Volume 104, Issue 8

Subduction is a component of plate tectonics, which is widely accepted as having operated in a manner similar to the present-day back through the Phanerozoic Eon. However, whether Earth always had plate tectonics or, if not, when and how a globally linked network of narrow plate boundaries emerged are matters of ongoing debate. Earth’s mantle may have been as much as 200–300 °C warmer in the Mesoarchean compared to the present day, which potentially required an alternative tectonic regime during part or all of the Archean Eon. Here we use a data set of the pressure ( P ), temperature ( T ), and age of metamorphic rocks from 564 localities that vary in age from the Paleoarchean to the Cenozoic to evaluate the petrogenesis and secular change of metamorphic rocks associated with subduction and collisional orogenesis at convergent plate boundaries. Based on the thermobaric ratio (TP), metamorphic rocks are classified into three natural groups: high T/P type ( T / P > 775 °C/GPa, mean T / P ~1105 °C/GPa), intermediate T/P type ( T / P between 775 and 375 °C/GPa, mean T / P ~575 °C/GPa), and low T/P type ( T / P < 375 °C/GPa, mean T / P ~255 °C/GPa). With reference to published thermal models of active subduction, we show that low T/P oceanic metamorphic rocks preserving peak pressures >2.5 GPa equilibrated at P – T conditions similar to those modeled for the uppermost oceanic crust in a wide range of active subduction environments. By contrast, those that have peak pressures <2.2 GPa may require exhumation under relatively warm conditions, which may indicate subduction of young oceanic lithosphere or exhumation during the initial stages of subduction. However, low T/P oceanic metamorphic rocks with peak pressures of 2.5–2.2 GPa were exhumed from depths where, in models of active subduction, the slab and overriding plate change from being decoupled (at lower P ) to coupled (at higher P ), possibly suggesting a causal relationship. In relation to secular change, the widespread appearance of low T/P metamorphism in the Neoproterozoic represents a “modern” style of cold collision and deep slab breakoff, whereas rare occurrences of low T/P metamorphism in the Paleoproterozoic may reveal atypical localized regions of cold collision. Low T/P metamorphism is not known from the Archean geological record, but the absence of blueschists in particular is unlikely to reflect secular change in the composition of the oceanic crust. In addition, the premise that the formation of lawsonite requires abnormally low thermal gradients and the postulate that oceanic subduction-related rocks register significantly lower maximum pressures than do continental subduction-related rocks, and imply different mechanisms of exhumation, are not supported. The widespread appearance of intermediate T/P and high T/P metamorphism at the beginning of the Neoarchean, and the subsequent development of a clear bimodality in tectono-thermal environments are interpreted to be evidence of the stabilization of subduction during a transition to a globally linked network of narrow plate boundaries and the emergence of plate tectonics.

We studied FeCO 3 using Fe K -edge X-ray absorption near-edge structure (XANES) spectroscopy at pressures up to 54 GPa and temperatures above 2000 K. First-principles calculations of Fe at the K -edge in FeCO 3 were performed to support the interpretation of the XANES spectra. The variation of iron absorption edge features with pressure and temperature in FeCO 3 matches well with recently reported observations on FeCO 3 at extreme conditions, and provides new insight into the stability of Fe-carbonates in Earth’s mantle. Here we show that at conditions of the mid-lower mantle, ~50 GPa and ~2200 K, FeCO 3 melts and partially decomposes to high-pressure Fe 3 O 4 . Carbon (diamond) and oxygen are also inferred products of the reaction. We constrained the thermodynamic phase boundary between crystalline FeCO 3 and melt to be at 51(1) GPa and ~1850 K. We observe that at 54(1) GPa, temperature-induced spin crossover of Fe 2+ takes place from low to high spin such that at 1735(100) K, all iron in FeCO 3 is in the high-spin state. A comparison between experiment and theory provides a more detailed understanding of FeCO 3 decomposition observed in X-ray absorption spectra and helps to explain spectral changes due to pressure-induced spin crossover in FeCO 3 at ambient temperature.

