Structure and lattice vibrations of Mg-Al spinel solid solution

X-ray structure refinements have been made for nonstoichiometric (MgO · 3Al₂O₃) and stoichiometric Mg-Al spinels. Several structure variations with chemical composition have been observed and are discussed in relation to Al substitution in tetrahedral sites. Infrared reflection and Raman spectra of the single crystal of the nonstoichiometric spinel (MgO · 3Al₂O₃) have been measured and analyzed. The results obtained are compared with those reported for the stoichiometric sample. From the infrared and Raman frequencies reported for the stoichiometric Mg-Al spinel, which are partly complemented with our results, the effective ionic charges of the ions in MgAl₂O₄ have been estimated on the basis of the rigid ion model.     

Spinel-anthophyllite rocks of the Kokchetav massif (northern Kazakhstan)

Spinel-anthophyllite rocks that may be classified as ultrabasic low-Ca spinel amphibolites have been first discovered in the Kokchetav collision zone (northern Kazakhstan). They outcrop 2 km west of Enbek-Berlyk Village among schists and quartzites and are closely associated with spinel harzburgites and garnet pyroxenites. The main hosted minerals are spinel (hercynite) and anthophyllite. The rocks bear magnetite-hornblende-spinel-anthophyllite pseudomorphs with rounded and polygonal sections, which might have been resulted from the replacement of garnet grains. The prismatic anthophyllite crystals and scarce olivine relics contain elongate parallel spinel inclusions resembling spinel-olivine syntactic intergrowths in the Enbek-Berlyk spinel harzburgites. The spinel-anthophyllite rocks are similar to the associated spinel harzburgites in CaO, MnO, SiO₂, and Al₂O₃ contents but are richer in FeO and poorer in MgO (F = FeO/(FeO + MgO) = 57% against 35% in the harzburgites). Geological, mineralogical, and geochemical data suggest that the spinel-anthophyllite rocks formed during the isochemical contact metamorphism of garnet-bearing spinel harzburgites, which contained more FeO and less MgO than garnet-free harzburgites of the same area. Variations in FeO and MgO contents in both types of harzburgites seem to be due to different chemical compositions of the chlorite protoliths of these rocks.     

辽宁瓦房店金伯利岩中尖晶石族矿物种类划分及指示意义

晶体矿物学理论认为不同成岩环境金伯利岩中尖晶石族矿物由于形成物理化学条件不同,其晶体结构和化学成分会发生明显的变化,通过对无矿、贫矿、富矿金伯利岩岩管中的尖晶石族矿物晶胞参数和化学成分的测定,研究尖晶石族矿物化学成分和晶胞参数变化与无矿、贫矿、富矿金伯利岩的内在关系,可以提高金伯利岩型金刚石矿床找矿效率。为了确定辽宁瓦房店金伯利岩中的尖晶石族矿物种属,探讨辽宁瓦房店金伯利岩中尖晶石族矿物化学成分和晶胞参数与金伯利岩含矿性关系,本文运用电子探针波谱仪对50件尖晶石族矿物中的MgO、FeO、TiO2、Al2O3、MnO及Cr2O3进行微区化学成分分析,运用单晶X射线衍射仪对136个尖晶石族矿物晶胞参数进行测定。数据统计显示:瓦房店金伯利岩中尖晶石族矿物为铬铁矿和镁铬铁矿,以化学分子式中A、B组主要阳离子占位特征为基础,可把矿区的尖晶石族矿物划分为10个亚种;如果用尖晶石族矿物化学成分中Cr2O3与(Cr2O3+Al2O3)含量的比值Cr'来表示尖晶石族矿物与金伯利岩含矿性的关系,金伯利岩岩体含矿性由富矿→中等含矿→贫矿,相应岩体中尖晶石族矿物Cr'值分别为89.5%、83.4%~87.1%和70.2%,逐渐变低;从无矿金伯利岩岩体→贫矿和中等含矿金伯利岩岩体→富矿金伯利岩岩体,金伯利岩体中第一世代尖晶石族矿物晶胞参数分别为0.831~0.832 nm、0.834~0.836 nm、0.837 nm,有逐渐变大的趋势。本文认为,辽宁瓦房店金伯利岩中第一世代尖晶石族矿物晶胞参数大小和Cr'参数可以作为判断辽宁瓦房店金伯利岩含矿性的指示标型。     

