Recent experimental and theoretical studies provide new information on the variety of reconstructive, displacive, electronic, and magnetic transformations induced by pressure in minerals that comprise the deep mantle and core. Laser-heating diamond-cell experiments indicate that the partitioning of Fe and Mg between coexisting (Mg,Fe)SiO3-perovskite and (Mg,Fe)O-magnesiowüstite strongly depends on pressure, temperature, bulk Fe/Mg ratio, and ferric content, and the perovskite stability field expands relative to magnesiowüstite + SiO2 on increasing pressure and temperature. The soft-mode transition in SiO2 to the CaCl2-type structure, first predicted by accurate theoretical methods, has been documented by both Raman scattering and single-crystal diffraction, and the transition to higher pressure forms has been examined. The phase diagram of FeO, including the rhombohedral and higher pressure NiAs transitions, has been determined, and the former has been studied in magnesiowüstites. Moreover, high-spin/low-spin transitions in FeO, as well as Fe2O3 and FeS, have been investigated by high-resolution x-ray emission spectroscopy to 150 GPa. For the core, laser-heating studies to above 150 GPa and 2500 K show that (hcp) <epsilon>-Fe exhibits a large P-T field of stability field. Radial x-ray diffraction measurements carried out at room temperature to above 200 GPa have been used to constrain the elasticity, rheology, and sound velocities of <epsilon>-Fe at core pressures. Comparison with seismic data indicates large temperature effects on the shear velocities, additional low velocity phases, and/or complicated textures for the inner core.
Diamonds are found in strongly shocked gneiss clasts within impact melt rocks (tagamites) and suevites of the Popigaï structure (Northern Siberia), in the zone immediately outside the central part (Koeberl et al., 1997).The translucent to black diamonds formed by shock transformation of the tabular graphite crystals of the preimpact target rocks: they display a layered texture, with twinning and striations, interpreted as inherited from this prograde phase transition, in agreement with trace elements and isotopic compositions. XRD showed the polycrystallinity of diamond aggregates (< 1µm) and the intergrowth with the hexagonal high-pressure polymorph lonsdaleite. In the black diamond grains TEM investigations allowed to image graphite flakes, which have no clear orientation relationship to the diamond and thus might have formed by retrograde phase transition. We performed Raman microspectroscopy studies on such sub-millimetric raw grains, using a conventional XY Dilor microspectrometer (Argon laser beam, 514,5 and 488 nm, X50 microscope objective lens, triple monochromator, CCD detector). The weak Raman signals were continually superimposed on an intense luminescence background, which nearly canceled them. The typical diamond vibration band (1332 cm-1) was recorded on some sample parts, with an important line broadening (up to 18 cm-1) as expected for nanocrystalline domains and no frequency shift. This band was frequently associated with a weak and broad signal in the 1560 -1600 cm-1 range which, in some cases, could be recorded alone, as observed generally for poorly organized carbon samples with graphite-like nanodomains. Both the microscopic structural observations and the spectroscopic data are consistent with an intergrowth of diamond-type crystalline zones and graphite-like regions, involving complex interfaces, in terms of sp2 versus sp3 bonding. In accordance with TEM, the texture could be interpreted in terms of retrograde phase transition from diamond to graphite during a slow cooling down of the target material after high pressure release.
Koeberl C, Masaitis VL, Shafranovsky GI, Gilmour I, Langenhorst F & Schrauder M, Geology, 25, 967-970, (1997).
Recent high-pressure investigations on silica lead to the recognition of phenomena like displacive phase transitions, amorphization transitions, and features induced by dynamic wave propagation. Molecular dynamics calculations also indicated the existence of several competitive very dense post-stishovite polymorphs and delineated their pressure stability ranges and their displacive transition behavior with increasing pressure (Kingma et al., 1995; Belonoshko et al., 1996; Dubrovinsky et al., 1997; Teter et al., 1998). On the other hand, dynamic experiments produced several dense silica polymorphs neither encountered in static experiments conducted at comparable maximum pressures (German et al., 1973; Sekine et al., 1987) nor predicted from molecular dynamics calculations. The origin of these polymorphs whether formed metastably at the estimated dynamic peak pressures or as quench phases formed during adiabatic decompression remained obscure. This issue is of particular importance, specifically if each of these phases obtained in dynamic experiments is a quench product of specific post-stishovite polymorphs. For example, is natural stishovite in shocked rocks a primary phase formed by phase transformation of quartz at the peak shock pressure required, or was it formed by back inversion of the nonquenchable CaCl2-structured silica during decompression?
Until very recently, coesite and stishovite were the only naturally reported dense polymorphs of silica from meteorite craters (Chao et al., 1962). We have been investigating naturally densified silica grains in the heavily shocked SNC meteorite Shergotty that is thought to be of Martian origin. The silica grains display an orthogonal intergrowth of several sets of lamellae of very dense silica glass and a new high-pressure orthorhombic polymorph of silica with a structure consistent with a Pca2 space group (Sharp et al., 1998). New investigations indicated the presence of an additional dense polymorph whose x-ray diffraction pattern could be indexed in terms of a monoclinic lattice that is related to the baddeleyite-type-structure with the cell parameters a = 4.375 Å, b = 4.584 Å, c = 4.708 Å, ß = 99.97, <rho> (calc.) = 4.30 g/cm3. The observed x-ray reflections could be explained by the baddeleyite-type structure. Both x-ray diffraction and HRTEM of the same grains indicate a multiphase intergrowth of the Pca2 polymorph, baddeleyite-structured polymorph, stishovite, and dense glass. The intergrowth is suggestive of an assemblage produced during decompression from a high shock state. The multicomponent silica grains display numerous radiating cracks initiating at their surfaces and penetrating deep into the Shergotty matrix (= 600 µm). This is indicative of volume increase due to expansion triggered by phase transitions during decompression. This interpretation is also supported by recent simulation calculations which suggest that an <alpha>-PbO2-structured silica (with space group Pnc2) that is closely related to the baddeleyite structure, could invert upon decompression to the baddeleyite type (P21/c) with seven-coordinated-silicon. It is conceivable that such an assemblage is a good candidate to explore for the presence of decompression quench phases and ultimately trace their very high-pressure parental phases.
Kingma K, Cohen R E, Hemley R J, Mao H-K, Nature, 374, 243-245, (1995).
Belonoshko AB, Dubrovinsky LS, Dubrovinsky NA, American Mineralogist, 81, 785-788, (1996).
Dubrovinsky LS, Saxena S, Lazor P, Ahuja R, Erikson O, Wills JM, Johansson B, Nature, 388, 362-365, (1997).
Teter DM, Hemley MA, Kresse G, Hafner J, Physycal Review Letters, 80, 2145-2148, (1998).
German VN, Podurets MA, Trunin RF, Soviet Physics - JETP, 37, 107, (1973).
Sekine T, Akaishi M, Setka N, Geochemica Cosmochemica Acta, 51, 379-381, (1987).
Using a combination of the Rigid Unit Mode (RUM) theory, molecular dynamics simulations, and neutron diffuse scattering, we have been able to describe the sequence of phase transitions in tridymite. The high-temperature phase transitions involve a simple sequence of displacive phase transitions in which the phases have simple group-subgroup symmetry relations, and the distortions accompanying each phase transition correspond to RUM distortions of each parent phase. The lower temperature phases break the sequence, and actually involve RUM distortions of the high-temperature hexagonal phase. The computer simulations and neutron diffuse scattering data show that the higher temperature phases have dynamic disorder of the orientations of the SiO4 tetrahedra in order to avoid formation of linear Si-O-Si bonds, as in cristobalite, and there are no domains of lower-symmetry phases. Our work also explains why there are so many phase transitions in tridymite (Pryde and Dove, 1998).
Pryde ALA & Dove MT, Physics and Chemistry of Minerals, 26, 171-179, (1998).
The wavenumber <nu> of the most intense A1 Raman mode of quartz displays a relatively large pressure dependence (<nu>/P)T coupled with a relatively moderate dependence on temperature (<nu>/T)P. Therefore, it can be potentially applied as pressure sensor in hydrothermal diamond-anvil cell studies of SiO2-saturated systems. In this study, we obtained information about the pressure dependence of this Raman line at high temperatures for which no data has previously been available. Furthermore, we redetermined wavenumbers and linewidths of the main Raman active modes at 1 bar pressure in the temperature range 24°C to 800°C.
A doubly polished quartz disk and, for P>1 bar, distilled water, which served as pressure medium, were placed in a hydrothermal diamond-anvil cell. A DILOR XY Raman microprobe was used to record spectra in the wavenumber range between 100 and 600 cm-1. Wavenumber offsets were corrected by measuring the wavenumbers of three plasmalines immediately after obtaining the quartz spectrum at the same conditions and comparing the results to the known wavenumbers of these lines (116.04 cm-1, 266.29 cm-1, and 520.30 cm-1). The pressures in the sample chamber at the experimental temperatures were determinined by measuring the liquid-vapor homogenization temperature of the pressure medium and calculating the pressure on the corresponding isochore at the experimental temperature. A Linkam TH 1500 heating stage was used instead of a diamond-anvil cell to check the reproducibility of the data at 1 bar pressure. These results were in close agreement with the data obtained using the diamond-anvil cell and with the available literature data.
At 1 bar, the wavenumber of the most intense Raman mode of quartz shows a slightly nonlinear decrease from n=464.1 cm-1 at 24°C to 457 cm-1 at 540°C, is then approximately constant between 540°C and 560°C, and then increases rapidly by about 1 cm-1 as the temperature approaches the <alpha>-ß quartz transition. A discontinuity occurs at that temperature. In the ß-quartz stability field, the observed wavenumber increase was slight and apparently linear to <nu>=459.5 cm-1 at 800°C. The width at the half maximum of this line at 1 bar broadens with temperature from about 7.5 cm-1 at 24°C to about 43 cm-1 at 800°C with a discontinuity at the transition temperature. The obtained pressure shifts for this line range from 0.86±0.06 cm-1 kbar-1 at 200°C to 0.92±0.05 cm-1kbar-1 at 500°C and are significantly higher than the value of 0.8 cm-1kbar-1 at 4K and 300 K reported in the literature. All other Raman modes of quartz are less suitable for application as a pressure sensor.
Silicate perovskite and (Mg,Fe)O are believed to constitute most of the lower mantle. Previous research indicates that iron and aluminium can be incorporated in the perovskite structure, where Fe3+/(sum)Fe depends primarily on the amount of aluminium. In this study we investigate the nature of the substitution mechanism of Fe3+ and Al3+ in the perovskite structure, particularly with regard to the occurrence of defects.
(Mg,Fe,Al)(Si,Al)O3 perovskites with varying amounts of Fe and Al were synthesised from orthopyroxenes in a multi-anvil press at conditions relevant to the lower mantle (26 GPa, ~1700°C, runtime 2-8 h) with different oxygen fugacities produced by using Re and Fe capsules. Resulting run products were characterised by X-ray diffraction and chemical composition was analysed with an electron microprobe. The Fe3+/(sum)Fe ratio was determined by 57Fe-Mössbauer spectroscopy and electron energy loss spectroscopy.
