus

H2O



727



ownloaded from


Diamond traps










Exp 2099 2098 2082 2083 2097

P (GPa) 25 25 35 35 45

T (8C) 700 600 600 700 700

SiO2 7247 7538 6902 7505 8847

TiO2 024 032 042 027 022

Al2O3 1478 1021 880 774 261

FeO 183 187 293 309 027

MnO 004 008 014 026 001

MgO 060 247 300 180 393

CaO 184 097 168 183 061

Na2O 578 786 1255 848 320

K2O 209 083 146 147 069

Sum 100 100 100 100 100

H2O 20 95 92 89 85

K2OH2O 0084 00004 00013 00018 00012

XMg 037 070 065 051 096

ASI 098 065 035 041 037

Alb 71 88 87 81 80

An 12 6 6 10 8

Or 17 6 7 9 11



728



ownloaded from




Garnet




05

1

15

2

25

3

Ln

KD

(G

rt-P

he)

75 8 85 9 95 10 1051T (degK) x 10000

45 GPa1

2

3

4

5

Na 2

O C

px

Mel

t

700 800 900 1000 1100T (˚C)


T (degC)

7008009001000

T (degC)

GH82

(a) (b) (c)



minus30

minus20

minus10

0

10

20

30

(Pca

lc -

Pex

p)

Pex

p x

100

600 700 800 900 1000 1100



729



ownloaded from






Phengite



730



ownloaded from










731



ownloaded from







2

4

600 800 1000

T (degC)

600 800 1000

T (degC)

N94

S04

A06

J99th

is s

tudy

biotite out

phen

gite

out

1

3

5

hydr

ous

glas

s

2

4

P (

GP

a)

P (

GP

a)

N94

S04

this

stu

dy

1

3

551015202530

8

N94

hydrous melt

(a) (b)

H2O



732



ownloaded from







733



ownloaded from




K2O

H2O

600 800 1000

2535

45

aqueous fluid


lavas


2

4

400 600 800 1000

T (degC)

P (

GP

a)

1) 2) 3) 4) 5)

N94

this

stu

dy

phen

gite

out

1

3

5

hydr

ous

glas

s

T (degC)


hydrous melt

aqueous fluid



0001

001

01

1

500 700 900

(a) (b)



734



ownloaded from









735



ownloaded from









736



ownloaded from











737



ownloaded from

















































738



ownloaded from















































739



ownloaded from




















740



ownloaded from


Solid phases




Fluid phases





35

45

P (

GP

a)

600 700 800 900 1000T (degC)

25

Grt Cpx CoePhe Gl


Phe F

Ctd Cpx CoePhe F

Am QtzPhe F


Qtz Kfsp Gl

non

quen

chab

le fl

uid

hydr

ous

glas

s

CoeQtz

Phe

ngite

Garnet

Amphibole

Biotite



723



ownloaded from



Garnet






Clinopyroxene


Gro

Alm Py

60 408010

10

20

30

10

30

30

50

50

700 800900

1000

0

002

004

006

008

01

Na

(pfu

)

0 001 002 003 004 005P (pfu)

253545

700ndash800

800ndash900

900ndash1000

Garnet

(a) (b)



724



ownloaded from



Phengite



04

05

06

07

08

Na

(pfu

)

500 600 700 800 900 1000 1100

T (degC)

45

35

25

3

0

005

01

015

02

Exc

ess

Al (

pfu

)

392 394 396 398 4high T

low T

(a) (b)

Clinopyroxene



725



ownloaded from


substitution

Other minerals






31

32

33

34

35

36

500 600 700 800 900 1000 1100

45

35

25

(a) (b)

Si (

pfu

)

T (degC)500 600 700 800 900 1000 1100

T (degC)

059

051

063

066

061

042

096

072

042

145

089

054

068

187

176

177

107

191

199

25

35

45

P (

GP

a)

