Petrology and Geochemistry of High-Pressure Rocks and Related Veins (SW Tianshan, China): Constraints on Fluid-Rock Interactions in Subduction Zones

Home , Aluminosilicate, Eclogite, Paragonite, Pyrope

Petrology and geochemistry of high-pressure rocks and related veins (SW Tianshan, China): Constraints on fluid-rock interactions in subduction zones

Petrologie und Geochemie von Hochdruckgesteinen und assoziierten Adern (SW Tianshan, China): Hinweise auf Fluid-Gesteinswechselwirkungen in Subduktionszonen

Der Naturwissenschaftlichen Fakultät

der Friedrich-Alexander Universität

Erlangen-Nürnberg

zur

Erlangung des Doktorgrades Dr. rer. nat.

vorgelegt von

Jilei Li

aus Yutai, Shandong, VR China

Als Dissertation genehmigt von der Naturwissenschaftlichen Fakultät der Friedrich-Alexander Universität Erlangen-Nürnberg

Tag der mündlichen Prüfung: 10.04.2014

Vorsitzender der Promotionsorgans: Prof. Dr. Johannes Barth

Gutachter: Prof. Dr. Reiner Klemd

Prof. Dr. Timm John

Abstract

High-pressure/low-temperature blueschist-facies to eclogite-facies rocks are widely regarded to represent exhumed fragments of subducted slabs. Therefore, metamorphic studies of eclogites and related high-pressure rocks may reveal important information about the P-T evolution of subducted crustal rocks at depth in subduction zones. In addition, vein systems in eclogite-facies rocks are the direct result of dehydration of oceanic crust and fluid activity in subduction zones. Thus their study provides direct insight into fluid processes and element mobility during fluid-rock interaction in subduction zones. The Chinese Tianshan (ultra-)high- pressure/low-temperature metamorphic belt represents a former oceanic subduction zone complex, in which eclogite-facies rocks and related high-pressure veins are commonly observed.

This thesis focuses on three major aspects that involve (1) the metamorphic evolution of carbonate-bearing eclogites/blueschists, (2) processes of fluid-rock interaction and related element mobility in subduction zones, and (3) the relationship of eclogite interlayers and surrounding metasedimentary rocks:

Carbonate-bearing eclogite and blueschist coexist in the same lithological sequence (Chapter 3), both of which consist of the same mineral assemblage but with various modal amounts. Mineral assemblages and textures show that they underwent an identical metamorphic evolution, while phase equilibrium modeling further indicates their equilibration at the same peak metamorphic conditions. This coexistence is believed to be due to different bulk-rock compositions. Three kinds of blueschist were formed during subduction and exhumation of the South Tianshan oceanic crust, including prograde- and retrograde-blueschist, and peak metamorphic blueschist coexisting with eclogite. Compositional zoning of dolomite was discovered in one carbonate-lawsonite-bearing eclogite (Chapter 4). The chemical zoning in dolomite is well defined by a continuous core–to–rim Mg increase and Fe–Mn decrease. Thermodynamic modeling demonstrates

that the Fe–Mg zoning of dolomite is largely temperature dependent and the dolomite is in equilibrium with garnet formed as a result of changing matrix compositions. This is the first report showing the possibility to retrieve information on the prograde metamorphic evolution of eclogites from high-pressure carbonates.

The study of fluid-rock interaction processes and mobilization of both transition metals and trace elements in subduction-zone fluids was undertaken using petrological- geochemical data of a sulfide-bearing HP vein and its massive lawsonite eclogite host rock (Chapter 5). The omphacite-dominated vein is enveloped by a garnet-poor, sulfide- bearing eclogite-facies reaction selvage. Textual evidence indicates that the selvage formed due to dissolution-precipitation processes as a consequence of fluid-rock interaction. Mass-balance calculations indicate that the reaction selvage experienced a significant depletion of the LILE (K–Rb–Ba) and a moderate depletion of the HREE and metal elements including Fe, Cu, Ni, Zn, Co, Cr, and Mn, along with a significant enrichment of CaO, Sr, Pb and S whereas the HFSE show no significant variations. Significant amounts of transitional metal elements were released into the fluid during the dissolution of white mica and the dissolution-precipitation behavior of garnet, omphacite, dolomite and sulfides. Thus the LILE and some transition metal elements (e.g., Fe, Cu, Ni and Zn) were mobilized during the dissolution-precipitation processes that led to the selvage formation. Accordingly the slab fluids are not only strongly enriched in LILE and depleted in HFSE, but also carry significant amounts of transition metals. It is most likely that slab fluids strongly contribute to the metal flux into the arc magma systems finally resulting in arc-related ore deposits.

Interlayered eclogite, marble and quartz-mica schist from a drill core are studied to constrain the metamorphic evolution of metabasalts and intercalcated metasediments (Chapter 6). Using garnet isopleths thermobarometry, pseudosections for the eclogite and quartz-mica schist reveal a common metamorphic history under high-pressure condition, which is also consistent with the conventional P–T estimates for the marble. The uniform P–T paths of those rocks show that they form an internally coherent high-pressure unit and not an ―internally coherent ultrahigh-pressure unit‖, which was ii

recently proposed for all eclogite-facies rocks from this area.

This thesis was aimed to enhance the understanding of the high-pressure metamorphic evolution of subducted oceanic slabs and fluid-rock interaction processes in subduction zones. Furthermore, it provides constraints on the geodynamic setting of the Tianshan high-pressure terrane. The highlights of the study include that: (1) blueschists coexist with eclogites at the same metamorphic conditions; (2) eclogite-facies dolomite has compositional zoning that allows to retrieve prograde metamorphic evolutions of eclogite-facies rocks; (3) dissolution-precipitation processes that occur during the fluid-rock interaction may trigger the transport of metal elements and this can ultimately contribute to the formation of arc-related ore deposits; and (4) the previously proposed tectonic subdivision in the northern ―internally coherent UHP unit‖ and the southern ―coherent HP unit‖ for the (ultra-)high-pressure/low-temperature Tianshan terrane is unlikely.

iii

Kurzfassung

Hochdruck/Niedertemperaturgesteine der Blauschiefer- bis Eklogit-Fazies werden im Allgemeinen als exhumierte Fragmente subduzierter Platten betrachtet. Folglich können Studien über Eklogite und assoziierte Hochdruckgesteine wichtige Informationen über die Druck-Temperatur (P-T) Entwicklung subduzierter krustaler Gesteine in der Tiefe liefern. Weiterhin stellen Ader-Systeme in eklogit-faziellen Gesteinen das direkte Ergebnis von Entwässerung ozeanischer Kruste und Fluidaktivität in Subduktionszonen dar. Aus diesem Grunde gibt deren Untersuchung einen direkten Einblick in Fluidprozesse, sowie die Mobilität verschiedener Elemente bei Fluid-Gesteins- Wechselwirkungen in Subduktionszonen. Das Vorkommen metamorpher (Ultra)Hochdruck/Niedertemperatur Gesteine im chinesischen Tianshan entspricht einem ehemaligen Subduktionszonen-Komplex, in welchem weithin eklogit-fazielle Gesteine und zugehörige Hochdruck-Adern zu finden sind.

Der Schwerpunkt dieser Arbeit liegt auf 3 Hauptthemen: (1) die metamorphe Entwicklung Karbonat-führender Eklogite und Blauschiefer, (2) Fluid-Gesteins- Wechselwirkungsprozesse und die damit einhergehende Mobilität verschiedener Elemente in Subduktionszonen, und (3) der Zusammenhang zwischen eklogitischen Lagen und den umgebenden Metasedimenten:

Karbonat-führende Eklogite und Blauschiefer kommen in der gleichen lithologischen Einheit nebeneinander vor (Kapitel 3). Dabei sind beide Gesteinsarten aus derselben Mineralparagenese aufgebaut, unterscheiden sich jedoch im Bezug auf die Modalgehalte der jeweiligen Minerale. Sowohl Mineralparagenesen, als auch Strukturen zeigen, dass beide Gesteine durch eine identische metamorphe Entwicklung gebildet wurden. Weiterhin impliziert eine thermodynamische Modellierung eine Äquilibration unter identischen maximalen Metamorphosebedingungen. Die Koexistenz der beiden verschiedenen Gesteine ist vermutlich einem unterschiedlichen Gesamtgesteinschemismus zuzuschreiben. Während der Subduktion und Exhumierung der ozeanischen Kruste im südlichen Tianshan wurden drei verschiedene Arten von

Blauschiefern gebildet, nämlich prograder und retrograder Blauschiefer, sowie Blauschiefer der unter maximalen Metamorphosebedingungen gebildet wurde. Letztgenannter kommt stabil neben Eklogit vor.

In einem Karbonat-Lawsonit-führenden Eklogit wurde eine chemische Zonierung in Dolomit gefunden (Kapitel 4). Diese ist vom Kern zum Rand durch einen systematischen Anstieg von Mg und eine Abnahme von Fe-Mn gekennzeichnet. Eine thermodynamische Modellierung zeigt, dass die Fe-Mg Zonierung von Dolomit überwiegend temperaturabhängig ist und sich Dolomit im Gleichgewicht mit Granat befindet, welcher sich wiederum als Ergebnis der sich ändernden Matrixzusammensetzung bildet. Dies ist die erste Arbeit, welche die Möglichkeit aufzeigt, aus Hochdruck-Karbonaten Informationen über die prograde metamorphe Entwicklung abzuleiten.

Mithilfe von petrologischen und geochemischen Daten einer Sulfid-führenden Hochdruckader und deren Umgebungsgestein – einem massiven Lawsonit-Eklogit – wurde eine Untersuchung über Fluid-Gesteins-Wechselwirkungen sowie die Mobilisierung von verschiedenen Übergangsmetallen und Spurenelementen in Subduktionszonenfluiden durchgeführt (Kapitel 5). Die überwiegend aus Omphazit aufgebaute Ader ist von einem Granat-armen, Sulfid-führenden eklogit-faziellen Reaktionssaum umgeben. Strukturell lässt sich nachweisen, dass sich der Saum aufgrund von Lösungs- und Ausfällungsprozessen – einer Konsequenz der Fluid-Gesteins-Wechselwirkung – gebildet hat. Berechnungen im Rahmen einer Massenbilanzierung zeigen, dass der Reaktionssaum deutlich an LILE (K, Rb, Ba) und mäßig an HREE sowie Metallen (u.a. Fe, Cu, Ni, Zn, Co, Cr und Mn) verarmt ist, allerdings sind CaO, Sr, P und S stark angereichert, wohingegen die HFSE keine deutlichen Änderungen aufweisen. Ein signifikanter Anteil von Elementen der Übergangsmetalle wurde während der Lösung von Hellglimmer und der Lösungs-Fällungs-Reaktionen von Granat, Omphazit, Dolomit und Sulfiden in das Fluid freigesetzt. Folglich wurden während der Lösungs-Fällungs-Prozesse, welche zur Bildung des Reaktionssaums führten, die LILE und einige Elemente der Übergangsmetalle (z.B. Fe, Cu, Ni und Zn) mobilisiert. Die Fluide der subduzierten

Platte hingegen sind nicht nur stark an LILE angereichert und an HFSE verarmt, sondern sind Träger signifikanter Mengen an Übergangsmetallen. Es ist demnach sehr wahrscheinlich, dass Fluide aus subduzierten Platten einen großen Anteil am Elementfluss von Metallen in die Magmen von Inselbögen haben, was schließlich zur Bildung von mit Inselbögen in Verbindung stehenden Erzlagerstätten führt.

Um die metamorphe Entwicklung von Metabasalten und zwischengelagerten Metasedimenten zu untersuchen wurden zwischengelagerte Eklogite, Marmore und Quarz-Glimmer-Schiefer eines Bohrkerns untersucht (Kapitel 6). Unter Zuhilfenahme der Granat-Isoplethen Thermobarometrie zeigen quantitative Netze (P-T Pseudosections) des Eklogites und des Quarz-Glimmer-Schiefers eine gemeinsame metamorphe Entwicklung unter Hochdruckbedingungen, was mit konventionellen P-T Abschätzungen des Marmors übereinstimmt. Die einheitlichen P-T Pfade dieser Gesteine zeigen, dass sie eine intern kohärente Hochdruck-Einheit bilden und nicht, wie kürzlich für alle eklogit-faziellen Gesteine dieser Gegend vorgeschlagen, eine „intern kohärente Ultra-Hochdruck-Einheit―.

Das Ziel dieser Arbeit war es, das Wissen über die metamorphe Entwicklung ozeanischer Platten und Fluid-Gesteins-Wechselwirkungen in Subduktionszonen unter Hochdruckbedingungen zu erweitern. Des Weiteren wurde der geodynamische Rahmen des Tianshan Hochdruck-Terranes untersucht. Die wichtigsten Erkenntnisse dieser Arbeit umfassen: (1) Blauschiefer kommen zusammen mit Eklogiten unter identischen metamorphen Bedingungen vor; (2) eklogit-fazieller Dolomit weist eine chemische Zonierung auf, welche es ermöglicht, die prograde metamorphe Entwicklung eklogit-fazieller Gesteine zu ermitteln; (3) Lösungs-Fällungs-Reaktionen die während Fluid-Gesteins-Wechselwirkungen auftreten können den Transport von Übergangsmetallen zur Folge haben, was schließlich zur Bildung von mit Inselbögen in Bezug stehenden Erzlagerstätten führen kann und (4) die früher postulierte Unterteilung des (Ultra)Hochdruck/Niedertemperatur Tianshan Terranes in eine nördliche „intern kohärente UHP Einheit― und eine südliche „kohärente HP Einheit― ist unwahrscheinlich.

Statement of candidate

I certify that the work in this thesis entitled “Petrology and geochemistry of high-pressure rocks and related veins (SW Tianshan, China): Constraints on fluid-rock interactions in subduction zones” has previously not been submitted for a degree nor has it been submitted as part of requirement for a degree to any other university or institution other than the Friedrich-Alexander Universität Erlangen-Nürnberg.

I also certify that the thesis is a new, original piece of research and it has been written by me. Any help and assistance that I have received in my research work and the preparation of the thesis itself have been appropriately acknowledged.

In addition, I certify that all information sources and literature used are indicated in the thesis.

This thesis contains material that has been published or submitted for publication in peer-reviewed ISI-journals, as follows:

Chapter 3 “Coexisting carbonate-bearing eclogite and blueschist in SW Tianshan, China: Petrology and phase equilibria” has been published in Journal of Asian Earth Sciences in 2012. My contribution to this publication consisted of sample collection (95%) and preparation (80%), data analyses (90%), pseudosection calculation (95%), interpretation (90%) and writing (85%), resulting in a total contribution of about 90%. Impact factor: 2.379 (2012).

Chapter 4 “Compositional zoning in dolomite from lawsonite-bearing eclogite (SW Tianshan, China): Evidence for prograde metamorphism during subduction of oceanic crust” has been published in American Mineralogist in 2014. My contribution to this publication consisted of sample collection (95%) and preparation (80%), data analyses (90%), pseudosection calculation (100%), interpretation (95%) and writing (95%), resulting in a total contribution of about 95%. Impact factor: 2.204 (2012).

vii

Chapter 5 “Fluid-mediated transition metal transport in subduction zones and its link to arc-related giant ore deposits: Constraints from a sulfide-bearing HP vein in lawsonite eclogite (Tianshan, China)” has been published in Geochimica et Cosmochimica Acta in 2013. My contribution to this publication consisted of sample collection (95%) and preparation (90%), data analyses (85%), pseudosection calculation (95%), mass balance calculation (90%), interpretation (70%) and writing (70%), resulting in a total contribution of about 80%. Impact factor: 3.884 (2012).

Chapter 6 “A common high-pressure metamorphic evolution of interlayered eclogites and metasediments from the „ultrahigh-pressure unit‟ of the Tianshan metamorphic belt in China” has been submitted to Lithos at the beginning of 2014. My contribution to this publication consisted of sample preparation (30%), data analyses (80%), pseudosection calculation (100%), interpretation (80%) and writing (80%), resulting in a total contribution of about 70%.

Erlangen, 13.01.2014

Jilei Li

viii

Complete publication list

Li J.L., Klemd R., Gao J., Jiang T., Song Y.H. (2014) A common high-pressure metamorphic evolution of interlayered eclogites and metasediments from the ‗ultrahigh-pressure unit‘ of the Tianshan metamorphic belt in China. Submitted to Lithos.

Klemd R., Hegner E., Bergmann H., Pfänder J.A., Li J.L., Hentschel F. (2014) Eclogitization of continental crust of the Aktyuz Complex during Late Palaeozoic plate collisions in the Northern Tianshan of Kyrgyzstan. Gondwana Research, DOI: http://dx.doi.org/10.1016/j.gr.2013.08.018.

Li J.L., Klemd R., Gao J., Meyer M. (2014) Compositional zoning in dolomite from lawsonite-bearing eclogite (SW Tianshan, China): Evidence for prograde metamorphism during subduction of oceanic crust. American Mineralogist 99, 206–217.

Li J.L., Gao J., John T., Klemd R., Su W. (2013) Fluid-mediated transition metal transport in subduction zones and its link to arc-related giant ore deposits: Constraints from a sulfide-bearing HP vein in lawsonite eclogite (Tianshan, China). Geochimica et Cosmochimica Acta 120, 326–362.

Li J.L., Klemd R., Gao J., Meyer M. (2012) Coexisting carbonate-bearing eclogite and blueschist in SW Tianshan, China: Petrology and phase equilibria. Journal of Asian Earth Sciences 60, 174–187.

Gao J., Klemd R., Qian Q., Zhang X., Li J.L., Jiang T., Yang Y. (2011) The collision between the Yili and Tarim blocks of the Southwestern Altaids: Geochemical and age constraints of a leucogranite dike crosscutting the HP-LT metamorphic belt in the Chinese Tianshan Orogen. Tectonophysics 499, 118–131.

Acknowledgments

My deepest gratitude goes first and foremost to Prof. Dr. Reiner Klemd, my supervisor, for his constant guidance and encouragement through all stages of my researches. I am further very greatful to Prof. Dr. Jun Gao and Prof. Dr. Timm John for their guidance and encouragement, as well as their patient discussions and suggestions on my papers.

Second, I would like to thank my colleagues Dr. Philipp A. Brandl, Dr. Helene Brätz, Xiaoxia Duan, Sarah Freund, Melanie Hertel, Prof. apl. Dr. Michael Joachimski, Manuel Keith, Dr. Xiaoxiao Ling, Melanie Meyer, Dr. Yadong Sun, Christoph Weinzierl and all my other colleagues at the GeoZentrum Nordbayern for sharing this exciting 3-years together. I would like to express my special gratitude to Melanie Meyer, who has instructed me on technical details and helped me on various aspects in the past three years.

Third, I also would like to thank all my Chinese friends in Erlangen and Germany, especially to Renneng Bi, Jiapeng Huang, Jin Luo, Jie Tong, Anjia Wang, Shuai Yan and Ling Zhou; you listened to me and helped me to get rid of loneliness in a foreign country. I cannot forget the wonderful time when we travelled in Germany and Europe.

My deepest thanks would go to my beloved wife Sha Liu, my lovely daughter Joanna, and my whole family for your loving considerations and whole-hearted support all through these years. You are the best motivation to push me moving on…

Finally I thank the Deutscher Akademischer Austauschdienst (DAAD), China Scholarship Council (CSC), and the Deutsche Forschungsgemeinschaft (DFG, KL 692/17-2, 3) for the scholarship and financial help supporting my PhD study in Germany.

I love you, Erlangen. I love you, Germany.

Contents

Abstract ...... i Kurzfassung ...... iv Statement of candidate ...... vii Complete publication list ...... ix Acknowledgments ...... x Contents ...... xi List of Figures ...... xv List of Tables ...... xvi 1 Introduction ...... 1 1.1 Subduction zone and related rocks ...... 1

1.2 HP-UHP metamorphism ...... 2

1.3 Nature of subduction zone fluids ...... 4

1.4 Tianshan (U)HP/LT metamorphic belt ...... 6

1.5 Research approachs and analytial techniques ...... 9

References ...... 10 2 Aims of the study ...... 17 2.1 Metamorphism of carbonate-bearing eclogites and blueschists ...... 17

2.2 Process of fluid-rock interaction and transport of metal elements ...... 18

2.3 Tectonic relationship between mafic and metasedimentary rocks in Tianshan ...... 19

References ...... 20 3 Coexisting carbonate-bearing eclogite and blueschist in SW Tianshan, China: Petrology and phase equilibria ...... 23 3.1 Abstract ...... 23

3.2 Introduction ...... 24

3.3 Geological setting ...... 25

3.4 Sample description and petrography ...... 28

3.4.1 Eclogite ...... 30

3.4.2 Blueschist ...... 30

3.5 Analytical methods ...... 33

3.6 Mineral chemistry ...... 34

3.6.1 Garnet ...... 34

3.6.2 Omphacite ...... 36

3.6.3 Amphibole ...... 36

3.6.4 White mica ...... 37

3.6.5 Carbonates ...... 37

3.6.6 Epidote/clinozoisite ...... 38

3.6.7 Accessory minerals ...... 38

3.7 Phase equilibria and P-T conditions ...... 39

3.7.1 Effective bulk-rock composition ...... 39

3.7.2 Pseudosection calculations ...... 41

3.8 Discussion ...... 44

3.8.1 Coexistence of eclogite and blueschist ...... 44

3.8.2 Carbonates in the Tianshan HP rocks ...... 47

3.8.3 Implications ...... 48

3.9 Conclusions ...... 49

Acknowledgements ...... 50

References ...... 50 4 Compositional zoning in dolomite from lawsonite-bearing eclogite (SW Tianshan, China): Evidence for prograde metamorphism during subduction of oceanic crust57 4.1 Abstract ...... 57

4.2 Introduction ...... 58

4.3 Geological setting and petrography ...... 59

4.4 Analytical methods ...... 62

4.5 Mineral chemistry ...... 63

4.6 Compositional zoning of dolomite ...... 65

4.6.1 Major elements ...... 65

4.6.2 Trace elements ...... 67

4.7 Thermodynamic modeling ...... 69

4.8 Discussion ...... 71

4.8.1 Formation of dolomite and its prograde compositional zoning ...... 71

4.8.2 The occurrence of magnesite in dolomite in the Tianshan eclogites ...... 75

4.9 Implications ...... 77

Acknowledgements ...... 78

References ...... 78

xii

5 Fluid-mediated metal transport in subduction zones and its link to arc-related giant ore deposits: Constraints from a sulfide-bearing HP vein in lawsonite eclogite (Tianshan, China) ...... 83 5.1 Abstract ...... 83

5.2 Introduction ...... 85

5.3 Geological context ...... 87

5.4 Sample description and petrography ...... 90

5.4.1 Sample description ...... 90

5.4.2 The wall-rock eclogite (WE) ...... 91

5.4.3 The reaction selvage (RS) ...... 93

5.4.4 The vein (V) ...... 95

5.5 Analytical methods ...... 96

5.6 Results ...... 100

5.6.1 Bulk geochemistry ...... 100

5.6.2 Mineral chemistry ...... 105

5.7 Discussion ...... 122

5.7.1 The formation of the vein-selvage system ...... 122

5.7.2 Chemical feedback of fluid-rock interaction ...... 131

5.7.3 Constraints on fluid sources ...... 139

5.7.4 Implications for the arc-related ore deposits ...... 142

5.8 Conclusions ...... 143

Acknowledgements ...... 144

References ...... 145 6 A common high-pressure metamorphic evolution of interlayered eclogites and metasediments from the „ultrahigh-pressure unit‟ of the Tianshan metamorphic belt in China ...... 155 6.1 Abstract ...... 155

6.2 Introduction ...... 156

6.3 Geological setting ...... 158

6.4 Sample description and petrography ...... 160

6.4.1 Eclogite ...... 161

6.4.2 Quartz-mica schist ...... 161

6.4.3 Marble ...... 163

6.5 Analytical methods ...... 163

xiii

6.6 Results ...... 164

6.6.1 Bulk rock chemistry ...... 164

6.6.2 Mineral chemistry ...... 165

6.7 Phase equilibria and P–T evolution ...... 169

6.7.1 Pseudosection calculation ...... 169

6.7.2 Pseudosection for eclogite sample KP1–6 ...... 172

6.7.3 Pseudosection for quartz-mica schist sample KP1–4b ...... 174

6.7.4 P–T constraints for marble sample KP1–7a ...... 175

6.8 Discussion ...... 176

6.8.1 Metamorphic evolution for interlayered eclogite, marble and mica schist ...... 176

6.8.2 Southern HP unit vs. northern UHP unit in the Tianshan? ...... 178

6.8.3 Geodynamic implications ...... 179

Acknowledgements ...... 181

References ...... 181

7 Summary and significance ...... 187 Appendix ...... a

xiv

List of Figures

1.1 Diagram of a typical subduction zone…………………………………………2 1.2 Coesite relicts in UHP rocks……………….………….……………………….4 1.3 HP veins in HP/LT rocks….………...……..………….……………………….5

3.1 Geological map of the Tianshan (U)HP/LT belt……………………...………27 3.2 Field photographs of the coexisting eclogites and blueschists……………..…29 3.3 Photomicrographs and BSE images of the eclogite.…………………………31 3.4 Photomicrographs and BSE images of the blueschist…………..……………32 3.5 Chemical compositions of garnet, omphacite and amphibole……...…………35 3.6 P-T pseudosections for eclogite and blueschist………………………….……42 4.1 Photomicrographs and BSE images of dolomite in eclogite…….……………61 4.2 X-ray maps of Ca, Fe, Mg and Mn in the dolomite...…………………………66 4.3 Chemical composition profile of the dolomite………..………………………67 4.4 X-ray maps of Ca, Fe, Mg and Mn of dolomite grains……….………………68 4.5 Trace elements in the zoned matrix dolomite…...…….………………………69 4.6 P–T pseudosection of the eclogite…………………….………………………72 5.1 Geological map of the Tianshan (U)HP/LT belt...…….………………………89 5.2 Outcrop photograph of the eclogite and the relevant vein.……………………90 5.3 Photomicrographs of the host eclogite, selvage and vein………..……………92 5.4 Occurrence of the relictic lawsonite inclusion in eclogites………...…………94 5.5 Mineral inclusions in pyrite in the selvage…………....………………………97 5.6 Diagrams of trace element in samples….……..………………..……………103 5.7 Mg, Fe, Ca and Mn X-ray maps of host eclogite garnet….…………………106 5.8 Mg, Fe, Ca and Mn X-ray maps of selvage and vein garnets………..………107 5.9 Chemical compositions of garnet, omphacite and amphibole….……………109 5.10 REE concentrations of minerals from WE, selvage and vein………..………113 5.11 Metal concentrations of Omp and Amp in WE, selvage and vein…………...116 5.12 Element maps of epidote and dolomite in WE and selvage…………………118 5.13 P–T pseudosection for the host eclogite…………………..…………………125 5.14 Schematic illustration of the formation of the vein-selvage system…………130 5.15 Concentration ratio diagram in the selvage………………………….………134 6.1 Geological map of the Tianshan (U)HP/LT belt.…….………………………159 6.2 Photography and schematic stratigraphic diagram of drill core sample….…160 6.3 Photomicrographs of the interlayered eclogites, marbles and mica-schists…162 6.4 Compositions of garnet, omphacite, amphibole and white mica…….………166 6.5 Calculation method of effective bulk composition……………..……………171 6.6 P–T pseudosections of the eclogite and the mica-schist……………………173 6.7 P–T paths of the eclogite and mica-schist…………………………………177 6.8 Various P–T paths of HP–UHP metamorphic rocks in Tianshan belt.………180

List of Tables

3.1 Major element composition of minerals in the eclogite…………….……36 3.2 Major element composition of minerals in the coexisting blueschist….……38 3.3 Whole-rock compositions of the eclogite and blueschist……..………………40 3.4 Comparisons of calculated results with measured results...... ………………44 4.1 Major element composition of minerals in eclogite……..……………………65 4.2 Trace element composition of dolomite and magnesite………………………68 4.3 Bulk composition for calculation and mineral assembalges…………………73 5.1 Modal abundances of minerals……………………….………………………93 5.2 Bulk major and trace elements of the WE, selvage and the vein…...….……102 5.3 Sr-Nd isotopic compositions of the eclogite, selvage and vein..…….………104 5.4 Major element composition of minerals in the WE, selvage and the vein...…111 5.5 Trace element composition of minerals in the WE, selvage and the vein...…114 5.6 Trace element composition of sulfides in the WE, selvage and the vein……121 5.7 Distributions of metal elements in the wall-rock eclogite...…………………137 6.1 Bulk major and trace element of samples………….………………………165 6.2 Major element composition of minerals in eclogite, marble and schist..……167 6.3 Effective bulk compositions used for pseudosection calculations…..………172 6.4 Temperatures of marble obtained from calcite-dolomite geothermometry….176

xvi

1. Introduction

1 Introduction

1.1 Subduction zone and related rocks

Subduction zones, which constitute the largest recycling system on our planet, play a key role in Earth‘s geodynamics, crustal evolution, as well as earthquakes and volcanism (Stern, 2002). Subduction zones are locations where sediments, oceanic crust, and mantle lithosphere return to and reequilibrate with Earth‘s mantle (Fig. 1-1). Oceanic materials such as pelagic and terrigenous sediments, altered and fresh basaltic oceanic crust, and mantle lithosperher enter the subduction factory as raw materials, and the products are arc magmatism, their solidified materials, and ultimately continental crust (Tatsumi, 2005). Aqueous fluids, which are released from oceanic crustal rocks via dehydration reactions during subduction, play the most critical role in this process (e.g., Austrheim, 1987; McCulloch and Gamble, 1991). Subduction zone-derived magmas (Fig. 1-1) are usually enriched in large ion lithophile elements (LILE) and depleted in high field strength elements (HFSE). These geochemical signatures are believed to be the result of metasomatized mantle wedge melting, which have been caused by infiltrating hydrous fluids that stem from the dehydration of subducted oceanic slabs (Gill, 1981; Tatsumi, 1989; McCulloch and Gamble, 1991). This hypothesis has been confirmed by both the systematic geochemical data obtained from arc igneous rocks (e.g., Ryan et al., 1995; Stolz et al., 1996; Plank and Langmuir, 1993) and high-pressure (HP) rocks from former subduction-zones (e.g., Scambelluri and Philippot, 2001; John et al., 2004; Hermann et al., 2006).

In subduction zones, with increasing temperature and pressure, the water in oceanic crust will be continuously released through a series of dehydration reactions, and the rock is ultimately transformed to eclogite. High-pressure/low-temperature (HP/LT) blueschist- to eclogite-facies terranes are widely regarded to represent exhumed fragments of subducted slabs. Therefore, metamorphic studies of eclogites and related

1. Introduction high-pressure rocks may reveal important information about their P-T evolution and associated tectonometamorphic processes at depth in subduction zones. Furthermore, high-pressure vein networks in eclogite-facies rocks are interpreted to represent fossilized fluid pathways within the subducted slabs and thus, investigations of these veins may provide important constraints on the flow of fluids, fluid-rock interaction processes and related element behavior in deeply subducted oceanic crust, and also provide an alternative mechanism for transporting H2O to deep level in subduction zones (e.g., Becker et al., 1999; Gao and Klemd, 2001; Widmer and Thompson, 2001; Spandler and Hermann, 2006; John et al., 2008).

Figure 1-1 Diagram of a typical subduction zone. Subduction zones are areas where one of the Earth's tectonic plates slips under another, raising mountain ranges along the margin sprinkled with volcanoes (e.g., Stern, 2002).

1.2 HP-UHP metamorphism

With respect to metamorphism, the most important feature of subduction zones is their low geothermal gradient. Thus the main model for subduction-zone metamorphism refers to high-pressure/low-temperature metamorphism taking places in rocks which are part of the down-going slab. High-pressure metamorphism is unique for two reasons:

1. Introduction

Firstly, the high pressures of metamorphism (often accompanied by low temperatures) usually yield interesting mineral assemblages (Ernst, 1971a). The typical HP/LT metabasites include blueschists with diagnostic minerals like glaucophane and lawsonite, and eclogites with diagnostic minerals like garnet, omphacite and rutile (Carswell, 1990). Secondly, the recognition of high-P/T metamorphism is very exciting from a tectonic perspective since it usually points to former presence of subduction zones (e.g., Coleman, 1971; Ernst, 1971b; 1975; Miyashiro, 1972). However, the preservation of high-P/T rocks is relatively rare because minerals stabilized under high P conditions are easily re-equilibrated and/or overprinted during their exhumation, thus the preservation of high-P/T rocks usually requires very rapid decompression during exhumation (e.g., Ernst, 1988).

Ultrahigh-pressure (UHP) metamorphism refers to a metamorphic process that occurs at P greater than ~28 kbar, which generate coesite- and/or diamond-bearing metamorphic rocks or rocks containing equivalent HP mineral assemblages. Since the first coesite discovery in an almost pure pyrope garnet (Fig. 1-2a) in a pyrope- quartzite from the Western Alps of Italy (Chopin, 1984) and in an eclogite from the Western Gneiss Region of Norway (Smith, 1984), more than 20 UHP metamorphic terranes have been reported worldwide (Carswell and Compagnoni, 2003; Liou et al., 2004, 2009), and UHP rocks have attracted intensive research efforts and have evolved as an important tool of geodynamic settings. The formation of UHP minerals requires pressures in excess of that expected within the crust of the Earth, thus it appears that large masses of lighter crustal rock have been subducted into the detpth of Earth‘s upper mantle. Most UHP rocks have undergone extensive retrograde metamorphism and preserve little or no UHP record. Commonly, only a few UHP minerals in eclogites reveal that the entire terrain was subducted to mantle depths. UHP index minerals are very rare and usually found as inclusions in unreactive host grains, such as garnet, jadeitic clinopyroxene or omphacite and zircon (Fig. 1-2; e.g., Chopin, 1984; Smith, 1984; Sobolev and Shatsky, 1990; Ye et al., 2000). Those strong host grains served as effective ―pressure cells‖ that preserved the UHP environment during exhumation of the rocks. Preservation of UHP

1. Introduction rocks appears to stem from their development in protoliths that were rapidly exhumed once buoyant forces took over in the refrigerated subducting plate (Ernst et al., 1997).

Figure 1-2 Coesite relicts in UHP rocks. (a) Relics of coesite (Coe) included in a pyrope (Prp) porphyroblast from Western Alps and partially converted to quartz (Qtz) along rims and fractures. Radial fractures, consequent to volume increase during the coesite to quartz inversion, are developed in the hosting pyrope. (b) Photograph showing coesite (Coe), omphacite (Omp), phengite (Phn) and titanite (Ttn) inclusions in zircon of granitic gneiss from Taohang, Sulu UHP terrane (after Ye et al., 2000).

1.3 Nature of subduction zone fluids

The calc-alkaline volcanic rocks situated along convergent plate margins are charactered by enrichment of LILE and depletion of HFSE. The genesis of these rocks has been attributed to the melting of the mantle wedge, triggered by the influx of subduction-zone fluids (e.g., Tatsumi, 1989; McCulloch and Gamble, 1991). However, the enriched LILE is either believed to stem from fluids and or melts derived from the dehydration and/or melting of sediments and oceanic crusts in subducting slabs or from shallow crustal contamination of arc lavas (Becker et al., 1999). Evidence from systemic geochemical studies of arc basalts and associated altered mantle xenoliths (e.g., Defant et al., 1990; Morris et al., 1990; Maury et al., 1992; Ryan et al., 1995; Schiano et al., 1995), theoretical phase diagram calculation of subduction zone (e.g., Schmidt and Poli, 1998; Hacker et al., 2003) and fluid-rock experiment under high-temperature and high-pressure (e.g., Brenan et al., 1995; Kessel et al., 2005; Manning et al., 2008) suggests that the subduction zone fluids have rather complex components, including aqueous fluid derived from dehydration of basaltic oceanic crust (e.g., Brenan et al., 1995; Ryan et al., 1995; Schmidt and Poli, 1998; Manning, 2004), aqueous fluid

1. Introduction released by devolatilization of sediments (e.g., Morris et al., 1990; Plank and Langmuir, 1993), hydrous melt producted by melting of mafic oceanic crust (e.g., Ryan et al., 1995), melt producted by melting of sediments (e.g., Hawkesworth et al., 1997; Hermann et al., 2006) and supercritical liquids from subducted oceanic crust (e.g., Kessel et al., 2005). However, the composition of subduction fluids obtained from indirect approaches is difficult to testify by direct studies due to the inaccessibility of samples at great depths.

Figure 1-3 HP veins in HP/LT rocks. (a) Outcrop photo of a garnet–quartz–phengite vein with associated green bleach zone in the host eclogite from New Caledonia (after Spandler and Hermann, 2006). (b) Outcrop photo of omphacite-dominated veinnets in the host blueschist from Chinese Tianshan. (c) Outcrop photo of an omphacite-dominated vein in host eclogite-facies rock from Chinese Tianshan. (d) Two generations of eclogitic veins (vein_I and vein_II) cutting through garnet and omphacite-bearing blueschist (after John et al., 2008).

However, HP-UHP metamorphic rocks and associated HP veins (Fig. 1-3) not only record the subduction and exhumation processes of oceanic crust, but also directly document the deep fluid activities in subduction zones; therefore HP-UHP terranes are excellent natural laboratories to study the processes of subduction fluids in the

1. Introduction

‘subduction factory’ (e.g., Hacker et al., 2003). In recent years, research of HP veins in eclogites has become a frontior study in the metamorphic field, and many creative results have been published on various aspects. The subduction-zone derived low salinity aqueous fluid is thought to contain volatiles and major element components including Si, Al, Na and Ca (Manning, 2004; Gao et al., 2007). The trace element budget transported by these fluids is believed to be quite complex. Some studies suggest a decoupling of water and trace element release during the dehydration of oceanic crust (e.g., Hermann et al., 2006; Scambelluri and Philippot, 2001; Spandler et al., 2004), whereas others are more focused on the reactivity of the released aqueous fluids which may allow these fluids to mobilize and transport certain trace elements while they are traveling through the slab towards the mantle wedge (e.g., Manning, 2004; John et al., 2004; Beinlich et al., 2010; Guo et al., 2012; Herms et al., 2012; John et al., 2012). It was reported that subduction-zone fluids may be responsible for the mobilization of considerable amounts of LILE, U and light rare earth elements (LREE) (McCulloch and Gamble, 1991; Hawkesworth et al., 1993; Scambelluri and Philippot, 2001; John et al., 2004; Manning, 2004). In some cases, it was also shown that they can mobilize and transport heavy rare earth elements (HREE) (Spandler and Hermann, 2006; John et al., 2008; Guo et al., 2012) and even HFSE which are usually believed to behave rather immobile in aqueous fluids (e.g., Rubatto and Hermann, 2003; Xiao et al., 2006; Gao et al., 2007; Zhang et al., 2008; Rapp et al., 2010).

1.4 Tianshan (U)HP/LT metamorphic belt

The Tianshan (Ultra)high-pressure/low-temperature [(U)HP/LT] metamorphic belt (extending for 1500 km) represents a subduction/collision zone in the South Tianshan Orogen (e.g., Gao et al., 2009, 2011;). In the Chinese Tianshan, the (U)HP/LT belt extends for at least 200 km along the South Central Tianshan suture and consists of a suite of metasedimentary, mafic and ultramafic rocks (Fig. 3-1a) that were interpreted as a paleo-accretionary wedge. It is mainly composed of blueschist-, eclogite- and greenschist-facies meta-sedimentary rocks and some mafic metavolcanic rocks with

1. Introduction

N-MORB, E-MORB, OIB and arc basalt affinities (Gao and Klemd, 2003) and was considered to represent an oceanic crust subduction zone.

Eclogite was first found in 1995 in Tianshan (U)HP/LT metamorphic belt (Gao, 1997). Eclogites occur in blueschist or mica schist as pods, boudins, thin layers or as massive blocks interpreted to represent a tectonic mélange (Gao et al., 1999; Gao and Klemd, 2003; van der Straaten et al., 2008; Wei et al., 2009). Most eclogites have experienced peak metamorphism estimated to range between 480 and 580 °C at 14–21 kbar at a regional scale (e.g., Klemd et al., 2002; Wei et al., 2003). UHP metamorphism was suggested in 2002 by the evidence from inclusions of coesite pseudomorphs in garnet, quartz exsolution lamellae in omphacite, and magnesite in eclogites (Zhang et al., 2002a, b). This raised a controversial discussion (Klemd, 2003; Zhang et al., 2003) since the evidence was ambiguous in the absence of coesite relicts and/or other UHP minerals. However, UHP metamorphism was subsequently verified by the presence of coesite inclusions in garnet, omphacite in eclogites and host mica schists (Zhang et al., 2005; Lü et al., 2008, 2009). This was confirmed by further identification of more coesite-bearing localities (e.g., Lü and Zhang, 2012; Yang et al., 2013) and by thermodynamic modeling (e.g., Wei et al., 2009; Tian and Wei, 2013). The discovery of coesite indicates that some of the eclogites and metapelites indeed experienced UHP metamorphism in Tianshan (U)HP/LT belt.

The age of peak metamorphism in Tianshan (U)HP/LT belt was also the topic of controversial discussions. It was first constrained at late Paleozoic by Sm–Nd isochron data (Omp-Gln-Grt-whole rock) and 40Ar/39Ar data of crossite yielding an age of 343 ± 44 Ma and 344 ± 1 Ma, respectively, indicating a Carboniferous age (Gao and Klemd, 2003). White mica Ar–Ar and Rb–Sr ages cluster at 310 Ma were interpreted to represent the exhumation stage (Klemd et al., 2005). However, SHRIMP U–Pb ages of 233–226 Ma obtained from zircon rims in eclogites were interpreted as the age of peak metamorphism at the Early Triassic (Zhang et al., 2007). Unfortunately, the absence of index mineral inclusions (such as omphacite, phengite) at zircon rim and ambiguous trace element pattern makes a peak Triassic metamorphic age rather unlikely. Some

1. Introduction researches argued that the new growth of zircon rim may due to a fluid-assisted recrystallization instead of a record of peak metamorphism (de Jong et al., 2009). Actually, a zircon U–Pb concordant crystallization age of 285 Ma obtained from a leucogranite dike crosscutting the Tianshan (U)HP/LT belt (Gao et al., 2011) gives an upper limit of peak metamorphism. The timing of peak metamorphism was determined by U-Pb SIMS ages of ca. 320 Ma on metamorphic zircon rims (containing omphacite inclusions) from eclogites (Su et al., 2010). This age is in agreement, within error, with multi-point Lu-Hf isochron ages from four blueschist- or eclogite-facies rocks yielding consistent garnet-growth ages of ca. 315 Ma (Klemd et al., 2011), SIMS U–Pb rutile age of 318 Ma from eclogite (Li et al., 2011), and SHRIMP zircon U–Pb age of 320 Ma obtained from a UHP metapelite (Yang et al., 2013). Those direct datings of eclogite-facies minerals suggest a Carboniferous subduction and collision.

The Tianshan is the first locality worldwide where primary fluids at blueschist to eclogite transition have been documented (Gao and Klemd, 2001). The ubiquitous presence of high-pressure vein networks in blueschists and eclogites generated during prograde and retrograde metamorphism reveals extensive fluid-rock interaction activities and fluid-mediated mass transport in a former oceanic subduction zone (Gao and Klemd, 2001; Gao et al., 2007; John et al., 2008; van der Straaten et al., 2008; Beinlich et al., 2010; Lü et al., 2012; Klemd, 2013). The major element composition of veins indicates that Si, Na and Ca-rich aqueous fluids were released during the blueschist to eclogite transition (Gao and Klemd, 2001). The presence of rutile in the segregations and transport veins indicates that HFSE such as Ti, Nb, and Ta were mobilized during the dehydration process and were transported over long-distances during the dehydration of blueschist to eclogite (Gao et al., 2007). Fluid-generated dehydration embrittlement of blueschist may lead to the infiltration of an external fluid and produce an eclogite-facies vein, which mobilized significant amounts of LILE, REE, and HFSE during fluid-rock interaction (John et al., 2008). Meanwhile, the chemistry of the fluids infiltrating during retrograde metamorphic overprint indicates that the fluid was derived from serpentinites (van der Straaten et al., 2008) or predominantly derived

1. Introduction from seawater-altered oceanic lithosphere (van der Straaten et al., 2012). In addition, the infiltration of a channelized Ca-rich fluid, caused eclogitization of the immediate wall-rocks, and trigger major and trace element mobilization as well as the decoupling of Zr-Hf from Nb-Ta (Beinlich et al., 2010). Vein formation occurred contemporaneously to peak metamorphism at age of ca. 320 Ma as inferred from U–Pb dating of vein rutile dating (Li et al., 2011) and Rb–Sr vein-whole rock data (John et al., 2012).

1.5 Research approachs and analytial techniques

The study of this Ph.D. thesis began with detailed and extensive field work in the southwest Tianshan, NW China. The sample collection was focused on fresh eclogites and associated veins, blueschists and metapelites. Both the eclogites and host rocks (interlayered blueschists and metapelites) were sampled and their relationship was recorded by sketches and photographs. Several samples were taken from the HP veins considering the rather heterogeneous mineral distributions in the vein. In addition, from the host eclogite towards the vein, a series of samples were collected along a profile in order to study the chemical variations and discuss the possible element mobilizations during fluid-rock interaction (refer to Beinlich et al., 2010). Thin sections of all rocks were made accordingly, as well as for the whole-rock powder of samples. Thin sections were examined by polarizing microscopy. Bulk rock major and trace element contents of all separated rocks were determined by X-ray fluorescence (XRF) and Inductively Coupled Plasma Mass Spectrometry (ICP-MS). Major element compositions of minerals were obtained by electron microprobe analysis (EPMA). X-ray maps for selected elements in certain minerals (such as garnet) were conducted in WDS mode. Trace element compositions of minerals were determined on polished thin sections using a Laser Ablation ICP-MS. To identify some critical minerals (such as lawsonite) Laser-Raman spectroscopy was performed. These data were used in combination with the petrographical observations for reconstruction of the P-T evolution of the rocks. P-T pseudosections, calculated with the thermodynamical calculation program Perple_X

1. Introduction

(Connolly, 1990; 2005), were dominantly used to constrain P-T conditions and paths, allowing the interpretion of the eclogite and related rocks relationship. The results did not only reveal the formation of compositional zoned minerals (for example dolomite in this study), but also provided information on the dehydration processes and the origin of fluids. In fluid-mediated selvages (of the veins), mass-balance calculations were promoted in order to estimate element mass gain or loss and further clarify whether there are systematic variations that occurred during fluid-rock interaction. In combining all the data described above, the author believes (and hopes) that systematic discussions and well-balanced intensive interpretations were implemented on each of the topics above.

