Volume 105, Issue 4

A major unsolved problem of the Proterozoic is the genesis and tectonic evolution of the massif type anorthosites. The idea of large-scale floating of plagioclase crystals in a basaltic magma chamber eventually generating massif type anorthosite diapirs from the floatation cumulates is not supported by observations of the major layered basic complexes of Proterozoic to Eocene age. In this paper, we test and propose a new genetic process of anorthosite diapirism through Rayleigh-Taylor instability. We have carried out a numerical modeling study of parallel, horizontal, multiple layers of norite and anorthosite, in a model layered basic complex, behaving like Newtonian or non-Newtonian power law fluids in a jelly sandwich model of the continental lithosphere. We have shown that in this pressure-temperature-rheology configuration the model lithosphere generates Rayleigh-Taylor instability, which triggers diapirism of the anorthosite. In our model, the anorthosite diapirs buoyantly rise through stages of simple, symmetrical upwelling and pronounced bulbous growth to a full-blown mushroom-like form. This is the growth path of diapirs in nearly all analog and numerical previous studies on diapirism. Our anorthosite diapirs fully conform to this path. Furthermore, we demonstrate that the progressive diapirism brings in striking internal changes within the diapir itself. In the process, the lowermost anorthosite layer rises displacing the upper norite and anorthosite layers as progressively stretched and isolated segments driven to the margin of the rising diapir—a feature commonly seen in natural anorthosite massifs. We propose that a large plume-generated basaltic magma chamber may be ponded at the viscous lower crust or ductile-plastic upper mantle or further down in the weaker mantle of the jelly sandwich type continental lithosphere. The magma may cool and crystallize very slowly and resolve into a thick-layered basic complex with anorthosite layers. Rheologically behaving like Newtonian or non-Newtonian power law fluids, the layers of the basic complex with built-in density inversions would generate RT (Rayleigh-Taylor) instability. The RT instability would trigger a buoyant rise of the unstable anorthosite from the layered complex. The upward driven anorthosite, accumulated as anorthosite plutons, would gradually ascend across the lower and middle crust as anorthosite diapirs.

The origin of banding patterns in malachite [Cu 2 CO 3 (OH) 2 ] is an enduring problem in geology. While the bright green, vivid colors of this mineral have been attributed to the presence of Cu, no specific process has been proposed that can explain the perfect circularly concentric banding and geometrical shapes in botryoidal malachite. These patterns of concentric equidistant laminations are comparable to those arising from chemically oscillating experiments using the classical reactants of the Belousov-Zhabotinsky (B-Z) reaction. Through optical microscopy and micro-Raman imaging, this contribution documents that the geometric centers of the self-similar geometric patterns are often composed of organic matter. Carbon isotopes and trace elements further suggest that non-biological decarboxylation reactions of biological organic matter took place during diagenesis. Hence, the morphological and chemical characteristics of chemically oscillating reactions offer a plausible explanation for the formation of botryoidal malachite and abiotic environmental decarboxylation reactions.

