T43E-01 13:40h
Lithospheric Delamination as a Process to Introduce Water Into the Mantle
Delamination of the lower continental lithosphere has been inferred geologically from crustal heat flow increase, rapid regional uplift, and the appearance of signature high-potassium magmas such as lamprophyres and leucitites. Seismic studies also support delamination in specific areas. Though delamination is the suggested process in many geologic settings, no previous quantified study of the dewatering and melting consequences of delamination have been made. We use numerical models to investigate delamination of the lower continental lithosphere, resulting upper mantle flow patterns, topographic expression, and most significantly, the potential for the delaminating material to dewater as it falls. A dense lower-lithospheric region may develop through melt injection and transformation into eclogitic phase assemblages, or through thickening and cooling of a lithospheric root, such as in an arc setting. Lower crustal and mantle compositions that result from arc magmatism may exceed asthenospheric density by 50 to 250 kg/m$^{3}$ (about 1 to 5% density contrast) (Kay and Kay, 1993; Jull and Kelemen, 2001). Density contrasts in this range are sufficient to drive gravitational Rayleigh-Taylor instabilities. As the dense lithosphere drops by Rayleigh-Taylor instability it pulls the topography down by tens to hundreds of meters below the undisturbed zero value. Asthenosphere moves upward and laterally to replace the mass of the instability, possibly resulting in initial dry adiabatic melt erupting during topographic subsidence. Numerical models indicate that this adiabatic melting may produce hundreds to thousands of cubic kilometers of primary melt. Eruption during subsidence, observed in some geologic regions, is inconsistent with simple melting in an upwelling. Eventually the lithosphere rebounds to topographic levels above its initial elevation because of the hot buoyant asthenosphere now filling the dome in the lithosphere. Any subsequent eruption will likely occur on an uplifted surface. In an extinct arc setting it is reasonable to assume that the lithosphere itself is hydrous. As the delaminating piece falls through the mantle and heats conductively water and other volatiles will be forced out of it, just as hydrous fluids rise from a subducting slabs as the slab moves out of the stability regimes of various low-pressure and low-temperature hydrous minerals. The phase relations of Schmidt and Poli (1998) indicate that the descending instability is heated more than sufficiently to dehydrate its outermost several kilometers even if it consists solely of hydrous peridotite. This dehydrating fluid will percolate upward and hydrate the asthenosphere. The temperatures in the heating, descending delaminated material are also sufficient to melt the outermost kilometer if the material consists of wet peridotite. The descending material itself would produce magma from wet peridotite, or about twice the volume of magma from wet basalt. This hydrous fluid therefore can have two effects: it can enter the overlying mantle and produce a laterally-heterogeneous hydrated mantle, which may melt, or it can allow parts of the delaminating material itself to melt. The process therefore provides a number of geochemically different source regions: the hydrated delaminating material itself; hydrated mantle material in the wake of the delamination; and dry mantle material upwelling beneath the delaminated lithosphere. Models indicate that, over a range of lithospheric thicknesses and mantle potential temperatures, these processes can create a wide range of primary melt volumes, up to tens of thousands of cubic kilometers.