For the incorporation of Fe3+ and Al3+ we propose two possible mechanisms. The first mechanism, Si4+ + Mg2+ <->2(Fe3+, Al3+), involves substitution of trivalent cations on both sites in the perovskite structure and requires no oxygen defects, while the second mechanism, 2Si4+ + O2- <-> 2(Fe3+, Al3+) + V[O]··, involves the creation of oxygen defects.
To distinguish between these two mechanisms, we used the microprobe data and the Fe3+/(sum)Fe ratio measured by Mössbauer spectroscopy to determine the amounts of the three components SiO2, (Mg,Fe)O and (Al2O3 + Fe2O3) in the perovskites. In a ternary diagram with the vectors for each substitution mechanism, we found evidence for both mechanism to operate simultaneously. Further evidence for the second substitution mechanism was seen in some Mössbauer spectra which contain a magnetic sextet corresponding to Fe3+ in a tetrahedral environment.
The presence of magnetically ordered IVFe3+ implies the presence of oxygen defects with some degree of order, similar to structural elements in the brownmillerite structure. These oxygen defects and the nature of their ordering are expected to have a significant influence on physical properties such as compressibility and electrical conductivity, and hence significant implications for our interpretation of geophysical data.
The substitution of Ti4+ by Fe3+ in CaTiO3 perovskite leads to the formation of nonstoichiometric compounds with the general formula CaTi1-xFexO3-x/2 (with 0 ¾ x ¾ 1), i.e introducing two Fe3+ atoms in Ti4+ sites results in the creation of one oxygen vacancy, in Kröger-Vink notation: 2TiTi + OO <-> 2FeTi´ + VO¨. We have studied the system CaTiO3-CaFeO2.5 with 0 ¾ x ¾ 0.5 at room pressure and high temperature in order to analyse the phase equilibria as a function of temperature and composition. Quenched samples were characterised by X-ray diffraction, Mössbauer spectroscopy and transmision electron microscopy.
The results show the existence of two different phase transitions. At low Fe contents the high temperature cubic phase (with disordered oxygen vacancies) undergoes displacive phase transitions from cubic to tetragonal to orthorhombic symmetry upon cooling. A strain analysis of samples annealed at high temperatures and with variable Fe content have shown that the phase transition is possibly first order. The critical temperature for this transition is strongly affected by the number of vacancies (dT/dx > 50°C/mol% of vacancies).
The second set of phase transitions involves ordering of the oxygen vacancies. Close to the CaTiO3 composition the perovskite structure is preserved and the oxygen vacancies seem to be randomly distributed, with Fe3+ in IV, V and VI coordinated sites as shown by Mössbauer spectroscopy. With increasing Fe content Mössbauer spectra indicate that short range ordering develops, generating more tetrahedrally- and octahedrally-coordinated Fe sites following the reaction 2FeV <-> FeIV + FeVI. In some samples a tweed microstructure has been observed in high-resolution TEM images. Finally, in samples of higher Fe content, long range ordering of tetrahedral and octahedral layers, perpendicular to the b direction, is indicated by TEM and XRD observations; Mössbauer spectra of these samples show only FeIV and FeVI sites. These results also reveal that the distortion of the VI site increases with increasing Fe composition while distortion of the IV site increases with increasing annealing temperature.
The results of a high temperature single crystal X-ray diffraction study of synthetic titanite within the stability field of the A2/a (C2/c) paraphase (T>500K) are reported. Intensity data were collected up to T=1000K using MoK<alpha> radiation at diffraction angles up to 60°2 <theta>. A break in the volume expansion of titanite near T=825K has been found. This discontinuity correlates with an effective volume contraction of the Ti-octahedron, as obtained from structure refinement. The volume anomaly is accompanied by a clear decrease in octahedral angle variance, small shifts in the O3A oxygen position and a slight change in the thermal behaviour of the Ti-O3A-Si bond angle near 825K. A discontinuity in the elongation of the thermal displacement ellipsoid of the Ti atom parallel to the a-direction also coincides with the volume anomaly. The temperature of the structural instability is in agreement with previous findings based on Hard Mode IR Spectroscopy (Zhang et al., 1997). Possible mechanisms causing the observed isosymmetric structural changes, their relation to the P21/a - A2/a order-disorder phase transition near 500K in titanite and similarities to the thermal behaviour of the isostructural malayaite, CaSnSiO5, are discussed.
Zhang M, Salje E & Bismayer U, American Mineralogist, 82, 30-35, (1997).
Lawsonite is a mineral found in metamorphic rocks which experienced high pressures at relatively low temperatures. Its structure is made up of chains of edge sharing AlO6 octahedra along the a-axis. These chains are interconnected by Si2O7 groups forming a framework with channels running parallel to a. Bigger channels contain water-molecules and Ca-atoms and in smaller channels the OH-groups are found. The water content of ~11wt% together with its for a water-bearing mineral rather high density of ca 3.3 g/cm3 gave rise to the idea that it might be responsible for the transport of water into the upper earth's mantle.
Libowitzky and Armbruster (1995) reported that lawsonite undergoes two phase transitions below room temperature. The symmetry changes at 273 K from Cmcm to Pmcn and at 155 K further reduces to space group P21cn. These changes seem to be related to a reorientation of H2O molecules and to reordering of hydrogen bonds.
A new study of spontaneous strain and micrsocopic structural variations has been carried out. Xray powder collected in the temperature region from 40 to 773 K using a focusing Guinier camera with monochromatized CuK<alpha>1 radiation in the temperature region from 40 to 773 K gave lattice constants from which spontaneous strain data were calculated. To take care of low temperature saturation effects a function of the type y=y0 + y1*<theta>*coth(<theta>/T) was fitted to the high temperature part of the data and extrapolated to low temperatures as a baseline. The resulting strains in the order of ± 0.003 give evidence of a highly anisotropic behaviour below ca 300 K. They show several features which may be correlated to more than two transitions.
Powder Infrared experiments were performed to reveal the response of different parts of the structure on a local scale to changes in temperature. KBr and CsI pellets were used for the MIR region and polyethylene pellets for the FIR region. The resulting spectra were analysed using the autocorrelation method. The effective linewidth as a function of temperature correlates partly to the strain data.
Libowitzky E & Armbruster T, American Mineralogist, 80, 1277-1285, (1995).
When magnesian cordierite is grown by annealing glass for short times at high temperatures, the initial crystals have a high degree of Al/Si disorder and hexagonal symmetry. Further annealing of such samples induces a progressive increase in Al/Si order through a sequence of structural states from hexagonal to modulated and, finally, to the equilibrium orthorhombic structure. This structural sequence has previously been characterised by transmission electron microscopy, vibrational spectroscopy and NMR spectroscopy (Putnis, 1980; Putnis et al., 1987; Güttler et al., 1989; Daniels et al., 1994, and references therein). In order to quantify the enthalpy changes associated with the Al/Si ordering, high temperature solution calorimetric measurements have been made on samples annealed for different times at 1400, 1345, 1290, 1250 and 1185°C. These show a distinct break in slope as a function of ln(annealing time) at approximately the point where the hexagonal crystals become modulated. States of order of the full suite of samples have been characterised by spontaneous strain analysis (using lattice parameters from conventional X-ray powder diffraction) and hard mode IR spectroscopy (using powder absorption spectra). The hexagonal to modulated transition is continuous. It is marked by a change in slope with ln(annealing time) of selected phonon frequencies, but no equivalent influence is seen in the non-symmetry-breaking strain parameters (<epsilon>1+<epsilon>2) and <epsilon>3. This behaviour can be rationalised in terms of two order parameters, Qt and qod, where Qt describes the partitioning of Al and Si between T1 and T2 sites and qod describes local scale ordering on the basis of orthorhombic symmetry. Qt appears to vary linearly through the hexagonal to modulated sequence while qod shows a marked change in trend at approximately the transition point. The symmetry-breaking strain (<epsilon>1-<epsilon>2) appears to vary in a non-linear manner with qod in the orthorhombic structure, while the enthalpy changes appear to be linear functions of (qod)2. Analysis of the IR spectra provides information on the evolution of these order parameters on length scales intermediate between the length scales of NMR spectroscopy (~nearest neighbour configurations) and X-ray diffraction (many unit cells).
Putnis A, Contrib. Min. Pet, 74, 135-141, (1980).
Putnis A, Salje E, Redfern SAT, Fyfe A & Strob H, Phys. Chem. Min, 17, 446-454, (1987).
Guttler B, Salje EKH & Putnis A, Phys. Chem. Min, 16, 365-373, (1989).
Daniels P, Wunder B, Sahl K & Schreyer W, Eur. J. Min, 6, 323-335, (1994).
The development of facilities to investigate optical second harmonic generation (SHG) in minerals is particularly exciting because this technique has the potential to reveal polar transitions in minerals that are difficult to study by traditional techniques. We have built an instrument capable of measuring the intensity of second harmonic light generated in polar crystals as a function of temperature and spatial position. This has been applied to the study of centric-acentric phase transitions in a number of phases, including quartz, tridymite and barium titanate. By studying small-grained powders we can overcome difficulties of optic phase matching, so that the degree of polar order can be measured rapidly and reliably in such materials over a wide temperature range (from room temperature up to 1000oC in our instrument). The method also reveals the existence of short-lived polar clusters in the high-temperature para-phase in all transitions we have studied. By translating single crystal wedges in the beam we are able to measure Maker fringes as a function of temperature, and hence determine the form of the SHG moduli below acentric-centric transitions. Our latest results from this instrument will be presented, demonstrating its powerful application to the study of mineral behaviour.
Natural zircon, ZrSiO4, is known to undergo a radiation-induced transformation from the original crystalline state to a metamict or highly amorphous state (Weber et al., 1994) as a consequence of the <alpha>-decay of radioactive impurities, commonly uranium and thorium. The radiation created defects, Frenkel pairs, have been investigated by means of X-rays (Cu K<alpha>1) diffraction measurements of the diffuse scattering close to the Bragg reflections (see Salje et al.). 3D-images of the diffuse scattering were obtained for several reflections in zircons with different degrees of damage, roughly with doses between 1-4 1018 <alpha>/g. The diffuse intensity was thus analysed along the three main crystallographic axes as a function of the damage. Our recent results are interpreted in the frame of existing theoretical models.
Weber WJ, Ewing RC & Wang LM, J. Mater. Res, 9, 688-698, (1994).