The calcium-silicate garnets, grossular (Ca 3 Al 2 Si 3 O 12 ), andradite (Ca3Fe23+Si3O12),$\left( \text{C}{{\text{a}}_{\text{3}}}\text{Fe}_{\text{2}}^{\text{3+}}\text{S}{{\text{i}}_{\text{3}}}{{\text{O}}_{\text{12}}} \right),$and their solid solutions [Ca3(Alx,Fel−x3+)2Si3O12],$\left[ \text{C}{{\text{a}}_{3}}{{\left( \text{A}{{\text{l}}_{\text{x}}},\text{Fe}_{\text{l}-\text{x}}^{3+} \right)}_{2}}\text{S}{{\text{i}}_{\text{3}}}{{\text{O}}_{\text{12}}} \right],$can incorporate various amounts of structural OH – . This has important mineralogical, petrological, rheological, and geochemical consequences and extensive experimental investigations have focused on the nature of “water” in these phases. However, it was not fully understood how OH - was incorporated and this has seriously hampered the interpretation of different research results. IR single-crystal spectra of several nominally anhydrous calcium silicate garnets, both “end-member” and solid-solution compositions, were recorded at room temperature and 80 K between 3000 and 4000 cm –1 . Five synthetic hydrogarnets in the system Ca 3 Al 2 (SiO 4 ) 3 -Ca 3 Al 2 (H 4 O 4 ) 3 - Ca3Fe23+(SiO4)3-Ca3Fe23+(H2O4)3$\text{C}{{\text{a}}_{\text{3}}}\text{Fe}_{\text{2}}^{\text{3+}}{{\left( \text{Si}{{\text{O}}_{\text{4}}} \right)}_{\text{3}}}\text{-C}{{\text{a}}_{\text{3}}}\text{Fe}_{\text{2}}^{\text{3+}}{{\left( {{\text{H}}_{\text{2}}}{{\text{O}}_{\text{4}}} \right)}_{\text{3}}}$were also measured via IR ATR powder methods. The various spectra are rich in complexity and show several OH - stretching modes at wavenumbers between 3500 and 3700 cm –1 . The data, together with published results, were analyzed and modes assigned by introducing atomic-vibrational and crystal-chemical models to explain the energy of the OH - dipole and the structural incorporation mechanism of OH - , respectively. It is argued that OH - is located in various local microscopic- and nano-size Ca 3 Al 2 H 12 O 12 - and Ca3Fe23+H12O12−$\text{C}{{\text{a}}_{\text{3}}}\text{Fe}_{\text{2}}^{\text{3+}}{{\text{H}}_{\text{12}}}{{\text{O}}_{{{12}^{-}}}}$like clusters. The basic substitution mechanism is the hydrogarnet one, where (H 4 O 4 ) 4 – ↔ (SiO 4 ) 4– , and various local configurations containing different numbers of (H 4 O 4 ) 4– groups define the cluster type. Some spectra also possibly indicate the presence of tiny hydrous inclusion phases, as revealed by OH - modes above about 3670 cm –1 . They were not recognized in earlier studies. Published proposals invoking different hypothetical “defects” and coupled-substitution mechanisms involving H + are not needed to interpret the IR spectra, at least for OH - modes above about 3560 cm –1 . Significant mineralogical, petrological, and geochemical consequences result from the analysis and are discussed in the accompanying Part II (this issue) of the investigation.

The nominally anhydrous, calcium-silicate garnets, grossular (Ca 3 Al 2 Si 3 O 12 ), andradite (Ca3Fe23+Si3O12),$\left( \text{C}{{\text{a}}_{\text{3}}}\text{Fe}_{\text{2}}^{\text{3+}}\text{S}{{\text{i}}_{\text{3}}}{{\text{O}}_{\text{12}}} \right),$schorlomite Ca3Ti24+Si,Fe23+O12$\left( \text{C}{{\text{a}}_{\text{3}}}\text{Ti}_{\text{2}}^{\text{4+}}\left[ \text{Si,Fe}_{\text{2}}^{\text{3+}} \right]{{\text{O}}_{\text{12}}} \right)$and their solid solutions can incorporate structural OH - , often termed “water.” The IR single-crystal spectra of several calcium silicate garnets were recorded between 3000 and 4000 cm –1 . Spectroscopic results are also taken from the literature. All spectra show various OH - stretching modes between 3500 and 3700 cm –1 and they are analyzed. Following the conclusions of Part I of this study, the garnets appear to contain local microscopic- and nano-size Ca 3 Al 2 H 12 O 12 - and Ca3Fe23+H12O12-$\text{C}{{\text{a}}_{\text{3}}}\text{Fe}_{\text{2}}^{\text{3+}}{{\text{H}}_{\text{12}}}{{\text{O}}_{\text{1}{{\text{2}}^{\text{-}}}}}$like domains and/or clusters dispersed throughout