Sodic amphiboles and pyroxenes from fenites in East Africa

During alkali metasomatism of the country-rock associated with ijolite-carbonatite complexes the development of sodic amphibole and/or pyroxene is characteristic. In this paper, some new chemical analyses of these minerals, together with published analyses from fenites of Kenya, Uganda and Tanzania, include those of co-existing pairs of amphibole and pyroxene. The common amphiboles of the fenites are magnesioarfvedsonites with 100 Mg: Mg+Fe+Mn ranging from 67 to 36. They co-exist with aegirines having 0.75 to 0.89 ions Fe⁺³. Most of these minerals are poor in Ca; co-existing pairs tend to show corresponding increases in Ca and in Fe⁺². In the syenitic fenites of Tororo and Budeda, considered to have formed at higher temperatures, the stable mineral is aegirine-augite. New analyses of richterite, magnesioarfvedsonite and aegirine from carbonate-rich rocks are also presented, and the relation between fenites and carbonatites is discussed.     

The distribution of certain elements among coexisting calcic pyroxenes, calcic amphiboles and biotites in skarns

Skarns composed predominantly of Ca pyroxene, Ca amphibole, and biotite occur locally in the Grenville metamorphic terrain of the Canadian pre-Cambrian Shield. The enclosing rocks are Grenville gneisses and limestones, gabbroic rocks, veined gneisses, and granitic rocks. The metamorphic grade of the skarns is identical to that of the enclosing rocks, namely the upper region of the amphibolite facies.

Twelve skarn specimens were selected for chemical study. An investigation is made of the concentrations of Al, Fe, Mg, Mn, Ti, Ca, Na, K, Ba, V, Cr, Zr, Y, and Sc in eleven Ca pyroxenes, ten Ca amphiboles, and eleven biotites.

Consideration is given to the distribution of elements among coexisting minerals. Regular relationships appear when

1. (1) the concentration of an element in a mineral is related to the concentration of the element in a coexisting mineral by a linear or non-linear function

2. (2) the linear or non-linear function is itself a function of the concentration of another element in one or both of the coexisting minerals. These relationships indicate that, for the most part, minerals within each skarn specimen closely approached a state of chemical equilibrium.

A preliminary attempt is made to explain the distribution relationships in terms of crystal chemistry.     

The crystallization of pyroxenes and amphiboles in some alkaline rocks and the presence of a pyroxene compositional gap

Ca- and Na-rich pyroxene-amphibole compositions and textures from a range of felsic alkaline rocks have been studied in detail. The data indicate that in a single sample, when amphibole crystallizes, a gap is observed between Ca- and Na-rich pyroxene compositions. This break in composition is analogous to Aoki's (1964) immiscibility gap between Ca-and Na-rich pyroxenes and can be overlooked when considering pyroxene compositions from a suite of rocks. The role of volatiles in governing the stability and composition of amphiboles is discussed. The presence of late crystallizing Na-rich pyroxene is related to the development of peralkinity in the late-stage melts.On extrusion, many alkaline rocks lose their volatiles and amphibole is absent. In these rocks complete zoning from Ca-rich to Na-rich pyroxene compositions are observed within the one sample.     

Chemical types of amphiboles from the Tyrnyauz deposit

Based on the results of microprobe analyses, six groups of calcium amphiboles were distinguished for the Tyrnyauz deposit. Group I (mainly edenite) includes amphiboles of hornfels and periskarn rocks. Groups II-VI (actinolite-hastingsite series) are related to post-skarn metasomatic rocks of the productive stage after hornfels (II), the tonalite and plagiogranite complex (III), ultrabasic rocks (IV), pyroxene-plagioclase periskarn rocks (V), and skarn (VI). Group IV is represented by actinolite; group VI, by hastingsite; groups II, III, and V, by hornblende varieties.     

Chemical etching of dislocations in forsterite

Chemical etching of pure melt-grown forsterite crystals is reported here for the first time. Cleaved forsterite crystals of known purity, and polished sections of the same crystals have been successfully etched revealing dislocations, subgrain boundaries, inclusions and growth imperfections.     