New orthorhombic phases on the join NiAl2O4 (spinel analog)-Ni2SiO4 (olivine analog): Stability and implications to mantle mineralogy

Three new crystalline phases differing in Si/Al ratio have been synthesized from compositions along the join NiAl₂O₄-Ni₂SiO₄. Four reversible univariant equilibria involving these new phases plus Ni₂SiO₄ (olivine) have been located within the P-T region studied (1 atm–40 kb, 1000–1700° C); an invariant point occurs near 22 kb, 1150°C.All three new phases are orthorhombic. Precession photographs and electron microprobe analyses yield the following information:Phase I: 5NiO·3Al₂O₃·SiO₂ = 3NiAl₂O₄·Ni₂SiO₄, Pmma, a=5.67, b=11.51, c=8.10 (Å)Phase II: 7NiO·3Al₂O₃·2SiO₂ = 3NiAl₂O₄· 2Ni₂SiO₄, Imma, a=5.66, b=17.32, c=8.11Phase III: 3NiO· Al₂O₃· SiO₂ = NiAl₂O₄·Ni₂SiO₄, Imma, a=5.68, b=11.49, c=8.12Comparison with known structures suggests that these three phases plus NiAl₂O₄ spinel and high pressure Ni₂SiO₄ spinel belong to a homologous series based on a cubic close oxygen packing of the formula: M_2n O _n}-1 (T _n O_3n+1) where M and T are octahedrally and tetrahedrally coordinated cations, respectively. When n=1 the formula for spinel is obtained; n = 2 for phase I and phase III, both similar to the beta-phase of orthosilicates; and n = 3 for phase II which is related to the manganostibite structure.Similar phase equilibria and structural relations may occur on other joins of the aluminateorthosilicate type. Furthermore, the occurrence of such structural modifications between the spinel (aluminate) and olivine (orthosilicate) compositions suggests that there could be a corresponding polymorphic series between the olivine and spinel forms of orthosilicates.     

Oxidation mechanisms in Al–Mg–Fe spinels. A second stage: α–Fe2O3 exsolution

A two-stage model of oxidation was devised to explain the observed variations in crystallographic parameters in two artificially oxidized natural spinels. In the first stage, oxygen is added to the crystal boundary as cations are preserved, with Fe rising in total valence and vacant sites being formed. In the second stage, oxygen is preserved and α−Fe₂O₃ intergrowth occurs, at the expense of the oxygen of the parent spinel structure. On the basis of this model, crystallochemical formulae were calculated and cations partitioned in the various conditions. It was found that, both before and after oxidation, the spinel site population varies continuously in the direction of an increase in random charge distribution, depending on the increase of heat to the crystals. This trend was found to be reversible. Cation vacancies produced during oxidation are distributed between tetrahedral site T and octahedral site M. Received: 12 June 1997 / Revised, accepted: 17 February 1998     

Calculations of pressure-induced phase transitions in mantle minerals

The crystal structures and energies of SiO₂ stishovite, MgO periclase, Mg₂SiO₄ spinel, and MgSiO₃ perovskite were calculated as a function of pressure with the polarization-included electron gas (PEG) model. The calculated pressures of the spinel to perovskite phase transitions in the Mg₂SiO₄ and MgSiO₃ systems are 26.0 GPa and 27.0 GPa, respectively, compared to the experimental zero temperature extrapolations of 27.4 GPa and 27.7 GPa. The two oxide phases are found to be the most stable form in the pressure range 24.5 GPa to 31.5 GPa, compared to the experimental zero temperature extrapolation of 26.7 GPa to 28.0 GPa. The volume changes associated with the phase transitions are in good agreement with experiment. The transition pressures calculated with the PEG model, which allows the ions to distort from spherical symmetry, are in much better agreement with experiment than those calculated with the modified electron gas (MEG) model, which constrains the ions to be spherical.     

