Frank R. Schilling

Partial melting below the magmatic arc in the central Andes deduced from geoelectromagnetic field experiments and laboratory data

Abstract Magnetotelluric and geomagnetic deep soudings in northern Chile revealed a pronounced high conductivity zone (HCZ). Below the Western Cordillera, which constitutes the present magmatic arc with active volcanism of the South American continental margin, conductivities in the range of 1 S/m are observed. The anomalously high conductivities in a broad depth range from approximately 20 km to at least 60 km, are interpreted in terms of partial melting. Other geophysical observations, such as a zone of low seismic velocities (LVZ) at similar depths, high heat flow values (> 100 mW/m2) and a pronounced negative anomaly in the residual gravity field, are also considered. Impedance spectroscopic laboratory experiments up to and in the temperature range of partial melting were performed under controlled oxygen fugacities. At sub-solidus temperatures, electrical behavior is described by defect electrons with an activation energy of 1.34 eV and a conductivity of 2.5 mS/m at 900°C. Model calculations using a modified-brick-layer model (MBL) were compared with experimental observations. A good agreement between calculations and experiments is achieved with an electrical resistivity of the melt phase of 7 S/m at 1250°C assuming an activation energy of 1 eV. The same MBL model is used to calculate melt proportions beneath the Western Cordillera. Between 14 and 27 vol.% of interconnected melt are necessary to explain the observed HCZ. The stability of the melt rich crust is explained by a dynamic melting-crystallisation behavior during crustal anatexis and by magma filled dikes.

Elastic Shear Anisotropy of Ferropericlase in Earth's Lower Mantle

Seismic shear anisotropy in the lowermost mantle most likely results from elastic shear anisotropy and lattice preferred orientation of its constituent minerals, including perovskite, post-perovskite, and ferropericlase. Measurements of the elastic shear anisotropy of single-crystal (Mg0.9Fe0.1)O up to 69 gigapascals (GPa) show that it increased considerably across the pressure-induced spin transition of iron between 40 and 60 GPa. Increasing iron content further enhances the anisotropy. This leads to at least 50% stronger elastic shear anisotropy of (Mg,Fe)O in the lowermost mantle compared to MgO, which is typically used in geodynamic modeling. Our results imply that ferropericlase is the dominant cause of seismic shear anisotropy in the lower mantle.

Seismic, gravity and petrological evidence for partial melt beneath the thickened Central Andean crust (21–23°S)

On the basis of seismic refraction investigations and gravimetric data we have modelled the crustal structure of the southern Central Andes (21–23°S). A pronounced variation in crustal parameters is seen in N-S- and W-E-crossing seismic profiles over the entire Andean orogene, characterized by a crustal thickness of up to 70 km under the magmatic arc and backarc, strongly reduced seismic velocities and a Bouguer minimum of −450 mGal. Anomalously low velocities of 5.9–6.0 km/s in the deeper crust of the Western Cordillera and Altiplano regions lead to an over-compensation of the Bouguer minima resulting in values of crustal densities higher than estimates based purely on seismic velocity measurements. In an attempt to reconcile these differences, the behavior of crystalline rocks based on published laboratory data was studied under varying pressure and temperature conditions up to the range of partial melting. If the temperature is increased above the melting point, a rapid decrease in seismic velocity is accompanied by a slow decrease in density. For the Central Andes, a good fit of the observed and calculated Bouguer anomalies is obtained if the densities of the rocks from the low-velocity zone (LVZ) beneath the Western Cordillera and the Altiplano are varied. Model calculations lead to a velocity-density relation for partial molten rocks that allows the melt proportions of rocks to be estimated. Model calculations indicate that 15–20 vol.% of basaltic to andesitic melt at depth is necessary to explain the LVZ and Bouguer anomaly beneath the arc and parts of the backarc. High heat flow values (100 mW/m2) support the idea that large areas of the deeper Andean crust are strongly weakened by the presence of partially molten rocks, resulting in reduced seismic velocities, with the Western Cordillera, the active volcanic arc of the Andean mountain range, acting as a ductile buffer between the two more rigid crustal blocks of the forearc and backarc regions.