an anhydrous “matrix.” The substitution mechanism is the hydrogarnet one, where (H 4 O 4 ) 4– ↔ (SiO 4 ) 4– , and various local configurations containing different numbers of (H 4 O 4 ) 4– groups define the cluster type. A single (H 4 O 4 ) group is roughly 3 Å across and most (H 4 O 4 )-clusters are between this and 15 Å in size. This model can explain the IR spectra and also other experimental results. Various hypothetical “defect” and cation substitutional mechanisms are not needed to account for OH - incorporation and behavior in garnet. New understanding at the atomic level into published dehydration and H-species diffusion results, as well as H 2 O-concentration and IR absorption-coefficient determinations, is now possible for the first time. End-member synthetic and natural grossular crystals can show similar OH - “band patterns,” as can different natural garnets, indicating that chemical equilibrium could have operated during their crystallization. Under this assumption, the hydrogarnet-cluster types and their concentrations can potentially be used to decipher petrologic (i.e., P - T - X ) conditions under which a garnet crystal, and the rock in which it occurs, formed. Schorlomites from phonolites contain no or very minor amounts of H 2 O (0.0 to 0.02 wt%), whereas Ti-bearing andradites from chlorite schists can contain more H 2 O (~0.3 wt%). Different hydrogarnet clusters and concentrations can occur in metamorphic grossulars from Asbestos, Quebec, Canada. IR absorption coefficients for H 2 O held in hydrogrossular-and hydroandradite-like clusters should be different in magnitude and this work lays out how they can be best determined. Hydrogen diffusion behavior in garnet crystals at high temperatures is primarily governed by the thermal stability of the different local hydrogarnet clusters at 1 atm.

Chromite is one of the earliest crystallized minerals from mafic melts and has been used as an important “petrogenetic indicator.” Its composition may be modified by interaction with intercumulate melt and adjacent minerals. Thus, chromite in mafic-ultramafic rocks contains clues to the geochemical affinity, evolution, and mantle source of its parent magmas. The Devonian Xiarihamu intrusion, located in the East Kunlun Orogenic Belt in the northern Tibet Plateau, China, hosts a very large disseminated Ni-Co sulfide deposit. This study focuses on geochemistry of the chromite enclosed in olivine of ultramafic rocks of the intrusion. Enrichments in Mg and Al in the rim of the chromite indicate only minor effects of alteration on the compositions of the chromite. The chromites enclosed in the olivines with forsterite percentage (Fo) lower than 87 are characterized by large variations in major and trace elements, such as large ranges of Cr·100/(Cr+Al) (Cr# = 15–47), Mg·100/(Mg+Fe 2 + ) (Mg# = 41–65), and Al 2 O 3 (= 26–53 wt%) as well as 380–3100 ppm V, 70–380 ppm Ga, and 1100–16300 ppm Zn. The chromites display positive correlations between Cr/(Cr+Al) and Ti, Mn, V, Ga, and Sc, inconsistent with fractional crystallization but indicative of an interaction between the chromites, intercumulate melts and hosting minerals. In contrast, chromites hosted in olivine with Fo > 87 in harzburgite have small variations in Cr# (ranging from 37 to 41), Mg# (48 to 51), and Al 2 O 3 (30 to 35 wt%) as well as restricted variation in trace elements, indicating relatively weak interaction with trapped liquid and adjacent phases; these compositions are close to those of the most primitive, earliest crystallized chromites. The most primitive chromite has similarities with chromite in mid-ocean ridge basalt (MORB) in TiO 2 and Al 2 O 3 contents (0.19–0.32 and 27.9–36.3 wt%, respectively) and depletion of Sc and enrichment of Ga and Zn relative to MORB chromite. The geochemistry of the chromite indicates a partial melting of the asthenospheric mantle that was modified by melts derived from the subduction slab at garnet-stable pressures.