The influence of ionic radii of cations and covalent forces on the unit cell dimensions of spinels

Fe-, Cr- and Al-spinels were synthesized and their unit cell sizes determined by means of X-rays. Differential thermal curves show that the magnetic inversion of Fe₂O₃ at 680° C accelerates the formation of the ferrites when the constituent oxides are heated together.A correlation can be made between ionic radii of cations and unit cell dimensions provided the effect of covalent forces in the lattice is taken into account. The values for ionic radii of cations as given byAhrens (1952) permit a better correlation than those ofGoldschmidt.A shrinkage of 0.01 Å in the unit cell size per 0.01 Å decrease in the ionic radius of the divalent cations was determined when spinels with the same cation arrangement in the same group were compared. A shrinkage of 0.027 Å in the unit cell size per 0.01 Å decrease in the ionic radius of the trivalent cations was determined in spinels having the same divalent cation and cation arrangement when the trivalent cations form the same type of bonds.The half-occupation of the 3d orbits in Mn²⁺ and Fe³⁺ causes abnormally high unit cell dimensions in spinels where these ions are incorporated in octahedral sites. This is attributed to the formation of electrovalent bonds by these ions. Variable forces of contraction in the lattice are revealed when the unit cell dimensions are correlated with the ionic radii of cations. The force of contraction can be satisfactorily explained as being due to covalent forces in the spinel structure. The magnitude of this force or the degree of covalence in the bonds increases in the following order of cations where these are situated in tetrahedral sites:The divalent transition element ions, Fe²⁺, Co²⁺ and Ni²⁺; the B-Sub-group element ions Cd²⁺ and Zn²⁺; Fe³⁺ in tetrahedral co-ordination.     

Formation of sapphirine during retrogression of a basic high-pressure granulite,Roan, Western Gneiss Region,Norway

Sapphirine-bearing rocks occur in the northern part of the Western Gneiss Region, Vestranden, central Norway. The sapphirine-bearing rocks are characterized by a high MgO/(MgO + FeO) ratio, high Al₂O₃, MgO and CaO, and low SiO₂ contents. These rocks form layers within larger complexes which originated as layered magmatic rocks. High PT-metamorphism produced a cpx+ky+gt assemblage. The P and T estimates are P = 14.5±2 kbar and T= 870±50° C. During retrogression, the high-P granulite assemblage broke down to form an intermediate-P granulite mineralogy comprising orthopyroxene, spinel, anorthite, andesine, sapphirine and corundum. Textural relationships suggest that sapphirine formed by the reaction: spinel+kyanite sapphirine+corundum, and probably also by a reaction between corundum, spinel and orthopyroxene. All reactions took place within the stability field of kyanite. Textural and micro-chemical relationships indicate equilibrium, conditions during the peak metamorphism, whereas pronounced disequilibrium characterizes the mineral associations formed during the early retrogression at low P _H2O. The investigation shows that parts of the northern segment of the Western Gneiss Region underwent a metamorphic evolution similar to the Caledonian one recorded from eclogite/granulite terrains further south.     

A mapping of the electron localization function for the silica polymorphs: evidence for domains of electron pairs and sites of potential electrophilic attack