References

Austrheim H. (1987) Eclogitization of lower crustal granulites by fluid migration through shear zones. Earth and Planetary Science Letters 81, 221–232. Becker H., Jochum K.P. and Carlson R.W. (1999) Constraints from high-pressure veins in eclogites on the composition of hydrous fluids in subduction zones. Chemical Geology 160, 291–308. Beinlich A., Klemd R., John T. and Gao J. (2010) Trace-element mobilization during Ca-metasomatism along a major fluid conduit: eclogitization of blueschist as a consequence of fluid–rock interaction. Geochimica et Cosmochimica Acta 74, 1892– 1922. Brenan J.M., Shaw H.F. and Ryerson F.J. (1995) Experimental-evidence for the origin of lead enrichment in convergent-margin magmas. Nature 378, 54–56. Carswell D.A. (1990) Eclogites and eclogite facies: definitions and classifications, In: Carswell, D.A. (Eds.), Eclogite Facies Rocks. Blackie and Son Ltd, New York, pp. 1–13. Carswell D.A. and Compagnoni R. (2003) Introduction with review of the definition, distribution and geotectonic significance of ultrahigh pressure metamorphism. In: Carswell D.A. and Compagnoni R. (Eds.), Ultrahigh Pressure Metamorphism. Eotvos University Press, Budapest, pp. 3–9. Chopin C. (1984) Coesite and pure pyrope in high-grade blueschists of the Western Alps: a first record and some consequences. Contributions to Mineralogy and Petrology 86, 107–118. Coleman R.G. (1971) Plate tectonic emplacement of upper mantle peridotites along continental edges. Journal of Geophysical Research 76, 1212–1222. Connolly J.A.D. (1990) Multivariable phase-diagrams: an algorithm based on generalized thermodynamics. American Journal of Science 290, 666–718. Connolly J.A.D. (2005) Computation of phase equilibria by linear programming: A tool

1. Introduction

for geodynamic modelling and its application to subduction zone decarbonation. Earth and Planetary Science Letters 236, 524–541. de Jong K., Wang B., Faure M., Shu L.S., Cluzel D., Charvet J., Ruffet G. and Chen Y. (2009) New 40Ar/39Ar age constraints on the Late Palaeozoic tectonic evolution of the western Tianshan (Xinjiang, northwestern China), with emphasis on Permian fluid ingress. International Journal of Earth Sciences 98, 1239–1258. Defant M.J. and Drummond M.S. (1990) Derivation of some modern arc magmas by melting of young subducted lithosphere. Nature 347, 662–665. Ernst W.G. (1971a) Do mineral parageneses reflect unusually high-pressure conditions of Franciscan metamorphism? American Journal of Science 270, 81–108. Ernst W.G. (1971b) Metamorphic zonations on presumably subducted lithospheric plates from Japan, California and the Alps. Contributions to Mineralogy and Petrology 34, 43–59. Ernst W.G. (1975) Systematics of large-scale tectonics and age progressions in Alpine and Circum-Pacific blueschist belts. Tectonophysics 26, 229–246. Ernst W.G. (1988) Tectonic history of subduction zones inferred from retrograde blueschist P-T paths. Geology 16, 1081–1084. Ernst W.G., Maruyama S. and Wallis S. (1997) Buoyancy-driven, rapid exhumation of ultrahigh-pressure metamorphosed continental crust. Proceedings of the National Academy of Sciences of the United States of America 94, 9532–9537. Gao J. (1997) The discovery of eclogites and its tectonic implication, Southwest Tianshan. Chinese Science Bulletin 42, 737–740 (in Chinese with English abstract). Gao J. and Klemd R. (2001) Primary fluids entrapped at blueschist to eclogite transition: evidence from the Tianshan meta-subduction complex in northwestern China. Contributions to Mineralogy and Petrology 142, 1–14. Gao J. and Klemd R. (2003) Formation of HP–LT rocks and their tectonic implications in the western Tianshan Orogen, NW China: geochemical and age constraints. Lithos 66, 1–22. Gao J., Klemd R., Zhang L., Wang Z. and Xiao X. (1999) P–T path of high-pressure/low-temperature rocks and tectonic implications in the western Tianshan Mountains, NW China. Journal of Metamorphic Geology 17, 621–636. Gao J., John T., Klemd R. and Xiong X.M. (2007) Mobilization of Ti–Nb–Ta during subduction: Evidence from rutile-bearing dehydration segregations and veins hosted in eclogite, Tianshan, NW China. Geochimica et Cosmochimica Acta 71, 4974–4996. Gao J., Long L.L., Klemd R., Qian Q., Liu D.Y., Xiong X.M., Su W., Liu W., Wang Y.T. and Yang F.Q. (2009) Tectonic evolution of the South Tianshan orogen and adjacent regions, NW China: geochemical and age constraints of granitoid rocks. International Journal of Earth Sciences 98, 1221–1238. Gao J., Klemd R., Qian Q., Zhang X., Li J.L. and Jiang T. (2011) The collision between the Yili and Tarim blocks of the Southwestern Altaids: Geochemical and age constraints of a leucogranite dike crosscutting the HP–LT metamorphic belt in the Chinese Tianshan Orogen. Tectonophysics 499, 118–131. Gill J. B. (1981) Orogenic andesites and plate tectonics. Springer Verlag, Berlin. Guo S., Ye K., Chen Y., Liu J., Mao Q. and Ma Y. (2012) Fluid–rock interaction and

1. Introduction

element mobilization in UHP metabasalt: Constraints from an omphacite-epidote vein and host eclogites in the Dabie orogen. Lithos 136, 145–167. Hacker B.R., Abers G.A. and Peacock S.M. (2003) Subduction factory 1. Theoretical

mineralogy, densities, seismic wave speeds, and H2O contents. Journal of Geophysical Research-Solid Earth 108, 2029. Hawkesworth C.J., Gallagher K., Hergt J.M. and McDermott F. (1993) Mantle and slab contributions in arc magmas. Annual Review of Earth and Planetary Sciences 21, 175–204. Hawkesworth C.J., Turner S.P., McDermott F., Peate D.W., vanCalsteren P. (1997) U-Th isotopes in arc magmas: Implications for element transfer from the subducted crust. Science 276, 551–555. Hermann J., Spandler C., Hack A. and Korsakov A.V. (2006) Aqueous fluids and hydrous melts in high-pressure and ultra-high pressure rocks: Implications for element transfer in subduction zones. Lithos 92, 399–417. Herms P., John T., Bakker R. J. and Schenk V. (2012) Evidence for channelized external fluid flow and element transfer in subducting slabs (Raspas Complex, Ecuador). Chemical Geology 310, 79–96. John T., Scherer E.E., Haase K. and Schenk V. (2004) Trace element fractionation during fluid-induced eclogitization in a subducting slab: trace element and Lu–Hf– Sm–Nd isotope systematics. Earth and Planetary Science Letters 227, 441–456. John T., Klemd R., Gao J. and Garbe-Schonberg C.D. (2008) Trace-element mobilization in slabs due to non steady-state fluid-rock interaction: Constraints from an eclogite-facies transport vein in blueschist (Tianshan, China). Lithos 103, 1–24. John T., Gussone N., Podladchikov Y.Y., Bebout G. E., Dohmen R., Halama R., Klemd R., Magna T. and Seitz H.M. (2012) Volcanic arcs fed by rapid pulsed fluid flow through subducting slabs. Nature Geoscience 5, 489–492. Kessel R., Schmidt M.W., Ulmer P. and Pettke T. (2005) Trace element signature of subduction-zone fluids, melts and supercritical liquids at 120-180 km depth. Nature 437, 724–727. Klemd R. (2003) Ultrahigh-pressure metamorphism in eclogites from the western Tianshan high-pressure belt (Xinjiang, western China) - Comment. American Mineralogist 88, 1153–1156. Klemd R. (2013) Metasomatism during high-pressure metamorphism: eclogites and blueschist-facies rocks, In: Harlov, D.E., Austrheim, H. (Eds.), Metasomatism and the chemical transformation of rock. Springer-Verlag Berlin Heidelberg 2012, pp. 351–413. Klemd R., Schröter F.C., Will T.M. and Gao J. (2002) P–T evolution of glaucophane-omphacite bearing HP–LT rocks in the western Tianshan Orogen, NW China: new evidence for 'Alpine-type' tectonics. Journal of Metamorphic Geology 20, 239–254. Klemd R., Bröcker M., Hacker B.R., Gao J., Gans P. and Wemmer K. (2005) New age constraints on the metamorphic evolution of the high-pressure/low-temperature belt in the western Tianshan mountains, NW China. Journal of Geology 113, 157–168. Klemd R., John T., Scherer E.E., Rondenay S. and Gao J. (2011) Changes in dip of

1. Introduction

subducted slabs at depth: Petrological and geochronological evidence from HP–UHP rocks (Tianshan, NW–China). Earth and Planetary Science Letters 310, 9–20. Li Q.L., Lin W., Su W., Li X.H., Shi Y.H., Liu Y. ang Tang G.Q. (2011) SIMS U-Pb rutile age of low-temperature eclogites from southwestern Chinese Tianshan, NW China. Lithos 122, 76–86. Liou J.G., Tsujimori T., Zhang R.Y., Katayama I. and Maruyama S. (2004) Global UHP metamorphism and continental subduction/collision: The Himalayan model. International Geology Review 46, 1–27. Liou J.G., Ernst W.G., Zhang R.Y., Tsujimori T. and Jahn B.M. (2009) Ultrahigh- pressure minerals and metamorphic terranes - The view from China. Journal of Asian Earth Sciences 35, 199–231. Lü Z. and Zhang L.F. (2012) Coesite in the eclogite and schist of the Atantayi Valley, southwestern Tianshan, China. Chinese Science Bulletin 57, 1467–1472. Lü Z., Zhang L.F., Du J.X. and Bucher K. (2008) Coesite inclusions in garnet from eclogitic rocks in western Tianshan, northwest China: Convincing proof of UHP metamorphism. American Mineralogist 93, 1845–1850. Lü Z., Zhang L., Du J. and Bucher K. (2009) Petrology of coesite-bearing eclogite from Habutengsu Valley, western Tianshan, NW China and its tectonometamorphic implication. Journal of Metamorphic Geology 27, 773–787. Lü Z., Zhang L., Du J., Yang X., Tian Z. and Xia B. (2012) Petrology of HP metamorphic veins in coesite-bearing eclogite from western Tianshan, China: Fluid processes and elemental mobility during exhumation in a cold subduction zone. Lithos 136, 168–186. Manning C.E. (2004) The chemistry of subduction-zone fluids. Earth and Planetary Science Letters 223, 1–16. Manning C.E., Wilke M., Schmidt C. and Cauzid J. (2008) Rutile solubility in albite-

H2O and Na2Si3O7-H2O at high temperatures and pressures by in-situ synchrotron radiation micro-XRF. Earth and Planetary Science Letters 272, 730–737. Maury R.C., Defant M.J. and Joron J.L. (1992) Metasomatism of the sub-arc mantle inferred from trace-elements in Philippine xenoliths. Nature 360, 661–663. McCulloch M.T. and Gamble J.A. (1991) Geochemical and geodynamical constraints on subduction zone magmatism. Earth and Planetary Science Letters 102, 358–374. Miyashiro A. (1972) Metamorphism and related magmatism in plate tectonics. American Journal of Science 272, 629–656. Morris J.D., Leeman W.P. and Tera F. (1990) The subducted component in island-arc lavas - constraints from Be isotopes and B-Be systematics. Nature 344, 31–36. Plank T. and Langmuir C.H. (1993) Tracing trace-elements from sediment input to volcanic output at subduction zones. Nature 362: 739–743. Rapp J.F., Klemme S., Butler I.B. and Harley S.L. (2010) Extremely high solubility of rutile in chloride and fluoride-bearing metamorphic fluids: An experimental investigation. Geology 38, 323–326. Rubatto D. and Hermann J. (2003) Zircon formation during fluid circulation in eclogites (Monviso, Western Alps): Implications for Zr and Hf budget in subduction zones. Geochimica et Cosmochimica Acta 67, 2173–2187.

1. Introduction

Ryan J.G., Morris J., Tera F., Leeman W.P. and Tsvetkov A. (1995) Cross-arc geochemical variations in the Kurile arc as a function of slab depth. Science 270, 625–627. Scambelluri M. and Philippot P. (2001) Deep fluids in subduction zones. Lithos 55, 213–227. Schiano P., Clocchiatti R., Shimizu N., Maury R.C., Jochum K.P. and Hofmann A.W. (1995) Hydrous, silica-rich melts in the sub-arc mantle and their relationship with erupted arc lavas. Nature 377, 595–600. Schmidt M.W. and Poli S. (1998) Experimentally based water budgets for dehydrating slabs and consequences for arc magma generation. Earth and Planetary Science Letters 163, 361–379. Smith D.C. (1984) Coesite in clinopyroxene in the Caledonides and its implications for geodynamics. Nature 310, 641–644. Sobolev N.V. and Shatsky V.S. (1990) Diamond inclusions in garnets from metamorphic rocks: a new environment for diamond formation. Nature 343, 742–746. Su W., Gao J., Klemd R., Li J.L., Zhang X., Li X.H., Chen N.S. and Zhang L. (2010) U-Pb zircon geochronology of Tianshan eclogites in NW China: implication for the collision between the Yili and Tarim blocks of the southwestern Altaids. European Journal of Mineralogy 22, 473–478. Spandler C., Hermann J., Arculus R. and Mavrogenes J. (2004) Geochemical heterogeneity and element mobility in deeply subducted oceanic crust; insights from high-pressure mafic rocks from New Caledonia. Chemical Geology 206, 21–42. Spandler C. and Hermann J. (2006) High-pressure veins in eclogite from New Caledonia and their significance for fluid migration in subduction zones. Lithos 89, 135–153. Stern R.J. (2002) Subduction zones. Reviews of Geophysics 40, 1012, doi:1010.1029/ 2001RG000108. Stolz A.J., Jochum K.P., Spettle B. and Hofmann A.W. (1996) Fluid- and melt-related enrichment in the subarc mantle: Evidence from Nb/Ta variations in island-arc basalts. Geology 24, 587–590. Tatsumi Y. (1989) Migration of fluid phases and genesis of basalt magmas in subduction zones. Journal of Geophysical Research 94, 4697–4707. Tatsumi Y. (2005) The subduction factory: how it operates in the evolving Earth. GSA Today 15, 4–10. Tian Z.L. and Wei C.J. (2013) Metamorphism of ultrahigh-pressure eclogites from the Kebuerte Valley, South Tianshan, NW China: phase equilibria and P–T path. Journal of Metamorphic Geology 31, 281–300. van der Straaten F., Schenk V., John T. and Gao J. (2008) Blueschist-facies rehydration of eclogites (Tian Shan, NW–China): Implications for fluid-rock interaction in the subduction channel. Chemical Geology 255, 195–219. van der Straaten F., Halama R., John T., Schenk V., Hauff F. and Andersen N. (2012) Tracing the effects of high-pressure metasomatic fluids and seawater alteration in blueschist-facies overprinted eclogites: Implications for subduction channel processes. Chemical Geology 292–293, 69–87.

1. Introduction

Wei C.J., Powell R. and Zhang L.F. (2003) Eclogites from the south Tianshan, NW

China: petrological characteristic and calculated mineral equilibria in the Na2O– CaO–FeO–MgO–Al2O3–SiO2–H2O system. Journal of Metamorphic Geology 21, 163–179. Wei C.J., Wang W., Clarke G.L., Zhang L.F. and Song S.G. (2009) Metamorphism of High/ultrahigh-pressure Pelitic-Felsic Schist in the South Tianshan Orogen, NW China: Phase Equilibria and P–T Path. Journal of Petrology 50, 1973–1991. Widmer T. and Thompson A.B. (2001) Local origin of high pressure vein material in eclogite facies rocks of the Zermatt-Saas Zone, Switzerland. American Journal of Science 301, 627–656. Xiao Y.L., Sun W.D., Hoefs J., Simon K., Zhang Z.M., Li S.G. and Hofmann A. W. (2006) Making continental crust through slab melting: Constraints from niobium-tantalum fractionation in UHP metamorphic rutile. Geochimica et Cosmochimica Acta 70, 4770–4782. Yang X., Zhang L., Tian Z. and Bader T. (2013) Petrology and U–Pb zircon dating of coesite-bearing metapelite from the Kebuerte Valley, western Tianshan, China. Journal of Asian Earth Sciences 70–71, 295–307. Ye K., Yao Y.P., Katayama I., Cong B.L., Wang Q.C. and Maruyama S. (2000) Large areal extent of ultrahigh-pressure metamorphism in the Sulu ultrahigh-pressure terrane of East China: new implications from coesite and omphacite inclusions in zircon of granitic gneiss. Lithos 52, 157–164. Zhang L., Ellis D.J. and JiangW. (2002a) Ultrahigh pressuremetamorphism in western Tianshan, China, part I: evidences from the inclusion of coesite pseudomorphs in garnet and quartz exsolution lamellae in omphacite in eclogites. American Mineralogist 87, 853–860. Zhang L.F., Ellis D.J., Williams S. and Jiang W.B. (2002b) Ultra-high pressure metamorphism in western Tianshan, China: Part II. Evidence from magnesite in eclogite. American Mineralogist 87, 861–866. Zhang L.F., Ellis D.J., Williams S. and Jiang W.B. (2003) Ultrahigh-pressure metamorphism in eclogites from the western Tianshan, China - Reply. American Mineralogist 88, 1157–1160. Zhang Z.M., Shen K., Sun W.D., Liu Y.S., Liou J.G., Shi C. and Wang J.L. (2008) Fluids in deeply subducted continental crust: Petrology, mineral chemistry and fluid inclusion of UHP metamorphic veins from the Sulu orogen, eastern China. Geochimica et Cosmochimica Acta 72, 3200–3228. Zhang L.F., Song S.G., Liou J.G., Ai Y.L. and Li X.P. (2005) Relict coesite exsolution in omphacite from Western Tianshan eclogites, China. American Mineralogist 90, 181– 186. Zhang L.F., Ai Y.L., Li X.P., Rubatto D., Song B., Williams S., Song S.G., Ellis D. and Liou J.G. (2007) Triassic collision of western Tianshan orogenic belt, China: Evidence from SHRIMP U-Pb dating of zircon from HP/UHP eclogitic rocks. Lithos 96, 266–280.

1. Introduction

2. Aims of the study

2 Aims of the study

2.1 Metamorphism of carbonate-bearing eclogites and blueschists

The Tianshan (U)HP/LT belt is interpreted to constitute a former oceanic subduction zone complex, which includes blueschists, eclogites and metasedimentary rocks. Carbonates are common minerals in some blueschists and eclogites in this terrane. High carbonate contents in oceanic metabasalts suggest that the precursor basaltic crust has undergone significant hydrothermal alteration prior to subduction (e.g., Staudigel, 2003). However, the petrology and metamorphic evolution of carbonate-bearing eclogites and blueschists have not yet been studied in detail in the Tianshan.

Eclogites are commonly interlayered with the blueschists as pods, boudins, thin layers or as massive blocks interpreted to represent a tectonic mélange (Gao et al., 1999; Gao and Klemd, 2003; van der Straaten et al., 2008; Wei et al., 2009). The blueschists were interpreted to have formed under prograde metamorphic conditions or during retrogression of the eclogites (e.g., Du et al., 2011; Gao and Klemd, 2003; Gao et al., 1999; Klemd et al., 2002; Lü et al., 2009; Wei et al., 2003, 2009). However, we investigated recently discovered eclogites and blueschists which are interlayered on a mm- to cm-scale, thereby raising the question about the relationship of those two lithologies as well as controls on the facies-defining mineral assemblages. Therefore in Chapter 3, we present the petrogical and mineral chemical data of coexisting carbonate-bearing eclogite and omphacite-bearing blueschist and use a pseudosection approach to study whether the eclogite and blueschist stabilized under the same metamorphic peak P-T conditions.

Compositional zoning is believed to be a distinctive feature that is commonly observed in petrographical and chemical studies of minerals, and it may provide important

2. Aims of the study information on the minerals‘ growth history and geological conditions during mineral formation. However, to our knowledge the compositional zoning of eclogite-facies dolomite has not been reported and thermodynamically modeled yet. Nonetheless it was already rarely observed in other metamorphic carbonates (cf., Jones and Ghent, 1971; Reinecke et al., 2000). In Chapter 4, we present concentric Fe–Mg zoning of dolomite from a HP lawsonite-carbonate-bearing eclogite. The detailed textural and thermodynamic modeling studies of the dolomite are conducted to study the formation of this compositional zoning as well as carbonate-phase transitions in subducted oceanic crust.

2.2 Process of fluid-rock interaction and transport of metal elements

Subduction zones not only control the crustal evolution, earthquakes and volcanism, but also control the formation of various types of metal ore-deposits (e.g., Sawkins, 1972; Mungall, 2002; Sillitoe, 2008; Sun et al., 2004). However, no consensus has been reached with regards to the source of the giant metal deposits formed in arc settings. The bulk of the metal flux into the arc magma sources is thought to have been predominantly derived from the partial melting of metasomatized asthenospheric mantle triggered by slab fluids (e.g., Pettke et al., 2010; Richards, 2011), whereby some metals may have been transferred from the subducting slab via dehydration fluids to economic arc-associated ore deposits (e.g., Hedenquist and Lowenstern, 1994). A model was proposed for the ore metal recycling process during which subduction-derived fluids transport soluble metals from the slab into the mantle wedge and thus are ultimately responsible for the formation of porphyry copper and epithermal Au deposits in arc magmas (McInnes et al., 1999). Nevertheless, the above hypotheses concerning the mobilization of ore-producing metals in subduction-zone fluids have not been verified yet by the direct study of fossilized subduction-zone derived HP-rocks and associated intra-slab fluid-flow structures such as HP-veins.

2. Aims of the study

Although pyrite has been reported to occur as daughter mineral entrapped in fluid inclusions in vein omphacite (Philippot and Selverstone, 1991) and as matrix mineral coexisting with vein omphacite and the immediate high-pressure host rocks (e.g., Gao and Klemd, 2001; Spandler and Hermann, 2006; Zhang et al., 2008), no detailed investigation has been undertaken concerning the mobilization behavior of transition metals such as Au, Fe, Cu, Ni and Zn during fluid-rock interaction in subduction zones. In Chapter 5, we report petrological and geochemical data from a sulfide-bearing high-pressure vein, its selvage and host eclogites (wall rock eclogite) from the Tianshan (U)HP/LT belt in order to discuss the mobilization of both transition metals and trace elements in subduction-zone fluids.

2.3 Tectonic relationship between mafic and metasedimentary rocks in Tianshan

In many metamorphic terranes worldwide, mafic eclogites commonly occur as pods, boudins, lenses or interlayers in surrounding metasediments and are interpreted as foreign tectonic slices or as integral part of a lithostratigraphic sequence (e.g., Ernst, 1970; Carswell, 1990; Federico et al., 2007; Davis and Whitney, 2008). In general, eclogite boudins/lenses record higher P–T conditions than surrounding metasedimentary country rocks, which is mainly based on the competence contrast of both rock types. However, some studies revealed that some metasediments also experienced HP to UHP metamorphism and shared a common metamorphic history with the enclosed/ interlayered eclogites (e.g., Spear and Franz, 1986; Klemd et al., 1991, 1994; Liu et al., 2001; Gross et al., 2008), suggesting that they may be subducted and exhumed as a coherent units. Thus, it is worthy to investigate the metamorphic evolution of both eclogite lenses/interlayers and associated host metasediments. These informations play an important role in understanding HP–UHP metamorphism during subduction and exhumation processes.

In addition, a recent study proposed two separate tectonic units, namely an internally

2. Aims of the study coherent UHP unit in the north and a coherent HP unit in the south, in the Chinese part of the Tianshan metamorphic belt. However, this ‗tectonic‘ division is based exclusively on thermodynamical modeling while a major tectonic lineament separating these two units was not found (Lü et al., 2012, 2013). Thus, further tectonic and petrological studies are still essentially needed in order to testify the reliability of the ‗tectonic division‘ model since several studies already reported HP condition for eclogites and blueschists from the proposed UHP unit (e.g., Gao et al., 2007; Li et al., 2013; Wei et al. 2009). Chapter 6 deals with eclogites intercalated with mica schists and marble layers in the ‗UHP unit‘, thereby providing a good opportunity to either test the hypothesis of coherent UHP vs. HP units by studying their respective metamorphic evolutions or whether the interlayered eclogites and surrounding metasediments have experienced the same metamorphic evolution as internally coherent unit.

References

Carswell D.A. (1990) Eclogites and eclogite facies: definitions and classifications, In: Carswell, D.A. (Ed.), Eclogite Facies Rocks. Blackie and Son Ltd, New York, pp. 1– 13. Du J., Zhang L., Lü Z. and Chu X. (2011) Lawsonite-bearing chloritoid-glaucophane schist from SW Tianshan, China: phase equilibia and P-T path. Journal of Asian Earth Sciences 42, 684–693. Davis P.B. and Whitney D.L. (2008) Petrogenesis and structural petrology of high-pressure metabasalt pods, Sivrihisar, Turkey. Contributions to Mineralogy and Petrology 156, 217–241. Ernst W.G. (1970) Tectonic Contact between the Franciscan Mélange and the Great Valley Sequence—Crustal Expression of a Late Mesozoic Benioff Zone. Journal of Geophysical Research 75, 886–901. Federico L., Crispini L., Scambelluri M. and Capponi G. (2007) Ophiolite melange zone records exhumation in a fossil subduction channel. Geology 35, 499–502. Gao J. and Klemd R. (2001) Primary fluids entrapped at blueschist to eclogite transition: evidence from the Tianshan meta-subduction complex in northwestern China. Contributions to Mineralogy and Petrology 142, 1–14. Gao J. and Klemd R. (2003) Formation of HP–LT rocks and their tectonic implications in the western Tianshan Orogen, NW China: geochemical and age constraints. Lithos 66, 1–22. Gao J., Klemd R., Zhang L., Wang Z. and Xiao X. (1999) P-T path of high-pressure/low-temperature rocks and tectonic implications in the western Tianshan Mountains, NW China. Journal of Metamorphic Geology 17, 621–636.

2. Aims of the study

Gao J., John T., Klemd R. and Xiong X.M. (2007) Mobilization of Ti–Nb–Ta during subduction: Evidence from rutile-bearing dehydration segregations and veins hosted in eclogite, Tianshan, NW China. Geochimica et Cosmochimica Acta 71, 4974–4996. Gross J., Burchard M., Schertl H.P. and Maresch W.V. (2008) Common high-pressure metamorphic history of eclogite lenses and surrounding metasediments: a case study of calc-silicate reaction zones (Erzgebirge, Germany). European Journal of Mineralogy 20, 757–775. Hedenquist J.W. and Lowenstern J.B. (1994) The role of magmas in the formation of hydrothermal ore deposits. Nature 370, 519–527. Jones J.W. and Ghent E.D. (1971) Zoned siderite porphyroblasts from the Esplanade range and northern Dogtooth mountains, British Columbia. American Mineralogist 56, 1910–1916. Klemd R., Matthes S. and Okrusch M. (1991) High-pressure relics in meta-sediments intercalated with the Weissenstein eclogite, Munchberg gneiss complex, Bavaria. Contributions to Mineralogy and Petrology 107, 328–342. Klemd R., Matthes S. and Schussler U. (1994) Reaction textures and fluid behaviour in very high pressure calc-silicate rocks of the Münchberg gneiss complex, Bavaria, Germany. Journal of Metamorphic Geology 12, 735–745. Klemd R., Schröter F.C., Will T.M. and Gao J. (2002) P–T evolution of glaucophane-omphacite bearing HP–LT rocks in the western Tianshan Orogen, NW China: new evidence for 'Alpine-type' tectonics. Journal of Metamorphic Geology 20, 239–254. Li J.L., Gao J., John T., Klemd R. and Su W. (2013) Fluid-mediated metal transport in subduction zones and its link to arc-related giant ore deposits: Constraints from a sulfide-bearing HP vein in lawsonite eclogite (Tianshan, China). Geochimica et Cosmochimica Acta 120, 326–362. Liu J.B., Ye K., Maruyama S.N., Cong B.L. and Fan H.R. (2001) Mineral inclusions in zircon from gneisses in the ultrahigh-pressure zone of the Dabie Mountains, China. Journal of Geology 109, 523–535. Lü Z., Zhang L., Du J. and Bucher K. (2009) Petrology of coesite-bearing eclogite from Habutengsu Valley, western Tianshan, NW China and its tectonometamorphic implication. Journal of Metamorphic Geology 27, 773–787. Lü Z., Bucher K., Zhang L. and Du J. (2012) The Habutengsu metapelites and metagreywackes in western Tianshan, China: metamorphic evolution and tectonic implications. Journal of Metamorphic Geology 30, 907–926. Lü Z., Bucher K. and Zhang L. (2013) Omphacite-bearing calcite marble and associated coesite-bearing pelitic schist from the meta-ophiolitic belt of Chinese western Tianshan. Journal of Asian Earth Sciences 76, 37–47. McInnes B.I.A., McBride J.S., Evans N.J., Lambert D.D. and Andrew A.S. (1999) Osmium isotope constraints on ore metal recycling in subduction zones. Science 286, 512–516. Mungall J.E. (2002) Roasting the mantle: Slab melting and the genesis of major Au and Au-rich Cu deposits. Geology 30, 915–918. Pettke T., Oberli F. and Heinrich C.A. (2010) The magma and metal source of giant

2. Aims of the study

porphyry-type ore deposits, based on lead isotope microanalysis of individual fluid inclusions. Earth and Planetary Science Letters 296, 267–277. Philippot P. and Selverstone J. (1991) Trace-element-rich brines in eclogitic veins – implications for fluid composition and transport during subduction. Contributions to Mineralogy and Petrology 106, 417–430. Reinecke T., Bernhardt H.J. and Wirth R. (2000) Compositional zoning of calcite in a high-pressure metamorphic calc-schist: clues to heterogeneous grain-scale fluid distribution during exhumation. Contributions to Mineralogy and Petrology 139, 584–606. Richards J.P. (2011) Magmatic to hydrothermal metal fluxes in convergent and collided margins. Ore Geology Reviews 40, 1–26. Sawkins F.J. (1972) Sulfide ore-deposits in relation to plate tectonics. Journal of Geology 80, 377–397. Sillitoe R.H. (2008) Major gold deposits and belts of the North and South American Cordillera: Distribution, tectonomagmatic settings, and metallogenic considerations. Economic Geology 103, 663–687. Spandler C. and Hermann J. (2006) High-pressure veins in eclogite from New Caledonia and their significance for fluid migration in subduction zones. Lithos 89, 135–153. Spear F.S. and Franz G. (1986) P-T evolution of metasediments from the Eclogite Zone, south-central Tauern-Window, Austria. Lithos 19, 219–234. Staudigel H. (2003) Hydrothermal alteration processes in the Oceanic crust. In H. Holland and K. Turekian, Eds., The Crust 3, 511–537, Treatise on Geochemistry, Elseiver, New York. Sun W.D., Arculus R.J., Kamenetsky V.S. and Binns R.A. (2004) Release of gold-bearing fluids in convergent margin magmas prompted by magnetite crystallization. Nature 431, 975–978. van der Straaten F., Schenk V., John T. and Gao J. (2008) Blueschist-facies rehydration of eclogites (Tian Shan, NW–China): Implications for fluid-rock interaction in the subduction channel. Chemical Geology 255, 195–219. Wei C.J., Powell R. and Zhang L.F. (2003) Eclogites from the south Tianshan, NW

3. Coexisting eclogite and blueschist

3 Coexisting carbonate-bearing eclogite and blueschist in SW Tianshan, China: Petrology and phase equilibria

Ji-Lei Lia, Reiner Klemda, Jun Gaob, Melanie Meyera aGeoZentrum Nordbayern, Universität Erlangen-Nürnberg, Schlossgarten 5a, D-91054 Erlangen, Germany bKey Laboratory of Mineral Resources, Institute of Geology and Geophysics, Chinese Academy of Sciences, P.O. Box 9825, Beijing 100029, China

3.1 Abstract

Carbonate-bearing eclogite and blueschist coexist in the same lithological sequence in the Tianshan (ultra-)high-pressure/low-temperature metamorphic belt, northwestern China. Both of them consist of the mineral assemblage garnet + omphacite + dolomite ± magnesite + phengite ± paragonite + epidote/clinozoisite + glaucophane ± barroisite/ (Mg-)katophorite + rutile/titanite + pyrite. The eclogite has an omphacite content of ca. 43 vol.% and a glaucophane content of < 1 vol.%. In contrast, the blueschist has an omphacite content of < 8 vol.% and a glaucophane content of ca. 43 vol.%. The blueschist occurs as bands or interlayers in the eclogite. Lawsonite pseudomorphs of epidote/clinozoisite + paragonite assemblages in garnet are commonly observed in the eclogite and blueschist. High carbonate contents in oceanic metabasalts suggest that the precursor basaltic crust has undergone significant hydrothermal alteration prior to subduction. Mineral assemblages and textures show that the omphacite-carbonate- bearing blueschist and the carbonate-bearing eclogite underwent an identical metamorphic evolution. Phase equilibrium modeling in the TiNCaKFMASCHO system further indicates that both the carbonate-bearing eclogite and the omphacite-

3. Coexisting eclogite and blueschist bearing blueschist equilibrated at peak metamorphic conditions of 540 - 565 oC and

21.8 - 23.1 kbar. Under these conditions calculated XCO2 isopleths reveal the presence of a mainly aqueous fluid phase (XCO2 ≤ 0.02) for the coexisting eclogite and omphacite-bearing blueschist. The stability of the high-pressure mineral assemblages in the eclogite and the omphacite-bearing blueschist is due to differences of the respective bulk-rock composition (the CaO content in particular). This study suggests that three different kinds of blueschist were formed during subduction and exhumation of the South Tianshan oceanic crust, including prograde blueschist formed prior to eclogitization, retrograde blueschist formed by rehydration of eclogite during exhumation in the subduction channel, and coexisting blueschist with eclogite under peak metamorphic conditions.

3.2 Introduction

Low temperature mafic eclogite commonly occurs as pods, boudins, lenses or interlayers in association with blueschist in many metamorphic high-pressure (HP) terranes worldwide and is thought to be the result of subduction-related plate convergence (e.g., Carswell, 1990; Davis and Whitney, 2008; Ernst, 1970; Gao et al., 1999). The blueschists in relation to the associated eclogites are usually interpreted to have formed during the prograde P-T path of the latter prior to peak metamorphic conditions (e.g., Beinlich et al., 2010; Caron and Pequignot, 1986; Gao and Klemd, 2001; Maruyama et al., 1996; Ridley, 1984), or may have formed by rehydration during the retrograde path of the eclogites (e.g., Carson et al., 2000; Davis and Whitney, 2008; Ernst, 1988; van der Straaten et al., 2008). However, the coexistence of omphacite-garnet bearing blueschist and eclogite under the same peak metamorphic conditions has been recognized in some high-pressure belts and is considered to be the result of different bulk-rock compositions (e.g., El-Shazly et al., 1997; Gomez Pugnaire et al., 1997; Oh et al., 1991; Schliestedt, 1986; Vitale Brovarone et al., 2011). Thermodynamic modeling also confirms the coexistence of eclogite- and blueschist-facies mineral assemblages due to differences in the bulk-rock composition

3. Coexisting eclogite and blueschist

(Vitale Brovarone et al., 2011; Wei and Clarke, 2011).

The Tianshan (ultra-)high-pressure/low-temperature ((U)HP/LT) belt is interpreted to constitute a former oceanic subduction zone complex, which includes blueschists, eclogites and metasedimentary rocks. Eclogites are commonly interlayered with the blueschists as pods, boudins, thin layers or as massive blocks interpreted to represent a tectonic mélange (Gao et al., 1999; Gao and Klemd, 2003; van der Straaten et al., 2008; Wei et al., 2009). The blueschists were interpreted to have formed under prograde metamorphic conditions or during retrogression of the eclogites (e.g., Du et al., 2011; Gao and Klemd, 2003; Gao et al., 1999; Klemd et al., 2002; Lü et al., 2009; Wei et al., 2003, 2009). A further type of eclogite forming selvages around crosscutting high-pressure veins in a prograde blueschist host was interpreted to be the result of a change in the Ca-concentration of the bulk rock composition (Beinlich et al., 2010). In the present study, however, we investigated recently discovered eclogites and blueschists which are interlayered on a mm- to cm-scale, thereby raising the question about the relationship of those two lithologies as well as controls on the facies-defining mineral assemblages.

In this article, we present the petrology of coexisting carbonate-bearing eclogite and omphacite-bearing blueschist (in the following referred to as eclogite and blueschist) in the Tianshan (U)HP/LT metamorphic belt, SW China. The petrography and mineral chemistry of the eclogite and blueschist were studied in detail and their P-T conditions and metamorphic evolution were modeled using a pseudosection approach in the TiNCaKFMASCHO system. These data were used to discuss the compositional effect on eclogite- and blueschist-facies mineral assemblages under the same P-T conditions in the Tianshan (U)HP/LT metamorphic belt. The thermodynamic modeling of the eclogite and blueschist sheds some light on the phase transitions which are believed to occur in seafloor altered oceanic metabasalts during subduction.

3.3 Geological setting

3. Coexisting eclogite and blueschist

The southern Tianshan Orogen is situated along the southwest margin of the Central Asian Orogenic Belt and records the final collision between the Tarim and Yili cratons (Gao et al., 1998, 2009; Han et al., 2011; Long et al., 2011; Xiao et al., 2009). The Chinese Tianshan (U)HP/LT metamorphic belt extends for at least 200 km along the South Central Tianshan suture and is thought to represent an accretionary wedge (Fig. 3-1a; Gao and Klemd, 2003). The (U)HP/LT belt is separated from the high-grade belt to the north and low-grade belt to the south by large-scale ductile shear zones. To the north bounded by the South Central Tianshan suture, the Central Tianshan Arc terrane is mainly composed of amphibolite- and granulite-facies rocks, island-arc-type volcanics and volcaniclastic rocks. To the south, the (U)HP/LT belt is overlain by a succession of Paleozoic sedimentary strata, which are considered to represent the passive continental margin of the Tarim block (e.g., Allen et al., 1992). It is mainly composed of greenschist-facies metasedimentary rocks and interlayered blueschists and eclogites, which show N-MORB, E-MORB, OIB and arc basalt affinities (Gao and Klemd, 2003; John et al., 2008). The entire sequence is interpreted to represent a tectonic mélange (Gao et al., 1995, 1999; van der Straaten et al., 2008; Wei et al., 2009). The greenschist-facies metasedimentary rocks are suggested to have been subducted with the metabasalts to depths of ca. 70 km and then rapidly exhumed suffering extensive retrogression (Wei et al., 2009). Eclogites intercalated with blueschist layers as pods, boudins, thin layers or large massive blocks (Fig. 3-1b) are scattered in the greenschist-facies metasediments and display only a minor retrograde overprint. Two distinct types of blueschist were identified: the first prograde type lacks a foliation, may contain omphacite and (retrograde) albite is rare or absent, while the second retrograde blueschist-type is albite-bearing, contains no/little omphacite and shows a well-preserved foliation (Gao and Klemd, 2003). The timing of peak metamorphism was determined by multi-point Lu-Hf isochron ages from four blueschist- or eclogite-facies rocks yielding consistent garnet-growth ages of ca. 315 Ma (Klemd et al., 2011). This age is in agreement, within error, with U-Pb SIMS ages of ca. 320 Ma on metamorphic zircon rims from eclogites (Su et al., 2010). White mica 40Ar-39Ar and Rb-Sr ages of eclogite-facies metavolcanic rocks and omphacite-bearing

3. Coexisting eclogite and blueschist blueschists cluster at ca. 311 Ma and were interpreted to represent a major post-eclogite-facies episode of cooling or recrystallization (Klemd et al., 2005).

Figure 3-1 Geological map of the western Tianshan (U)HP/LT metamorphic belt in northwestern China, modified after van der Straaten et al. (2008). a: Regional tectonic map; b: Local geological map showing the area where samples were taken. Coesite-bearing eclogite or albite-schist localities from Lü et al. (2008, 2009, 2012a, b) are indicated by red stars.

Peak metamorphic conditions of most eclogites were estimated to range between 480 and 580 °C at 14-21 kbar at a regional scale (e.g., Klemd et al., 2002; Wei et al., 2003). However, ultrahigh-pressure (UHP) peak metamorphism was reported for some eclogites and controversially discussed (e.g., Klemd, 2003; Zhang et al., 2003a). Relict coesite exsolution in omphacite from one eclogite sample was found and regarded as

3. Coexisting eclogite and blueschist evidence for UHP conditions in the entire Tianshan high-pressure belt (Zhang et al., 2005). The occurrences of quartz or coesite exsolution in eclogitic pyroxenes are reported in many HP-UHP localities worldwide, but their geobarometric interpretation remains problematic (e.g., Liou et al., 2009). The description of minute, relict coesite exsolution in eclogitic pyroxenes from the Chinese Western Tianshan in particular is unreliable due to problems with the estimation of the Ca-Eskola-component (Klemd, 2003; Page et al., 2005), the mechanisms generating the silica lamellae in those eclogitic pyroxenes (Day and Mulcahy, 2007) and the absence of Raman spectra of standard omphacite from the Tianshan host eclogites (cf., Fig. 2e in Zhang et al., 2005). However, UHP metamorphism was confirmed by the presence of coesite in garnet, and P-T conditions were estimated to 27-33 kbar at 570-630 °C for eclogitic mica schists and 24-27 kbar at 470-510 °C for eclogites from several localities in the Tianshan area (Lü et al., 2008, 2009, 2012a). Thus, the intimate interlayering of HP and UHP rocks on a meter scale (Lü et al., 2009) was interpreted to be due to juxtaposition processes during subduction and exhumation in the subduction channel (Klemd et al., 2011). However, the Tianshan blueschists are less intensively investigated in contrast to the eclogites. The peak metamorphic conditions of some prograde blueschists were estimated to range between 530 and 550 °C at 19 kbar (John et al., 2008), 480 and 540 °C at 21 ± 1.5 kbar (Beinlich et al., 2010), 499 and 519 °C at 23.2-25.3 kbar (Du et al., 2011), while equilibrium P-T conditions of 580 ± 10 °C at 15 ± 1 kbar (van der Straaten et al., 2008) and of 430-530 °C at 15-20 kbar (Du et al., 2011) were calculated for retrograde blueschists. In addition, carbonate-bearing eclogites and blueschists are commonly observed in the Tianshan.

3.4 Sample description and petrography

In the Tianshan (U)HP/LT metamorphic belt, in places, eclogites and blueschists occasionally are intimately interlayered on a cm-scale and display textural characteristics of coexistence, instead of prograde or retrograde transformations. In some cases, glaucophane-free eclogites alternate with omphacite-free blueschists

3. Coexisting eclogite and blueschist which are separated by sharp recrystallized interfaces indicating different protoliths (Fig. 3-2a and c). In other cases, the contact zone between omphacite-bearing blueschists and glaucophane-bearing eclogites is transitional, indicating a heterogeneous chemical composition of a common protolith (Fig. 3-2b, d, e). The here investigated samples were collected from one large eclogite block (0.5 × 1 m) which includes an omphacite-bearing blueschist band (Fig. 3-2d and e) from the upper part of Atantayi River, a branch of the Kepuerte River (Fig. 3-1b).

Figure 3-2 Field photographs and microphotographs of the coexisting eclogites and blueschists. BS = blueschist; BS-EC= blueschist-eclogite transition zone; EC = eclogite. (a) Parallel interlayers of glaucophane-free eclogites and omphacite-free blueschists with rather sharp boundaries. (b) Transitional zone of amphibole-dominated high-pressure bands in eclogitic rocks. (c) Omphacite-dominated eclogite interbedded in garnet-bearing glaucophanite with sharp boundaries. (d) Carbonate-bearing eclogite contains an omphacite-bearing blueschist band with transitional boundaries. (e) Microphotograph taken from (d) showing the relationship between the carbonate-bearing eclogite and omphacite-bearing blueschist.

A ~20 cm wide omphacite-carbonate-bearing blueschist band is interbedded in a massive carbonate-bearing eclogite host (Fig. 3-2d). Generally, the contact zone

3. Coexisting eclogite and blueschist between the blueschist and the eclogite is transitional, and the omphacite content decreases and the glaucophane content increases from the eclogite towards the blueschist (Fig. 3-2d and e). Eclogite and blueschist have the same mineral assemblage which, however, varies significantly concerning the modal content. The main prograde and peak mineral phases are garnet, omphacite, glaucophane, dolomite, epidote, phengite, paragonite and rutile, while Ca-amphiboles and titanite are post peak or retrograde phases.

3.4.1 Eclogite

The eclogite consists of garnet (ca. 18%), omphacite (ca. 43%), dolomite (ca. 18%), phengite (ca. 14%), epidote (ca. 3%), glaucophane (ca. 1%), quartz (ca. 1%) and accessory minerals such as chlorite, rutile, titanite and pyrite (Fig. 3-3a and b). Dolomite bands in association with phengite define a weak foliation in the eclogite. Garnet porphyroblasts, which usually have an inclusion-free rim, contain omphacite, dolomite, rutile, apatite and occasionally chlorite as inclusions in the mantle-core region (Fig. 3-3). In addition, box-shaped clinozoisite/epidote + paragonite assemblages probably after lawsonite also are common inclusions in garnet (Fig. 3-3c and d). Coesite relics or quartz pseudomorphs after coesite were not found in the eclogite and blueschist. Omphacite is the main matrix mineral and also occurs as inclusion in the mantle zone of the garnet (Fig. 3-3a and b). The BSE image reveals two different omphacites namely a lighter type (Omp1) and a darker type (Omp2) in the eclogite (Fig. 3-3c). Omp1 occurs as euhedral matrix mineral and as inclusion in dolomite and garnet, and in places it is intergrown and replaced along the rims by

Omp2 (Fig. 3-3c). Glaucophane mostly occurs in the matrix and occasionally shows a retrograde barroisitic rim; it also is found as inclusion in the garnet cores. Rutile occurs as the main Ti-phase, and is occasionally rimmed by titanite.

3.4.2 Blueschist

The blueschist is composed of garnet (ca. 23%), glaucophane (ca. 43%), omphacite

3. Coexisting eclogite and blueschist

Figure 3-3 Representative photomicrographs and back scattered electron (BSE) images of the the carbonate-bearing eclogite. (a, b) Structure and mineral assemblages of carbonate-bearing eclogite. The garnet core contains lawsonite pseudomorphs and abundant omphacite inclusions in the garnet mantle, with an inclusion-free (except for rutile) garnet rim. Grt=garnet, Omp=omphacite, Gln=glaucophane, Phn=phengite, Dol=dolomite, Cz=clinozoisite, Pg=paragonite, Ep=epidote and Py=pyrite. (c) Garnet porphyroblast with omphacite and dolomite inclusions and clinozoisite/epidote+paragonite assemblages as pseudomorphs after lawsonite. Two types of matrix omphacite are distinguished namely light Omp1 and dark intergranular Omp2. Dolomite commonly shows a dark rim and a light core. (d) Garnet with clinozoisite/epidote+paragonite pseudomorphs after lawsonite.

(ca. 8%), dolomite (ca. 11%), phengite (ca. 13%), minor calcic amphibole and epidote, as well as accessory minerals like rutile, titanite, apatite and pyrite (Fig. 3-4a). Orientated dolomite, glaucophane and phengite crystals mark a weak foliation. Garnet porphyroblasts contain dolomite, omphacite, glaucophane (occasionally rimmed by calcic amphibole) and rutile inclusions (Fig. 3-4b, e, h). The euhedral omphacite and glaucophane inclusions in the garnet have straight grain boundaries and appear to be in textural equilibrium (Fig. 3-4b). Clinozoisite/epidote + paragonite box-shaped inclusions which are believed to be pseudomorphs after lawsonite (see above) were

3. Coexisting eclogite and blueschist also observed in blueschist garnets (Fig. 3-4e). Euhedral matrix glaucophane has a size

3. Coexisting eclogite and blueschist

◀ Figure 3-4 Representative photomicrographs and BSE images of the omphacite-bearing blueschist. (a) Structure and mineral assemblages of omphacite-bearing blueschist. The matrix mineral assemblage mainly consists of glaucophane, phengite, dolomite, omphacite and minor Ti-phases. Glaucophane occasionally has a barroisite rim. Bar=barroisite. (b) Garnet porphyroblast contains glaucophane, omphacite and dolomite inclusions. Glaucophane inclusions in garnet are generally euhedral and show textual equilibrium with omphacite. (c) Euhedral glaucophane contains small omphacite grains and shows textual equilibrium with omphacite in the matrix. (d) Omphacite contains glaucophane inclusions and displays glaucophane-omphacite transformation texture. (e) Garnet porphyroblast in blueschist contains clinozoisite/epidote + paragonite assemblages as pseudomorphs after lawsonite. (f, g) Euhedral glaucophane, either in matrix or as inclusions in omphacite, displays the same compositional zoning, consisting of light core, dark mantle and light rim. Matrix omphacite shows light Omp1 and dark intergranular Omp2. (g) Omphacite inclusions in glaucophane and vice versa in the blueschist matrix. (h) Garnet includes mineral assemblages of omphacite, phengite, barroisite, quartz and euhedral glaucophane, which either displays the same compositional zoning as the glaucophane in matrix. of 1-3mm in its longest dimensions (Fig. 3-4a) and contains small omphacite inclusions (Fig. 3-4c and g). BSE images commonly show that both the glaucophane inclusions in garnet and matrix glaucophane have an identical compositional zoning constituting a light core, a dark mantle and a light thin rim (Fig. 3-4f and g). Matrix omphacite also contains glaucophane inclusions (Fig. 3-4d and g) indicating that both minerals have grown contemporaneously. The common absence of replacement textures such as sutured and curved grain boundaries between omphacite and glaucophane indicates that they are in textural equilibrium (Fig. 3-4c). Omphacite in the blueschist also occurs as lighter type (Omp1) and darker type (Omp2) omphacite (Fig. 3-4f). In places, glaucophane is occasionally rimmed by calcic amphiboles like barroisite and/or katophorite (Fig. 3-4a and 5d) and rutile by titanite (Fig. 3-4f and g).

3.5 Analytical methods

Two representative samples, one eclogite and one blueschist, were selected for a detailed mineral chemical study. Minerals were analyzed with a JEOL JXA-8200 electron microprobe at GeoZentrum Nordbayern of the University Erlangen-Nürnberg, Erlangen, Germany. The accelerating voltage was 15 kV with 15 nA beam current, 3 μm beam spot, and 10-30s counting time. Natural minerals and synthetic oxides were used as standards, and a program based on the ZAF procedure was used for data

3. Coexisting eclogite and blueschist correction. Representative microprobe analyses for the two samples are presented in Table 3-1 and 2.

Major element compositions were determined by X-ray fluorescence spectrometry (XRF) on fused glass disks, using a Rigaku100e at the Guangzhou Institute of Geochemistry, Chinese Academy of Sciences. Loss of ignition (LOI) was determined prior to major element analyses using a pre-ignition method after heating the samples to 1000 °C. Analyses of rock standard reference materials (GSR-1, GSR-2, and GSR-3) indicated that analytical uncertainties for most major oxides are < 2%, for MnO and

P2O5 are < 5%, for the totals are within 100 ± 1%.

3.6 Mineral chemistry

3.6.1 Garnet

Idioblastic garnet in the eclogite shows distinctive prograde growth zoning with a core composition of Alm65-67Prp7-9Grs22-24Sps2-4 with XFe of 0.89-0.92 (XFe = Fe/(Fe+Mg)), a mantle composition of Alm60-65Prp8-10Grs22-26Sps2-4 with XFe of 0.84-0.88 and a rim composition of Alm58-60Prp11-13Grs27-28Sps1-2 with XFe of 0.82-0.84 (Table 3-1). The almandine and spessartine contents decrease from core to rim while the grossular and pyrope contents increase indicating prograde growth (Fig. 3-5a and e).

The garnet in the blueschist shows a somewhat different compositional zoning. The garnet core composition is Alm~66Prp~8.5Grs~23Sps~1.9 with XFe of 0.89, the mantle composition Alm65-68Prp10-11.5Grs19-22Sps1.4-1.7 with XFe of 0.85-0.87 and the rim composition Alm60-64Prp11-13Grs21-25Sps0.8-1.8 with XFe of 0.84-0.85 (Table 3-2). In contrast to the eclogite garnet, the almandine content slightly increases from core to mantle and drastically decreases at the rim, while the grossular content decreases slightly from core to mantle and drastically increases at the rim. The pyrope content continuously increases from core to rim, while the spessartine content may slightly decrease from core to rim (Fig. 3-5b, c and e).

3. Coexisting eclogite and blueschist

Figure 3-5 The chemical compositions of garnet, omphacite and amphibole. (a) Compositional zoning of garnet in the eclogite. Alm=almandine, Gross=grossular, Spess=spessartine. (b, c) Zoning profiles of garnet in the omphacite-bearing blueschist. (d) Zoning profile of XGln through glaucophane with a barroisite/katophorite rim in the omphacite-bearing blueschist. (e) The chemical composition of garnet displays that garnet cores have higher almandine and spessartine, lower grossular and pyrope components than the garnet rims. Ec = eclogite; Bs = omphacite-bearing blueschist. (f) Compositional ternary classification diagram of omphacite after Morimoto et al. (1988). (g, h, i) Chemical classification of amphibole in the eclogite and omphacite-bearing blueschist (after Leake et al., 1997).

3. Coexisting eclogite and blueschist

3.6.2 Omphacite

The chemical composition of the omphacite from the eclogite and the blueschist is similar (Fig. 3-5f). The omphacite types Omp1 and Omp2 with different BSE images are also different in composition (Table 3-1 and 2). Omp1 has higher FeO (7.81 wt.%),

Na2O (8.43 wt.%) contents and lower MgO (5.98 wt.%), CaO (10.93 wt.%) contents compared to Omp2 (FeO = 5.84 wt.%, Na2O = 7.29 wt.%, MgO = 7.85 wt.% and CaO

= 12.89 wt.%). In addition Omp1 has a higher jadeite component (e.g., 46.1) than Omp2 (41.8)-(Table 3-1).

Table 3-1 Representative major element composition of minerals in the eclogite

2+ 2+ (i) XFe = Fe /(Fe +Mg); Alm: almandine, Grs: grossular, Pyp: pyrope, Sps: spessartine, WEF: wollastonite+enstatite+ferrosilite, JD: jadeite, AE: aegirine, Ps: pistacite. (ii) c: core of mineral, m: mantle of mineral, r: rim of mineral. (iii) Fe3+ was calculated assuming stoichiometric mineral compositions.