Chemical etching of deformation sub-structures in quartz

Chemical etching of dislocations has been studied in natural and synthetic quartz single crystals, in deformed synthetic quartz and in naturally and experimentally deformed quartzites. The ability of different etchants to produce polished or preferentially etched surfaces on quartz is described. Dislocation etching was achieved on all crystal planes examined by using a saturated solution of ammonium bifluoride as the etchant. Appropriate etching times were determined for etching quartzites for grain size, subgrain boundaries, deformation lamellae, dislocations and twins. Growth and polished surfaces of synthetic single crystal quartz were similarly etched and dislocation etch pits, characteristic of various orientations were found. The use of ammonium bifluoride proved to be expecially advantageous for the basal plane, producing a polished surface with etch pits, suitable for dislocation etch pit counting. “Double” etch pits have been found on Dauphiné twin boundaries on the basal plane and the first order prism, using this etchant. Slip lines and deformation bands were suitably etched on deformed synthetic crystal surfaces for identification of the slip planes. Other acidic etchants have been explored and their application to the study of deformation structures in quartz crystals is discussed.     

Compositional relations among pyroxenes,amphiboles and other mafic phases in the Oslo Region plutonic rocks: Publication No. 115 in the Norwegian Geotraverse project

Mafic minerals in samples from older and younger intrusions in the Permian Oslo rift were analysed to determine temperatures and oxygen fugacities of crystallization and possible genetic relations between different rock types. Applications of various geothermometers and oxygen barometers to plutonic rocks have been assessed in connection with this work.Pyroxenes in the peralkaline rocks define a continuous trend from augites to pure aegirines. This trend was found to be the result of fractional crystallization of plagioclase (and augite), suggesting that the peralkaline rocks formed from alkaline initial liquids.     

Oxygen diffusion in amphiboles

The kinetics of oxygen isotope self-diffusion in natural samples of hornblende, tremolite, and richterite have been measured. Samples were run under hydrothermal conditions using ¹⁸O enriched water. Profiles of ${}^{18}\text{O}\text{(}{}^{16}\text{O +}{}^{18}\text{O)}$vs depth into the crystal were obtained using an ion microprobe; the depths of sputtered holes were measured using an optical interferometer. At 1000 bars (100 MPa) water pressure, the activation energies (Q) and pre-exponential factors (D₀) for diffusion parallel to c are: D₀(cm²/sec) Q (kcal/gm-atom) T (°C) Hornblende 1⁺²⁰_?1 × 10^?741 ± 6 650–800 Tremolite 2⁺³⁰_?2× 10^?8 39 ± 5 650–800 Richterite 3⁺⁵_?2 × 10^?4 57 ± 2 650–800The diffusion coefficient (D) for hornblende at 800°C and 1000 bars water pressure measured parallel to the c crystallographic direction is at least ten times greater than that parallel to the a or b directions. An increase in water pressure from 200 to 2000 bars increases D by a factor of 2.7 for diffusion parallel to c at 800°C. The D value for hornblende at 800°C is about 0.01 that for quartz and 0.001 that for anorthite. As a result, closure temperatures for oxygen exchange in natural primary amphiboles are significantly higher than for quartz or feldspars. It is unlikely that amphiboles will exchange oxygen isotopes by diffusion under most crustal conditions.     

Thermodynamic analysis of quadrilateral pyroxenes

The thermodynamics of quadrilateral pyroxene solutions have been analysed with the ternary non-convergent site-disorder model developed in Part I. Solution parameters have been refined for this model with a non-linear least squares technique to fit experimentally determined phase equilibria for the assemblages orthopyroxene-clinopyroxene (opx-cpx), pigeonite-augite (pig-aug), and opx-pig-aug. Calculated phase relations agree, within error, with experiments. Predicted cation distributions and enthalpies of solution for opx are also in agreement with measurements. Predicted cation distributions for cpx are more disordered than indicated by most available measurements. Two types of pyroxene thermometers are presented: (a) a single-pyroxene (aug) thermometer, and (b), a two-pyroxene thermometer which approximates the temperature of an observed pig-aug or cpx-opx assemblage as that of its nearest model tieline. Pyroxene pairs from two granulite suites, whose compositions were projected to the ternary system by the method of Lindsley and Andersen (1983), yield temperatures that are 25° C higher by thermometers (a) and (b) than determined from Lindsley's (1983) graphical thermometer. Temperatures calculated for projected compositions of coexisting pig-aug are within 20° C of run temperatures in experiments by Grove and Bryan (1983).     