The electron localization function, η, evaluated for first-principles geometry optimized model structures generated for quartz and coesite, reveals that the oxide anions are coordinated by two hemispherically shaped η-isosurfaces located along each of the SiO bond vectors comprising the SiOSi angles. With one exception, they are also coordinated by larger banana-shaped isosurfaces oriented perpendicular to the plane centered in the vicinity of the apex of each angle. The hemispherical isosurfaces, ascribed to domains of localized bond-pair electrons, are centered ～0.70 Å along the bond vectors from the oxide anions and the banana-shaped isosurfaces, ascribed to domains of localized nonbonding lone-pair electrons, are centered ～0.60 Å from the apex of the angle. The oxide anion comprising the straight SiOSi angle in coesite is the one exception in that the banana-shaped isosurface is missing; however, it is coordinated by two hemispherically shaped isosurfaces that lie along the bond vectors. In the case of a first-principles model structure generated for stishovite, the oxide anion is coordinated by five hemispherically shaped η-isosurfaces, one located along each of the three SiO bond vectors (ascribed to domains of bonding-electron pairs) that are linked to the anion with the remaining two (ascribed to domains of nonbonding-electron pairs) located on opposite sides of the plane defined by three vectors, each isosurface at a distance of ～0.5 Å from the anion. The distribution of the five isosurfaces is in a one-to-one correspondence with the distribution of the maxima displayed by experimental Δρ and theoretical ??²ρ maps. Isosurface η maps calculated for quartz and the (HO) ₃ SiOSi(OH) ₃ molecule also exhibit maxima that correspond with the (3,?3) maxima displayed by distributions of ??²ρ. Deformation maps observed for the SiOSi bridges for the silica polymorphs and a number of silicates are similar to that calculated for the molecule but, for the majority, the maxima ascribed to lone-pair features are absent. The domains of localized nonbonding-electron pair coordinating the oxide anions of quartz and coesite provide a basis for explaining the flexibility and the wide range of the SiOSi angles exhibited by the silica polymorphs with four-coordinate Si. They also provide a basis for explaining why the SiO bond length in coesite decreases with increasing angle. As found in studies of the interactions of solute molecules with a solvent, a mapping of η-isosurfaces for geometry-optimized silicates is expected to become a powerful tool for deducing potential sites of electrophilic attack and reactivity for Earth materials. The positions of the features ascribed to the lone pairs in coesite correspond with the positions of the H atoms recently reported for an H-doped coesite crystal.     

ELF isosurface maps for the Al2SiO5 polymorphs

This study examines the electron localization function (ELF) isosurfaces of the Al₂SiO₅ polymorphs kyanite, sillimanite, and andalusite to see how differences in coordination and geometry of the cations and anions affect the ELF isosurfaces. Examination of the ELF isosurfaces indicates that their shapes are dependent on the coordination and geometry of the oxygen atoms and are not sensitive to coordination of the surrounding cations. Of the 18 crystallographically distinct oxygen atoms in the Al₂SiO₅ polymorphs, 13 are bonded to two aluminum atoms and one silicon atom (Al₂–O–Si) and are associated with two different ELF isosurface shapes. The shape of the ELF isosurface is dependent on the distance at which the oxygen atom lies from a plane defined by the three surrounding cations: at a distance greater than 0.2 Å the ELF can be defined as horseshoe-shaped and at a distance less then 0.2 Å it can be described as concave hemispherical. This feature is also seen in the ELF isosurfaces for the oxygens bonded to three aluminum atoms (Al₃–O) where the isosurfaces can be defined as trigonally toroidal and uniaxially trigonally toroidal. The changes in the ELF isosurfaces for the three coordinated oxygens are also indicative of changes in hybridization. The ELF isosurface for the two-fold coordinated oxygen (Al–O–Si) has a large mushroom-shaped isosurface along the Al–O bond and a concave hemispherical isosurface along the Si–O. The four-fold coordinated oxygen (Al₄–O) contains two concave hemispherical isosurfaces along the shorter Al–O bonds and a banana-shaped isosurface, which encompasses the longer Al–O bonds. In addition, this study shows the homeomorphic relationship between the ELF isosurfaces and electron density difference maps with respect to number and arrangement of domains.     