Phe

out

05 075 10 15 20TiO2

Phengite



726



ownloaded from

2

4

6

8

10

SiO

2A

l 2O

3

500 600 700 800 900 1000 1100

T (degC)

500 600 700 800 900 1000 1100

T (degC)

800degC 45 GPa

0

05

1

15

Na 2

OA

l 2O

3

aqueous fluid

hydrous melt

aqueous fluid

hydrous melt

01

1

10

K2O

Na 2

O

0

25

5

75

10

125

K2O

500 600 700 800 900 1000 1100

T (degC)

500 600 700 800 900 1000 1100

T (degC)

An

Ab Or

tonalite

trondhjemite

granodiorite

granite

(95)

(92)

20

(89)

(85)

16

25

10

20

29

5

10

1111

16

6

8 6

25

35

45

P (

GP

a)

600 700 800 900 1000 1100

T (degC)

51015202530

(a) (b)

(c)(d)

(f)(e)


aqueous fluid

hydrous melt

wet

sol

idus

H2O



727



ownloaded from


Diamond traps










Exp 2099 2098 2082 2083 2097

P (GPa) 25 25 35 35 45

T (8C) 700 600 600 700 700

SiO2 7247 7538 6902 7505 8847

TiO2 024 032 042 027 022

Al2O3 1478 1021 880 774 261

FeO 183 187 293 309 027

MnO 004 008 014 026 001

MgO 060 247 300 180 393

CaO 184 097 168 183 061

Na2O 578 786 1255 848 320

K2O 209 083 146 147 069

Sum 100 100 100 100 100

H2O 20 95 92 89 85

K2OH2O 0084 00004 00013 00018 00012

XMg 037 070 065 051 096

ASI 098 065 035 041 037

Alb 71 88 87 81 80

An 12 6 6 10 8

Or 17 6 7 9 11



728



ownloaded from




Garnet




05

1

15

2

25

3

Ln

KD

(G

rt-P

he)

75 8 85 9 95 10 1051T (degK) x 10000

45 GPa1

2

3

4

5

Na 2

O C

px

Mel

t

700 800 900 1000 1100T (˚C)


T (degC)

7008009001000

T (degC)

GH82

(a) (b) (c)



minus30

minus20

minus10

0

10

20

30

(Pca

lc -

Pex

p)

Pex

p x

100

600 700 800 900 1000 1100



729



ownloaded from






Phengite



730



ownloaded from










731



ownloaded from







2

4

600 800 1000

T (degC)

600 800 1000

T (degC)

N94

S04

A06

J99th

is s

tudy

biotite out

phen

gite

out

1

3

5

hydr

ous

glas

s

2

4

P (

GP

a)

P (

GP

a)

N94

S04

this

stu

dy

1

3

551015202530

8

N94

hydrous melt

(a) (b)

H2O



732



ownloaded from







733



ownloaded from




K2O

H2O

600 800 1000

2535

45

aqueous fluid


lavas


2

4

400 600 800 1000

T (degC)

P (

GP

a)

1) 2) 3) 4) 5)

N94

this

stu

dy

phen

gite

out

1

3

5

hydr

ous

glas

s

T (degC)


hydrous melt

aqueous fluid



0001

001

01

1

500 700 900

(a) (b)



734



ownloaded from









735



ownloaded from









736



ownloaded from











737



ownloaded from

















































738



ownloaded from















































739



ownloaded from




















740



ownloaded from



Garnet






Clinopyroxene


Gro

Alm Py

60 408010

10

20

30

10

30

30

50

50

700 800900

1000

0

002

004

006

008

01

Na

(pfu

)

0 001 002 003 004 005P (pfu)

253545

700ndash800

800ndash900

900ndash1000

Garnet

(a) (b)



724



ownloaded from



Phengite



04

05

06

07

08

Na

(pfu

)

500 600 700 800 900 1000 1100

T (degC)

45

35

25

3

0

005

01

015

02

Exc

ess

Al (

pfu

)

392 394 396 398 4high T

low T

(a) (b)