Approximately 700 diamond crystals were identified in volcanic (mainly pyroclastic) rocks of the Tolbachik volcano, Kamchatka, Russia. They were studied with the use of SIMS, scanning and transmission electron microscopy, and utilization of electron energy loss spectroscopy and electron diffraction. Diamonds have cube-octahedral shape and extremely homogeneous internal structure. Two groups of impurity elements are distinguished by their distribution within the diamond. First group, N and H, the most common structural impurities in diamond, are distributed homogeneously. All other elements observed (Cl, F, O, S, Si, Al, Ca, and K) form local concentrations, implying the existence of inclusions, causing high concentrations of these elements. Most elements have concentrations 3–4 orders of magnitude less than chondritic values. Besides N and H, Si, F, Cl, and Na are relatively enriched because they are concentrated in micro- and nanoinclusions in diamond. Mineral inclusions in the studied diamonds are 70–450 nm in size, round- or oval-shaped. They are represented by two mineral groups: Mn-Ni alloys and silicides, with a wide range of concentrations for each group. Alloys vary in stoichiometry from MnNi to Mn 2 Ni, with a minor admixture of Si from 0 to 5.20–5.60 at%. Silicides, usually coexisting with alloys, vary in composition from (Mn,Ni) 4 Si to (Mn,Ni) 5 Si 2 and Mn 5 Si 2 , and further to MnSi, forming pure Mn-silicides. Mineral inclusions have nanometer-sized bubbles that contain a fluid or a gas phase (F and O). Carbon isotopic compositions in diamonds vary from –21 to –29‰ δ 13 C VPDB (avg. = –25.4). Nitrogen isotopic compositions in diamond from Tolbachik volcano are from –2.32 to –2.58‰ δ 15 N Air . Geological, geochemical, and mineralogical data confirm the natural origin of studied Tolbachik diamonds from volcanic gases during the explosive stage of the eruption.

Cristobalite is a low-pressure, high-temperature SiO 2 polymorph that occurs as a metastable phase in many geologic settings, including as crystals deposited from vapor within the pores of volcanic rocks. Such vapor-phase cristobalite (VPC) has been inferred to result from silica redistribution by acidic volcanic gases but a precise mechanism for its formation has not been established. We address this by investigating the composition and structure of VPC deposited on plagioclase substrates within a rhyolite lava flow, at the micrometer to nanometer scale. The VPC contains impurities of the form [AlO 4 /Na + ] 0 —coupled substitution of Al 3+ charge-balanced by interstitial Na + —which are typical of cristobalite. However, new electron probe microanalysis (EPMA) element maps show individual crystals to have impurity concentrations that systematically decline from crystal cores-to-rims, and atom probe tomography reveals localized segregation of impurities to dislocations. Impurity concentrations are inversely correlated with degrees of crystallinity [observed by electron backscatter diffraction (EBSD), hyperspectral cathodoluminescence, laser Raman, and transmission electron microscopy (TEM)], such that crystal cores are poorly crystalline and rims are highly ordered tetragonal α-cristobalite. The VPC-plagioclase interfaces show evidence that dissolution-reprecipitation reactions between acidic gases and plagioclase crystals yield precursory amorphous SiO 2 coatings that are suitable substrates for initial deposition of impure cristobalite. Successive layers of cubic β-cristobalite are deposited with impurity concentrations that decline as Al-bearing gases rapidly become unstable in the vapor cooling within pores. Final cooling to ambient temperature causes a displacive transformation from β→α cristobalite, but with locally expanded unit cells where impurities are abundant. We interpret this mechanism of VPC deposition to be a natural proxy for dopant-modulated Chemical Vapor Deposition, where halogen-rich acidic gases uptake silica, react with plagioclase surfaces to form suitable substrates and then deposit SiO 2 as impure cristobalite. Our results have implications for volcanic hazards, as it has been established that the toxicity of crystalline silica is positively correlated with its purity. Furthermore, we note that VPC commonly goes unreported, but has been observed in silicic lavas of virtually all compositions and eruptive settings. We therefore suggest that despite being metastable at Earth’s surface, cristobalite may be the most widely occurring SiO 2 polymorph in extrusive volcanic rocks and a useful indicator of gas-solid reaction having occurred in cooling magma bodies.