Phase transformations in descending plates and implications for mantle dynamics

The role of phase transformations in a mantle of pyrolite composition is reviewed in the light of recent experimental data. The pyroxene component of pyrolite transforms to the garnet structure at 300–350 km whilst olivine transforms to beta-Mg₂SiO₄ near 400 km. Between about 500 and 550 km, beta-Mg₂SiO₄ probably transforms to a partially inverse spinel structure whilst the CaSiO₃ component of the complex garnet solid solution exsolves and transforms to the perovskite structure. The major seismic discontinuity near 650–700 km is probably caused by disproportionation of Mg₂SiO₄ spinel into periclase plus stishovite. At a slightly greater depth, the remaining magnesian garnet transforms to the corundum or ilmenite structure. Finally, at a depth probably in the vicinity of 800–1000 km, the (Mg,Fe)SiO₃ component of the ilmenite phase transforms to a perovskite structure whilst stishovite and some of the periclase recombine to form perovskite also. The mineral assemblage so formed is about 4% denser than mixed oxides (MgO + FeO + A1₂O₃ + CaO + stishovite) isochemical with pyrolite. The above sequence of phase transformations in pyrolite provides a satisfactory general explanation of the elastic properties and density distribution in the mantle. In particular, there is no evidence requiring an increase of FeO/(FeO + MgO) ratio with depth.The depths at which major phase transformations occur in subducted lithosphere differ from those in ‘normal’ mantle. These differences are caused by two factors: (1) Temperatures within sinking plates are much lower than in surrounding mantle to depths of 700 km or more. (2) Irreversible chemical differentiation of pyrolite occurs at oceanic ridges. Lithosphere plates so formed consist of a layer of basaltic rocks underlain successively by layers of harzburgite, lherzolite, and pyrolite slightly depleted in highly incompatible elements (e.g. La, Ba, Rb, U). The phase-transformation behaviour of the first three of these layers differs from that of pyrolite.The effects of these and other factors connected with phase transformations on the dynamics of plate subsidence are discussed. It appears quite likely that plates penetrate the 650–700 km discontinuity, largely because the slope of the spinel disproportionation is probably positive, not negative as generally supposed. The former basaltic oceanic crust probably sinks deeply into the lower mantle, whilst the former harzburgite component of the plate may collect above the perovskite transition boundary. Phase transformations may thus serve as a kind of filter, leading to increased and irreversible mantle heterogeneity with time.The possible roles of phase transformations in causing deep-focus earthquakes and introducing water into the mantle in subduction zones are also briefly discussed.     

地幔橄榄岩捕掳体中尖晶石筛状边的成因

山东的无棣大山和栖霞方山是两个以霞石岩为主的新生代火山,含有大量的地幔橄榄岩捕掳体,橄榄岩捕掳体中的尖晶石常见黑色反应边。在BSE图像上,这些反应边根据结构的不同又可以进一步分为两种:筛状边和均一边。有时两种反应边在同一颗尖晶石中共存并相互过渡。尖晶石捕掳晶和捕掳体中与寄主岩直接接触的尖晶石常发育均一边;发育筛状边的尖晶石则一般不直接与寄主岩接触,但位于捕掳体的边缘。尖晶石的筛状边呈多孔状,为富Cr尖晶石或铬铁矿。筛状边成分上与核部(Cr_2O_3=7.52%~36.75%,Cr#=7.80~44.20,Mg#=44.70~74.48)区别明显,具有高的Cr_2O_3含量(23.42%~65.96%)、Cr#值(78.97~92.49)以及低的Mg#值(17.22~43.02)。另外,筛状边相对其核部(TiO 2=0.00%~0.53%,MnO=0.04%~0.35%)还显示偏高的TiO 2(0.20%~3.60%)和MnO(0.29%~1.93%)。在筛状边附近存在富Al_2O_3(8.00%~17.57%)和MgO(17.89%~26.02%)的玻璃。尖晶石的均一边无孔洞但多发育裂理,成分上与核部突变,并以富TiO 2(20.90%~6.64%)和FeO T(70.79%~53.92%)为特征,最外部为钛磁铁矿。均一边具有明显的成分分带,表现为由内至外TiO 2、FeO T含量逐渐增高,Al_2O_3(0.04%~16.34%)、Cr_2O_3(0.77%~25.69%)和MgO(0.50%~7.16%)含量逐渐降低。尖晶石的两种反应边与寄主岩密切的空间关系说明其成因与寄主岩浆有关。虽然尖晶石筛状边的Cr#值(79.0~92.5)明显高于核部(7.8~44.2),但是其相对核部偏高的TiO 2和MnO含量,以及显著降低的Mg#值均与单纯的部分熔融趋势不吻合。根据尖晶石两种反应边的结构特征、成分特征和相互的空间关系,我们认为筛状边与均一边是岩浆与尖晶石反应不同阶段的产物。当橄榄岩捕掳体被岩浆捕获后,由于成分上的不均衡两者之间将发生溶解反应。反应过程中,尖晶石中不同元素向熔体迁移的速率区别明显,其中Cr是最难熔的元素也是迁移最慢的元素。在反应的初始阶段,由于易溶组份Al、Mg等元素随熔体迁出,尖晶石发育筛状边,并同时伴随Cr#值的显著升高和Mg#值的显著降低。此时,寄主岩浆中的Ti、Mn等元素也部分扩散进入筛状边中的残留尖晶石。如果熔体足够多,反应得以持续进行,筛状边中的Cr最终也将被熔体带走。在反应产生的混合熔体中金属氧化物最早达到饱和并结晶。随着反应的持续进行,结晶出的矿物将继续生长、变粗,并相互连接,直至孔洞消失形成均一边。同时伴随着矿物的不断结晶,溶解界面上的熔体成分也不断变化,从而使均一边具有明显的成分分带。在橄榄岩捕掳体内部,由于渗进捕掳体中的岩浆有限,渗透岩浆与尖晶石之间的反应多限于早期阶段,反应产物为筛状边。在捕掳体边缘,与寄主岩浆直接接触的尖晶石由于参与反应的岩浆量足够多,因此一般产生反应的终端产物———均一边。     