Granite-related wolframite-quartz veins are the world’s most important tungsten mineralization and production resource. Recent progress in revealing their hydrothermal processes has been greatly facilitated by the use of infrared microscopy and laser ablation-inductively coupled plasma-mass spectrometry (LA-ICP-MS) analysis of both quartz- and wolframite-hosted fluid inclusions. However, owing to the paucity of detailed petrography, previous fluid inclusion studies on coexisting wolframite and quartz are associated with a certain degree of ambiguity. To better understand the fluid processes forming these two minerals, free-grown crystals of intergrown wolframite and quartz from the giant Yaogangxian W deposit in South China were studied using integrated in situ analytical methods including cathodoluminescence (CL) imaging, infrared microthermometry, Raman microspectroscopy, and fluid inclusion LA-ICP-MS analysis. Detailed crystal-scale petrography with critical help from CL imaging shows repetition of quartz, wolframite, and muscovite in the depositional sequence, which comprises a paragenesis far more complex than previous comparable studies. The reconstruction of fluid history in coexisting wolframite and quartz recognizes at least four successive fluid inclusion generations, two of which were entrapped concurrently with wolframite deposition. Fluctuations of fluid temperature and salinity during precipitation of coexisting wolframite and quartz are reflected by our microthermometry results, according to which wolframite-hosted fluid inclusions do not display higher homogenization temperature or salinity than those in quartz. However, LA-ICP-MS analysis shows that both primary fluid inclusions in wolframite and quartz-hosted fluid inclusions associated intimately with wolframite deposition are characterized by strong enrichment in Sr and depletion in B and As compared to quartz-hosted fluid inclusions that are not associated with wolframite deposition. The chemical similarity between the two fluid inclusion generations associated with wolframite deposition implies episodic tungsten mineralization derived from fluids exhibiting distinct chemical signatures. Multiple chemical criteria including incompatible elements and Br/Cl ratios of fluid inclusions in both minerals suggest a magmatic-sourced fluid with the possible addition of sedimentary and meteoric water. Combined with microthermometry and Raman results, fluid chemical evolution in terms of B, As, S, Sr, W, Mn, Fe, and carbonic volatiles collectively imply fluid phase separation and mixing with sedimentary fluid may have played important roles in wolframite deposition, whereas fluid cooling and addition of Fe and Mn do not appear to be the major driving factor. This study also shows that fluid inclusions in both wolframite and coexisting quartz may contain a substantial amount of carbonic volatiles (CO 2 ± CH 4 ) and H 3 BO 3 . Ignoring the occurrence of these components can result in significant overestimation of apparent salinity and miscalculation of LA-ICP-MS elemental concentrations. We suggest that these effects should be considered critically to avoid misinterpretation of fluid inclusion data, especially for granite-related tungsten-tin deposits.

