Sequential crystal overproduction triggering Mg-Cr-Ti-V-P-MREE- enrichment in a single-pulse tholeiitic mafic sill in the Central Iberian Zone, Spain

Highlights

• Titanium and P are immobile elements and Ti-P enrichment is decoupled from fluids.

• Immiscibility and Fenner evolutionary trends do not explain the Cr-Ti-P enrichment.

• Bowen trend at low fO₂ plus accumulation explain Cr-Ti-P enrichment.

• Crystal overproduction yields extra latent heat at the arrival of Cpx, Ilm, and Ap.

• Extra latent heat promotes postcumulus processes in bodies with high undercooling.

Abstract

We study the causes of Mg-Cr-Ti-V-P-MREE enrichment in a 134 m thick mafic sill emplaced in a single magma pulse. Whole-rock chemistry indicates that TiO₂ (up to 5.51 wt%) and P₂O₅ (up to 1.60 wt%) enrichment occurred in a closed system. Immiscibility processes during magmatic crystallization and interaction with late fluids seem to have played little or no role for the Ti-P distribution. Liquid lines of descent for anhydrous and hydrous conditions reveal that the melt did not follow a Fenner trend (anhydrous), which may lead to strong Ti or P enrichment, but a Bowen trend (hydrous), during crystallization at low fO₂ (∆FMQ-2) and low water contents (0.5% H₂O). The Mg-Cr-Ti-V-P-MREE-rich samples are cumulates and their composition is offset from the liquid line of descent due to postcumulus compaction and compositional convection. Theoretical formulation of compaction and compositional convection demonstrates that classical postcumulus scenarios of strong undercooling (sill <150 m thick) and high accumulation rates (ca. 1 m/y), using uniform cooling and anhydrous viscosity values, properly explain the observed Ti enrichment, but not the P and Cr enrichment. Latent heat released during overproduction of clinopyroxene, plagioclase, ilmenite, and apatite may result in non-uniform cooling with periods of near-isothermal conditions. Non-uniform cooling may locally reduce the accumulation rate, which temporarily enhances the efficiency of compaction and compositional convection and leads to the development of MgCr, TiV, or P-MREE-rich cumulates. The efficiency of postcumulus processes declines once the latent heat does not compensate heat loss to the country rock, leading to cumulates with high intergranular liquid fractions and without Cr, Ti, and P enrichment. High water contents in the upper level of the sill facilitate postcumulus processes and migration of residual liquids and fluids upwards, leading to the formation of abundant pegmatoid pockets and the development of a sandwich horizon that does not follow the liquid line of descent.

Introduction

The emplacement history of thick (>50 m) mafic bodies can be unraveled from mineral- and whole-rock compositional profiles (Boudreau and Simon, 2007; Latypov and Egorova, 2012; Woodruff et al., 1995). These profiles may have a complex geometry if the magma had been emplaced in multiple pulses, showing characteristic geochemical fingerprints, such as a depletion in transition elements (e.g., Mg, Cr, Ni) within each new magma batch (Jesus et al., 2014; Luan et al., 2014; Maghdour-Mashhour and Shabani, 2017; Zhang et al., 2012; Zieg and Marsh, 2012). Multiple reinjections of new melt typically translate into an incomplete record of the evolutionary history of the magma, limiting the information on those processes that lead to the enrichment of economically interesting elements (e.g., Ti, V, P, REE), especially those related to late crystallizing minerals such as apatite. The economic importance of these mafic bodies has led to considerable research in the past years, with a great advance in the knowledge about the processes producing this enrichment (Charlier et al., 2006, Charlier et al., 2010; Namur and Charlier, 2012; Tollari et al., 2008; Zhang et al., 2012). Thin bodies (<50 m thick) commonly do not show multiple melt injection and do not undergo remarkable chemical and mineralogical changes, resulting in intrusions without mineral enrichments of economic interest. However, there are rare sill-like diabase bodies thicker than 50 m, emplaced in one single magma pulse and with strong enrichment in some elements (López-Moro et al., 2007a). These bodies can be distinguished from multi-pulse sills by the following geochemical features (Latypov, 2003; Maghdour-Mashhour and Shabani, 2017): (i) S-shaped and Z-shaped compositional profiles for transition and incompatible elements, respectively, and (ii) identical composition at the upper and lower chilled margins, which is essentially indistinguishable from the average composition of the mafic body. These sill-like bodies allow studying in detail the cooling and evolution of the magma at the place of emplacement, easily deciphering the liquid line of descent of the melt. These bodies may exhibit significant Ti enrichment that may be higher (up to 5.5% TiO₂, López-Moro et al., 2007a) than in layered mafic intrusions with multiple phases of replenishment (e.g., up to 4% TiO₂ in the Beja Layered Gabbroic Sequence in Portugal, Jesus et al., 2014). Sills emplaced as single magma pulses represent natural laboratories to study the cooling history of a melt, and the origins of Fe, Ti, P, and REE enrichment, although the processes related to the latter are not well understood.