Clustering of cations in amphiboles

The OH^? stretching frequencies of clino-amphiboles are known to depend on the cations to which the OH^? is co-ordinated, and the intensities of the corresponding infra-red (ir) absorption bands have been used to obtain evidence as to the occupancy of the M1 and M3 sites. However, the possible effects on the method arising from clustering together of like cations, or of the mutual avoidance of like cations (anti-clustering), have not hitherto been explicitly analysed. A model is set up which permits these to be systematically explored. The results are exemplified for occupation of M1 and M3 by (Fe, Mg). Clustering of Fe increases the frequency of both Fe₃ and Mg₃ triads, though not equally, and the effect of anti-clustering is in the opposite sense. The effect on the frequencies of MgFe₂ and Mg₂Fe triads is much more complicated and the sense of the changes depends on both the overall Fe content and on the degree of clustering or anti-clustering.     

Statistical chemistry of calcic amphiboles

Principal components analysis is used to study the chemistry of 639 calcic amphiboles. Eigenvectors representing multiple partial correlation coefficients give various sets of substitutional relationships. The relative significance of each set can be noted by the percent variation of the data it represents. The highest percent variation (36%) is associated with the substitutions $$Si + Mg \rightleftharpoons Al^{IV} + Al^{VI} + Ti + Fe^{3 + } + Fe^{2 + } + Na + K$$ . Other expected substitutions among the ions such as Al^IV + Na ? Si, the positive correlation between Al^IV and Al^VI etc. are shown statistically. The substitution of Al in T ₁ and T ₂ positions imposes an ordering in the M ₁, M ₂ and M ₃ sites. Variability of OH in the amphiboles is found to be significant. There is no definite correlation between OH and Fe³⁺ but OH and Ti are positively correlated. Under certain conditions and provided the concentration of Al^IV does not change significantly, Fe and Mg may be assumed to mix ideally in the amphibole solid solution.     

Thermodynamic study of calcic amphiboles

The paper reports original thermochemical data on six natural amphibole samples of different composition. The data were obtained by high-temperature melt solution calorimetry in a Tian–Calvet microcalorometer and include the enthalpies of formation from elements for actinolite Ca_1.95(Mg_4.4Fe _0.5 ²⁺ Al₀₁)[Si_8.0O₂₂](OH)₂(–12024 ± 13 kJ/mol) and Ca_2.0(Mg_2.9Fe _1.9 ²⁺ Fe _0.2 ³⁺ )[Si_7.8Al_0.2O₂₂](OH)₂, (–11462 ± 18 kJ/mol), and Na_0.1Ca_2.0(Mg_3.2Fe _1.6 ²⁺ Fe _0.2 ³⁺ )[Si_7.7Al_0.3O₂₂](OH)₂ (–11588 ± 14 kJ/mol); for pargasite Na_0.5K_0.5Ca_2.0-(Mg_3.4Fe _1.8 ²⁺ Al_0.8)[Si_6.2Al_1.8O₂₂](OH)₂ (–12316 ± 10 kJ/mol) and Na_0.8K_0.2Ca_2.0(Mg_2.8Fe _1.3 ³⁺ Al_0.9) [Si_6.1Al_1.9O₂₂](OH)₂ (–12 223 ± 9 kJ/mol); and for hastingsite Na_0.3K_0.2Ca_2.0(Mg_0.4Fe _1.3 ²⁺ Fe _0.9 ³⁺ Al_0.2) [Si_6.4Al_1.6O₂₂](OH)₂ (?10909 ± 11 kJ/mol). The standard entropy, enthalpy, and Gibbs free energy of formation are estimated for amphiboles of theoretical composition: end members and intermediate members of the isomorphic series tremolite–ferroactinolite, edenite–ferroedenite, pargasite–ferropargasite, and hastingsite.     