Quantifying partial melt fraction in the crust beneath the central andes and the Tibetan plateau

Abstract An interdisciplinary approach is used to quantify partial melt fractions and to infer the origin and distribution (melt structure) of melts located in the crust beneath the Central Andes and the Tibetan plateau. In these areas field observations of Low Velocity Zones (LVZ) and High Conductivity Zones (HCZ), which are commonly attributed to partial melting, are used to quantify melt fractions. Additional information is obtained from νP/νS ratios, seismic attenuation data, and heat flow density and gravity anomalies. These data accompanied by thermal modelling suggest that melts of mainly crustal origin are interconnected through dykes and veins. Experimental results and model calculations indicate that the minimum fraction of melt necessary to describe the LVZs and HCZs in the Central Andes and the Tibetan plateau is approximately 20 vol.%, and the melt has a non-ideal interconnectivity.

Partial Melting in the Central Andean Crust: a Review of Geophysical, Petrophysical, and Petrologic Evidence

The thickened crust of the Central Andes is characterized by several first-order geophysical anomalies that seem to reflect the presence of partial melts. Magnetotelluric and geomagnetic deep-sounding studies in Northern Chile have revealed a high conductivity zone (HCZ) beneath the Altiplano Plateau and the Western Cordillera, which is extreme both in terms of its size and integrated conductivity of > 20000 Siemens. Furthermore, this region is characterized by an extremely high seismic attenuation and reduced seismic velocity. The interrelation between the different petrophysical observations, in combination with petrological and heat-flow density studies, strongly indicates a huge area of partially molten rocks that is possibly topped with a thin, saline fluid film. The average melt fraction is deduced to be ∼20 vol.%, which agrees with typical values deduced from eroded migmatites. Based on the distribution and geochemical composition of Pliocene to Quaternary silicic ignimbrites in this area, this zone is thought to be dominated by crustally-derived rhyodacite melts with minor andesitic contribution. An interconnected melt distribution — typical for migmatites - would satisfy both the magnetotelluric and seismic observations. The high melt fraction in this mid-crustal zone should lead to strong weakening, which may be a main cause for the development of the flat topography of the Altiplano Plateau.

Single crystal elasticity of lawsonite

Abstract The single-crystal elastic moduli of lawsonite [CaAl2(Si2O7)(OH)2·H2O] were measured by Brillouin spectroscopy at ambient conditions. The Voigt-Reuss-Hill averaged aggregate elastic moduli are KS = 125(2) GPa and μ = 52(2) GPa, for the adiabatic bulk modulus and shear modulus, respectively. Our acoustic results resolve discrepancies between the bulk moduli obtained in earlier compression studies. Lawsonite has distinctive acoustic properties, being characterized by extremely high shear elastic anisotropy (74%), a high VP to VS ratio (VP/VS = 1.94), and a large Poisson’s ratio (σ = 0.318)

The influence of OH in coesite on the kinetics of the coesite-quartz phase transition

Abstract Metastable coesite is an important pressure indicator for ultrahigh-pressure rocks. However, in many cases coesite does not survive exhumation, but reacts back to quartz. Although it was shown experimentally that incorporation of H in coesite increases with increasing pressure, most coesite relics found in nature are essentially dry (i.e., OH concentrations are below detection limit, <1 wt ppm H2O). Thus, does the incorporation of H promote the back-reaction of coesite to quartz during exhumation? The influence of intrinsic OH on the kinetics of the coesite-quartz phase transition was determined using synthetic “dry” coesite with ≈ 10 wt ppm H2O and “wet” coesite with ≈ 105 wt ppm H2O. TEM analysis of the quenched samples proved the presence and absence of water in the .wet. and .dry. samples, respectively. The kinetics of the coesite-quartz transition was investigated in-situ using the multi-anvil apparatus MAX 80 at the Hamburger Synchrotron Radiation Laboratory (HASYLAB). The transition rates were measured by observing changes in selected diffraction line intensities as a function of time. The transformation and growth rates were derived using Cahns model of nucleation and growth at grain boundaries. Under the same experimental conditions the transformation rate of the “wet” coesite is more than ten times higher than that of the “dry” coesite. This difference may explain why OH-bearing natural coesite is rare. This study reveals the importance of structurally bound OH for the kinetics of phase transitions of nominally anhydrous minerals.