26 Mg tracer diffusion coefficients were determined in single crystals of pure synthetic forsterite (Mg 2 SiO 4 ). Isotopically enriched powder sources both acted as the 26 Mg source and buffered the activities of silica (aSiO2)$\left( {{a}_{\text{Si}{{\text{O}}_{\text{2}}}}} \right)$at forsterite + protoenstatite (Mg 2 Si 2 O 6 ) (high aSiO2)$\left. {{a}_{\text{Si}{{\text{O}}_{2}}}} \right)$and forsterite + periclase (MgO) (low aSiO2).$\left. {{a}_{\text{Si}{{\text{O}}_{\text{2}}}}} \right).$Experiments were conducted at atmospheric pressure between 1250 and 1600 °C, and at oxygen fugacities (fO2S)$({{f}_{{{\text{O}}_{\text{2}}}}\text S)}$between 10 –12 bars (CO-CO 2 mix) and 10 –0.7 bars (air). The resulting diffusion profiles were measured along the three principal crystallographic axes ( a , b , and c ; ||[100], ||[010], ||[001]) using laser ablation–inductively coupled plasma–mass spectrometry (LA-ICP-MS), with a quadrupole mass spectrometer. These measurements were corroborated by ion microprobe using the sensitive high resolution ion microprobe-reverse geometry (SHRIMP-RG) instrument. Mg tracer diffusion is anisotropic, with D [001] > D [010] > D [100] , the difference in diffusion coefficients varying by about one order of magnitude at a given temperature with crystallographic orientation. Diffusion is faster in protoenstatite-buffered than periclase-buffered conditions, again with around one order of magnitude difference in diffusivity between buffering conditions. There is no apparent effect of fO2${{f}_{{{\text{O}}_{\text{2}}}}}$on diffusion. A global fit to all data, including data from Chakraborty et al. (1994) and Morioka (1981) yields the relationship: log10D=log10D0(m2s−1)+0.61(±0.03)log10aSiO2+−359(±10)kJ/mol2.303RT$${{\log }_{10}}D={{\log }_{10}}{{D}_{0}}\left( {{m}^{2}}{{s}^{-1}} \right)+0.61\left( \pm 0.03 \right){{\log }_{1}}_{0}{{a}_{\text{Si}{{\text{O}}_{\text{2}}}}}+\frac{-359\left( \pm 10 \right){\text{kJ}}/{\text{mol}}\;}{2.303\text{R}T}$$ where log 10 D 0 is –3.15 (±0.08), –3.61 (±0.02), and –4.01 (± 0.05) m 2 s –1 for the [001], [010], and [100] directions, respectively (1 s.d.). The LA-ICP-MS technique reproduces diffusion coefficients determined by SHRIMP-RG, albeit with slightly different absolute values of isotope ratios. This shows that LA-ICP-MS, which is both accessible and rapid, is a robust analytical method for such tracer diffusion studies.

