«DISS. ETH NO. 20810 Elasticity and preferred orientations of ultra high pressure hydrous phases: implications for seismic anisotropy and deep water ...»
DISS. ETH NO. 20810
Elasticity and preferred orientations of ultra high pressure hydrous phases:
implications for seismic anisotropy and deep water recycling
in subduction zones
A dissertation submitted to
for the degree of
Doctor of Sciences
Angelika Dorothea Rosa
Diplom-Mineralogin, Ruprecht-Karls-Universität Heidelberg
Born on the 19th November 1983
in Munich, Germany
accepted on the recommendation of
Prof. Dr. Carmen Sanchez Valle ETH Zurich examiner Prof. Dr. Sébastien Merkel Université de Lille co-examiner Prof. Dr. Thomas S. Duffy Princeton University co-examiner Acknowledgements Acknowledgements First I would like to express my sincere gratitude to my thesis supervisor Prof. Dr. Carmen Sanchez-Valle, for giving me the opportunity to join this research project and for her time, enthusiasm and continuous support during this PhD thesis. I would also like to express my deepest thanks to Prof.
Dr. Merkel for his support, strong efforts during the synchrotron runs and fruitful discussions all along this thesis and to Prof. Dr. Thomas S. Duffy for accepting to examine this manuscript. I would like to acknowledge Prof. Dr. Max W. Schmidt and Prof. Dr. Peter Ulmer for supporting this thesis work and sharing with me their deep scientific and technical expertise.
I am grateful to Jingyun for her invaluable support, fruitful discussions and her major contribution to this thesis. In particular, I appreciated to learn from her the delicate sample preparation methods and to benefit from her experience in Brillouin scattering techniques and her approach to solve difficult technical problems. I would like to thank Ashima for her enthusiasm and strong efforts for synthesizing large single-crystals and for her great friendship. I also would like to express my gratitude to Arno for his important implication in the sample synthesis. I would like to express my deepest thank to Sujoy. His experience in high-pressure multi-anvil press techniques was invaluable for this thesis and made a great part of this work possible. Furthermore, I thank him for his support, incitements and fruitful discussions. I am particularly grateful for the assistance given by Dr. Michael Wörle and for his kindness. I also wish to acknowledge the technical support and the precious advices provided by Hansjörg Frei. The assistance byAndreas, Urs Graber, Peter Nievergelt, Eric Reusser, Lydia Zehnder and Bruno Zuercher was also greatly appreciated. I greatly appreciated the help of Rita, Rohit and Ute with the piston cylinder experiments. A special thank belongs to Claudia and Ursula for sorting out all my administrative formalities.
I would like to offer my deepest and warmest gratitude to my parents and brothers to whom I dedicate this work. Thank you for your love and invaluable support throughout all my life. I warmly thank Prof. Dr. Rainer Altherr and Prof. Dr. Ronald Miletich-Pawliczek for their strong support and for passing their passion in Earth science during my studies in Heidelberg. A special and warm thanks goes to Mohamed for all the good moments, his support, the fun and his passion for science that inspired me greatly.
I also would like to thank my friends in Germany: Julia, Helena, Melanie, Andrea and Gesine for sharing with me all situations in life. I would like to warmly thank Dominik and Andreas for their support, efforts, and all the funny and unforgettable adventures. A special thank goes to Sasha for all the fun and the precious and close friendship. I wish to thank various people for making the last four years a unique experience and for all the good moments, Rohit, Davide, Mattia, Omar, Remco, Rita, Santanu, Lukas, Stefanie, Sonja, Barbara, Chloé, Anastasia, Achille, Maike, Daniel, Tobias, Daniele, Ettore, Ying-Jui, Tamara, Ute, Dave, Esther and Mathew. A special thank goes to my office colleagues Davide and Marion for their help and the good friendship, to Shahrzad for her advices. I also wish to acknowledge Wim, Zoltan, Lucie, György and Christian for their scientific support and Isabelle, Carole, Caroline, Shawn and Nico for their support at the synchrotron and for their enthusiasm.
Finally, I want to thank Rohit, Tobias, both Benjamins, Jaap, Mauro, Matthias, Tilman, Samuel, Thibault, Philip, Martin, Rita, Sasha, Davide for all the good moments in the dolomites and in the lab.
Abstract Water plays a key role in earth dynamic processes including plate tectonics, melt generation and earthquakes. A detailed knowledge of plausible water transport mechanisms to great depth is therefore of central importance to understand the origin and nature of these processes.