Salje EKH, Chrosch J & Ewing RC, Am. Min., to be published
In recent years the high-pressure behavior of aluminosilicates, especially zeolites, has been an object of much interest. Analcime is a highpore framework aluminosilicate, of the ANA structure type. The secondary building units are the 4-, 6- and 8-membered rings. Earlier Hasen and Finger (1979) reported X-ray study of 4 phase transitions under pressure below 30 kbar in natural analcime (Golden, Colorado). The presence of H2O molecules in pores of natural analcime moves the phase transition pressure to higher values and complicates the observation of the process, so we have chosen an anhydrous analcime for our study.
The Raman experiment was conducted in a diamond anvil cell with hydrostatic pressure of medium (glycerine) up to 37 kbar, the pressure in apparate being defined by the ruby line shift. Raman spectra were recorded on the OMARS 89 DILOR spectrometer combined with OLYMPUS microscope. Single crystals of dehydrated analcime (initial formula Na1.88[Al1.88Si4.12O12] x 2H2O, from Nidym River, East Siberia), monoclinic, placed in the diamond anvil cell, exhibited two phase transitions at 4 and 12 kbar, each of them is revisable with considerable hysteresis (~2 kbar). The well-defined picture of phase transitions was obtained for nontwinned crystal of ~50x150x190 µm size. At 1 kbar complete darkening was observed. A gradual lightning of the crystal (phase I) was observed in crossed polarizators with pressure increasing up to 4 kbar. At the transition point (4 kbar) a sharp zone of new phase II appeared moving from edges to the center of the crystal with gradual attenuation. Phase II was completely dark at 4 kbar and gradually lightened with pressure increasing. At 12 kbar the second crystal darkening in crossed polarizators occurs. Like in the two previous phases, slight darkening was observed in III-phase without crossed polarizators at the transition point. This darkening is supposed to be a result of crystal strain occurred at phase boundary. New phase formation was fixed by changing in Raman spectra. At 3.8 kbar phases I-II were observed. Phase II was detected visually (under optic microscope) and by strong in the shift of Raman bands with pressure. It was clearly seen in a strong shift of the original strong doublet 480, 500 cm-1 of bending O-T-O vibrations (T=Al, Si). At phase transition II-III this doublet was substituted by a triplet (500, 525, 560 cm-1).
The disturbance of crystal darkening with pressure increasing is supposed to be due to rotation of optical axes at different angles caused by inhomogeneous strain appeared in the crystal. New high-pressure phase releases macroscopic strains. Phase transitions in dehydrated analcime are connected with 4-membered rings deformations. Dehydrated analcime is monoclinic (I2/a) and 4- membered rings deformation may cause the symmetry reduction to triclinic.
Several phase transitions occur in CaTiO3 perovskite at temperatures above 1000°C. There is a progression from orthorhombic (Pbnm) to cubic (Pm3 m) symmetry. The range of possible space groups for transformations from Pbnm to Pm3m have been calculated; and various structures have been previously observed by X-ray and neutron scattering; and inferred from calorimetry. Changes in wave-speed are expected to accompany these changes in structure.
High temperature (to 1300°C) ultrasonic determinations of wavespeed in polycrystalline CaTiO3 have resulted in the observation of discontinuities in wavespeed at 1100°C, 1200°C and 1300°C. These changes in wavespeed can be associated with known phase transitions.
Seismic frequency determinations of shear-wave speed and dissipation have not observed the effects of phase transitions. Instead grain-boundary sliding accommodated by diffusion creep results in a strong decrease in shear-wave speed and increase in attenuation at temperatures ~900°C, well below the onset of the first symmetry transition. It is possible that grain boundary films have a significant effect on the creep rate.
A series of feldspars of composition Ca1-xPbxAl2Si2O8, topologically isostructural with anorthite, was synthesized from melt and by solid-state reaction, and subsequently isothermally annealed in the temperature range 900-1100°C. The relevant cell parameters were measured at room temperature. In the Pb-rich region of the join An-PbF, a continuous transition from triclinic to monoclinic symmetry was observed at a composition near An10PbF90, on increasing Pb content. This symmetry increase from I-1 to I2/c corresponds to a zone-centre ferroelastic transition, analogous to that observed along the join CaAl2Si2O8 - SrAl2Si2O8 (An-SrF) (McGuinn and Redfern, 1994).
In situ HT X-ray powder spectra were performed and the duration of the heating runs was chosen so that a negligible change in the degree of Al-Si order would be expected to occur. A sample with nominal composition An80PbF20, triclinic at room temperature, undergoes on heating the same ferroelastic transition from I-1 to I2/c. The transition temperature (Tc) varies significantly if the sample is treated with different annealing times. Furthermore, samples with longer annealing times (at least 4 days at T = 1000°C) show a continuous (second-order) behaviour (Tc about 620°C; ß = 0.49(2)), while samples annealed for shorter times, show a discontinuous (first-order) behaviour and a significantly lower Tc. This behaviour is similar to that observed in intermediate Ca-Sr feldspars (Tribaudino et al., 1993; McGuinn and Redfern, 1997; Dove and Redfern, 1997). As a conclusion, it appears that in Ca-Pb feldspars the degree of Al-Si order, related to the annealing duration, does indeed couple strongly to the order parameter for the ferroelastic process. It must be stressed that compositional fluctuations and an inhomogeneous distribution of the non-tetrahedral cations can play a significant role.
Dove MT & Redfern SAT, Amer. Mineral, 82, 8-15, (1997).
McGuinn MD & Redfern SAT, Amer. Mineral, 79, 24-30, (1994).
McGuinn MD & Redfern SAT, Eur. J. Mineral, 9, 1159-1172, (1997).
Tribaudino M, Benna P & Bruno E, Phys. Chem. Minerals, 20, 221-227, (1993).
Previous investigation has shown that in anorthite the space group at room T is P-1, while triclinic Sr-rich samples along the join CaAl2Si2O8 - SrAl2Si2O8 (An-SrF) display a I-1 space group (Tribaudino et al., 1995). This suggests that a zone boundary phase transition from P-1 to I-1 occurs at room T due to Sr substitution in Al-Si ordered anorthite. The onset of the phase transition P-1 to I-1 with increasing Sr substitution and hence increasing cation size in the non-tetrahedral cation was therefore investigated in samples with almost constant Qod and Al-Si ratio fixed at 1:1. The evolution of the reflections violating I-1 space group with T and increasing Sr content was as well measured. Two sets of samples, with composition from An100 to An50SrF50, were synthesised from gel and annealed by dry and hydrothermal treatment, respectively at 1450°C for 96 h and at 600°C and 2 kb for 300 h.
At room T a P-1 stability field is present up to An80SrF20, revealed by sharp 'c' type (h+k even, l odd) reflections in well ordered hydrothermal samples. The 'c' antiphase domain size is 100 - 200 Å for both An100 and An80SrF20. A P-1 configuration is however still present up to An50SrF50 in all samples, indicated by weak and diffuse 'c' reflections.
HT in situ TEM observation was performed on natural well ordered anorthite and on hydrothermal An80SrF20. In anorthite sharp 'c' reflections disappear at about 240°C, leaving only a diffuse component which becomes apparent in the proximity of the transition and disappears only at T much higher than Tc. These results are in agreement with previous observations (Van Tendeloo et al., 1989). In An80SrF20 the sharp component in 'c' reflections coexists with a diffuse component already at room T. At about 150°C the sharp component disappears, while the diffuse component disappears completely around 350°C, and could not be found at higher temperatures even in overexposed photographs. An I2 vs T plot suggests a Tc = 155°C for An80SrF20, with a significant decrease in the Tc respect to anorthite (Tc = 240°C, Redfern and Salje, 1992). The slope in the I2 vs T plot in An80SrF20 is lower than in An100.
The above results for An80SrF20 are in agreement with the Redfern and Salje (1992) model for the transition from P-1 to I-1 in anorthite that interprets the diffuse component beyond the transition as related to Ca fluctuations that decay rapidly as T becomes higher than Tc.
Redfern SAT & Salje E, Phys. Chem. Min., 18, 526-533, (1992).
Tribaudino M, Benna P & Bruno E, Amer. Mineral., 80, 907-915, (1995).
Van Tendeloo G, Ghose S & Amelincx S, Phys. Chem. Minerals, 16, 311-319, (1989).
Grain shape fabrics and optical microstructures of embedded K-feldspar single grains in quartz have been investigated in a centimeter-scale shear zone in the Teshima granite, the Ryoke metamorphic belt, Japan. Optical microstructures in quartz are consistent with the accommodation of some intracrystalline plastic strains by dislocation creep. However, embedded K-feldspar grains have no evidence for dislocation creep. Instead, they seem to be accommodated by diffusion creep. Phase boundary morphology between K-feldspar and quartz shows that phase boundary cusps consistently point along the foliation and parallel to the mineral lineation (i.e. in directions of inferred finite extension) which implies that phase boundary motion and cusp formation occurred during deformation. However, the morphology of quartz- K-feldspar phase boudaries changes gradually as the grain size of embedded K-feldspar grains decreases, so that the cusp disappears to form a flat crystalline facet whose morphology is similar in certain respects to the phase boundary morphology of metals annealed experimentally under hydrostatic stress conditions. Furthermore, this phase transition appears to vary with shear strain, since the critical grain size for faceting in K-feldspar grains are smaller at the region of higher strains than those at lower strains. Therefore, this transition cannot be explained by either diffusion creep or static annealing alone. We propose that phase boundary motion resulted from a competing process between diffusion-accommodated intracrystalline strain and annealing to reduce the interfacial free energy. We present a model for shape change of inclusion mineral during contemporaneous deformation and annealing in two-dimensional case and discuss a possibility to use the phase boundary transition to estimate strain rate during ductile shearing as a paleo-piezometer.
Mackinawite (tetragonal Fe1+xS) is an important diagenetic iron sulfide phase in many anoxic marine environments, and also occurs as a corrosion product of ferrous metals in sulfidic systems in the petroleum and nuclear industries. Well-crystallized mackinawite is difficult to synthesize quickly from aqueous solutions, but can be readily synthesized from metallic Fe and aqueous sulfide. The product, though, undergoes changes in the stacking arrangement of the mackinawite layers in the first few minutes after formation on the iron surface. We believe the process involves previously-undescribed details of the sulfide corrosion mechanism.
Experimental low-temperature electrocrystallization of mackinawite on body-centered cubic alpha-iron in aqueous sulfide solutions produces sufficient product in a few minutes that the structural changes during ripening can be followed by powder X-ray diffraction analyses. Observed anomalies in the powder diffractograms of young mackinawite include unusually large d(001) of up to 5.37Å (cf. 5.03Å for normal mackinawite) and systematic variations in the I(200)/I(112) intensity ratio. Calculated diffractograms of model mackinawite-like structures with various offsets parallel mk[100](001) of adjacent S-Fe-S sheets reproduce the experimental results, and extended Hückel calculations were used to follow electronic interactions with ripening.
Our results indicate a newly-described corrosion mechanism is involved:
1) The high degree of crystallinity of mackinawite produced is the result of epitaxial nucleation on <alpha>-iron, with mk[100](001) parallel Fe[110](001).