The alteration of magmatic monazite and its consequences for monazite geochronology are explored in granitoids from the western part of the Ryoke belt (Iwakuni-Yanai area, SW Japan). Biotite-granite samples were collected in two plutons emplaced slightly before the main tectono-metamorphic event: the first one, a massive granite (Shimokuhara) adjoins schistose rocks affected by greenschist facies metamorphism; and the second, a gneissose granite (Namera) adjoins migmatitic gneiss that experienced upper-amphibolite facies conditions. Despite contrasting textures, the granite samples have similar mineral modes and compositions. Monazite in the massive granite is dominated by primary domains with limited secondary recrystallization along cracks and veinlets. It is variably replaced by allanite+apatite±xenotime±Th-U-rich phases. The outermost rims of primary domains yield a weighted average 206 Pb/ 238 U date of 102 ± 2 Ma while the Th-U phases show Th-U-Pb dates of 58 ± 5 and 15 to 14 ± 2–3 Ma. Monazite in the gneissose granite preserves sector- or oscillatory-zoned primary domains cross-cut by secondary domains enriched in Ca, Y, U, P, and containing numerous inclusions. The secondary domains preserve concordant 206 Pb/ 238 U dates spreading from 102 ± 3 to 91 ± 2 Ma while primary domain analyses are commonly discordant and range from 116 to 101 Ma. Monazite alteration textures in the two granites chiefly reflect differences in their post-magmatic histories. In the massive granite, monazite replacement occurred via a nearly stoichiometrically balanced reaction reflecting interaction with an aqueous fluid enriched in Ca+Al+Si±F during hydrothermal alteration of the granitic assemblage, likely below 500 °C. In the gneissose granite, a small amount of anatectic melt, probably derived from the neighboring metasedimentary rocks, was responsible for pseudomorphic recrystallization of monazite by dissolution-reprecipitation above 600 °C. Regardless of whether monazite underwent replacement or recrystallization, primary monazite domains preserve the age of magmatic crystallization for both plutons (102 ± 2 and 106 ± 5 Ma). Conversely, the age of monazite alteration is not easily resolved. Monazite replacement in the massive granite might be constrained using the Th-U-rich alteration products; with due caution and despite probable radiogenic Pb loss, the oldest date of 58 ± 5 Ma could be ascribed to chloritization during final exhumation of the granite. The spread in apparently concordant 206 Pb/ 238 U dates for secondary domains in the gneissose granite is attributed to incomplete isotopic resetting during dissolution-reprecipitation, and the youngest date of 91 ± 2 Ma is considered as the age of monazite recrystallization during a suprasolidus metamorphic event. These results reveal a diachronous, ca. 10 Ma-long high-temperature (HT) history and an overall duration of about 15 Ma for the metamorphic evolution of the western part of the Ryoke belt.

Smamite, Ca 2 Sb(OH) 4 [H(AsO 4 ) 2 ]·6H 2 O, is a new mineral species from the Giftgrube mine, Rauenthal, Sainte-Marie-Aux-Mines ore-district, Haut-Rhin department, France. It is a supergene mineral found in quartz-carbonate gangue with disseminated to massive tennantite-tetrahedrite series minerals, native arsenic, Ni-Co arsenides, and supergene minerals picropharmacolite, fluckite, and pharmacolite. Smamite occurs as lenticular crystals growing in aggregates up to 0.5 mm across. The new mineral is whitish to colorless, transparent with vitreous luster and white streak; non-fluorescent under UV radiation. The Mohs hardness is ~3½; the tenacity is brittle, the fracture is curved, and there is no apparent cleavage. The measured density is 2.72(3) g/cm 3 ; the calculated density is 2.709 g/cm 3 for the ideal formula. The mineral is insoluble in H 2 O and quickly soluble in dilute (10%) HCl at room temperature. Optically, smamite is biaxial (–), α = 1.556(1), β = 1.581(1), γ = 1.588(1) (white light). The 2 V (meas) = 54(1)°; 2 V (calc) = 55.1°. The dispersion is weak, r > v . Smamite is non-pleochroic. Electron microprobe analyses provided the empirical formula Ca 2.03 Sb 0.97 (OH) 4 [H 1.10 (As 1.99 Si 0.01 O 4 ) 2 ]·6H 2 O. Smamite is triclinic, P 1¯$\overline{1}$, a = 5.8207(4), b = 8.0959(6), c = 8.21296(6) Å, α = 95.8343(7)°, β = 110.762(8)°, γ = 104.012(7)°, V = 402.57(5) Å 3 , and Z = 1. The structure ( R obs = 0.027 for 1518 I >3σ I reflections) is based upon {Ca 2 (H 2 O) 6 Sb(OH) 4 [H(AsO 4 ) 2 ]} infinite chains consisting of edge-sharing dimers of Ca(H 2 O) 3 O 2 (OH) 2 polyhedra that share edges with Sb(OH) 4 O 2 octahedra; adjacent chains are linked by H-bonds, including one strong, symmetrical H-bond with an O–H bond-length of ~1.23 Å. The name “smamite” is based on the acronym of the Sainte-Marie-aux-Mines district.