Clinopyroxene



725



ownloaded from


substitution

Other minerals






31

32

33

34

35

36

500 600 700 800 900 1000 1100

45

35

25

(a) (b)

Si (

pfu

)

T (degC)500 600 700 800 900 1000 1100

T (degC)

059

051

063

066

061

042

096

072

042

145

089

054

068

187

176

177

107

191

199

25

35

45

P (

GP

a)

Phe

out

05 075 10 15 20TiO2

Phengite



726



ownloaded from

2

4

6

8

10

SiO

2A

l 2O

3

500 600 700 800 900 1000 1100

T (degC)

500 600 700 800 900 1000 1100

T (degC)

800degC 45 GPa

0

05

1

15

Na 2

OA

l 2O

3

aqueous fluid

hydrous melt

aqueous fluid

hydrous melt

01

1

10

K2O

Na 2

O

0

25

5

75

10

125

K2O

500 600 700 800 900 1000 1100

T (degC)

500 600 700 800 900 1000 1100

T (degC)

An

Ab Or

tonalite

trondhjemite

granodiorite

granite

(95)

(92)

20

(89)

(85)

16

25

10

20

29

5

10

1111

16

6

8 6

25

35

45

P (

GP

a)

600 700 800 900 1000 1100

T (degC)

51015202530

(a) (b)

(c)(d)

(f)(e)


aqueous fluid

hydrous melt

wet

sol

idus

H2O



727



ownloaded from


Diamond traps










Exp 2099 2098 2082 2083 2097

P (GPa) 25 25 35 35 45

T (8C) 700 600 600 700 700

SiO2 7247 7538 6902 7505 8847

TiO2 024 032 042 027 022

Al2O3 1478 1021 880 774 261

FeO 183 187 293 309 027

MnO 004 008 014 026 001

MgO 060 247 300 180 393

CaO 184 097 168 183 061

Na2O 578 786 1255 848 320

K2O 209 083 146 147 069

Sum 100 100 100 100 100

H2O 20 95 92 89 85

K2OH2O 0084 00004 00013 00018 00012

XMg 037 070 065 051 096

ASI 098 065 035 041 037

Alb 71 88 87 81 80

An 12 6 6 10 8

Or 17 6 7 9 11



728



ownloaded from




Garnet




05

1

15

2

25

3

Ln

KD

(G

rt-P

he)

75 8 85 9 95 10 1051T (degK) x 10000

45 GPa1

2

3

4

5

Na 2

O C

px

Mel

t

700 800 900 1000 1100T (˚C)


T (degC)

7008009001000

T (degC)

GH82

(a) (b) (c)



minus30

minus20

minus10

0

10

20

30

(Pca

lc -

Pex

p)

Pex

p x

100

600 700 800 900 1000 1100



729



ownloaded from






Phengite



730



ownloaded from










731



ownloaded from







2

4

600 800 1000

T (degC)

600 800 1000

T (degC)

N94

S04

A06

J99th

is s

tudy

biotite out

phen

gite

out

1

3

5

hydr

ous

glas

s

2

4

P (

GP

a)

P (

GP

a)

N94

S04

this

stu

dy

1

3

551015202530

8

N94

hydrous melt

(a) (b)

H2O



732



ownloaded from







733



ownloaded from




K2O

H2O

600 800 1000

2535

45

aqueous fluid


lavas


2

4

400 600 800 1000

T (degC)

P (

GP

a)

1) 2) 3) 4) 5)

N94

this

stu

dy

phen

gite

out

1

3

5

hydr

ous

glas

s

T (degC)


hydrous melt

aqueous fluid



0001

001

01

1

500 700 900

(a) (b)



734



ownloaded from









735



ownloaded from









736



ownloaded from











737



ownloaded from

















































738



ownloaded from















































739



ownloaded from




















740



ownloaded from



Phengite



04

05

06

07

08

Na

(pfu

)

500 600 700 800 900 1000 1100

T (degC)