2) Adjacent mackinawite-like sheets form with an initial offset of ca. 2.027Å in the mk[100] direction because of the locations of Fe atoms in the <alpha>-iron {002} planes. As the sheets detach from the iron, they shift laterally into the normal mackinawite positions, but in doing so, cause filled S 3 pz lone pair orbitals in adjacent sheets to first approach, then move away from one another.
3)The unusually large (001) and variable d-spacing observed in young synthetic mackinawite formed from alpha-iron is caused by the repulsion between the S 3 pz lone pairs in adjacent S-Fe-S layers in the distorted mackinawite structure as lone pairs in adjacent sheets move past one another during ripening.
4) The anomalous intensity ratios observed for the (200) and (112) X-ray diffraction peaks arise from {001} layer offsets controlled by the {002} layer arrangement in <alpha>-iron.
Phenomenological observations on natural series allowed to think that ammonium-bearing illite/muscovite could be formed, at temperatures above 200°C, by reaction of kaolinite with ammonia derived from thermal degradation of organic nitrogen (Daniels and Altaner 1990; Sucha et al, 1994). More recent isotopic studies on the anthracite field of Pennsylvania showed that during anthracitization, the organic nitrogen content decreases rapidly while organic nitrogen isotopic composition does not change with increasing rank: 15Norg = 4.5 ± 1.5‰ (Ader et al, 1998a), and that ammonium-rich-illite in high rank coals have an isotopic composition higher than the isotopic composition of the organic matter: average 15NNH4 = 9‰ (Ader et al. 1998b). The striking contrast between the rapid and sharp decrease of organic nitrogen content and the invariance of its isotopic composition during the whole anthracitization, associated with the formation of ammonium-rich illite, suggests that ammonia is an important product of the denitrogenation process. The discrepancy between the isotopic composition of the total nitrogen released and that of ammonium-illite, suggests that isotopically light nitrogen (molecular, ammonium and/or ammonia) escaped from the coal.
In the present work, gold cell experiments were performed to simulate meta-anthracitization. Anthracite samples were first heated in sealed gold cell at 530°C-600°C, 0.2 GPa during 7 days or 15 days. The simulation showed that, as in natural samples, organic N/C ratio decreases while organic 15N remains unchanged. In a second type of experiments, pure kaolinite with carbazole (taken as a model of organic nitrogen) was heated in sealed gold cell at 530°C, 0.2 GPa during 7 days. As shown by XRD, ammonium-illite was readily synthesised, wether water was previously added or not. Experimental modeling demonstrate that ammonium-illite/muscovite can derive from thermal maturation of organic nitrogen and kaolinite assemblages, as inferred from phenomenological observation on natural series.
Ader M, Boudou J-P, Javoy M, Goffé B, Daniels E, Org. Geochem, in press, (1998a).
Ader M, Javoy M, Boudou J-P, Hieronymus B, Roux J, Daniels E, Min. Mag, 62A, 13, (1998b).
Daniels EJ, Altaner SP, Amer. Min, 75, 825-839, (1990).
Sucha V, Kraus I, Madejova J, Clay Minerals, 29, 269-377, (1994).
Structural transformations study of fine grained manganese minerals was conducted in three directions: a) phases transitions study in native conditions; b) high temperature heating in dry air; c) experimental modelling at high temperatures and pressures in different media (neutral, acid, alkaline).
The object of our investigations is fine-grained minerals from pelagic ferromanganese nodules. United structural basis of fine dispersional manganese oxides (Mn-octahedranal layers) make possible their mutual phases transitions. Following structural transformations are observe in nature most frequently: birnessite-asbolan; buserite1-birnessite; buserite1-buserite 2; todorokite-asbolan.
Under high temperature heating in dry conditions manganese phases are changed. Its way to transform is dependent on which mineral it represents. Formation of jacobsite is essential for all samples. If nodules are enriched in iron then gematite is formed, if they are enriched in manganese then bixsbyite is formed.
On the basis of hydrothermal experiments fields of crystalline phase stability were revealed. Under hydrothermal conditions weakly crystalline minerals and roentgenamorphous phases transform to phase with stable structures. Products of experimental modelling in neutral media are devided into mineral phases with two structure types: olivine structure type (tephroite, fayalite) and spinel structure type (hausmanite, magnetite, ferrites and manganites of nickel and cobalt). Under hydrothermal conditions in acid media hematite, jacobsite, tephroite, braunite and hausmanit are formed. Under hydrothermal conditions in alkaline media are formed there following associations of mineral phases: asbolan; asbolan-braunite; asbolan-hausmanite; jakobsite-tephroite.
The temperature dependence of the intracrystalline Fe,Mg partitioning (KD) in two olivine crystals (Fa11) separated from the Acapulco meteorite was determined by single crystal X-ray structure analyses. A point-charge model was employed to account for the charge accumulations on the Si-O bonds. The crystals were equilibrated at 750°C, 650°C, and 550°C. The results yield
lnKD=0.345(60)-204(53)/T,
which is in qualitative agreement with earlier work. Comparison of these data with the higher temperature data of Artioli et al. (1995) suggests an unusual temperature variation of the Fe,Mg distribution within two temperature regimes. In the low T regime (T<800°C), Fe orders into the M1 site with increasing temperature whereas it tends to concentrate in the M2 site at T>850°C. Cooling rate calculations have been performed to assess the potential of olivine as a geospeedometer. The required expression of the disordering rate constant, kdis, for the Fe,Mg exchange process has been derived from the Fe,Mg interdiffusion data of Chakraborty (1997) and the theoretical formulation of Ganguly and Tazzoli (1994) relating interdiffusion and order-disorder kinetics:
kdis[d-1]=(1.68±0.25)*1015 exp[-(54.2±4.3)][kcal/mol]/RT
at f(O2)=10-12 bar. However, uncertainties in the temperature variation of KD and kdis remain and make it difficult at present to establish olivine as a geospeedometer.
Artioli G, Rinaldi R, Wilson CC & Zanazzi PF, American Mineralogist, 80, 197-200, (1995).
Chakraborty S, Journal of Geophysical Research, 102, 12,317-12,331, (1997).
Ganguly J & Tazzoli V, American Mineralogist, 79, 930-937, (1994).
An olivine single crystal (Fo93) from a mantle xenolith in kimberlite ("Udachnaya" pipe, Siberia) was annealed in air at 700°C, 9h. Oxidation products were studied by EMPA, FTIR, UV-spectroscopy, electron energy-loss spectroscopy EELS and TEM. Ferric iron and hydroxyl were detected in UV and in IR spectra. Neither hematite, magnetite nor laihunite were revealed in TEM. Dislocations in olivine decorated with inclusions of several hundred nm in size were observed. The inclusions are predominantly of hexagon-like shape, and elongated parallel to [001] of the olivine matrix. The coherent interface between inclusion and matrix is parallel to (010), ± (011) and ± (011) of the olivine matrix. Cations of the inclusions are Mg, Fe and Si with an Si/(Mg+Fe) ratio of 0.28-0.38. Inclusions are enriched in Fe with respect to the olivine matrix. They are heterogeneous with a rim close to the composition to the olivine matrix and a Fe-rich Mg-poor core. Often, sharp core/rim boundaries are observed with moiré contrast in the core. EELS spectra of inclusions exhibit a pre-peak at 528 eV indicative for Fe3+ (van Aken et al, 1994) and/or OH (Wirth, 1987). A positive correlation between the intensity of the pre-peak at 528 eV and the intensity of the Fe-L3,2 peak was observed. SAED and CBED diffraction patterns, with the electron beam oriented along [100], [211] and [311] directions of the olivine matrix, show a superposition of reflections from different phases: olivine matrix (m); SiO2 beta-cristobalite (ß-cr) and FeOOH feroxyhite (f) in the core of inclusions; a secondary olivine (ol) in the rim of inclusions. The following orientation relationships are observed: (100)ol(m) // (001)FeOOH // (111)ß-cr // (100)(ol); [010]ol(m) // [1-10]FeOOH // [110]ß-cr // [010](ol). The complex microstructure of the inclusions is interpreted as solid state transformation of a precursor phase such as hydrous Mg-silicate (HMS). It is speculated that the inclusions are an exsolution of HMS from the olivine matrix saturated by OH-bearing point defects accumulated at dislocations. The transformation of HMS could have occurred during experimental annealing and is assumed to involve oxidation and partial dehydration of the precursor phase, accompanied by phase separation inside inclusions and Mg-Fe exchange between Fe-free HMS inclusions and the Fe-bearing olivine matrix.
van Aken PA, Liebscher B & Styrsa VJ, Phys Chem Minerals, 25, 494-498, (1998).
Wirth R, Phys Chem Minerals, 24, 561-568, (1987).
In the Erzgebirge Crystalline Complex, metabasaltic eclogite and garnet peridotite bodies occur within high grade granulitic gneisses. The peridotite lenses (up to several 100 m) sporadically contain small boudins of meta-rodingite as well as folded layers of garnet pyroxenite. The coarse-grained pyroxenite consists of diopsidic clinopyroxene (di67jd17-19hed05), pyropic garnet (pyr70alm20grs07), magnesian orthopyroxene (en88) and minor rutile. Applying the 'Al-in-opx' barometer, the cpx-opx solvus thermometer, and the opx-grt and cpx-grt exchange thermometers, a high-pressure (HP) equilibrium stage at about 850-900°C and 33 kbar can be inferred, matching the PT estimates for both the host peridotite as well as the eclogite. Eclogite shows a clockwise PT path whereas the mantle-derived ultramafic rocks experienced an anticlockwise PT evolution with an early stage at higher temperature and lower pressure compared to the HP equilibrium stage. This is inferred from inclusions of early Cr-Al spinel within garnet and exsolution features in clinopyroxene. Spinel inclusions are zoned with increasing Cr from core to rim which is ascribed to the formation of garnet at the expense of spinel due to the reaction opx + cpx + sp = g + fo in response to increasing pressure. Exsolution features in clinopyroxene are complex. Early exsolution of orthopyroxene points to a higher temperature prior to the HP stage. Lamellae with orthopyroxene (0.3-1 µm) are close to the (100) planes of diopside. During decompression and further cooling pigeonite lamellae are formed. Most lamellae (0.05-0.1 µm) are oriented close to the (100) planes, with about 20° deviation. In addition, z-shaped pigeonite lamellae were observed, with an orientation close to (001) in the central part (0.1-0.2 µm) changing to '100' at both ends. The average thickness of the '100 tails' is in the same range as that of the simple '100' lamellae. Judging from the 20° inclination, the average composition and the calculations of Skrotzki et al. (1991) an exsolution temperature around 700°C can be estimated for the '100' pigeonite lamellae.
Skrotzki W, Müller WF & Weber K, J Eur Mineral, 3, 39-61, (1991).