Subsolidus Phase Relations in the System MgO-SiO_2-Cr-O and Thermodynamic Properties of Related Cr~(2+)-Containing End-members

Subsolidus phase relations have been determined in the systems SiO2-Cr-O and MgO-SiO2-Cr-O in equilibrium with metallic Cr, at 1100 to 1500℃ and 0 to 2.88 GPa. The results show that there are no ternary phases in the SiO2-Cr-O system at these conditions, i.e., only the assemblage eskolaite-Cr-metal-quartz (or tridymite) is found. In the MgO-containing system, however, extensive substitution of Cr2+ for Mg is observed in (Mg, Cr2+)2SiO4 olivine, (Mg, Cr2+)2Si2O6 pyroxene, and (Mg, Cr2+)Cr2O4 spinel. Cr3+ levels in olivine and pyroxene are below detection limits. The pyroxene is orthohombic at XCrPx2+ < 0.2, monoclinic at higher XCrPx2+ . Thestructure of the spinels becomes tetragonally distorted at XCr2+Sp >0.2. The experimental datahave been fitted to a thermodynamic model, and the authors obtained the mixing parameter (W) of Mg-Cr2+ in olivine, pyroxene and spinel, and the relation between temperatures and free energies of formation for the end-members: Cr2+-olivine (Cr2SiO4), Cr2+-pyroxene (Cr2Si2O6)     

Vibrational spectroscopy of aluminate spinels at 1 atm and of MgAl2O4 to over 200 kbar

Single-crystal Raman and infrared reflectivity data including high pressure results to over 200 kbar on a natural, probably fully ordered MgAl₂O₄ spinel reveal that many of the reported frequencies from spectra of synthetic spinels are affected by disorder at the cation sites. The spectra are interpreted in terms of factor group analysis and show that the high energy modes are due to the octahedral internal modes, in contrast to the behavior of silicate spinels, but in agreement with previous data based on isotopic and chemical cation substitutions and with new Raman data on gahnite (～ ZnAl₂O₄) and new IR reflectivity data on both gahnite and hercynite (～Fe_0.58Mg_0.42Al₂O₄). Therefore, aluminate spinels are inappropriate as elastic or thermodynamic analogs for silicate spinels. Fluorescence sideband spectra yield complementary information on the vibrational modes and provide valuable information on the acoustic modes at high pressure. The transverse acoustic modes are nearly pressure independent, which is similar to the behavior of the shear modes previously measured by ultrasonic techniques. The pressure derivative of all acoustic modes become negative above 110 kbar, indicating a lattice instability, in agreement with previous predictions. This lattice instability lies at approximately the same pressure as the disproportionation of spinel to MgO and Al₂O₃ reported in high temperature, high pressure work.     