3.6.3 Amphibole

Na-rich and Ca-rich amphiboles were found in both eclogite and blueschist (Table 3-1 and 2). The sodic amphibole, which is classified as glaucophane (Fig. 3-5g, Leake et. al., 1997), is occasionally rimmed by sodic-calcic amphiboles, namely barroisite and/or katophorite/magnesiokatophorite (Fig. 3-5d, h and i). In the blueschist, the Na2O content of glaucophane varies from 6.8 wt.% to 7.3 wt.% and that of the barroisite and

3. Coexisting eclogite and blueschist katophorite/magnesiokatophorite from 3.6 wt.% to 4.8 wt.%, respectively. The glaucophane in blueschist displays a complex compositional zoning. The dark mantle contains higher MgO (11.36 wt.%) and lower FeO (8.37 wt.%) than the light core (MgO = 9.75 wt.%; FeO = 11.13 wt.%) and rim (MgO = 10.20 wt.%; FeO = 10.33 wt.%), whereas the CaO, NaA-site increase and NaB-site decreases from core to rim (Fig. 3-4f and g, Table 3-2). The zoning of the glaucophane including the occasional barroisite or katophorite/magnesiokatophorite outer rims is displayed by XFe (XFe = Fe/(Fe+Mg)), which is around 0.37 in core domain, drops abruptly to 0.28-0.32 in the mantle, increases to core-level at the glaucophane rim, and increases sharply to 0.40-0.41 or 0.42-0.45 in the outer rim of barroisite and katophorite, respectively (Fig. 3-5d). This compositional glaucophane zoning behavior suggests prograde growth of the dominant core-mantle- inner rim region and retrograde growth of the thin outer rim.

3.6.4 White mica

Phengite, which is the main type of white mica both in the eclogite and blueschist, occurs in the matrix and as inclusion in garnet and omphacite, whereas paragonite frequently occurs as box-shaped intergrowth with epidote/clinozoisite in garnet forming pseudomorphs possibly after lawsonite (Fig. 3-3 and 4). The Si-content of phengite clusters at 3.4-3.5 pfu both in eclogite and blueschist while the Na-content of paragonite at 0.87-0.88 pfu and the Si-content at 3.0-3.12 pfu (Table 3-1 and 2).

3.6.5 Carbonates

The carbonates in the eclogite and the blueschist are predominantly dolomite, while minor calcite occurs as inclusion in garnet, and magnesite is embraced by dolomite both of which occur as inclusions in garnet and in the matrix. The dolomite contains an ankerite component between 20-30 mol. % while magnesite contains a siderite endmember with XFe of 0.41-0.50 (Table 3-1 and 2). Minor MnO (< 1 wt.%) is measured in the carbonates. In the eclogite, however, matrix dolomite has a rim that

3. Coexisting eclogite and blueschist contains higher Mg and lower Fe than the dominant core domain (Fig. 3-3c, Table 3-1).

Table 3-2 Representative major element composition of minerals in the coexisting blueschist

3.6.6 Epidote/clinozoisite

Matrix epidote group minerals have a high Fe2O3 content (> 7 wt.%) and are usually epidote (optically negative), both in the eclogite and blueschist (Table 3-1 and 2). However, the eclogite garnet mainly contains clinozoisite inclusions (optically positive,

Fe2O3 < 7 wt.%) which are intimately intergrown with paragonite while epidote occasionally occurs as interstitial grains between clinozoisite and paragonite (Fig. 3-3c and d). In the blueschist garnet, both clinozoisite and epidote were found to be

* 3+ intimately intergrown with paragonite (Fig. 3-4e). The Ps ratio [XPs=100 Fe / (Fe3++Al)] of epidote varies from 20 to 24, while clinozoisite has a lower Ps ratio of 11-14 (Table 3-1 and 2).

3.6.7 Accessory minerals

Rutile contains minor FeO and Cr2O3, commonly less than 0.5 wt.%. The TiO2 content of titanite ranges between 37 wt.% and 38 wt.% and the Al2O3 content is ca. 2 wt.% (Table 3-1 and 2). Apatite inclusions in eclogite garnet have a F content of ~3.5 wt.%. Most sulphides are classified as pyrite, with a minor Ni content of < 1 wt.%.

3. Coexisting eclogite and blueschist

3.7 Phase equilibria and P-T conditions

Based on the mineral assemblages and compositions, the model system TiNCaKF-

MASCHO (TiO2-Na2O-CaO-K2O-FeO-MgO-Al2O3-SiO2-CO2-H2O-O (Fe2O3)) was chosen to calculate P-T pseudosections for the eclogite and blueschist. MnO, which predominantly occurs in garnet, was neglected. However, Fe2O3 was included in the system because it plays an important role in enhancing amphibole stabilities, especially at higher metamorphic conditions (e.g., Warren and Waters, 2006; Diener and Powell 2012).

Pseudosection calculations were performed using the Perple_X software package (Connolly, 1990, 2005; version 6.6.7) and the internally consistent thermodynamic database of Holland and Powell (1998 and update). The following solid-solution models were used: Gt(HP) for garnet (Holland and Powell, 1998), Omph(GHP) for omphacite (Green et al., 2007), cAmph(DP) for amphibole (Diener et al., 2007), Mica(CHA) for phengite (Coggon and Holland, 2002), Chl(HP) for chlorite (Holland and Powell, 1998), Ep(HP) for epidote (Holland and Powell, 1998), Do(HP) for dolomite-ankerite (Holland and Powell, 1998), and F for H2O-CO2 fluid solution (Connolly and Trommsdorff, 1991). The magnesite-siderite solution model M(HP) (Holland and Powell, 1998) was not taken, since it significantly changes the mineral modal compositions when compared with the microscopically derived modal compositions. Thermodynamic data for H2O, CO2, and their mixtures were calculated using the CORK equation of state (Holland and Powell, 1991, 1998).

3.7.1 Effective bulk-rock composition

The bulk-rock composition of metamorphic rocks may change during its P-T evolution for instance due to element fractionation during prograde garnet growth (e.g., Evans, 2004). Therefore it is critical to generate an effective bulk composition for phase equilibrium calculations in metamorphic rocks (e.g., Stüwe, 1997). The detailed effective bulk composition approach used in this study is based on the method

3. Coexisting eclogite and blueschist described by Marmo et al. (2002) and Evans (2004). The effective bulk compositions, which approximate those of the XRF analyses, used for the eclogite and blueschist modelling in the TiNCaKFMASCHO system are given in Table 3-3. The compositional differences between the effective blueschist composition and the whole rock XRF analysis are thought to be due to the occasional presence of mineral enriched and mineral depleted textural domains which occurred in the thin sections from which the effective bulk-rock composition was generated (cf., Wei et al., 2009; Groppo and Castelli, 2010). For example, the lower content of CaO in the effective blueschist composition compared to that of the whole rock XRF analysis is thought to be due to textural domains which are relatively depleted in dolomite and epidote. H2O and CO2 are commonly considered as saturated phase components for saturated fluid phases when undertaking thermodynamic modelling with carbonate-bearing metamorphic rocks. However, the specification of H2O and CO2 as thermodynamic components gives more realistic results for phase diagram modelling (e.g., Kerrick and Connolly,

2001). The CO2 and H2O contents can also be calculated using the modal abundances of carbonaceous and hydrous minerals in the samples to be investigated. It is generally thought that subducted oceanic metabasalts undergo very limited decarbonation during subduction along a low temperature geotherm (Kerrick and Connolly, 2001). Thus the calculated CO2 values were taken as bulk CO2 contents of the respective model system and assumed to remain constant during the whole metamorphic evolution. However, the H2O content of the rock decreases along the prograde P-T path and a pseudosection calculated with the water content of the present-day bulk rock must erroneously

Table 3-3 Whole-rock compositions of the carbonate-bearing eclogite and omphacite-carbonate- bearing blueschist from the South Tianshan (U)HP/LT belt

FeO* = total iron

3. Coexisting eclogite and blueschist influence the stabilities of hydrous minerals during the prograde P-T trajectory. Therefore in the present study we used the water content from average carbonate-bearing MOR metabasalt (Super Composite, Staudigel et al., 1989) as the effective bulk H2O values thereby reflecting the bulk water content prior to significant dehydration during subduction.

3.7.2 Pseudosection calculations

The TiNCaKFMASCHO P-T pseudosection calculated for the eclogite is presented in Fig. 3-6a. Omphacite, dolomite, and phengite are stable phases over the whole shown P-T range, while siderite occurs only in the high pressure-low temperature area, garnet is absent in the low temperature area and lawsonite is absent in the low pressure-high temperature area. The fluid phase-in isograd is parallel to the P-axis above 17 kbar at 520 oC. The chlorite- and glaucophane-out isograds share this trend although at somewhat higher temperatures of 545 and 565 oC, respectively. The mineral volume and compositional isopleths of solid solution phases vary dramatically between the fluid-absent and fluid-present fields (Fig. 3-6b). The estimated mineral modal amounts

(garnet ~17 vol.%, glaucophane 1-2 vol.%) and measured mineral compositions (Xprp 0.11-0.13 in the garnet rim, Si content 3.4-3.5 in phengite) are consistent with a P-T range of 545 - 565 oC and 21.8 - 23.1 kbar in the stability field of Grt + Omp + Dol + Phn + Gln + Law ± Qtz + F. This narrow P-T range (see Fig. 3-6a) is considered to represent peak metamorphic conditions for the eclogite. The calculated modal proportions of garnet, omphacite, glaucophane, dolomite and phengite are in accordance with those estimated from thin sections at peak P-T condition of 555 oC and 22.5 kbar. In addition, the XPrp, the XGrs and the Si content in phengite are in accordance with or slightly lower than those values calculated by the thermodynamic modelling (Table 3-4). As suggested by the thermodynamic modelling, lawsonite should have been stable at the peak P-T conditions, which is supported by the lawsonite pseudomorphs consisting of box-shaped clinozoisite/epidote + paragonite assemblages in garnet (see section 3.4.1, Fig. 3-3c and d). However, lawsonite has

3. Coexisting eclogite and blueschist

Figure 3-6 P-T pseudosections for carbonate-bearing eclogite (a) and blueschist (d) in the system TiNCaKFMASCHO calculated with Perple_X. The bulk rock compositions are given in Table 3-3. ab = albite; amp = amphibole; arag = aragonite; cal = calcite; chl = chlorite; dol = dolomite; ep = epidote; f = fluid; grt = garnet; gln = glaucophane; law = lawsonite; mag = magnesite; omp = omphacite; qtz = quartz; rt = rutile; sid = siderite; ta = talc; ttn = titanite. The pseudosections are contoured to constrain the peak metamorphic P-T ranges (dark circles) with isopleths of garnet (XPrp and volume percent), glaucophane/omphacite (vol.%) and Si in phengite (pfu) of the carbonate-bearing eclogite (b) and blueschist (e), respectively. The isopleths of XCO2 for the fluid phase show that the CO2 activities are very low for Tianshan carbonate-bearing eclogite and blueschist. Due to the almost complete absence of retrograde mineral phases, only an approximate post-peak path showing isothermal decompression was calculated. The P-T range (dark circles in c, f) coincides with the titanite-bearing assemblage and is in accordance with the post-peak decompressional path of (U)HP pelitic-felsic schists (Wei et al., 2009) in the Tianshan metamorphic belt (white stars). The results suggest that the Tianshan (U)HP/LT eclogites and coexisting blueschist have experienced an isothermal decompression (d) during rapid exhumation.

3. Coexisting eclogite and blueschist experienced significant consumption during the prograde P-T path (e.g., lawsonite occupies ca. 12 vol.% at 520 oC and 19 kbar, and ca. 5 vol.% at the peak conditions of ca. 555 oC and 22.5 kbar; Fig. 3-6). The breakdown of lawsonite as well as that of chlorite and glaucophane produced garnet and omphacite during prograde eclogitization of the blueschist. The XCO2 isopleths indicate a very low CO2 activity (0.014) for the equilibrium fluid phase at peak conditions (Fig. 3-6b).

The calculated TiNCaKFMASCHO P-T pseudosection for the blueschist is presented in Fig. 3-6d. In comparison with the pseudosection calculated for the eclogite, the stability field of glaucophane is enlarged while that of lawsonite is reduced. This is thought to be due to the different bulk-rock compositions, especially the lower CaO content in the blueschist (Table 3-3). Lawsonite is only stable under T < 560 oC and P > 16 kbar in the calculated P-T range. The fluid phase-in isograd is similar to that in the eclogite pseudosection (i.e., parallel to the P-axis). The estimated volume percent of garnet (~23%), omphacite (~8%) and glaucophane (~43%), as well as measured XPrp in the garnet rim (0.11-0.13) and the Si content in phengite (3.4-3.5) constrains the peak metamorphic conditions for the blueschist between 540 and 555 oC at 21.8 - 22.8 kbar (Fig. 3-6e) within the stability field of Gln + Grt + Omp + Dol + Phn ± Law ± Qtz + F (Fig. 3-6d). Therefore, the peak P-T range of the blueschist overlaps that of the host eclogite. The calculated mineral modal amounts at the peak P-T condition of 548 oC and 22.3 kbar perfectly agree with those estimated from the thin sections. In addition, the calculated compositions (XPrp and Si in phengite) are close to the measured EMPA values, while the measured XGrs is somewhat lower than the calculated value (Table 3-4). Lawsonite, which was found only as box-shaped pseudomorphs (see section 3.4.2, Fig. 3-4e) in the studied thin section, is thought to have been (almost) completely consumed during the prograde P-T path seeing that the peak metamorphic mineral stability field is close to the lawsonite-out reaction (Fig. 3-6d). The glaucophane zoning is not reflected in the pseudosection; however, the increasing Ca content from core to rim suggests a prograde growth due to the consumption of lawsonite. This prograde glaucophane zoning has distinctly different chemical characteristics when

3. Coexisting eclogite and blueschist compared with retrograde glaucophane zoning (cf., Du et al., 2011). As is the case for the eclogite, the XCO2 isopleths of the blueschist also reveal that the fluid phase has a very low CO2 activity (0.012) at peak metamorphic conditions (Fig. 3-6e).

Table 3-4 Comparisons of calculated peak mineral modal proportions and compositions with those observed and measured for the carbonate-bearing eclogite and omphacite-carbonate- bearing blueschist

Due to the almost complete absence of retrograde mineral phases, only an approximate post-peak path showing isothermal decompression was calculated. The Ca content in the barroisite and katophorite rimming glaucophane (1.1-1.3) and Si content in the paragonitic mica (3.0-3.12) match P-T conditions at about P=11-13 kbar and T=530-560 oC (Fig. 3-6c and f). This P-T range coincides with the titanite(after rutile)-bearing assemblage and is in accordance with the post-peak decompressional path of (U)HP pelitic-felsic schists (Wei et al., 2009) in the Tianshan metamorphic belt (Fig. 3-6d).

3.8 Discussion

3.8.1 Coexistence of eclogite and blueschist

Carson et al. (2000) summarized the relationship of associated high-pressure blueschists and eclogites. The formation of high-pressure blueschists may (1) precede peak metamorphism and thus constrain the prograde path of eclogites, (2) have occurred at the same P-T conditions than that of eclogites or (3) post-date peak metamorphism and thus help to define the retrograde path of the associated eclogites. The here investigated eclogite and blueschist from the Tianshan (U)HP/LT metamorphic belt are thought to have coexisted under the same peak metamorphic conditions of 540 - 565 oC and 21.8 - 23.1 kbar. This coexistence is confirmed by the

3. Coexisting eclogite and blueschist following petrographic observations indicating that glaucophane was a stable peak metamorphic phase in the blueschist: (1) texturally equilibrated glaucophane and omphacite inclusions preserved in garnets (Fig. 3-4b); (2) omphacite inclusions in matrix glaucophane porphyroblasts (Fig. 3-4c and g); (3) glaucophane inclusions in omphacites (Fig. 3-4d and g); and (4) texturally equilibrated matrix glaucophane and omphacite (Fig. 3-4c). Moreover, identical glaucophane compositional zoning in matrix glaucophane and glaucophane inclusions in both omphacite and garnet (Fig. 3-4f, g and h) supports the contemporaneous coexistence of glaucophane and omphacite. The presence of coexisting pyroxenes and amphiboles in eclogites and associated coarse-grained blueschists were also reported from other high-pressure complexes (e.g., Brown and Bradshaw, 1979). Although occasionally barroisite is thought to occur in equilibrium with garnet and omphacite at eclogite-facies conditions (e.g., Gomez Pugnaire et al., 1997), in the here studied blueschist the above presented evidence suggests that the barroisite and the katophorite suite along the rim of the glaucophane grew during decompression (cf., Gao et al., 1999; Wei et al., 2009).

Several phase equilibrium modelling studies have revealed that eclogites and blueschists may coexist under the same high-pressure metamorphic conditions in various high-pressure terranes (e.g., El-Shazly et al., 1997; Gomez Pugnaire et al., 1997; Oh et al., 1991; Vitale Brovarone et al., 2011). Calculated phase equilibria of lawsonite eclogites using a MORB composition support the possible coexistence of low-T eclogites and blueschists in high-pressure terranes (Wei and Clarke, 2011). Phase equilibria in this study indicate that the eclogite and the blueschist reached the same peak conditions at 540 - 565 oC and 21.8 - 23.1 kbar, supporting the possible coexistence of eclogites and blueschists under high-pressure conditions. However, the peak P-T stability field in the pseudosections of both rock types suggests the stability of small amounts (< 5 vol.%) of lawsonite, which are present only as pseudomorphs in the investigated samples. Lawsonite relics were also identified in glaucophane in chloritoid-glaucophane schist from the same area (Du et al., 2011). Furthermore, box-shaped inclusions of epidote/clinozoisite + paragonite assemblages occur in garnet

3. Coexisting eclogite and blueschist and are interpreted to present pseudomorphs after lawsonite in eclogite and blueschist from the western Tianshan (e.g., Beinlich et al., 2010; Lü et al., 2009; van der Straaten et. al., 2008). Consequently, the modelling results and the textural observations suggest that lawsonite was a stable phase (although with a low modal content) at peak eclogite-facies metamorphism, but underwent significant consumption during prograde reactions and was ultimately replaced during exhumation.

Previous studies from other HP-UHP metamorphic belts considered the stable coexistence of eclogite and omphacite-bearing blueschist to be the result of different bulk-rock chemical compositions (e.g., El-Shazly et al., 1997; Gomez Pugnaire et al., 1997; Maruyama et al., 1996; Oh et al., 1991; Reinsch, 1979; Schliestedt, 1986; Vitale Brovarone et al., 2011). For example, Gomez Pugnaire et al. (1997) proposed that the

Na2O/(Na2O + CaO) ratio significantly influences the stability of the respective mineral assemblages under high-pressure conditions. These authors suggested that a lower Na2O/(Na2O + CaO) ratio favors an eclogite mineral assemblages while a higher ratio favors a blueschist mineral assemblages. However, Vitale Brovarone et al. (2011) argued that the CaO/(CaO + FeO) ratio plays an even more important role than the

Na2O/(Na2O + CaO) ratio and that a higher CaO/(CaO + FeO) ratio favors eclogite mineral assemblages while a lower ratio favors blueschist mineral assemblages. Wei and Clarke (2011) also suggested that lawsonite-eclogite assemblages are a result of a characteristic bulk compositions involving a high X(CaO) (=CaO⁄(CaO + MgO + FeO*

+ MnO + Na2O)) while lawsonite blueschist assemblages are a result of bulk compositions involving a low X(CaO). In the present study, the blueschist contains higher FeO and lower CaO compared to the eclogite (Table 3-3). In addition, the CaO/(CaO + FeO) ratio (0.38) and the X(CaO) = 0.22 of the blueschist are lower than those of the eclogite with 0.64 and 0.39, respectively. Thus the data presented here are in accordance with the assumption that Ca-rich bulk-rock compositions favor eclogitic mineral assemblages while Ca-poor bulk-rock compositions favor blueschist mineral assemblages at same metamorphic conditions (cf., Vitale Brovarone et al., 2011; Wei and Clarke, 2011). In addition, the CaO-poor characteristic of blueschist in contrast to

3. Coexisting eclogite and blueschist eclogite cannot be the result of a fluid influx during retrograde metamorphic conditions since several studies have shown that Ca-rich fluids dominate the subduction fluids in the Tianshan (U)HP/LT belt (e.g., Gao and Klemd, 2001; Beinlich et al., 2010; Lü et al., 2012a).

3.8.2 Carbonates in the Tianshan HP rocks

Carbonate-bearing eclogites and blueschists are commonly observed in the Tianshan (U)HP/LT belt. High carbonate content in metabasalts suggests that the precursor oceanic crust has undergone significant hydrothermal alteration prior to subduction. Magnesite and calcite (suggested to have formerly been aragonite) inclusions in dolomite were used to calculate peak metamorphic conditions for some carbonate-bearing eclogites (Zhang et al., 2002). The peak metamorphic temperatures and pressures were calculated to be 560-600 C, 4.95-5.07 GPa and are based on the calcite-dolomite geothermometer and the equilibrium calculation of the reaction dolomite = magnesite + aragonite, respectively (Zhang et al., 2003b). However, equilibrium conditions between the carbonates are doubtful and in addition dolomite was interpreted as a secondary mineral which not only replaced calcite and magnesite, but also matrix glaucophane, clinozoisite and paragonite (Klemd, 2003; see also Smit et al., 2008). Experimental studies of tholeiitic basalts involving H2O-CO2 mixed fluids suggest that hydrates and carbonates coexist over a large P-T range buffering the fluid composition and that amphibole coexists with dolomite and magnesite just at P ≥ 18 kbar (Molina and Poli, 2000). Furthermore, in multi-component Fe-bearing systems, the presence of magnesite, dolomite, or both (at pressures within the dolomite + magnesite stability field) can be strictly related to the amount of volatiles fixed to the rock after ocean floor metamorphism; thus magnesite/dolomite cannot be used as a pressure marker, without a strict control of the relative amounts of volatile and non-volatile components (Poli et al., 2009). In the present study, magnesite was identified as inclusions in dolomite and pseudosection modellings suggest both carbonate-bearing eclogites and blueschists formed under high-pressure conditions.

3. Coexisting eclogite and blueschist

The contoured XCO2 isopleths show that the CO2 activity in the fluid phase for Tianshan carbonate-bearing eclogite and blueschist was very low (XCO2 ≤ 0.02) during metamorphism.

3.8.3 Implications

Blueschist and eclogite belts occur worldwide and mark the locations of former subduction zones (Maruyama et al., 1996). Eclogites occurring as boudins, blocks and interlayers in the blueschists were widely found in various HP-UHP metamorphic belts. Pressure, temperature, bulk rock and fluid compositions are the principal factors controlling the type of the metamorphic mineral assemblages produced. The juxtapositions of blueschists and eclogites are usually regarded as being responsible for different equilibrium P-T conditions that are derived for those rocks (e.g., Brown and Bradshaw, 1979; Davis and Whitney, 2006; Ridley, 1984). However, coexisting blueschists and eclogites, which have formed under the same P-T conditions, are thought to have been stabilized by differences in the bulk rock composition (e.g., Gomez Pugnaire et al., 1997; Oh et al., 1991; Vitale Brovarone et al., 2011). Such compositional variations in the eclogites and blueschists are either ascribed to different protolith chemistries before metamorphism (El-Shazly et al., 1997) or are interpreted to be the result of variable fluid infiltrations during prograde metasomatism or retrograde overprinting during subduction/exhumation processes (e.g., Beinlich et al., 2010; El-Shazly et al., 1997; van der Straaten et al., 2008).

In the Tianshan (U)HP/LT belt, two distinct types of blueschist were already described by Gao and Klemd (2003): An omphacite-garnet bearing blueschist, which is characterized by the absence of a foliation and albite, was interpreted to have formed during prograde metamorphism. The second type is an albite-bearing foliated blueschist, which was interpreted to have formed during retrograde metamorphic conditions. A typical prograde blueschist metabasalt in the Tianshan consists of glaucophane + garnet ± omphacite + epidote/clinozoisite + phengite/paragonite + dolomite, as well as accessory minerals (cf. Fig. 3b in John et al., 2008). The peak

3. Coexisting eclogite and blueschist metamorphic conditions of the prograde blueschist were estimated to range from 530-550 °C at 19 kbar (John et al., 2008) to 510 ± 30 °C at 21 ± 1.5 kbar (Beinlich et al., 2010). Retrograde blueschists were formed by fluid-rock interaction during uplift within the subduction channel (van der Straaten et al., 2008), and the P-T conditions were estimated at 580 ± 10 °C at 15 ± 1 kbar (van der Straaten et. al., 2008) or 430-530 °C at 15-20 kbar (Du et al., 2011). The present paper presents a third type of blueschist, which coexists with eclogite under high-pressure metamorphism. The peak metamorphic conditions of coexisting blueschist with eclogite in the presence of an

o aqueous fluid phase (XCO2 ≤ 0.02) were estimated at 540 - 565 C and 21.8 - 23.1 kbar using the pseudosection modelling approach. Thus three different types of blueschists occur in the Tianshan subduction complex: a prograde blueschist type formed prior to eclogitization, a retrograde blueschist type formed by rehydration of eclogite during exhumation, and a third blueschist type coexisting with eclogite. Moreover, the coexisting eclogite and blueschist have experienced post-peak rapid isothermal decompression, as also observed for other South Tianshan eclogite-facies rocks (Klemd et al., 2002; Zhang et al., 2003b).

3.9 Conclusions

1. Textural, thermodynamic and compositional evidence shows that the carbonate-bearing eclogite and blueschist underwent the same metamorphic evolution, and the stability of the different mineral assemblages at peak metamorphic conditions is due to the differences in the respective bulk-rock composition. Higher CaO contents of the protolith rocks favor eclogitic mineral assemblages while lower CaO contents favor blueschist mineral assemblages.

2. Pseudosection modelling suggests that both the eclogite and the blueschist are stable at peak metamorphic conditions of 540 - 565 oC and 21.8 - 23.1 kbar during which the eclogite and blueschist were in equilibrium with an aqueous fluid phase

with a very low CO2 activity (XCO2 ≤ 0.02).

3. Coexisting eclogite and blueschist

3. Three types of blueschists were recognized in the Tianshan subduction complex, including prograde, retrograde, and eclogite-facies blueschist.

Acknowledgements

This research was supported by the Deutsche Forschungsgemeinschaft (KL692/17-3), the ‗National Natural Science Foundation of China‘ (41025008), and the State Key Project for Basic Research of China (2007CB411302). The authors appreciate the constructive comments and suggestions of H.-P. Schertl and an anonymous reviewer which clearly helped to improve the manuscript. T. John is thanked for valuable discussions on the pseudosection modeling approach. Furthermore, we thank J.G. Liou for his editorial handling and comments.

References

Allen M.B., Windley B.F. and Zhang C. (1992) Palaeozoic collisional tectonics and magmatism of the Chinese Tien Shan, central Asia. Tectonophysics 220, 89–115. Beinlich A., Klemd R., John T. and Gao J. (2010) Trace-element mobilization during Ca-metasomatism along a major fluid conduit: eclogitization of blueschist as a consequence of fluid-rock interaction. Geochimica et Cosmochimica Acta 74, 1892– 1922. Brown E.H. and Bradshaw J.Y. (1979) Phase-relations of pyroxene and amphibole in greenstone, blueschist and eclogite of the Franciscan Complex, California. Contributions to Mineralogy and Petrology 71, 67–83. Caron J.M. and Pequignot G. (1986) The transition between blueschists and lawsonite-bearing eclogites based on observations from Corsican metabasalts. Lithos 19, 205–218. Carson C.J., Clarke G.L. and Powell R. (2000) Hydration of eclogite, Pam Peninsula, New Caledonia. Journal of Metamorphic Geology 18, 79–90. Carswell D.A. (1990) Eclogites and eclogite facies: definitions and classifications, In: Carswell, D.A. (Eds.), Eclogite Facies Rocks. Blackie and Son Ltd, New York, pp. 1–13. Coggon R. and Holland T.J.B. (2002) Mixing properties of phengitic micas and revised garnet-phengite thermobarometers. Journal of Metamorphic Geology 20, 683–696. Connolly J.A.D. (1990) Multivariable phase-diagrams - an algorithm based on generalized thermodynamics. American Journal of Science 290, 666–718. Connolly J.A.D. and Trommsdorff V. (1991) Petrogenetic grids for metacarbonate rocks: pressure-temperature phase-diagram projection for mixed-volatile systems. Contributions to Mineralogy and Petrology 108, 93–105.

3. Coexisting eclogite and blueschist

Connolly J.A.D. (2005) Computation of phase equilibria by linear programming: A tool for geodynamic modelling and its application to subduction zone decarbonation. Earth and Planetary Science Letters 236, 524–541. Davis P.B. and Whitney D.L. (2006) Petrogenesis of lawsonite and epidote eclogite and blueschist, Sivrihisar Massif, Turkey. Journal of Metamorphic Geology 24, 823–849. Davis P.B. and Whitney D.L. (2008) Petrogenesis and structural petrology of high-pressure metabasalt pods, Sivrihisar, Turkey. Contributions to Mineralogy and Petrology 156, 217–241. Day H.W. and Mulcahy S.R. (2007) Excess silica in omphacite and the formation of free silica in eclogite. Journal of Metamorphic Geology 25, 37–50. Diener J.F.A., Powell R., White R.W. and Holland T.J.B. (2007) A new thermodynamic

model for clino- and orthoamphiboles in the system Na2O-CaO-FeO-MgO-Al2O3- SiO2-H2O-O. Journal of Metamorphic Geology 25, 631–656. Diener J.F.A. and Powell R. (2012) Revised activity–composition models for clinopyroxene and amphibole. Journal of Metamorphic Geology 30, 131–142. Du J., Zhang L., Lü Z. and Chu X. (2011) Lawsonite-bearing chloritoid-glaucophane schist from SW Tianshan, China: phase equilibia and P-T path. Journal of Asian Earth Sciences 42, 684–693. El-Shazly A.K., Worthing M.A. and Liou J.G. (1997) Interlayered eclogites, blueschists and epidote amphibolites from NE Oman: a record of protolith compositional control and limited fluid infiltration. Journal of Petrology 38, 1461– 1487. Ernst W.G. (1970) Tectonic Contact between the Franciscan Mélange and the Great Valley Sequence—Crustal Expression of a Late Mesozoic Benioff Zone. Journal of Geophysical Research 75, 886–901. Ernst W.G. (1988) Tectonic history of subduction zones inferred from retrograde blueschist P-T paths. Geology 16, 1081–1084. Evans T.P. (2004) A method for calculating effective bulk composition modification due to crystal fractionation in garnet-bearing schist: implications for isopleth thermobarometry. Journal of Metamorphic Geology 22, 547–557. Gao J., He G., Li M., Xiao X., Tang Y., Wang J. and Zhao M. (1995) The mineralogy, petrology, metamorphic PTDt trajectory and exhumation mechanism of blueschists, south Tianshan, northwestern China. Tectonophysics 250, 151–168. Gao J., Li M.S., Xiao X.C., Tang Y.Q. and He G.Q. (1998) Paleozoic tectonic evolution of the Tianshan Orogen, northwestern China. Tectonophysics 287, 213– 231. Gao J., Klemd R., Zhang L., Wang Z. and Xiao X. (1999) P-T path of high-pressure/low-temperature rocks and tectonic implications in the western Tianshan Mountains, NW China. Journal of Metamorphic Geology 17, 621–636. Gao J. and Klemd R. (2001) Primary fluids entrapped at blueschist to eclogite transition: evidence from the Tianshan meta-subduction complex in northwestern China. Contributions to Mineralogy and Petrology 142, 1–14. Gao J. and Klemd R. (2003) Formation of HP-LT rocks and their tectonic implications

3. Coexisting eclogite and blueschist

in the western Tianshan Orogen, NW China: geochemical and age constraints. Lithos 66, 1–22. Gao J., Long L.L., Klemd R., Qian Q., Liu D.Y., Xiong X.M., Su W., Liu W., Wang Y.T. and Yang F.Q. (2009) Tectonic evolution of the South Tianshan orogen and adjacent regions, NW China: geochemical and age constraints of granitoid rocks. International Journal of Earth Sciences 98, 1221–1238. Gomez Pugnaire M.T., Karsten L. and SanchezVizcaino V.L. (1997) Phase relationships and P-T conditions of coexisting eclogite-blueschists and their transformation to greenschist-facies rocks in the Nerkau Complex (Northern Urals). Tectonophysics 276, 195–216. Green E., Holland T. and Powell R. (2007) An order-disorder model for omphacitic pyroxenes in the system jadeite-diopside-hedenbergite-acmite, with applications to eclogitic rocks. American Mineralogist 92, 1181–1189. Groppo C. and Castelli D. (2010) Prograde P-T Evolution of a Lawsonite Eclogite from the Monviso Meta-ophiolite (Western Alps): Dehydration and Redox Reactions during Subduction of Oceanic FeTi-oxide Gabbro. Journal of Petrology 51, 2489–2514. Han B.F., He G.Q., Wang X.C. and Guo Z.J. (2011) Late Carboniferous collision between the Tarim and Kazakhstan-Yili terranes in the western segment of the South Tian Shan Orogen, Central Asia, and implications for the Northern Xinjiang, western China. Earth-Science Reviews 109, 74–93. Holland T. and Powell R. (1991) A Compensated-Redlich-Kwong (CORK) equation

for volumes and fugacities of CO2 and H2O in the range 1 bar to 50 kbar and 100-1600 °C. Contributions to Mineralogy and Petrology 109, 265–273. Holland T.J.B. and Powell R. (1998) An internally consistent thermodynamic data set for phases of petrological interest. Journal of Metamorphic Geology 16, 309-343. John T., Klemd R., Gao J. and Garbe-Schonberg C.D. (2008) Trace-element mobilization in slabs due to non steady-state fluid-rock interaction: Constraints from an eclogite-facies transport vein in blueschist (Tianshan, China). Lithos 103, 1–24. Kerrick D.M. and Connolly J.A.D. (2001) Metamorphic devolatilization of subducted oceanic metabasalts: implications for seismicity, arc magmatism and volatile recycling. Earth and Planetary Science Letters 189, 19–29. Klemd R., Schröter F.C., Will T.M. and Gao J. (2002) P-T evolution of glaucophane- omphacite bearing HP-LT rocks in the western Tianshan Orogen, NW China: new evidence for 'Alpine-type' tectonics. Journal of Metamorphic Geology 20, 239–254. Klemd R. (2003) Ultrahigh-pressure metamorphism in eclogites from the western Tianshan high-pressure belt (Xinjiang, western China) - Comment. American Mineralogist 88, 1153–1156. Klemd R., Bröcker M., Hacker B.R., Gao J., Gans P. and Wemmer K. (2005) New age constraints on the metamorphic evolution of the high-pressure/low-temperature belt in the western Tianshan mountains, NW China. Journal of Geology 113, 157–168. Klemd R., John T., Scherer E.E., Rondenay S. And Gao J. (2011) Changes in dip of subducted slabs at depth: Petrological and geochronological evidence from HP-UHP

3. Coexisting eclogite and blueschist

rocks (Tianshan, NW-China). Earth and Planetary Science Letters 310, 9–20. Leake B.E., Woolley A.R., Arps C.E.S., Birch W.D., Gilbert M.C., Grice J.D., Hawthorne F.C., Kato A., Kisch H.J., Krivovichev V.G., Linthout K., Laird J., Mandarino J.A., Maresch W.V., Nickel E.H., Rock N.M.S., Schumacher J.C., Smith D.C., Stephenson N.C.N., Ungaretti L., Whittaker E.J.W. and Guo Y.Z. (1997) Nomenclature of amphiboles: Report of the subcommittee on amphiboles of the International Mineralogical Association, commission on new minerals and mineral names. American Mineralogist 82, 1019–1037. Liou J.G., Ernst W.G., Zhang R.Y., Tsujimori T. and Jahn B.M. (2009) Ultrahigh-pressure minerals and metamorphic terranes - The view from China. Journal of Asian Earth Sciences 35, 199–231. Long L., Gao J., Klemd R., Beier C., Qian Q., Zhang X., Wang J. and Jiang T. (2011) Geochemical and geochronological studies of granitoid rocks from the Western Tianshan Orogen: Implications for continental growth in the southwestern Central Asian Orogenic Belt. Lithos 126, 321–340. Lü Z., Zhang L.F., Du J.X. and Bucher K. (2008) Coesite inclusions in garnet from eclogitic rocks in western Tianshan, northwest China: Convincing proof of UHP metamorphism. American Mineralogist 93, 1845–1850. Lü Z., Zhang L., Du J. and Bucher K. (2009) Petrology of coesite-bearing eclogite from Habutengsu Valley, western Tianshan, NW China and its tectonometamorphic implication. Journal of Metamorphic Geology 27, 773–787. Lü Z., Zhang L., Du J., Yang X., Tian Z. and Xia B. (2012) Petrology of HP metamorphic veins in coesite-bearing eclogite from western Tianshan, China: Fluid processes and elemental mobility during exhumation in a cold subduction zone. Lithos 136, 168–186. Lü Z. and Zhang L.F. (2012) Coesite in the eclogite and schist of the Atantayi Valley, southwestern Tianshan, China. Chinese Science Bulletin 57, 1467–1472. Marmo B.A., Clarke G.L. and Powell R. (2002) Fractionation of bulk rock composition due to porphyroblast growth: effects on eclogite facies mineral equilibria, Pam Peninsula, New Caledonia. Journal of Metamorphic Geology 20, 151–165. Maruyama S., Liou J.G. and Terabayashi M. (1996) Blueschists and Eclogites of the World and Their Exhumation. International Geology Review 38, 485–594. Molina J.F. and Poli S. (2000) Carbonate stability and fluid composition in subducted

oceanic crust: an experimental study on H2O-CO2-bearing basalts. Earth and Planetary Science Letters 176, 295–310. Morimoto N., Fabries J., Ferguson A.K., Ginzburg I.V., Ross M., Seifert F.A., Zussman J., Aoki K. and Gottardi G. (1988) Nomenclature of pyroxenes. American Mineralogist 73, 1123–1133. Oh C.W., Liou J.G. and Maruyama S. (1991) Low-Temperature Eclogites and Eclogitic Schists in Mn-rich Metabasites in Ward Creek, California; Mn and Fe Effects on the Transition between Blueschist and Eclogite. Journal of Petrology 32, 275–302. Page F.Z., Essene E.J. and Mukasa S.B. (2005) Quartz exsolution in clinopyroxene is not proof of ultrahigh pressures: Evidence from eclogites from the Eastern Blue

3. Coexisting eclogite and blueschist

Ridge, Southern Appalachians, USA. American Mineralogist 90, 1092–1099. Poli S., Franzolin E., Fumagalli P. and Crottini A. (2009) The transport of carbon and hydrogen in subducted oceanic crust: An experimental study to 5 GPa. Earth and Planetary Science Letters 278, 350–360. Reinsch D. (1979) Glaucophanites and eclogites from Val-Chiusella, Sesia-Lanzo Zone, Italian Alps. Contributions to Mineralogy and Petrology 70, 257–266. Ridley J. (1984) Evidence of a Temperature-dependent ‗Blueschist‘ to ‗Eclogite‘ Transformation in High-pressure Metamorphism of Metabasic Rocks. Journal of Petrology 25, 852–870. Schliestedt M. (1986) Eclogite blueschist relationships as evidenced by mineral equilibria in the high-pressure metabasic rocks of Sifnos, Cycladic islands., Greece. Journal of Petrology 27, 1437–1459. Smit M.A., Bröcker M. and Scherer E.E. (2008) Aragonite and magnesite in eclogites

from the Jaeren nappe, SW Norway: disequilibrium in the system CaCO3-MgCO3 and petrological implications. Journal of Metamorphic Geology 26, 959–979. Staudigel H., Hart S.R., Schmincke H.U. and Smith B.M. (1989) Cretaceous ocean crust at DSDP Sites 417 and 418: Carbon uptake from weathering versus loss by magmatic outgassing. Geochimica et Cosmochimica Acta 53, 3091–3094. Stüwe K. (1997) Effective bulk composition changes due to cooling: a model predicting complexities in retrograde reaction textures. Contributions to Mineralogy and Petrology 129, 43–52. Su W., Gao J., Klemd R., Li J.L., Zhang X., Li X.H., Chen N.S. and Zhang L. (2010) U-Pb zircon geochronology of Tianshan eclogites in NW China: implication for the collision between the Yili and Tarim blocks of the southwestern Altaids. European Journal of Mineralogy 22, 473–478. van der Straaten F., Schenk V., John T. and Gao J. (2008) Blueschist-facies rehydration of eclogites, Tian Shan, NW-China.: Implications for fluid-rock interaction in the subduction channel. Chemical Geology 255, 195–219. Vitale Brovarone A., Groppo C., Hetényi G., Compagnoni R. and Malavieille J. (2011) Coexistence of lawsonite-bearing eclogite and blueschist: phase equilibria modelling of Alpine Corsica metabasalts and petrological evolution of subducting slabs. Journal of Metamorphic Geology 29, 583–600. Warren C.J. and Waters D.J. (2006) Oxidized eclogites and garnet-blueschists from Oman: P-T path modelling in the NCFMASHO system. Journal of Metamorphic Geology 24, 783–802. Wei C.J., Powell R. and Zhang L.F. (2003) Eclogites from the south Tianshan, NW China: petrological characteristic and calculated mineral equilibria in the

Na2O-CaO-FeO-MgO-Al2O3-SiO2-H2O system. Journal of Metamorphic Geology 21, 163–179. Wei C.J., Wang W., Clarke G.L., Zhang L.F. and Song S.G. (2009) Metamorphism of High/ultrahigh-pressure Pelitic-Felsic Schist in the South Tianshan Orogen, NW China: Phase Equilibria and P-T Path. Journal of Petrology 50, 1973–1991. Wei C.J. and Clarke G.L. (2011) Calculated phase equilibria for MORB compositions: a reappraisal of the metamorphic evolution of lawsonite eclogite. Journal of

3. Coexisting eclogite and blueschist

Metamorphic Geology 29, 939–952. Xiao W.J., Windley B.F., Huang B.C., Han C.M., Yuan C., Chen H.L., Sun M., Sun S. and Li J.L. (2009) End-Permian to mid-Triassic termination of the accretionary processes of the southern Altaids: implications for the geodynamic evolution, Phanerozoic continental growth, and metallogeny of Central Asia. International Journal of Earth Sciences 98, 1189–1217. Zhang L.F., Ellis D.J., Williams S. and Jiang W.B. (2002) Ultra-high pressure metamorphism in western Tianshan, China: Part II. Evidence from magnesite in eclogite. American Mineralogist 87, 861–866. Zhang L.F., Ellis D.J., Williams S. and Jiang W.B. (2003a) Ultrahigh-pressure metamorphism in eclogites from the western Tianshan, China - Reply. American Mineralogist 88, 1157–1160. Zhang L.F., Ellis D.J., Arculus R.J., Jiang W.B. and Wei C.J. (2003b) 'Forbidden zone' subduction of sediments to 150 km depth - the reaction of dolomite to magnesite plus aragonite in the UHPM metapelites from western Tianshan, China. Journal of Metamorphic Geology 21, 523–529. Zhang L.F., Song S.G., Liou J.G., Ai Y.L. and Li X.P. (2005) Relict coesite exsolution in omphacite from Western Tianshan eclogites, China. American Mineralogist 90, 181–186.

3. Coexisting eclogite and blueschist

4. Compositional zoning in dolomite

4 Compositional zoning in dolomite from lawsonite-bearing eclogite (SW Tianshan, China): Evidence for prograde metamorphism during subduction of oceanic crust

4.1 Abstract

Dolomite with compositional zoning was discovered in carbonate-lawsonite-bearing eclogites in the Tianshan (ultra-)high-pressure/low-temperature metamorphic belt, northwestern China. The eclogite-facies dolomite occurs as matrix porphyroblast and as inclusion in garnet, both of which display the same chemical zoning pattern. The dolomite contains inclusions of calcite (probably after aragonite), magnesite, glaucophane, lawsonite (and its pseudomorphs), allanite, epidote, paragonite, phengite and omphacite. The chemical zoning in dolomite is well defined by a continuous core– to–rim Mg increase and Fe–Mn decrease. The concentrations of transition metal elements, REE and Y also decrease from core to rim of the dolomite. Thermodynamic modeling demonstrates that the Fe–Mg zoning of dolomite is largely temperature dependent and, thus, is interpreted as prograde growth zoning, which developed during subduction of carbonate-bearing oceanic crust. It is suggested that dolomite in equilibrium with garnet formed as a result of changing matrix compositions due to increasing temperatures. In addition, thermodynamic modeling demonstrates that during

4. Compositional zoning in dolomite subduction at high-pressure conditions prograde-formed aragonite and dolomite were transformed to dolomite and magnesite. Furthermore, Fe-rich magnesite inclusions in matrix dolomite and in dolomite inclusions in garnet are shown to have formed during high-pressure conditions prior to peak metamorphic conditions and, therefore caution is warranted using Fe-bearing magnesite occurrences in eclogite-facies rocks as an unambiguous ultrahigh pressure indicator as previously suggested.

4.2 Introduction

Carbonates are common minerals in some blueschists and eclogites in (ultra-)high-pressure/low-temperature ((U)HP/LT) metamorphic terranes. Coesite inclusions in dolomite in eclogite-facies calc-silicate rocks and metabasalts from the Dabie ultrahigh-pressure (UHP) belt were interpreted as evidence for the subduction of continental crust (including sediments) to mantle depths (>100 km) and that coexisting dolomite and magnesite are stable mineral phases under UHP conditions (Schertl and Okay, 1994; Zhang and Liou, 1996). Coexisting magnesite, aragonite and calcite inclusions in dolomite have also been reported to occur in eclogites from several high-pressure (HP) and UHP metamorphic terranes and are regarded to be an UHP indicator (e.g., Wang and Liou, 1993; Zhang and Liou, 1994; Zhang et al., 2003). However, the coexistence of magnesite and dolomite as well as magnesite inclusions in dolomite in blueschists and eclogites were also interpreted to have formed under HP conditions (e.g., Klemd, 2003; Smit et al., 2008; Li et al., 2012).

Compositional zoning is a distinctive feature that is commonly observed in petrographical and chemical studies of minerals, and it may provide important information on the minerals‘ growth history and geological conditions during mineral formation. Compositional zoning, involving the variations in both trace and major elements, is most suitable for studying the dynamics of crystal growth (Tracy, 1982). Different types of compositional zoning in dolomite, including concentric zoning, sector zoning and oscillatory zoning, were studied through imaging techniques such as cathodoluminescence (CL), back scattered electron imaging (BSE) and X-ray

4. Compositional zoning in dolomite topographs in some sedimentary rocks (e.g., Farr, 1989; Reeder and Prosky, 1986; Reeder, 1991; Wogelius et al., 1992; Shore and Fowler, 1996). However, to our knowledge the compositional zoning of eclogite-facies dolomite has not been reported and thermodynamically modeled yet. Nonetheless it was rarely observed in other metamorphic carbonates (cf., Jones and Ghent, 1971; Reinecke et al., 2000).

In the present study, we present concentric Fe–Mg zoning of dolomite from HP lawsonite-carbonate-bearing eclogites in the southwest Tianshan (U)HP/LT metamorphic belt. The petrography and mineral chemistry of the dolomite and its inclusions were studied in detail by means of electron microprobe and laser ablation inductively coupled plasma mass spectrometry (LA–ICP–MS). Furthermore we conducted detailed textural and thermodynamic modeling studies of the dolomite and its inclusions in order to shed some light on the carbonate-phase formations and transitions in subducted oceanic crust.

4.3 Geological setting and petrography

The Chinese Tianshan (U)HP/LT metamorphic belt extends for >200 km along the South Tianshan suture zone separating the Yili (-Central Tianshan) block to the north and the Tarim block to the south (Fig. 3-1a; Gao et al., 1998, 2011; Han et al., 2011). It mainly consists of blueschist-, eclogite- and greenschist-facies meta-sedimentary and some mafic metavolcanic rocks, the latter of which have N-MORB, E-MORB, OIB and arc basalt affinities (Gao et al., 1999; Gao and Klemd, 2003; John et al., 2008). Peak metamorphism of some eclogites was estimated to have occurred under high-pressure conditions between 14 and 23 kbar at 480 and 580°C (e.g., Gao et al., 1999; Klemd et al., 2002; Wei et al., 2003; Li et al., 2012), while coesite relicts were identified in other eclogites or eclogite-facies rocks indicating that these rocks underwent UHP metamorphism (Zhang et al., 2005; Lü et al., 2008, 2009; Lü and Zhang, 2012). The intimate interlayering of HP and UHP rocks on a meter scale (Lü et al., 2009) was interpreted to be due to juxtaposition processes during subduction and exhumation in the subduction channel (Klemd et al., 2011). The timing of peak metamorphism was

4. Compositional zoning in dolomite determined by multi-point Lu–Hf isochron ages from four blueschist- or eclogite-facies rocks yielding consistent garnet-growth ages of ca. 315 Ma (Klemd et al., 2011). This age is in agreement, within error, with U–Pb SIMS ages of ca. 320 Ma on metamorphic zircon rims from eclogites (Su et al., 2010). White mica 40Ar–39Ar and Rb–Sr ages of eclogite-facies metavolcanic rocks and omphacite-bearing blueschists cluster at ca. 311 Ma and were found to represent a major post-eclogite-facies episode of cooling or recrystallization (Klemd et al., 2005).

Marbles and carbonate-bearing HP metabasalts, metapelites are commonly observed in the Tianshan (U)HP/LT metamorphic belt. The occurrence of magnesite and calcite inclusions in dolomite in eclogites and metapelites was considered as evidence for UHP metamorphism (Zhang et al., 2002, 2003), but the petrological interpretation of carbonate/silicate mineral assemblages in (U)HP rocks is often highly controversial due to disequilibrium, crystallographical or textural considerations (e.g., Klemd et al., 1994; Klemd, 2003; Smit et al., 2008; Hammouda et al., 2011). More recently, interlayered dolomite- and magnesite-bearing eclogite and blueschist in the Tianshan were interpreted to have coexisted at the same peak metamorphic HP conditions due to different bulk-rock compositions (Li et al., 2012).

This study focuses on the chemical zoning of dolomite from a HP carbonate-lawsonite-bearing eclogite (sample L0910, details see chapter 5). The peak metamorphic P–T conditions of the eclogite are estimated at 580–590 ºC and 2.2–2.4 GPa. The eclogite consists of garnet (ca. 16 vol.%), omphacite (ca. 45%), carbonate (ca. 13%), white mica (ca. 13%), epidote (ca. 5%), glaucophane (ca. 3%), sulfide (ca. 2%), quartz (ca. 1%) and accessory minerals such as rutile, titanite, lawsonite and apatite. Porphyroblastic garnet contains barroisite/glaucophane, epidote, paragonite, quartz and albite in the core domain and omphacite, dolomite/magnesite and rutile in the mantle domain while the rim domain usually is inclusion-free. Occasionally, lawsonite and magnesite were found to be enclosed by dolomite inclusions in garnet porphyroblasts. Omphacite is often orientated parallel to the weak foliation as defined by phengite. Furthermore, some glaucophane inclusions were detected in the omphacite cores.