An excellent opportunity to investigate the influence of the PTT factor on compositional and structural features of rock-forming, trace and ore minerals is provided by the Kola Superdeep Borehole (KSDB), which stripped a 6.84 km deep section of sedimentary-volcanic rocks of the Early Proterozoic Pechenga Structure and the depth interval from 6.84 to 12.26 km, which produced cores of Archaean TTG-gneisses (45%), high-alumina gneisses (20%), amphibolites (30%) and granites (5%). In the context of a new IGCP Project-408 "Comparison of composition, structure and physical properties of rocks and minerals in the Kola Superdeep Borehole and their homologues on the surface" (see Episodes, 1998, vol. 21, N 2, p.111), it is particularly important that all the rocks present in many kilometers of core samples have their homologues in outcrops of the Pechenga region. Rocks on the surface and rocks at depth are considered to be homologous as they were initially of the same composition and structure, occupy the same stratigraphic position, and originated in the same PT-conditions; they are not completely identical due to different posterior evolution and, especially, the exhumation way. Specifically, rocks and minerals from the core have become tempered as a result of fast lifting from depth to the surface, and they have preserved the record of thermodynamic conditions of the middle crust. To obtain data of this kind, the Project investigates fine features of mineral composition, the degree of mineral nonstoichiometry, exsolution structures, cation and magnetic ordering, the forms of water and various gaseous substances, etc. There is still much work to be done to examine such phenomena as phase transitions, order-disorder transformations, isomorphic capacity, etc. in silicates, alumosilicates, sulphides and other minerals, and researchers from different countries, especially those who have access to state-or-the-art analytical techniques, are most welcome to collaborate in the Project.
Study of the structure of non-crystalline solid carbon, which is responsible for the most essential properties of shungite rocks (shungite occurrences in the north-western part of the lake Onega, Karelia), is of interest because of a number of its physico-chemical properties underlying its many perspective applications.
Weak structural organization of carbon in shungite and its roentgenoamorphity account for the fact that high-resolution methods of direct observation are the most perspective for shungite study. Among these methods are scanning tunneling microscopy (STM) and atomic force microscopy (AFM). They allow to obtain microtopographic portrays of the surface up to atomic resolution.
STM and AFM were applied to study highly-carbonaceous (to 98%) vein shungite from various deposits (Shunga, Tchebolaksha, Maxovo, Zazhogino). Shunga and Maxovo rocks are altered primary shungites (Maxovo shungites are characterized by a higher temperature and Tchebolaksha are characterized by a higher pressure of formation compared to Shunga ones), Zazhogino shungites were redeposited, these factors being responsible for the different structural organizations.
We relied on the obtained images of the surface, which is a combination of hills with a tendency of similar orientation, to perform statistical analysis of linear dimensions of the elements of shungite structure. The nanostructures observed on the surface were interpreted in terms of V.V. Kovalevsky's model on fullerene-like supermolecular formations in shungite, known as globules. Surface images of Tchebolaksha shungites hardly show any globules, the profiles displaying blocks varying in shape and size, the latter ranging from one to several hundred, even thousand nanometers. The images disclosed different organization of globules in Shunga and Maxovo shungites: the globules of Shunga in the former are typically aggregated as lenticular accumulations about 200-300 nm in size, elongated globules displaying a tendency to similar orientation within the clusters. Characteristic of Maxovo shungites is the globules' alignment in chains, sticking together to form layers. Globular organization of Zazhogino shungites is more close-packed than of Maxovo and Shunga ones.
The zoned muscovite was identified by backscattered electron imaging and electron microprobe analysis in samples of the tin-bearing muscovite granite from Ervedosa showing some metasomatic effects.
This Hercynian granite is peraluminous (A/CNK>=1.13) and of S-type . It belongs to a sequence ranging from muscovite-biotite granite to muscovite granite. This sequence defines a Rb-Sr whole-rock isochron of 327±9 Ma and initial 87Sr/86Sr ratio of 0.7155±0.0018. The muscovite granite was derived from a muscovite-biotite granite magma by fractional crystallization of plagioclase, biotite, quartz and K-feldspar, but the behaviour of LIL elements was controlled by magmatic fluids (Gomes, 1996).
Stanniferous quartz veins cut Silurian mica schists and the Hercynian granites and are spatially linked to the muscovite granite. Sn was exploited in the Ervedosa mine. Ore-forming fluids altered hydrothermally this granite particularly at the quartz vein walls. It shows replacements of K-feldspar by albite and also of albite by K-feldspar, kaolinization of K-feldspar, chloritization of biotite, zoned muscovite, white mica formed by alteration of plagioclase and white mica replacing the other minerals and around the quartz ocels. So the altered granite contains a variety of postmagmatic white micas (e.g. Neiva, 1992; Dempster et al., 1994). The hydrothermally altered samples show enrichment in Na2O, P2O5, Sn, Nb, Zn, Sr and impoverishment in MgO, K2O, Li, Ni, Zr, Pb and REE (Gomes, 1996).
In the zoned muscovite flakes, the rim is generally richer in Fe+Mg, Rb and poorer in Alvi, Aliv + Alvi, Na than the core. The diagrams of Aliv+Alvi versus Fe+Mg and Ti - (Fe+Mg) - Na show the best graphical distinctions between rim and core. In the TiO2 - MgO - total Fe2O3 diagram of Monier et al. (1984), some of the core compositions fall in the field of magmatic muscovite, while the others and the rim compositions plot in the field of late- to post-magmatic muscovite. The Al-rich core represents the recrystallized magmatic muscovite and the more celadonitic overgrowth formed during mineralizing fluid infiltration and retrogression of low biotite content.
Acknowledgements. We are grateful to Prof. B. J. Wood and Dr. J. C. Schumacher for the EUGF-Bristol facility, contract ERBFMGECT980128 to analyze the zoned muscovite.
Dempster TJ, Tanner PWG, Ainsworth P, Amer, Mineral, 79, 536-544, (1994).
Gomes MEP, Unpublished Ph. D. thesis. Univ. Trás-os-Montes e Alto Douro, Portugal, (1996).
Monier G, Mergoil-Daniel J, Labernardière H, Bull. Mineral, 107, 55-68, (1984).
Neiva AMR, Memórias Notícias Publ. Mus. Lab. Mineral. Geol. , Univ. Coimbra, 113, 75-91, (1992).
Micas from the mantle rocks: the sanidine lamproites of transitional type (1720±8 m.y) and olivine lamproites (1230±5 m.y.) of the Baltic shield (Nikitina et al., 1998) are investigated by methods of the electron microprobe analysis, Mössbauer and Infrared spectroscopy. Phlogopites from sanidine lamproites can be assigned to phlogopite-biotite isomorphous series and contain 1-4wt.% TiO2, 3-13wt.% FeO, 17-25wt.% MgO, 1-4wt.% Fe2O3, 12-15wt.% Al2O3, 41-43wt.% SiO2. Phlogopites from olivine lamproites are the members of phlogopite-tetraferripholopite isomorphous series. They are depleted in Al2O3 (3-12wt.%) and enriched in TiO2 (3-6wt.%). The contents of the basic oxides are 39-44wt.% SiO2, 3-5wt.% FeO, 3-12 wt.% Fe2O3, 21-26wt.% MgO. The degree of oxidation state of iron in micas from olivine lamproites (Fe3+/Fe2+ = 0.57-2.00) and lamproites of transitional type (Fe3+/Fe2+ = 0.21-0.36) is much greater than one in micas from core rocks. In micas of phlogopite-tetraferripholopite series Fe3+ is distributed between tetrahedral (IV) and octahedral (VI) positions in the relation Fe3+ IV/Fe3+VI = 0.60-2.29. At deficiency Si and Al (< 4.0 f.u.) the free tetrahedral positions are filled by both cations Fe3+ and Ti4+, even at the high contents Fe3+, sufficient for occupation tetrahedral site deficiency. A part of cations Fe3+ and Ti occupy octahedral positions as well. In micas of phlogopite-biotite series Fe3+ occupies only octahedral position M2. Tetrahedral positions which have been not populated Si4+ and Al3+, are filled by Ti4+. The character of Fe2+ distribution on octahedral positions M1 and M2 in phenocrysts and groundmass plates is different. In phenocrysts disordering distribution of Fe2+ cations is observed. In groudmass plates Fe2+ is ordered in position M1. The significant quantity of vacancies in interlayer position and in octahedral layer and presence of molecular water in various energy positions are established.In micas a quantity of H2O, chemically active gases CO2, N2 (up to 200 ccSTP/g) and inert gases (Ar) is higher in comparison with micas from core rocks. So, the micas from lamproites are characterized by a high degree of crystal structure defects shown in a high degree of cations disordering, in significant quantity of vacancies in octahedral layer and interlayer position. The vacancies, apparently, were used as effective structural traps for fluid components.
This work was supported by RFFI grant 96-05-64685.
Nikitina L.P., Levskiy L.K., Beliatsky B.V., Juravlev V.A., Lepekhina E.N., Antonov A.V., Abs. of Inter. Conf. Genesis problems of magmatic and metamorphic rocks, S.-Petersburg, p.p 116-117, (1998).
The curvature or the tubular morphology often observed in serpentines is generally explained by a structural mismatch between the tetrahedral and the octahedral sheets, such a mismatch being directly related to the relative sizes of the tetrahedral and octahedral cations. In order to test this assumption, Ni3(Si,Ge)2O5(OH)4 serpentine syntheses have been performed, with progressive substitution for Si by the larger Ge ion, in tetrahedra. Five syntheses (Ge substitution rates 0%, 36%, 43%, 65% and 100%) have been studied by high-resolution transmission electron microscopy and analytical electron microscopy.
In the 0%-Ge sample, the crystals show a cylindrical "tile" shape ("roman tile"), with a curvature radius of about 10nm. In the 36%-Ge sample, two types of serpentines are observed: (i) curved structures, of "roman tile" or tubular shape, with 20 to 40 nm curvature radius, their tetrahedral composition ranging from Si90Ge10 to Si75Ge25; (ii) plane structures, with hexagonal or triangular shape (view normal to the layers), and Si75Ge25 to Si45Ge55 tetrahedral composition. In all other samples(43%, 65% and 100%-Ge), only plane crystals of hexagonal shape are present. Their tetrahedral compositions (for the 43% and 65%-Ge samples) are spread out from Si75Ge25 to Si15Ge85.
The preceding results show that the two structural morphologies are directly related to the chemical composition: (i) curved structures, such as "roman tiles" or tubes, occur for a tetrahedral composition ranging from Si100 Ge0 to Si75 Ge25; (ii) plane structures, for Si75 Ge25 to Si0 Ge100 composition. In other words, owing to the greater size of Ge cations, the mismatch between the octahedral and the tetrahedral sheets is overcome - thus leading to plane structures - when the (tetrahedral) Ge substitution rate passes beyond the Si75 Ge25 value.