45

35

25

3

0

005

01

015

02

Exc

ess

Al (

pfu

)

392 394 396 398 4high T

low T

(a) (b)

Clinopyroxene



725



ownloaded from


substitution

Other minerals






31

32

33

34

35

36

500 600 700 800 900 1000 1100

45

35

25

(a) (b)

Si (

pfu

)

T (degC)500 600 700 800 900 1000 1100

T (degC)

059

051

063

066

061

042

096

072

042

145

089

054

068

187

176

177

107

191

199

25

35

45

P (

GP

a)

Phe

out

05 075 10 15 20TiO2

Phengite



726



ownloaded from

2

4

6

8

10

SiO

2A

l 2O

3

500 600 700 800 900 1000 1100

T (degC)

500 600 700 800 900 1000 1100

T (degC)

800degC 45 GPa

0

05

1

15

Na 2

OA

l 2O

3

aqueous fluid

hydrous melt

aqueous fluid

hydrous melt

01

1

10

K2O

Na 2

O

0

25

5

75

10

125

K2O

500 600 700 800 900 1000 1100

T (degC)

500 600 700 800 900 1000 1100

T (degC)

An

Ab Or

tonalite

trondhjemite

granodiorite

granite

(95)

(92)

20

(89)

(85)

16

25

10

20

29

5

10

1111

16

6

8 6

25

35

45

P (

GP

a)

600 700 800 900 1000 1100

T (degC)

51015202530

(a) (b)

(c)(d)

(f)(e)


aqueous fluid

hydrous melt

wet

sol

idus

H2O



727



ownloaded from


Diamond traps










Exp 2099 2098 2082 2083 2097

P (GPa) 25 25 35 35 45

T (8C) 700 600 600 700 700

SiO2 7247 7538 6902 7505 8847

TiO2 024 032 042 027 022

Al2O3 1478 1021 880 774 261

FeO 183 187 293 309 027

MnO 004 008 014 026 001

MgO 060 247 300 180 393

CaO 184 097 168 183 061

Na2O 578 786 1255 848 320

K2O 209 083 146 147 069

Sum 100 100 100 100 100

H2O 20 95 92 89 85

K2OH2O 0084 00004 00013 00018 00012

XMg 037 070 065 051 096

ASI 098 065 035 041 037

Alb 71 88 87 81 80

An 12 6 6 10 8

Or 17 6 7 9 11



728



ownloaded from




Garnet




05

1

15

2

25

3

Ln

KD

(G

rt-P

he)

75 8 85 9 95 10 1051T (degK) x 10000

45 GPa1

2

3

4

5

Na 2

O C

px

Mel

t

700 800 900 1000 1100T (˚C)


T (degC)

7008009001000

T (degC)

GH82

(a) (b) (c)



minus30

minus20

minus10

0

10

20

30

(Pca

lc -

Pex

p)

Pex

p x

100

600 700 800 900 1000 1100



729



ownloaded from






Phengite



730



ownloaded from










731



ownloaded from







2

4

600 800 1000

T (degC)

600 800 1000

T (degC)

N94

S04

A06

J99th

is s

tudy

biotite out

phen

gite

out

1

3

5

hydr

ous

glas

s

2

4

P (

GP

a)

P (

GP

a)

N94

S04

this

stu

dy

1

3

551015202530

8

N94

hydrous melt

(a) (b)

H2O



732



ownloaded from







733



ownloaded from




K2O

H2O

600 800 1000

2535

45

aqueous fluid


lavas


2

4

400 600 800 1000

T (degC)

P (

GP

a)

1) 2) 3) 4) 5)

N94

this

stu

dy

phen

gite

out

1

3

5

hydr

ous

glas

s

T (degC)


hydrous melt

aqueous fluid



0001

001

01

1

500 700 900

(a) (b)



734



ownloaded from









735



ownloaded from









736



ownloaded from











737



ownloaded from

















































738



ownloaded from















































739



ownloaded from




















740



ownloaded from


substitution

Other minerals






31

32

33

34

35

36

500 600 700 800 900 1000 1100

45

35

25

(a) (b)

