Effects of open and closed system oxidation on texture and magnetic response of remelted basaltic glass

Abstract

[1] As part of an experimental and observational study of the magnetic response of submarine basaltic glass (SBG), we have examined, using ion backscattering spectrometry (RBS), transmission and scanning electron microscopy, energy dispersive X-ray spectrometry, and surface X-ray diffraction, the textures wrought by the controlled, open and closed system oxidation of glasses prepared by the controlled environment remelting and quenching of natural SBG. Initial compositions with ∼9 wt % FeO* were melted at 1430°C with the oxygen fugacity buffered at fayalite-magnetite-quartz; melts were cooled at a rate of 200°C min −1 near the glass transition (T g = 680°C). In open system experiments, where chemical exchange is allowed to occur with the surrounding atmosphere, polished pieces of glass were reheated to temperatures both below and above T g for times 1-5000 h; undercooled melts were oxidized at 900°C and 1200°C for 18 and 20 h, respectively. RBS demonstrates unequivocally that the dynamics of open system oxidation involves the outward motion of network-modifying cations. Oxidation results in formation of a Fe-, Ca-, and Mg-enriched surface layer that consists in part of Ti-free nanometer-scale ferrites; a divalentcation-depleted layer is observed at depths >1 mm. Specimens annealed/oxidized above T g have magnetizations elevated by 1-2 orders of magnitude relative to the as-quenched material; this does not appear to be related to the surface oxidation. Quenched glass (closed system, i.e., no chemical exchange between sample and atmosphere) exhibits very fine scale chemical heterogeneities that coarsen with time under an electron beam; this metastable amorphous immiscibility is the potential source for the nucleation of ferrites with a wide range of Ti contents, ferrites not anticipated from an equilibrium analysis of the bulk basalt composition.

Timing of magnetite formation in basaltic glass: Insights from synthetic analogs and relevance for geomagnetic paleointensity analyses

“…In quickly cooled samples, no oxides are visible and a continuous spherulitic zone is absent, although small, discontinuous spherulitic areas are associated with the ends of the plagioclase crystals (Figure 2b). At a ∼30 nm scale, glass generated in a companion experiment [ Burgess et al , 2010] is completely uniform away from the crucible walls. However, in high‐magnification TEM amorphous immiscibility and (very sparse) nanocrystals are present [ Burgess et al , 2010, Figure 2].…”

“…At a ∼30 nm scale, glass generated in a companion experiment [ Burgess et al , 2010] is completely uniform away from the crucible walls. However, in high‐magnification TEM amorphous immiscibility and (very sparse) nanocrystals are present [ Burgess et al , 2010, Figure 2]. Diffraction patterns are not indexable, but indicate a cubic phase.…”

“…In contrast to speculations by Heller et al [2002], it is clear that the low‐Ti magnetite is a primary, not secondary, phase. Results from a companion study [ Burgess et al , 2010] further demonstrate that the magnetic properties of this phase do not significantly alter with reheating (in air) at temperatures below the glass transition temperature (T g ). Although indirect, the data strongly suggest that the low‐Ti phase forms predominantly at T > T g (680°C), which is well above the magnetite Curie temperature (580°C).…”

“…We explore the effects of temperature, melt composition, oxygen fugacity ( f O 2 ), and cooling rate on the magnetic mineral assemblage in rapidly cooled basaltic melts. A companion study [ Burgess et al , 2010] additionally examines the possible role of posteruption oxidation in modification of the quenched glass. For comparison, we examine zero age, young (10s of kyr) and much older (up to ∼100 Myr) natural samples.…”

Contribution of multidomain titanomagnetite to the intensity and stability of Mars crustal magnetic anomalies

“…Both M‐type and T‐type samples contain devitrification textures within the glass, with the amount of glass devitrification increasing with anneal time (Figure ). We speculate that stable‐single‐domain (SSD) iron oxides below the resolution of the electron microprobe are forming during devitrification, leading to more PSD‐like hysteresis loops, enhanced TRM acquisition, and harder AF demagnetization spectra in our annealed samples, which has also been observed in natural and synthetic basaltic glasses [ Burgess et al ., ; Bowles et al ., ]. The high amount of glass in these samples is a function of quenching the pre‐anneal starting materials.…”