It is admitted that water is most efficiently transported to depth via subduction of hydrated slab material. Seismic anomalies observed in deep slabs in turn have often been related to hydration. Indeed, numerous studies have pointed out the potential of hydrous phases to contribute to seismic anomalies including low velocity zones and shear wave splittings. The focus of this thesis lies therefore on the quantitative interpretation of seismic observations in deep slabs in terms of the hydration state. In order to understand the origin of seismic anomalies a cross-disciplinary approach from seismology, petrology and mineral physics is required. Therefore, the mechanical properties of water carriers and their stability fields have to be known. This thesis aims to experimentally investigate the mechanical properties of the high-pressure dense hydrous magnesium silicate phases namely superhydrous phase B (Mg10Si3H4O18) and phase D (ideal composition MgSi2H6O2). Both phases can account for up to 50 modal wt.% in very hydrous peridotites and may therefore significantly modify the seismic properties of an aggregate. In addition, the layered structure of ShyB and phase D may easily align in a non-hydrostatic stress field due to the anisotropic bonding strength and therefore form strong lattice preferred orientation textures (LPO) which are accepted as the main cause for seismic shear anisotropy in deep regions of the slab. The velocity contrast and the seismic anisotropy plausibly generated by these hydrous phases could be used to place tighter constrains on the hydration state of deep slabs.
We have synthesized high-quality single-crystals and powders in the multi-anvil apparatus (at up to 24 GPa and 1300 °C) and utilize diamond anvil cell techniques combined with Brillouin spectroscopy and synchrotron X-ray diffraction to precisely determine respectively their single-crystal elastic properties and deformation mechanisms up to 65 GPa, spanning the stability field of these phases in the slab. The results were used to model seismic velocities at relevant conditions in order to relate them to observed seismic velocities.
Ultimately, implications regarding the amount of water transported and stored in deep subduction zones were drawn.
The single-crystal elastic tensors of magnesium pure phase D (Mg1.1Si1.9H2.4O6) and of Al- and Fe-bearing phase D (Mg1.0Fe0.11Al0.03Si1.9H2.5O6) have been determined at ambient conditions. The obtained adiabatic bulk moduli are respectively KS0 = 154.8(3.2) GPa and µ = 104.3(2.1) GPa and KS0 = 158.4(3.9) GPa and µ = 104.3(2.7) GPa for Mg-phase D and AlFephase D. In addition, the single-crystal equation of state of Magnesium pure phase D has been determined up to 65 GPa using diamond anvil cell technique combined with synchrotron Xray diffraction. The obtained isothermal Birch-Murnaghan equation state parameters are KT0 = 151.4(1.2) GPa and K’= 4.9(1) and V0 = 85.80(5) Å3. In contrast with the results from previous powder X-ray diffraction studies, the compression curve obtained in this work does not display irregularities or discontinuities that could be ascribed to for hydrogen bonding symmetrization transition in the structure. The results on the single-crystal elasticity and EoS
solve the discrepancies between previously reported EoS parameters of phase D from Brillouin, static compression studies on powders and first principle studies.
The texture formation in polycrystalline phase D has been determined up to 48 GPa using radial X-ray diffraction diamond anvil cell techniques. The effect of Fe- and Alsubstitutions on the plastic properties has been investigated using three different phase D compositions. All samples preferentially aligned with the crystallographic c-axis parallel the compression axis resulting in a dominant 0001 texture that occurred already at low strains followed by a subsidiary 10-10 texture that evolved at higher strains. Therefore the plastic properties of phase D are only slightly affected by cation substitution, similarly to the elastic properties. The observed textures have been correlated mainly to dominant basal slip and first order pyramidal slip using visco-plastic self consistent (VPSC) modelling.
The single-crystal elastic tensor of superhydrous B (Mg10.4Si3.1H2.7O18) has been determined at ambient condition and up to 15 GPa using Brillouin scattering. By fitting the data to a third order finet equation of state the obtained adiabatic bulk elastic moduli are KS = 150(2) GPa and µ = 99(1) GPa and their pressure derivatives are (∂Ks/∂P)T0 = 4.2(2) and (∂µ/∂P)T0 = 1.40(5). These results resolve discrepancies to recent computational studies and static compressional studies using powder and single crystals.
The obtained elastic properties on ShyB and phase D have been combined to model the evolution of seismic velocities in a hydrous peridotite with various water contents and with increasing pressure. We modeled the contrast between the seismic wave velocities of a dry peridotite to hydrous peridotite containing either ShyB or phase D. The results show that hydration of peridotitic assemblies could lead to negative velocity anomalies at transition zone and lower mantle pressures. The strength of the negative velocity anomalies depend however crucially on the stable phase assembly and the hydration state and therefore on the abundance of hydrous phases.
