Ichiro Sunagawa

A model for mixing basaltic and dacitic magmas as deduced from experimental data

It has been demonstrated experimentally that basaltic and dacitic magmas can be easily mixed to form both banded dacite and homogeneous andesite in less than a few hours. The presence of phenocrysts larger than 0.5 mm increased considerably the efficiency of mixing. Flow patterns in the experimental system were visualized using Pt spheres, which indicated that convection occurs in basalt melt, but not in dacite melt. The Reynolds numbers of the basaltic and dacitic melts in the experimental system were calculated to be about 10−3 and 10−6, respectively. Mixing proceeds initially by mechanical mixing of the two magmas in a large scale, but later by coupling interfacial convection and mutual diffusion. Thus, depending on the depth where vesiculation and following disruption of the magma occurs, banded pumice, homogeneous pumice and homogeneous andesite lava are erupted. The observed textures of mixed rocks of Plinian type eruption and the limiting occurrence of banded pumice are satisfactorily accounted for on this model.

Growth and morphology of diamond crystals under stable and metastable contitions

Abstract The growth of diamonds under stable and metastable conditions are discussed in the light of an analysis of the morphological characteristics of natural diamonds, and of synthetic diamonds grown under stable and labile conditions.

Effect of stirring on crystallization kinetics of basalt: texture and element partitioning

Isothermal crystallization experiments on basalt have been carried out using an infrared heating furnace to investigate the effect of stirring. When stirring was not applied (static experiment), the results agreed well with previous experiments. But when stirring was applied and a flow of Reynolds number=10−3∼−4 was present (dynamic experiment), considerably different results were obtained, especially in respect to the nucleation rate and the morphology of crystals. At ΔT=25° C essentially similar results were obtained on the nucleation rates and morphologies of crystals in both static and dynamic experiments. However, at supercoolings larger than 45° C, nucleation density increased drastically in dynamic experiments reaching up to ten times as large as that in static experiments. Crystals of plagioclase and clinopyroxene were small and adapted acicular morphology regardless of ΔT in dynamic experiments, and hyalopilitic textures were formed. A TTT-diagram shows that the nucleation incubation time is shorter in dynamic experiments than in static experiments. No compositional difference in major elements was found in plagioclase and clinopyroxene produced in both static and dynamic experiments. However, minor element concentrations, e.g., Mg in plagioclase and Ti, Al in clinopyroxene, were found to increase with both ΔT and flow velocity. All these results imply that although chemical diffusion in the melts did not play an important role in the dynamic experiments, interface kinetics were important. It is suggested that hyalopilitic texture commonly seen in natural basalt is mainly due to flow in magma.

Observations of the influence of stress fields on the shape of growth and dissolution spirals

Abstract Examples are presented which show the influence of stress fields around screw dislocations upon the shape of growth and dissolution spirals. When a growth spiral originates from an independent single screw dislocation with a large Burgers vector, the spiral step shows a change of curvature from negative to positive as it escapes from the centre of the spiral. The degree of change of curvature varies depending on the type of stress field. In case of a strong stress field a stable hollow core is formed in the spiral centre, whose diameter depends on the stress field. When a composite spiral is formed from a group of screw dislocations, a depressed area may occur in the spiral centre due to stress. In the case of dissolution spirals, although a hollow core appears, a change of curvature is not observed. The relation between the hollow cores thus formed and the outcrops of the hollow tubes in crystals observed in SiC or YIG crystals is also discussed.

Twin lamellae as possible self-perpetuating step sources

Abstract Introduction of a (111) twin lamella by a series of parallel stacking faults in a fcc crystal produces two types of twinned area on the smooth (111), (111) and (111) growth surfaces. One type has an atomic configuration corresponding to a rough {100} surface, and thus can act as a self-perpetuating step source. Cooperation of the rough twinned areas and the re-entrant corners will produce a plate-like crystal.