We conducted a series of monazite and xenotime dissolution experiments in a boron-bearing silicic melt at 1000–1400 °C and 800 MPa in a piston-cylinder apparatus. We present new measurements of monazite and xenotime solubility in hydrous (~3 wt% water), boron-bearing rhyolitic melts, as well as the diffusivities of the essential structural constituents of those minerals (LREE, P, and Y). We compare our results to the previous studies and discuss the implications of this study on the understanding of natural, silicic (granitic/rhyolitic) systems. We propose one equation describing the relationship between the solubility of xenotime and temperature in hydrous rhyolitic melts: ln⁡Y=18.3±0.3−125499±3356RT$$\begin{array}{} \displaystyle \ln Y=18.3\pm0.3-\frac{125499\pm 3356}{{\rm R}T} \end{array} $$ and another for monazite: lnΣLREE=18.6±1.5−129307±18163RT.$$\begin{array}{} \displaystyle \ln\! \Sigma{\rm LREE} =18.6\pm1.5-\frac{129307\pm 18163}{{\rm R}T}. \end{array} $$ In the presence of sufficient phosphorous, the concentration of LREE needed for monazite saturation is within the uncertainty of the Y concentration needed from xenotime saturation and indicates that in the case of equilibrium crystallization the mineral that forms will only depend on the availability of LREE and Y and HREE. Given the similarity of the solubility of xenotime to that of monazite, we propose that previously published models of monazite solubility in silicic melts can potentially be applied to xenotime, and could, like monazite, serve as a geothermometer. In the case of disequilibrium crystallization in front of rapidly growing crystals, Y will diffuse faster than LREE and xenotime will only crystallize when LREE are depleted. We also found that the diffusion of Y is greater than the diffusion of P from dissolving xenotime, unlike the similar diffusivities of LREE and P during monazite dissolution. The significant difference between Y and P diffusivities suggests that the components forming xenotime diffuse as separate entities rather than molecular complexes. The dissolution of phosphates (monazite, xenotime, apatite) in hydrous, silicic melts with the addition of boron leads to liquid-liquid immiscibility at high temperatures where the saturation values of P and either light rare earth elements or Y are in the weight percent range. Immiscibility is not observable at low, magmatic temperatures, most probably due to the lower concentrations of P 2 O 5 necessary for phosphate saturation at these conditions; however addition of other components, notably F, may result in liquid-liquid immiscibility at magmatic temperatures.

The relatively recent entry of field emission electron microprobes into the field of microanalysis provides another tool for the study of small features of interest (e.g., mineral and melt inclusions, exsolution lamellae, grain boundary phases, high-pressure experimental charges). However, the critical limitation for accurate quantitative analysis of these submicrometer- to micrometer-sized features is the relationship between electron beam potential and electron scattering within the sample. To achieve submicrometer analytical volumes from which X-rays are generated, the beam accelerating voltage must be reduced from 15–20 to ≤ 10 kV (often 5 to 7 kV) to reduce the electron interaction volume from ~3 to ~0.5 μm in common geological materials. At these low voltages, critical K α X-ray lines of transition elements such as Fe are no longer generated, so L X-ray lines must be used. However, applying the necessary matrix corrections to these L lines is complicated by bonding and chemical peak shifts for soft X-ray transitions such as those producing the Fe L α X-ray line. It is therefore extremely challenging to produce accurate values for Fe concentration with this approach. Two solutions have been suggested, both with limitations. We introduce here a new, simple, and accurate solution to this problem, using the common mineral olivine as an example. We also introduce, for the first time, olivine results from a new analytical device, the Extended Range Soft X-ray Emission Spectrometer.

We present the results of a comprehensive study on the concentrations of first-row transition elements (FRTE: Sc, Ti, V, Cr, Mn, Fe, Co, Ni, Cu, and Zn), as well as Ga and Ge, in liquidus olivine from 2.7–3.5 Ga old Al-undepleted and Al-depleted komatiites from the Kaapvaal and Zimbabwe Cratons in southern Africa, the Yilgarn Craton in Australia, and the Superior Craton in Canada. The sample set includes komatiites that remained sulfur-undersaturated upon emplacement, as well as komatiites that reached sulfide saturation owing to assimilation of crustal sulfur. All olivine grains display concentrations of Mn, Zn, Ge, Co, Fe, Mg, and Ni similar to the Bulk Silicate Earth (BSE) values, with significant negative anomalies in Sc, Ti, V, Cr, Ga, and Cu. Olivine from the studied Al-depleted komatiites displays on average higher 100×Ga/Sc ratios (>5) than olivine from Al-undepleted komatiites (≤5). Because garnet preferentially incorporates Sc over Ga, the data suggest that elevated Ga/Sc ratios in komatiitic olivine are indicative of garnet retention in the source region of komatiites, highlighting the potential of olivine trace element chemistry as a proxy for the depth of komatiite melting and separation of the magma from the melting residue. Copper concentrations in the studied olivine grains are controlled by sulfur saturation of the host komatiite during olivine crystallization. Olivine from sulfur-undersaturated komatiite systems displays Cu concentrations mostly between 1 and 10 ppm, whereas olivine from sulfide-bearing komatiites has Cu contents of <0.5 ppm. Because komatiites contain some of the world’s highest metal tenor magmatic Ni-Cu sulfide deposits, the Cu variability in olivine as a function of the sulfide-saturation state highlights a potential application of olivine chemistry in the exploration for sulfide ore deposits. Olivine from the Paleo-Archean (3.5–3.3 Ga) komatiites displays overall higher V/Sc ratios (V/Sc = 2.1 ± 0.96; 2 S.D.) than olivine from their Neo-Archean (2.7 Ga) counterparts (V/Sc = 1.0 ± 0.81, 2 S.D.). Vanadium and Sc behave similarly during partial melting of the mantle and are similarly compatible in majorite garnet. However, V is redox-sensitive and its compatibility in olivine increases as the system becomes less oxidized, whereas Sc is redox-insensitive. We argue that olivine from the studied Paleo-Archean komatiites crystallized from more reduced magmas than their Neo-Archean counterparts. Elevated Fe/Mn ratios in olivine from Paleo-Archean komatiites mimic the V/Sc signatures and are interpreted to reflect that Fe 2+ is more compatible in olivine than Fe 3+ . These results imply that V/Sc and Fe/Mn in komatiitic olivine may potentially provide insight into the evolution of the oxidation state of the Archean mantle. Additional studies that integrate the chemistry of komatiitic olivine with those of relict interstitial glass and melt/fluid inclusions are encouraged to fully understand and quantify the potential of FRTE in olivine as a proxy for the oxidation state of the mantle sources of komatiite magmas.