Si (

pfu

)

T (degC)500 600 700 800 900 1000 1100

T (degC)

059

051

063

066

061

042

096

072

042

145

089

054

068

187

176

177

107

191

199

25

35

45

P (

GP

a)

Phe

out

05 075 10 15 20TiO2

Phengite



726



ownloaded from

2

4

6

8

10

SiO

2A

l 2O

3

500 600 700 800 900 1000 1100

T (degC)

500 600 700 800 900 1000 1100

T (degC)

800degC 45 GPa

0

05

1

15

Na 2

OA

l 2O

3

aqueous fluid

hydrous melt

aqueous fluid

hydrous melt

01

1

10

K2O

Na 2

O

0

25

5

75

10

125

K2O

500 600 700 800 900 1000 1100

T (degC)

500 600 700 800 900 1000 1100

T (degC)

An

Ab Or

tonalite

trondhjemite

granodiorite

granite

(95)

(92)

20

(89)

(85)

16

25

10

20

29

5

10

1111

16

6

8 6

25

35

45

P (

GP

a)

600 700 800 900 1000 1100

T (degC)

51015202530

(a) (b)

(c)(d)

(f)(e)


aqueous fluid

hydrous melt

wet

sol

idus

H2O



727



ownloaded from


Diamond traps










Exp 2099 2098 2082 2083 2097

P (GPa) 25 25 35 35 45

T (8C) 700 600 600 700 700

SiO2 7247 7538 6902 7505 8847

TiO2 024 032 042 027 022

Al2O3 1478 1021 880 774 261

FeO 183 187 293 309 027

MnO 004 008 014 026 001

MgO 060 247 300 180 393

CaO 184 097 168 183 061

Na2O 578 786 1255 848 320

K2O 209 083 146 147 069

Sum 100 100 100 100 100

H2O 20 95 92 89 85

K2OH2O 0084 00004 00013 00018 00012

XMg 037 070 065 051 096

ASI 098 065 035 041 037

Alb 71 88 87 81 80

An 12 6 6 10 8

Or 17 6 7 9 11



728



ownloaded from




Garnet




05

1

15

2

25

3

Ln

KD

(G

rt-P

he)

75 8 85 9 95 10 1051T (degK) x 10000

45 GPa1

2

3

4

5

Na 2

O C

px

Mel

t

700 800 900 1000 1100T (˚C)


T (degC)

7008009001000

T (degC)

GH82

(a) (b) (c)



minus30

minus20

minus10

0

10

20

30

(Pca

lc -

Pex

p)

Pex

p x

100

600 700 800 900 1000 1100



729



ownloaded from






Phengite



730



ownloaded from










731



ownloaded from







2

4

600 800 1000

T (degC)

600 800 1000

T (degC)

N94

S04

A06

J99th

is s

tudy

biotite out

phen

gite

out

1

3

5

hydr

ous

glas

s

2

4

P (

GP

a)

P (

GP

a)

N94

S04

this

stu

dy

1

3

551015202530

8

N94

hydrous melt

(a) (b)

H2O



732



ownloaded from







733



ownloaded from




K2O

H2O

600 800 1000

2535

45

aqueous fluid


lavas


2

4

400 600 800 1000

T (degC)

P (

GP

a)

1) 2) 3) 4) 5)

N94

this

stu

dy

phen

gite

out

1

3

5

hydr

ous

glas

s

T (degC)


hydrous melt

aqueous fluid



0001

001

01

1

500 700 900

(a) (b)



734



ownloaded from









735



ownloaded from









736



ownloaded from











737



ownloaded from

















































738



ownloaded from















































739



ownloaded from




















740



ownloaded from

2

4

6

8

10

SiO

2A

l 2O

3

500 600 700 800 900 1000 1100

T (degC)

500 600 700 800 900 1000 1100

T (degC)