4. Compositional zoning in dolomite

◀ Figure 4-1 Representative photomicrographs and BSE images of dolomite in carbonate- lawsonite-bearing eclogite in the Tianshan (U)HP/LT metamorphic belt. All=allanite, Cal =calcite, Chl=chlorite, Dol=dolomite, Ep=epidote, Gln=glaucophane, Grt=garnet, Law=lawsonite,

Mags.s.=magnesite-siderites.s., Omp=omphacite, Phn=phengite, Pg=paragonite, Py=pyrite, Qtz=quartz, Ttn=titanite. (a, b) Dolomite inclusions in garnet and coarse-grained matrix dolomite, both enclosing Mags.s. inclusions. (c) Mags.s. inclusions in dolomite containing glaucophane inclusions. (d) The zoned porphyroblastic dolomite contains inclusions of calcite, chlorite, phengite, paragonite, allanite and omphacite. The calcite was identified by the representative Raman shifts (see inset) at 154 cm-1, 279 cm-1, 713 cm-1 and 1085 cm-1. (e) Glaucophane inclusions in dolomite. (f) Phengite, allanite/epidote, omphacite and quartz inclusions in matrix dolomite. (g) Lawsonite inclusions and its pseudomorph (epidote+paragonite assemblages) in the core domain of matrix dolomite. (h) Calcite-chlorite- dolomite inclusions in garnet in nearby carbonate-bearing eclogite, see also Fig. 3-3d.

Significantly, the eclogite contains ca. 13 vol.% carbonates, mainly dolomite and minor magnesite and calcite (Table 4-1). The dolomite occurs as idioblastic/subidioblastic coarse-grained porphyroblastic matrix mineral and as inclusion in garnet (Fig. 4-1). The matrix dolomite (0.2–1 mm in diameter) contains inclusions of omphacite, glaucophane, lawsonite, phengite, paragonite, epidote, allanite, chlorite, calcite, magnesite and quartz (Fig. 4-1), while dolomite inclusions in garnet contain magnesite, paragonite and lawsonite inclusions (Fig. 4-1a–b; cf., Li et al., 2013). Magnesite occurs as inclusion in matrix dolomite and in dolomite inclusions in garnet (Fig. 4-1a–c). Calcite intergrown with chlorite, phengite and paragonite occurs in the core of idioblastic matrix dolomite only (Fig. 4-1d). Magnesite occasionally contains glaucophane inclusions (Fig. 4-1c). Epidote, which usually has an allanite core, occurs as matrix mineral and as inclusion in dolomite porphyroblasts (Fig. 4-1d and f). Lawsonite and its pseudomorphs (epidote-paragonite assemblages) were also identified as inclusions in matrix dolomite (Fig. 4-1g).

4.4 Analytical methods

In situ major element compositions of carbonates and inclusion minerals were obtained from polished thin sections by electron microprobe analysis (JEOL JXA 8200) at the GeoZentrum Nordbayern (GZN) of the University Erlangen–Nürnberg, Erlangen, Germany. Quantitative major element analyses were performed with an acceleration

4. Compositional zoning in dolomite voltage of 15 kV, a beam current of 15 nA, a beam diameter of 3 μm and 10–30s counting time, while qualitative mapping of Ca, Fe, Mg and Mn in dolomite was conducted in WDS with an acceleration voltage of 15 kV, a beam current of 230 nA, a 3–5 μm pixel size and dwell time of 100 ms. Natural minerals and synthetic oxides were used as standards, and a program based on the ZAF procedure was used for data correction. Representative microprobe analyses for the minerals in this study are presented in Table 4-1. In order to determine the nature of the CaCO3 polymorph occurring as inclusions in dolomite and garnet, Laser-Raman spectroscopy was performed at Department of Materials Science and Engineering in the same university.

The laser beam (wavelength of 533 nm) was focused on the CaCO3 inclusions by means of 50× objectives of a polarizing microscope. The laser spot size was focused to 1 µm.

In situ trace element analyses of zoned dolomite and inclusions were performed by LA– ICP–MS at the GZN using a single collector quadrupole Agilent 7500i ICP–MS equipped with an UP193Fx Argon Fluoride New Wave Research Excimer laser ablation system. The glass reference material NIST SRM 612 was used as standard for external calibration. LA–ICP–MS measurements were conducted using a spot size of 25µm in diameter, a laser frequency of 15 Hz and 0.63 GW/cm2 and a fluence of 3.32 J/cm2. The carrier gas consists of a mixture of 0.65 l/min helium and 1.10 l/min argon. Acquisition time was 20 s for the background and 25 s for the mineral analysis. The Ca-content of the carbonates determined by EMP analysis was used as internal standard. Reproducibility and accuracy, which were determined for NIST SRM 610, are usually <8% and <6%. The trace element concentrations were calculated by GLITTER Version 3 (van Achterbergh et al., 2000). Representative trace element analyses of the carbonate minerals are given in Table 4-2.

4.5 Mineral chemistry

Garnet has a prograde compositional growth zoning with a continuous increase of the pyrope component from 4.9 to 11.4 % and a decrease of the spessartine component from 4.2 to 1.6 % from the core towards the rim (Table 4-1). Omphacite inclusions in garnet

4. Compositional zoning in dolomite and dolomite and matrix omphacite have a similar jadeite component of between 32.8 mol.% and 46.5 mol.%. The Si–content of phengite is between 3.45–3.51 pfu and the Na-content of paragonite at ca. 0.80 pfu (Table 4-1). The glaucophane and epidote-group minerals display rather uniform compositions either as matrix or inclusion minerals (Table 4-1). For more detailed microprobe data concerning the silicates in the eclogite see chapter 5.

The dolomite studied here is a dolomite-ankerite solid solution (dolomites.s.) that

2+ contains 20–35 mol.% ankerite [CaFe(CO3)2] component due to the substitution of Fe for Mg (Table 4-1). This type of dolomite always contains small but variable amounts of Mn and less commonly minor Sr, Pb, Ni and Zn (Table 4-2). LA–ICP–MS data show that the dolomite contains high amounts of transition metal elements (e.g., Co: 18–60 ppm; Ni: 54–264 ppm and Zn: 114–147 ppm), Sr (449–1633 ppm), Pb (2.17–27.8 ppm), P (23–28 ppm) and minor Li, Ba, REE and Y (Table 4-2). The magnesite inclusions in dolomite are magnesite-siderite solid solutions (magnesites.s.) with a high siderite

[FeCO3] component (44–47 mol.%) and about 1 mol.% CaCO3 and even less MnCO3 (Table 4-1). Magnesite hosts higher amounts of transition metal elements (e.g., Co: 127–252 ppm; Ni: 181–533 ppm and Zn: 161–449 ppm) but much lower Sr (0.110–5.17 ppm), Pb, HREE and Y concentrations than the dolomite (Table 4-2). The HFSE concentrations in dolomite and magnesite are always below the detection limit (Table

4-2). The CaCO3 phase occurring as inclusions in dolomite is calcite according to Raman spectrometry (Fig. 4-1d) and contains minor Mg, Fe and Mn contents (Table 4-1). However, the stability of aragonite instead of calcite at HP/LT metamorphic conditions is suggested by the thermodynamic modeling (see below) which is in agreement with experimental studies (cf., Carlson, 1980). Thus, the calcite inclusions are thought to have formed at the expense of former aragonite at low-P and high-T during post peak metamorphic conditions (e.g., Carlson, 1980; Proyer et al., 2008).

4. Compositional zoning in dolomite

Table 4-1 Representative major element composition of minerals in dolomite-bearing eclogite

4.6 Compositional zoning of dolomite

4.6.1 Major elements

The zoning of dolomite is petrographically displayed by a change in color from core to rim in the BSE images (here displayed by changes of the gray intensity, Fig. 4-1d). The dolomite core contains inclusions of calcite (possibly after aragonite), magnesites.s.,

4. Compositional zoning in dolomite phengite, paragonite, epidote, lawsonite, epidote and paragonite intergrowth (possibly after lawsonite) and occasionally omphacite. Omphacite, phengite and allanite inclusions occur in the darker rim domains (Fig. 4-1d and f). The Ca content of dolomite is rather homogeneous with a slight increase at the outer-rim (Figs. 4-2 and 4-3a). The Fe content continuously decreases from core (12.66 wt.%) to rim (7.51 wt.%)

Figure 4-2 X-ray intensity maps of Ca, Fe, Mg and Mn in the dolomite porphyroblast from Fig. 4-1d. The compositional zoning is displayed by core–to–rim gradually increasing Mg and decreasing Fe and Mn, whereas Ca is rather homogeneous (for profile A–B see Fig. 4-3a). Red color refers high amounts whereas black color refers low amounts. The numbered circles (1–15) represent LA–ICP– MS spots, the data of which are listed in Table 4-2.

while the Mg content increases from 13.29 wt.% to 16.97 wt.% (Table 4-1). Thus XFe (=Fe/(Fe+Mg)) decreases accordingly (Fig. 4-3a). The Mn content decreases from the dolomite core (0.53 wt.%) towards the inner-rim (0.18 wt.%) while it increases at the outermost rim (0.40 wt.%). In addition, a subtle oscillatory Mn zoning occurs in the inner-rim domain (Fig. 4-2). Magnesium and Fe show a good negative correlation in the Mg vs. Fe plot (Fig. 4-3b). Interestingly, Mn shows a negative correlation with Mg and

4. Compositional zoning in dolomite a positive correlation with Fe when neglecting the outer-rim Mn composition (Fig. 4-3c–d). The large scale mapping images document that the dolomite inclusions in garnet and matrix dolomite grains share the same compositional core–to–rim zoning pattern; however, the dolomite inclusions in garnet lack the (mantle-) rim-domain in contrast to the matrix dolomite (Fig. 4-4).

Figure 4-3 Chemical composition profile of the dolomite porphyroblast from Fig. 4-2. (a) Fe and Mn decrease from core to rim while Mg increases and Ca is constant. Towards the outermost rim Ca and Mn increase slightly. (b) Good negative correlation of Mg and Fe. (c, d) Mn shows a negative correlation with Mg and a positive correlation with Fe, respectively.

4.6.2 Trace elements

The Sr and Pb concentrations decrease gradually from the dolomite core to the mantle and increase from the mantle to the inner-rim with an abrupt decrease at the outer-rim (Fig. 4-5a). Lithium, Ba and Sr display similar distribution patterns throughout the zoning profile (Fig. 4-5a). Manganese and Co exhibit a general decrease from core to rim and a slight oscillatory zoning at the inner-rim and an increase at the outermost rim

(Fig. 4-5b). Zinc displays a flat pattern. Vanadium, Cr and Ni concentrations decrease

4. Compositional zoning in dolomite

Figure 4-4 X-ray maps of Ca, Fe, Mg and Mn of dolomite grains from Fig. 4-1a–b. The dolomite inclusions in garnet show the same compositional zoning as the core (-mantle) domain of matrix dolomite. Table 4-2 Representative trace element composition (in ppm) of dolomite and magnesite. Aanlyses 1-15 refer to the profile in dolomite (Fig. 4-2 Mn).

4. Compositional zoning in dolomite from core to mantle and then increase towards the rim with a significant drop at the outermost rim (Fig. 4-5b). In general, the dolomite core contains higher REE and Y concentrations than the mantle and rim (Fig. 4-5c and d). The chondrite-normalized REE diagram displays a MREE (especially Eu) enrichment relative to the LREE and HREE (Fig. 4-5d). The REE patterns presented here are in accordance with those of dolomite from other Tianshan eclogites (van der Straaten et al., 2008) but are quite different with the flat REE patterns in dolomite from Central Dabie coesite-bearing eclogites (Sassi et al., 2000). This discrepancy is believed to be due to different chemical bulk compositions of the protoliths (e.g. oceanic crust vs. continental crust) and/or the different rock mineralogy, which strongly controls the trace element distributions. The magnesite shows a relative HREE enrichment compared to the LREE and MREE (Fig. 4-5e).

4.7 Thermodynamic modeling

In order to obtain information on the carbonate-phase transitions as well as compositional variation of carbonates during HP/LT metamorphism, thermodynamic modeling of the carbonate-bearing eclogite L0910 was undertaken in the

NMnCaKFMASCHO (Na2O–MnO–CaO–K2O–FeO–MgO–Al2O3–SiO2–CO2–H2O–

Fe2O3) model system using Perple_X software (Connolly, 1990, 2005) and internally consistent thermodynamic dataset (Holland and Powell, 1998; and update) based on the effective bulk-rock composition (Table 4-3). Mineral solid-solution models are Gt(WPH) for garnet (White et al., 2007), Omph(GHP) for omphacite (Green et al., 2007), Amph(DPW) for amphibole (Dale et al., 2005), Mica(CHA) for phengite (Coggon and Holland, 2002), Chl(HP) for chlorite (Holland and Powell, 1998), Ep(HP) for epidote

(Holland and Powell, 1998), and F for H2O–CO2 fluid solution (Connolly and Trommsdorff, 1991). In particular, the recently developed solid-solution model odCcMS(EF) for ternary Ca–Fe–Mg carbonates (Franzolin et al., 2011) was used to determine compositional variations in dolomite, the transition of carbonate minerals and the effect of Fe on carbonate stabilities during subduction conditions. For more details

4. Compositional zoning in dolomite concerning the pseudosection calculations see chapter 5.

Figure 4-5 Trace elements in the zoned matrix dolomite porphyroblast from Fig. 4-2 and the magnesite inclusions in dolomite. (a–c) Trace element concentrations along a profile from core to rim in the dolomite. (a) Li–Ba–Sr–Pb. (b) Transition metal elements. (c) REE and Y. (d) Chondrite-normalized rare earth element patterns. The dolomite core contains higher REE (especially HREE) contents than the rim. (e) Chondrite-normalized rare earth element patterns of magnesite inclusions in dolomite. Normalization values are after Sun and McDonough (1989).

A calculated phase equilibrium diagram is presented in Figure 4-6a. Omphacite, phengite and garnet are stable phases over the whole shown P–T range, while lawsonite occurs in the low–T range from which epidote-group minerals are absent. Amphibole is absent in the HT/HP stability fields. Dolomite is stable over a wide P–T range, while magnesite occurs in the higher-pressure and aragonite in the LP/LT fields only (Fig.

4. Compositional zoning in dolomite

4-6b). The modal amounts of dolomite, aragonite, magnesite, chlorite, amphibole, omphacite, lawsonite and garnet, along with the XFe isopleths of magnesites.s. and dolomites.s. were contoured in the modeled pseudosection (Fig. 4-6). The modal amount of dolomite increases and that of aragonite decreases with increasing P and T in the aragonite-dolomite transition area (Fig. 4-6a). The modal amount of magnesite decreases and that of dolomite increases with increasing T in the magnesite-bearing P–T

o range of 550 to 600 C (Fig. 4-6a). The XFe isopleths of magnesites.s. are parallel to the P-axis and decrease in value from low to high temperatures in the amphibole-absence field (Fig. 4-6b). Similarly, the XFe isopleths of dolomites.s. are parallel to those of magnesites.s. and also decrease from low to high temperatures in this field (Fig. 4-6b).

The chemical changes as displayed by the XFe contents of the dolomites.s. are in agreement with the compositional variations observed in the zoned dolomites.s. in the lawsonite-bearing eclogite (Figs. 4-2 and 4-3a). The P–T paths of the studied eclogite and nearby eclogites (details see chapter 3 and 5) indicate that the low temperature/low pressure aragonite was progradely transformed to dolomites.s. and then to dolomites.s. and magnesites.s. during increasing P–T conditions (Fig. 4-6b, Table 4-3). The modal amount of omphacite (and of dolomites.s. and magnesites.s.) increases at the expense of chlorite and amphibole with increasing P–T conditions (Fig. 4-6b–d). Garnet growth at the expense of lawsonite started at temperatures between 550 and 600oC (Fig. 4-6f and g).

4.8 Discussion

4.8.1 Formation of dolomite and its prograde compositional zoning

Significant amounts of carbonates are introduced into mafic oceanic crust by hydrothermal alteration prior to subduction (e.g., Staudigel, 2003). The crystallographically-bound CO2 that preferentially occurs in carbonates in the slab rocks is carried down to great mantle depths without prominent decarbonation (Kerrick and

4. Compositional zoning in dolomite

Connolly, 2001). Dolomitic carbonates are generally stable at pressures of up to 2–7 GPa in mafic eclogites (e.g., Yaxley and Green, 1994; Molina and Poli, 2000; Poli et al., 2009). High-pressure experimental data on the carbonate stability based on a basaltic composition in the presence of a H2O–CO2 mixed fluid demonstrate that calcite is stable at P ≤ 1.4 GPa, dolomite at P between 1.4 and 1.8 GPa, and dolomite and magnesite at P ≥ 1.8 GPa (Molina and Poli, 2000). These results are in accordance with the presented modeling which suggests the stability of aragonite at T ≤ 450oC, P ≤ 1.7 GPa, of

Figure 4-6 (a) P–T pseudosection (using Perple_X) in the system NMnCaKFMASCHO for the eclogite based on an effective bulk composition. The pseudosection is contoured for aragonite, dolomites.s. and magnesites.s. modal abundances (vol.%). (Amp = amphibole; Arag = aragonite; Chl = chlorite; Dol = dolomite; Ep = epidote; F = fluid; Grt = garnet; Law = lawsonite; Mag = magnesite; Omp = omphacite; Phn = phengite; Q = quartz; Rhc = rhodochrosite; Zoi = zoisite). For more mineral assemblages see Table 4-3. (b) The P–T path (for details see Li et al., 2012, 2013) indicates that carbonates in subducted oceanic crust undergo the aragonite → dolomite → dolomite+magnesite transitions during prograde subduction zone metamorphism. The pseudosections are contoured with XFe isopleths in magnesites.s. and dolomites.s.. The isopleths are largely T-dependent in area C: the Fe-content decreases and the Mg-content increases from low-T to high-T, which is in agreement with the measured compositional zoning in eclogite-facies dolomite of this study. (c–g) The contoured modal amounts (vol.%) of chlorite (c), amphibole (d), omphacite (e), lawsonite (f) and garnet (g).

4. Compositional zoning in dolomite

Table 4-3 Bulk composition used for pseudosection calculation and mineral assembalges from Fig. 4-6a

Sample SiO2 Al2O3 Na2O K2O FeO Fe2O3 MgO CaO MnO CO2 H2O L0910 43.39 12.71 3.44 1.19 7.24 1.9 6.73 12.76 0.11 5.77 4.68 Field A: A: Omp Phn Chl Amp Dol Mag Law Rhc F A1: Amp Omp Grt Phn Chl Arag Law Q F B: Amp Omp Phn Chl Dol Law Arag Rhc F A2: Amp Omp Grt Phn Chl Dol Arag Law Q F C: Amp Omp Grt Phn Chl Arag Law Ep Q F A3: Amp Omp Grt Phn Chl Dol Arag Law F D: Amp Omp Grt Phn Chl Dol Arag Law Ep Q F E: Amp Omp Grt Phn Chl Dol Arag Law Ep F Field B: F: Amp Omp Grt Phn Chl Arag Ep Q F B1: Amp Omp Grt Phn Chl Dol Law F G: Amp Omp Grt Phn Chl Dol Arag Ep Q F B2: Omp Grt Amp Phn Chl Dol Mag Law F H: Amp Omp Grt Phn Chl Dol Arag Ep F B3: Omp Grt Amp Phn Dol Mag Law F I: Amp Omp Grt Phn Dol Law F B4: Omp Grt Amp Phn Dol Mag Law F Q J: Amp Omp Grt Phn Dol Law Ep F K: Amp Omp Grt Phn Dol Law Ep Q F Field C: L: Amp Omp Grt Phn Dol Mag Law Ep Q F C1: Omp Grt Phn Dol Mag Law Q F M: Amp Omp Grt Phn Dol Mag Ep Q F C2: Omp Grt Phn Dol Mag Law Ep Q F N: Omp Grt Phn Dol Law Q F C3: Omp Grt Phn Dol Mag Ep Q F O: Amp Omp Grt Phn Dol Zoi Q F dolomite at 1.7 ≤ P ≤ 2.1 GPa and of dolomite and magnesite at even higher pressures (e.g., P ≥ 2.1 GPa) (Fig. 4-6).

The occurrence of calcite (after aragonite)-chlorite inclusions in dolomite core domains in the studied sample (Fig. 4-1d) and calcite (after aragonite)- chlorite-dolomite inclusions in garnet cores (Fig. 4-1h) in adjacent eclogite (details see chapter 3) suggest the prograde formation of dolomite at LT–LP conditions (400 ≤ T ≤ 480oC, 1.5 ≤ P ≤ 2.1 GPa) according to the continuous reaction:

Arag + Chl + amp → Dols.s.1 + Omp + H2O (1), this reaction corresponds to a decrease in the modal amounts of aragonite, chlorite and amphibole and an increase of dolomite and omphacite in the pseudosection (see hatched area A; Fig. 4-6a and c–e). During increasing P–T conditions dolomite partly reacts to magnesite (hatched area B) according to the reaction:

Dols.s.1 + Chl + Amp → Mags.s. + Omp ± Grt + H2O (2), this reaction is displayed by increasing modal amounts of magnesite and omphacite

4. Compositional zoning in dolomite

(and garnet) which formed at the expense of chlorite and amphibole along the P–T path (hatched area B in the pseudosection, Fig. 4-6a–e). This is supported by glaucophane inclusions in magnesite (Fig. 4-2c). The lawsonite and magnesites.s. inclusions in matrix dolomite (Figs. 4-1c and g) and in dolomite inclusions in garnet (details see chapter 5) suggest that with increasing temperatures and pressures dolomite formed at the expense of magnesite (e.g., 550 ≤ T ≤ 600oC, P ≥ 2.1 GPa) according to the continuous reaction:

Mags.s. + Dols.s.1+ Law + Omp → Dols.s.2 + Grt + Qtz + H2O (3), corresponding with gradually decreasing modal abundances of magnesite and lawsonite (Fig. 4-6a and f) and increasing modal amounts of dolomite and garnet (Fig. 4-6a and g) in the pseudosection (hatched area C; Fig. 4-6b). The change in the modeled chemical composition (XFe) of the dolomite (Fig. 4-6b) is in good agreement with that of the

EPM-analyses of the dolomites.s.2 (Table 4-1, No. 4–8; Fig. 4-3a). In addition, the measured chemical composition of the magnesites.s. inclusions (XFe ≈ 0.47) in dolomite

(Table 4-1, No. 9–10) corresponds with that of the modeled magnesites.s. (XFe =0.40– 0.50) in the pseudosection (area B–C; Fig. 4-6b).

Chemical zoning in metamorphic minerals relates to the changing effective bulk rock composition and the precipitating fluid during changing pressures and/or temperatures (e.g., Tracy, 1982). In case of a closed chemical system chemical zoning is directly related to the availability of reactants and the degree of reaction completion. Reaction (3) suggests that dolomite and garnet formed simultaneously. The garnet in eclogite L0910 shows distinctive prograde growth zoning with continuously decreasing Fe- and Mn-contents and a gradually increasing Mg-content from core to rim, respectively (details see chapter 5). X-ray maps and composition profiles display that the dolomite porphyroblasts exhibit the same compositional zoning as the garnets (Figs. 4-2 and 4-3a). The symmetric concentric chemical zoning occurs both in matrix dolomite and dolomite inclusions in garnet (Fig. 4-4) suggesting that the zoning is the product of prograde growth processes. This assumption is supported by the following evidence: 1) the dolomite contains mineral inclusions of glaucophane, lawsonite (and its

4. Compositional zoning in dolomite pseudomorphs), chlorite, allanite, epidote, white mica and omphacite (Fig. 4-1) indicating that the growth of dolomite may have occurred during the blueschist-facies to eclogite-facies stages; 2) the thermodynamic modeling reveals that the Fe substitution of Mg in dolomite decreases during the prograde P–T evolution (Fig. 4-6), which is in agreement with the observed Fe–Mg zoning in dolomite; and 3) matrix dolomite grains and dolomite inclusions in garnet have the same core (-mantle) compositional zoning pattern (Fig. 4-4).

Decreasing Fe and increasing Mg contents from core to rim of the dolomite are believed to be the result of a changing matrix bulk rock composition during temperature increase as revealed by the thermodynamic modeling (Fig. 4-6). The prograde chlorite and amphibole breakdown provides the Mg for the dolomite formation and at the same time the garnet crystallization reduces the Fe concentration in the system. However, the core–to–inner rim Mn decrease of the dolomite is thought to be the result of fractional crystallization, which was already proposed for similar zoned siderite porphyroblasts from greenschist-facies phyllites of the Esplanade Range and Northern Dogtooth Mountains in British Columbia (Jones and Ghent, 1971). Trace element zoning of dolomite such as decreasing core–to–rim concentrations of Sr, Pb, transition metal elements, REE (Fig. 4-5a–d) and Y (cf., garnet, Hollister, 1966), is believed to be further evidence for fractional crystallization of the dolomite. The abrupt Mn increase at the outer-rim of the dolomite accompanied by large compositional changes of the Fe- and Mg-contents (Figs. 4-2 and 4-3) along with the sudden variations of trace elements (Fig. 4-5a–c), may be attributed to variations in the equilibrium fluid composition or to disequilibrium crystallization during the last porphyroblastic growth stages (cf., Jones and Ghent, 1971).

4.8.2 The occurrence of magnesite in dolomite in the Tianshan eclogites

The occurrence of magnesite ± calcite (former aragonite) inclusions in dolomite was

4. Compositional zoning in dolomite interpreted as evidence for UHP metamorphism in excess of 5 GPa for eclogite- facies-rocks from the Tianshan orogen (Zhang et al., 2002, 2003). These authors referred to the equilibrium reaction dolomite = magnesite + aragonite in order to deduce these exceptionally high UHP conditions (cf., Hammouda et al., 2011, and references therein). The stability of magnesite at ultrahigh pressure conditions (e.g., >5GPa) was confirmed by a series of experiments (Hammouda et al., 2011, and references therein) and some field observations (e.g., Zhang and Liou, 1994; Messiga et al., 1999). On the other hand, Fe-rich magnesites.s. and dolomites.s. were experimentally shown to be stable at HP as well as UHP conditions in mafic Fe-rich eclogites. However, the influence of Fe on the carbonate structure and thus its stability was not discussed in detail by these studies (e.g., Dasgupta et al., 2004; Yaxley and Brey, 2004), although recently it was experimentally shown that the Fe-content plays an decisive role in effecting the ordering-disordering state of dolomite (Franzolin et al., 2012) and, thus, has a considerable effect on the magnesite-aragonite-dolomite stability relationship (Hammouda et al., 2011). In Fe-bearing systems, experimental results suggest that the nucleation of aragonite and magnesite occurs prior to that of dolomite at relatively low-pressure/low-temperature conditions. Furthermore, a relatively high Fe proportion is expected to shift the aragonite + magnesites.s.↔ dolomite reaction toward higher temperatures, enlarging the stability field of aragonite and magnesites.s. (Franzolin et al.,

o 2012). For instance, aragonite and magnesites.s. are stable between 400 and 450 C at ca. 2 GPa and react to highly disordered dolomite only at temperatures > 450oC in Fe-rich systems (see Fig. 10 in Hammouda et al., 2011). In the present study, the application of a new thermodynamic solid-solution model for Ca–Mg–Fe carbonate provides an opportunity to investigate the transition of carbonate minerals during subduction zone metamorphism. In the modeled complex system using a particular bulk chemistry (Fig. 4-6a) aragonite occurs in the low temperature/low pressure stability fields in the phase equilibrium diagram while magnesites.s. appears only at P > 2.0 GPa (Fig. 4-6) indicating that Fe-rich magnesite (in equilibium with dolomite) is stable under high-pressure conditions (P ≤ 2.5 GPa, T ≤ 600oC), but is expected to react out to dolomite at temperatures > 600oC (Fig. 4-6).

4. Compositional zoning in dolomite

Thus, the reported magnesites.s. inclusions in dolomites.s.2 (Table 4-1) from the Tianshan

HP eclogites may have formed prograde in equilibrium with dolomites.s.1 and not during decompression as previously proposed (Zhang et al., 2003). This interpretation is also supported by glaucophane inclusions in magnesite (Fig. 4-1c) and the occurrence of dolomite-rimmed magnesite inclusions in the core domain of prograde garnet porphyroblasts (Fig. 4-1a–b; see also Li et al., 2012). These textural observations indicate that Fe-magnesite formed at the expense of glaucophane between 500 and 560oC at 2.0 ≤ P ≤ 2.3 GPa and was then transformed to dolomite prior to or simultaneously with the pervasive garnet growth during increasing P–T conditions. Consequently caution is warranted in interpreting magnesite occurrences in dolomite in eclogite facies rocks as unambiguous evidence for UHP metamorphism (see also Smit et al., 2008). Alternatively, fluid flow along micro cracks may enter the inner part of garnet (or other container minerals) and thus may change the carbonate composition of inclusions forming pseudomorph after older minerals. In addition, it should be kept in mind that magnesite and dolomite assemblages cannot be treated as an absolute pressure indicator without a strict control of the relative amounts of volatile and non-volatile components in multi-component Fe-bearing systems (Poli et al., 2009).

4.9 Implications

Compositionally zoning was observed in dolomite in a lawsonite-bearing eclogite from the western Tianshan HP–UHP metamorphic belt. Evidence from mineral inclusions, thermodynamic modeling and textures suggest that the zoning formed during prograde dolomite growth under HP metamorphism. To our knowledge the compositional zoning of carbonate is very rare in metamorphic rocks, and this is the first report showing the possibility to retrieve the prograde metamorphic evolution from high-pressure carbonates, which are generally believed to record only the latest stages of the metamorphic evolution of eclogite-facies rocks. The core–to–rim Mg–Fe–Mn zoning of dolomite formed contemporaneously with garnet thereby suggesting that the dolomite, like garnet, also experienced fractional crystallization during the HP metamorphism.

4. Compositional zoning in dolomite

Therefore more attention should be paid to the carbonates when considering mineral equilibrium and effective bulk-rock composition calculations in HP–UHP carbonate-bearing rocks. In addition, our findings show that the chemical zoning in dolomite is largely temperature dependent and thus the analysis of major and trace elements in carbonates from blueschists and eclogites may bear important thermobarometric implications.

The present study shows that carbonates in subducted oceanic crust may experience gradual aragonite → dolomite → dolomite+magnesite transition stages during blueschist- to eclogite-facies conditions. The occurrence of magnesites.s. inclusions in dolomite enclosed in garnet indicates that the magnesites.s. had formed already under HP conditions. Fe content in magnesite determines its stability at lower pressures. Thus, caution is warranted using magnesite occurrences in eclogite-facies rocks as unambiguous evidence for UHP metamorphism, in particular in a multi-component Fe-bearing system.

Acknowledgements

This study was funded by the Deutsche Forschungsgemeinschaft (KL692/17-3) and the National Natural Science Foundation of China (41025008). We are grateful to Dr. H. Brätz for assistance with the LA–ICP–MS and to Dr. H.X. Bao for help with the Raman measurements. Dr. T. John is thanked for discussion on various topics of the manuscript. The first author thanks the DAAD for the scholarship supporting his PhD study in Germany. Constructive reviews by S. Poli and H.P. Schertl are greatly appreciated. E. Ghent is thanked for his comments and editorial work. This publication is a contribution to IGCP Project 592.

References

Carlson W.D. (1980) The calcite-aragonite equilibrium: effects of Sr substitution and anion orientational disorder. American Mineralogist 65, 1252–1262. Coggon R. and Holland T.J.B. (2002) Mixing properties of phengitic micas and revised garnet-phengite thermobarometers. Journal of Metamorphic Geology 20, 683–696.

4. Compositional zoning in dolomite

Connolly J.A.D. (1990) Multivariable phase-diagrams: an algorithm based on generalized thermodynamics. American Journal of Science 290, 666–718. Connolly J.A.D. (2005) Computation of phase equilibria by linear programming: A tool for geodynamic modelling and its application to subduction zone decarbonation. Earth and Planetary Science Letters 236, 524–541. Connolly J.A.D. and Trommsdorff V. (1991) Petrogenetic grids for metacarbonate rocks: pressure-temperature phase-diagram projection for mixed-volatile systems. Contributions to Mineralogy and Petrology 108, 93–105. Dale J., Powell R., White R.W., Elmer F.L. and Holland T.J.B. (2005) A thermodynamic

model for Ca-Na clinoamphiboles in Na2O-CaO-FeO-MgO-Al2O3-SiO2-H2O-O for petrological calculations. Journal of Metamorphic Geology 23, 771–791. Dasgupta R., Hirschmann M.M. and Withers A.C. (2004) Deep global cycling of carbon constrained by the solidus of anhydrous, carbonated eclogite under upper mantle conditions. Earth and Planetary Science Letters 227, 73–85. Farr M.R. (1989) Compositional zoning characteristics of late dolomite cement in the cambrian bonneterre formation, Missouri: Implications for parent fluid migration pathways. Carbonate Evaporite 4, 177–194. Franzolin E., Schmidt M.W. and Poli S. (2011) Ternary Ca-Fe-Mg carbonates: subsolidus phase relations at 3.5 GPa and a thermodynamic solid solution model including order/disorder. Contributions to Mineralogy and Petrology 161, 213–227. Franzolin E., Merlini M., Poli S. and Schmidt M.W. (2012) The temperature and compositional dependence of disordering in Fe-bearing dolomites. American Mineralogist 97, 1676–1684. Gao J. and Klemd R. (2003) Formation of HP-LT rocks and their tectonic implications in the western Tianshan Orogen, NW China: geochemical and age constraints. Lithos 66, 1–22. Gao J., Li M.S., Xiao X.C., Tang Y.Q. and He G.Q. (1998) Paleozoic tectonic evolution of the Tianshan Orogen, northwestern China. Tectonophysics 287, 213–231. Gao J., Klemd R., Zhang L., Wang Z. and Xiao X. (1999) P-T path of high-pressure/low-temperature rocks and tectonic implications in the western Tianshan Mountains, NW China. Journal of Metamorphic Geology 17, 621–636. Gao J., Klemd R., Qian Q., Zhang X., Li J.L. and Jiang T. (2011) The collision between the Yili and Tarim blocks of the Southwestern Altaids: Geochemical and age constraints of a leucogranite dike crosscutting the HP-LT metamorphic belt in the Chinese Tianshan Orogen. Tectonophysics 499, 118–131. Green E., Holland T. and Powell R. (2007) An order-disorder model for omphacitic pyroxenes in the system jadeite-diopside-hedenbergite-acmite, with applications to eclogitic rocks. American Mineralogist 92, 1181–1189. Hammouda T., Andrault D., Koga K., Katsura T. and Martin A.M. (2011) Ordering in double carbonates and implications for processes at subduction zones. Contributions to Mineralogy and Petrology 161, 439–450. Han B.-F., He G.-Q., Wang X.-C. and Guo Z.-J. (2011) Late Carboniferous collision between the Tarim and Kazakhstan-Yili terranes in the western segment of the South Tian Shan Orogen, Central Asia, and implications for the Northern Xinjiang, western

4. Compositional zoning in dolomite

China. Earth-Science Reviews 109, 74–93. Holland T.J.B. and Powell R. (1998) An internally consistent thermodynamic data set for phases of petrological interest. Journal of Metamorphic Geology 16, 309–343. Hollister L.S. (1966) Garnet Zoning: An Interpretation Based on the Rayleigh Fractionation Model. Science 154, 1647–1651. John T., Klemd R., Gao J. and Garbe-Schonberg C.D. (2008) Trace-element mobilization in slabs due to non steady-state fluid-rock interaction: Constraints from an eclogite-facies transport vein in blueschist (Tianshan, China). Lithos 103, 1–24. Jones J.W. and Ghent E.D. (1971) Zoned siderite porphyroblasts from the Esplanade range and northern Dogtooth mountains, British Columbia. American Mineralogist 56, 1910–1916. Kerrick D.M. and Connolly, J.A.D. (2001) Metamorphic devolatilization of subducted oceanic metabasalts: implications for seismicity, arc magmatism and volatile recycling. Earth and Planetary Science Letters 189, 19–29. Klemd R. (2003) Ultrahigh-pressure metamorphism in eclogites from the western Tianshan high-pressure belt (Xinjiang, western China) - Comment. American Mineralogist 88, 1153–1156. Klemd R., Matthes S. and Schüssler U. (1994) Reaction textures and fluid behaviour in very high-pressure calc-silicate rocks of the Münchberg gneiss complex, Bavaria, Germany. Journal of Metamorphic Geology 12, 735–745. Klemd R., Schröter F.C., Will T.M. and Gao J. (2002) P-T evolution of glaucophane- omphacite bearing HP-LT rocks in the western Tianshan Orogen, NW China: new evidence for 'Alpine-type' tectonics. Journal of Metamorphic Geology 20, 239–254. Klemd R., Bröcker M., Hacker B.R., Gao J., Gans P. and Wemmer K. (2005) New age constraints on the metamorphic evolution of the high-pressure/low-temperature belt in the western Tianshan mountains, NW China. Journal of Geology 113, 157–168. Klemd R., John T., Scherer E.E., Rondenay S. and Gao J. (2011) Changes in dip of subducted slabs at depth: Petrological and geochronological evidence from HP-UHP rocks (Tianshan, NW-China). Earth and Planetary Science Letters 310, 9–20. Li J.L., Klemd R., Gao J. and Meyer M. (2012) Coexisting carbonate-bearing eclogite and blueschist in SW Tianshan, China: petrology and phase equilibria. Journal of Asian Earth Sciences 60, 174–187. Li J.L., Gao J., Klemd R., John T. and Su W. (2013) Fluid-mediated metal transport in subduction zones and its link to arc-related giant ore deposits: Constraints from a sulfide-bearing HP vein in lawsonite eclogite (Tianshan, China). Geochimica et Cosmochimica Acta 120, 326–362. Lü Z. and Zhang, L.F. (2012) Coesite in the eclogite and schist of the Atantayi Valley, southwestern Tianshan, China. Chinese Science Bulletin 57, 1467–1472. Lü Z., Zhang L.F., Du J.X. and Bucher K. (2008) Coesite inclusions in garnet from eclogitic rocks in western Tianshan, northwest China: Convincing proof of UHP metamorphism. American Mineralogist 93, 1845–1850. Lü Z., Zhang L.F., Du J.X. and Bucher K. (2009) Petrology of coesite-bearing eclogite from Habutengsu Valley, western Tianshan, NW China and its tectonometamorphic implication. Journal of Metamorphic Geology 27, 773–787.

4. Compositional zoning in dolomite

Messiga B., Kienast J.R., Rebay G., Riccardi M.P. and Tribuzio R. (1999) Cr-rich magnesiochloritoid eclogites from the Monviso ophiolites (Western Alps, Italy). Journal of Metamorphic Geology 17, 287–299. Molina J.F. and Poli, S. (2000) Carbonate stability and fluid composition in subducted

oceanic crust: an experimental study on H2O-CO2-bearing basalts. Earth and Planetary Science Letters 176, 295–310. Poli S., Franzolin E., Fumagalli P. and Crottini A. (2009) The transport of carbon and hydrogen in subducted oceanic crust: An experimental study to 5 GPa. Earth and Planetary Science Letters 278, 350–360. Proyer A., Mposkos E., Bazicitis I. and Hoinkes G. (2008) Tracing high-pressure metamorphism in marbles: Phase relations in high-grade aluminous calcite-dolomite

marbles from the Greek Rhodope massif in the system CaO-MgO-Al2O3-SiO2-CO2 and indications of prior aragonite. Lithos 104, 119–130. Reeder R.J. (1991) An overview of zoning in carbonate minerals. In C.E. Barker and O.C. Kopp, Eds., Luminescence Microscopy: Quantitative and Qualitative Aspects. SEPM Short Course Notes 25, 77–82. Reeder R.J. and Prosky J.L. (1986) Compositional sector zoning in dolomite. Journal of Sedimentary Petrology 56, 237–247. Reinecke T., Bernhardt H.J. and Wirth R. (2000) Compositional zoning of calcite in a high-pressure metamorphic calc-schist: clues to heterogeneous grain-scale fluid distribution during exhumation. Contributions to Mineralogy and Petrology 139, 584–606. Sassi R., Harte B., Carswell D.A. and Yujing, H. (2000) Trace element distribution in Central Dabie eclogites. Contributions to Mineralogy and Petrology 139, 298–315. Schertl H.P. and Okay A.I. (1994) A coesite inclusion in dolomite in Dabie-Shan, China: petrological and theological significance. European Journal of Mineralogy 6, 995– 1000. Shore M. and Fowler A.D. (1996) Oscillatory zoning in minerals: A common phenomenon. Canadian Mineralogist 34, 1111–1126. Smit M.A., Bröcker M. and Scherer E.E. (2008) Aragonite and magnesite in eclogites

from the Jaeren nappe, SW Norway: disequilibrium in the system CaCO3-MgCO3 and petrological implications. Journal of Metamorphic Geology 26, 959–979. Staudigel H. (2003) Hydrothermal alteration processes in the Oceanic crust. In H. Holland and K. Turekian, Eds., The Crust 3, 511–537, Treatise on Geochemistry, Elseiver, New York. Su W., Gao J., Klemd R., Li J.L., Zhang X., Li X.H., Chen N.S. and Zhang L. (2010) U-Pb zircon geochronology of Tianshan eclogites in NW China: implication for the collision between the Yili and Tarim blocks of the southwestern Altaids. European Journal of Mineralogy 22, 473–478. Sun S.S. and McDonough W.F. (1989) Chemical and isotopic systematics of oceanic basalts: implications for mantle composition and processes. In A.D. Saunders and M.J. Norry, Eds., Magmatism in the Ocean Basins. Geological Society Special Publication 42, 313–345. Tracy R.J. (1982) Compositional zoning and inclusions in metamorphic minerals. In

4. Compositional zoning in dolomite

J.M. Ferry, Eds., Characterization of Metamorphism Through Mineral Equilibria. Reviews in Mineralogy and Geochemistry 10, 355–397. van Achterbergh E., Ryan C. G. and Griffin W. L. (2000) GLITTER: on-line interactive data reduction for the Laser Ablation ICP–MS Microprobe. Ninth Annual V. M. Goldschmidt Conference, Cambridge. 7215. van der Straaten F., Schenk V., John T. and Gao J. (2008) Blueschist-facies rehydration of eclogites (Tian Shan, NW-China): Implications for fluid-rock interaction in the subduction channel. Chemical Geology 255, 195–219. Wang X.M. and Liou J.G. (1993) Ultra-high-pressure metamorphism of carbonate rocks in the Dabie Mountains, central China. Journal of Metamorphic Geology 11, 575– 588. Wei C.J., Powell R. and Zhang, L.F. (2003) Eclogites from the south Tianshan, NW

China: petrological characteristic and calculated mineral equilibria in the Na2O-CaO- FeO-MgO-Al2O3-SiO2-H2O system. Journal of Metamorphic Geology 21, 163–179. White R.W., Powell R. and Holland T.J.B. (2007) Progress relating to calculation of partial melting equilibria for metapelites. Journal of Metamorphic Geology 25, 511– 527. Wogelius R.A., Fraser D.G., Feltham D.J. and Whiteman, M.I. (1992) Trace-element zoning in dolomite: Proton microprobe data and thermodynamic constraints on fluid compositions. Geochimica et Cosmochimica Acta 56, 319–334. Yaxley G.M. and Green D.H. (1994) Experimental demonstration of refractory carbonate-bearing eclogite and siliceous melt in the subduction regime. Earth and Planetary Science Letters 128, 313–325. Yaxley G.M. and Brey G.P. (2004) Phase relations of carbonate-bearing eclogite assemblages from 2.5 to 5.5 GPa: implications for petrogenesis of carbonatites. Contributions to Mineralogy and Petrology 146, 606–619. Zhang L.F., Ellis D.J., Williams S. and Jiang W.B. (2002) Ultra-high pressure metamorphism in western Tianshan, China: Part II. Evidence from magnesite in eclogite. American Mineralogist 87, 861–866. Zhang L., Ellis D.J., Arculus R.J., Jiang W. and Wei C. (2003) 'Forbidden zone' subduction of sediments to 150 km depth - the reaction of dolomite to magnesite plus aragonite in the UHPM metapelites from western Tianshan, China. Journal of Metamorphic Geology 21, 523–529. Zhang L.F., Song S.G., Liou J.G., Ai Y.L. and Li X.P. (2005) Relict coesite exsolution in omphacite from Western Tianshan eclogites, China. American Mineralogist 90, 181– 186. Zhang R.Y. and Liou J.G. (1994) Significance of magnesite paragenesis in ultrahigh-pressure metamorphic rocks. American Mineralogist 79, 397–400. Zhang R.Y. and Liou J.G. (1996) Coesite inclusions in dolomite from eclogite in the southern Dabie Mountains, China: The significance of carbonate minerals in UHPM rocks. American Mineralogist 81, 181–186.

5. Metal transport in subduction fluids

5 Fluid-mediated metal transport in subduction zones and its link to arc-related giant ore deposits: Constraints from a sulfide- bearing HP vein in lawsonite eclogite (Tianshan, China)

Ji-Lei Lia, Jun Gaob, Timm Johnc, Reiner Klemda, Wen Sub aGeoZentrum Nordbayern, Universität Erlangen–Nürnberg, Schlossgarten 5a, D-91054 Erlangen, Germany bKey Laboratory of Mineral Resources, Institute of Geology and Geophysics, Chinese Academy of Sciences, P.O. Box 9825, Beijing 100029, China cInstitut für Mineralogie, Universität Münster, Corrensstr. 24, D-48149 Münster, Germany

5.1 Abstract

High-pressure (HP) veins in eclogites provide insight into element mobility during fluid-rock interaction in subduction zones. Here, we present a petrological- geochemical study of a sulfide-bearing HP vein and its massive lawsonite eclogite host rock from the Chinese Tianshan (ultra-)high-pressure/low-temperature metamorphic belt. The omphacite-dominated vein is enveloped by a garnet-poor, sulfide-bearing eclogite-facies reaction selvage. Lawsonite, garnet, omphacite, glaucophane and other HP minerals occur as inclusions in pyrite porphyroblasts of the selvage rock, indicating that the selvage formed prograde under eclogite-facies conditions. Relicts of wall-rock garnet (Grt_I) cores in recrystallized selvage garnet (Grt_II) close to the wall rock and the Ca distribution in Grt_II, which images the overgrown selvage matrix, indicate that the

5. Metal transport in subduction fluids selvage formed due to dissolution-precipitation processes as a consequence of fluid-rock interaction of the wall rock eclogite with the vein-forming fluid. The peak metamorphic P–T conditions of the wall-rock eclogite are estimated at ca. 590 ºC and 23 kbar. Mass-balance calculations indicate that the reaction selvage experienced: (1) a depletion of the large-ion lithophile elements (K–Rb–Ba) of up to 100% relative to their concentrations in the wall-rock eclogite; (2) a moderate depletion of the HREE and some transition metal elements including Fe, Cu, Ni, Zn, Co, Cr, and Mn (10%–40%); (3) a significant enrichment of CaO (up to 50%–80%), Sr (up to ~200%), Pb (up to ~170%) and S (up to ~210%); (4) a slight to moderate enrichment of the LREE (10– 20%) and MREE (0%–40%); whereas (5) the HFSE show no significant variations.

The chemical changes in the selvage suggest that the fluid, which caused the dissolution of the wall-rock and the precipitation of the selvage assemblage while the vein formed, was probably a mixture of an ―internal‖ fluid derived from the prograde dehydration (e.g., lawsonite breakdown) of the wall rock and an ―external‖ fluid most likely derived from dehydrating altered oceanic crust located in stratigraphically lower units of the subducting slab. The external fluid introduced Ca, S, Sr, Pb and at least parts of the LREE and the MREE into the selvage, whereas some elements, such as the remaining LREE and MREE, may have been derived from the wall rock eclogite via diffusional transport into the selvage. The enrichment of Ca and L-MREE is displayed by the abundant growth of selvage epidote. In contrast, the dissolution of garnet and phengite released significant amounts of HREE and LILE (K–Rb–Ba) into the passing fluid, because the chemical changes within the selvage prevented the formation of a mineral assemblage with sufficiently high bulk-fluid partition coefficients for these elements. Significant amounts of transitional metal elements were released into the fluid during the dissolution of white mica and the dissolution-precipitation behavior of garnet, omphacite, dolomite and sulfides. Thus the LILE and HREE along with some transition metal elements (e.g., Fe, Cu, Ni and Zn) were mobilized during the dissolution-precipitation processes that led to the selvage formation. Accordingly the slab fluids are not only strongly enriched in LILE and depleted in HFSE, but also carry

5. Metal transport in subduction fluids significant amounts of transition metals. It is most likely that slab fluids strongly contribute to the metal flux into the arc magma systems finally resulting in giant arc-related ore deposits.

5.2 Introduction

Subduction zones, which constitute the largest recycling system on our planet, play a key role in Earth‘s geodynamics, crustal evolution, as well as earthquakes and volcanism. Fluids may have played the most critical role in the above-mentioned geodynamic processes (e.g., Austrheim, 1987; McCulloch and Gamble, 1991; Labrousse et al., 2010; Richards, 2011). Subduction zone processes also control the formation of various types of metal ore-deposits (e.g., Sawkins, 1972; Glasby, 1996; Kerrich et al., 2005). Subduction zone-derived magmas are usually enriched in large ion lithophile elements (LILE) and depleted in high field strength elements (HFSE). These geochemical signatures are believed to be the result of metasomatized mantle wedge melting. The metasomatism is interpreted to have been caused by infiltrating hydrous fluids that stem from the dehydration of subducted oceanic slabs (Gill, 1981; Tatsumi, 1989; McCulloch and Gamble, 1991). This hypothesis was subsequently confirmed by both the systematic geochemical data obtained from arc igneous rocks (e.g., Ryan et al., 1995; Stolz et al., 1996; Plank and Langmuir, 1993) and high-pressure rocks from former subduction-zones (Scambelluri and Philippot, 2001; John et al., 2004; Hermann et al., 2006; Beinlich et al., 2010; Konrad–Schmolke et al., 2011).

In contrast, no consensus has been reached with regards to the source of the giant metal deposits formed in arc settings. The bulk of the metal flux into the arc magma sources is thought to have been predominantly derived from the partial melting of metasomatized asthenospheric mantle triggered by slab fluids (e.g., Pettke et al., 2010; Richards, 2011), whereby some metals may have been transferred from the subducting slab via dehydration fluids to economic arc-associated ore deposits (e.g., Hedenquist and Lowenstern, 1994). It is further suggested that Fe, Cu, Pb, Zn, Sn and Mo of the South American magmatic arc-related ore deposits were derived from the descending slab and

5. Metal transport in subduction fluids the overlying mantle wedge (Sillitoe, 1972). Additionally, Cu may partition into the slab-derived fluids which then enrich the magma source ultimately producing Cu deposits at convergent margins (Stolper and Newman, 1994; de Hoog et al., 2001a). Also S, Pb, As, and Sb may have been transferred into arc magma sources by fluids derived from the down-going slab (Alt et al., 1993; Noll et al., 1996; Metrich et al., 1999; de Hoog et al., 2001b). A model was proposed for the ore metal recycling process during which subduction-derived fluids transport soluble metals from the slab into the mantle wedge and thus are ultimately responsible for the formation of porphyry copper and epithermal Au deposits in arc magmas (McInnes et al., 1999). Nevertheless, the above hypotheses concerning the mobilization of ore-producing metals in subduction-zone fluids have not been verified by the direct study of fossilized subduction-zone derived HP-rocks and associated intra-slab fluid-flow structures such as HP-veins.