Equilibrium coexistence of three amphiboles

Electron probe and wet chemical analyses of amphibole pairs from the sillimanite zone of central Massachusetts and adjacent New Hampshire indicated that for a particular metamorphic grade there should be a restricted composition range in which three amphiboles can coexist stably. An unequivocal example of such an equilibrium three amphibole rock has been found in the sillimanite-orthoclase zone. It contains a colorless primitive clinoamphibole, space group P2₁/m, optically and chemically like cummingtonite with blue-green hornblende exsolution lamellae on (100) and (¯101) of the host; blue-green hornblende, space group C2/m, with primitive cummingtonite exsolution lamellae on (100) and (¯101) of the host; and pale pinkish tan anthophyllite, space group Pnma, that is free of visible exsolution lamellae but is a submicroscopic intergrowth of two orthorhombic amphiboles. Mutual contacts and coarse, oriented intergrowths of two and three host amphiboles indicate the three grew as an equilibrium assemblage prior to exsolution. Electron probe analyses at mutual three-amphibole contacts showed little variation in the composition of each amphibole. Analyses believed to represent most closely the primary amphibole compositions gave atomic proportions on the basis of 23 oxygens per formula unit as follows: for primitive cummingtonite (Na_0.02Ca_0.21 ^– Mn_0.06Fe²⁺ _2.28Mg_4.12Al_0.28) (Al_0.17Si_7.83), for hornblende (Na_0.35Ca_1.56Mn_0.02Fe_1.71Mg_2.85Al_0.92) (Al_1.37Si_6.63), and for anthophyllite (Na_0.10Ca_0.06Mn_0.06Fe_2.25Mg_4.11Al_0.47) (Al_0.47Si_7.53). The reflections violating C-symmetry, on X-ray single crystal photographs of the primitive cummingtonite, are weak and diffuse, and suggest a partial inversion from a C-centered to a primitive clinoamphibole. Single crystal photographs of the anthophyllite show split reflections indicating it is an intergrowth of about 80% anthophyllite and about 20% gedrite which differ in their b crystallographic dimensions. Split reflections are characteristic of all analyzed orthorhombic amphiboles so far examined from Massachusetts and New Hampshire except the most aluminous gedrites, and the relative intensity of the gedrite reflections is roughly proportional to the degree of Na and Al substitution. Thin sections of a few of these anthophyllite specimens show lamellae parallel to (010) that are just resolved with a high power objective.Publication approved by the Director, U.S. Geological Survey.     

Thermochemistry of some pyroxenes and related compounds

The enthalpies of formation of a number of crystalline silicates from the oxides at 986 K were determined by oxide melt solution calorimetry. The values of ΔH°_{f, 986}, in kcal/mol, are as follows: MgCaSi₂O₆, ? 34.3 ± 0.4; CoCaSi₂O₆, ? 26.7 ± 0.5; NiCaSi₂O₆, ? 27.1 ± 0.5; MnSiO₃, ? 6.3 ± 0.3; Mn₂SiO₄, ? 12.2 ± 0.3. In addition, for MnSiO₃ (rhodonite)→ MnSiO₃ (pyroxmangite), ΔH°₉₈₆ = + 0.06 ± 0.3₃kcal/mol and for MgCaSi₂O₆ (diopside) = MgCaSi₂O₆ (glass), ΔH°₉₈₆ = + 21.0 ± 0.3 kcal/ mol. For hedenbergite, FeCaSi₂O₆, ΔG°₁₃₅₀ = ?25.6 ± 1.5 kcal/mol. In terms of pyroxene phase equilibria and crystal chemistry, our thermochemical data support the generally accepted crystallographic arguments that (a) the C2/c clinopyroxene structure increases in stability with decreasing size of the ion occupying the Ml site in the MCaSi₂O₆ series, and (b) the energy (and enthalpy) differences between orthopyroxene, clinopyroxene, and pyroxenoid structures are generally quite small and often less than 500 cal/mol in magnitude.     

Characteristics of asbestiform and non-asbestiform calcic amphiboles

In terms of morphology there are four major types of calcic amphibole; massive, prismatic, finely acicular and asbestos. Representatives of each of these types have been examined by optical microscopy, X-ray diffraction, scanning and transmission electron microscopy, and electron probe microanalysis. Massive specimens (nephrite) consist of randomly oriented clusters of fine, roughly lath-shaped, sub-microscopic crystals; within each cluster the lath lengths (z) are approximately aligned but neighbouring laths are rotated with respect to one another. Finely acicular specimens (“byssolites”) have well-formed crystals bounded mainly by {110} (100) and (010) faces and characteristically have striations parallel to their lengths. Asbestiform varieties range from finer (flexible) to coarser (more brittle) specimens and many specimens contain a mixture of fine and coarse fibrils. The fibrils in a bundle are aligned parallel to z but are in a range of azimuthal orientations. It is inferred that they are formed by multiple independent nucleation and growth parallel to z rather than through parting or cleavage on {110} planes. (100) defect or twin planes, or on (010) planar defects.