800degC 45 GPa

0

05

1

15

Na 2

OA

l 2O

3

aqueous fluid

hydrous melt

aqueous fluid

hydrous melt

01

1

10

K2O

Na 2

O

0

25

5

75

10

125

K2O

500 600 700 800 900 1000 1100

T (degC)

500 600 700 800 900 1000 1100

T (degC)

An

Ab Or

tonalite

trondhjemite

granodiorite

granite

(95)

(92)

20

(89)

(85)

16

25

10

20

29

5

10

1111

16

6

8 6

25

35

45

P (

GP

a)

600 700 800 900 1000 1100

T (degC)

51015202530

(a) (b)

(c)(d)

(f)(e)


aqueous fluid

hydrous melt

wet

sol

idus

H2O



727



ownloaded from


Diamond traps










Exp 2099 2098 2082 2083 2097

P (GPa) 25 25 35 35 45

T (8C) 700 600 600 700 700

SiO2 7247 7538 6902 7505 8847

TiO2 024 032 042 027 022

Al2O3 1478 1021 880 774 261

FeO 183 187 293 309 027

MnO 004 008 014 026 001

MgO 060 247 300 180 393

CaO 184 097 168 183 061

Na2O 578 786 1255 848 320

K2O 209 083 146 147 069

Sum 100 100 100 100 100

H2O 20 95 92 89 85

K2OH2O 0084 00004 00013 00018 00012

XMg 037 070 065 051 096

ASI 098 065 035 041 037

Alb 71 88 87 81 80

An 12 6 6 10 8

Or 17 6 7 9 11



728



ownloaded from




Garnet




05

1

15

2

25

3

Ln

KD

(G

rt-P

he)

75 8 85 9 95 10 1051T (degK) x 10000

45 GPa1

2

3

4

5

Na 2

O C

px

Mel

t

700 800 900 1000 1100T (˚C)


T (degC)

7008009001000

T (degC)

GH82

(a) (b) (c)



minus30

minus20

minus10

0

10

20

30

(Pca

lc -

Pex

p)

Pex

p x

100

600 700 800 900 1000 1100



729



ownloaded from






Phengite



730



ownloaded from










731



ownloaded from







2

4

600 800 1000

T (degC)

600 800 1000

T (degC)

N94

S04

A06

J99th

is s

tudy

biotite out

phen

gite

out

1

3

5

hydr

ous

glas

s

2

4

P (

GP

a)

P (

GP

a)

N94

S04

this

stu

dy

1

3

551015202530

8

N94

hydrous melt

(a) (b)

H2O



732



ownloaded from







733



ownloaded from




K2O

H2O

600 800 1000

2535

45

aqueous fluid


lavas


2

4

400 600 800 1000

T (degC)

P (

GP

a)

1) 2) 3) 4) 5)

N94

this

stu

dy

phen

gite

out

1

3

5

hydr

ous

glas

s

T (degC)


hydrous melt

aqueous fluid



0001

001

01

1

500 700 900

(a) (b)



734



ownloaded from









735



ownloaded from









736



ownloaded from











737



ownloaded from

















































738



ownloaded from















































739



ownloaded from




















740



ownloaded from


Diamond traps










Exp 2099 2098 2082 2083 2097

P (GPa) 25 25 35 35 45

T (8C) 700 600 600 700 700

SiO2 7247 7538 6902 7505 8847

TiO2 024 032 042 027 022

Al2O3 1478 1021 880 774 261

FeO 183 187 293 309 027

MnO 004 008 014 026 001

MgO 060 247 300 180 393

CaO 184 097 168 183 061

Na2O 578 786 1255 848 320

K2O 209 083 146 147 069

Sum 100 100 100 100 100

H2O 20 95 92 89 85

K2OH2O 0084 00004 00013 00018 00012

XMg 037 070 065 051 096

ASI 098 065 035 041 037

Alb 71 88 87 81 80

An 12 6 6 10 8

Or 17 6 7 9 11



728



ownloaded from




Garnet




05

1

15

2

25

3

Ln

KD

(G

rt-P

he)