Continuous dehydration of subducted oceanic crust takes place from 10–20 km to about 70–300 km (e.g., Schmidt and Poli, 1998), whereas the transition from ―wet‖ blueschist into ―dry‖ eclogite at a depth of ca. 70 km is thought to be the most important subduction-zone fluid releasing process with respect to the oceanic crust (Peacock, 1993). The subduction-zone derived low salinity aqueous fluid is thought to contain volatiles and major element components including Si, Al, Na and Ca (Manning, 2004; Gao et al., 2007). The trace element budget transported by these fluids is believed to be quite complex (cf., Klemd, 2013). Some studies suggest a decoupling of water and trace element release during the dehydration of oceanic crust (e.g., Hermann et al., 2006; Scambelluri and Philippot, 2001; Spandler et al., 2004), whereas others are more focused on the reactivity of the released aqueous fluids which may allow these fluids to mobilize and transport certain trace elements while they are traveling through the slab towards the mantle wedge (e.g., Manning, 2004; John et al., 2004; Beinlich et al., 2010; Guo et al., 2012; Herms et al., 2012; John et al., 2012). It is reported that subduction-zone fluids may be responsible for the mobilization of considerable amounts of LILE, U and light rare earth elements (LREE) (McCulloch and Gamble, 1991;

5. Metal transport in subduction fluids

Hawkesworth et al., 1993; Scambelluri and Philippot, 2001; John et al., 2004; Manning, 2004). In some cases, it is also shown that they can mobilize and transport heavy rare earth elements (HREE) (Spandler and Hermann, 2006; John et al., 2008; Guo et al., 2012) and even HFSE which are usually believed to behave rather immobile in aqueous fluids (e.g., Rubatto and Hermann, 2003; Xiao et al., 2006; Gao et al., 2007; Zhang et al., 2008; Rapp et al., 2010). Compared with the state-of-the-art concepts listed above, little is known about the mobilization behavior of transition metals such as Au, Fe, Cu, Ni and Zn during fluid-rock interaction in subduction zones even though they often form huge ore deposits associated with arc settings (e.g., Richards, 2011).

High-pressure vein networks in eclogite-facies rocks are interpreted to represent fossilized fluid pathways within the subducted slabs and thus are thought to record the fluid-rock interaction and related element behavior during fluid flow (e.g., Becker et al., 1999; Gao and Klemd, 2001; Widmer and Thompson, 2001; John and Schenk, 2003; Spandler and Hermann, 2006; John et al., 2008). Although pyrite was reported to occur as daughter mineral entrapped in fluid inclusions in vein omphacite (Philippot and Selverstone, 1991) and as matrix mineral coexisting with vein omphacite and the immediate high-pressure host rocks (e.g., Gao and Klemd, 2001; Spandler and Hermann, 2006; Zhang et al., 2008), no detailed investigation has been undertaken concerning the effect of subduction-zone fluid flow on the first series of transition metal elements (labeled as transition metal elements below). We report petrological and geochemical data from a sulfide-bearing high-pressure vein, its selvage and host eclogites (wall rock eclogite) from the Tianshan (ultra-)high-pressure/low-temperature ((U)HP/LT) belt in order to discuss the mobilization of both transition metals and trace elements in subduction-zone fluids.

5.3 Geological context

The southern Tianshan Orogen, which extends west-east for about 2500 km from Uzbekistan, Tajikistan, Kyrgyzstan, Kazakhstan to Xinjiang in northwestern China, is situated along the southwest margin of the Central Asian Orogenic Belt (e.g., Jahn et al.,

5. Metal transport in subduction fluids

2000) and marks the final collision between the Tarim and Yili blocks (Gao et al., 2009; Xiao et al., 2009; Han et al., 2011; Long et al., 2011). A (U)HP/LT metamorphic belt, extending for 1500 km from Kekesu, Akeyazi and Changawuzi in NW China, Atbashi of Kirghizia to Fan Karategin of Tajikistan in the south Tianshan Orogen, represents a subduction/collision zone in the South Tianshan Orogen (e.g., Gao et al., 1995, 2009, 2011; Hegner et al., 2010; Volkova and Budanov, 1999). In the Chinese Tianshan, the (U)HP/LT belt extends for at least 200 km along the South Central Tianshan suture and consists of a suite of metasedimentary, mafic and ultramafic rocks (Fig. 5-1a) that were interpreted as a paleo-accretionary wedge. It is mainly composed of blueschist-, eclogite- and greenschist-facies meta-sedimentary rocks and some mafic metavolcanic rocks with N-MORB, E-MORB, OIB and arc basalt affinities (Gao and Klemd, 2003; John et al., 2008). Blueschists occur within greenschist-facies metasediments as small discrete bodies, lenses, bands and thick layers. Eclogites are interlayered with the blueschist or mica schist layers as pods, boudins, thin layers or as massive blocks interpreted to represent a tectonic mélange (Gao et al., 1999; Gao and Klemd, 2003; van der Straaten et al., 2008; Wei et al., 2009).

Most eclogites have experienced peak metamorphism estimated to range between 480 and 580 °C at 14–21 kbar at a regional scale (e.g., Klemd et al., 2002; Wei et al., 2003). Ultrahigh-pressure (UHP) conditions (confirmed by the presence of coesite) of 27–33 kbar at 570–630 °C have been reported for eclogitic mica schists and 24–27 kbar at 470–510 °C for eclogites from several localities (Lü et al., 2008, 2009). These P–T estimates are in accordance with petrological studies on meta-sedimentary rocks claiming that the Tianshan (U)HP/LT belt is predominantly composed of HP pelitic-felsic schists with minor UHP pelitic-felsic slices or blocks (Wei et al., 2009). The intimate interlayering of high- and ultrahigh-pressure rocks suggests that the rocks were derived from varying depths from the descending slab and then juxtaposed during exhumation in the subduction channel (Klemd et al., 2011). Multi-point Lu–Hf isochron age dating (omphacite-garnet-whole rock) of eclogites from the Chinese Tianshan suggests an age of ca. 315 Ma for the peak of eclogite-facies metamorphism (Klemd et

5. Metal transport in subduction fluids al., 2011). Vein formation occurred contemporaneous to peak metamorphism at about 317 ± 5 Ma as inferred from Rb–Sr vein-whole rock data (John et al., 2012). White mica Ar–Ar and Rb–Sr ages cluster at 310 Ma, which is interpreted to date exhumation of the high-pressure rocks (Klemd et al., 2005; Wang et al., 2011).

Figure 5-1 Geological map of the western Tianshan (U)HP/LT metamorphic belt in northwestern China, modified after Li et al. (2012). (a): Regional tectonic map; (b): Local geological map showing the area from which samples were taken.

The ubiquitous presence of high-pressure vein networks in blueschists and eclogites generated during prograde and retrograde metamorphism reveals extensive fluid-rock interaction activities and fluid-mediated mass transport in a former oceanic subduction zone (Gao and Klemd, 2001; Gao et al., 2007; John et al., 2008, 2012; van der Straaten et al., 2008; Beinlich et al., 2010).

5. Metal transport in subduction fluids

5.4 Sample description and petrography

A striking feature of the vein and its host rock reported here is that both of them contain considerable amounts of sulfides (predominantly pyrite and minor chalcopyrite). This sample (L082-5) stems from the upper area of the Atantayi River (42°29′43″, 81°16′11″), which is a branch of the Kepuerte River (Fig. 5-1b). The sulfide-bearing samples occur as loose, meter-sized blocks, which were derived from the adjacent steep mountain slopes (Fig. 5-2a). The in situ outcrops are either located in adjacent steep walls or are covered by glaciers.

Figure 5-2 Outcrop photograph of the sampled eclogite boulder and the relevant vein (a) and the hand specimen used in this study (b). (a) The vein with the selvage cuts through host eclogite. (b) A vein consisting of pyrite, omphacite, epidote and quartz is enveloped by a pyrite-bearing eclogitic selvage in host eclogite. (c) A sketch map shows the whole rock samples (separated by blue lines) taken for chemical composition analysis along a profile with varying distance to the vein; a: the wall-rock eclogite; b, c, d, e: the selvage; f: the vein rock.

5.4.1 Sample description

The 3~8 cm wide sulfide and omphacite-bearing vein crosscutting massive eclogite (Fig. 5-2a) extends almost perpendicular to the host rock foliation. The interface between the vein and the immediate wall rock is characterized by the occurrence of a ca. 6 cm wide reaction selvage. The wall rock-selvage contact appears to be corrosive while the boundary between the selvage and the vein appears to be sharp in the hand-specimen (Fig. 5-2b) but is jagged on the microscopic scale (Fig. 5-3g). Fibrous omphacite grew almost perpendicular to the vein wall and is intergrown with the fine-grained omphacite

5. Metal transport in subduction fluids of the selvage. The modal volume and grain size of the sub- to idioblastic sulfides in the selvage gradually increase over a distance of several centimeters towards the vein (Fig. 5-2b).

5.4.2 The wall-rock eclogite (WE)

The wall-rock eclogite consists of garnet (average modal content: 16 vol.%), omphacite (45 vol.%), dolomite (13 vol.%), white mica (13 vol.%), epidote (5 vol.%), glaucophane (3 vol.%), sulfide (2 vol.%), quartz (1 vol.%), minor lawsonite and accessory minerals such as rutile, titanite and apatite (Table 5-1). Sulfide, dominantly pyrite, occurs as fine-grained inclusion in garnet and as subidioblastic, medium-grained matrix mineral (Fig. 5-3a and c). The large matrix pyrite grains contain omphacite, dolomite and rutile inclusions (Fig. 5-3d). Chalcopyrite usually occurs as small-sized particles on the edges of pyrite grains (Fig. 5-3d). Rarely glaucophane forms inclusions in garnet cores (Fig. 5-3b) whereas some glaucophane grains are randomly distributed in the matrix. In places, glaucophane and phengite define a weak foliation of the WE. Porphyroblastic garnet (1.5–4 mm in diameter) contains blueschist-facies and eclogite-facies mineral inclusions. The blueschist-facies inclusion assemblage includes barroisite/glaucophane, epidote, paragonite, dolomite, quartz and albite, all of which are found in the core domain of garnet porphyroblasts (Figs. 5-3b and 5-4a). In contrast, the eclogite-facies inclusion assemblage, which includes omphacite, dolomite/magnesite and rutile, predominantly occurs in the mantle and rim domains of the garnet porphyroblasts (Figs. 5-3b and 5-4a). In addition, rare zircon grains (<50 µm) occur as inclusion in the rim domain (Fig. 5-3b). Occasionally, lawsonite and magnesite were found to be enclosed by dolomite inclusions in garnet porphyroblasts (Fig. 5-4) and in matrix dolomite. Omphacite is the dominant matrix mineral of the WE. Orientated tubular two-phase fluid inclusions occasionally occur in the core of the omphacite (Fig. 5-3d). Fibrous omphacite is orientated parallel to the weak foliation as defined by white mica (Fig. 5-3d). Furthermore, some glaucophane inclusions were detected in the omphacite cores. Fine-grained rutile grains occur as prolonged aggregations which are often orientated

5. Metal transport in subduction fluids parallel to the weak foliation of the WE (Fig. 5-3d).

5. Metal transport in subduction fluids

◀Figure 5-3 Representative photomicrographs of the host eclogite, the selvage and the vein; (a) Garnet porphyroblast in the host eclogite contains omphacite, epidote, apatite, dolomite and pyrite inclusions; Dol=dolomite, Ap=apatite, Ep=epidote, Grt=garnet, Omp=omphacite, Phn=phengite, Py=pyrite, Rt=rutile, Ttn=titanite, Zr=zircon. (b) The blueschist-facies assemblage consisting of barroisite/glaucophane, paragonite and albite occurs in the core of the idiomophic garnet porphyroblast in the WE, and the eclogite-facies assemblage consisting of omphacite, dolomite and rutile occurs in the mantle and rutile inclusions in the rim; Gln=glaucophane, Bar= barroisite, Ab=albite, Pg=paragonite. (c) Pyrite occurs as inclusion in garnets and fine-grained matrix constituent in the WE. (d) The matrix pyrite in the WE contains omphacite, dolomite and rutile inclusions and its long axis is orientated parallel to the foliation as defined by lepidoblastic phengites and fibrous omphacites; Cpy=chalcopyrite. (e) Fine-grained garnet, chalcopyrite and omphacite are intergrown with the lawsonite-bearing pyrite porphyroblast in the selvage; Law=lawsonite. (f) Orientated primary two-phase tubular fluid inclusions occur in the core of both omphacite and epidote. (g) The fibrous omphacite in the vein-wall is solid-inclusion free and almost perpendicular to the jagged boundary between the vein and the selvage; Qtz=quartz. (h) In general, minerals located in the vein center distribute parallel to the vein. Pyrites in the vein occur as aggregates in places and contain omphacite inclusions. Rare garnet occurs in the vein.

5.4.3 The reaction selvage (RS)

The eclogite-facies reaction selvage is composed of garnet, omphacite, white mica, epidote, dolomite, sulfides (mostly pyrite), rutile/titanite and apatite (Table 5-1). The modal abundance and the grain size of most minerals display relatively regular variations with decreasing distance towards the vein (Table 5-1). For instance, the abundance of garnet and white mica decreases drastically towards the vein (Table 5-1), from 10 vol.% to 0.5 vol.% garnet, and 5 to almost 0 vol.% white mica, while the omphacite abundance decreases from 55 to 45 vol.% (and rises back to 55% closest to

Table 5-1 Modal abundances of minerals in vol.% *

*mineral modes were determined from the thin section by point counting on the basis of petrographic observations. The uncertainty of modal abundances of minerals was estimated to be less than 10% by repeating the same operations. DfV= distance from vein

5. Metal transport in subduction fluids

Figure 5-4 Back scattered electron (BSE) images of a garnet prophyroblast and a relictic lawsonite inclusion in host eclogites. (a) Garnet contains omphacite, quartz and epidote + paragonite (pseudomorph after lawsonite) inclusions in the core and omphacite + dolomite inclusions in the mantle. (b) Enlarged area marked by the rectangle in (a) displays dolomite inclusions. (c) Close up of the rectangle area in (b) shows lawsonite and magnesite enclosed by dolomite inclusions. Mag= magnesite. (d) Representative laser Raman spectra of lawsonite are at 562 cm-1, 695 cm-1 and 937 cm-1. the vein). The modal content of epidote increases from 10 to 25 vol.% in the selvage. The size of garnet grains decreases (0.2–0.8 mm in diameter) in comparison with the host garnet (1.5–4 mm). In contrast, the abundance of pyrite and its grain size increases from 2 to 6 vol.% and 1 mm to nearly 10 mm, respectively, towards the vein (Fig. 5-2b). Whereas the apatite mode changes only moderately (from ~0.5 to 2 vol.%) towards the vein. The abundance of both rutile and titanite abruptly decreases from 2 vol.% and 0.5 vol.%, respectively, to almost 0 vol.% close to the vein (Table 5-1). Few fine-grained, xenoblastic chalcopyrite grains are intergrown with pyrite and the other eclogite-facies minerals (Fig. 5-3e).

5. Metal transport in subduction fluids

Porphyroblastic pyrite contains several mineral inclusions such as chalcopyrite, garnet, omphacite, lawsonite, epidote/clinozoisite, phengite/paragonite, quartz, glaucophane, barroisite, chlorite, albite, quartz, dolomite, rutile and titanite (Fig. 5-3e and Fig. 5-5a–f). Lawsonite, identified by Raman spectrometry (Fig. 5-5g), occurs as inclusions in pyrite and coexists with omphacite (Fig. 5-5b) and epidote-paragonite- quartz assemblages (Fig. 5-5b and f) or chlorite. Occasionally, lawsonite has titanite and quartz inclusions (Fig. 5-5c and f). Garnet inclusions in pyrite are ca. 0.5 mm in diameter and are rarely enveloped by a thin rim of albite, glaucophane, barroisite and chlorite (Fig. 5-5d and e). Omphacite grows perpendicular to the selvage-vein boundary and rarely contains fine-grained (<10 µm) zircon inclusions. Primary two-phase tubular fluid inclusions occur in the core of both omphacite and epidote with the orientation parallel to the X-axis of the host mineral (Fig. 5-3f). Epidote porphyroblasts contain omphacite, dolomite, pyrite, phengite, rutile and zircon inclusions. Minor fibrous phengite occurs in the matrix or as inclusion in pyrite and epidote. Rutile also occurs as fine-grained inclusions in garnet, epidote and pyrite, or as elongated aggregates, parallel to the foliation (Fig. 5-3d).

5.4.4 The vein (V)

The vein has a relatively heterogeneous mineral modal content. For instance, dolomite and pyrite often are locally concentrated (Figs. 5-2 and 5-3h); epidote usually occurs as coarse grains in the middle (Fig. 5-2b) and apatite along the edge of the vein (Fig. 5-3g). The mineral assemblage of the investigated part of the vein (Fig. 5-2) is given in Table 5-1. Omphacite fibers (up to 1 cm in length) and quartz, which constitute ca. 60 vol.% of the vein minerals, are homogenously distributed in the vein. Fibrous, randomly-oriented omphacite is usually free of solid inclusions (Fig. 5-3g); however, the core occasionally contains two-phase tubular aqueous fluid inclusions. Pyrite and minor chalcopyrite occur as concentrated aggregates heterogeneously distributed in the vein (Fig. 5-3h). They are intimately intergrown with omphacite, quartz, epidote and apatite and frequently contain omphacite inclusions (Fig. 5-3h). Dolomite

5. Metal transport in subduction fluids porphyroblasts have a grain size of between 1 and 3 mm. Minor, fine-grained glaucophane is associated with omphacite, pyrite and epidote. Garnet rarely occurs (<0.1 vol.%) in the vein and the grain size (0.2–0.6 mm in diameter) is much smaller than that of the garnet in the wall-rock eclogite (Fig. 5-3h). Zircon grains were not observed in the vein and vein minerals.

5.5 Analytical methods

The bulk rock major element compositions were determined by X-ray fluorescence spectrometry (XRF) on fused glass disks, using a Rigaku100e at the Guangzhou Institute of Geochemistry, Chinese Academy of Sciences (GIGCAS) in Guangzhou, China. The loss on ignition (LOI) was determined prior to major element analyses using a pre-ignition method after heating the samples to 1000 °C. The analyses of rock standard reference materials (GSR-1, GSR-2, and GSR-3) indicated that analytical uncertainties for most major oxides are <2%, for MnO and P2O5 are <5%, and the totals are 100 ± 1% (Table 5-2).

The bulk rock trace element compositions were analyzed by inductively coupled plasma mass spectrometry (ICP–MS, PE Elan 6000) at the same Institute of the GIGCAS in Guangzhou. Whole rock powders were digested in a screw-top Teflon beaker by

HF/HNO3/HClO4 (2:2:1) mixture at 100 °C for 3 days, and evaporated off at 140 °C.

The residue was then digested with a HF/HNO3/HClO4 (2:2:1) mixture in a screw-top PTFE-lined stainless steel bomb at 190 °C for 48 hours, and the evaporated residue was dissolved using 4ml HNO3 (50%, volume ratio) in a screw-top PTFE-lined stainless steel bomb at 170 °C for another 4 hours. The sample solution was then diluted with

HNO3 (3%, volume ratio) to a sample/solution weight ratio of 1/2000. Rh (10 ppb) standard solution was added as an internal standard at a weight ratio of 1:1, while rock reference standards (BHVO-2, AGV-2, W-2, GSR-2 and GSR-3) were chosen as external calibration standards for calculating the element concentrations in the measured samples. The analyses for the international rock standards BHVO-2 (Hawaiian Basalt, USGS), AGV-2 (Andesite, USGS), GSR-2 (andesite) and GSR-3 (basalt) are given in

5. Metal transport in subduction fluids

Figure 5-5 BSE images of mineral inclusions in pyrite in the selvage. (a) Separate lawsonite, omphacite and dolomite inclusions scatter in pyrite porphyroblasts. The rectangle shows the location of (b). (b) Lawsonite is intergrown with omphacite and epidote in pyrite. (c) Lawsonite contains titanite + quartz and titanite + rutile + quartz inclusions. (d) Both blueschist-facies and eclogite-facies mineral assemblages are present as inclusions. Garnet occurs as inclusion in pyrite and matrix component; Cal=calcite. (e) Garnet inclusion in pyrite enveloped by the chlorite + glaucophane + barroisite + omphacite + albite assemblage; Chl=chlorite. (f) The intergrowth of lawsonite, epidote, quartz and paragonite assemblages suggests the breakdown reaction of lawsonite: 4 Law + Jd = 2 Ep + Pg + Qtz + 6 H2O. (g) Representative laser Raman spectra of lawsonites in pyrite are at 562 cm-1, 694 cm-1 and 939 cm-1.

5. Metal transport in subduction fluids

Table 5-2. The total procedure blanks were processed in the same way and relative standard deviations (RSD) are within ± 5% for most trace elements.

Whole rock powder samples for Sr and Nd isotopic analyses were dissolved in Teﬂon bombs after being spiked with 87Rb, 84Sr, 147Sm and 150Nd tracers prior to

HF+HNO3+HClO4 dissolution. Rb–Sr, Sm–Nd were separated using conventional ion exchange procedures and measured using a Finnigan MAT262 multi-collector mass spectrometer at Institute of Geology and Geophysics, Chinese Academy of Sciences (IGGCAS) in Beijing, China. Detailed descriptions of the analytical techniques are given in Chu et al. (2009). Measured 87Sr/86Sr and 143Nd/144Nd ratios were corrected for mass-fractionation using 86Sr/88Sr=0.1194 and 146Nd/144Nd=0.7219, respectively. Standard NBS-987 Sr and JNdi-1 Nd were measured at 87Sr/86Sr=0.710263±10 (2σ) and 143Nd/144Nd=0.512096±14 (2σ), respectively. USGS reference material BCR-2 was measured to monitor the accuracy of the analytical procedures, with the results of: 45.79 ppm Rb, 325.5 ppm Sr and 87Sr/86Sr=0.701051±11 (2σ); 6.196 ppm Sm, 27.14 ppm Nd and 143Nd/144Nd=0.512887±13 (2σ). These values are comparable with those reported in Chu et al. (2009) as well as with internationally accepted reference values (GeoREM, http://georem.mpch-mainz.gwdg.de).

In situ analyses by laser ablation inductively-coupled plasma mass spectrometry (LA– ICP–MS) were made on thin sections in the GeoZentrum Nordbayern (GZN) of the University Erlangen–Nürnberg, Erlangen, Germany. The analyses were performed with a single collector quadrupole Agilent 7500i ICP–MS equipped with an UP193Fx Argon Fluoride Fast New Wave Research Excimer laser ablation system. For sulfides, Po724

B2 SRM (standard for Au and PGEs, Memorial University Newfoundland), (Fe, Ni)1-xS (standard for Ni and Re, University Münster; Wohlgemuth-Ueberwasser et al., 2007) and MASS-1 (Polymetal sulfide for V, Mn, Co, Cu, Zn, As, Se, Mo, Ag, Sb, Te, W, Pb and Bi, USGS) were used as reference materials for external calibration. Reproducibility for SRM is <7% and for MASS-1 <10%, and the accuracy for PGEs was tested by ablating the SRM FeNiS standard (<20%). LA–ICP–MS measurements were conducted using a spot size of 50 µm in diameter, a laser frequency of 15 Hz and

5. Metal transport in subduction fluids

0.32 GW/cm2 and a fluence of 1.63 J/cm2. For smaller grains a 25 µm spot was used. The carrier gas consists of a mixture of 0.65 l/min helium and 1.06 l/min argon. Acquisition time was 20 s for the background and 20 s for the mineral analysis. Signal quantiﬁcation was carried out by GLITTER using sulfur as internal standard determined by means of the electron microprobe. For more analytical details concerning the sulfides see Osbahr et al. (2012). The glass reference material NIST SRM 612 was used as standard for the silicates for external calibration. LA–ICP–MS measurements were conducted using a spot size of 25 µm in diameter, a laser frequency of 15 Hz and 0.50 GW/cm2 and a fluence of 2.48 J/cm2. The carrier gas consists of a mixture of 0.65 l/min helium and 1.06 l/min argon. Acquisition time was 20 s for the background and 25 s for the mineral analysis. The internal standards such as Si for the silicates, Ca for the Ca-phases and Ti for the Ti-phases were determined by EMP analysis. Reproducibility and accuracy, which were determined for NIST SRM 610, are usually <8% and <6%. The trace element concentrations were calculated by GLITTER Version 3.0 (van Achterbergh et al., 2000). For more analytical details concerning the silicate analyses see Schulz et al. (2006).

Major element compositions of minerals were obtained by electron microprobe analysis at the IGGCAS (JEOL JXA 8100) and the GZN (JEOL JXA 8200). The quantitative analyses were performed using wavelength dispersive spectrometers with an acceleration voltage of 15 kV, a beam current of 15 nA, a 3 μm beam size and 10–30s counting time. Natural minerals and synthetic oxides were used as standards, and a program based on the ZAF procedure was used for data correction. X-ray maps for selected elements in garnet, dolomite and epidote were conducted in WDS mode at GZN with an acceleration voltage of 15 kV, a beam current of 180–220 nA, a 2–8 μm pixel size and dwell time of 100 ms.

To accurately identify the lawsonite inclusions in pyrite and garnet, Laser-Raman spectroscopy was performed at IGGCAS using a Renishaw Raman MKI-1000 system equipped with a CCD detector and an Ar ion laser. The laser beam with a wavelength of 514.5 nm was focused on the lawsonite inclusions through 50× and 100× objectives of a

5. Metal transport in subduction fluids light microscope. The laser spot size was focused to 1 µm. The reproduction of spectra for the same spot is better than 0.2 cm-1.

5.6 Results

5.6.1 Bulk geochemistry

Bulk-rock major and trace element compositions from the host eclogite, the selvage and the vein of the most homogeneous sample (L082-5) are presented in Table 5-2.

5.6.1.1 The wall-rock eclogite

The wall-rock eclogite has the largest distance to the vein (L082-5a) and is thus believed to represent the least altered sample. The SiO2 content of the WE sample is 38.48 wt.% that is somewhat lower than that of a typical basalt. The CaO (12.43 wt.%) and Li (22.6 ppm) concentrations of the WE are much higher than that of unaltered MORB and ocean island basalts (OIB), indicating that the precursor basalt underwent significant low-temperature weathering and/or hydrothermal alteration the latter of which preceding subduction (Bouman et al., 2004). Seafloor alteration likely caused the increase in carbonate abundance (dolomite up to 14 vol.%, Table 5-1) and the decrease of the SiO2 content of the WE. The Na2O content is 2.58 wt.%, the K2O content 2.25 wt.% and the TiO2 content 1.42 wt.% (Table 5-2). In agreement with the high carbonate content, the WE also has a high LOI value of 9.11 wt.%. The chondrite-normalized rare earth elements (REE) pattern (Fig. 5-6a) indicates an enriched mantle source for the protolith of the WE, with a total LREE concentrations of ~70 times chondrite, a

(La/Sm)N value of 2.5, a (La/Yb)N value of 5.1 and a Nb/La ratio of ~1.02. The

(La/Sm)N vs. Nb/La relationship suggests an OIB affinity for the precursor rocks (John et al., 2003). The MORB-normalized trace-element pattern shows a scattered distribution (Fig. 5-6b). The LILE (Rb, Ba, K, Pb, Sr and Li) display prominent positive anomalies while the HFSE (Ti, Nb, Ta, Zr, Hf) and P only show minor negative anomalies. The Nb/Ta ratio of the WE is ~14 and Zr/Hf ~42, and are thus within the

100

5. Metal transport in subduction fluids accepted range for oceanic basalts (e.g., Pfänder et al., 2007). Furthermore, the wall-rock eclogite also contains considerable contents of some transition metals, e.g., 179 ppm V, 660 ppm Cr, 308 ppm Ni, 122 ppm Cu and 99.8 ppm Zn. The concentrations of these transition metals are similar to those of MORBs (Sun et al., 1979) and OIBs (Dupuy et al., 1988).

5.6.1.2 The reaction selvage

Four samples were taken for a detailed geochemical study from various distances to the vein (Fig. 5-2c). According to their decreasing distance to the vein the investigated samples are L082-5b: 6.2 cm to 4.3 cm away, L082-5c: 4.3 cm to 2.4 cm; L082-5d: 2.4 cm to ca. 1 cm; L082-5e: ca. 1cm to 0 cm. The bulk chemical data from the selvage exhibit highly variable SiO2 contents (L082-5b= 39.49 wt.%, L082-5c = 35.77 wt.%, L082-5d= 33.33 wt.%, L082-5e= 38.47 wt.%; Table 5-2), which are believed to be due to the high but variable modal abundance of dolomite in the selvage (Table 5-1). Most samples have similar TiO2 contents between 1.47 wt.% (L082-5b) and 1.22 wt.% (L082-5d) but with a much lower content 0.68 wt. % (L082-5e) close to the vein and a relatively constant Na2O content (average 2.4±0.4 wt.%). The K2O content gradually decreases towards the vein from 2.17 wt.% (L082-5b) to 0.02 wt.% (L082-5e) and CaO increases from 13.43 wt.% (L082-5b) to 19.24 wt.% (L082-5e). The chondrite-normalized REE patterns of the selvage samples have negative slopes displaying LREE enrichment compared to the HREE, similar to that of the wall-rock eclogite (Fig. 5-6a). The selvage is more depleted in HREE relative to those in the WE and has a (La/Sm)N value of 2.1–2.4 and a (La/Yb)N value varying between 6.2–10.2. The MORB-normalized trace-element patterns show negative LILE (Rb, Ba, K) anomalies for selvage samples c–e (Fig. 5-6b). The four samples along the profile crosscut the selvage (Fig. 5-2c) and display regular compositional variations such as an increase of CaO, S, Sr, Pb and a decrease of LILE (Rb, Ba, and K) and HREE (Tm, Yb, and Lu) concentrations towards the vein (Fig. 5-6c and h). The HFSE concentrations (Ti, Nb, Ta, Zr and Hf) of most selvage samples are similar to those of the WE, whereas they decease abruptly nearest to the vein (Fig. 5-6d). The Nb/Ta ratio of the selvage is

101

5. Metal transport in subduction fluids

15.1±0.2 whereas the Zr/Hf ratio varies from 37 to 46. The concentration of some transition metals (Fe, Cu, Ni and Zn) of the selvage are somewhat lower than those of the WE, e.g., average 11.33 wt.% vs. 13.55 wt.% FeO*, 79.1 ppm vs. 122 ppm Cu, 232 ppm vs. 308 ppm Ni and 84.7 ppm vs. 99.8 ppm Zn (Table 5-2). Although the variation

Table 5-2 Bulk geochemical data, major and trace elements, of the wall-rock eclogite (WE), the reaction selvage (RS) and the vein (V). The major elements are given in wt.%, and the trace elements in ppm

*Total iron; n.d.: not determined.

102

5. Metal transport in subduction fluids

Figure 5-6 Chondrite-normalized rare earth element patterns (a) and MORB-normalized trace element spider diagrams (b) of the wall-rock eclogite, the selvage and the vein. Normalization values are after Sun and McDonough (1989). (c–h) Chemical compositions along a profile in the selvage with variable distances towards the vein, (c) LILE + CaO (wt.%) + S (wt.%); (d) HFSE; (e) Transition metal elements; (f) LREE; (g) MREE; (h) HREE. in Cu are somewhat ambiguous due to the heterogeneous distribution of chalcopyrite, a continuous decrease in concentration of the other transition metals Mn, Co, Ni and Zn is apparent (Fig. 5-6e).

5.6.1.3 The vein

The SiO2 content of the vein is 66.54 wt.%, higher than that of all other samples from

* the WE and the selvage. However, Al2O3, MgO, CaO, FeO , K2O and LOI of the vein are significantly lower than those of the WE and the selvage (Table 5-2). The chondrite-normalized REE pattern of the vein, which is about one order of magnitude lower than that of the WE and the selvage (Fig. 5-6a), shows a negative slope with a

103

5. Metal transport in subduction fluids

(La/Sm)N value of 2.4 and a (La/Yb)N value of 12.5. In fact most trace elements have concentrations about one order of magnitude lower than those of the WE and the selvage (Fig. 5-6b) except the LILE, Th, P and Li. The MORB-normalized trace-element pattern of the vein shows that both Pb and Li have ~10 times higher concentrations than the average MORB, whereas the concentrations of K, Rb and Ba are less (0.2–0.8 times) than those of the average MORB. Significant negative anomalies of HFSE are also displayed. The concentrations of the transition metals are somewhat lower than those of the WE and the selvage, e.g., 7.12 wt.% vs. 13.55 wt.% and 11.33 wt.% FeO*, 46.2 ppm vs. 122 ppm and 79.1 ppm Cu, 91.5 ppm vs. 308 ppm and 232 ppm Ni (Table 5-2).

5.6.1.4 Sr–Nd isotopes

The Sr and Nd isotopic compositions of the WE, the selvage and the vein are presented in Table 5-3. The initial Sr and Nd isotopic values were calculated for an age of 315 Ma, which has been interpreted to date the peak of metamorphism (e.g., Klemd et al., 2011). The initial 87Sr/86Sr value of the vein is 0.706527 which is similar to that of the WE and the selvage, 0.706303 and 0.706827 respectively. The εNd(t) value of the vein (0.29±0.24) is identical within error to that of the WE and the selvage (0.75±0.24 and 0.39±0.24, respectively, Table 5-3).

Table 5-3 Sr-Nd isotopic compositions of the wall-rock eclogite, reaction selvage and vein. 87 86 Age Rb Sr 87 86 Sr/ Sr 87 86 Sample Loc. Rb/ Sr ( Sr/ Sr)I (Ma) (ppm) (ppm) 2σ L082-5a WE 315 54.7 308 0.513865 0.708606±9 0.706303 L082-5c RS 315 8.86 795 0.032233 0.706971±11 0.706827 L082-5f V 315 0.197 214 0.002665 0.706539±10 0.706527 147 144 143 144 143 144 Sm Nd Sm/ Nd Nd/ Nd ( Nd/ Nd)I εNd(t) L082-5a WE 3.76 16.6 0.136772 0.512553±12 0.512271 0.75±0.24 L082-5c RS 4.83 19.1 0.153577 0.512569±12 0.512252 0.39±0.24 L082-5f V 0.54 1.78 0.183895 0.512627±12 0.512247 0.29±0.24

104

5. Metal transport in subduction fluids

5.6.2 Mineral chemistry

5.6.2.1 Garnet

Compositional maps of wall-rock eclogite garnet (Grt_I) indicate prograde growth zoning with a continuous increase of Mg from the core towards the rim, a Fe content that is fairly constant in the core, decreases continuously towards the mantle and reaches the lowest value at the rim, whereas the Ca content displays a reverse pattern (Figs. 5-7 and 5-9a). The oscillatory spessartine content is highest in the core and lowest at the rim (Fig. 5-7). Wall-rock eclogite garnet has core compositions of about Alm66–

67Prp4–6Grs23–25Sps4–5 with XFe of 0.92–0.93%, mantle compositions of Alm65–67Prp6–

9Grs21–23Sps3–4 with XFe of 0.87–0.89 and rim compositions of Alm59–60Prp11–12Grs27–

29Sps1–2 with XFe of 0.83–0.85 (Table 5-4). When normalized against chondrite, garnet displays a REE pattern with a relative depletion in LREE and a relative enrichment of HREE (Fig. 5-10a). The prograde growth of WE garnet is also reflected by the concentric zoning of the (M-)HREE and Y concentrations which decrease from the core to the rim with Yb: 38.4 ppm in the core, 6.52 ppm in the mantle and 1.62 ppm in the rim, Y: 603 ppm in the core, 81.0 ppm in the mantle and 14.9 ppm in the rim. Ti and Ni also decrease from the core to the rim with Ti: 827 ppm in the core, 612 ppm in the mantle and 350 ppm in the rim, Ni: 2.25 ppm in the core, 1.71 ppm in the mantle and 0.995 ppm in the rim. In contrast to these elements Cr and Zn show moderate increasing concentrations from the core to the rim (Cr: 428 ppm in the core, 456 ppm in the mantle and 553 ppm in the rim, Zn: 47.7 ppm in the core, 66.1 ppm in the mantle and 77.3 ppm in the rim; Table 5-5).

Selvage garnet (Grt_II) has a distinctly different chemical compositional zoning compared to that of the WE garnet, i.e. Grt_I (Fig. 5-8c–e vs. 5-7 and Fig. 5-9c vs. 9a). In contrast to Grt_I, which displays a continuous core–to–rim Fe, Mn decrease and an Mg increase (Figs. 5-7 and 5-9a), Grt_II profiles show rather uniform Fe, Mg and Ca contents (Figs. 5-8c–e and 5-9c). Grt_II has overall constant Fe contents, whereas Mg contents are only constant in the central domain and somewhat lower in the rim. The

105

5. Metal transport in subduction fluids

Figure 5-7 Mg, Fe, Ca and Mn X-ray maps of host eclogite garnet with a perfectly undisturbed compositional zoning profile. Continuous core–to–rim oscillatory Mn zoning is found in the garnet. most striking feature of Grt_II is displayed by the Ca maps which show patchy variations that are monitoring the shape of overgrown matrix minerals (Fig. 5-8b–f). In general, the idioblastic Grt_II has a similar composition compared with the rim domain of Grt_I in the WE (Table 5-4, Figs. 5-8, 5-9a and c). It is noteworthy that the relationship between Grt_I (WE garnet) and Grt_II (selvage garnet) is best documented by the garnet (Fig. 5-8b) located along the boundary of the selvage and the wall-rock eclogite. The garnet contains two (corrosive) irregular-shaped cores (relicts of Grt_I) which show the same composition as the Grt_I core (Figs. 5-8b vs. 8a and 5-7, and Fig. 5-9b vs. 9a), and distinctive mantle and rim domains (Grt_II) which are chemically similar to those of garnet grains in the selvage (Fig. 5-8b vs. 8c–e and Fig. 5-9b vs. 9c). The irregular shape of the cores was likely caused by resorption during reaction with the vein-forming fluid. The corrosive boundary between Grt_I and Grt_II is reflected by the

106

5. Metal transport in subduction fluids

Mg, Fe and Mn maps and –although not as clear– the Ca map (Fig. 5-8b) all of which

107

5. Metal transport in subduction fluids

◀Figure 5-8 Mg, Fe, Ca and Mn X-ray maps of WE garnet (a), selvage garnets (b–e) and vein garnet (f). For every element the color bar values of all garnet grains were adjusted in order to allow the comparison of the relative garnet compositions. (a) The WE garnet (Grt_I) contains abundant mineral inclusions and displays regular prograde elemental zoning, e.g., continuous Mg increase and Mn decrease from the core to the rim despite the faint oscillatory Mn zoning in the mantle domain. (b) The garnet best displaying the intimate relationship of wall-rock eclogite and selvage shows a corrosive core domain (Grt_I) that is equivalent to the WE garnet and a well-shaped mantle-rim domain (Grt_II) that is chemically identical to the selvage and vein garnets. (c–e) The selvage garnet (Grt_II), which occasionally contains relicts of the WE Grt_I core, reveals somewhat decreasing Mn contents from the core to the mantle with a slight increase towards the Mn rim. (f) The rare vein garnet (Grt_II) displays the same core–to–rim zoning and shows chemical equilibrium with selvage garnet. Some patchy variations in Ca are highlighted by dotted circles and arrows. display an abrupt change of content (Fig. 5-9b). Grt_I relicts are absent with increasing distance from the wall rock-selvage boundary towards the vein (Fig. 5-8c and e) or very rare (Fig. 5-8d). The trace-element pattern also reveals a relative depletion of LREE and a relative enrichment of HREE in the chondrite-normalized REE diagram (Fig. 5-10a). The concentrations of HREE and Y of Grt_II, which slightly decrease from the core to the rim, range between those of the mantle and the rim of Grt_I from the WE (Fig. 5-10a). V and Cr show slightly decreasing concentrations from the core to the rim while Co and Zn exhibit an inverse behavior (e.g., Cr: 311 ppm in the core, 473 ppm in the mantle and 698 ppm in the rim, Zn: 73.1 ppm in the core, 65.8 ppm in the mantle and 61.6 ppm in the rim; Table 5-5). In the selvage, garnet inclusions in pyrite have similar compositions as the matrix garnets.

Vein garnet is compositionally indistinguishable from Grt_II (Figs. 5-8f and 5-9d). Trace-element concentrations of the vein garnet are also similar to those of Grt_II (Table 5-5) displaying LREE depletion and HREE enrichment patterns in the chondrite-normalized REE diagram. In addition the vein garnet shows a slight REE, Y and Co decrease and Cr increase from the core to the rim (Fig. 5-10a, Table 5-5).

5.6.2.2 Omphacite

Omphacites from the WE, the selvage and the vein show a similar wide range of compositions. The vein omphacite has the highest jadeite component with up to 55 mol.%

108

5. Metal transport in subduction fluids

(Fig. 5-9f). The jadeite component (32.4 mol.% to 45.7 mol.%) of omphacite inclusions in matrix garnet and pyrite is almost identical to that of WE matrix omphacite (32.8 mol.% to 46.5 mol.%). The selvage omphacite has a similar jadeite component between

109

5. Metal transport in subduction fluids

◀Figure 5-9 Chemical compositions of garnet, omphacite and amphibole. (a) Compositional zonings of garnet in the wall-rock eclogite (Fig. 5-7) showing continuous prograde growth. Alm=almandine, Gross=grossular, Spess=spessartine. (b) Compositional profile of the garnet at the boundary between the WE and the selvage (Fig. 5-8b), in which from the corrosive Grt_I core to the Grt_II mantle and rim shows an abrupt compositional change. (c) Profile of the selvage garnet (Fig. 5-8c) showing rather uniform compositions from the core to the rim. (d) Compositional profile of the vein garnet that is in chemical equilibrium with the selvage garnet. (e) Chemical ternary composition diagram of garnet in the vein, the selvage and the WE eclogite. (f) Compositional ternary classification diagram of omphacite after Morimoto et al. (1988) shows that omphacites in the WE, the selvage and the vein have a similar jadeite component. (g, h) Chemical composition of amphibole in the WE, the selvage and the vein (classification after Leake et al., 1997).

31.8 mol.% and 48.0 mol.%. In the vein, the jadeite component of the omphacite shows a wider range from 31.5 mol.% to 55.0 mol.% than that of the omphacite inclusions in vein pyrite ranging from 37.1 mol.% to 50.5 mol.% (Table 5-4).

The concentrations of LILE and HFSE in omphacite are mostly below the detection limit (Table 5-5). However, omphacite is the main carrier of Li, the concentration of which shows little variation for the omphacites from the WE (e.g., 59.9 ppm), the selvage (45.1 ppm) and the vein (51.7 ppm). Moreover, omphacite carries considerable amounts of transition metals. Vein omphacite has relatively lower concentrations of most analyzed transition metals compared to the WE and selvage omphacites (e.g., V and Cr), but Cu is similar throughout the profile (Fig. 5-11a). The Co and Ni concentrations in omphacite from the WE are much higher than those from the selvage and the vein, whereas the Zn concentration increases from the WE via the selvage to the vein (Fig. 5-11a). No trace element zoning was found in the selvage and vein omphacite.

5.6.2.3 Amphibole

Sodic amphibole inclusions in garnet and matrix Na-amphibole have similar chemical compositions and are classified as glaucophane or crossite according to the classification of Leake et al. (1997) (Fig. 5-9g). Barroisite occasionally replaces glaucophane or omphacite along their rims in the WE and the selvage (Fig. 5-9h). Glaucophane is unzoned with AlIV and NaA–site varying from 0.2 to 0.4 and 0.1 to 0.3

110

5. Metal transport in subduction fluids

111

5. Metal transport in subduction fluids

112

5. Metal transport in subduction fluids pfu. The AlVI values range between 1.4 and 1.7 pfu. Fluorine and Cl contents are below the detection limit (Table 5-4).

The concentrations of most trace elements in the sodic and sodic-calc amphibole are below the detection limit, except Li and Ti. The Li concentration is 8.89 ppm and the Ti concentration 300 ppm in glaucophane. The barroisite has slightly higher concentrations of Ti, Sr and Pb in comparison with glaucophane (Table 5-5). Furthermore, amphibole

Figure 5-10 Chondrite-normalized REE concentrations of minerals from the WE, the selvage and the vein. (a) Garnet. c=core; m=mantle; r=rim. (b) Lawsonite. (c) Epidote in the selvage. (d) Epidote in the WE and the vein. (e) Titanite. (f) Apatite. Normalization-values are taken from Sun and McDonough (1989).

113

5. Metal transport in subduction fluids

114

5. Metal transport in subduction fluids

115

5. Metal transport in subduction fluids contains considerable concentrations of transition metals, which show concentration variations along the profile towards the vein. In general, the concentrations of V, Cr, Ni and Co of glaucophane decease from the WE to the vein (Fig. 5-11b). In contrast, the Zn and Cu concentrations of glaucophane increase slightly from the WE to the vein (Fig. 5-11b).

5.6.2.4 Lawsonite

Lawsonite occurs as inclusions in pyrite or garnet with a composition close to the ideal formula of CaAlSi2O7(OH)2·H 2O (Table 5-4). The FeO content of lawsonite inclusions in pyrite varies from 0.79 wt.% to 1.62 wt.%, while that of lawsonite enclosed by garnet is 0.88 wt.% (Table 5-4). The presence of lawsonite is also confirmed by the characteristic Raman spectroscopic peak at 562 cm-1, 695 cm-1 and 937 cm-1 (Fig. 5-4d and Fig. 5-5g).

The concentrations of most LILE and HFSE are below the detection limits for lawsonite (Table 5-5). However, it contains considerable concentrations of REE (∑REE range between 274 and 189 ppm), Sr (1566–1133 ppm), Pb (19.9–5.28 ppm), Y (30.3–28.1 ppm) Ti (576– 398 ppm) and some transition metals (e.g., V between 449 and 341 ppm, Cr between 913 and 146 ppm). Lawsonite displays chondrite- normalized REE pattern with a negative slope, with high LREE and lower HREE concentrations (Fig. Figure 5-11 Transition metal element concentrations (average values) of omphacite (a) and amphibole (b) 5-10b). In general, the REE patterns in the WE, the selvage and the vein.

116

5. Metal transport in subduction fluids are similar to that of selvage epidote (Fig. 5-10b and c), but the LREE content is higher and the HREE content lower (Fig. 5-10b and c) than those of the associated epidote inclusions in pyrite (Fig. 5-5f).

5.6.2.5 Epidote group minerals

Most epidote group minerals of all samples have high Fe2O3 contents (~10 wt.%) and are classified as epidote, whereas a few inclusions in WE garnet are clinozoisite (=Cz) with lower Fe2O3 contents (~5.56 wt.%) and inclusions in WE epidote/dolomite are allanite (Fig. 5-12a). The allanite inclusions in WE epidote/dolomite may contain significant LREE up to ca. 10 wt.% LREE (Table 5-5). The Al content decreases while Fe and to a minor extent Mn increase correspondingly from the core to the rim of the WE epidote (Fig. 5-12e), the selvage epidote (Fig. 5-12c, d and f) as well as the vein epidote. The epidote (and clinozoisite) is an important carrier of Sr, Y, Pb, P and REE, with Sr concentrations between 869 and 2828 ppm (557 ppm in Cz), Y between 18.3 and 49.7 ppm (371 ppm in Cz), Pb between 17.2 and 41.7 ppm (5.17 ppm in Cz) and P between 122 and 393 ppm (69.8 ppm in Cz) (Table 5-5). The ∑REE concentrations of epidote vary from 25.4 ppm to 393 ppm and clinozoisite in garnet has 2689 ppm. Additionally, epidote also contains considerable concentrations of some transition metals such as V and Cr (266–520 ppm and 747–1541, respectively). The selvage epidote is zoned with regards to the trace elements. The Ti content decreases from the core to the rim while most other trace elements increase, including the REE, Y, Ba, Zr, Hf, U, Pb and the transition metal elements V, Cr, Mn, Co, Ni and Zn.

Although the epidote in the WE, selvage and vein does not have significant compositional variations of major elements, its chondrite-normalized REE patterns reveal two distinct types. The first type (Ep_I) displays a positive slope with a relative depletion of the LREE compared to the HREE, whereas the second type (labeled as Ep_II) has an inverse pattern (Fig. 5-10d). No variation in Sr and Pb was observed in these two types. The WE contains both types of epidote (Ep_I and Ep_II) (Fig. 5-10d)

117

5. Metal transport in subduction fluids while the selvage only contains Ep_II (Fig. 5-10c). Both Ep_I and Ep_II occur in the vein and porphyroblastic epidote has Ep_II cores and Ep_I rims (Fig. 5-10d).

Figure 5-12 BSE images and element maps showing the microstructure of epidote and dolomite in the WE and the selvage. (a) WE epidote has a REE-rich allanite core; (b) BSE image of selvage epidote (Ep_II); (c, d) Al and Fe X-ray maps of selvage epidote indicating decreasing Al and increasing Fe contents from the core to the rim; (e) Compositional zoning of WE epidote; (f) Compositional zoning of selvage epidote; (g) BSE image of WE dolomite; (h) Mg X-ray map of WE dolomite indicating the compositional zoning with core–to–rim increasing Mg content; (i) BSE image of selvage dolomite including patchy sectors; (j) Mg X-ray map of selvage dolomite monitoring the high Fe, Mn and low Mg patchy variations.

5.6.2.6 White mica

Two types of white mica, phengite and minor paragonite, occur as matrix minerals and as inclusions in garnet in the WE. Minor contents of phengite and paragonite also occur

118

5. Metal transport in subduction fluids in the selvage matrix and as inclusions in the selvage pyrite. The Si–content of phengite in WE clusters at 3.45–3.50 pfu and the Na-content of paragonite at 0.85–0.94 pfu. The Si–content of selvage phengite (3.41–3.48 pfu) is similar to that of the WE.

The HFSE and REE concentrations in phengite and paragonite are generally below the detection limit (Table 5-5). Phengite has much higher concentrations of Cs, Rb and Ba than paragonite and lower Sr and Pb contents (e.g., 9.19 ppm vs. 404 ppm Sr and 0.973 ppm vs. 2.21 ppm Pb). Furthermore, phengite contains Li (average 8.26 ppm) and some transition metals above the detection limit like average 241 ppm V, 905 ppm Cr, 10.4 ppm Co, 206 ppm Ni, and 106 ppm Zn. Paragonite also carries considerable concentrations of Cr and Cu (e.g., 2061 ppm and 2.61 ppm, respectively).