According to the theory of elastic thin plates, the curvature radius R of the plate is related to the difference between the surface stresses <sigma>+ and <sigma>- (on the two sides of the plate):
1/R= 6(1-<nu>)(<sigma>+ - <sigma>-)/Eh2
where E is the Young's modulus, <nu> the Poisson's ratio, and h the thickness of the plate. In the case of a single serpentine layer, 0.7-nm thick, the value <sigma>+-<sigma>- = 1 J/m2 - considered as a reasonable example for Si-rich serpentines - would then produce a curvature with a radius R of about 17nm.
The preceding value belongs to the range R = 10 to 40 nm, observed in the "tiles" and tubes of the studied samples. It may be concluded that these "tiles" and tubes are probably formed by stacking of such elastically curved single serpentine layers. This process requires the creation of stacking defects, which make the whole crystal "rigid".
For more Ge-substitued serpentines (above Si75 Ge25), <sigma>+-<sigma>- probably vanishes, and the stacking of plane single layers leads to plane serpentine crystals.
Transformations and interstratifications in clay minerals have been studied both experimentally, using high-resolution transmission electron microscopy (HRTEM) and analytical electron microscopy (AEM), and theoretically, with the help of lattice energies calculations.
HRTEM-AEM examinations have been performed on layered minerals involved in the two geological important transformations (i) of smectite into illite (by hydrothermal process) and (ii) of smectite into kaolinite (by tropical weathering). Detailed observations reveal the various mechanisms which are responsible for these transformations: (a) dissolution (of smectite) and crystallization (of illite); (b) a solid-state mechanism in which one layer (of smectite) is progressively transformed into one layer (of illite or kaolinite); (c) another solid-state mechanism in which one layer (of kaolinite) is intercalated in smectite (Amouric and Olives, 1991, 1998).
Mixed-layer minerals (illite-smectite or kaolinite-smectite) are obtained as intermediate products of such transformations. The mixed-layering sequences are completely disordered (R0 type) in the less transformed samples (i.e., the more smectitic ones), and present an increasing local order (R1 and R2 types) as the transformation evolves (i.e., in more illitized or kaolinitized samples).
A new theoretical approach, based on lattice energies calculations, is used to study the structure and stability of mixed-layer minerals. Electrostatic lattice energy calculations have been made with our "overlap method" (Olives, 1986), of more rapid convergence than the classical Ewald's or Bertaut's methods. This method leads to the determination of the illite-smectite mixed-layer structure, as corresponding to the minimum of the energy. This structure may be described as the stacking of O0.5 TIT O0.5 "layers" (I, T and O are respectively the interlayer, tetrahedral and octahedral levels). In the case of the perfectly ordered ... isis ... sequence (i = illite layer, s = smectite layer), the arrangement is thus ... Om (TIT)i Om (TIT)s ... (m = middle composition between i and s). These considerations then clearly lead to a polar model for TOT layers, which display a Ti Om Ts structure. Besides, the lattice energy of the preceding interlayered ... isis ... is found to be nearly equal to the middle value between the lattice energies of illite and smectite. The stability of interstratified illite-smectite is then similar to that of the two-phases illite and smectite association.
As a consequence of the preceding model, the 1 smectite layer 1 illite layer solid-state mechanism, responsible for the interlayered sequences, consists of the transformation of one interlayer level I and the two adjacent TO0.5 units.
Amouric Mand Olives J, Eur. J. Mineral, 3, 831, (1991).
Amouric Mand Olives J, Clays and Clay Miner, 46, (1998).
Olives J, Acta Cryst. A 42, 340, (1986).
The new data on chemical composition of metasomatic and metamorphic staurolites from the high-alumina rocks were obtained. The content of the minor elements vary largely. The influence of the chemical composition of the rocks on the geochemistry of staurolites is great, especially on the content of Zr, Y, Nb, Cr, V. The variations of chemical composition of staurolites are functions of mineral paragenesis. The presence of garnet, kyanite, Mg-chlorite has a profound effect on composition of staurolites. Metasomatic and metamorphic staurolites are similar in their composition. High content of Zr and FeO in staurolites is a result of the ferriferous metasomatism with alkaline tendency. The staurolites from garnet-staurolite metasomatites are characterized by lowering content of MnO, Cr, Ni, Co. The distribution of Co, V, Cu between coexisting staurolites and garnets is close to the ideal.
Phase Transformations lead to characteristic microstructures which serve as indicators for past transformation events. One of the most common transformation mechanism in minerals related to ferroelastic phase transitions and the formation of twin structures. The twin wall patterns include individual twin walls and the formation of needle domains as the most elementary steps in the construction of hierarchical wall structures. More complex patterns include junction and tweed patterns (Salje et al. 1998). The domain patterns provide fast diffusion paths for characteristic elements. The typical case of fast oxygen diffusion along twin walls in perovskite structures is discussed. The patterns are also charateristic for subsequent annealing processes: individual twin walls have no decay mechanism while needle domains are destroyed via the retraction of the needle tip. Prolonged anneal leads to comb patterns with parallel twin domains and no junctions. Short anneal often related to coarse tartan patterns. Examples for various patterns formation mechanisms are illustrated by examples.
Salje EKH, Buckley A, Van Tendeloo G, Ishibashi Y, Am Minerl, 83, 811-822, (1998).
Improper ferroelastics of the palmierite-type, Pb3(PO4)2, and isostructural mixed crystals Pb3(P1-xAsxO4)2 and (Pb1-xMx)3(PO4)2 form W and W' domain walls at temperatures below the ferroelastic transition R3barm - C2/c. Such walls correspond to ferroelastic deformation states according to the supergroup-subgroup relation and represent pseudo-symmetry elements in the monoclinic phase, respectively pseudo-mirror planes and pseudo-binary axis. The resulting domain size covers all length scales. The trajectories of the walls are described by the model developed by Salje and Ishibashi (1996).
The cleavage plane (100) of the palmierite type samples is non-planar across a W wall which separates two ferroelastic orientation states. The spontaneous strain, <epsilon>s, determines the angle between the (100) plane and the W wall with index (113) or (113) joining two domains on either side of the wall. In pure lead phosphate this angle was found to be 178.55°. It is increased in lead phosphate mixed crystals where es is decreased compared to Pb3(PO4)2. The topology of single W walls as well as needle domains is studied using tapping mode AFM. X-ray diffuse scattering reveals diffraction signals stemming from the domain walls. The trajectories of W walls are studied using TEM. Light scattering experiments based on hard mode spectroscopy reveal the superposition of phonons stemming from local modes of the walls.
Salje EKH & Ishibashi Y, J. Phys.: Condensed Matter, 8, 8477-8495, (1996).
Many natural minerals and synthetic materials display twin microstructures resulting from displacive phase transitions. These microstructures may temporarily be removed from the sample by heating above the relevant transition temperature, though the twinning generally returns on subsequent cooling.
In anorthoclase, the spatial distributions of twins before and after brief annealing above TC are often identical. This property appears to be a common feature in many materials which undergo ferroelastic phase transitions, and is known as "twin memory". The atomic mechanisms responsible for this twin memory may be investigated by studying the annealing regimes required to remove the memory effect; how long must a sample be annealed, and at what temperature, to induce "twin amnesia".
In this study, high-resoulution X-ray diffraction has been used to investigate twin memory and twin amnesia in anorthoclase. The principal advantage of using X-ray methods, rather than the more conventional electron microscopy techniques, is that diffraction is a significantly less invasive technique. In addition, diffraction allows for the study of a larger sample volume, improving the statistical properties of the resulting data.
In anorthoclase, the primary constraint on twin amnesia is thermodynamic, rather than kinetic. The critical temperature to induce amnesia correlates well with the top of the (Na, K) solvus in disordered alkali feldspar. For this reason, the proposed mechanism for twin memory involves the segregation of alkali cations in thin lamellae at the twin boundaries. The lamellae initially form because of the enhanced kinetics and thermodynamic stability of exsolution on the twin boundaries rather than in the bulk material. However, once the lamellae have formed, they remain stable above the displacive transition temperature, and may act to pin the twins which form on subsequent cooling.
Zircon (ZrSiO4) is a commonly occurring accessory mineral whose remarkable durability is attested to by its ability to survive over 4 billion years of earth history. Because of its ability to incorporate actinides substituting for zirconium, zircon is widely used for radiometric age dating and has been proposed as a stable waste form for excess Pu-239 derived from the dismantling of nuclear weapons. In the present experiments, zircon was irradiated in-situ in a transmission electron microscope using 800 keV Kr+ ions. Experiments were also performed on hafnon (HfSiO4) and thorite (ThSiO4) to simulate the effects of a neutron poison or an actinide on the crystalline-to-amorphous (c-a) transition. At temperatures below 600°C, the materials became amorphous in two stages as an increasing function of temperature. A model was derived to relate the amorphization dose to the sample temperature, from which we can predict the c-a transition for natural zircons used for radiometric age dating and for Pu-doped zircons proposed for nuclear waste disposal. Our model was validated by examining natural zircons spanning a wide range of age and uranium content. Above 600°C, however, a new effect was observed - the ABO4 orthosilicates decomposed under irradiation into their component oxides (e.g., tetragonal ZrO2 and amorphous SiO2 in the case of zircon). These results may explain the common observation of ZrO2 crystallites in radiation-damaged natural zircons and suggest an optimum temperature "window" between approximately 400 and 600°C within which Pu-doped zircon will neither amorphize nor decompose into the component oxides. This is the first example of an irradiation-induced phase decomposition in a nonmetal - a phenomenon that sheds light on the atomic-scale process leading to amorphization of complex ceramic materials in general.
Natural zircon (ZrSiO4) often shows amorphous regions incorporated in a crystalline matrix. Such "metamict" zircons suffered radiation damage from radiogenic impurities such as U, Hf, Th. The stability of zircon with respect to irradiation determines the applicability as actinite bearing host material for nuclear waste disposal. Annealing of irradiated zircon, in a view to cure irradiation damages, promotes the unmixing of SiO2 and ZrO2 phases in the amorphous micro-domains, i.e. the formation of small zirconia crystallites in contact with amorphous silica. This unmixing process are expected to strongly influence the physico-chemical properties of zircon but their detailed structural characteristics are not understood. Metamict zircon samples from Sri Lanka were annealed at high temperature (typically 1400 K for 1 hour in an N2 flow). The annealed samples, as well as the starting material, were characterized by analytical transmission electron microscopy (ATEM). The goal of the investigation was to characterize the ZrO2/SiO2 unmixing process with the aim of understanding how radiation damages in zircon can be cured by annealing. The starting material consists of a highly metamict single crystal containing islands of residual crystalline zircon with typical diameters of ca.10 nm. The amorphous phase forms an interconnected network throughout the sample. In electron diffraction patterns, the amorphous phase display diffuse rings together with spots from the remaining zircon crystals. The recovered samples show evidence of growth of the small initial zircon islands. They now form a network within which ZrO2 microcrystallites (about 10-20 nm in size) and silica-rich domains are embedded. The ZrO2 microcrystallites are randomly oriented as revealed by the corresponding sharp rings observed on diffraction patterns. The microcrystallites are thus not in epitaxy with the zircon host. They however often exhibit Moiré fringes resulting from an interface boundary between the zirconia and the zircon host. This observation suggests that the zirconia nucleated at the edges of the zircon amorphous domains. The SiO2-rich domains exhibit very irregular shapes in the annealed samples. Further characterization of the amorphous domains in terms of connectivity and composition will be discussed.