75 8 85 9 95 10 1051T (degK) x 10000

45 GPa1

2

3

4

5

Na 2

O C

px

Mel

t

700 800 900 1000 1100T (˚C)


T (degC)

7008009001000

T (degC)

GH82

(a) (b) (c)



minus30

minus20

minus10

0

10

20

30

(Pca

lc -

Pex

p)

Pex

p x

100

600 700 800 900 1000 1100



729



ownloaded from






Phengite



730



ownloaded from










731



ownloaded from







2

4

600 800 1000

T (degC)

600 800 1000

T (degC)

N94

S04

A06

J99th

is s

tudy

biotite out

phen

gite

out

1

3

5

hydr

ous

glas

s

2

4

P (

GP

a)

P (

GP

a)

N94

S04

this

stu

dy

1

3

551015202530

8

N94

hydrous melt

(a) (b)

H2O



732



ownloaded from







733



ownloaded from




K2O

H2O

600 800 1000

2535

45

aqueous fluid


lavas


2

4

400 600 800 1000

T (degC)

P (

GP

a)

1) 2) 3) 4) 5)

N94

this

stu

dy

phen

gite

out

1

3

5

hydr

ous

glas

s

T (degC)


hydrous melt

aqueous fluid



0001

001

01

1

500 700 900

(a) (b)



734



ownloaded from









735



ownloaded from









736



ownloaded from











737



ownloaded from

















































738



ownloaded from















































739



ownloaded from




















740



ownloaded from




Garnet




05

1

15

2

25

3

Ln

KD

(G

rt-P

he)

75 8 85 9 95 10 1051T (degK) x 10000

45 GPa1

2

3

4

5

Na 2

O C

px

Mel

t

700 800 900 1000 1100T (˚C)


T (degC)

7008009001000

T (degC)

GH82

(a) (b) (c)



minus30

minus20

minus10

0

10

20

30

(Pca

lc -

Pex

p)

Pex

p x

100

600 700 800 900 1000 1100



729



ownloaded from






Phengite



730



ownloaded from










731



ownloaded from







2

4

600 800 1000

T (degC)

600 800 1000

T (degC)

N94

S04

A06

J99th

is s

tudy

biotite out

phen

gite

out

1

3

5

hydr

ous

glas

s

2

4

P (

GP

a)

P (

GP

a)

N94

S04

this

stu

dy

1

3

551015202530

8

N94

hydrous melt

(a) (b)

H2O



732



ownloaded from







733



ownloaded from




K2O

H2O

600 800 1000

2535

45

aqueous fluid


lavas


2

4

400 600 800 1000

T (degC)

P (

GP

a)

1) 2) 3) 4) 5)

N94

this

stu

dy

phen

gite

out

1

3

5

hydr

ous

glas

s

T (degC)


hydrous melt

aqueous fluid



0001

001

01

1

500 700 900

(a) (b)



734



ownloaded from









735



ownloaded from









736



ownloaded from











737



ownloaded from

















































738



ownloaded from















































739



ownloaded from




















740



ownloaded from






Phengite



730



ownloaded from










731



ownloaded from







2

4

600 800 1000

T (degC)

600 800 1000

T (degC)

N94

S04

A06

J99th

is s

tudy

biotite out

phen

gite

out

1

3

5

hydr

ous

glas

s

2

4

P (

GP

a)

P (

GP

a)

N94

S04

this

stu

dy

1

3

551015202530

8

N94

hydrous melt

(a) (b)

H2O



732



ownloaded from







733



ownloaded from




K2O

H2O

600 800 1000

2535

45

aqueous fluid


lavas


2

4

400 600 800 1000

T (degC)

P (

GP

a)

1) 2) 3) 4) 5)

N94

this

stu

dy

phen

gite

out

1

3

5

hydr

ous

glas

s

T (degC)


hydrous melt

aqueous fluid



0001

001

01

1

500 700 900

(a) (b)