5.6.2.7 Dolomite

Dolomite in the present study is a dolomite-ankerite solid solution (Table 5-4). The WE dolomite shows compositional zoning (Li et al., 2013) with increasing Mg (Fig. 5-12g and h) and decreasing Fe contents from the core (XFe ~0.35) to the mantle (XFe ~0.26) and the rim (XFe ~0.20) (Table 5-4). However, the vein dolomite has a rather constant and uniform XFe value of around 0.20, similar to that of the WE dolomite rim. The selvage dolomite grains are in equilibrium with vein dolomite (XFe ~0.20) but contain some patchy-like sectors (Fig. 5-12i and j) with higher Fe and lower Mg contents (XFe ~ 0.25) (Table 5-4). The MnO contents of the patchy sectors in the selvage dolomite (0.89–1.37 wt.%) and the dolomite inclusions in the selvage pyrite (e.g., 1.13 wt.%) are much higher than that of the WE matrix dolomite (0.15–0.58 wt.%), the predominant domain of the selvage dolomite (0.43–0.58 wt.%), and the vein dolomite (0.42–0.65 wt.%; (Table 5-4). Most trace-element concentrations are below 1 ppm or the detection limit except Sr (average 212 ppm) and Pb (average 1.87 ppm). The Sr and Pb concentrations of dolomite in the WE are slightly higher than those in dolomite from the selvage and the vein. Moreover, the dolomite displays high contents of some transition metals (Table 5-5) like Co (8.18 to 57.4 ppm), Ni (110 to 358 ppm) and Zn (177 to 227 ppm). The selvage dolomite shows lower Co and Ni concentrations (8.18 and 110 ppm,

119

5. Metal transport in subduction fluids respectively) than the WE dolomite (36.6 and 250 ppm, respectively). In addition, the Mn-enriched patchy sector of the selvage dolomite contains lower Co and Ni concentrations (e.g., 0.434 and 23.5 ppm, respectively) and higher REE (above detection limit), Y (1.53 ppm), Sr (1147 ppm) and Pb (12.4 ppm) concentrations than the other domains of the selvage dolomite (Fig. 5-12i and j, Table 5-5).

5.6.2.8 Rutile/Titanite

The HFSE contents in the selvage rutile are similar to that of the WE with Nb and Ta varying from 897 to 1428 ppm and 36.8 to 96.1 ppm, respectively. The Nb/Ta ratio of rutile varies from 14.2 to 20.8 in the WE and from 17.8 to 27.2 in the selvage. The Zr and Hf contents of rutile range from 19.2 to 103 ppm and from 0.75 to 4.4 ppm, respectively. The Zr/Hf ratio of rutile varies between 21.4 and 39.8 ppm in the WE, and 27.8 and 31.6 ppm in the selvage. The Nb/Ta ratios positively correlate with the Zr/Hf ratios in rutile. Rutile also hosts some transition metals like V (747 to 847 ppm), Cr (645 to 1582 ppm), Cu (0.66 to 2.02 ppm) and Zn (3.39 to 10.7 ppm) (Table 5-5).

The TiO2 content of titanite ranges between 34 and 39 wt.% and the Al2O3 content between 1.71 and 2.42 wt.%. Titanite also contains minor fluorine varying from 0.28 to 0.56 wt.% (Table 5-4) and hosts considerable amounts of HFSE and REE. The Nb concentration varies from 58 to 88 ppm, Ta from 6.69 to 14.6 ppm, Zr from 5.37 to 6.91 ppm and Hf from 0.512 ppm to 0.595 ppm (Table 5-5). The Nb/Ta ratio varies from 6.02 to 8.68 and the Zr/Hf ratios within a narrow range from 10.1 to 11.8. The HFSE in titanite are lower than those in rutile, but the Nb/Ta ratios also show a positive correlation with Zr/Hf ratios in the former. Titanite in the WE displays a relative enrichment (chondrite-normalized) of the middle rare earth elements (MREE) compared to the HREE and the LREE (Fig. 5-10e). However, titanite from the selvage shows an enrichment of both the MREE and the HREE relative to the LREE (Fig. 5-10e). Furthermore, titanite also contains some transition metals like V (up to 239 ppm), Cr (up to 290 ppm) and Mn (up to 104 ppm) (Table 5-5).

120

5. Metal transport in subduction fluids

5.6.2.9 Apatite

Apatite has a CaO content of ~55 wt.%, a P2O5 content of ~40 wt.%, and a F content of ~3.6 wt.% (Table 5-4). The Cl content usually is below the detection limit. The LILE, HFSE and transition metal element concentrations of apatite mostly are also below the detection limit (Table 5-5). However, apatite contains considerable amounts of Sr (1271–2561 ppm), Pb (4.42–21.1 ppm), Y (160–211 ppm) and MREE (Fig. 5-10f).

Table 5-6 Representative trace element composition of sulfides in the wall-rock eclogite, reaction selvage and vein (in ppm) Sample L0910 L0910 L082-5 L082-5 L082-5 L082-5 L082-5 L082-5 L082-5 Location WE WE RS RS V V WE RS RS Mineral Py-c Py-r Py-c Py-r Py-c Py-r Cpy Cpy Cpy No. 10 11 13 12 4 3 4 8 10 Au 0.008 0.007 bdl bdl bdl bdl 0.061 0.311 0.248 Ni 1465 2447 807 1280 3162 2617 64.4 1096 4793 V bdl 0.011 bdl bdl 0.016 bdl bdl 9.45 16.1 Cr bdl bdl bdl bdl bdl 0.394 4.67 120 223 Mn 0.078 bdl bdl 0.057 bdl bdl bdl 24.9 122 Co 2484 974 274 883 209 137 87.4 13.1 24.7 Cu 0.262 0.407 0.431 0.468 0.933 1.09 - - - Zn 0.257 0.330 0.243 0.201 0.295 0.150 35.4 64.3 225 As 21.2 22.3 3.30 42.2 11.0 8.98 0.545 9.60 17.0 Se 3.43 5.33 2.92 13.4 7.59 9.33 30.0 59.8 66.7 Mo bdl bdl bdl bdl 0.011 bdl 0.105 56.7 236 Ag bdl bdl bdl bdl bdl bdl 28.8 14.9 45.1 Sb bdl bdl bdl bdl bdl bdl 0.045 0.558 1.81 Te bdl bdl bdl bdl bdl bdl 0.407 1.93 7.79 W bdl bdl bdl bdl bdl bdl bdl 0.155 1.11 Pb 0.009 bdl 0.017 bdl bdl bdl 8.57 21.0 28.0 Bi bdl bdl bdl bdl bdl bdl bdl 1.98 0.268

5.6.2.10 Sulfides

Pyrite is the dominant sulfide and has a S content of ~54 wt.%, a Fe content of ~46 wt.% and a minor Ni content of <1 wt.% (Table 5-6). Minor chalcopyrite has a S content of ~35 wt.%, a Fe content of ~31 wt.% and a Cu content of ~34 wt.% (Table 5-6). Pyrite is only zoned in Ni and Co with 1465 ppm Ni in the core vs. 2447 ppm in the rim in the WE, 807 ppm vs. 1280 ppm Ni in the selvage, 3162 ppm vs. 2617 ppm Ni in the vein;

121

5. Metal transport in subduction fluids

2484 ppm Co in the core vs. 974 ppm in the rim in the WE, 274 ppm vs. 883 ppm Co in the selvage and 209 ppm vs. 137 ppm Co in the vein (Table 5-6). The Cu and Zn concentrations of pyrite are generally lower than 1 ppm. The vein pyrite contains up to 1 ppm Cu which is <0.5 ppm in pyrite from the WE and the selvage. Furthermore, chalcopyrite also hosts up to 4793 ppm Ni, 223 ppm Cr, 24.7 ppm Co, 225 ppm Zn, 0.248 ppm Au and 45.1 ppm Ag (Table 5-6). The Au and PGE (Ru, Rh, Pd, Os, Ir, Pt) concentrations generally are below the detection limit.

5.7 Discussion

5.7.1 The formation of the vein-selvage system

5.7.1.1 P–T history of the WE-selvage-vein system

Vein-network systems in eclogite-facies rocks represent fossilized fluid pathways and record the presence of a free fluid phase at high-pressure P–T conditions. They may provide vital information on fluid-rock interaction processes in oceanic subduction zones (e.g., Gao and Klemd, 2001; John and Schenk, 2003; Spandler and Hermann, 2006; John et al., 2008, 2012; Spandler et al., 2011; Klemd, 2013). However, the exact P–T conditions at which the vein minerals formed often remain largely unconstrained due to the lack of critical mineral assemblages for geothermobarometry. Accordingly, the P–T conditions of vein formation and related fluid-rock interaction are often constrained via geothermobarometrical studies of the immediate wall rocks of the veins.

In the present study the WE contains an eclogite-facies mineral assemblage in which mineral inclusions in garnet clearly reflect the prograde evolution of the rock. The eclogite-facies selvage assemblage contains only rarely blueschist-facies indicative minerals such as glaucophane (Fig. 5-3b). Pyrite from the WE and selvage contains various kinds of mineral inclusions, such as lawsonite, epidote, chlorite, paragonite, glaucophane, quartz, phengite, titanite, omphacite, garnet, rutile and dolomite (Figs. 5-4 and 5-5) which are thought to record a transitional process from blueschist- to

122

5. Metal transport in subduction fluids eclogite-facies conditions. Garnet porphyroblasts in the WE contain a similar mineral inclusion assemblage (Fig. 5-3b, Fig. 5-4a), suggesting that both minerals grew during a common prograde metamorphic evolution at the blueschist-to-eclogite transition. In addition, dolomite inclusions in the core and mantle domains of the WE garnet and pyrite (Fig. 5-3a–d) and the rather homogeneous distribution (no mode change from the vein towards the WE) of matrix dolomite suggest that the dolomite formed in the oceanic slab via seafloor alteration prior to subduction rather than during vein metasomatism. Compositionally, the matrix omphacite and phengite, garnet and dolomite rims in the WE, the selvage, and the vein are very similar (Fig. 5-9a–f; Table 5-4). This suggests that the WE, the selvage and the vein equilibrated at the same prograde to peak metamorphic conditions. Consequently, a major influx of external fluids during retrograde exhumation-related processes seems to be very unlikely. In fact, exhumation-related fluid flow results in metasomatism along pathways surrounding and not crosscutting the dismembered massive mafic rocks derived from the descending slab as was shown by field observations (for details see van der Straaten et al., 2008) and theoretical considerations (Ague, 2007).

In order to constrain the P–T evolution of the wall-rock eclogite phase equilibrium modeling was primarily undertaken in the NMnCaKFMASCHO (Na2O–MnO–CaO–

K2O–FeO–MgO–Al2O3–SiO2–CO2–H2O–Fe2O3) model system. TiO2 was not considered in the calculation because it is mostly incorporated in Ti-dominated phases.

MnO and Fe2O3 are included since they affect the stability fields of garnet and amphibole, respectively. Pseudosection calculations were performed using the Perple_X software package (Connolly, 1990, 2005; version 6.6.7) and the internally consistent thermodynamic database of Holland and Powell (1998 and update). The following solid-solution models were used: Gt(HP) for garnet (Holland and Powell, 1998), Omph(GHP) for omphacite (Green et al., 2007), cAmph(DP) for amphibole (Diener et al., 2007), Mica(CHA) for phengite (Coggon and Holland, 2002), Chl(HP) for chlorite (Holland and Powell, 1998), Ep(HP) for epidote (Holland and Powell, 1998), Do(HP) for dolomite–ankerite solid solution (Holland and Powell, 1998), and F for H2O–CO2

123

5. Metal transport in subduction fluids fluid solution (Connolly and Trommsdorff, 1991). Lawsonite and quartz are pure end-number phases. Bulk compositions obtained by XRF analysis are not suitable for the phase equilibrium modeling due to effects such as crystal fractionation (e.g., zoned garnet) and retrograde metamorphism and thus the rock may have only domain-like equilibria (e.g., Evans, 2004; Konrad–Schmolke et al., 2008). The effective bulk composition used in this study was calculated by integrating mineral compositions and modal abundance data for the phases present, in which only half of the garnet mode is taken into account. The bulk rock compositions generated with this approach are appropriate to model the phase equilibria from the pre-peak to the peak stage of metamorphism (Wei et al., 2009 and references therein). The effective bulk composition calculated for pseudosection modeling is given in Figure 5-13a, while the H2O content was set at 4.68 wt.% which is in excess for saturating the mineral assemblages in all shown P–T fields (e.g., Chen et al., 2013). For more details concerning the pseudosection calculations see Li et al. (2012).

A calculated phase equilibrium diagram for an effective bulk composition is presented as Figure 5-13 for the pressure–temperature range of 400–600 °C and 15–25 kbar. Omphacite, phengite, dolomite and fluid are stable phases in all stability fields. However, the omphacite and garnet modal abundances increase with increasing P and T conditions, while the amphibole, lawsonite and chlorite abundances decrease. Garnet-in, chlorite-out, quartz-in and amphibole-out isograds appear almost pressure insensitive at T = ca. 400°C, 530°C, 540°C and 560°C, respectively, in the lawsonite-present field. Siderite occurs in the HP/LT field and aragonite in the LP/LT field. Lawsonite starts to disappear when the temperature rises above 475 °C. Moreover, the pseudosection was contoured for the lawsonite modal abundance, suggesting that the stability of lawsonite is largely T-dependent, i.e., it rapidly breaks down when the temperature exceeds 500°C (Fig. 5-13a). However, minor discrepancies with regard to the epidote and amphibole stability fields between the pseudosection and sample observation may be due to the

Fe2O3 estimation, modeling uncertainty or decompression replacement.

124

5. Metal transport in subduction fluids

Figure 5-13 (a) P–T pseudosection (using Perple_X) in the system NMnCaKFMASCHO for the host eclogite calculated with an effective bulk composition. The pseudosection is contoured for the lawsonite modal abundance. The solid dots represent evolution stages of the fluid in Figure 5-14. (amp = amphibole; arag = aragonite; chl = chlorite; dol = dolomite; ep = epidote; f = fluid; grt = garnet; law = lawsonite; omp = omphacite; qtz = quartz; rhc = rhodochrosite; sid = siderite). (b) The pseudosections are contoured with grossular, pyrope and spessartine isopleths and Si isopleths in phengite (p.f.u) to constrain the P–T path of the carbonate-bearing eclogite. The peak metamorphic conditions (gray field) were determined at 580–590 ºC and 22.5–23.5 kbar using the compositions of WE garnet rim (solid gray dots). The prograde P–T path is estimated on the basis of garnet core-mantle compositions (open dots).

Isopleths (mole fractions) of spessartine, pyrope and grossular and Si (per formula unit) in phengite were modeled for the P–T range of 500–600 °C and 20–25 kbar to determine the peak metamorphic conditions of the WE (Fig. 5-13b). The Si isopleths of phengite show that Si in phengite increases with increasing pressure in the lawsonite stability field (Fig. 5-13b). The spessartine contents decrease with increasing temperature. However, the modeling results show that the amphibole-out and lawsonite-out isograds play important roles in determining the distribution of some isopleths. For example, the pyrope isopleths are temperature dependent and the Mg– content increases with increasing temperatures in the amphibole stability field. Whereas the grossular isopleths are pressure dependent and the Ca–content decreases with increasing pressure in the amphibole/lawsonite-present field, but respectively switch to be temperature dependent and increase with increasing temperature after the disappearance of either amphibole or lawsonite (Fig. 5-13b). In the wall-rock eclogite,

125

5. Metal transport in subduction fluids the grossular component in the garnet rims varies from 0.27 to 0.29, the pyrope component from 0.11 to 0.12, the spessartine component from 0.01 to 0.02 and Si (p.f.u) in phengite from 3.40 to 3.50 (Table 5-4) thereby constraining the peak metamorphic P– T conditions between 580–590 ºC and 22.5–23.5 kbar. These results are in accordance with the calculated P–T conditions of eclogite located nearby (Li et al., 2012) and comparable to other HP rocks in the study area (e.g., Klemd et al., 2002; Wei et al., 2003, 2009). The P–T path was likewise deduced from the garnet core–to–mantle composition profile (Fig. 5-13b).

5.7.1.2 Evolution of selvage-vein system

In subduction zones, seawater altered oceanic crust experiences dehydration with increasing P–T conditions, in particular at the blueschist- to eclogite-facies transition (Peacock, 1993). Large amounts of water are liberated from dehydration reactions, which include (e.g., Evans, 1990; Gao and Klemd, 2001):

9 Chl + 4 Pg + 16 Qtz = 13 Prp + 2Gln + 38 H2O

Gln + Pg = Prp +3 Jd + 2 Qtz + 2 H2O

12 Law + Gln = Prp + 2 Pg +6 Ep + 5 Qtz + 20 H2O

4 Law + Jd = 2 Ep + Pg + Qtz + 6 H2O

The prograde breakdown of hydrous minerals (e.g., lawsonite and glaucophane) release significant amounts of fluids while the volume of the newly formed minerals decreases (i.e., densification reactions), resulting in the dynamic formation of a fluid-filled porosity (John et al., 2012). The increase of pore fluid pressures may result in an enhanced interconnected porosity (permeability). Fractures allow the release of fluid pressure and thereby may create a pressure gradient focusing the fluid into these fracture zones within the dehydration domain (e.g., Yardley et al., 1986; John et al., 2008). Ultimately, the fluid produced by dehydration may generate an interconnected vein network when sufficient volume of fluids is liberated to allow fracturing-related

126

5. Metal transport in subduction fluids vein formation (Miller et al., 2003; John et al., 2008). Intense fluid-rock interaction may occur in the immediate wall-rock (selvage) along the main flow structures, which acts as major fluid conduits (e.g., Ague, 2003; Zack and John, 2007; Beinlich et al., 2010). The metasomatic-metamorphic fluid-rock interaction creates a transient increase of the porosity due to reaction in progress (John et al., 2012), which in turn facilitates mass transfer.

Fluids provide a medium to transport major and trace elements and induce re-equilibration through dissolution of parent phases and precipitation of product phases. The studied selvage formation process is considered to be accompanied by the dissolution of the WE minerals and precipitation of selvage mineral assemblages (cf., Putnis and Putnis, 2007; Putnis and John, 2010) as a consequence of the interaction between the WE and the externally derived vein-forming fluid. The selvage contains less garnet in comparison with the wall rock eclogite. The selvage garnet encloses corrosive relicts of WE garnet (Grt_I ) which shows a gradual decrease in size with increasing distance to the WE promoting the garnet as the best record of the dissolution-precipitation processes. In addition, the REE pattern and the major-trace element zonings of selvage epidote are different when compared to those of the WE epidote (Figs. 5-10c–d and 5-12a–f) suggesting that the selvage epidote also experienced dissolution-precipitation processes. This is also suggested for the selvage dolomite as is shown by the uniform XFe value in the selvage dolomite and the patchy distribution of compositional sectors instead of the XFe core–to–rim zoning seen in the WE dolomite (Fig. 5-12g–j). Dissolution–precipitation processes liberate major and trace elements from the parent mineral phases and results, after fluid-mediated diffusional or advective transport of these elements, in a complete or partial incorporation into product mineral phases. This depends on whether the system is open or closed, which itself may relate to the overall dimension of the considered system (cf., Putnis and John, 2010). Accordingly, the partitioning of trace elements among solid parent, solid product and involved fluid phases are believed to strongly depend on the

127

5. Metal transport in subduction fluids nature and transport properties of the interfacial fluid rather than on closed-system equilibrium partitioning coefficients (cf., Putnis and John, 2010).

The timing of the formation of the selvage-vein system is thought to be best recorded by the corrosive replacement of Grt_I close to the wall rock-selvage boundary (Fig. 5-8b). The newly formed selvage Grt_II is interpreted to have grown at the expense of Grt_I, i.e., by replacement processes (Putnis and John, 2010), while the WE garnet (Grt_I) was not/little affected (Fig. 5-7/8a) and the prograde zoning continuously developed during this period. The Grt_I rims in WE (Fig. 5-7), the rims of Grt_II replacement growth on Grt_I (e.g, Fig. 5-8a and b) and the rims of fully new precipitated selvage and vein garnet (Grt_II, Fig. 5-8c–f) have almost identical compositions (Fig. 5-9a–e) which also counts for the omphacite and the phengite in the WE, the selvage and the vein (e.g., Fig. 5-9f). This suggests that the vein-selvage-host rock system formed prograde close to the peak metamorphic conditions (e.g., Spandler and Hermann, 2006; Beinlich et al., 2010). The formation of the vein-selvage-host rock system is described by the following 4 stages (Fig. 5-14):

a) Stage T0: at the blueschist- to eclogite-facies transition hydrous minerals such as glaucophane, lawsonite and chlorite react to form garnet and omphacite thereby significant amounts of fluids are transferred from solid into a dynamically forming porosity. The increasing fluid pressure may enhance the interconnectivity of the porosity, hence the permeability within the WE.

b) Stage T1: the WE experiences an ongoing eclogitization while an external fluid forms the vein. The externally derived vein-forming fluid and the locally derived WE fluid interact and due to their compositional differences diffusional transport causes local changes in the bulk composition (solid + fluid) which triggers mineral reactions within the immediate wall-rock (i.e., selvage formation). The WE minerals, including omphacite (Omp_I), phengite, dolomite (Dol_I), epidote (Ep_I), lawsonite (Law_I), sulfide (sulfide_I) and garnet (Grt_I), start to dissolve due to the continuous change in the bulk composition of the local system (solid + fluid) and

128

5. Metal transport in subduction fluids

the new equilibrium mineral assemblage precipitates. A large numbers of elements formerly stored in solids are exchanged with the fluid.

c) Stage T2: the reaction front propagates further into the WE and the selvage domain widens. The selvage assemblages, including newly formed omphacite (Omp_II), dolomite (Dol_II), epidote (Ep_II), lawsonite (Law_II), sulfide (sulfide_II) and garnet (Grt_II), are precipitating rapidly at the reaction front as a consequence of the altered bulk composition. Only small amounts of lawsonite and glaucophane form due to the rather high temperature (Fig. 5-13a).

d) Stage T3: the mass action within the connected fluid system largely comes to an end and the selvage mineral assemblage is establishing equilibrium on a larger scale, including the Grt_II overgrowth of Grt_I relicts as a ―frozen in‖ record of the involved dissolution-precipitation process. This final equilibration occurs close to the peak metamorphic conditions, the newly formed Grt_II shows similar chemical zoning compared to that of the WE garnet rims (Fig. 5-8), and also omphacite, dolomite, epidote, sulfides have similar bulk compositions throughout the vein-selvage-host rock system. After nucleation, Grt_II rapidly overgrows the selvage and vein matrix, which is indicated by the Ca concentration distribution in Grt_II (Fig 5-8b–f). The Ca distribution in Grt_II is believed to be transport controlled. The Ca zoning hence indicates that garnet grows faster than the surface diffusion on the propagating garnet surface. The Mn zoning is interpreted to be mainly supply controlled, while the Fe and Mg contents are thermodynamically buffered by the exchanging assemblages (cf., Chakraborty and Ganguly, 1992; Ganguly et al., 1998; Carlson, 2002). Different types of newly formed minerals like lawsonite, garnet, omphacite and dolomite are enclosed also by pyrite.

The microstructural and mineral chemical evidence indicates that the vein-selvage-wall rock system formed during the prograde metamorphic evolution while still being part of a coherent slab and consequently experienced the same peak metamorphic conditions of 580–590 ºC and 22.5–23.5 kbar. During the whole fluid-rock interaction period the

129

5. Metal transport in subduction fluids element transport and mobilization persisted and accompanied the dissolution- precipitation processes.

Figure 5-14 Schematic illustration of the formation of the vein-selvage system and element mobilization during metasomatism of wall-rock eclogite. The left part addresses the selvage formation processes and the chemical feedbacks during the fluid-rock interaction, while the right column shows the concentration variations of locally derived Fluid_1 vs. externally derived Fluid_2 along the selvage profile (see text). (a) Snapshot of the eclogitization process of the wall rock eclogite (Stage T0) showing that Fluid_1, formerly stored in solids, is liberated into a dynamically forming porosity. (b) Stage T1: the external Ca- and S-rich fluid (Fluid_2) in contact with the WE system indicates chemical exchange with Fluid_1. The chemical exchange between the fluids changes the composition of the local bulk system which in turn triggers the fluid-rock interaction

130

5. Metal transport in subduction fluids that leads to the selvage formation. Some elements such as LILE, HREE and metal elements are released into the passing fluid phase accompanying mineral dissolutions (e.g., Grt_I, Omp_I, Dol_I, Ep_I, Law_I and sulfide_I). (c) Stage T2: The newly formed selvage assemblage (including Omp_II, Dol_II, Ep_II, Law_II and sulfide_II) is precipitated contemporaneously and partly formed due to the Ca, Sr, Pb and S increases under pre-peak P–T conditions. Thereby L(M)REE are also transferred into the new minerals. The reaction front is progressively moving into the WE, while the replacement of WE minerals continues. Some corrosive Grt_I relicts are preserved in the position where the effect of fluid-rock interaction decreases. (d) Stage T3: the newly formed selvage assemblages rapidly precipitate and Grt_I is enclosed by Grt II as is indicated by the Ca zoning of the latter. Newly formed selvage lawsonite, garnet, omphacite, dolomite are enclosed by the fast forming pyrite. The mixed fluid consisting of internal dehydration-derived fluid and externally introduced fluid eventually leaves the WE-selvage-vein system and the final vein assemblages precipitate. The evolution stages of the fluid are labeled in the pseudosection diagram (Fig. 5-13a).

5.7.2 Chemical feedback of fluid-rock interaction

Although continuous fluid liberation may take place down to 300 km depth in the subduction zone, the major pulse of fluid release of the upper oceanic crust occurs under blueschist-eclogite-facies conditions, corresponding to 70–100 km depth in subduction zones (Peacock, 1993; Schmidt and Poli, 1998). Limited element mobilization and redistribution on a small-scale mainly occurs during pervasive, internally-buffered intraslab fluid flow enhanced by local devolatilization reactions (e.g., Spandler et al., 2003; Spandler and Hermann, 2006; Gao et al., 2007; John et al., 2008) while extensive mass transport within slabs is associated with channelized, externally-derived fluid flow (e.g., Molina et al., 2004; John et al., 2004, 2008; Beinlich et al., 2010; Spandler et al., 2011; Penniston-Dorland et al., 2012; Herms et al., 2012). This flow may occur pulse-like at high fluxes and short durations (Dragovic et al., 2012; John et al., 2012). Previous studies suggest that the budget of different element groups in eclogites are controlled by certain minerals, e.g., Li by omphacite (Manning, 1998; Marschall et al., 2006; Beinlich et al., 2010), the LILE by phengite (Zack et al., 2001), the HFSE by rutile (Stalder et al., 1998; Schmidt et al., 2009), the HREE by garnet (Stalder et al., 1998) and the REE and Ca by lawsonite, apatite and epidote-group minerals (Sorensen and Grossman, 1989; Tribuzio et al., 1996; Feineman et al., 2007). Suppressed formation of these minerals can dramatically enhance the mobility of major and trace elements during fluid-rock interaction (John et al., 2004).

131

5. Metal transport in subduction fluids

5.7.2.1 Mass-balance calculations

Mass-balance calculations related to metasomatic selvages derived by fluid-rock interaction provide a quantitative estimate of element mass transfer compared with the qualitative judgment according to the absence, decrease and increase in the modal abundance of a certain mineral (e.g., Ague, 1991, 2011). This calculation will further clarify whether there are systematic variations with respect to a geochemical frame, which may be otherwise covered by a metasomatic overall mass gain or loss. To constrain the successful mass balance analysis, the ―precursor‖ rock (unaltered by metasomatism) and the geochemical frame (immobile elements during metasomatism) have to be chosen as references (Gresens, 1967; Ague, 2011). This is complicated due to the heterogeneous bulk chemistry and mineral distribution in the studied host eclogite. In the present study, we used the host eclogite (least altered sample) which is ca. 7.7 cm (labeled as (a)) away from the vein as the reference rock and the samples along the profile (6.2 cm (b), 4.3 cm (c), 2.4 cm (d) and 1 cm (e) towards the vein) were used for the mass balance calculation in order to obtain information on elemental loss or gain during fluid-rock interaction. Due to their low solubility in aqueous fluids the HFSE are commonly employed to define the geochemical reference frame. Nb–Ta are discarded from the potential frame reference candidates in this case since sample (e) shows visible lower modal rutile and the selvage rutile contains lower Nb–Ta concentrations in comparison with the WE rutile. Additionally minor rutile occurs in the vein, and mobilization of Ti–Nb–Ta has been reported as a consequence of extreme chemical gradients between the fluid phase and the immediate wall rock (cf., Xiao et al., 2006; Gao et al., 2007; Zhang et al., 2008; Rapp et al., 2010). Although the mobility of Zr was reported for F-rich hydrothermal fluids (e.g., Rubin et al., 1993) and Cl-rich fluids or melts (e.g., Higashino et al., 2013), Zr is still prevalent being taken as an immobile reference frame element during mass-balance calculations (e.g., Ague, 2003, 2011; Gao et al., 2007; Penniston-Dorland and Ferry, 2008). Zirconium and Hf were chosen as reference elements in the present study since the existence of fine-grained zircon grains in the host and the selvage and their absence in the vein suggest that both Zr and Hf are

132

5. Metal transport in subduction fluids rather immobile in the considered system and in addition the reactive fluid phase is thought to be neither F- nor Cl-rich. This is further supported by the constant Zr and Hf concentrations along the profile from 7.7 cm (a) to 1 cm (e) towards the vein (Fig. 5-6d). In addition, the Zr/Hf ratio of the wall-rock eclogite and the selvage samples varies from 37 to 46 and is within the range of typical oceanic basalts indicating the absence of a fractionation between Zr and Hf. The concentrations of Zr and Hf obtained from sample (a)–(e) result in an average Zr value of 106 ppm with a RSD of 9.5% and an average Hf value of 2.56 ppm with RSD=9.6%. Thus a representative estimate of 10% uncertainty introduced by the heterogeneity in the pre-metamorphic bulk composition and the analytical error associated with the measured concentrations of immobile elements is thought to be realistic (e.g., Penniston-Dorland and Ferry, 2008). The rinv

altered rock precursor values (Ci /Ci ; Ague, 2003) are 0.91 (Zr) and 0.93 (Hf) for the sample (b), 0.91 (Zr) and 1.03 (Hf) for sample (c), 1.06 (Zr) and 0.98 (Hf) for sample (d) and 0.83

(Zr) and 0.90 (Hf) for sample (e). The average rinv value of 0.92 indicates an overall rock mass gain of 9% (±0.9%) for sample (b). Sample (c) shows evidence for a minor mass gain of 3% (±0.3%) and sample (d) displays a minor mass loss of 2% (±0.2%), whereas the overall rock mass gain of sample (e) is up to 16% (±1.6%) indicating that no extensive mass loss or gain was experienced by the samples during fluid-rock interaction.

Although the samples of the selvage display relatively little overall mass change during the fluid-rock interaction, most of the trace elements show strong compositional changes throughout the sampled profile (Fig. 5-15) and only round numbers are quoted because of the uncertainties of the method. Major elements (Si, Al, Mg and Na) show mass variations between +20% (b) to –20% (d) in the selvage (Fig. 5-15b). Significant mass losses of LILE (K, Rb and Ba) reach up to ~100% for (d) and (e), whereas enormous gains of Sr, Pb and S are apparent (up to 170–210% for (d)). Calcium and V have also experienced significant mass gains in the selvage, ~50% and 35%, respectively (Fig. 5-15c). Most selvage samples did not experience a mass loss or gain for Ti, Nb and Ta, however a mass loss of ~40% for those elements is recorded by the

133

5. Metal transport in subduction fluids sample closest to the vein (Figs. 5-6d and 5-15d). With respect to the REE, both LREE and MREE of most selvage samples (except for e) experienced a mass gain of ca. 10–20% and 0% ~ 40% respectively (Fig. 5-15e and f), in contrast, the HREE (Er, Tm, Yb and

134

5. Metal transport in subduction fluids

◀Figure 5-15 Concentration ratio diagram illustrating the mass gain and loss of the eclogitic rocks in the selvage with variable distance towards the vein. Concentration ratios are derived from mass-balance calculations relative to the least altered wall rock eclogite (L082-5a) and Zr–Hf as immobile reference elements. The estimate of the concentration ratio for immobile elements is indicated by the thick horizontal black line, which is marked by 0% mass change. The gray horizontal lines show the range of the immobile reference frame and symbols falling within this range indicate no conclusive evidence for mass gain or loss. Error bars (2σ) are shown when they are larger than the size of symbol. Positive deviations from the reference line indicate a mass-gain while negative deviations represent a mass-loss. The diagrams usually display an abrupt change close to the vein (ca. 1cm) probably partly as a result of the vein composition participation in sample L082-5e due to the jagged boundary. (a) Major mineral modal abundances (vol.%) variations from host eclogite towards the vein; (b) A gradual Si–Mg–Na mass loss in the selvage except sample (e); (c) Significant mass gains of CaO, Sr, Pb, S and V and extreme mass losses of LILE (K–Rb–Ba); (d) Almost constant HFSE values in the selvage except Ti–Nb–Ta in sample (e); (e) Mass changes of LREE; (f) Mass gains of MREE; (g) Gradual HREE mass losses towards the vein; (h) Mass losses of transition metal elements.

Lu) display a gradually increasing mass loss from ~10% (b) to ~40% (e) (Fig. 5-15g). With regard to transition metal elements, Fe, Cu, Ni, Zn, Co, Cr and Mn have experienced mass losses of 10% to 40%, although the mass losses decrease for the sample close to the vein (Fig. 5-15h). In general, the chemical composition of the profile throughout the selvage, from the wall rock to the vein, may reflect a hybrid composition of the wall rock and the vein-forming fluid. The selvage generally displays element concentration profiles with a gradual increase of Ca, Sr, Pb, S, LREE and MREE, whereas Si–Mg–Na, LILE, HREE and transition metal elements decrease significantly from the unaltered host eclogite towards the vein (Fig. 5-15). This suggests that Ca, Sr, Pb, S, LREE and MREE may have been added to the metasomatic system and Si–Mg–Na, LILE, HREE and transition metal elements may have been scavenged and mobilized during the fluid-rock interaction.

5.7.2.2 Element mobility during fluid-rock interaction

The modal content of white mica (mostly phengite) decreases from ca. 10 vol.% in the WE (a), 5 vol.% in the selvage (b), 3 vol.% in the selvage (c), 1 vol.% in the selvage (d) to almost zero close to (e) and within the vein (Fig. 5-15a, Table 5-1) indicating that phengite has been dissolved during the selvage formation and that the LILE (K, Rb, Cs

135

5. Metal transport in subduction fluids and Ba) were drained into the fluid channel (Fig. 5-14). The LILE may have been transported out of the investigated system beyond the exposed vein since no LILE-hosting mineral is observed in the vein. The significant mass losses of LILE (up to ~100% (d)) are also in agreement with their high mobility in aqueous fluids. The successive breakdown of lawsonite (Fig. 5-13a) not only released considerable amounts of water into the wall rock system, but also supplied significant quantities of elements including Ca, Al, Sr, Pb and LREE for the precipitation of epidote in the selvage and vein (Fig. 5-15). The LREE-rich Ep_II probably formed during the breakdown of lawsonite and inherited the LREE budget. Nevertheless, the breakdown of lawsonite did not satisfy all the demand of Ca, Sr, Pb and REE necessary for the epidote crystallization as is indicated by the very high modal amount of epidote and the mass gains of these elements in the selvage (Fig. 5-15; cf., Widmer and Thompson, 2001). The dissolution of garnet, whose modal content gradually decreases from the WE to the vein (Fig. 5-15a, Table 5-1), may have liberated additional Al and Ca as well as Fe, Mn and HREE into the fluid (Fig. 5-14b and c). Ca and Al were incorporated into the epidote in the selvage and the vein. The vein epidote has a chondrite-normalized REE pattern displaying a negative slope in the Ep_II core and a positive slope of enriched HREE relative to LREE in the Ep_I rim (Fig. 5-10d), suggesting that the HREE might have been delivered from dissolving Grt_I and incorporated into the rims of vein epidote. Thus the commonly thought immobile HREE in metamorphic fluids (e.g., Stalder et al., 1998) were mobilized during the formation of the vein-selvage system, which was also observed in other HP vein studies from eclogite-facies terranes (e.g., Spandler and Hermann, 2006; Gao et al., 2007; Zhang et al., 2008; Guo et al., 2012). Compared with the LILE and REE, the HFSE (Ti, Nb, Ta, Zr and Hf) behaved relatively immobile (Fig. 5-15d). However, the drastic Ti, Nb and Ta decreases close to the vein may be the result of incorporation of vein material in sample (e) due to the jagged selvage-vein boundary.

According to the mass balance calculation (Fig. 5-15h), most of transition metal elements (Fe, Cu, Ni, Zn, Mn, Co, Cr) experienced losses of 10% to ~ 40%, whereas

136

5. Metal transport in subduction fluids vanadium displays a gain of ~35%. The calculated distributions of transition metals among the different high-pressure minerals of the wall-rock eclogite are based on the bulk rock element concentrations, the element concentrations of the minerals and their modal abundances (Table 5-7). As expected, Ti is mostly incorporated in rutile and titanite with 84% and 12% of the bulk Ti, respectively. 70% of the bulk V is mainly hosted in omphacite and 11% bulk V in phengite (similar values below). Chromium occurs in omphacite (45%), garnet (20%) and phengite (20%), whereas Mn is predominantly hosted by garnet (57%) and dolomite (36%). Iron is incorporated in several minerals such as garnet (40%), omphacite (27%), dolomite (15%) and pyrite (13%), whereas Co is hosted by pyrite (76%) and omphacite (16%). Nickel is incorporated in omphacite (43%), pyrite (25%), phengite (15%) and dolomite (13%). Copper is almost exclusively hosted by chalcopyrite. Zinc occurs in omphacite (56%), dolomite (21%) and garnet (11%) (Table 5-7). The garnet (Grt_I) and phengite dissolutions may have liberated significant amount of Fe, Mn, Cr, Ni and Zn into the fluid. Iron was captured by pyrite when externally-derived S infiltrated into the metasomatic system, and it also was incorporated in the idioblastic selvage and vein epidote as well as in the vein dolomite and omphacite (Fig. 5-14). Manganese mainly occurs in dolomite (especially the high Mn patchy sectors) while Cr is incorporated in the selvage and vein epidote and the vein omphacite. Nickel occurs in the mega-pyrite

Table 5-7 Distributions (wt. %) of transitional metal elements in the wall-rock eclogite* Min. Fe Mn V Cr Ti Co Ni Cu Zn Grt 39.7* 56.9 5.0 20.1 0.4 3.5 0.1 - 11.1 Omp 27.2 6.3 70.4 45.0 1.0 15.8 43.4 0.2 55.7 Gln 0.8 0.1 1.6 1.5 - 0.3 2.4 - 2.6 Dol 14.7 35.6 0.1 0.2 0.3 3.8 12.5 - 20.8 Phn 2.1 - 11.2 20.1 1.6 1.0 14.5 - 9.1 Rt - - 5.8 4.0 84.0 - - - 0.1 Ttn - - 0.6 0.3 12.4 - - - - Ep 1.7 1.1 5.3 8.8 0.3 - 0.1 - 0.5 Py 13.4 - - - - 75.7 25.5 - - Cpy 0.4 - - 0.1 - - 1.5 99.8 0.2 *The values are calculated based on transition metal concentrations in different minerals (data from both EPM analysis and mineral trace elements measurement) in combination of the mineral modal abundances from thinsection point counting. The italic bold font refers to the predominant distributions (>10 wt.%).

137

5. Metal transport in subduction fluids crystal and dolomite, whereas Zn predominantly in vein omphacite and dolomite (Fig. 5-14). Omphacite is thought to host most of the transition metal elements in the WE. It hosts about 27% of the bulk Fe, 70% V, 45% Cr, 16% Co, 43% Ni and 56% Zn, whereas dolomite carries 15% of the bulk Fe, 36% Mn, 13% Ni and 21% Zn (Table 5-7). The Co and Ni contents in the selvage omphacite and dolomite are much lower than those in the WE (Table 5-5) suggesting their re-equilibration during the selvage formation and that Co and Ni partitioned into the fluid. Cobalt and Ni may have been re-captured by the selvage pyrite, the vein pyrite and the dolomite, and part of the Ni by the vein omphacite (Fig. 5-14). Furthermore, since the selvage pyrite and chalcopyrite host different Co and Ni concentrations (even in the core and rim), which is in contrast to the WE pyrite and chalcopyrite (Table 5-6), the sulfides in the selvage are also thought to have experienced dissolution-precipitation during the metasomatic selvage formation (Fig. 5-14). This process may contribute to the mobilization of Cu and partly of Co–Ni during fluid-rock interaction, considering that about 76% Co and 26% Ni occur in pyrite and that Cu is almost completely controlled by chalcopyrite in the WE (Table 5-7).

In summary, all elements were at least redistributed during the metasomatic selvages formation as a consequence of the replacement character of the underlying mechanism (Putnis and Putnis, 2007; Putnis and John, 2010). In this context the transition metals and the trace elements (except the HFSE) were partly scavenged by the vein-forming fluid. This fluid seems to have been a Sr–Pb–S-bearing Ca-rich aqueous fluid phase that is represented now by the pyrite-bearing eclogitic vein. Figure 5-14 illustrates how the WE fluid-system (fluid-filled interconnected porosity) is thought to have interacted with the vein-forming fluid during the formation of the reaction selvage. The locally derived dehydration fluid (Fluid_1) interacted with the external transport fluid (Fluid_2) at the vein-WE contact and the decreasing chemical gradients in the fluid developed with time (Fig. 5-14). The mineral assemblage in the WE transformed due to changing bulk compositions at the reaction front, which had been progressively propagating into the WE thereby increasing the size of the selvage and fluid-rock interaction domain. Importantly, the material flux from the WE into the selvage most likely occurred via

138

5. Metal transport in subduction fluids diffusional transport as long as the vein was fluid filled (Fig. 5-14) which may have contributed not only to mineral growth and element redistributions (e.g., Widmer and Thompson, 2001) but also to the strong chemical gradients and ultimate element loss observed in the selvage closer to the vein (sample (e) in Fig. 5-15). Calcium, S, Sr, Pb and L(M)REE from the combined fluid system (Fluid_1 + Fluid_2) were captured, predominately by the abundant epidote and pyrite which had been formed by the replacement reaction during the selvage formation (e.g., Widmer and Thompson, 2001; Herms et al., 2012; Guo et al., 2012). However, significant amounts of LILE and HREE and transitional metal elements left the WE-selvage system due to the dissolution of white mica as well as dissolution-precipitation of garnet, omphacite, dolomite and sulfide.

5.7.3 Constraints on fluid sources

The metasomatic reaction selvage and related veins were either formed by internally-buffered fluids derived from the surrounding wall rocks (e.g., Philippot and Selverstone, 1991; Nadeau et al., 1993; Widmer and Thompson, 2001; Gao and Klemd, 2001; Spandler and Hermann, 2006) or externally (to the host rocks) derived fluids (e.g., John and Schenk, 2003; Molina et al., 2004; John et al., 2008, 2012; Beinlich et al., 2010), or a combination of both (e.g., Spandler et al., 2011; Herms et al., 2012). The blueschist-facies and eclogite-facies mineral assemblages preserved in the pyrite of the selvage (Fig. 5-5) strongly suggest that the selvage and thus the vein formed at the blueschist- to eclogite-facies transition. A limited retrograde alteration overprint is indicated by the rare occurrence of late stage barroisite, albite and chlorite surrounding the garnet and omphacite, which occur around or as inclusions in the mega-pyrite crystal. Prior to the fluid-rock interaction, the wall-rock blueschist was already partly converted to eclogite (Fig. 5-14a) and dehydration was progressively conducted triggering a continuous formation of a fluid-filled porosity (e.g., John et al., 2012). The similarity of the mineral assemblages and mineral compositions among the WE, the selvage and the vein also requires the combination of a significant amount of locally

139

5. Metal transport in subduction fluids derived material to maintain a similar effective bulk composition and identical P–T conditions during formation (selvage and vein) and equilibration (WE, selvage, and vein)

(e.g., Widmer and Thompson, 2001). In addition, neodymium isotopic data (εNd(t)) obtained for the selvage (0.39±0.24) and vein (0.29±0.24) are identical within error to that of the WE (0.75±0.24) (Table 5-3), which is also in agreement with a locally derived fluid or a fluid that had a similar source (seawater altered oceanic crust) or a combination of both. The infiltration of an external fluid is expected to cause a slight decrease of the εNd(t) value which is displayed by the decreasing values from the WE towards the vein.

However, the significant mass gains of Ca, Sr, Pb and S (Fig. 5-15) indicate that externally derived S- and Ca-rich fluids may have made a vital contribution to the metasomatism of the selvage. Even though evidence for local sources (WE) of the selvage and vein mineral-forming material was found, the massive mass gains and losses obtained during fluid-rock interaction was often interpreted to represent robust evidence for the influx of external fluids (Molina et al., 2004; Hermann et al., 2006; John et al., 2008; Beinlich et al., 2010; Spandler et al., 2011; Herms et al., 2012). The modal contents of both epidote and pyrite increase along a profile from the WE (ca. 5 vol.% and 1.5 vol.% respectively for (a)), the selvage (10 vol.% and 2 vol.% for (b), 20 vol.% and 4 vol.% for (c) and 25 vol.% and 5 vol.% for (d)), suggesting the infiltration of an externally derived S- and Ca-rich fluid (Fig. 5-14). Thus, the fluid involved in the fluid-rock interaction of the selvage may have been a mixture of internal-derived fluids released by dehydration reactions within the WE and an infiltrating external fluid.

In general, except from the locally derived and dehydration related fluid, three components are believed to contribute to the source of the vein-forming fluids namely metasediments, altered oceanic crust and the serpentinized portion of the underlying mantle in a subduction zone (e.g., John et al., 2004; Bebout, 2007; Beinlich et al., 2010; Spandler et al., 2011; Herms et al., 2012; Klemd, 2013). In the present study, the serpentinized part of the underlying mantle (Scambelluri et al., 1997; John and Schenk, 2003) as sole source rock for the metasomatising fluid is ruled out since the breakdown

140

5. Metal transport in subduction fluids of serpentine releases Mg, Cr and Ni (van der Straaten et al., 2008; Spandler et al., 2011) instead of Ca, Sr and Pb which are enriched in the reactive fluid phase (see above).

Considering the lack of LILE in the selvage and vein, large amounts of sediment-derived fluid addition to the vein-forming fluid can be excluded since LILE are relatively enriched in fluids derived from metasediments in comparison to other potential slab fluid sources (e.g., Plank and Langmuir, 1993). Furthermore, it is believed that the contribution of metasediments was not significant because subducted sediments usually have a pronounced negative neodymium value (εNd(t) < –5; Plank and Langmuir,

1998) which would have reduced the εNd(t) values of the selvage (under extreme fluid/rock ratios) and the vein. Thus a sole sedimentary source cannot explain the εNd(t) of the selvage and the vein which is identical to or slightly lower than that of the wall-rock eclogite. A possible metasedimentary contribution is thought to have been very limited and estimated to be less than 6% considering an εNd(t) value between –5.5 to –9.4 for reported sediments in the Tianshan subduction complex (Zeng et al., 2003).

Accordingly, a reasonable source for the Ca, Sr, Pb and S-bearing fluid seems to be dehydrating altered oceanic basaltic crust situated stratigraphically below the vein-hosting eclogite. Ca-rich fluids were also involved in the fluid-mediated eclogitization of selvages around HP veins in the Tianshan and Ecuador and are interpreted to have been dominantly derived from the dehydration of oceanic crustal metabasalts (Tianshan, Beinlich et al., 2010; John et al., 2012) with some contributions from deserpentinization and sedimentary components (Ecuador, Herms et al., 2012). Altered oceanic basalts can also contain considerable amounts of sulfides derived during seafloor alteration (Alt et al., 1989; Ohmoto, 1996), which may decompose and satisfy the S requirement for the Ca, Sr, Pb and S-bearing fluid released during prograde metamorphism. Thus the external fluid is considered to be largely derived from altered oceanic basaltic crust with little or without the influence of sedimentary fluid sources.

141

5. Metal transport in subduction fluids

5.7.4 Implications for the arc-related ore deposits

Up to now, several studies have been conducted to understand the links between subduction, arc magmatic processes and formation of arc-related ore deposits (e.g., Mungall, 2002; Sillitoe, 2008; Sun et al., 2004). However, the primary source(s) for the arc magmas, and therefore for arc-associated metal deposits, is still not well constrained since the characteristics of arc magmas may ascribe to multiple sources including subducting oceanic basaltic crust and subducted seafloor sediments (via dehydration fluids or melting), the asthenospheric mantle wedge between the subducting and overriding plates and the upper plate lithosphere (e.g., Richards, 2011). The subduction metasomatized mantle wedge above subduction slabs is considered to be the most important source for arc-associated ore deposits (e.g., Kepezhinskas et al., 2002; Pettke et al., 2010; Richards, 2011). However, there is some evidence for slab-derived fluid contributions of metal elements to the mantle wedge and volcanic magmas (e.g., Sillitoe, 1972; Hedenquist and Lowenstern, 1994; Stolper and Newman, 1994; Noll et al., 1996; McInnes et al., 1999; de Hoog et al., 2001a; Dale et. al., 2009). Furthermore some ore-forming fluids in the fore arc are considered to have been formed by prograde metamorphism of the subducting slab (Kerrich et al., 2005). Even though not much is known about the volumetric extent and efficiency of mobilization of slab-derived metals into the mantle wedge, fluid metasomatism clearly represents one viable mechanism for metal transfer into arc magma sources (cf. Richards, 2011). In the present study, the transition metals including Fe, Cu, Ni, Zn, Co and Mn have been depleted by 10% to 40% in the selvage according to the mass balance calculation (Fig. 5-15) indicating that both siderophile (Fe, Co, Ni and Mn) and chalcophile (Cu and Zn) elements were released during the metasomatic selvage formation into the vein-forming fluid. Such siderophile and chalcophile element-bearing fluids may ascend upwards towards the mantle wedge along major conduits of slab fluid release (Zack and John, 2007; John et al., 2012), fertilize the overlying mantle wedge, trigger partial melting and finally access the metal source to arc magmas and arc-related ore deposits. Thus, the pyrite-bearing HP vein and related host lawsonite-bearing eclogites indicate that ore-producing metals (together

142

5. Metal transport in subduction fluids with the LILE and HREE) can be mobilized by subduction-zone fluids, thereby confirming that some metals were transferred from subducting slabs into the arc magma source (e.g., Hedenquist and Lowenstern, 1994; Stolper and Newman, 1994; McInnes et al., 1999; Kerrich, 2005; Groves et al., 2010; Richards, 2011). This conclusion is further supported by the occurrence of VMS-type Pb–Zn deposits (e.g., Keketale deposit in the Altai, Wan et al., 2010), Alaskan-type Cu–Ni–Co deposits (e.g., Xiangshan deposit in the Tianshan, Han et al., 2010) and magmatic Fe deposits (e.g., Zhibo deposit in the Tianshan, Zhang et al., 2012) in arc settings in the Central Asia Orogenic Belt which has undergone multi-subductional and accretionary orogenic processes (Xiao et al., 2010).