T43E-02 13:55h
The role of hydration in the eclogitization of the lower crust of central Chile and Argentina
Instead of marking a sharp boundary between the crust and mantle, the Moho beneath the western Sierras Pampeanas, Precordillera, and Andes in the flat slab region of central Chile and Argentina appears as a diffuse boundary that has been difficult to detect on receiver functions. This observation differs from the clear strong Moho signal detected by both P and S wave receiver functions across much of the eastern Sierras Pampeanas by the CHARGE PASSCAL array. The differences in crustal signature appear to correlate with a major suture zone between the eastern and western Sierras Pampeanas; to the east of which, a strong Moho can be clearly detected near 40 km depth, however to the west, we find only low-amplitude broad arrivals around 50-60 km depth and the presence of mid-crustal arrivals. Interestingly, this change in crustal thickness is not accompanied by a similar change in surface elevations. Numerous recent findings of weak, or absent, Moho signals have been attributed to hydration of the mantle wedge reducing mantle velocities and thereby diminishing the impedance contrast across the Moho. Our findings differ from other places where the Moho has been found to be weak as the anomalous region extends over 400 km to the east of the trench compared to less that 100 km found elsewhere. Ascribing the expansive weak Moho to more extensive serpentization in the flat slab region is inconsistent with detailed tomographic images of the upper mantle that identify increased shear wave speeds and low Vp/Vs, inconsistent with extensive hydration. It instead appears that the diminished impedance contrast across the Moho is more likely to result from increased lower crustal velocities. The increased thickness of the crust below the western Sierras Pampeanas, Precordillera, and Andes results in the lower crust reaching the eclogite stability field. Yet, as others have demonstrated elsewhere, rocks under these P-T conditions will metastably remain as granulite until exposed to hydrous fluids, at which point they undergo rapid eclogitization. The fluids present in the sediments and crust associated with the Juan Fernandez ridge, the subduction of which has been proposed to initiate the current flat subduction geometry, could provide the needed hydration to transform the metastable granulite into eclogite. The low Vp/Vs anomaly in the mantle is consistent with a dehydrated serpentinite, suggesting fluids transited through the mantle to reach the crust but no longer reside there. A dense lower crust composed of eclogite can readily explain the lack of a sharp Moho signal as well as account for the lack of correlation between crustal thickness and surface elevation.
T43E-03 14:10h
Eclogitic Pseudotachylytes in Subducted Oceanic Crust: are Intermediate-Depth Earthquakes Producing Pathways for Fluid Flow?
Subduction of oceanic crust and the underlying lithospheric mantle leads to densification and dehydration of the downgoing slab. Associated processes are slab pull which forces the subduction, intermediate-depth (70-300 km) seismicity, and arc magmatism. These processes are somehow linked because intermediate-depth earthquakes usually occur where dehydration reactions are expected to take place and furthermore released slab fluids trigger arc magmatism. Hence, it is suggested that high fluid pressures can cause dehydration embrittlement which may result in earthquakes. Another possibility to generate intermediate-depth earthquakes in slabs is shear instability. Such a process requires that very narrow shear zones and high shear strain-rates exist in slabs. However, up to now most of the knowledge concerning intermediate-depth earthquakes and their relations to intraslab fluid flow is based on theoretical approaches, experiments and interpretations of seismic data. Field evidences, such as pseudotachylytes (the only doubtless geological evidence for paleo-earthquakes), confirming the suggested hypotheses are rare. For the first time we now can present evidence that imply that first frictional failure with melting occurred. In eclogites of a paleo-subduction zone textural features occur, which are usually found in quenched melts of active volcanoes. We interpret them as representing former frictional melts and thus pseudotachylytes. The seismic failure was followed nearly contemporaneously by infiltration of an external fluid. Long-standing fluid flow produced hydraulic fracturing and continuous vein formation in the eclogites during ongoing burial. Furthermore, during their passage through the rocks the external fluids leached most of those trace elements from the eclogites which are required to produce the characteristic slab component of arc magmas.
T43E-04 14:25h
Constraining Mantle Discontinuity Structure Beneath the Laramie Array in SE Wyoming
The transition zone 'water - filter' model of Bercovici and Karato predicts that a layer of depleted mantle melt may pool above the 410 - km discontinuity, causing a persistent lateral negative velocity gradient above the discontinuity. Previous receiver function analysis of the mantle transition zone [Dueker, Fee in review] from the Billings Array (Montana) is suggestive of such a velocity gradient. We will present a receiver function analysis of the 410 and 660 km discontinuities underneath the Laramie Teleseismic Array (SE Wyoming) and examine evidence for the existence of the negative velocity gradient above the 410 discontinuity. The Laramie Array consisted of a dense (2 - km station spacing) deployment of 30 broadband seismometers in operation from Sept. 2000 to May 2001. Previous work with the Laramie Array data examined the Archean-Proterozoic Cheyenne Belt suture zone [Dueker, Yuan in preparation] as part of the CD-ROM project [Karlstrom et al., 2002].