Conversion of hydrotalcite (Ht) to saponite was observed by hydrothermal alkaline alteration of metal oxides. The conversion was through a pathway of hydration-dissolution-precipitation. It involved several critical steps, including the construction of Ht from metal oxides, dissolution of Al 3+ from Ht, condensation of metasilicate anions with Ht, and finally crystallization of saponite. The condensation was favored by relatively low Mg/Al ratios of Ht, along with high concentrations of Al 3+ and silicate oligomers in the environment, resulting in highly crystalline saponite. The latter conversion was greatly accelerated by the isomorphous substitution of Al 3+ for Si 4+ in silicate oligomers. The substitution generated the extra negative charge and led to the aforementioned condensation with Ht surface, thereby promoting the formation of saponite TOT layers. During the process, CO 2 is an indispensable component. Initially intercalated as CO32−$\begin{array}{} \displaystyle {\rm CO_{3}^{2-}} \end{array}$ to form Ht, CO 2 was subsequently eliminated from the solid phase, and saponite formed when the layer charge was reversed. Thus, this study presents a novel formation mechanism of saponite from metal oxides via hydrotalcite and contributes to a better understanding of the crystallization, chemical stability, and transformation of Ht to saponite. The results are also relevant to evaluating metal availability and carbon cycling on the surface of the Earth.

Australasian Muong Nong-type tektites from the locality centered at 16.46150° N, 106.48917° E in Laos contain sporadic spherical heterogeneous sulfide inclusions less than 10 μm in diameter, which have been identified by electron probe microanalyzer and electron backscatter diffraction to represent a mixture of rare mineral shenzhuangite with a pyrrhotite polytype (possibly troilite). Contrary to type shenzhuangite found in the shocked L6 chondrite Suizhou, the mineral embedded in the tektite glass is nearly free of copper, making the composition close to its expected end-member that has not yet been found in nature; the empirical formula closest to ideal end-member composition based on four atoms per formula unit is Ni 1.007 Fe 0.998 Cu 0.016 Co 0.058 S 1.922 . The described occurrence also represents the first find of shenzhuangite in terrestrial material. Presented is the Raman spectrum for shenzhuangite with the tentative assignment of spectral bands based on the analogy with synthetic chalcopyrite-structured phases. The chemical composition of shenzhuangite close to NiFeS 2 is not consistent with any stable phase in the Fe–Ni–S system up to melting temperature. Available data so far on phase relations in this system do not allow unambiguous interpretation of conditions under which the sulfide association within inclusions had formed.