Activities of spiral growth hillocks on the (111) faces of barium nitrate crystals growing in an aqueous solution

In situ measurements of normal growth rates R, slopes p, and step advancing rates v = R / p of the spiral growth hillocks generated from two types of dislocations, screw (b = [111]) and mixed (b = 〈110〉), were made on the (111) faces of Ba(NO3)2 crystals growing from an aqueous solution at different flow velocity, u. These two types of dislocations were identified also in situ by stress birefringence microscopy. It was found that: (1) The growth hillocks originated from the screw dislocations grow faster than those from the mixed dislocations, and the tendency is more pronounced as u increases (Rs/Rm = 1.3 when u = 40 cm/s). (2) As u increases, the step advancing rates become equal for the two types, in spite of the difference of step heights (hs/hm = 1.5). (3) The growth hillocks generated from the screw dislocations are steeper under any growth conditions than those from the mixed dislocations. It was also demonstrated that the dislocations outcropped near the periphery of a growing face were twice more active as growth centres than those at the centre of the face, due to the higher supersaturation along the periphery of the face. It is concluded that the direct integration model explains the result on the step movement better than the surface diffusion model. It is also concluded that the increase of the interstep spacing and therefore the edge free energy of steps are not proportional to the increase of the step height.

The influence of stress on spiral growth

Abstract A differential equation and boundary conditions describing rotating spirals of stationary shape are presented. If the dislocation stress field is strong enough to allow the formation of hollow dislocation cores (radius: r hc ) both the analytical approximation as the phase plane analysis of the differential equation indicate the occurrence of three kinds of spiral: (i) global spirals, running from the dislocation centre ( r =0) to infinity, (ii) inner spirals, running from r =0 to r hc , and (iii) outer spirals running from r = r hc to r = ∞. If the hollow core is kinetically and thermodynamically possible, only types (ii) and (iii) occur. From the solution of the differential equation, the angular velocity ω 1 , and thus the spiral growth or dissolution rate, is obtained numerically. The ω 1 values indicate the sudden occurrence of steep and macroscopic etch pits below a certain critical undersaturation. Also it is found that stress opposes growth slightly but favours dissolution considerably. Starting at high undersaturation, the shape of a spiral at decreasing undersaturation is as follows: closely spaced dissolution spiral, hollow core inner and outer dissolution spiral, hollow core inner and outer growth spiral, growth spiral with depressed centre en finally widely spaced archimedean-type growth spiral. Some observations of the above mentioned spiral types are shown.

Laboratory growth of sector zoned clinopyroxenes in the system CaMgSi2O6-CaTiAl2O6

Sector zoning has been experimentally reproduced in CaMgSi2O6-CaTiAl2O6 clinopyroxene crystals by isothermal crystallization using seed crystals. Element partitioning in different growth sectors and between the core and rim portions in single crystals was analysed in relation to growth rate R and degree of supercooling Δ T. The TiO2 and Al2O3 contents increase with increase in R and Δ T, but when they are compared between different sectors in a single crystal grown at the same Δ T, they correlate negatively with R. The order of faces in respect of contents of TiO2 and Al2O3 is (100)>(110)≳(010)≳(111) at ΔT= 13° C and 18° C but changes to (110)>(100)>(010)>(111) at Δ T= 25° C. The growth mechanism is concluded to be controlled by interface kinetics at ΔT= 13–25° C for all these faces, while at Δ T=45° C this relation holds for (100) and (010) faces, but not for (110) and (111), based on the growth rate versus supercooling relation and surface microtopographic observations. The interface kinetics play the essential role in the formation of sector zoning, when the layer growth mechanism takes place.

Stacking faults as self-perpetuating step sources

Abstract Atomic configurations of sub-steps created on a (111) growth surface of an fcc crystal by stacking faults with fault vectors 1 6 〈112〉 and 1 3 〈111〉 show that they can act as self-perpetuating step sources. Growth kinetics provided by such sub-steps have been analyzed, and it has been shown that the 2D heterogeneous nucleation barrier at the sub-steps is always smaller than that of conventional 2D nucleation. It has also been shown that the 2D heterogeneous mono-nuclear and birth-and-spread nucleation along the sub-steps should provide a growth rate exceeding the one provided by conventional 2D nucleation at low and high supersaturations.

Role of buoyancy driven convection in aqueous solution growth; A case study of Ba(NO3)2 crystal

Abstract Using Schlieren technique and Mach-Zehnder interferometry, the diffusion boundary layer and buoyancy driven convection around a growing barium nitrate crystal from the aqueous solution were visualized and their effect upon the growth kinetics of crystals was investigated in relation to the bulk supersaturation. It was demonstrated that buoyancy driven convection plumes behaved differently depending on the bulk supersaturation, and that their behavior gave a definitive effect upon the growth rates.