5.8 Conclusions

1. An externally derived fluid caused Ca–S metasomatism in the immediate lawsonite-bearing wall rock eclogite in a cold subduction zone during prograde to the peak eclogite-facies metamorphism at 580–590 ºC and 22.5–23.5 kbar.

2. Petrological and geochemical data imply that two distinct sources supplied fluids and elements that led to the metasomatic selvage formation and the precipitation of the vein: 1) an internal source derived from the breakdown of hydrous minerals (e.g., lawsonite, glaucophane and white mica) in the wall rock; and 2) an external S-bearing Ca-rich fluid mainly derived from the dehydration of altered oceanic basalts from lower parts of the descending slab.

3. Both, transition metal and trace elements (except HFSE) are mobilized during the fluid-rock interaction leading to the formation of the pyrite-bearing eclogitic vein. The LILE (K, Rb, Ba) were most effectively scavenged from the selvage (depleted up to 100%). The HREE and transition metals are significantly mobilized with a mass loss ranging from 10% to 40%. However, Sr and Pb were enriched up to ~200%, and LREE and MREE show considerable enrichments of 0 to 40% as well.

4. The formation of the vein-selvage system is related to mineral dissolution-precipitation processes. The wall-rock eclogite minerals along the

143

5. Metal transport in subduction fluids

pathway are dissolved due to chemical changes in the selvage induced by the external fluid, while the selvage minerals precipitated with adjusted abundances and compositions at pre-peak to peak P–T conditions. Significant element gains and losses occurred during their dissolution-precipitation processes, which are related to the chemical differences between the internally derived dehydration fluid in the wall-rock eclogite and the externally derived vein-forming fluid.

5. The dissolution of white mica in the selvage caused LILE (K, Rb, Cs and Ba) loss and along with the dissolution-precipitation behavior of garnet, omphacite, dolomite and sulfides during fluid-rock interaction considerable amounts of transition metals were released into the passing fluid. The dissolution of garnet may have also released substantial amounts of HREE into the fluid.

6. The suite of elements mobilized during fluid-rock interaction induced by the dehydration of subducted slabs in subduction zones not only resembles that of island arc magmas which have chemical characteristics of strong LILE enrichment and strong HFSE depletion, but also indicates that the slab-derived fluids may be effective in transporting metals (e.g., Fe, Cu, Ni and Zn) from the slab into the overlying mantle wedge and the arc crust. This suggests a recycling behavior of ore-forming metals at subduction zones and provides an alternative source to the widely believed asthenospheric mantle magma source for metal enrichment in arc magmas and arc-related ore deposits as was proposed by former workers.

Acknowledgements

This study was funded by the National Natural Science Foundation of China (41025008, 41172066), the State Key Project for Basic Research of China (2007CB411302), and the Deutsche Forschungsgemeinschaft (KL 692/17–2, 3; 23-1). We are grateful to T.P. Zhao for help with XRF and ICP–MS whole-rock analyses, Q. Mao, Y.G. Ma and M. Meyer for help with the electron microprobe analyses, and H. Brätz for assistance with the LA–ICP–MS. The first author thanks DAAD for the scholarship supporting his PhD

144

5. Metal transport in subduction fluids study in Germany. Thorough and constructive reviews by S. Penniston-Dorland and two other anonymous reviewers are highly appreciated and clearly helped to improve the manuscript. M. Norman is thanked for his editorial help. This publication is a contribution to IGCP Project 592.

References

Ague J.J. (1991) Evidence for major mass-transfer and volume strain during regional metamorphism of pelites. Geology 19, 855–858. Ague J.J. (2003) Fluid infiltration and transport of major, minor, and trace elements during regional metamorphism of carbonate rocks, Wepawaug Schist, Connecticut, USA. American Journal of Science 303, 753–816. Ague J.J. (2007) Models of permeability contrasts in subduction zone melange: Implications for gradients in fluid fluxes, Syros and Tinos Islands, Greece. Chemical Geology 239, 217–227. Ague J.J. (2011) Extreme channelization of fluid and the problem of element mobility during Barrovian metamorphism. American Mineralogist 96, 333–352. Alt J.C., Anderson T.F. and Bonnell L. (1989) The geochemistry of sulfur in a 1.3 km section of hydrothermally altered oceanic-crust, DSDP-Hole-504B. Geochimica et Cosmochimica Acta 53, 1011–1023. Alt J.C., Shanks W.C. and Jackson M.C. (1993) Cycling of sulfur in subduction zones: the geochemistry of sulfur in the Mariana-island arc and back-arc trough. Earth and Planetary Science Letters 119, 477–494. Austrheim H. (1987) Eclogitization of lower crustal granulites by fluid migration through shear zones. Earth and Planetary Science Letters 81, 221–232. Bebout G.E. (2007) Metamorphic chemical geodynamics of subduction zones. Earth and Planetary Science Letters 260, 373–393. Becker H., Jochum K.P. and Carlson R.W. (1999) Constraints from high-pressure veins in eclogites on the composition of hydrous fluids in subduction zones. Chemical Geology 160, 291–308. Beinlich A., Klemd R., John T. and Gao J. (2010) Trace-element mobilization during Ca-metasomatism along a major fluid conduit: eclogitization of blueschist as a consequence of fluid–rock interaction. Geochimica et Cosmochimica Acta 74, 1892– 1922. Bouman C., Elliott T. and Vroon P.Z. (2004) Lithium inputs to subduction zones. Chemical Geology 212, 59–79. Carlson W.D. (2002) Scales of disequilibrium and rates of equilibration during metamorphism. American Mineralogist 87, 185–204. Chakraborty S. and Ganguly J. (1992) Cation diffusion in aluminosilicate garnets: experimental determination in spessartine-almandine diffusion couples, evaluation of effective binary diffusion coefficients, and applications. Contributions to Mineralogy and Petrology 111, 74–86.

145

5. Metal transport in subduction fluids

Chen Y., Ye K., Wu T.F. and Guo S. (2013) Exhumation of oceanic eclogites: thermodynamic constraints on pressure, temperature, bulk composition and density. Journal of Metamorphic Geology 31, 549-570. Chu Z.Y., Wu F.Y., Walker R.J., Rudnick R.L., Pitcher L., Puchtel I.S., Yang Y.H. and Wilde S.A. (2009) Temporal Evolution of the Lithospheric Mantle beneath the Eastern North China Craton. Journal of Petrology 50, 1857–1898. Coggon R. and Holland T.J.B. (2002) Mixing properties of phengitic micas and revised garnet–phengite thermobarometers. Journal of Metamorphic Geology 20, 683–696. Connolly J.A.D. (1990) Multivariable phase-diagrams: an algorithm based on generalized thermodynamics. American Journal of Science 290, 666–718. Connolly J.A.D. (2005) Computation of phase equilibria by linear programming: A tool for geodynamic modelling and its application to subduction zone decarbonation. Earth and Planetary Science Letters 236, 524–541. Connolly J.A.D. and Trommsdorff V. (1991) Petrogenetic grids for metacarbonate rocks: pressure–temperature phase-diagram projection for mixed-volatile systems. Contributions to Mineralogy and Petrology 108, 93–105. Dale C.W., Burton K.W., Pearson D.G., Gannoun A., Alard O., Argles T.W. and Parkinson I.J. (2009) Highly siderophile element behaviour accompanying subduction of oceanic crust: whole rock and mineral-scale insights from a high-pressure terrain. Geochimica et Cosmochimica Acta 73, 1394–1416. de Hoog J.C.M., Mason P.R.D. and van Bergen M.J. (2001a) Sulfur and chalcophile elements in subduction zones: Constraints from a laser ablation ICP–MS study of melt inclusions from Galunggung Volcano, Indonesia. Geochimica et Cosmochimica Acta 65, 3147–3164. de Hoog J.C.M., Taylor B.E. and van Bergen M.J. (2001b) Sulfur isotope systematics of basaltic lavas from Indonesia: implications for the sulfur cycle in subduction zones. Earth and Planetary Science Letters 189, 237–252. Diener J.F.A., Powell R., White R.W. and Holland T.J.B. (2007) A new thermodynamic

model for clino- and orthoamphiboles in the system Na2O–CaO–FeO–MgO–Al2O3– SiO2–H2O–O. Journal of Metamorphic Geology 25, 631–656. Dragovic B., Samanta L.M., Baxter E.F. and Selverstone J. (2012) Using garnet to constrain the duration and rate of water-releasing metamorphic reactions during subduction: An example from Sifnos, Greece. Chemical Geology 314, 9–22. Dupuy C., Barsczus H.G., Liotard J.M. and Dostal J. (1988) Trace element evidence for the origin of ocean island basalts: an example from the Austral Islands (French Polynesia). Contributions to Mineralogy and Petrology 98, 293–302. Evans B.W. (1990) Phase relations of epidote-blueschists. Lithos 25, 3–23. Evans T.P. (2004) A method for calculating effective bulk composition modification due to crystal fractionation in garnet-bearing schist: implications for isopleth thermobarometry. Journal of Metamorphic Geology 22, 547–557. Feineman M.D., Ryerson F.J., DePaolo D.J. and Plank T. (2007) Zoisite-aqueous fluid trace element partitioning with implications for subduction zone fluid composition. Chemical Geology 239, 250–265. Ganguly J., Cheng W.J. and Chakraborty S. (1998) Cation diffusion in aluminosilicate

146

5. Metal transport in subduction fluids

garnets: experimental determination in pyrope-almandine diffusion couples. Contributions to Mineralogy and Petrology 131, 171–180. Gao J. and Klemd R. (2001) Primary fluids entrapped at blueschist to eclogite transition: evidence from the Tianshan meta-subduction complex in northwestern China. Contributions to Mineralogy and Petrology 142, 1–14. Gao J. and Klemd R. (2003) Formation of HP–LT rocks and their tectonic implications in the western Tianshan Orogen, NW China: geochemical and age constraints. Lithos 66, 1–22. Gao J., He G.Q., Li M.S., Xiao X.C., Tang Y.Q., Wang J. and Zhao M. (1995) The mineralogy, petrology, metamorphic PTDt trajectory and exhumation mechanism of blueschists, south Tianshan, northwestern China. Tectonophysics 250, 151–168. Gao J., Klemd R., Zhang L., Wang Z. and Xiao, X. (1999) P–T path of high-pressure/low-temperature rocks and tectonic implications in the western Tianshan Mountains, NW China. Journal of Metamorphic Geology 17, 621–636. Gao J., John T., Klemd R. and Xiong X.M. (2007) Mobilization of Ti–Nb–Ta during subduction: Evidence from rutile-bearing dehydration segregations and veins hosted in eclogite, Tianshan, NW China. Geochimica et Cosmochimica Acta 71, 4974–4996. Gao J., Long L.L., Klemd R., Qian Q., Liu D.Y., Xiong X.M., Su W., Liu W., Wang Y.T. and Yang F.Q. (2009) Tectonic evolution of the South Tianshan orogen and adjacent regions, NW China: geochemical and age constraints of granitoid rocks. International Journal of Earth Sciences 98, 1221–1238. Gao J., Klemd R., Qian Q., Zhang X., Li J.L. and Jiang T. (2011) The collision between the Yili and Tarim blocks of the Southwestern Altaids: Geochemical and age constraints of a leucogranite dike crosscutting the HP–LT metamorphic belt in the Chinese Tianshan Orogen. Tectonophysics 499, 118–131. Gill J. B. (1981) Orogenic andesites and plate tectonics. Springer Verlag, Berlin. Glasby G.P. (1996) The relation between ore-deposit formation and subduction: Mass balance considerations. Island Arc 5, 396–406. Green E., Holland T. and Powell R. (2007) An order-disorder model for omphacitic pyroxenes in the system jadeite–diopside–hedenbergite–acmite, with applications to eclogitic rocks. American Mineralogist 92, 1181–1189. Gresens R.L. (1967) Composition-volume relationships of metasomatism. Chemical Geology 2, 47–55. Groves D.I., Bierlein F.P., Meinert L.D. and Hitzman M.W. (2010) Iron oxide Copper- Gold (IOCG) Deposits through Earth History: Implications for Origin, Lithospheric Setting, and Distinction from other epigenetic Iron Oxide Deposits. Economic Geology 105, 641–654. Guo S., Ye K., Chen Y., Liu J., Mao Q. and Ma Y. (2012) Fluid–rock interaction and element mobilization in UHP metabasalt: Constraints from an omphacite-epidote vein and host eclogites in the Dabie orogen. Lithos 136, 145–167. Han B.F., He G.Q., Wang X.C. and Guo Z.J. (2011) Late Carboniferous collision between the Tarim and Kazakhstan-Yili terranes in the western segment of the South Tian Shan Orogen, Central Asia, and implications for the Northern Xinjiang, western China. Earth-Science Reviews 109, 74–93.

147

5. Metal transport in subduction fluids

Han C., Xiao W., Zhao G., Ao S., Zhang J., Qu W. and Du A. (2010) In-situ U–Pb, Hf and Re–Os isotopic analyses of the Xiangshan Ni–Cu–Co deposit in Eastern Tianshan (Xinjiang), Central Asia Orogenic Belt Constraints on the timing and genesis of the mineralization. Lithos 120, 547–562. Hawkesworth C.J., Gallagher K., Hergt J.M. and McDermott F. (1993) Mantle and slab contributions in arc magmas. Annual Review of Earth and Planetary Sciences 21, 175–204. Hedenquist J.W. and Lowenstern J.B. (1994) The role of magmas in the formation of hydrothermal ore deposits. Nature 370, 519–527. Hegner E., Klemd R., Kröner A., Corsini M., Alexeiev D.V., Iaccheri L.M., Zack T., Dulski P., Xia X. and Windley B.F. (2010) Mineral ages and P–T conditions of Late Paleozoic high-pressure eclogite and provenance of melange sediments from Atbashi in the south Tianshan orogen of Kyrgyzstan. American Journal of Science 310, 916– 950. Hermann J., Spandler C., Hack A. and Korsakov A.V. (2006) Aqueous fluids and hydrous melts in high-pressure and ultra-high pressure rocks: Implications for element transfer in subduction zones. Lithos 92, 399–417. Herms P., John T., Bakker R. J. and Schenk V. (2012) Evidence for channelized external fluid flow and element transfer in subducting slabs (Raspas Complex, Ecuador). Chemical Geology 310, 79–96. Higashino F., Kawakami T., Satish-Kumar M., Ishikawa M., Maki K., Tsuchiya N., Grantham G.H. and Hirata T. (2013) Chlorine-rich fluid or melt activity during granulite facies metamorphism in the Late Proterozoic to Cambrian continental collision zone—An example from the Sør Rondane Mountains, East Antarctica. Precambrian Research 234, 229–246. Holland T.J.B. and Powell R. (1998) An internally consistent thermodynamic data set for phases of petrological interest. Journal of Metamorphic Geology 16, 309–343. Jahn B.M., Wu F.Y. and Chen B. (2000) Granitoids of the Central Asian Orogenic Belt and continental growth in the Phanerozoic. Transactions of the Royal Society of Edinburgh: Earth Sciences 91, 181–193. John T. and Schenk V. (2003) Partial eclogitisation of gabbroic rocks in a late Precambrian subduction zone (Zambia): prograde metamorphism triggered by fluid infiltration. Contributions to Mineralogy and Petrology146, 174–191. John T., Schenk V., Haase K., Scherer E. and Tembo F. (2003) Evidence for a Neoproterozoic ocean in south-central Africa from mid-oceanic-ridge-type geochemical signatures and pressure-temperature estimates of Zambian eclogites. Geology 31, 243–246. John T., Scherer E.E., Haase K. and Schenk V. (2004) Trace element fractionation during fluid-induced eclogitization in a subducting slab: trace element and Lu–Hf– Sm–Nd isotope systematics. Earth and Planetary Science Letters 227, 441–456. John T., Klemd R., Gao J. and Garbe-Schonberg C.D. (2008) Trace-element mobilization in slabs due to non steady-state fluid-rock interaction: Constraints from an eclogite-facies transport vein in blueschist (Tianshan, China). Lithos 103, 1–24. John T., Gussone N., Podladchikov Y.Y., Bebout G. E., Dohmen R., Halama R., Klemd

148

5. Metal transport in subduction fluids

R., Magna T. and Seitz H.M. (2012) Volcanic arcs fed by rapid pulsed fluid flow through subducting slabs. Nature Geoscience 5, 489–492. Kepezhinskas P., Defant M.J. and Widom E. (2002) Abundance and distribution of PGE and Au in the island-arc mantle: implications for sub-arc metasomatism. Lithos 60, 113–128. Kerrich R., Goldfarb R. and Richards J.P. (2005) Metallogenic provinces in an evolving geodynamic framework. In Economic Geology and the Bulletin of the Society of Economic Geologists, One Hundredth Anniversary Volume 1905–2005 (eds. J. W. Hedenquist, J. F. H. Thompson, R. J. Goldfarb and J. P. Richards). pp. 1097–1136. Klemd R., Schröter F.C., Will T.M. and Gao J. (2002) P–T evolution of glaucophane-omphacite bearing HP–LT rocks in the western Tianshan Orogen, NW China: new evidence for 'Alpine-type' tectonics. Journal of Metamorphic Geology 20, 239–254. Klemd R., Bröcker M., Hacker B.R., Gao J., Gans P. and Wemmer K. (2005) New age constraints on the metamorphic evolution of the high-pressure/low-temperature belt in the western Tianshan mountains, NW China. Journal of Geology 113, 157–168. Klemd R., John T., Scherer E.E., Rondenay S. and Gao J. (2011) Changes in dip of subducted slabs at depth: Petrological and geochronological evidence from HP–UHP rocks (Tianshan, NW–China). Earth and Planetary Science Letters 310, 9–20. Klemd R. (2013) Metasomatism during high-pressure metamorphism: eclogites and blueschist-facies rocks. In: Metasomatism and the Chemical Transformation of Rocks (eds. D. E. Harlov and H. Austrheim). Springer-Verlag, Berlin Heidelberg. pp. 351– 413. Konrad–Schmolke M., O'Brien P.J., de Capitani C. and Carswell D.A. (2008) Garnet growth at high- and ultra-high pressure conditions and the effect of element fractionation on mineral modes and composition. Lithos 103, 309–332. Konrad–Schmolke M., O'Brien P.J. and Zack T. (2011) Fluid Migration above a Subducted Slab–Constraints on Amount, Pathways and Major Element Mobility from Partially Overprinted Eclogite-facies Rocks (Sesia Zone, Western Alps). Journal of Petrology 52, 457–486. Labrousse L., Hetenyi G., Raimbourg H., Jolivet L. and Andersen T.B. (2010) Initiation of crustal-scale thrusts triggered by metamorphic reactions at depth: Insights from a comparison between the Himalayas and Scandinavian Caledonides. Tectonics 29, TC5002, doi:10.1029/2009TC002602. Leake B.E., Woolley A.R., Arps C.E.S., Birch W.D., Gilbert M.C., Grice J.D., Hawthorne F.C., Kato A., Kisch H.J., Krivovichev V.G., Linthout K., Laird J., Mandarino J.A., Maresch W.V., Nickel E.H., Rock N.M.S., Schumacher J.C., Smith D.C., Stephenson N.C.N., Ungaretti L., Whittaker E.J.W. and Guo Y.Z. (1997) Nomenclature of amphiboles: Report of the subcommittee on amphiboles of the International Mineralogical Association, commission on new minerals and mineral names. American Mineralogist 82, 1019–1037. Li J.L., Klemd R., Gao J. and Meyer M. (2012) Coexisting carbonate-bearing eclogite and blueschist in SW Tianshan, China: petrology and phase equilibria. Journal of Asian Earth Sciences 60, 174–187.

149

5. Metal transport in subduction fluids

Li J.L., Klemd R., Gao J. and Meyer M. (2013) Compositional zoning in dolomite from lawsonite-bearing eclogite (SW Tianshan, China): Evidence for prograde metamorphism during subduction of oceanic crust. American Mineralogist 99, 206– 217. Long L., Gao J., Klemd R., Beier C., Qian Q., Zhang X., Wang J. and Jiang T. (2011) Geochemical and geochronological studies of granitoid rocks from the Western Tianshan Orogen: Implications for continental growth in the southwestern Central Asian Orogenic Belt. Lithos 126, 321–340. Lü Z., Zhang L.F., Du J.X. and Bucher K. (2008) Coesite inclusions in garnet from eclogitic rocks in western Tianshan, northwest China: Convincing proof of UHP metamorphism. American Mineralogist 93, 1845–1850. Lü Z., Zhang L., Du J. and Bucher K. (2009) Petrology of coesite-bearing eclogite from Habutengsu Valley, western Tianshan, NW China and its tectonometamorphic implication. Journal of Metamorphic Geology 27, 773–787. Manning C.E. (1998) Fluid composition at the blueschist - eclogite transition in the

model system Na2O–MgO–Al2O3–SiO2–H2O–HCl. Swiss Bulletin of Mineralogy and Petrology 78, 225–242. Manning C.E. (2004) The chemistry of subduction-zone fluids. Earth and Planetary Science Letters 223, 1–16. Marschall H.R., Altherr R., Ludwig T., Kalt A., Gmeling K. and Kasztovszky Z. (2006) Partitioning and budget of Li, Be and B in high-pressure metamorphic rocks. Geochimica et Cosmochimica Acta 70, 4750–4769. McCulloch M.T. and Gamble J.A. (1991) Geochemical and geodynamical constraints on subduction zone magmatism. Earth and Planetary Science Letters 102, 358–374. McInnes B.I.A., McBride J.S., Evans N.J., Lambert D.D. and Andrew A.S. (1999) Osmium isotope constraints on ore metal recycling in subduction zones. Science 286, 512–516. Metrich N., Schiano P., Clocchiatti R. and Maury R.C. (1999) Transfer of sulfur in subduction settings: an example from Batan Island (Luzon volcanic arc, Philippines). Earth and Planetary Science Letters 167, 1–14. Miller S.A., van der Zee W., Olgaard D.L.and Connolly J.A.D. (2003) A fluid-pressure feedback model of dehydration reactions: experiments, modelling, and application to subduction zones. Tectonophysics 370, 241–251. Molina J.F., Poli S., Austrheim H., Glodny J. and Rusin A. (2004) Eclogite-facies vein systems in the Marun-Keu complex (Polar Urals, Russia): textural, chemical and thermal constraints for patterns of fluid flow in the lower crust. Contributions to Mineralogy and Petrology 147, 484–504. Morimoto N., Fabries J., Ferguson A.K., Ginzburg I.V., Ross M., Seifert F.A., Zussman J., Aoki K. and Gottardi G. (1988) Nomenclature of pyroxenes. American Mineralogist 73, 1123–1133. Mungall J.E. (2002) Roasting the mantle: Slab melting and the genesis of major Au and Au-rich Cu deposits. Geology 30, 915–918. Nadeau S., Philippot P. and Pineau F. (1993) Fluid inclusion and mineral isotopic compositions (HCO) in eclogitic rocks as tracers of local fluid migration during

150

5. Metal transport in subduction fluids

high-pressure metamorphism. Earth and Planetary Science Letters 114, 431–448. Noll P.D., Newsom H.E., Leeman W.P. and Ryan J.G. (1996) The role of hydrothermal fluids in the production of subduction zone magmas: Evidence from siderophile and chalcophile trace elements and boron. Geochimica et Cosmochimica Acta 60, 587– 611. Ohmoto H. (1996) Formation of volcanogenic massive sulfide deposits: The Kuroko perspective. Ore Geology Reviews 10, 135–177. Osbahr I., Klemd R., Oberthür T., Brätz H. and Schouwstra R. (2012) Platinum-group element distribution in base-metal sulfides of the Merensky Reef from the eastern and western Bushveld Complex, South Africa. Mineralium Deposita 48, 211–232. Peacock S.M. (1993) The importance of blueschist - eclogite dehydration reactions in subducting oceanic-crust. Geological Society of America Bulletin 105, 684–694. Penniston-Dorland S.C. and Ferry J.M. (2008) Element mobility and scale of mass transport in the formation of quartz veins during regional metamorphism of the Waits River Formation, east-central Vermont. American Mineralogist 93, 7–21. Penniston-Dorland S.C., Bebout G.E., Pogge von Strandmann P.A.E., Elliott T. and Sorensen S.S. (2012) Lithium and its isotopes as tracers of subduction zone fluids and metasomatic processes: Evidence from the Catalina Schist, California, USA. Geochimica et Cosmochimica Acta 77, 530–545. Pettke T., Oberli F. and Heinrich C.A. (2010) The magma and metal source of giant porphyry-type ore deposits, based on lead isotope microanalysis of individual fluid inclusions. Earth and Planetary Science Letters 296, 267–277. Pfänder J. A., Münker C., Stracke A. and Mezger K. (2007) Nb/Ta and Zr/Hf in ocean island basalts – Implications for crust-mantle differentiation and the fate of Niobium. Earth and Planetary Science Letters 254, 158–172. Philippot P. and Selverstone J. (1991) Trace-element-rich brines in eclogitic veins – implications for fluid composition and transport during subduction. Contributions to Mineralogy and Petrology 106, 417–430. Plank T. and Langmuir C.H. (1993) Tracing trace-elements from sediment input to volcanic output at subduction zones. Nature 362: 739–743. Plank T. and Langmuir C.H. (1998) The chemical composition of subducting sediment and its consequences for the crust and mantle. Chemical Geology 145, 325–394. Putnis A. and John T. (2010) Replacement Processes in the Earth's Crust. Elements 6, 159–164. Putnis A. and Putnis C.V. (2007) The mechanism of reequilibration of solids in the presence of a fluid phase. Journal of Solid State Chemistry 180, 1783–1786. Rapp J.F., Klemme S., Butler I.B. and Harley S.L. (2010) Extremely high solubility of rutile in chloride and fluoride-bearing metamorphic fluids: An experimental investigation. Geology 38, 323–326. Richards J.P. (2011) Magmatic to hydrothermal metal fluxes in convergent and collided margins. Ore Geology Reviews 40, 1–26. Rubatto D. and Hermann J. (2003) Zircon formation during fluid circulation in eclogites (Monviso, Western Alps): Implications for Zr and Hf budget in subduction zones. Geochimica et Cosmochimica Acta 67, 2173–2187.

151

5. Metal transport in subduction fluids

Rubin J.N., Henry C.D. and Price J.G. (1993) The mobility of zirconium and other ―immobile‖ elements during hydrothermal alteration. Chemical Geology 110, 29–47. Ryan J.G., Morris J., Tera F., Leeman W.P. and Tsvetkov A. (1995) Cross-arc geochemical variations in the Kurile arc as a function of slab depth. Science 270, 625–627. Sawkins F.J. (1972) Sulfide ore-deposits in relation to plate tectonics. Journal of Geology 80, 377–397. Scambelluri M. and Philippot P. (2001) Deep fluids in subduction zones. Lithos 55, 213–227. Scambelluri M., Piccardo G.B., Philippot P., Robbiano A. and Negretti L. (1997) High salinity fluid inclusions formed from recycled seawater in deeply subducted alpine serpentinite. Earth and Planetary Science Letters 148, 485–499. Schmidt A., Weyer S., John T. and Brey G.P. (2009) HFSE systematics of rutile-bearing eclogites: New insights into subduction zone processes and implications for the earth's HFSE budget. Geochimica et Cosmochimica Acta 73, 455–468. Schmidt M.W. and Poli S. (1998) Experimentally based water budgets for dehydrating slabs and consequences for arc magma generation. Earth and Planetary Science Letters 163, 361–379. Schulz B., Klemd R. and Brätz H. (2006) Host rock compositional controls on zircon trace-element signatures in metabasites from the Austroalpine basement. Geochimica et Cosmochimica Acta 70, 697–710. Sillitoe R.H. (1972) Relation of Metal Provinces in Western America to Subduction of Oceanic Lithosphere. Geological Society of America Bulletin 83, 813–818. Sillitoe R.H. (2008) Major gold deposits and belts of the North and South American Cordillera: Distribution, tectonomagmatic settings, and metallogenic considerations. Economic Geology 103, 663–687. Sorensen S.S. and Grossman J.N. (1989) Enrichment of trace elements in garnet mphibolites from a paleo-subduction zone: Catalina Schist, southern California. Geochimica et Cosmochimica Acta 53, 3155–3178. Spandler C. and Hermann J. (2006) High-pressure veins in eclogite from New Caledonia and their significance for fluid migration in subduction zones. Lithos 89, 135–153. Spandler C., Hermann J., Arculus R. and Mavrogenes J. (2003) Redistribution of trace elements during prograde metamorphism from lawsonite blueschist to eclogite facies; implications for deep subduction-zone processes. Contributions to Mineralogy and Petrology 146, 205–222. Spandler C., Hermann J., Arculus R. and Mavrogenes J. (2004) Geochemical heterogeneity and element mobility in deeply subducted oceanic crust; insights from high-pressure mafic rocks from New Caledonia. Chemical Geology 206, 21–42. Spandler C., Pettke T. and Rubatto D. (2011) Internal and External Fluid Sources for Eclogite-facies Veins in the Monviso Meta-ophiolite, Western Alps: Implications for Fluid Flow in Subduction Zones. Journal of Petrology 52, 1207–1236. Stalder R., Foley S.F., Brey G.P. and Horn I. (1998) Mineral aqueous fluid partitioning of trace elements at 900–1200 degrees C and 3.0–5.7 GPa: New experimental data

152

5. Metal transport in subduction fluids

for garnet, clinopyroxene, and rutile, and implications for mantle metasomatism. Geochimica et Cosmochimica Acta 62, 1781–1801. Stolper E. and Newman S. (1994) The role of water in the petrogenesis of Marina trough magmas. Earth and Planetary Science Letters 121, 293–325. Stolz A.J., Jochum K.P., Spettle B. and Hofmann A.W. (1996) Fluid- and melt-related enrichment in the subarc mantle: Evidence from Nb/Ta variations in island-arc basalts. Geology 24, 587–590. Sun S.S. and McDonough W.F. (1989) Chemical and isotopic systematics of oceanic basalts: implications for mantle composition and processes. In Magmatism in the Ocean Basins (eds. A.D. Saunders and M. J. Norry). Blackwell, London. pp. 313– 345. Sun S.S., Nesbitt R.W. and Sharaskin A.Y. (1979) Geochemical characteristics of mid-ocean ridge basalts. Earth and Planetary Science Letters 44, 119–138. Sun W.D., Arculus R.J., Kamenetsky V.S. and Binns R.A. (2004) Release of gold-bearing fluids in convergent margin magmas prompted by magnetite crystallization. Nature 431, 975–978. Tatsumi Y. (1989) Migration of fluid phases and genesis of basalt magmas in subduction zones. Journal of Geophysical Research 94, 4697–4707. Tribuzio R., Messiga B., Vannucci R. and Bottazzi P. (1996) Rare earth element redistribution during high-pressure-low-temperature metamorphism in ophiolitic Fe-gabbros (Liguria, northwestern Italy): Implications for light REE mobility in subduction zones. Geology 24, 711–714. van Achterbergh E., Ryan C.G. and Griffin W.L. (2000) GLITTER: on-line interactive data reduction for the Laser Ablation ICP–MS Microprobe, Ninth Annual V. M. Goldschmidt Conference, Cambridge. 7215. van der Straaten F., Schenk V., John T. and Gao J. (2008) Blueschist-facies rehydration of eclogites (Tian Shan, NW–China): Implications for fluid-rock interaction in the subduction channel. Chemical Geology 255, 195–219. Volkova N.I. and Budanov V.I. (1999) Geochemical discrimination of metabasalt rocks of the Fan-Karategin transitional blueschist/greenschist belt, South Tianshan, Tajikistan: seamount volcanism and accretionary tectonics. Lithos 47, 201–216. Wan B., Zhang L. and Xiao W. (2010) Geological and geochemical characteristics and ore genesis of the Keketale VMS Pb–Zn deposit, Southern Altai Metallogenic Belt, NW China. Ore Geology Reviews 37, 114–126. Wang B., Jahn B.M., Lo C.H., Shu L.S., Wu C.Y., Li K.S. and Wang F. (2011) Structural analysis and (40)Ar/(39)Ar thermochronology of Proterozoic rocks in Sailimu area (NW China): Implication to polyphase tectonics of the North Chinese Tianshan. Journal of Asian Earth Sciences 42, 839–853. Wei C.J., Powell R. and Zhang L.F. (2003) Eclogites from the south Tianshan, NW

153

5. Metal transport in subduction fluids

China: Phase Equilibria and P–T Path. Journal of Petrology 50, 1973–1991. Widmer T. and Thompson A.B. (2001) Local origin of high pressure vein material in eclogite facies rocks of the Zermatt-Saas Zone, Switzerland. American Journal of Science 301, 627–656. Wohlgemuth–Ueberwasser C.C., Ballhaus C., Berndt J., Paliulionyte V.S.N. and Meisel T. (2007) Synthesis of PGE sulfide standards for laser ablation inductively coupled plasma mass spectrometry (LA–ICP–MS). Contributions to Mineralogy and Petrology 154, 607–617. Xiao W.J., Huang B.C., Han C.M., Sun S. and Li J.L. (2010) A review of the western part of the Altaids: A key to understanding the architecture of accretionary orogens. Gondwana Research18, 253–273. Xiao W.J., Windley B.F., Huang B.C., Han C.M., Yuan C., Chen H.L., Sun M., Sun S. and Li J.L. (2009) End-Permian to mid-Triassic termination of the accretionary processes of the southern Altaids: implications for the geodynamic evolution, Phanerozoic continental growth, and metallogeny of Central Asia. International Journal of Earth Sciences 98, 1189–1217. Xiao Y.L., Sun W.D., Hoefs J., Simon K., Zhang Z.M., Li S.G. and Hofmann A. W. (2006) Making continental crust through slab melting: Constraints from niobium-tantalum fractionation in UHP metamorphic rutile. Geochimica et Cosmochimica Acta 70, 4770–4782. Yardley B.W.D. (1986) Is there water in the deep continental crust. Nature 323, 111– 111. Zack T. and John T. (2007) An evaluation of reactive fluid flow and trace element mobility in subducting slabs. Chemical Geology 239, 199–216. Zack T., Rivers T. and Foley S.F. (2001) Cs–Rb–Ba systematics in phengite and amphibole: an assessment of fluid mobility at 2.0 GPa in eclogites from Trescolmen, Central Alps. Contributions to Mineralogy and Petrology 140, 651–669. Zeng L., Saleeby J. B. and Gao J. (2003) Sr and Nd isotopic constraints on the protoliths of the Chinese Tianshan UHP metamorphic complex, West China. Eos, Trans. AGU, 84(47), Fall Meet. Suppl., F1305 (Abstr.). Zhang X., Tian J., Gao J., Klemd R., Dong L., Fan J., Jiang T., Hu C. and Qian Q. (2012) Geochronology and geochemistry of granitoid rocks from the Zhibo syngenetic volcanogenic iron ore deposit in the Western Tianshan Mountains (NW–China): Constraints on the age of mineralization and tectonic setting. Gondwana Research 22, 585–596. Zhang Z.M., Shen K., Sun W.D., Liu Y.S., Liou J.G., Shi C. and Wang J.L. (2008) Fluids in deeply subducted continental crust: Petrology, mineral chemistry and fluid inclusion of UHP metamorphic veins from the Sulu orogen, eastern China. Geochimica et Cosmochimica Acta 72, 3200–3228.

154

6. A common P-T evolution of eclogites and metasediments

6 A common high-pressure metamorphic evolution of interlayered eclogites and metasediments from the ‘ultrahigh-pressure unit’ of the Tianshan metamorphic belt in China

Ji-Lei Li1, Reiner Klemd1, Jun Gao2, Tuo Jiang2 and Yun-Hui Song3

1GeoZentrum Nordbayern, Universität Erlangen–Nürnberg, Schlossgarten 5a, D-91054 Erlangen, Germany ([email protected]) 2Key Laboratory of Mineral Resources, Institute of Geology and Geophysics, Chinese Academy of Sciences, P.O. Box 9825, Beijing 100029, China 3The seventh Geological Brigade of Xinjiang Bureau of Geology and Mineral Resources, Wusu 833000, China

6.1 Abstract

Petrological and mineralogical data of interlayered eclogite, marble and quartz-mica schist from a drill core are used to constrain the metamorphic evolution of metabasalts and intercalated metasediments in the Tianshan (ultra-)high-pressure/ low-temperature metamorphic belt, NW China. The eclogite mainly consists of varying amounts of garnet, omphacite, quartz and epidote, the marble of calcite (>95 vol.%) with minor silicate inclusions such as zoisite and phengite, and the schist of quartz, mica and minor calcite, chlorite, albite and garnet. Using garnet isopleths thermobarometry, pseudosection calculations for the eclogite and quartz-mica schist reveal a common metamorphic history under HP condition of both rock types that is also consistent with

155

6. A common P-T evolution of eclogites and metasediments the P–T estimations for the marble using conventional thermometry and barometry. The uniform P–T paths of the interlayered eclogite, marble and mica-schist shows that they are an internally coherent HP unit. Thus mafic protoliths and surrounding sediments are believed to have undergone the same metamorphic evolution. In addition, the present HP rocks collected from the northern part of the Tianshan metamorphic belt are in contrast with the recently proposed scheme of a northern ―internally coherent UHP unit‖ and a southern ―coherent HP unit‖.

6.2 Introduction

In many metamorphic terranes worldwide, low temperature–high pressure (HP) metamafic rocks, such as eclogites, commonly occur as pods, boudins, lenses or interlayers in surrounding metasediments and are interpreted as foreign tectonic slices or as integral part of a lithostratigraphic sequence (e.g., Ernst, 1970; Carswell, 1990; Gao et al., 1999; Federico et al., 2007; Davis and Whitney, 2008; Klemd et al., 2011). In general, eclogite boudins/lenses record higher P–T conditions than the surrounding metasedimentary country rocks, which is interpreted to be a consequence of the competence contrast of both rock types. In the competent eclogites, which often escape deformation, interface kinetics and rates of diffusion are too sluggish to enhance (retrograde) metamorphic reactions occurring in the more incompetent metasediments (e.g., Rubie and Thompson, 1984). However, some studies revealed that some metasediments also experienced HP to ultra-high-pressure (UHP) metamorphism and shared a common metamorphic history with the enclosed/interlayered eclogites (e.g., Spear and Franz, 1986; Klemd et al., 1991, 1994; Liu et al., 2001; Gross et al., 2008), suggesting that they may be subducted and exhumed as a coherent unit. Thus, the study of the metamorphic evolution of both (U)HP eclogite lenses/interlayers and associated host metasediments may provide important information in understanding HP–UHP metamorphism during subduction and exhumation processes and the respective geodynamic setting.

The Tianshan (ultra-)high-pressure/low-temperature ((U)HP/LT) belt represents a

156

6. A common P-T evolution of eclogites and metasediments former oceanic subduction zone complex (Fig. 6-1a), which includes metabasalts (e.g., eclogites and blueschists) and metasedimentary rocks (e.g., mica schists and marbles). Eclogites occur commonly as pods, boudins, and thin layers or as massive blocks in metasediments and were interpreted to represent a tectonic mélange (Gao et al., 1999; Gao and Klemd, 2003; van der Straaten et al., 2008; Wei et al., 2009). A recent study proposed two separate tectonic units (Fig. 6-1b), namely an internally coherent UHP unit in the north and a coherent HP unit in the south, in the Chinese part of the Tianshan metamorphic belt. However, this ‗tectonic‘ division is based exclusively on thermodynamical modeling while a major tectonic lineament separating these two units was not found (Lü et al., 2012, 2013). Thus, further tectonic and petrological studies are still essentially needed in order to testify the reliability of the ‗tectonic division‘ model since several previous studies areported HP condition for eclogites and blueschists from the proposed UHP unit (e.g., Gao et al., 2007; Li et al., 2013; Wei et al., 2009). The present study is concerned with eclogites intercalated with mica schist and marble layers in the ‗UHP unit‘, thereby providing a good opportunity not only to test the hypothesis of coherent UHP vs. HP units by studying their respective metamorphic evolutions, but also whether the different rock types have experienced the same metamorphic evolution as internally coherent unit (Lü et al., 2012, 2013) or whether they represent tectonic slices from different depths which were juxtaposed during exhumation in the subduction zone (e.g., Klemd et al., 2011).

In this study, we present: (1) a detailed petrographical and petrological investigation of interlayered eclogite, marble and mica schist, which were derived from a drill core Hole-ZK2011. (2) Phase equilibria of one eclogite and one mica schist sample using a pseudosection approach in the MnNCKFMASH(O) system based on their effective bulk rock compositions (EBCs). (3) Garnet isopleth thermobarometry of eclogites and mica schists and conventional geothermobarometry for the calcite marbles. These data were used to discuss an integrated subduction and exhumation process of the interlayered eclogites and metasedimentary rocks and the respective geodynamic setting in the Chinese part of the Tianshan (U)HP/LT metamorphic belt.

157

6. A common P-T evolution of eclogites and metasediments

Mineral abbreviations are after Whitney and Evans (2010).

6.3 Geological setting

The southern Tianshan Orogen, which is situated along the southwest margin of the Central Asian Orogenic Belt, records the final collision between the Tarim and Yili cratons (e.g., Gao et al., 1998, 2009; Xiao et al., 2009). The Chinese Tianshan (U)HP/LT metamorphic belt extends for at least 200 km along the South Central Tianshan suture zone and consists of a suite of metasedimentary, mafic and ultramafic rocks that were interpreted as a paleo-accretionary wedge (Fig. 6-1a; Gao and Klemd, 2003). The (U)HP/LT belt is separated from the north and south units by two large-scale ductile shear zones, the South Central Tianshan suture and the Northern Haerkeshan Peak Fault, respectively (e.g., Gao and Klemd, 2003). It is mainly composed of greenschist-facies metasedimentary rocks and interlayered metabasalts. Eclogites, with N–MORB, E–MORB, OIB and arc basalt affinities (Gao and Klemd, 2003; John et al., 2008), are scattered in the greenschist-facies metasediments as pods, boudins, thin layers or large massive blocks (Fig. 6-1b). The greenschist-facies metasedimentary rocks are believed to have been subducted with the metabasalts to depths of ca. 70 km and then rapidly exhumed suffering extensive retrogression (Wei et al., 2009). A 319 ± 3 Ma peak metamorphic U–Pb age was obtained for metamorphic zircon rims from eclogites (Su et al., 2010). This age was confirmed, within error, by multi-point Lu–Hf isochron ages (garnet-growth age 315 ± 3 Ma) from four blueschist- or eclogite-facies rocks (Klemd et al., 2011) and a SHRIMP U–Pb zircon age of 320 ± 4 Ma from one UHP mica schist (Yang et al., 2013). The major post-eclogite-facies episode of cooling or recrystallization is revealed by 40Ar–39Ar and Rb–Sr ages of ca. 311 Ma for white mica from eclogite-facies metavolcanic rocks and omphacite-bearing blueschists (Klemd et al., 2005).

Peak metamorphic conditions for eclogites were estimated to range between 480 and 580 °C at 1.4–2.1 GPa14–21 kbar on a regional scale (e.g., Klemd et al., 2002; Wei et al., 2003). The presence of coesite inclusions in garnet suggests that some of the

158

6. A common P-T evolution of eclogites and metasediments

Figure 6-1 Geological map of the western Tianshan (U)HP/LT metamorphic belt in northwestern China, modified after Li et al. (2013). (a): Regional tectonic map; (b): Local geological map showing the area from which samples were taken. The dashed line shows the deduced boundary of north UHP unit and south HP unit by Lü et al. (2012). eclogites have experienced UHP metamorphism (Lü et al., 2009; Lü and Zhang, 2012). The intimate interlayering of HP and UHP rocks on a meter scale (Lü et al., 2009) was interpreted to be due to juxtaposition processes during subduction and exhumation in the subduction channel (Klemd et al., 2011). Furthermore, the identification of coesite relicts and thermodynamic modeling of metapelites indicate that some of the metasediments were also underwent UHP metamorphic conditions (e.g., Lü et al., 2008; Wei et al., 2009; Lü and Zhang, 2012; Yang et al., 2013). The petrology and metamorphic evolution of carbonate-bearing eclogite and metasediment were investigated in detail (Zhang et al., 2003; Li et al., 2012, 2013, 2014), however, little

159

6. A common P-T evolution of eclogites and metasediments attention has been given to marble although it usually occurs closely associated with mica schist (Lü et al., 2013).

6.4 Sample description and petrography

The HP–UHP metamorphic belt in the Kepuerte area is dominated by pelitic-felsic schists, enclosing various sizes of eclogite lenses or interlayers (Fig. 6-1b). In order to clarify the relationship between eclogites and country rocks, six samples were collected between -16.8 m to -17.6 m from drill core Hole-ZK2011 (Fig. 6-2) at the junctional area of Kepuerte River and Atantayi River (Fig. 6-1b). The collected drill core samples include multiple interlayers of eclogite with metasedimentary rocks such as mica-schists, marbles and minor quartzite. Individual layers range in thickness from 5 cm (quartzite) up to 20 cm (eclogite) and display rather sharp and undisturbed contacts (Fig. 6-2). The eclogite and metasediment interlayers are believed to occur as coherent sequence since the same lithological sequences are observed at the same depth in the drill core Hole-102 that was drilled about 10 meters away. No tectonic contacts have been identified between the eclogite and the metasedimentary rocks up to now. In the present study we investigated two samples (KP1–5 and KP1–6) from the eclogite layer in drill core Hole-ZK2011, in addition, two marble samples (KP1–4a and KP1–4b) and two quartz-mica schist samples (KP1–4b and KP1–7b) were studied from positions above and below the eclogite layer, respectively (Fig. 6-2).

Figure 6-2 Photography and schematic stratigraphic diagram of drill core sample showing the interlayered eclogite, marble and mica-schist.

160

6. A common P-T evolution of eclogites and metasediments

6.4.1 Eclogite

Both eclogite samples KP1–5 and KP1–6 have a similar modal composition consisting of garnet (ca. 9 vol. %), omphacite (ca. 33 vol. %), quartz (10–30 vol. %), epidote (10– 30 vol. %), amphibole (7%), calcite (3%) and accessory minerals such as phengite, pyrite, rutile and titanite (Fig. 6-3a–c). Garnet porphyroblasts (grain size ranging from 0.5 to 2 mm) contain quartz, calcite, chlorite and box-shaped epidote-paragonite-albite intergrowths as inclusions in the core-mantle domain, and omphacite, quartz inclusions in the mantle-rim region. The box-shaped epidote-paragonite-albite intergrowths (Fig. A1) are believed to be pseudomorphs after lawsonite (cf. Klemd et al., 2002). Post-peak metamorphic garnet inclusions are mainly Ca-amphibole, albite and minor chlorite, which mainly occur along cracks. Omphacite and quartz are the main matrix minerals. Generally, sodic-calcic to calcic amphiboles occur as retrograde minerals with albite and grow at the expense of matrix omphacite (Fig. 6-3c). Minor phengite either occurs as matrix mineral or as inclusion in garnet porphyroblasts. Accessory titanite occasionally contains rutile inclusions.

6.4.2 Quartz-mica schist

The quartz-mica schist samples KP1–4b and KP1–7b are composed of garnet (ca. 0.5 vol. %), quartz (ca. 40%), phengitic muscovite (ca. 40%), paragonite (ca. 5%), calcite (ca. 5%), chlorite (ca. 5%), albite (ca. 2%) and accessory dolomite, apatite, pyrite and rutile (Fig. 6-3d–f). Chlorite, calcite and quartz assemblages form pseudomorphs after former garnet porphyroblasts, which contain –usually in the core domain– minor discrete garnet relicts (Fig. 6-3d and e). The modal amount and size of the garnet pseudomorphs reveal that the quartz-mica schist initially contained 3–5 vol. % garnet with the size of 0.5–1.5mm in diameter at peak metamorphic conditions. The garnet relicts are rather inclusion-free while phengite, quartz, dolomite and rutile inclusions occur in the garnet pseudomorphs. Occasionally, however, the appearance of small albite inclusion in the garnet pseudomorphs may indicate the former presence of

161

6. A common P-T evolution of eclogites and metasediments

162

6. A common P-T evolution of eclogites and metasediments

◀Figure 6-3 Representative photomicrographs of the interlayered eclogite, marble and quartz- mica-schist. (a) Photomicrograph of the eclogite consisting of garnet, omphacite, quartz, epidote and amphibole. (b) Garnet porphyroblast contains quartz, and other mineral assemblages. (c) Garnet porphyroblast with a compositional profile showing the growth zoning, and omphacite that is locally replaced by retrograde calcic amphibole+albite assemblages. (d) Garnet pseudomorph presented by chlorite-calcite-quartz assemblages in quartz-mica-schist. (e) Occasional garnet relicts are preserved as discrete fragments in pseudomorph in quartz-mica-schist. (f) Matrix minerals in quartz-mica-schist and early HP index minerals are not preserved. (g, h) Marble contains minor dolomite, phengitic muscovite (Ph_I and Ph_II), paragonite, zoisite, quartz, chlorite, kaolinite and rutile/titanite inclusions. Primary phengite (Ph_I) commonly occurs as isolated mineral between calcite grains and Ph_II generally occurs as secondary mineral replacing early zoisite. omphacite. Quartz and phengite are the dominant matrix minerals, both of which define the good foliation in the mica schist (Fig. 6-3f).

6.4.3 Marble

The here investigated marble samples KP1–4a and KP1–7a are dominated by calcite (> 95 vol. %) and contains only minor dolomite, phengitic muscovite, paragonite, zoisite, quartz, feldspar, chlorite, kaolinite and rutile/titanite inclusions. Coarse-grained granoblasts (0.5–2 mm) of calcite occur homogeneously distributed in the rock matrix, while silicates form isolated mineral assemblages within the calcite matrix (Fig. 6-3g and h). Two textural types of phengitic muscovite are distinguished: primary phengite (Ph_I) is tabular, idioblastic and usually occurs as isolated mineral in the calcite matrix (Fig. 6-3g), while later xenoblastic phengite (Ph_II) usually is intimately intergrown with zoisite and calcite relicts (Fig. 6-3g and h). The other late minerals are paragonite, kaolinite, albite and orthoclase, which are associated with Ph_II or occur as inclusions in Ph_II, or along cracks within zoisite (Fig. 6-3g and h). Frequently, chlorite is intimately intergrown with Ph_II and calcite in places separating them (Fig. 6-3g). In addition, dolomite is intergrown with zoisite and Ph_II and quartz generally occurs as isolated inclusion in calcite (Fig. 6-3h).