The obtained plastic properties on phase D and modeled slip systems in ShyB were used to simulate the micro-texture and the seismic velocities in uniaxial compressed hydrous peridotites. We calculated the strength of a deformed hydrous peridotite aggregate by combining the obtained plastic and elastic properties of phase D. Phase D displays the lowest strength among coexisting phases, including Mg-perovskite, Ca-perovskite and periclase.
Phase D might therefore accommodate most of the strain under deformation regimes and forms the strongest texture in hydrous peridotite in deep subducted slabs. Basal slip has been inferred as the most plausible slip systems active in ShyB owing to its layered structure although experimental confirmation is yet to be provided. The simulations reveal that the amount of shear wave splitting displayed by a seismic wave crucially depends on the orientation of the ray path whereas the maximum splitting is observed in the plane perpendicular to the maximum strain axis. We found that the amount of splitting depends on the hydration state, the assumed shear stress, the strain distribution and the total strain which controls the slip systems active in the phase.
The results were applied to interpret seismic anomalies including negative velocity zones and shear anisotropies observed in the deep Tonga slab in terms of hydration. The
calculated velocities successfully predict all the observed anomalies in Tonga with a hydrated peridotite of 1.2 wt.% throughout the transition zone. This thesis has provided important new data on the elasticity and deformation mechanisms of two crucial hydrous phases. These data are important for the understanding of the fate of water in subducted slabs. The results point out that the abundance of phase D and ShyB introduce seismic anomalies including low velocities and high shear anisotropies which can be seismically identified. This thesis showed that the obtained physical properties can be used to interpret observed seismic anomalies beyond the transition zone in terms of hydration.
Zusammenfassung Wasser spielt eine Hauptrolle in den dynamischen Prozessen der Erde einschliesslich der Platentektonik, Erzeugung von Schmelzen und Erdbeben. Ein detailiertes Verständnis von plausiblen Wassertransportmechanismen zu grossen Tiefen ist daher von zentraler Bedeutung, um die Herkunft und Charakteristika dieser Prozesse zu verstehen. Es ist Allgemein anerkannt, dass das Wasser effizient via Subduction von hydratisiertem Plattenmaterial in die Tiefe transportiert wird. Seismischen Anomalien, die in tiefen subduzierten Platten beobachtet wurden, wurden hingegen oft mit Hydratisierung in Beziehung gesetzt. Tatsächlich haben mehrere Studien auf das Potential wasserhaltiger Phasen hingewiesen zu seismischen Anomalien beizusteuern, einschliesslich niedriger Geschwindigkeitszonen und Scherwellenaufspaltungen. Der Schwerpunkt dieser Doktorarbeit liegt daher auf der quantitativen Interpretation der seismischen Beobachtungen in tiefen Platten bezüglich zu deren Hydratisierungsgrad. Um den Ursprung der seismischen Anomalien zu verstehen benötigt man einen multidisziplinären Ansatz von Seismologie, Petrologie und Mineralphysik. Die mechanischen Eigenschaften von Wasserträgern in subduzierten Platten und deren Stabilitätsfelder müssen bekannt sein. Diese Doktorarbeit strebt die experimentelle Untersuchung der mechanischen Eigenschaften der Hochddruckphasen der dichten wasserhaltigen Magnesiumsilikate Phasen einschliesslich der Phase D (ideale Zusammensetzung Mg10Si3H4O18) und der super-wasserhaltige Phase B (ShyB, Mg10Si3H4O18) an. Beide Phasen können bis zu 50 modale Gew. % in einem sehr wasserhaltigen Peridotite ausmachen und daher möglicherweise beachtlich das seismische Signal beeinflussen. Zusätzlich zeigen die lagigen Strukturen von ShyB und Phase D auf Grund der anisotropen Bindungsstärken im Gitter eine Tendenz sich schnell und leicht in einem nicht-hydrostatischen Spannungsfeld auszurichten. Daher könnten diese Phasen starke präferentielle Gitterausrichtungstexturen ausbilden, die als der Hauptgrund für beobachtete seismische Scherwellenanisotropien akzeptiert sind. Die seismischen Geschwindigkeitsunterschiede und Scherwellenanisotropien, die möglicherweise durch die Anwesenheit dieser Phasen verursacht werden, könnten daher benutz werden um den Hydratisierungsstand tiefer subduzierter Platten enger einzugrenzen.