Kroupaite (IMA 2017-031), ideally KPb 0.5 [(UO 2 ) 8 O 4 (OH) 10 ]·10H 2 O, is a new uranyl-oxide hydroxyl-hydrate mineral found underground in the Svornost mine, Jáchymov, Czechia. Electron-probe microanalysis (WDS) provided the empirical formula (K 1.28 Na 0.07 ) Σ1.35 (Pb 0.23 Cu 0.14 Ca 0.05 Bi 0.03 Co 0.02 Al 0.01 ) Σ0.48 [(UO 2 ) 7.90 (SO 4 ) 0.04 O 4.04 (OH) 10.00 ]·10H 2 O, on the basis of 40 O atoms apfu. Sheets in the crystal structure of kroupaite adopt the fourmarierite anion topology, and therefore kroupaite belongs to the schoepite-family of minerals with related structures differing in the interlayer composition and arrangement, and charge of the sheets. Uptake of dangerous radionuclides ( 90 Sr or 135 Cs) into the structure of kroupaite and other uranyl-oxide hydroxy-hydrate is evaluated based on crystal-chemical considerations and Voronoi-Dirichlet polyhedra measures. These calculations show the importance of these phases for the safe disposal of nuclear waste.

The unit-cell parameter a of antigorite (usually expressed as the polysome m value) has been determined as a function of temperature ( T ) and pressure ( P ) in the range of 600–650 °C, 25–45 kbar in weeklong piston-cylinder experiments. A well-characterized natural antigorite (with m = 16 and less abundant m = 15) was used as a starting material that coexisted with olivine, chlorite, Ti-humite, and aqueous fluid at run conditions. Transmission electron microscope (TEM) measurements on selected focused ion beam (FIB) wafers showed that antigorite m values after the experiments varied between 14 and 22. More than 40 punctual analyses for each run condition were acquired to determine the range and the primary m value. The most frequent antigorite m -value decreased systematically from 17–19 at 600 °C to 15–16 at 650 °C. The spacing of the m -isolines is getting narrower as the antigorite breakdown reaction is approached. The topology of the m -isolines is similar to that previously characterized for the simple MgO-SiO 2 -H 2 O (MSH) system. However, the isolines are shifted to about 50–100 °C higher temperatures due to the incorporation of Al into antigorite. Powder samples and FIB wafers of natural antigorite from the Tianshan UHP belt (China) with peak metamorphic conditions of ~35 kbar, ~520 °C were also investigated with TEM. Low Al-antigorite formed at peak metamorphic conditions displays a peak m value of 20–21, whereas high-Al antigorite formed during isothermal decompression displays a lower m value of 19. Combination of our results with the published data of m values from metamorphic antigorite that experienced various conditions allowed construction of a P-T - m diagram that can be used in future studies to better constrain formation conditions of serpentinites. The decrease of m values and the increase of Al in antigorite with increasing temperature result in small, continuous dehydration whereby the H 2 O content of antigorite changes from 12.4 to 12.1 wt%. Therefore, it is expected that a pore fluid is present during the prograde deformation of serpentinites. TEM observations showed that antigorite adjusted its Al content by segregation of chlorite at the nanoscale. Together with the observation that multiple m values are always present in a single sample, this result indicates that full equilibration of antigorite at the micrometer-scale is rare, with important implications for the interpretation of geochemical signatures obtained by in situ techniques.