Acknowledgments: this work results from a joint research program supported by the European TMR network on Mineral Transformations (contract number ERB-FMRX-CT97-0108).
Crystalline zircon (ZrSiO4) undergoes a transition to the metamict or amorphous state through the damage caused by decaying radioactive elements incorporated into Zr sites. The degree of damage caused in zircon per <alpha> dose, in terms of the fraction of the total number of atoms displaced, appears to vary depending on which experimental technique is used to assess it. This is most probably related to the fact that characteristic distance scales are associated with different techniques and these require order to extend over these different length scales for it to be observed. Our motivation for using NMR in this well-studied area is that the NMR signal is simply a sum of the signals from all 29Si atoms in the sample and does not depend on the existence of any long range order. We have carried out some absolute quantification of 29Si zircon spectra to verify the observability of NMR signals from all silicon sites in the sample and obtain the proportions of damaged and undamaged regions. NMR is a very local technique being sensitive to the second coordination sphere around Si. We have used this to examine the nature of the connectivity in amorphised or glass-like regions in Zircon using the large amount of 29Si NMR chemical shift data on silicate glasses.
For samples with a cumulative radiation dose of less than 0.5 x1018 <alpha>-decays/g, we observe only a single sharp peak at a chemical shift of -81.6 ppm, this is characteristic of Si in isolated SiO4 tetrahedra as expected in the zircon structure. For samples with 1-2 x1018<alpha>-decays/g, we also observe an additional broad peak centred at -90 ppm due to Si atoms in the amorphised regions. We have checked that the breadth of this peak is due to an inhomogeneous distrbution of Si Qn species and not a homogeneously broadened line through spin-echo experiments out to 8 ms (a line width of 0.5 ppm). This means the amorphised region contains a range of polymerised SiO4 tetrahedra with an average connectivity of Q2, somewhat similar to a rapidly quenched metasilicate glass. Quantification experiments and T1 measurements reveal that we are observing about 80% of the silicon in these zircons and that for radiation doses of 1.2 and 1.8 x1018 <alpha>-decays/g we observe 28% and 34% of Si to be in the amorphised regions, respectively. Thus we appear to be observing more damage per <alpha>-decay than other experimental techniques.
Natural clinopyroxenes of intermediate composition are commonly exsolved into lamellae of Ca-rich and Ca-poor composition, i.e., augite or diopside with space group C2/c and pigeonite with space group P21/c at room temperature. We used the experimental data of McCallister (1978) on the coarsening kinetics of exsolution lamellae in order to model their growth in low-Fe clinopyroxenes of the meteorite Allende and thus to estimate the cooling rates of chondrules (Weinbruch and Müller, 1995). Inconsistencies of the data of McCallister (1978) motivated us to repeat his experiments. Our starting material was a homogeneous clinopyroxene of the composition Di54.1En45.9 which was synthesized from glass at 30 kbar and 1550°C in a belt apparatus (Institute of Mineralogy, University Frankfurt a.M., G. Brey). The samples were studied by transmission electron microscopy (TEM). The starting clinopyroxene had a basis-centered (C) bravais lattice. Isothermal heating experiments were performed at 1100°C (1 h to 4320 h), 1200°C (1 h to 720 h) and 1300°C (1 h to 720 h). At 1100°C, tweed patterns with contrast modulations near (001) and (100) appeared after heating of 3 hours. The periodicity (wavelength) of the modulations on (001) was about 12 nm and remained constant until annealing times of 72 hours. After that the wavelength of the lamellae on (001) increased up to 60 - 100 nm after 4320 h of annealing. We define the onset of the increase of the wavelength as the beginning of coarsening. The identification of pigeonite and diopside even at the early stages of exsolution was enabled by lattice imaging. Between 3 and 72 hours of annealing, remarkable changes of the exsolution pattern took place. After 3 hours, small islands with a primitive structure (pigeonite) already appear within the C-phase, as visualized by (100) lattice fringes with a spacing of 0.9 nm. Then, after annealing for about 12 h, modulations on (100) are forming preferentially, displaying lamellae on (100), with (100) lattice fringes within the matrix of diopside. Between 24 and 72 hours heating, this stage is followed by the preferred development of exsolution lamellae on (001) which largely overprint the earlier formed microstructures. Rarely, some exsolution lamellae on (100) also coarsen. It is interesting to note that even at the early exsolution stages areas with different space groups have developed. At annealing temperatures of 1200°C and 1300°C, similar exsolution processes take place but at a shorter time scale. The initial wavelengths are about 22 nm for 1200°C and 32 nm for 1300°C. After 360 h annealing, the mean wavelength was about 80 nm for 1200°C and 125 nm for 1300°C. The coarsening observed in our experiments was significantly faster than that found by McCallister (1978).
McCallister RH, Contr Mineral Petrol, 65, 327-331, (1978).
Weinbruch S, Müller WF, Geochim Cosmochim Acta, 59, 3221-3230, (1995).
Clinopyroxenes (Cpx) diopside-augite (Ca(Mg,Fe)Si2O6 - (Ca,Na)(Mg,Fe,Al,Ti)(Si,Al)2O6) series from mantle rocks - a transitional type lamproites and olivine lamproites of Baltic shield (Nikitina at al, 1998) were studied using mössbauer, infrared spectroscopy together with electron microprobe. X-ray diffraction analyses show no sign of impurities or inhomogenity in Cpx. The degree of iron oxydation state in lamproite Cpx is much greater than in Cpx from metamorphic and effusive rocks (Nikitina et al, 1978). The amount of Fe3+ in lamproites Cpx is 20 - 68% of the total iron and Fe3+ occupies M1 octahedral positions and replace silicon in tetrahedral (T) positions, as well. The substitution of Si by Fe3+ indicates in Cpx structure, together with aegirine component, the presence of ferri-Tschermak's component - CaFe3+Fe3+SiO6, the existance of wich was supposed by Morimoto (Morimoto, 1988). In Cpx structure the large cations Ca and Na occupy M2 position and the amount of vacancies (1-Ca-Na), suitable for Fe2+ and Mg occupance, is small. For Cpx lamproites we establish the preferable occupance of this vacancies by Fe2+. The same M2 occupation vacancies character is observed in effusive rocks Cpx (Nikitina et al., 1978), wich show much greater ferrous iron disorder degree in lamproites and effusive clinopyroxenes, then in metamorphic ones. The occupance of M2 position by Ca, Na, Fe and Mg leads to the differences in cation enviroment composition for iron in M1 octahedron, which presents the additional doubletes in mössbauer spectra. In infrared spectra of Cpx the absorption bands of OH (3700 - 3530 cm-1) and molecular water (3440 cm-1) stretching vibrations are presented. The differences in position and number of absorption bands are connected with character of substitutions in cristallographic positions M1, M2 and T in structure of Cpx. Thus, in lamproitic clinopyroxene structures the large degree of iron oxidation and disordering of Fe3+ between octahedron and tetrahedron and Fe2+ between M1 and M2 positions are established and, also, the presence of OH-group and molecular water are detected. This work was supported by RFFI grant 96-05-64685.
Morimoto N., Amer. Miner., 73, 535-550, (1988).
Nikitina L.P., Ekimov S.P., Maslenikov A.V., Osherovich E.Z., Babushkina M. S., Jemay A.X., Cation distribution and thermodynamic of iron-magnesium silicate solid solutions, 1-239, (1978).
Nikitina LP, Levskiy LK, Beliatsky BV, Juravlev VA, Lepekhina EN, Antonov AV, Abs. of Inter. Conf. Genesis problems of magmatic and metamorphic rocks, S. -Petersburg, 116-117, (1998).
Clinopyroxenes (Cpx) with low Ca or Na-contents e.g. enstatite, ferrosilite, pigeonite or kanoite, exhibit a complicated polymorphism with pressure and temperature. At ambient conditions most of these pyroxenes have primitive space groups, monoclinic P21/c or orthorhombic Pbca, which are characterized by two different [Si2O6]-chain angles. While phase transitions involving orthopyroxenes are reconstructive, clinopyroxenes undergo displacive phase transitions at high-temperature as well as at high pressure in which both chains become equal and the symmetry increases to C2/c. An in-situ powder diffraction study of kanoite in a heated diamond anvil cell (DAC) using synchotron radiation demonstrated that the high-temperature form (HT C2/c; chains stretched) and the high pressure form (HP C2/c; chains strongly kinked) have distinct stability fields in the P/T space.
A high-pressure single-crystal X-ray study of the P21/c - HP C2/c transition shows that it is first order. At the transition pressure both phases coexist in a "single-crystal", separated by a phase boundary region observable as a white line. We were able to measure the cell dimensions of both phases simultaneously and varied stepwise their volume proportions in the transition region. Since the small volume change applied on the sample chamber was fully consumed by the volume change of the phase transition, the pressure remained constant during the phase transition. A possible interpretation is that the transition is martensitic and we are observing the effect of pressure buffering in the DAC. A strain energy contribution is assumed to stabilize the two phase region, breaking the classical Gibbs phase rule by providing an extra degree of freedom.
In Li-pyroxenes the high-temperature phase transition is shifted to below room temperature, with the consequence that LiFe3+, LiSc3+ and LiIn3+-cpx are reported with C2/c-symmetries at ambient conditions and may be regarded as equivalent to HT C2/c forms. Since the transition temperature increases with pressure, we performed a single-crystal X-ray study on LiSc3+-Cpx in a DAC. The h+k = odd reflections, characteristic for the P21/c symmetry, appeared at approx. 1 GPa. Nonlinearities in cell parameters are significant, but much weaker than those reported in temperature dependent P21/c - HT C2/c phase transitions (e.g. pigeonite). Structure refinements at 2.2 and 4.8 GPa display different O3-O3´-O3´´ angles of the A and B chains. A similar evolution of cell parameter vs. pressure is reported for ZnSiO3-Cpx in which a second, much stronger nonlinearity at higher pressure indicates the transition from P21/c to HP C2/c.
