The Rustenburg Layered Suite formed as a stack of mush with transient magma chambers

Abstract

The Rustenburg Layered Suite of the Bushveld Complex of South Africa is a vast layered accumulation of mafic and ultramafic rocks. It has long been regarded as a textbook result of fractional crystallization from a melt-dominated magma chamber. Here, we show that most units of the Rustenburg Layered Suite can be derived with thermodynamic models of crustal assimilation by komatiitic magma to form magmatic mushes without requiring the existence of a magma chamber. Ultramafic and mafic cumulate layers below the Upper and Upper Main Zone represent multiple crystal slurries produced by assimilation-batch crystallization in the upper and middle crust, whereas the chilled marginal rocks represent complementary supernatant liquids. Only the uppermost third formed via lower-crustal assimilation–fractional crystallization and evolved by fractional crystallization within a melt-rich pocket. Layered intrusions need not form in open magma chambers. Mineral deposits hitherto attributed to magma chamber processes might form in smaller intrusions of any geometric form, from mushy systems entirely lacking melt-dominated magma chambers.

Magnetite layer formation in the Bushveld Complex of South Africa

“…where m is the melt viscosity (~160 Pa•s) 3,51 , D is the effective chemical diffusivity in the melt, k is the permeability of crystal mush, g is the gravitational acceleration, and ∆´ is the density difference between the interstitial melt and overlaying magma. The diffusion coefficient of Cr in melt is estimated as 3.2×10 -12 m •s - 1 47 , and adopted as the D here.…”

“…Here we assume that the MML is crystallized from a thick magma sheet, and that magnetite is the sole crystalline phase during the formation of the crystal mush of the MML (Methods). In Figure 1 we show models of fractional crystallization of magnetite from a homogeneous body of melt containing 57 ppm Cr 3,25 , with fO2 between zero and one log unit below the quartz-fayalite-magnetite oxygen buffer (i.e., -1 < QFM < 0) 20,26-27 using a magnetite-melt partition coefficient of 200 at 1150 °C 14,24 . If the initial thickness of the overlying magma is set to 120 m as shown (Table 1), on the order of the vertical distance to the next magnetite layer up section 17,20,28 , then the exponential decrease in Cr between 10 and 40 cm is reproduced by the model but in the upper portions of all of the profiles the model predicts much greater depletion of Cr than is observed.…”

“…The sizes of the primary grains (~0.5-2 mm) are similar with that of disseminated magnetite in associated cumulate rocks, but far less than the crystals in silicate-free regions (~5-20 mm) 13 , indicating extensive overgrowth, compaction, and annealing to form the adcumulates. If it formed by a process of fractional crystallization as a loose pile of settled crystals [3][4] or as a loose arrangement of chains of magnetite crystals nucleated against others in situ 16,[18][19] and accumulated faster than the rate of compaction, the initial magnetite cumulate may have had a high porosity (~0.52) 31 . Compaction is driven by the large density contrast between solid (~4955 kg•m -3 ) and liquid (~2683 kg•m -3 ) (Methods).…”

“…Because it is impossible to observe directly the processes that produce layered plutonic mafic rocks, the nature of these processes must be inferred from the record of composition, mineral mode, and texture of the final products of emplacement, crystallization, and cooling of the magma 1 . Even the most basic of questions, such as whether or not crystals form within the melt and settle to the floor [2][3] or are nucleated and grow in situ (i.e., by heterogeneous nucleation in direct physical contact with the solid base) [4][5] , remain open to controversy after a century of dedicated research effort. Progress will be difficult without resolution of such fundamental questions.…”

“…The Upper Zone of the Bushveld Complex crystallized from an iron-rich basaltic andesite magma at the top of the Rustenburg Layered Suite in what appears to have been a meltdominated sill hundreds of km wide and at least several hundred m deep . The evolution of this magma and its cumulates can be modeled broadly as the product of a protracted process of fractional crystallization at the scale of the entire Upper Zone 3,17,20 , however the Upper Zone contains at least 26 magnetitite layers that are ~0.1-10 m thick and concordant with the igneous layering. The formation of these magnetite layers cannot be accommodated by a simple process of fractional crystallization from a single homogeneous body of liquid the size of the entire Upper Zone.…”