T43E-05 14:40h
The Water Cycle Within Subducting Lithospheric Slabs
A significant amount of water can be transported deep into Earth's upper mantle by descending lithosphere: several minerals can indeed bring it down to the `transition zone', where spinels then dissolve it and transport it further down. Because lower mantle minerals can only incorporate a very small fraction of that water, it is released at the bottom of the transition zone and enters a `Slab Water Cycle' that is indeed mostly located within the subducting slab and between the bottom of the transition zone and Earth's surface. That released water percolates upward within the slab, and (a) is incorporated into descending spinel, (b) eventually ascends above the transition zone and up to the hydrosphere, and/or, (c) leaks into adjacent mantle. Percolating water readily accounts for deep-focus earthquakes and for some features of intermediate focus earthquakes. It should also have a significant impact on location and geochemistry of fluids and melts generated in the upper mantle
T43E-06 14:55h
The structure and composition of the 10-$\AA$ phase
10-$\AA$ phase is a high-pressure hydrous phase stable up to at least 9 GPa. It forms from the hydration of talc above $\sim$ 5 GPa, and is therefore a candidate for transporting significant amounts of H$_{2}$O deep into the Earth's mantle in subduction zones. However, its equilibrium structure and composition are uncertain. A fixed MgO:SiO$_{2}$ ratio of 3:4 has long been assumed, but the H$_{2}$O content appears to depend on run duration, increasing with time up to a likely maximum of x = 2 in the proposed formula Mg$_{3}$Si$_{4}$O$_{10}$(OH)$_{2}$.xH$_{2}$O. We have obtained new structural and compositional data of a sample of well-crystallised deuterated 10-$\AA$ phase, which reveal an unexpected structural complexity, with implications for its stability in the Earth and capacity to store H$_{2}$O in subduction zones. The sample was synthesised in a multi-anvil device at 6.5 GPa, 600 $\deg$C, 400 hr from a stoichiometric mixture of Mg(OD)$_{2}$ and SiO$_{2}$ glass. $^{29}$Si MAS NMR spectroscopy reveals the presence of a significant Q$^{2}$ Si peak in addition to the normal Q$^{3}$ Si, indicating that approximately 1/16 of the expected talc-like Si sites are not occupied by Si. $^{2}$H NMR spectroscopy distinguishes at least 3 different types of D sites, one of which is highly mobile and is likely due to molecular D$_{2}$O, with the others due to OD groups. IR and Raman spectra show several more OH environments than found in talc, suggesting that charge balancing of the Si deficit involves additional OH groups (0.8 extra OH p.f.u.). We propose that the Si "vacancies" are hydrogarnet-type groups within the tetrahedral sheets. Electron microprobe analysis shows a Si deficit of 0.17 atoms p.f.u., implying that the Si "vacancies" occur as single entities rather than as clusters. No superlattice reflections are visible in hk0 sections of single-crystal XRD patterns; however, this does not rule out ordering of vacancies within individual sheets. On the other hand, 3-fold superlattice reflections occur along c*, superimposed on a modulated diffuse scattering profile.