Spectroscopic characterisation and crystal field calculations of varicoloured kyanites from Loliondo, Tanzania

Orange, ochre-coloured, light green and dark blue varieties of kyanite, ideally Al₂SiO₅, from Loliondo, Tanzania, have been characterised by electron microprobe analysis and polarised infrared and optical absorption spectroscopy. All colour varieties show elevated Fe contents of 0.39 to 1.31 wt.% FeO, but Ti contents only in the range of the EMP detection limit. Orange and ochre-coloured crystals have Mn contents of 0.23 and 0.06 wt.% MnO, respectively, the dark blue kyanite contains 0.28 wt.% Cr₂O₃, while the light green sample is nearly free from transition metal cations other than Fe. Polarised infrared spectra reveal OH defect concentrations of 3 to 17 wt.ppm H₂O with structural OH defects partially replacing the O_B (O2) oxygen atoms. Polarised optical absorption spectra show that the colour of all four varieties is governed by crystal field d-d transitions of trivalent cations, i.e. Fe³⁺ (all samples), Mn³⁺ (orange and ochre) and Cr³⁺ (blue kyanite), replacing Al in sixfold coordinated triclinic sites of the kyanite structure. Intervalence charge transfer, the prevalent colour-inducing mechanism in ‘usual’ (Cr-poor) blue kyanites, seems to play a very minor, if any, role in the present samples. Crystal field calculations in both a ‘classic’ tetragonal and in the semiempirical Superposition Model approach, accompanied by distance- and angle-least-squares refinements, indicate that Fe³⁺ preferably occupies the Al4 site, Cr³⁺ prefers the Al1 and Al2 sites, and Mn³⁺ predominantly enters the Al1 site. In each case specific local relaxation effects were observed according to the crystal chemical preferences of these transition metal cations. Furthermore, the high values obtained in the calculations for the interelectronic repulsion parameter Racah B correspond to a high ionic contribution to Me³⁺–O bonding in the kyanite structure. In the particular case of the blue sample, band positions specifically related to the high Racah B value enable this ‘unusual’ type of blue colouration of kyanite solely due to Cr³⁺ cations.     

Spinel minerals in transitional and alkali basaltic glasses from Iceland

Crystallization of spinel minerals in transitional and alkali basalts from Iceland can be related to the FeO, MgO, TiO₂ and Cr contents of the coexisting melt. Chromian spinel occurs in glasses in which TiO₂ is less than 2.8 wt.% and the weight ratio FeO/MgO is less than 2.0, whereas titanomagnetite occurs when the same parameters are greater than 4 wt.% and 2.7, respectively. In addition, chromian spinel only occurs in basalts with Cr greater than 200 ppm. It is suggested that chromian spinel crystallizes, together with olivine, from liquids with olivine liquidus temperatures ranging from above 1,200° C to approximately 1,150° C. A discontinuity in spinel crystallization follows until below 1,100° C, where titanomagnetite starts to crystallize. Compositional variations in chromian spinel attached to, or included, in homogeneous olivine phenocrysts, however, cannot be related to equilibrium relations. Textural relations suggest homogeneous nucleation for titanomagnetite, whereas chromian spinel nucleates heterogeneously, dependent on growth of olivine phenocrysts. The composition of chromian spinels cannot in detail be related to physical and compositional parameters of the average melt, but may be related to local compositional relations in the melt adjacent to growing crystals. Such compositional variation around growing olivine crystals may be the prime reason for the non-equilibrium precipitation of included chromian spinels.     

Crystal chemistry of natural and synthetic lead oxohalides: II. Crystal structure of Pb7O4(OH)4Br2

Crystals of lead oxobromide Pb₇O₄(OH)₄Br₂ have been synthesized by hydrothermal method. The structure of the new compound has been studied with X-ray single-crystal diffraction analysis. The compound is monoclinic, space group C112₁; unit-cell dimensions are a = 5.852(4), b = 13.452(7), c = 19.673(9) Å, γ = 90.04°, V = 1548.7(15) ų. The structure has been solved by direct methods and refined to R ₁ = 0.1138 for 1847 observed Pb₇O₄(OH)₄Br₂ unique reflections. The structure contains seven symmetrically independent bivalent Pb atoms. The coordination polyhedrons of Pb are strongly distorted due to stereochemical activity of unshared electron pair 6s ². Oxygen atoms are tetrahedrally coordinated by four Pb²⁺ cations with the formation of oxocentered tetrahedrons OPb₄. The compound is based on [O₂Pb₃]²⁺ double chains formed by OPb₄ tetrahedrons. (OH)Pb₂ dimers combine the [O₂Pb₃]²⁺ chains into 3D framework. Channels in the framework are parallel to [100] and are occupied by Br anions.     