In this study, we have measured the compressional and shear wave velocities of (Mg 1.77 Fe 0.22 Ca 0.01 ) Si 2 O 6 natural orthopyroxene up to 13.5 GPa and 873 K using ultrasonic interferometry in conjunction with in situ synchrotron X-ray diffraction and imaging techniques. Previous acoustic experiments on orthoenstatite (OEn) MgSiO 3 indicated that both compressional and shear velocities ( V P and V S ) of OEn undergo continuous velocity softening above 9 GPa at room temperature, which has been attributed to the phase transition from OEn to the metastable, high-pressure clinoenstatite HPCEn2. For the first time, our results suggest that pressure-induced velocity softening can occur in natural orthopyroxene at high-temperature conditions relevant to the Earth’s cold subduction zones. Estimates of the impedance and velocity contrasts between orthopyroxene (Opx) and high-pressure clinopyroxene (HPCpx) have been calculated, and the possibility of this phase transformation being a plausible candidate for seismic X-discontinuities at depth around 250–350 km is re-evaluated.

The migrating fluid-mineral interface provides an opportunity for the uptake of trace elements as solid solutions in the newly formed crystal lattice during the non-equilibrium growth of the crystal. However, mineral nanoparticles could precipitate directly from the interfacial fluid when it evolves to a supersaturated situation. To better understand the role of mineral nanoparticles in this scenario, this study focuses on a well-documented magnetite with oscillatory zoning from a skarn deposit by using high-resolution transmission electron microscopy (TEM). Our results show that the Al concentration in magnetite measured on a micrometer-scale is caused by three different effects: Al solid solution, Al-rich nanometer-sized lamellae, and zinc spinel nanoparticles in the host magnetite. Here, we propose a genetic relationship among the three different phases mentioned above. At first, a continuous increase of the Al concentration in the interfacial fluid can be incorporated into the crystal lattice of magnetite forming a solid solution. During cooling in a later stage, aluminum in magnetite is oversaturated and exsolution of hercynite (Al-rich lamellae) occurs from the host magnetite. If the Al concentration at the fluid-magnetite interface still increases during further growth of magnetite, the substitution of Fe by Al has gradually reached saturation so that aluminum cannot be incorporated in the magnetite crystal structure any longer. Using the magnetite lattice as a template, nucleation of abundant zinc spinel nanoparticles occurs. This will, in turn, lead to a gradual depletion of Al concentration in the interfacial fluid until the available ions for zinc spinel nucleation and growth have been used up. As a result, the migrating fluid-magnetite interface will enrich the Al concentration in the interfacial fluid until the available ion concentration is sufficient for nucleation of zinc spinel phase again. The fluidmineral interface in this mechanism has been repeatedly utilized during crystal growth, providing an efficient way for the uptake of trace element from a related undersaturated bulk fluid.

The electronic spin transition of iron has been shown to strongly affect many thermoelastic properties of the host mineral. However, the response of melting temperatures to the spin transition remains largely unexplored. Here, we study the melting of lower mantle minerals, ferropericlase and bridgmanite, using Lindemann’s Law. This empirical law predicts a negligible melting temperature depression for Earth-relevant bridgmanite but a substantial depression for Earth-relevant ferropericlase across the spin transition of iron, consistent with extant experimental results. This melting depression can be explained within the framework of Lindemann’s Law for a Debye-like solid. The transition of iron from high- to low-spin configuration reduces the molar volume and the bulk modulus of the crystal, leading to a decrease in Debye frequency and consequently lowering the melting temperature. Thermodynamically, the melting depression likely derives from a more negative Margules parameter for a liquid mixture of high- and low-spin end-members as compared to that of a solid mixture. This melting depression across the spin transition of iron may be the process responsible for the formation of a deep molten layer during the crystallization of a magma ocean in the past, and a reduced viscosity layer at present.