The {110} cleavage in amphiboles is well reported but (100) features are rarely mentioned in the literature. Our observations reveal the importance of (100) as a cleavage or parting as well as the tendency in nephrites, byssolites and asbestos towards a lath-like (parallel to z) morphology with flattening on (100). In the latter varieties therefore, the y-direction is that of second fastest crystal growth, after z.

When subjected to moderate grinding, the comminution of asbestos fibres proceeds more by separation of fibrils and less by fracturing to shorter lengths as compared with prismatic and byssolite specimens. Prolonged grinding does, however, shorten lengths of even the least brittle asbestos.

Transmission electron microscopy revealed extensive sub-grain boundaries and dislocation networks (suggesting a deformation history) in all prismatic and nephrite specimens. Fine multiple (100) twinning was observed in asbestos but not in other varieties. Although chain-width defects [on (010)], with visibility enhanced by beam damage, were most abundant in nephrites and fibrous tremolites, there appears to be no completely consistent relationship between such features and morphological type.

Electron probe analyses showed that specimens that contain more than a very small amount of aluminium do not have asbestiform habit. Asbestos specimens also have lower contents of Mn, Na and K and have formulae closer to the ideal Ca₂(Mg,Fe)₅Si₈O₂₂(OH)₂. Small departures from this in asbestos involve Na in the A site compensated by Na for Ca rather than Al for Si whereas the reverse is true in byssolites. Chemical substitutions in prismatic specimens are much less constrained.

The characteristics of the four morphological sub-groups correlate reasonably well with what is known of their geological environments.     

Crystallochemistry and origin of pyroxenes in komatiites

We present a detailed mineralogical and major- and trace-element study of pyroxenes in two Archean komatiitic flows in Alexo, Canada. The pyroxenes in spinifex-textured lavas commonly are zoned with cores of magnesian pigeonite and rims of augite. Concentrations of incompatible trace elements are low in pigeonite and jump to higher values in the augite mantles, a variation that can be modelled using accepted partition coefficients and assuming crystallization from komatiitic liquids. Crystallization sequences are very different in different parts of both flows. In the flow top, the sequence is olivine followed by augite: deeper in the spinifex sequence, pigeonite crystallizes after olivine, followed by augite; in lower cumulates, orthopyroxene or augite accompany olivine. In spinifex lavas, pigeonite crystallizes sooner than would be predicted on the basis of equilibrium phase relations. We propose that contrasting crystallization sequences depend on the position in the flow and on the conditions of crystal growth. In the flowtop, rapid cooling causes quench crystallization. Deeper in the spinifex layer, constrained growth in a thermal gradient, perhaps augmented by Soret differentiation, accounts for the early crystallization of pigeonite. The cumulus minerals represent a near-equilibrium assemblage. Augites in Al-undepleted Archean komatiites in various localities in Canada and Zimbabwe have high moderate to high Wo contents but their Mg# (Mg/(Mg + Fe) are lower than in augites in komatiites from Barberton, South Africa. We attribute the combination of high Wo and high Mg# in Barberton rocks to the unusually high CaO/Al₂O₃ of these Al-depleted komatiites.     

Multiple-chain and other unusual faults in amphiboles

A combination of high resolution transmission electron microscopy and computer simulation has revealed the existence of multiple-chain and other faults in nephrite jade. Attention is drawn to the subunit cell reorganization involved with many of these faults, which generally appear to be free from severe structural strain.     