Thermal and rheological properties of granodioritic rocks from the Central Andes, North Chile

Abstract The thermal and elastic behavior of three granodioritic rocks from the North Chilean Coastal Cordillera has been studied at atmospheric pressure and at temperatures from 25°–1000°C. Properties investigated are thermal conductivity, thermal diffusivity, heat capacity, thermal expansion and elastic properties as a function of pressure as well as textural and structural changes during heating. Using the thermophysical data the thermal conductivities as a function of temperature and pressure are presented. Thermal conductivity decreases from 2.71 W/(m K) at surface conditions to 1.66 W/(m K) at 1 GPa and 800°C. Using the thermal conductivity as a function of temperature and pressure, temperature-depth distributions are calculated for different surface heat flow values. It is shown that conductive heat transport alone is not sufficient to explain a heat flow of > 100 mW/m2 often observed in magmatic arcs. It is thus concluded that convective heat transport plays an important role. Models of the temperature distribution and of rheologic patterns with depth in the Andean crust have been constructed. It is suggested that the brittle-ductile transition occurs at depths less than 20 km.

Temperature distribution in piston-cylinder assemblies Numerical simulations and laboratory experiments

Knowledge of the temperature distribution in piston-cylinder assemblies is desirable for equilibrium studies and experiments under transient conditions. The accurate determination of equilibrium properties needs a homogeneous temperature within the sample and transient experiments often require a defined thermal gradient. Knowledge of the temperature difference between the sample and thermocouple is another important constraint for quantitative experiments. To this end, the temperature distribution within various different piston-cylinder assembly designs was modeled with a specially designed 3D-Finite-Difference (3D-FD) program and compared to laboratory observations. For the 3D-FD simulation, the temperature and pressure dependence of the thermal properties of piston-cylinder-assembly materials (NaCl, CaF2, pyrophyllite, Au, graphite, NiCr-alloy) was considered. Furthermore, the T -dependent resistivity of graphite was used to model the local heat generation of the graphite heater. Experimentally determined and modeled temperature distributions are in good agreement. This indicates that the 3D-FD program is useful and an appropriate tool in the design of virtual piston-cylinder assemblies to be used in homogeneous temperature-distribution or defined thermal gradient experiments. The influences of temperature, pressure, assembly design, assembly materials (CaF2, NaCl), stepped versus straight wall heater, and presence versus absence of gold capsules on the temperature distribution within piston-cylinder assemblies are modeled and discussed. Different piston-cylinder configurations are presented, optimized for equilibrium studies and transient experiments, focusing on low and predefined T -gradients, respectively.

A standard-free pressure calibration using simultaneous XRD and elastic property measurements in a multi-anvil device

A key question to all high-pressure research arises from the reliability of pressure standards. There is some indication and discussion of an uncertainty of 10–20 % for higher pressures in all standards. Independent and simultaneous investigation of the dynamical (ultrasonic interferometry of elastic wave velocities) and static (XRD-measurement of the pressure-induced volume decline) compressibility on a sample reveal the possibility of a standard-free pressure calibration and, consequently an absolute pressure measurement. Ultrasonic interferometry is used to measure velocities of elastic compressional and shear waves in the multi-anvil high-pressure device MAX80 at HASYLAB Hamburg enabling simultaneous XRD and ultrasonic experiments. Two of the six anvils were equipped with lithium niobate transducers of 33.3 MHz natural frequency. NaCl was used as pressure calibrant, using the EoS of Decker (1971), and sample for ultrasonic interferometry. From the ultrasonic wave velocity data we calculated the compressibility of NaCl as a function of pressure independent from NaCl-pressure calibrant. The results were compared with data of static compression experiments up to 5 GPa (Bridgman, 1940) and up to 30 GPa (Birch, 1986) using experimental data from Boehler & Kennedy (1980) and Fritz et al. (1971). At 1.2 GPa and 5.3 GPa the results of static compression data agree with our velocity-derived compressibility data. In the range between 2 and 4GPa our dynamical data have 1.5–3% higher values. The pressure revealed according to Decker (1971) is in accordance to our standard-free pressure calibration. Consequently, up to 8 GPa the NaCl pressure standard has a reliability of at least 1%. However, there is some evidence that at higher pressures the inaccuracy of the NaCl standard becomes ≫ 1%.