Epidote spherulites are identified in a greenschist facies metavolcanic breccia enclosing a body of coesite-bearing eclogite at Ganghe in the Dabie ultrahigh-pressure metamorphic belt, east-central China. The epidote spherulites are formed by fibrous, radially arranged, and rare earth element (REE)-rich epidote crystals (ΣREE = 0.13–0.36 (or slightly higher) cations per formula unit, cpfu) and interfibrillar REE-poor epidote (ΣREE ≤ 0.10 cpfu). Some of the epidote spherulites are overgrown by radially arranged euhedral epidote crystals, which also form aggregates around preexisting quartz, plagioclase, and/or epidote. The epidote grains in such aggregates display oscillatory zoning, with REE content varying from a negligible amount to about 0.44 cpfu. Epidote also occurs as REE-poor individual euhedral crystals about the radial epidote aggregates or form loose clusters of randomly oriented crystals. Thermodynamic modeling of the mineral assemblages in the plagioclase pseudomorphs and in the matrix shows that they formed at greenschist facies metamorphic conditions (435–515 °C and 5–7 kbar). The epidote spherulites and radial euhedral epidote aggregates, however, do not belong to these assemblages and are non-equilibrium textures. They imply crystal growth under large degrees of supersaturation, with relatively low ratios of the diffusion rate ( D ) to the crystal growth rate ( G ). At low D / G ratios, spiky interfaces are favorable for diffusion-controlled growth and the resultant texture is a collection of spikes around a growth center, forming a spherulite. The change of epidote texture from spherulite to radial euhedral crystal aggregate implies a decrease of supersaturation and an increase of D / G , such that the crystal morphology was controlled by its crystallographic structure. The crystallization of the individual epidote grains corresponds to a further drop of supersaturation and a further increase of the D / G ratio, approaching to the equilibrium conditions. Transiently higher P-T conditions are inferred from the spherulite-forming reactions, relative to the P - T estimates for the equilibrium assemblages. The fibrous crystals in the spherulites having relatively large interfacial energies would inevitably adjust their shapes to equilibrium ones with low interfacial energies if the P - T -H 2 O conditions were maintained for a sufficiently long period of time. The non-equilibrium epidote aggregates likely formed in response to P - T and fluid pulses, possibly related to seismicity.

A new (Mg 0.5 Fe0.53+$\begin{array}{} \displaystyle {\rm Fe_{0.5}^{3+}} \end{array}$)(Si 0.5 Al0.53+$\begin{array}{} \displaystyle {\rm Al_{0.5}^{3+}} \end{array}$)O 3 LiNbO 3 -type phase was synthesized at 27 GPa and 2000 K under highly oxidized conditions using an advanced multi-anvil apparatus. Single crystals for this phase are 0.2–0.3 mm in dimension and maroon in color. They crystallize in a noncentrosymmetric structure with space group R 3 c and lattice parameters of a = b = 4.8720(6) Å, c = 12.898(2) Å, and V = 265.14(8) Å 3 . Fe 3+ and Al 3+ cations substitute into A (Mg 2+ ) and B (Si 4+ ) sites through charge-coupled substitution mechanism, respectively. The distortion of B O 6 ( B = Si 0.5 Al0.53+$\begin{array}{} \displaystyle {\rm Al_{0.5}^{3+}} \end{array}$) octahedra is 1.6 times higher than that of AO 6 (A = Mg 0.5 Fe0.53+$\begin{array}{} \displaystyle {\rm Fe_{0.5}^{3+}} \end{array}$) octahedra. This phase is probably recovered from bridgmanite at lower-mantle conditions by a diffusionless transition because of the displacement of A cations and distortion of B O 6 octahedra on releasing pressure. Bridgmanite can thus contain the FeAlO 3 component (50 mol%) beyond previously reported solubility limit (37 mol%). The present study shows that the Earth’s most abundant elements form a new Fe 3+ - and Al 3+ -rich LiNbO 3 -type compound from bridgmanite at lower mantle conditions. This new compound provides a new insight into the complicated crystal chemistry of LiNbO 3 -type phase/ bridgmanite and constrains the pressure and temperature conditions for shocked meteorites.