A variety of magmatic and fluid-related models have been put forward to explain the chemical variation across mafic intrusions (e.g. Fenner, 1929; Woodruf et al. 1995; Veksler et al., 2007). Depending on starting composition, water content, and oxygen fugacity fO₂, different magmatic trends develop. Tholeiitic magma may evolve to either silica-rich iron-poor melts (Bowen trend, Bowen, 1928) or iron-rich and silica-poor melts (Fenner trend, Fenner, 1929). This dichotomy has been explained in terms of oxygen fugacity (Yigang et al., 2003): a low oxygen fugacity lowers the onset of crystallization of FeTi oxides leading to prolonged FeTi enrichment and a decrease of silica in the melt (e.g. Osborn, 1959; Toplis and Carroll, 1995). In contrast, a high oxygen fugacity leads to early formation of FeTi oxides, which decreases the FeTi and increases the Si contents early in the melt evolution (Botcharnikov et al., 2008; Jang et al., 2001). Other processes that may result in high Fe and Ti concentrations are: (i) Accumulation processes after FeTi oxides saturation (e.g. Jesus et al., 2014). (ii) Immiscibility of a tholeiitic melt evolving from basalt to rhyolite (e.g. Philpotts, 2008; Veksler et al., 2007). Note, liquid immiscibility may result in compositional variations that resemble a Fenner trend (Jakobsen et al., 2011). (iii) Hydrothermal alteration (e.g. Virginia and Palisade sills in eastern United States; Woodruf et al. 1995; Block et al., 2015), either related to exsolved magmatic fluids that may redistribute metals partitioning into the fluid (e.g. Fe; Martin and Piwinskii, 1969) or aqueous fluid that may scavenge and redistribute metals (e.g. Fe, Ti, Sc; Davidson and Wyllie, 1968; Belkin, 1988; Woodruf et al., 1995; Block et al., 2015).

The largest swarm of tholeiitic diabase bodies of the Iberian Massif occurs in the La Codosera syncline (Fig. 1a). Some of these mafic bodies are up to 400 m thick. One of the bodies, The Negro Villar Sill (NVS) is special as it was emplaced in one single magma pulse (López-Moro et al., 2007a) and is >50 m thick (namely, 135 m). It is the diabase body with the highest contents of TiO₂ (up to 5.51 wt%) and P₂O₅ (up to 1.60 wt%) worldwide (compilation in GEOROC database). The cause of Fe-Ti-P enrichment, however, has so far not been explored. The main aim of this study is to constrain the process that resulted in the exceptional Ti and P enrichment of the NVS rocks. To achieve this, we use extensive modeling of the major and trace element composition of NVS rocks and characterize apatite by cathodoluminescence microscopy and spectroscopy. The liquid line of descent is modeled using the MELTS software (Ghiorso and Sack, 1995) for different fO₂ and water contents, in combination with mass balance to check whether the TiP enrichment is due to immiscibility, fractionation along Fenner or Bowen trends, or accumulation processes. The role of postcumulus processes (compaction and compositional convection) is evaluated based on the interstitial liquid fraction (Tegner et al., 2009) and theoretical models (McKenzie, 1984; Sparks et al., 1985). Thermodynamic constraints allow us to assess the relationship between thermal buffering due to latent heat released by new liquidus phases (Morse, 2011), distribution of the interstitial liquid fraction, and enrichment in Cr-Ti-P-REE in a sill with strong undercooling.