734



ownloaded from









735



ownloaded from









736



ownloaded from











737



ownloaded from

















































738



ownloaded from















































739



ownloaded from




















740



ownloaded from










731



ownloaded from







2

4

600 800 1000

T (degC)

600 800 1000

T (degC)

N94

S04

A06

J99th

is s

tudy

biotite out

phen

gite

out

1

3

5

hydr

ous

glas

s

2

4

P (

GP

a)

P (

GP

a)

N94

S04

this

stu

dy

1

3

551015202530

8

N94

hydrous melt

(a) (b)

H2O



732



ownloaded from







733



ownloaded from




K2O

H2O

600 800 1000

2535

45

aqueous fluid


lavas


2

4

400 600 800 1000

T (degC)

P (

GP

a)

1) 2) 3) 4) 5)

N94

this

stu

dy

phen

gite

out

1

3

5

hydr

ous

glas

s

T (degC)


hydrous melt

aqueous fluid



0001

001

01

1

500 700 900

(a) (b)



734



ownloaded from









735



ownloaded from









736



ownloaded from











737



ownloaded from

















































738



ownloaded from















































739



ownloaded from




















740



ownloaded from







2

4

600 800 1000

T (degC)

600 800 1000

T (degC)

N94

S04

A06

J99th

is s

tudy

biotite out

phen

gite

out

1

3

5

hydr

ous

glas

s

2

4

P (

GP

a)

P (

GP

a)

N94

S04

this

stu

dy

1

3

551015202530

8

N94

hydrous melt

(a) (b)

H2O



732



ownloaded from







733



ownloaded from




K2O

H2O

600 800 1000

2535

45

aqueous fluid


lavas


2

4

400 600 800 1000

T (degC)

P (

GP

a)

1) 2) 3) 4) 5)

N94

this

stu

dy

phen

gite

out

1

3

5

hydr

ous

glas

s

T (degC)


hydrous melt

aqueous fluid



0001

001

01

1

500 700 900

(a) (b)



734



ownloaded from









735



ownloaded from









736



ownloaded from











737



ownloaded from

















































738



ownloaded from















































739



ownloaded from




















740



ownloaded from







733



ownloaded from




K2O

H2O

600 800 1000

2535

45

aqueous fluid


lavas


2

4

400 600 800 1000

T (degC)

P (

GP

a)

1) 2) 3) 4) 5)

N94

this

stu

dy

phen

gite

out

1

3

5

hydr

ous

glas

s

T (degC)


hydrous melt

aqueous fluid



0001

001

01

1

500 700 900

(a) (b)



734



ownloaded from









735



ownloaded from









736



ownloaded from











737



ownloaded from

















































738



ownloaded from















































739



ownloaded from




















740



ownloaded from




K2O

H2O

600 800 1000

2535

45

aqueous fluid


lavas


2

4

400 600 800 1000

T (degC)

P (

GP

a)

1) 2) 3) 4) 5)

N94

this

stu

dy

phen

gite

out

1

3

5

hydr

ous

glas

s

T (degC)


hydrous melt

aqueous fluid



0001

001

01

1

500 700 900

(a) (b)



734



ownloaded from









735



ownloaded from









736



ownloaded from











737



ownloaded from

















































738



ownloaded from















































739



ownloaded from




















740



ownloaded from









735



ownloaded from









736



ownloaded from











737



ownloaded from

















































738



ownloaded from















































739



ownloaded from




















740



ownloaded from









736



ownloaded from











737



ownloaded from

















































738



ownloaded from















































739



ownloaded from




















740



ownloaded from











737



ownloaded from

















































738



ownloaded from















































739



ownloaded from




















740



ownloaded from

















































738



ownloaded from















































739



ownloaded from




















740



ownloaded from















































739



ownloaded from




















740



ownloaded from




















740



ownloaded from

sediment melts at sub-arc depths: an experimental study

Documents