Sulfur isotopic zoning in apatite crystals: A new record of dynamic sulfur behavior in magmas

Abstract

The mobility and geochemical behavior of sulfur in magmas is complex due to its multi-phase (solid, immiscible liquid, gaseous, dissolved ions) and multi-valent (from S²⁻ to S⁶⁺) nature. Sulfur behavior is closely linked with the evolution of oxygen fugacity (fO₂) in magmas; the record of fO₂ evolution is often enigmatic to extract from rock records, particularly for intrusive systems. We apply a novel method of measuring S isotopic ratios in zoned apatite crystals that we interpret as a record of open-system magmatic processes. We interrogate the S concentration and isotopic variations preserved in multiple apatite crystals from single hand specimens from the Cadiz Valley Batholith, CA via electron microprobe and ion microprobe. Isotopic variations in single apatite crystals ranged from 0 to 3.8‰ δ³⁴S and total variation within a single hand sample was 6.1‰ δ³⁴S. High S concentration cores yielded high isotopic ratios while low S concentration rims yielded low isotopic ratios. We discuss a range of possible natural scenarios and favor an explanation of a combination of magma mixing and open-system, ascent-driven degassing under moderately reduced conditions: fO₂ at or below NNO+1, although the synchronous crystallization of apatite and anhydrite is also a viable scenario. Our conclusions have implications for the coupled S and fO₂ evolution of granitic plutons and suggest that in-situ apatite S isotopic measurements could be a powerful new tool for evaluating redox and S systematics in magmatic systems.

Introduction

The behavior of sulfur in magmas is inexorably linked to oxygen fugacity (fO₂) due to the wide valence state variability of S, from S²⁻ to S⁶⁺. Sulfur in magmas can exist as solid sulfate or sulfide phases, immiscible sulfide liquids, as ions in solution, or as a multi-phase gas and can therefore have complex evolution in magmatic rocks. Thus, the evolution of S phases during magmatic processes is challenging to constrain in natural systems. This is particularly true in ancient systems where the gas component is lost. Furthermore, for the intrusive parts of magmatic systems, complex mixing behaviors obscure the relative chronological constraints typically provided by melt inclusion-matrix glass comparisons applied to mafic extrusive rocks (Wallace and Edmonds, 2011, Cottrell and Kelley, 2011, Moussallam et al., 2014). The bulk ratio of plutonic to volcanic materials in continental arcs, for example, is up to 10:1 (White et al., 2006), indicating that the crustal-scale S phase distribution and related magma fO₂ evolution are dominated by intrusive processes.

Dissolved S in magmas exists primarily in two valence states, sulfide (S²⁻) and sulfate (S⁶⁺), although intermediate valence ions may also be present (Métrich et al., 2009, Konecke et al., 2017). The transition from sulfide to sulfate (or vice-versa) as the dominant magmatic S species occurs over the range of fO₂ typically observed in terrestrial magmas, from approximately nickel-nickel oxide buffer values of NNO+1 to NNO+3 (this corresponds to the quartz-fayalite-magnetite buffer [QFM] to QFM+2), suggesting that many magmas likely contain multi-valent S speciation (Carroll and Rutherford, 1988, Jugo et al., 2005, Jugo et al., 2010, Klimm et al., 2012). In addition to its sensitivity to fO₂ changes, S also exhibits a strong isotopic fractionation as a consequence of bonding environments in different valence states of S²⁻ and S⁶⁺ that can be several per mil (‰) at magmatic temperatures, with S²⁻ relatively enriched in ³²S and S⁶⁺ relatively enriched in ³⁴S (Harrison and Thode, 1957, Sakai, 1968). The utility of direct measurement of this isotopic fractionation in volcanic glasses and melt inclusions is currently being explored via laboratory experiments and in natural sample suites (Mandeville et al., 2009, Moussallam et al., 2014, Fiege et al., 2014). Alternatively, in a scenario in which multiple valence states are present in a magma, an isotopic fractionation associated with each species will be inherited by crystallizing minerals that incorporate S as a major element (such as sulfide and sulfate minerals) or as a trace element. Efforts are underway to combine these records in volcanic systems (Beaudry et al., 2015). However, as noted, the non-linear, sometimes cyclical nature of the evolution of intrusive igneous systems (Miller et al., 2007, Cooper and Kent, 2014, Barboni et al., 2016) that crystallize to the solidus makes interpreting melt inclusion data challenging.

Apatite is recognized as a zoned carrier of S (Streck and Dilles, 1998, Tepper and Kuehner, 1999, Parat et al., 2002, Chambefort et al., 2008, Van Hoose et al., 2013, Stock et al., 2015) (Fig. 1), but calculating magma S concentrations from apatite S concentrations is hampered by several complications. Sulfate has a much higher partition coefficient in apatite then sulfide, causing the bulk partition coefficient of S into apatite to be highly fO₂ dependent, in addition to dependencies on temperature (Peng et al., 1997, Parat and Holtz, 2005). This effect is layered on variability of the solubility of the multiple S species at changing pressure, temperature, compositional and fO₂ environments (Carroll and Rutherford, 1988, Wallace and Carmichael, 1992, Ducea et al., 1994, Ducea et al., 1999, Clemente et al., 2004, Liu et al., 2007, Jugo, 2009, Baker and Moretti, 2011, Klimm et al., 2012).

In contrast, this study capitalizes on the partitioning of S into apatite as a high concentration trace element (0.1–1 wt.% SO₃) (Ohmoto, 1986, Peng et al., 1997, Parat and Holtz, 2005), to make the first measurements of ³²S and ³⁴S isotopes at high precision in situ via secondary ion mass spectrometry (SIMS) (Fig. 1). Exploration of S concentration in concert with isotopic ratios yields unique information about the S behavior variations and the processes that drive them. This record has the added advantage of good preservation in plutonic rocks, such as the ones utilized in this study, and constraints provided by the relative chronology of crystal zoning.

Samples for this study are from the Cadiz Valley Batholith, a Cretaceous upper crustal California-type batholith (Howard, 2002, Barth et al., 2004). This batholith is representative of a family of intrusive suites, the observation of which has been integral to modern understanding of magmatic processes (Bateman and Chappell, 1979, Kistler et al., 1986, Coleman and Glazner, 1997, Paterson et al., 2011). These magmas were of the ‘cool, wet’ affinity (Deering et al., 2010) thus their generally low temperatures provide the highest likelihood of preserving mineral zoning that was minimally disturbed by diffusion. Indeed, fine-scale S zoning with sharp boundaries as shown by Streck and Dilles (1998) and found in this study suggest that that S substitution in apatite is highly resistant to diffusion (Fig. 1).

The mineral apatite is stable over a wide magma compositional range (Harrison and Watson, 1984, Piccoli and Candela, 2002) and is found in abundance information environments as diverse as lunar basalts and Yellowstone rhyolites (Boyce et al., 2010, Bindeman and Valley, 2001). The type of record documented here should be present in any system that is sufficiently oxidized to contain sulfate and that crystallizes apatite.