AbstractThe Izu–Bonin–Mariana volcanic arc is situated at a convergent plate margin where subduction initiation triggered the formation of MORB-like forearc basalts as a result of decompression melting and near-trench spreading. International Ocean Discovery Program (IODP) Expedition 352 recovered samples within the forearc basalt stratigraphy that contained unusual macroscopic globular textures hosted in andesitic glass (Unit 6, Hole 1440B). It is unclear how these andesites, which are unique in a stratigraphic sequence dominated by forearc basalts, and the globular textures therein may have formed. Here, we present detailed textural evidence, major and trace element analysis, as well as B and Sr isotope compositions, to investigate the genesis of these globular andesites. Samples consist of $$\hbox {K}_2\hbox {O}$$
K
2
O
-rich basaltic globules set in a glassy groundmass of andesitic composition. Between these two textural domains a likely hydrated interface of devitrified glass occurs, which, based on textural evidence, seems to be genetically linked to the formation of the globules. The andesitic groundmass is Cl rich (ca. $$3000\, \mu \hbox {g/g}$$
3000
μ
g/g
), whereas globules and the interface are Cl poor (ca. $$300\, \mu \hbox {g/g}$$
300
μ
g/g
). Concentrations of fluid-mobile trace elements also appear to be fractionated in that globules and show enrichments in B, K, Rb, Cs, and Tl, but not in Ba and W relative to the andesitic groundmass, whereas the interface shows depletions in the latter, but is enriched in the former. Interestingly, globules and andesitic groundmass have identical Sr isotopic composition within analytical uncertainty ($$^{87}\hbox {Sr}/^{86}\hbox {Sr}$$
87
Sr
/
86
Sr
of $$0.70580 \pm 10$$
0.70580
±
10
), indicating that they likely formed from the same source. However, globules show high $$\delta ^{11}$$
δ
11
B (ca. + 7$$\permille$$
‱
), whereas their host andesites are isotopically lighter (ca. – 1 $$\permille$$
‱
), potentially indicating that whatever process led to their formation either introduced heavier B isotopes to the globules, or induced stable isotope fractionation of B between globules and their groundmass. Based on the bulk of the textural information and geochemical data obtained from these samples, we conclude that these andesites likely formed as a result of the assimilation of shallowly altered oceanic crust (AOC) during forearc basaltic magmatism. Assimilation likely introduced radiogenic Sr, as well as heavier B isotopes to comparatively unradiogenic and low $$\delta ^{11}\hbox {B}$$
δ
11
B
forearc basalt parental magmas (average $$^{87}\hbox {Sr}/^{86}\hbox {Sr}$$
87
Sr
/
86
Sr
of 0.703284). Moreover, the globular textures are consistent with their formation being the result of fluid-melt immiscibility that was potentially induced by the rapid release of water from assimilated AOC whose escape likely formed the interface. If the globular textures present in these samples are indeed the result of fluid-melt immiscibility, then this process led to significant trace element and stable isotope fractionation. The textures and chemical compositions of the globules highlight the need for future experimental studies aimed at investigating the exsolution process with respect to potential trace element and isotopic fractionation in arc magmas that have perhaps not been previously considered.
Rare Earth, Trace Element and Stable Isotope Fractionation of Carbonatites at Kruidfontein, Transvaal