T43E-07 15:10h
Searching for the Mechanism of Dehydration in Clinopyroxene
Clinopyroxene (cpx) is a major constituent of the Earth's crust and mantle. From experimental studies as well as from studies of natural systems it is well known that cpx can store hydrogen in its structure with concentrations ranging from 30 to 3000 ppm H$_{2}$O (by weight). The highest concentrations are found in clinopyroxenes that come from high pressure region (e.g. Rossman 1996). But some of our high pressure candidates show concentrations as low as 30-50 ppm H$_{2}$O (Koch-M\"{u}ller et al., 2004). Does this reflect the mantle heterogeneity in water content or did the cpx loose their water during ascent? To shed light on this geochemically and petrologically important question we investigated cpx's from a wide range of occurrences by FTIR-, M\"{o}ssbauer- and optical spectroscopy, and single crystal X-ray diffraction. Water concentrations ranged from 60 to 700 ppm H$_{2}$O (by weight). In accordance with literature data we found Fe$^{2+}$ at M1 as well as on M2, and assigned in a first step Fe$^{3+}$ to M1. We simulated hydrogen loss through heating cpx in air up to 800 $\deg$C and analyzed the stepwise heated samples by the above mentioned methods. Depending on the type of OH defect, the grade of dehydration with increasing temperature is quite different. In samples relatively poor in Fe$^{3+}$ ($< $0.1 Fe$^{3+}$ pfu), hydrogen associated with vacancies at M2 (OH bands around 3450 cm$^{-1}$) starts to leave the structure at about 550 $\deg$C and are completely gone at 700 $\deg$C. Hydrogen concentrations associated with Al$^{3+}$ at the tetrahedral site (OH bands around 3525 cm$^{-1}$, Koch-M\"{u}ller et al., 2004) remain completely unaffected by heat treatment up to 700 $\deg$C. But all hydrogen vanished at about 775 $\deg$C. However, this is different in more Fe$^{3+}$-rich samples (about 0.2 Fe$^{3+}$ pfu). Here we observe a very intense OH band at 3515 cm$^{-1}$ plus shoulder at 3450 cm$^{-1}$. This intense band exhibit a different polarisation and behaves also different in its response to pressure and heating in comparison to those of the above mentioned samples. We assume that this band is due to vibrations of an OH dipole associated with Fe$^{3+}$ at M1 and a vacancy either at M1 or M2. The OH decrease during heating is in all samples positively correlated with decrease in Fe$^{2+}$ and increase in Fe$^{3+}$. Our data indicate that dehydration occurs through iron oxidation exclusively at the M2 site following the reaction:Fe$^{2+}$$_{M2} + OH$^{-}$ = Fe$^{2+}$$_{M2} + 1/2 H$_{2}$. Whereby the amount of Fe$^{2+}$ at M1 seems to remain unchanged. This conclusion is supported by the results of Skogby and Rossman (1989) who found reversibly that in cpx heated in hydrogen H incorporation was correlated with increase of Fe$^{2+}$ in the M2 site. Thus, if dehydration of cpx is indeed correlated with Fe$^{3+}$ at M2, the amount of Fe$^{3+}$ at M2 could give us an good estimate of the lost hydrogen. References Rossman G.R. (1996) Phys. Chem. Minerals, 23, 299-304. Koch-M\"{u}ller M., Matsyuk S.S., Wirth R. (2004) Am. Mineral., 89, 921-931. Skogy H., Rossman G.R. (1989) Am. Mineral., 74, 1059 - 1069.
T43E-08 15:25h
Phase transformation of water under the mantle conditions
Water plays very important role in the physics and chemistry of the planetary interiors. The H$_2$O incorporated in the mantle minerals may change their physical properties such as electrical conductivity, viscosity, and transformation kinetics. Despite numerous theoretical studies, its behavior at high pressures and temperatures (e.g., corresponding to deep planetary interiors) is still poorly understood. Recent studies have led to growing information about the polymorphism of H$_2$O at high pressures and at modest temperatures (300 K and below up to 100 GPa). The hydrogen bonding in the system gives rise to myriad phases (some 15) including stable and metastable crystalline and amorphous and possibly liquid) phases. We have conducted several experiments at high pressures and temperatures using {\it in situ} Raman spectroscopy and synchrotron diffraction techniques. Using both laser heating and external heating techniques we have observed a new transformation in both H$_2$O and D$_2$O at the mantle conditions. The transformation occurs at comparable pressures but lower temperatures to the predicted transition of ice VII to a superionic phase. It is also reversible; upon decompression normal ice VII is recovered at ~20 GPa and 300 K. This unexpected behavior along the melting curve overlaps the conditions present in the interiors of the Neptune and Uranus. As such, the transformation in water should be considered in understanding the magnetic fields, internal structure, and chemistry occurring at depth within the icy planets. More importantly, the existence of a new {\it P-T} phase may have the implications in understanding the solubility of H$_2$O in the mantle minerals, affecting the interpretation of the amount of water in the lower mantle and the physical properties of the hydrous minerals in the interior of the Earth.