Mechanisms of exsolution in sodic pyroxenes

The free energy curves for simple binary solid solutions with limited miscibility or atomic ordering have been combined to predict the phase relations and exsolution mechanisms for a system in which both ordering and exsolution are possible. The nature of the ordering process affects which exsolution mechanisms may be used. If the ordering is second (or higher) order in character then continuous mechanisms predominate and a ‘conditional spinodal’ (Alien and Cahn, 1976) can be described which operates between ordered and disordered end members. For a first order case, the ordered phase can only precipitate a disordered phase by nucleation and growth. Microstructures in omphacites observed by transmission electron microscopy include exsolution lamellae and antiphase domains and the relations between them in selected specimens have been used to interpret the exsolution mechanisms which operated under geological conditions. It appears that most omphacites undergo cation ordering, and then remain homogeneous or exsolve a disordered pyroxene by spinodal decomposition. The predominance of continuous mechanisms has been used to indicate that the C2/c→P2/n transformation may be second (or higher) order in character. A possible phase diagram for jadeite-augite is presented. It is based on the idea that there should be limited miscibility between the disordered end members at low temperatures and that the cation ordering at intermediate compositions (omphacite) is superimposed on a solvus. It is adequate to explain many of the observed microstructures and fits with petrographic evidence of broad two phase fields between impure jadeite and omphacite and between omphacite and sodic augite. The effect of adding acmite is analogous to increasing temperature so that the phase relations for jadeite-acmite-augite can also be predicted.     

Metabasite amphiboles of the Scottish Dalradian

Blue-green hornblendes are observed in metabasite assemblages throughout the chlorite, biotite and garnet zones of the southwest Scottish Highlands. Actinolites are common in more Mg-rich metabasites in these zones. At low grade, hornblendes are relatively edenite-rich, and may sometimes occur together with a more Mg-rich, Al-poor actinolite. Within the garnet zone, hornblendes are pargasitic, showing extensive tschermakite substitution. Textural and chemical evidence do not indicate the presence of any miscibility gap between hornblende and actinolite within the chlorite to garnet zones in the southwest Highlands. The occurrence of hornblende-actinolite pairs in metabasites of the Scottish Dalradian, and perhaps also in other metamorphic terrains, is considered to reflect the incomplete chemical equilibration of lower grade actinolitic amphibole during prograde metamorphism, rather than a miscibility gap. The paucity of amphibole compositions intermediate between hornblende and actinolite in many metamorphic terrains is thought to reflect the rapid but continuous change of stable amphibole compositions in metabasites over a small range of increased metamorphic grade.     

Thermochemistry of jadeite—diopside pyroxenes

The enthalpies of solution of seven synthetic clinopyroxenes on the join CaMg₂Si₂O₆ (diopside)-NaAlSi₂O₆ (jadeite), of two natural low-Fe ordered omphacites near the 1:1 composition, and of a nearly pure natural jadeite, were measured in molten Pb₂B₂O₅ at 970 K. Enthalpies of solution of the natural omphacites experimentally disordered at 1350°C and 30 kbar were also measured.The synthetic clinopyroxenes have positive excess enthalpies of mixing, which can be expressed by a symmetrical function ΔH_mix = W_HX_JdX_Di, with W_H = 7250 ±950 calories. The enthalpy of disordering of the two natural omphacites averages 1.8 kcal, which is nearly the same as the excess enthalpy of mixing of a 1:1 disordered pyroxene.The interaction parameter, W_H, can be shown to be essentially equivalent to ΔG° of the reciprocal reaction: CaMgSi₂O₆ + NaAlSi₂O₆ = CaAlSi₂O⁺₆ + NaMgSi₂O^?₆ M-site cation distribution data of natural omphacites heat-treated at 1000°C (Aldridgeet al., 1978) lead to ΔG° = 7200 cal for the above reaction, in good agreement with the calorimetric W_H. The reciprocal solution theory with ΔG° = 7200 cal predicts closely the activities of NaAlSi₂OP₆ in jadeite-diopsides found from phase equilibrium measurements at 600°C (Holland, 1979a) and is nearly equivalent to an entropically ideal two site mixing model with a (fictive) W_H of 5800 cal.Jadeite-diopside solid solutions near the 1:1 composition at temperatures of 1000–1500 K are ‘pseudoideal’; that is, they have nearly the free energies of ideal one-site mixtures (Ganguly, 1973). If the order-disorder transition is nearly first-order at about 1000 K, as suggested by Fleetet al. (1978), the pseudo-ideality holds also for ordered omphacites at least somewhat below 1000 K.