The temperature dependence of cation disordering in Fe-free Mn-cummingtonite has been measured between room temperature and 973 K by in situ time-of-flight neutron powder diffraction, at the POLARIS diffractometer, ISIS neutron spallation source, Rutherford Appleton Laboratory. At low temperatures the site scattering over the four M sites shows that Mn is highly ordered onto the large M4 site. We observe the onset of disordering of Mn and Mg between the M4 site and the M(1,2,3) sites at 823 K. The degree of disorder above this temperature increase with T, by some 5% over 150 K, with Mn in M4 decreasing with temperature. At the same time the beta angle of the unit cell increases anomalously, as do the c and b cell edges. The high neutron scattering contrast between Mn and Mg allows us to refine site occupancies to high accuracy as a function of temperature, and shows promise for future studies of amphibole order-disorder by this method using hydrothermal furnaces. The influence of the low-temperature displacive transition was not observed, due to restricted instrumental beam time, but the spontaneous strain due to this transition is clearly seen in the a and c cell parameters below 400 K.
Weathered metapelites containing coexisiting, altered carpholite and muscovite have been studied by scanning and analytical-transmission electron microscopy in order to contrast and compare the products and mechanisms of weathering. Up to mm-sized, elongated magnesiocarpholite crystals coexist with muscovite in quartz segregations in the Verrucano metasediments from the northern Apennines (Italy). The country rock is a highly weathered metapelite coated with Fe hydroxides. TEM images and SAED patterns demonstrate that primary magnesiocarpholite and muscovite are well-crystallized with few defects. Fractures occur parallel to (100) in magnesiocarpholite. The fractures are the sites of the weathering products halloysite and iron hydroxide. The latter has a texture in TEM images resembling that of typical goethite. Halloysite occurs as hollow tubes 400 Å in average diameter, with 100 Å -thick walls. Tubes in a given section have variable orientations which are apparently unrelated to the parent magnesiocarpholite structure. In sharp contrast, original muscovite crystals consist primarily of alternating packets of layers of kaolinite and muscovite. The layers of kaolinite are parallel to those of original muscovite, as consistent with direct replacement of muscovite layers by kaolinite. Small numbers of I/S layers, some with the 20- Å periodicity of R1 I/S, also occur as a replacement of muscovite layers. Microtextures imply that weathering of magnesiocarpholite occurs first by production of defect-rich domains along fractures followed by dissolution at fracture-mineral interfaces. The fractures serve as pathways for fluids with dissolved reactants and products. Halloysite and iron hydroxide precipitate directly from solution. Because the structure of magnesiocarpholite is not related closely to that of a clay mineral, the reaction must involve almost complete breakdown of the structure components of magnesiocarpholite, and crystallization of products with unrelated orientation and with the disordered structure of the halloysite. The direct replacement of muscovite by more highly ordered kaolinite (and I/S) is apparently accomodated by the similarity in major components of their sheet structures, implying the possibility that at least limited portions of the muscovite structure are retained. The changes in structure and composition required by direct replacement of muscovite by kaolinite imply a dissolution-neocrystallization process, but one occurring only locally at the muscovite-kaolinite interface. I/S, which requires some interlayer K, formed locally in muscovite but not magnesiocarpholite. The contrasting mineralogical, orientation, and textural relations thus occur because halloysite crystallized as a pore-filling mineral in fractures within the structurally dissimilar magnesiocarpholite, but kaolinite (and I/S) formed as a direct layer replacement due to the similarity in structures. This is a specific example of the general rule involving textures where phyllosilicates replace other phyllosilicates on one hand, or structurally unrelated silicates on the other.
Spectroscopic methods permit an insight into mineral transformations on a molecular level. Among these methods electron paramagnetic resonance (EPR) spectroscopy is a powerful tool to detect paramagnetic species in low concentrations and to provide information on their ligand fields. Spectral changes during phase transformation can be used as fingerprint in order to get a better understanding of structural conversions in minerals.
Continuous wave (cw) and pulsed EPR spectroscopies were used to monitoring structure-bound Fe(III) and V(IV) in vermiculite during the thermal conversion into enstatite. The overall transformation occurred at about 800°C as determined by XRD and FTIR. The Fe(III) cw EPR spectrum showed that this cation was coordinated in octahedral layers. During transformation Fe(III) was incorporated into the enstatite structure. The eight-line V(IV) spectrum was stable upon heating to 500°C. The two-pulse electron spin echo envelope modulation exhibited Al and protons in the vicinity of V(IV). This suggests that the paramagnetic centers were located in octahedral sites. During the thermal break down of the vermiculite, V(IV) was metastable and became completely oxidized. The simultaneous generation of a metastable defect structure indicated a transformation via a biopyribole.
Some bacteria contain intracellular nanometer-scale crystals of greigite (Fe3S4) that cause the bacteria to be oriented in magnetic fields. The magnetic sensor helps the bacteria find and maintain an optimal position in aquatic environments that contain vertical concentration gradients. Our transmission electron microscope (TEM) observations show that ferrimagnetic greigite in these bacteria forms from nonmagnetic mackinawite (tetragonal FeS) and possibly cubic FeS. The precursors apparently transform into greigite over a period of days to months.
Greigite crystals typically show non-uniform, blotchy contrast in the TEM and contain planar defects that can be remnants of the precursor structure. Some crystals seem to be in the process of transforming from mackinawite to greigite; such crystals contain bands parallel to (222) that have spacings that are consistent with the distance between cubic close-packed layers in mackinawite. The two structures are oriented with respect to one another such that the S substructure is continuous across the interfaces.
The sulfide crystals can contain up to 10 at% Cu, depending on the sampling locality and independent of the type of bacteria. The Cu content of mackinawite appears to affect the kinetics of its conversion. The complete conversion from mackinawite to greigite took place within 17 days after sample collection in a crystal that had no detectable Cu. On the other hand, another crystal converted only three months after sample collection; this crystal had a few at% Cu in its structure.
Since these microorganisms live where there are relatively high concentrations of H2S, it is likely that the mackinawite to greigite transformation is significantly faster in the living bacteria than in the dead cells that were dried and stored on a TEM grid. The time between two cell divisions can be estimated as anywhere between hours and a few days; if the conversion of mackinawite to greigite is as slow in the natural environment as we observed in the dead cells, the bacterium would end up having only nonmagnetic sulfides that are of no use for its magnetotaxis. It remains to be determined whether the speed of the phase transition is enhanced by biological processes in the natural environment.
Although it is generally known that Cu,Fe diffusion in many sulphides is relatively fast and diffusion coefficients for Cu are similar to those found for mass transport in solid fast ion conductors (Berger and Bucur, 1996), the implications of this to exsolution in systems such as bornite-digenite have not been explicitly explored. We have carried out a series of experiments on synthetic samples along the bornite-digenite join using the time-of-flight (TOF) high-resolution neutron powder diffraction station (HRPD) at the ISIS Spallation Source. The very high resolution of this instrument (d/d = 8 x 10-4 over the entire TOF range) showed conclusively that all samples quenched from the synthesis temperature of 600° Berger R. and Bucur R.V. C were completely exsolved to bornite plus low digenite (5a superlattice) during quenching. Although quenched samples appear homogeneous in reflected light, even at the highest resolutions, annealing for longer times below the solvus temperature results in typical 'basket-weave' exsolution textures. The neutron diffraction patterns of annealed samples were essentially identical to the quenched samples, indicating that annealing merely coarsens the existing phases.
The lattice parameters of the coexisting phases on quenching are consistent with end-member compositions. The sharpness of the neutron diffraction peaks indicates that the exsolution is on a scale of at least several hundred Å. On heating, the lattice parameter changes and the discontinuities associated with phase transitions can be followed in situ. The coalescence of the lattice parameters, and hence the formation of the high temperature solid solution can be readily observed, as can the phase separation on cooling.
The rapid kinetics enables equilibrium to be maintained during heating and cooling in a differential scanning calorimeter. The phase relations as a function of composition could be determined from the thermal anomalies associated with crossing phase boundaries. By correlating the DSC and the neutron diffraction results a new phase diagram for the bornite-digenite join was determined. The new phase diagram differs substantially from that previously published (Kullerud, 1960, Morimoto and Kullerud, 1966). For example, at bulk compositions of 50 and 75 mol percent bornite the solvus lies at 140o and 190°C respectively, compared with published temperatures of over 300°C for both compositions.
Berger R & Bucur RV, Solid State Ionics, 89, 269-278, (1996).
Kullerud G, Carnegie Inst. Washington Yearbook, 59, 114-116, (1960).
Morimoto N & Kullerud G, Zeitschrift für Kristallographie, 123, 235-254, (1966).
Injection of juvenile basaltic magma within dacitic magma chambers is believed to trigger tremendous explosive volcanic eruptions (Nevado del Ruiz, Colombia, 1986 or Pinatubo, Philippines, 1991). The magma-mixing model is supported by the finding of chemically zoned mineral phases within the erupted magma. Here we report results of mineralogical, rockmagnetic and micromagnetic investigations which were performed on a series of samples from the following locations (vulcanos): andestitic pumices from Nevado del Ruiz (Colombia), dacitic pumices from Pinatubo (Philippines) and Haruna (Japan) of holocenic age. Amongst other phases reported in literature, we also found a clear chemical zonation in the Fe-Ti oxide phases, namely in members of the hematite-ilmenite series (Fe2O3 - FeTiO3) revealing intermediate chemical composition (0.5 < X(ilm) < 0.6). These minerals have been identified as the main carriers of the NRM (natural magnetic remanence) of these rocks. Crystals of hematite-ilmenite are tabular with sizes of ca. 100 µm in length. Chemical analyses of hematite-ilmenite by electron microprobe demonstrate the existence of two distinct compositions in the cores and rims, respectively. The cores display compositions of X(ilm) = 0.58 whereas the rims of ca. 10-20 µm thickness are more iron-rich with X(ilm)= 0.52. Cores and rims are connected by a small diffusion zone of ca. 5-10 µm thickness. Minor amounts of Al2O3-, MnTiO3- and MgTiO3-components are always present up to 4 mole percent. Al2O3-, MnTiO3- component do not show any zonation but a clear zonation of geikielite compound (MgTiO3) was observed behaving antipathetic to hematite component. Magnetic domain observations by Bitter-technique and magneto-optical Kerr effect allow the identification of the SR-TRM carrying phases. The rim and diffusion zone reveales the typical magnetic domain structures of a weak-ferromagnetic phase, similar to hematite. Consequently hematite-ilmenite solid solution of the rim composition displays cation disorder. The core shows the typical micro-magnetic pattern of a ferrimagnetic phase similar to cobalt or magnetoplumbite, indicating cation order. The occurrence of disordered and ordered regions in a single crystal implies the existence of an antiphase boundary. The two magnetic phases are of multi-domain character divided by a magnetic domain wall-like boundary located at the contact core/diffusion zone. Magnetic moments of the core are oriented parallel to c-axis whereas in the rim and diffusion zone they are orientated perpendicular to c-axis. Our studies revealed that more or less strong deviations of the NRM/TRM vector from the direction of the ambient magnetic field or even SR-TRM can be attributed to the above mentioned chemically zoned hematite-ilmenite particles and strong magnetic interaction (exchange coupling) between these phases.
We present the results of a kinetic study of cation ordering in the inve