6.5 Analytical methods

The bulk rock major element compositions were determined by X-ray fluorescence

163

6. A common P-T evolution of eclogites and metasediments spectrometry (XRF) on fused glass disks, using a Philips PW 2400 XRF spectrometer at the GeoZentrum Nordbayern (GZN), University Erlangen–Nürnberg, Erlangen, Germany. The loss on ignition (LOI) was determined prior to major element analyses using a pre-ignition method after heating the samples (1g) for 12h at 1300 °C. The major elements SiO2, TiO2, Al2O3, Fe2O3, MnO, MgO, CaO, Na2O, K2O, and P2O5 along with a subset of trace elements (Ba, Cr, Ga, Nb, Ni, Pb, Rb, Sr, Yh, V, Y, Zn, Zr) were analyzed. Generally, precision and accuracy were better than 0.8 and 1 % (2σ) for most elements and 0.5 % for SiO2. Whole rock major element compositions are summarized in Table 6-1.

Major element compositions of minerals were obtained at the GZN using a JEOL JXA– 8200 microprobe. The microprobe is equipped with five wavelength-dispersive spectrometers (WDS). For all measured elements in various minerals an acceleration voltage of 15 kV, a probe current of 15 nA and a probe diameter of 3 µm were used. Both natural and synthetic oxides were used as standards and data corrections were conducted using a ZAF procedure. Element mapping of garnet was conducted in WDS (Ca, Fe, Mn, Mg, Al) and EDS (energy-dispersive spectrometer) mode with an acceleration voltage of 15 kV, a beam current of 100 nA, a 6 µm pixel size and dwell time of 100 ms.

6.6 Results

6.6.1 Bulk rock chemistry

Bulk-rock major element compositions from the eclogite, the quartz-mica schist and the marble samples are presented in Table 6-1. The eclogite samples have basaltic compositions with SiO2 = 53.7–55.7 wt.%, Na2O = 2.8 wt.%, K2O = 0.4–0.6 wt.% and

TiO2 = 0.5–1.1 wt.% (Table 6-1). The SiO2 content is slightly higher than normal Tianshan eclogites due to its higher quartz amounts (cf. John et al., 2008). The average loss on ignition (LOI) is 3.9 wt.%. Compared to the eclogite samples, the quartz-mica schist samples have higher SiO2 (60.3–61.1 wt.%) and K2O (4.6–4.8 wt.%) contents due

164

6. A common P-T evolution of eclogites and metasediments to the higher abundance of quartz and phengite, and lower CaO (2.6–3.6 wt.%) and

Na2O (0.4-0.9 wt.%) contents (Table 6-1). The average LOI is 6 wt.%. The bulk-rock major composition of the marble is rather monotonous with a CaO content of ca. 55.3 wt.% and a LOI of 41 wt.%. Minor SiO2 ranges between 0.95 and 1.4 wt.%, MgO

T between 1.1 and 1.2 wt.%, and even less Al2O3 and FeO (Table 6-1).

Table 6-1 Bulk major and trace element of samples eclogite mica schist marble Sample KP1-5 KP1-6 KP1-4b KP1-7b KP1-4a KP1-7a

SiO2 55.74 53.66 60.33 61.13 0.95 1.43

TiO2 1.05 0.54 0.63 0.67 0.04 0.04

Al2O3 12.33 13.97 14.68 14.67 0.27 0.32

Fe2O3* 7.27 9.26 5.77 5.47 0.33 0.45 MnO 0.07 0.18 0.04 0.06 0.06 0.05 MgO 4.60 4.16 4.44 2.53 1.21 1.11 CaO 10.50 11.67 2.57 3.62 55.25 55.27

Na2O 2.80 2.80 0.39 0.85 0.14 0.13

K2O 0.58 0.43 4.80 4.56 0.08 0.12

P2O5 0.01 0.01 0.10 0.13 0.05 0.05 LOI 4.81 3.07 5.94 6.02 41.55 40.94 Total 99.76 99.75 99.70 99.71 99.93 99.91 *Total iron for XRF data; LOI: loss on ignition.

6.6.2 Mineral chemistry

6.6.2.1 Eclogite

Idioblastic garnet in the eclogite samples shows distinctive growth zoning with a core composition of Alm67.0–68.7Prp7.2–7.6Grs18.4–19.7Sps5.3–6.0 and XFe of 0.89–0.91, a mantle composition of Alm66.4–70.4Prp7.9–8.8Grs19.3–21.0Sps2.1–4.7 and XFe of 0.89–0.90 and a rim composition of Alm62.0–64.6Prp11.4–12.1Grs22.8–24.9Sps1.6–2.2 and XFe of 0.84–0.88 (Table 6-2). The spessartine content decreases from core to rim while the grossular and pyrope contents increase indicating prograde growth (Fig. A1; cf., Hollister, 1966). The almandine content shows a slight increase from core to the garnet mantle domains and then decreases towards the rim (Fig. 6-4a). At the outermost rim (less than 50 µm in width), the garnet grain is characterized by a distinct composition involving a decrease

165

6. A common P-T evolution of eclogites and metasediments of almandine and pyrope contents, and an increase in grossular and spessartine contents (Fig. 6-4a, Table 6-2), which may be due to resorption and subsequent garnet growths (cf. Meyer et al., 2013).

Figure 6-4 Chemical compositions of garnet, omphacite, amphibole and white mica. (a) Compositional zoning of garnet in the eclogite showing continuous prograde growth. (b) Compositional profile of the garnet in the quartz-mica-schist (Fig. 6-3e). (c) Compositional ternary classification diagram of omphacite in the eclogite. (d) Chemical composition of amphibole in the eclogite. (e, f) Phengite and paragonite compositions in the studied samples.

166

6. A common P-T evolution of eclogites and metasediments

167

6. A common P-T evolution of eclogites and metasediments

The clinopyroxene in the eclogite samples is omphacite (Fig. 6-4c; Table 6-2) according to the classification of Morimoto et al. (1988). Omphacite has a jadeite component ranging between 41 and 48. The amphiboles have Mg/(Mg+Fe2+) ratios of 0.70–0.78 and range between actinolites and magnesio-hornblende in composition (Fig. 6-4d) according to the classification of Leake et al. (1997). Phengite, which is the dominant white mica, has Si contents in the range of 3.36–3.45 per formula unit (p.f.u.), while paragonite has a Na-content of 0.76 p.f.u. and a Si-content of 3.14 p.f.u. (Fig. 6-4e–f;

Table 6-2). Matrix epidote group minerals have a Fe2O3 content of < 2 wt.% and are thus usually zoisite, while inclusions in garnet and zoisite porphyroblasts usually are

3+ 3+ epidote with a Fe2O3 content > 6 wt.% (Table 6-2). The Ps ratio [XPs = Fe / (Fe + Al)] of zoisite varies from 0.02 to 0.04, while epidote has a higher Ps ratio of 0.14–0.21 (Table 6-2, cf., Franz and Liebscher, 2004).

6.6.2.2 Quartz-mica schist

Compositional core–to–rim profiles on the best preserved garnet relicts (Fig. 6-3e) reveal that garnet in the quartz-mica schists also displays particular growth zoning (Fig.

6-4b), reflected by a core composition of Alm69.0–70.2Prp6.5–7.6Grs11.1–14.3Sps12.0–9.1 and

XFe of 0.91, a mantle composition of Alm69.0–70.2Prp7.8–9.6Grs17.9–19.6Sps2.0–4.9 and XFe of

0.88–0.90 and a rim composition of Alm66.9–69.1Prp10.4–12.6Grs18.9–20.2Sps0.9–1.7 and XFe of 0.84–0.87 (Table 6-2). Similar to the garnets in the eclogite samples, the Mn content in garnet decreases from core to rim while the Ca and Mg contents increase indicating prograde growth. The Fe content is almost constant from the core to the mantle domains and then decreases towards the rim (Fig. 6-4b).

Chlorite in the garnet pseudomorphs generally has slightly higher XFe ratios (0.40–0.49) than that in the matrix (0.35–0.38) (Table 6-2). Occasionally, some chlorite bands

(width less than 20 µm in size) in the garnet pseudomorphs have a XFe ratio up to 0.62 indicating disequilibrium conditions during chlorite growth. The Si-content of phengite clusters at 3.35–3.58 p.f.u. and partly overlaps with that in the eclogites (Fig. 6-4e–f; Table 6-2), while paragonite has a Na-content of 0.76–0.88 p.f.u. and a Si-content of

168

6. A common P-T evolution of eclogites and metasediments

3.12–3.16 p.f.u.. Calcite contains a small amount of MgO (3–4 mol.%), FeO (2–3 mol.%) and MnO (1–2 mol.%), while dolomite has an ankerite component between 20 and 30 mol.%. Apatite has an F content of 3–4 wt.%.

6.6.2.3 Marble

Calcite in marbles is generally pure calcite with minor MgO (1–2 mol.%), FeO (0.2–0.6 mol.%) and MnO (0.1–0.3 mol.%) contents (Table 6-2). Dolomite contains significant amounts of the ankerite component; however, the dolomite rim often has a lower Fe and higher Mg than the core domain, which is similar to the zoning in dolomite observed in HP eclogites in the Tianshan metamorphic belt (Li et al., 2012, 2014).

3+ 3+ Zoisite has a Fe2O3 content < 1 wt.% and the Ps ratio [XPs = Fe / (Fe + Al)] is usually less than 0.02 (Table 6-2). Compared to the texturally primary phengite (Ph_I), the later phengite (Ph_II) contains less SiO2 (50.39 vs. 53.29 wt.%) and MgO (1.64 vs. 3.71 wt.%) and more Al2O3 (31.65 vs. 28.75 wt.%) and FeO (2.28 vs. 0.54 wt.%) (Table 6-2). The Si-content of Ph_I ranges between 3.41 and 3.44 p.f.u. similar to that of PH_I in the eclogite and quartz-mica-schist samples, while the Si-content of Ph_II clusters at 3.31– 3.36 p.f.u. (Fig. 6-4e–f; Table 6-2). In addition, paragonite has a Na-content of 0.83 p.f.u. and a Si-content of 3.15 p.f.u. (Fig. 6-4e–f; Table 6-2). Almost pure albite and orthoclase occur in and around Ph_II. Chlorite generally has slightly higher XFe values (ca. 0.56) than that in the mica-schists. Kaolinite shows a composition close to stoichiometric kaolinite and contains minor MgO and K2O contents (Table 6-2).

6.7 Phase equilibria and P–T evolution

6.7.1 Pseudosection calculation

According to the mineral assemblages and bulk rock compositions, the model system

MnNCKFMASH(O) [MnO–Na2O–CaO–K2O–FeO–MgO–Al2O3–SiO2–H2O(–O)] was chosen to calculate P–T pseudosections for the eclogite and quartz-mica-schist. Quartz and a fluid phase assumed to be pure H2O, were considered to be in excess. TiO2, which

169

6. A common P-T evolution of eclogites and metasediments

predominantly occurs in rutile/titanite, was neglected. CO2 was also omitted from our calculation because of the minor amounts of carbonate in rocks. MnO and Fe2O3 were included for the calculation of the pseudosection for one eclogite sample since both affect the stability fields of garnet and amphibole, respectively. Since naturally-

3+ occurring oceanic eclogites have higher Fe contents than unaltered MORB, XFe3+ (Fe3+/(Fe2+ + Fe3+)) was set at 0.36 following the suggestion of Rebay et al. (2010). In addition, a similar value was also used by Chen et al. (2013) for pseudosection modeling of oceanic eclogites. However, Fe2O3 was not included in the model system for the quartz-mica-schist, which is justified by the almost complete absence or scarcity of Fe3+-bearing minerals in this rock.

It is critical to generate an effective bulk composition (e.g., Stüwe, 1997) for phase equilibrium calculations in metamorphic rocks, since the bulk rock composition of metamorphic rocks may change during its P–T evolution for instance due to element fractionation during prograde garnet growth (e.g., Evans, 2004). In this study, bulk compositions effectively reacting during core–mantle–rim stages of garnet growth have been calculated following the method described by Evans (2004) and Gaidies et al. (2006). This method requires a strong correlation between the Mn concentration and Fe, Mg, Ca concentrations in garnet (Fig. 6-5c–d), based on a Rayleigh fractionation model of Mn during garnet growth (Hollister, 1966). Garnet zoning has been divided into three growth stages (core–mantle–rim) on the basis of the Mn content (Fig. 6-5a–b). The garnet core with the highest MnO content was considered to represent the primary

170

6. A common P-T evolution of eclogites and metasediments

Figure 6-5 (a, b) Mn X-ray map of the garnet from the eclogite (KP1-6) and BSE images of the garnet from the mica-schist (KP1-4b) showing shape and distribution of the three growth domains (a–b–c) of garnet used for the modelling of bulk-composition fractionation. (c, d) Relationships between selected compositional variables in garnet. The curvilinear correlations confirm that the sample is suitable for the determination of the fractionation effects on the bulk rock composition according to the method proposed by Evans (2004) and Gadies et al. (2006). (e, f) Plots of measured MnO, FeO, MgO and CaO vs. calculated modal garnet based on profile analyses on garnet from the eclogite (e) and mica schist (f). The concentrations of MnO, FeO, MgO and CaO in garnet to be subtracted from the initial bulk-composition, are then calculated by integrating the trendline for each plot between the limits of the lowest and highest modal value of garnet for each stage (a-b-c), respectively.

171

6. A common P-T evolution of eclogites and metasediments garnet crystallization when the effective bulk composition corresponds to the whole-rock composition (XRF result). Changes in the bulk rock chemistry as a result of preferential partitioning in garnet, which were subtracted from the initial bulk-rock composition, are calculated by integrating the fitting-curve for each element between the limits of the lowest and highest modal value of garnet for each stage (a–b–c), respectively (Fig. 6-5e–f, Gaidies et al., 2006; Groppo and Castelli, 2010). The representative effective bulk compositions of core–mantle–rim stages calculated for pseudosections are summarized in Table 6-3. The P–T conditions of each stage were constrained using the method of garnet isopleth thermobarometry (cf., Gaidies et al., 2006; Groppo and Castelli, 2010; Meyer et al., 2013).

Table 6-3 Effective bulk compositions used for pseudosection calculations (wt. %) Sample eclogite KP1-6 mica schist KP1-4b Stage a (c*) b (m) c (r) a (c) b (m) c (r)

SiO2 55.85 56.57 58.03 65.25 65.44 65.83

Al2O3 14.54 14.49 14.01 15.88 15.84 15.76 FeO 5.55 5.29 4.23 5.61 5.44 5.09

Fe2O3 3.47 3.31 2.65 MnO 0.19 0.15 0.06 0.06 0.04 0.01 MgO 4.33 4.41 4.59 4.80 4.82 4.86 CaO 12.15 12.34 12.72 2.78 2.77 2.72

Na2O 2.91 2.99 3.21 0.42 0.42 0.43

K2O 0.45 0.46 0.49 5.19 5.23 5.30 Total 100.00 100.00 100.00 100.00 100.00 100.00 *c: core, m: mantle, r: rim.

6.7.2 Pseudosection for eclogite sample KP1–6

The MnNCKFMASHO P–T pseudosections calculated for eclogite sample KP1–6 during bulk fractionation is presented in Fig. 6-6a–c. Garnet, omphacite, phengite and quartz/coesite are stable phases under all shown P–T conditions. Amphibole is absent in the HT–HP field. Lawsonite stabilizes in the LT–HP field, while the epidote group minerals appear in the HT–LP field. Along with the changes of the effective bulk composition corresponding to the element fractionation during garnet growth, the field with two kinds of omphacite is enlarged towards higher temperature, while the talc-

172

6. A common P-T evolution of eclogites and metasediments

Figure 6-6 P–T pseudosection calculations of the eclogite KP1-6 (a–c) and the mica-schist KP1-4b (d–f) based on the effective bulk compositions from Table 6-3. The garnet-in line is plotted as a heavy grey line, while compositional isopleths of Grs and Prp contents in garnet are plotted as dashed lines, and the estimated P–T range at individual stage are overlap by shaded circles.

173

6. A common P-T evolution of eclogites and metasediments bearing assemblage is stable at higher pressure only. Chlorite is stable under lower pressures of up to 23 kbar during peak metamorphic conditions (Fig. 6-6a–c). However, the amphibole-out, lawsonite-out and epidote-in isograds follow the change of the EBC (Fig. 6-6a–c).

The measured mineral composition of the garnet cores (Prp: 7.2–7.6 and Grs: 18.4–19.7) predicts that garnet nucleation started at a P–T range of 480–490 ºC and 25.5–26.5 kbar in the stability field of Grt + Omp + Gln + Ph + Lws + Tlc + Qz (Fig. 6-6a). The garnet mantle compositions (Prp: 7.9–8.8 and Grs: 19.3–21.0) constrain similar P–T conditions of 495–505 ºC and 24–26 kbar and are in equilibrium with the same mineral assemblages as the garnet core (Fig. 6-6b). However, the garnet rim compositions (Prp: 11.4–12.1 and Grs: 22.8–24.9) are consistent with a P–T range of 530–540 ºC and 24– 26 kbar in the stability field of Grt + Omp + Gln + Ph + Lws + Qz (Fig. 6-6c). Therefore, the reconstruction of the P–T path for the eclogite sample using the garnet isopleth thermobarometry method revealed that the eclogite has experienced persistent heating under HP condition before and during initial exhumation (Fig. 6-7), which was followed by initial isothermal decompression as is indicated by the abrupt increasing Ca component in the outermost rim of garnet (Wei et al., 2009; Li et al., 2012).

6.7.3 Pseudosection for quartz-mica schist sample KP1–4b

The calculated MnNCKFMASH P–T pseudosections for the quartz-mica schist sample KP1–4b is shown in Fig. 6-6d–f. Phengite and quartz/coesite are stable phases under all the shown P–T ranges. Chlorite stabilizes under LT–LP conditions while omphacite is present under HT–HP conditions. However, a second (low-jadeite) clinopyroxene is also stable under very low temperatures and pressures. The distribution patterns of amphibole, lawsonite and epidote are similar with those calculated for the eclogite pseudosections. Along with minor changes of the effective bulk composition, the mineral stability fields in the pseudosections (Fig. 6-6d–f) are rather similar due to the small modal amount of garnet allowing only limited element fractionation during the prograde P–T path of the schist. However, the garnet-in isograd was calculated to occur

174

6. A common P-T evolution of eclogites and metasediments at high P–T conditions (Fig. 6-6d–f). This is a reflection of the incorporated MnO content in the system that plays an important role in determining the stability of garnet at lower temperature especially at lower pressure conditions (e.g., Droop and Harte, 1995; Konrad-Schmolke et al., 2008).

According to the pyrope and grossular components of the garnet relicts in the quartz-mica schist, the garnet core with Prp=6.5–7.6 and Grs=11.1–14.3 corresponds to an initial growth at 460–470 ºC and 24.5–25.5 kbar in the Grt + Qz + Ph + Chl + Gln + Lws ± Tlc stability field (Fig. 6-6d). The garnet mantle with Prp=7.8–9.6 and Grs=17.9– 19.6 constrains P–T conditions at 490–510 ºC and 24–25 kbar in the Grt + Qz + Ph + Chl + Gln + Lws + Tlc stability field (Fig. 6-6e), while the garnet rim with Prp=10.4– 12.6 and Grs=18.9–20.2 are in agreement with a P–T range of 510–520 ºC and 24–25 kbar in the Grt + Qz + Ph + Gln + Lws + Tlc stability field (Fig. 6-6f). Thus a consistent P–T path for quartz-mica schist and the eclogite is defined, monitoring a distinct thermal relaxation at the beginning of exhumation (Fig. 6-7).

6.7.4 P–T constraints for marble sample KP1–7a

Due to its monotonous mineral assemblages and lack of critical index mineral (such as omphacite, Lü et al. 2013), the exact P–T path of the marble samples are impossible to reconstructed using thermodynamic modeling. However, the P–T conditions of marble sample KP1–7a was constrained by using conventional calcite-dolomite thermometry and phengite barometry. The calcite-dolomite pairs in marble sample KP1–7a suggest that the temperature (pressure independent) range between 430 ºC and 560 ºC (Table

6-4), considering the MgCO3 content in calcite (surrounding dolomite) and the effect of

FeCO3 on calcite-dolomite thermometry (Anovitz and Essene, 1987).

The silica content in phengite is pressure dependent in silicate rocks; therefore the Si content in phengite coexisting with a limiting mineral assemblage (e.g. K-feldspar, biotite, quartz) is used to constrain metamorphic pressures for a given temperature (e.g., Massonne and Schreyer, 1987). Although caution is warranted when phengite

175

6. A common P-T evolution of eclogites and metasediments compositions are the only indicator for metamorphic pressures when lacking the typical equilibrium HP assemblages or other critical index minerals (e.g. Bröcker et al., 2004), however, in the presented study the lithostratigraphic sequence of cm-scale interlayered marbles and mica-schists and the absence of tectonic contacts suggests that they most probably experienced the same metamorphic evolution as a coherent unit (cf., Klemd et al., 1991). This has been confirmed by the thermodynamic modeling of interlayered marble and mica-schists from another location in the Tianshan (U)HP/LT belt (Lü et al., 2013). Thus, the similar range of Si content of phengites in marbles (3.41–3.44 p.f.u. of Ph_I), mica-schists (3.35–3.58 p.f.u.) and eclogites (3.36–3.45 p.f.u.) indicates that the marble probably was part of the initial lithostratigraphic sequences and shared a common metamorphic evolution with the eclogites and mica-schists (Fig. 6-7).

Table 6-2 Representative temperatures of marble obtained from calcite-dolomite geothermometry Sample KP1-7b Cal FeO 1.81 0.53 0.57 0.80 0.47 0.22 MnO 0.12 0.08 0.12 0.22 0.13 0.60 MgO 1.60 1.87 1.62 1.02 0.78 0.15 CaO 52.16 49.86 54.46 50.23 57.98 57.61 Total 55.69 52.33 56.76 52.26 59.36 58.57 Fe 0.050 0.068 0.016 0.016 0.024 0.024 Mn 0.003 0.003 0.002 0.003 0.007 0.003 Mg 0.080 0.100 0.098 0.079 0.054 0.088 Ca 1.864 1.764 1.877 1.899 1.911 1.872 Cations 2 2 2 2 2 2

XFe 0.025 0.035 0.008 0.008 0.012 0.012

XMg 0.040 0.052 0.049 0.040 0.027 0.044 o TMg ( C)* 758 807 797 755 683 777 o TFe-Mg ( C)* 512 560 531 492 432 516 * Anovitz and Essene, 1987

6.8 Discussion

6.8.1 Metamorphic evolution for interlayered eclogite, marble and mica schist

The garnet isopleths thermobarometry for the eclogite sample KP1–6 defines a P–T path

176

6. A common P-T evolution of eclogites and metasediments from 480–490 ºC at 25.5–26.5 kbar (garnet core) via 495–505 ºC at 24–26 kbar (garnet mantle) to 530–540 ºC at 24–26 kbar (garnet rim). The relictic garnet from the mica-schist KP1–4b records a similar P–T path from 460–470 ºC at 24.5–25.5 (core) via 490–510 ºC at 24–25 kbar (mantle) to 510–520 ºC and 24–25 kbar (rim). Furthermore, conventional calcite-dolomite thermometry and phengite barometry constrain the peak metamorphism of the marble sample KP–7b at 430–560 ºC at 24–26 kbar. Consequently, the thermodynamic modeling and geothermobarometric calculations described above suggest that the interlayered eclogite, marble and mica-schist shared a common metamorphic evolution during P–T conditions of 450–550 ºC at 24–26 kbar (Fig. 6-7). This indicates that the interlayered eclogite, marble, and mica schist occurred as an internally coherent unit sharing at large parts of a common subduction and exhumation process.

Figure 6-7 P–T paths of the eclogite and mica-schist extracted from the pseudosections (Fig. 6-6) showing that the eclogite and metasediments share a common metamorphic history. The P–T conditions of marble estimated using conventional thermobarometer are marked as dashed box.

In addition, a retrograde overprint affected the different rock types of the sequence at different degrees during exhumation. The more competent eclogites (see discussion above) were least affected showing only minor calcic amphibole-albite replacement of omphacite (Fig. 6-3c), and a resorption of the outermost garnet rims as displayed by

177

6. A common P-T evolution of eclogites and metasediments increasing Mn contents (Fig. 6-6a). However, the quartz-mica-schists display a severe retrograde overprint, because not only the HP index minerals (such as talc, lawsonite and probably glaucophane) have been entirely erased from the samples (Fig. 6-3f and 6-6d–f), but also garnet was locally almost completely replaced by the chlorite ± calcite (Fig. 6-3d and e). The retrograde overprint in the marble is exhibited by the replacement of early zoisite by muscovite (Ph_II), paragonite, kaolinite and/or chlorite (Fig. 6-3g and h), similar to retrograded HP marbles from the Eclogite Zone in Tauern Window (Spear and Franz 1986).

6.8.2 Southern HP unit vs. northern UHP unit in the Tianshan?

Increasing numbers of eclogites and metapelite rocks in the Chinese part of the Tianshan (U)HP/LT metamorphic belt were considered to have experienced UHP metamorphism as shown by the actual presence of coesite or its former presence deduced by thermodynamic modeling (e.g., Lü et al., 2008, 2009, 2012, 2013; Wei et al., 2009; Lü and Zhang, 2012; Yang et al., 2013). According to pseudosection modeling by means of garnet isopleths thermobarometry, two tectonic units were proposed to occur in the Tianshan (U)HP/LT metamorphic belt: an internally coherent UHP unit in the north as well as a coherent HP unit in the south (Lü et al., 2012). However, considering the restricted modeling evidence and related uncertainties in pseudosection calculations in Lü et al. (2012) such as the employment of different thermodynamic databases and the neglectance of Fe3+, the tectonic division is unconvincing due to the absence of. a major tectonic lineament between the anticipated HP and UHP units (e.g., Gao et al., 1998, 2009; Gao and Klemd, 2003). Therefore, in the absence of a major lineament the occurrence of some coesite-bearing eclogites and metasediments in the northern unit does not mean that it acted as a coherent UHP unit, because HP (non-UHP) eclogites and blueschists were also reported from this unit (e.g. Li et al., 2012, 2013, 2014).

2+ Furthermore, Lü et al. (2012) use the XGrs = Ca / (Fe + Ca + Mg + Mn) value of garnet cores (and mantles) as only decisive criterion for defining HP or UHP metamorphism. This is due to the fact that the grossular content is largely pressure-dependent and the

178

6. A common P-T evolution of eclogites and metasediments

UHP conditions are only recorded in the garnet core and/or mantle in Chinese Tianshan eclogites. In general, if the XGrs value of garnet cores cluster at ca. <0.15 (for schists it may be even lower), the calculated peak conditions of a specific eclogite very likely reveal UHP conditions, while HP conditions are indicated if the XGrs value of garnet cores is > 0.20 when using garnet isopleths thermobarometry (Lü et al., 2012). However, the garnet (core) composition depends on several factors, such as partition coefficients, P–T conditions and bulk-rock compositions. Consequently, it is unreliable to use the

XGrs value of garnets in eclogites or metasediments as the only evidence for UHP metamorphism. In addition, in UHP/LT eclogites, the XGrs value of garnets cannot constrain pressures unambiguously because it is largely affected by the bulk rock composition at peak metamorphic conditions (Wei et al., 2013).

In the present study, however, the absence of any UHP mineral and the thermodynamic modeling of samples taken from so-called ―northern UHP unit‖ shows that the eclogite and quartz-mica schist just experienced HP (not UHP) metamorphism (see above), which makes the division of a northern UHP unit and a southern HP unit unlikely. The intimate interlayering of high- and ultrahigh-pressure rocks in the northern and southern units suggests that the rocks may be derived from varying depths from the descending slab and then juxtaposed during exhumation in the subduction channel (Klemd et al., 2011).

6.8.3 Geodynamic implications

A striking feature in the northern ‗UHP unit‘ of the Chinese Tianshan (U)HP/LT metamorphic belt is the mixed chaotic occurrence of HP eclogites and metasedimentary rocks (e.g., Klemd et al., 2002; Wei et al., 2003, 2009; Li et al., 2012, 2013; this study) with UHP eclogites and metapelites (e.g., Lü et al., 2009, 2012; Wei et al., 2009; Tian and Wei, 2013; Yang et al., 2013). The previous metamorphic thermodynamic modeling studies reveal that those HP and UHP rocks have experienced different P–T paths (Fig. 6-8). The various P–T paths are hardly considered to be a result of inconsistencies between the thermobarometry data of HP and UHP rocks since they were investigated

179

6. A common P-T evolution of eclogites and metasediments using the same thermodynamic modeling approach, the same computer software, and the same internally consistent thermodynamic datasets (cf., Wei et al., 2009). This implies that metamorphism of the metapelites and the various eclogites occurred in different depths of the subduction zone. Thus the eclogites and metasediments exhibit different P–T paths, consistent with the chaotic incorporation in a subduction channel (e.g., Cloos and Shreve, 1988; Gerya et al., 2002; Federico et al., 2007). Furthermore, numerical modeling shows that lower viscosities in the subduction channel resulted in smaller diameters of coherent units with increasing tectonic juxtaposition during exhumation (Gerya et al., 2002).

Figure 6-8 Various P–T paths derived from different HP–UHP metamorphic rocks in the northern ‘UHP’ part of the Tianshan (U)HP/LT metamorphic belt. The solid lines with arrows refer to eclogites whereas the dashed lines refer to metapelites. All P-T path were deduced by thermodynamic modeling.The shadow areas cover the estimated maximum pressure conditions from some HP and UHP metasedimentary rocks. The different P–T paths indicate those rocks were juxtaposed and exhumed in a subduction channel.

180

6. A common P-T evolution of eclogites and metasediments

However, in our case, the uniform P–T path of the interlayered eclogite, marble and mica-schist suggests that they are a local internally coherent HP unit during subduction and exhumation (e.g., Klemd et al., 1991; Gross et al., 2008). Thus, it is believed that in the Tianshan (U)HP/LT metamorphic belt some mafic basalt fragments intermingled with the sediments and subducted together into a great depth. The sediments may have been derived from subduction erosion processes (von Huene and Scholl, 1991). Thus it is believed that the meta-mafic and surrounding metasedimentary rocks have experienced integrated subduction and exhumation processes. In addition, our calculated P–T paths show a long pervasive heating after subduction during peak pressure conditions of this internally coherent HP unit, which is differs from P–T paths modeled for other (U)HP rocks in the Tianshan (Fig. 8). This suggests that some of the Tianshan HP rocks may experience thermal relaxation (i.e. longer duration) in the subduction channel prior to rapid decompression which is in accordance with peak metamorphic ages ranging from 324 to 312 Ma (see above).

Acknowledgements

The National Natural Science Foundation of China (41025008) and the Deutsche Forschungsgemeinschaft (KL692/17-3) funded this study. We are grateful to M. Hertel for help with the XRF and H. Brätz for assistance with the LA–ICP–MS whole-rock analyses, and to M. Meyer for help with the electron microprobe analyses. The first author thanks the CSC–DAAD for the scholarship supporting his PhD study in Germany.

References

Anovitz L.M. and Essene E.J. (1987) Phase Equilibria in the System CaCO3-MgCO3- FeCO3. Journal of Petrology 28, 389–414. Bröcker M., Bielin D., Hacker B. and Gans P. (2004) High-Si phengite records the time of greenschist facies overprinting: implications for models suggesting mega- detachments in the Aegean Sea. Journal of Metamorphic Geology 22, 427–442. Carswell D.A. (1990) Eclogites and eclogite facies: definitions and classifications. in: Carswell D.A .(ed) Eclogite Facies Rocks. Blackie and Son Ltd, New York, pp 1–13.

181

6. A common P-T evolution of eclogites and metasediments

Chen Y., Ye K., Wu T.F. and Guo S. (2013) Exhumation of oceanic eclogites: thermodynamic constraints on pressure, temperature, bulk composition and density. Journal of Metamorphic Geology 31, 549–570. Cloos M. and Shreve R.L. (1988) Subduction-Channel Model of Prism Accretion, Melange Formation, Sediment Subduction, and Subduction Erosion at Convergent Plate Margins: 1. Background and Description. Pure and Applied Geophysics 128, 455–500. Coggon R. and Holland T.J.B. (2002) Mixing properties of phengitic micas and revised garnet-phengite thermobarometers. Journal of Metamorphic Geology 20, 683–696. Connolly J.A.D. (1990) Multivariable phase-diagrams: an algorithm based on generalized thermodynamics. American Journal of Science 290, 666–718. Connolly J.A.D. (2005) Computation of phase equilibria by linear programming: A tool for geodynamic modeling and its application to subduction zone decarbonation. Earth and Planetary Science Letters 236, 524–541. Dale J., Powell R., White R.W., Elmer F.L. and Holland T.J.B. (2005) A thermodynamic

model for Ca-Na clinoamphiboles in Na2O-CaO-FeO-MgO-Al2O3-SiO2-H2O-O for petrological calculations. Journal of Metamorphic Geology 23, 771–791. Davis P.B. and Whitney D.L. (2008) Petrogenesis and structural petrology of high-pressure metabasalt pods, Sivrihisar, Turkey. Contributions to Mineralogy and Petrology 156, 217–241. Droop G.T.R. and Harte B. (1995) The effect of Mn on the phase relations of medium-grade pelites: Constraints from natural assemblages on petrogenetic grid topology. Journal of Petrology 36, 1549–1578. Ernst W.G. (1970) Tectonic Contact between the Franciscan Mélange and the Great Valley Sequence—Crustal Expression of a Late Mesozoic Benioff Zone. Journal of Geophysical Research 75, 886–901. Evans T.P. (2004) A method for calculating effective bulk composition modification due to crystal fractionation in garnet-bearing schist: implications for isopleth thermobarometry. Journal of Metamorphic Geology 22, 547–557. Federico L., Crispini L., Scambelluri M. and Capponi G. (2007) Ophiolite mélange zone records exhumation in a fossil subduction channel. Geology 35, 499–502. Franz G. and Liebscher A. (2004) Physical and chemical properties of the epidote minerals - An introduction. Epidotes 56, 1–81. Fuhrman M.L. and Lindsley D.H. (1988) Ternary-feldspar modeling and thermometry. American Mineralogist 73, 201–215. Gaidies F., Abart R., De Capitani C., Schuster R., Connolly J.A.D. and Reusser E. (2006) Characterization of polymetamorphism in the Austroalpine basement east of the Tauern Window using garnet isopleth thermobarometry. Journal of Metamorphic Geology 24, 451–475. Gao J. and Klemd R. (2003) Formation of HP-LT rocks and their tectonic implications in the western Tianshan Orogen, NW China: geochemical and age constraints. Lithos 66, 1–22. Gao J., Klemd R., Zhang L., Wang Z. and Xiao X. (1999) P-T path of high-pressure/ low-temperature rocks and tectonic implications in the western Tianshan Mountains,

182

6. A common P-T evolution of eclogites and metasediments

NW China. Journal of Metamorphic Geology 17, 621–636. Gao J., Li M.S., Xiao X.C., Tang Y.Q. and He G.Q. (1998) Paleozoic tectonic evolution of the Tianshan Orogen, northwestern China. Tectonophysics 287, 213–231. Gao J., Long L.L., Klemd R., Qian Q., Liu D.Y., Xiong X.M., Su W., Liu W., Wang Y.T. and Yang F.Q. (2009) Tectonic evolution of the South Tianshan orogen and adjacent regions, NW China: geochemical and age constraints of granitoid rocks. International Journal of Earth Sciences 98, 1221–1238. Gerya T.V., Stockhert B. and Perchuk A.L. (2002) Exhumation of high-pressure metamorphic rocks in a subduction channel: A numerical simulation. Tectonics 21, 1506, 10.1029/2002tc001406. Green E., Holland T. and Powell R. (2007) An order-disorder model for omphacitic pyroxenes in the system jadeite-diopside-hedenbergite-acmite, with applications to eclogitic rocks. American Mineralogist 92, 1181–1189. Groppo C. and Castelli D. (2010) Prograde P-T Evolution of a Lawsonite Eclogite from the Monviso Meta-ophiolite (Western Alps): Dehydration and Redox Reactions during Subduction of Oceanic FeTi-oxide Gabbro. Journal of Petrology 51, 2489– 2514. Gross J., Burchard M., Schertl H.P. and Maresch W.V. (2008) Common high-pressure metamorphic history of eclogite lenses and surrounding metasediments: a case study of calc-silicate reaction zones (Erzgebirge, Germany). European Journal of Mineralogy 20, 757–775 Holland T.J.B. and Powell R. (1998) An internally consistent thermodynamic data set for phases of petrological interest. Journal of Metamorphic Geology 16, 309–343. Hollister L.S. (1966) Garnet Zoning: An Interpretation Based on the Rayleigh Fractionation Model. Science 154, 1647–1651. John T., Klemd R., Gao J. and Garbe-Schonberg C.D. (2008) Trace-element mobilization in slabs due to non steady-state fluid-rock interaction: Constraints from an eclogite-facies transport vein in blueschist (Tianshan, China). Lithos 103, 1–24. Klemd R., Bröcker M., Hacker B.R., Gao J., Gans P. and Wemmer K. (2005) New age constraints on the metamorphic evolution of the high-pressure/low-temperature belt in the western Tianshan mountains, NW China. Journal of Geology 113, 157–168. Klemd R., John T., Scherer E.E., Rondenay S. and Gao J. (2011) Changes in dip of subducted slabs at depth: Petrological and geochronological evidence from HP-UHP rocks (Tianshan, NW-China). Earth and Planetary Science Letters 310, 9–20. Klemd R., Matthes S. and Okrusch M. (1991) High-pressure relics in meta-sediments intercalated with the Weissenstein eclogite, Munchberg gneiss complex, Bavaria. Contributions to Mineralogy and Petrology 107, 328–342. Klemd R., Matthes S. and Schussler U. (1994) Reaction textures and fluid behaviour in very high pressure calc-silicate rocks of the Münchberg gneiss complex, Bavaria, Germany. Journal of Metamorphic Geology 12, 735–745. Klemd R., Schröter F.C., Will T.M. and Gao J. (2002) P-T evolution of glaucophane-omphacite bearing HP-LT rocks in the western Tianshan Orogen, NW China: new evidence for 'Alpine-type' tectonics. Journal of Metamorphic Geology 20, 239–254.

183

6. A common P-T evolution of eclogites and metasediments

Konrad-Schmolke M., O'Brien P.J., de Capitani C. and Carswell D.A. (2008) Garnet growth at high- and ultra-high pressure conditions and the effect of element fractionation on mineral modes and composition. Lithos 103, 309–332. Leake B.E., Woolley A.R., Arps C.E.S., Birch W.D., Gilbert M.C., Grice J.D., Hawthorne F.C., Kato A., Kisch H.J., Krivovichev V.G., Linthout K., Laird J., Mandarino J.A., Maresch W.V., Nickel E.H., Rock N.M.S., Schumacher J.C., Smith D.C., Stephenson N.C.N., Ungaretti L., Whittaker E.J.W. and Guo Y.Z. (1997) Nomenclature of amphiboles: Report of the subcommittee on amphiboles of the International Mineralogical Association, commission on new minerals and mineral names. American Mineralogist 82, 1019–1037. Li J.L., Gao J., John T., Klemd R. and Su W. (2013) Fluid-mediated metal transport in subduction zones and its link to arc-related giant ore deposits: Constraints from a sulfide-bearing HP vein in lawsonite eclogite (Tianshan, China). Geochimica et Cosmochimica Acta 120, 326–362. Li J.L., Klemd R., Gao J. and Meyer M. (2012) Coexisting carbonate-bearing eclogite and blueschist in SW Tianshan, China: petrology and phase equilibria. Journal of Asian Earth Sciences 60, 174–187. Li J.L., Klemd R., Gao J. and Meyer M. (2014) Compositional zoning in dolomite from lawsonite-bearing eclogite (SW Tianshan, China): Evidence for prograde metamorphism during subduction of oceanic crust. American Mineralogist 99, 206– 217. Liu J.B., Ye K., Maruyama S.N., Cong B.L. and Fan H.R. (2001) Mineral inclusions in zircon from gneisses in the ultrahigh-pressure zone of the Dabie Mountains, China. Journal of Geology 109, 523–535. Lü Z., Bucher K. and Zhang L. (2013) Omphacite-bearing calcite marble and associated coesite-bearing pelitic schist from the meta-ophiolitic belt of Chinese western Tianshan. Journal of Asian Earth Sciences 76, 37–47. Lü Z., Bucher K., Zhang L. and Du J. (2012) The Habutengsu metapelites and metagreywackes in western Tianshan, China: metamorphic evolution and tectonic implications. Journal of Metamorphic Geology 30, 907–926. Lü Z., Zhang L., Du J. and Bucher K. (2009) Petrology of coesite-bearing eclogite from Habutengsu Valley, western Tianshan, NW China and its tectonometamorphic implication. Journal of Metamorphic Geology 27, 773–787. Lü Z. and Zhang L.F. (2012) Coesite in the eclogite and schist of the Atantayi Valley, southwestern Tianshan, China. Chinese Science Bulletin 57, 1467–1472. Lü Z., Zhang L.F., Du J.X. and Bucher K. (2008) Coesite inclusions in garnet from eclogitic rocks in western Tianshan, northwest China: Convincing proof of UHP metamorphism. American Mineralogist 93, 1845–1850. Massonne H.J. and Schreyer W. (1987) Phengite geobarometry based on the limiting assemblage with K-feldspar,phlogopite and quartz. Contributions to Mineralogy and Petrology 96, 212–224. Meyer M., Klemd R. and Konopelko D. (2013) High-pressure mafic oceanic rocks from the Makbal Complex, Tianshan Mountains (Kazakhstan & Kyrgyzstan): Implications for the metamorphic evolution of a fossil subduction zone. Lithos 177,

184

6. A common P-T evolution of eclogites and metasediments

207–225. Morimoto N., Fabries J., Ferguson A.K., Ginzburg I.V., Ross M., Seifert F.A., Zussman J., Aoki K. and Gottardi G. (1988) Nomenclature of pyroxenes. American Mineralogist 73, 1123–1133. Rebay G., Powell R. and Diener J.F.A. (2010) Calculated phase equilibria for a morb composition in a P-T range, 450-650 degrees C and 18-28 kbar: the stability of eclogite. Journal of Metamorphic Geology 28, 635–645. Spear F.S. and Franz G. (1986) P-T evolution of metasediments from the Eclogite Zone, south-central Tauern-Window, Austria. Lithos 19, 219–234. Stüwe K. (1997) Effective bulk composition changes due to cooling: a model predicting complexities in retrograde reaction textures. Contributions to Mineralogy and Petrology 129, 43–52. Su W., Gao J., Klemd R., Li J.L., Zhang X., Li X.H., Chen N.S. and Zhang L. (2010) U-Pb zircon geochronology of Tianshan eclogites in NW China: implication for the collision between the Yili and Tarim blocks of the southwestern Altaids. European Journal of Mineralogy 22, 473–478. Tian Z.L. and Wei C.J. (2013) Metamorphism of ultrahigh-pressure eclogites from the Kebuerte Valley, South Tianshan, NW China: phase equilibria and P–T path. Journal of Metamorphic Geology 31, 281–300. van der Straaten F., Schenk V., John T. and Gao J. (2008) Blueschist-facies rehydration of eclogites (Tian Shan, NW-China): Implications for fluid-rock interaction in the subduction channel. Chemical Geology 255, 195–219. von Huene R. and Scholl D.W. (1991) Observations at convergent margins concerning sediment subduction, subduction erosion, and the growth of continental crust. Reviews of Geophysics 29, 279–316. Wei C.J., Powell R. and Zhang L.F. (2003) Eclogites from the south Tianshan, NW

China: petrological characteristic and calculated mineral equilibria in the Na2O-CaO- FeO-MgO-Al2O3-SiO2-H2O system. Journal of Metamorphic Geology 21, 163–179. Wei C.J., Wang W., Clarke G.L., Zhang L.F. and Song S.G. (2009) Metamorphism of High/ultrahigh-pressure Pelitic-Felsic Schist in the South Tianshan Orogen, NW China: Phase Equilibria and P-T Path. Journal of Petrology 50, 1973–1991. Wei C.J., Tian Z.L. and Zhang L.F. (2013) Modelling of peak mineral assemblages and P-T conditions for high-pressure and ultrahigh-pressure eclogites. Chinese Science Bulletin 58, 2159–2164, doi: 10.1360/972013-605. Whitney D.L. and Evans B.W. (2010) Abbreviations for names of rock-forming minerals. American Mineralogist 95, 185–187. Xiao W.J., Windley B.F., Huang B.C., Han C.M., Yuan C., Chen H.L., Sun M., Sun S. and Li J.L. (2009) End-Permian to mid-Triassic termination of the accretionary processes of the southern Altaids: implications for the geodynamic evolution, Phanerozoic continental growth, and metallogeny of Central Asia. International Journal of Earth Sciences 98, 1189–1217. Yang X., Zhang L., Tian Z. and Bader T. (2013) Petrology and U–Pb zircon dating of coesite-bearing metapelite from the Kebuerte Valley, western Tianshan, China. Journal of Asian Earth Sciences 70–71, 295–307.

185

6. A common P-T evolution of eclogites and metasediments

Zhang L., Ellis D.J., Arculus R.J., Jiang W. and Wei C. (2003) 'Forbidden zone' subduction of sediments to 150 km depth - the reaction of dolomite to magnesite plus aragonite in the UHPM metapelites from western Tianshan, China. Journal of Metamorphic Geology 21, 523–529.

186

7. Summary and significance

7 Summary and significance

This thesis contributes to enhance the understanding of the high-pressure metamorphic evolution of carbonate-bearing metabasalts and fluid-rock interaction processes in subduction zones. The Chinese Tianshan HP-UHP metamorphic belt is interpreted to constitute a former oceanic subduction zone complex. Carbonate-bearing high-pressure eclogites and blueschists are commonly observed in this type of terrane, suggesting that the precursor basaltic oceanic crust has undergone significant hydrothermal alteration prior to subduction. However, metamorphism and P–T evolution of carbonate-bearing metabasalts have not been studied in detail in the Tianshan because the carbonate minerals are hardly attracted geologists‘ attentions and are usually ignored or treated as accessory minerals in phase diagram modeling due to the lack of good thermodynamic models and the complexity of the CO2-bearing system.

In the here presented thesis I studied coexisting carbonate-bearing eclogite and carbonate-omphacite-bearing blueschist using a pseudosection approach involving a

CO2-bearing system with fixed carbon oxide and water amounts (Chapter 3). Phase equilibrium modeling indicates that the eclogite and blueschist stabilized under the same peak metamorphic P-T conditions. This coexistence is believed to be due to different bulk-rock compositions. The present study constitutes the first attempt of pseudosection calculation in CO2-bearing system for Tianshan carbonate-bearing metabasalts. Carbonates in the Tianshan metabasalts are dominantely Fe-bearing dolomite. Interestingly, dolomite porphyroblast in some eclogites shows chemical compositional zoning, which is well defined by a continuous core–to–rim Mg increase and Fe–Mn decrease (Chapter 4). Thermodynamic modeling employing a new-developed Ca–Mg–Fe carbonate solution model demonstrates that the prograde growth zoning of dolomite is largely temperature dependent and that the carbonate underwent gradual aragonite → dolomite → dolomite+magnesite transitions during oceanic crust subduction. To the best of my knowledge the compositional zoning of

187

7. Summary and significance carbonate is very rare in metamorphic rocks, and this is the first study that was able to retrieve the prograde metamorphic evolution from high-pressure carbonates.

In addition, some carbonate-bearing host eclogites contain carbonate-sulfide-bearing high-pressure veins. Petrological and geochemical data of a sulfide-bearing vein, its selvage and host eclogites were studied in order to discuss the possible mobilization of both transition metals and trace elements in subduction-zone fluids (Chapter 5). Mixed fluids, consisting of an ―internal‖ and an ―external‖ fluid, trigger the dissolution- precipitation processes in the immediate wall-rock along the vein. LILE, HREE and some transition metal elements (e.g., Fe, Cu, Ni and Zn) were mobilized during those processes that led to the selvage formation. Thus one of the most significant results of the present study is that slab fluids can carry and transport significant amounts of transition metals. It is most likely that slab fluids strongly contribute to the metal flux into the arc magma systems finally resulting in the formation of arc-related ore deposits.

Furthermore, this thesis provides constraints on the geodynamic setting of the Tianshan high-pressure terrane. Using garnet isopleths thermobarometry, pseudosections indicate that the interlayered eclogite, marble and quartz-mica schist from a drill core have a common metamorphic evolution under high-pressure condition (Chapter 6). Thus the previously proposed tectonic subdivision in a northern ―internally coherent ultrahigh-pressure unit‖ and a southern ―coherent high-pressure unit‖ for the Tianshan terrane is unlikely. Further studies and detailed field mapping are essentially required to clarify the relationships of high-pressure and ultrahigh-pressure rocks as well as their exhumation mechanism.

188

Appendix

A1 Perple_X

Perple_X, written and maintained by Jamie Connolly (ETH Zürich), is a collection of Fortran77 programs for calculating and displaying phase diagrams, phase equilibria, and thermodynamic data. Website: http://www.perplex.ethz.ch/

Table A1 Solid solution models* of mineral phases used in this thesis

Symbol Phase Formula

3+ Amph(DPW) clinoamphibole Ca2(y+u+v)Nau+2(w+z)[MgxFe1-x]7-3u-2v-4(w+z)Fe 2zAl4y+3v+2wSi8-

(y+v)O22(OH)2, u+v+w+y+z≤1

3+ cAmph(DP) clinoamphibole Ca2(y+u+v)Nau+2(w+z)[MgxFe1-x]7-3u-2v-4(w+z)Fe 2zAl4y+3v+2wSi8-

(y+v)O22(OH)2, u+v+w+y+z≤1

Chl(HP) chlorite [MgxFewMn1-x-w]5-y+zAl2(1+y-z)Si3-y+zO10(OH)8, x+w≤1

Do(HP) dolomite CaMgxFe1-x(CO3)2

Ep(HP) epidote Ca2Al3-2xFe2xSi3O12OH

F fluid (H2O)x(CO2)1–x feldspar feldspar KyNaxCa1-x-yAl2-x-ySi2+x+yO8, x+y≤1

Gt(HP) garnet Fe3xCa3yMg3zMn3(1-x-y-z)Al2Si3O12, x+y+z≤1

Gt(WPH) garnet [FexCayMgzMn1-x-y-z]3[Fe1-wAlw]2Si3O12, x+y+z≤1

Mica(CHA) white mica KyCaxNa1-x-y(Mg1-vFev)zMgw TiwAl3+x-w-zSi3-x+zO10(OH)2,

x+y≤1, w+z≤y odCcMS(EF) carbonate CaxMgyFe1-x-yCO3, x+y≤1

2+ 3+ Omph(GHP) clinopyroxene Nay+w[CaMgxFe (1–x)]1–y–wAlyFe wSi2O6 *Details and citations of solid solution models see http://www.perplex.ethz.ch/

Appendix

Figure A1 Mg, Fe, Ca and Mn X-ray maps of garnet from interlayered eclogite (Sample KP1-6; chapter 6)