Garnet and spinel lherzolite assemblages in MgO–Al2O3–SiO2 and CaO–MgO–Al2O3–SiO2: thermodynamic models and an experimental conflict

The recent publication of an updated thermodynamic dataset for petrological calculations provides an opportunity to illustrate the relationship between experimental data and the dataset, in the context of a new set of activity–composition models for several key minerals. These models represent orthopyroxene, clinopyroxene and garnet in the system CaO–MgO–Al₂O₃–SiO₂ (CMAS), and are valid up to 50 kbar and at least 1800 °C; they are the first high‐temperature models for these phases to be developed for the Holland & Powell dataset. The models are calibrated with reference to phase‐relation data in the subsystems CaO–MgO–SiO₂ (CMS) and MgO–Al₂O₃–SiO₂ (MAS), and will themselves form the basis of models in larger systems, suitable for calculating phase equilibria in the crust and mantle. In the course of calibrating the models, it was necessary to consider the reaction orthopyroxene + clinopyroxene + spinel = garnet + forsterite in CMAS, representing a univariant transition between simple spinel and garnet lherzolite assemblages. The high‐temperature segment of this reaction has been much disputed. We offer a powerful thermodynamic argument relating this reaction to the equivalent reaction in MAS, that forces us to choose between good model fits to the data in MAS or to the more recent data in CMAS. We favour the fit to the MAS data, preserving conformity with a large body of experimental and thermodynamic data that are incorporated as constraints on the activity–composition modelling via the internally consistent thermodynamic dataset.     

Structure and chemistry of (111) twin boundaries in MgAl<Subscript>2</Subscript>O<Subscript>4</Subscript> spinel crystals from Mogok

The atomic scale structure and chemistry of (111) twins in MgAl₂O₄ spinel crystals from the Pinpyit locality near Mogok (Myanmar, formerly Burma) were analysed using complementary methods of transmission electron microscopy (TEM). To obtain a three-dimensional information on the atomic structure, the twin boundaries were investigated in crystallographic projections and Using conventional electron diffraction and high-resolution TEM (HRTEM) analysis we have shown that (111) twins in spinel can be crystallographically described by 180° rotation of the oxygen sublattice normal to the twin composition plane. This operation generates a local hcp stacking in otherwise ccp lattice and maintains a regular sequence of kagome and mixed layers. In addition to rotation, no other translations are present in (111) twins in these spinel crystals. Chemical analysis of the twin boundary was performed by energy-dispersive X-ray spectroscopy (EDS) using a variable beam diameter (VBD) technique, which is perfectly suited for analysing chemical composition of twin boundaries on a sub-nm scale. The VBD/EDS measurements indicated that (111) twin boundary in spinel is Mg-deficient. Quantitative analyses of HRTEM (phase contrast) and HAADF-STEM (Z-contrast) images of (111) twin boundary have confirmed that Mg²⁺ ions are replaced with Be²⁺ ions in boundary tetrahedral sites. The Be-rich twin boundary structure is closely related to BeAl₂O₄ (chrysoberyl) and BeMg₃Al₈O₁₆ (taaffeite) group of intermediate polysomatic minerals. Based on these results, we conclude that the formation of (111) twins in spinel is a preparatory stage of polytype/polysome formation (taaffeite) and is a result of thermodynamically favourable formation of hcp stacking due to Be incorporation on the {111} planes of the spinel structure in the nucleation stage of crystal growth. The twin structure grows as long as the surrounding geochemical conditions allow its formation. The incorporation of Be induces a 2D-anisotropy and exaggerated growth of the crystal along the (111) twin boundary.