August 2017 Issue Table of Contents
Subduction zones are geologically dramatic features, with much of the drama being driven by the movement of water. The “light and lively” nature of boron, coupled with its wide variations in isotopic composition shown by the different geo-players in this drama, make it an ideal tracer for the role and movement of water during subduction. The utility of boron ranges from monitoring how the fluids that are expelled from the accretionary prism influence seawater chemistry, to the subduction of crustal material deep into the mantle and its later recycling in ocean island basalts.
Keywords: subduction zones, water, serpentinite, phengite, crustal recycling
At subduction zones, oceanic crust and some of its sedimentary overburden descend into the mantle (Fig. 1). Water is carried down with the slab in sediment pore water and in hydrous minerals in the sediments and crust. As pressures and temperatures increase in the descending slab, the pore water rapidly returns to the oceans, while some mineral-associated water will be carried deeper. Eventually, the pressure and temperature will increase sufficiently to cause the hydrous mineral phases to release almost all of their remaining water, thus hydrating the overlying mantle wedge and forming serpentinites. These serpentinites may then rise as diapirs in the forearc or be dragged deeper into the subduction zone. At greater depths and temperatures, further release of hydrous fluids initiates partial melting in the overlying mantle wedge. This magma rises through the subarc mantle to form volcanoes and plutonic rocks. After it has lost almost all its water and other volatile components, the slab is subducted deep into the Earth’s mantle. It may then mingle with the deep mantle and be incorporated into the melts that form intraplate ocean island basalts (OIBs).
There are several attributes of boron that make it a sensitive tracer of these processes, especially in tracing the role of fluids. Compared to the mantle, boron is enriched in continental crust by several orders of magnitude. This enrichment is inherited by the terrestrial sediments that lie on the subducted crust. Boron is also very soluble, so it is one of the few elements to have a higher concentration in seawater (4.5 ppm) than in the mantle (<0.1 ppm). Thus, the pore waters that make up ~50% of the volume of the sediments on the subducted crust carry a substantial boron inventory, with some of this boron adsorbed on clay minerals. Prior to subduction, most oceanic crust has experienced millions of years of interaction with seawater circulating through the upper crust. During high temperature water–rock interaction at mid-ocean ridge crests, boron is leached from the rock. As the crust moves away from the ridge crest, the temperature of the circulating seawater drops and the behaviour of boron reverses, with seawater boron taken up into secondary hydrous minerals. The extent of this reaction is such that the concentration of boron rises from values of <1 ppm in fresh oceanic crust to up to 100 ppm in the altered ocean crust that enters the subduction zone.
Boron possesses another property that enhances its utility in this setting – the ratio of its stable isotopes, 10B and 11B, which is expressed as:
where the standard is the NIST SRM 951 boric acid that is used in most studies (e.g. Palmer and Swihart 1996). The d11B values vary greatly between the various components of the subduction zone (Fig. 1). For example, the d11B of seawater is +40‰, as compared to −7‰ in the mantle that is the source of mid-ocean ridge basalt (MORB) (Marschall et al. 2017). These large differences allow the d11B of solid and fluid phases to be used to further constrain boron sources and pathways in subduction zones.
The objective here is to review our understanding of boron cycling in subduction zones, including the nature of boron-hosting phases in the subducted slab and whether their stability under the changing pressure–temperature–chemical conditions may allow crustal boron to be recycled into the mantle. The process is of particular interest because the high boron concentrations and distinct d11B in slab phases relative to the mantle make boron a potentially sensitive tracer of crustal recycling into the deep mantle.
THE BORON INVENTORY OF THE SUBDUCTING SLAB
The major boron reservoir in marine sediments is clay minerals, within which boron is partitioned between a lattice-bound fraction and an adsorbed fraction derived from seawater. The boron concentration in the two fractions varies, but is typically 80–150 ppm in the lattice-bound fraction and 10–30 ppm in the adsorbed fraction. The d11B value of the lattice-bound fraction largely reflects that of the crustal protolith (−5‰ to +5‰), whereas fractionation of seawater boron isotopes during adsorption onto clays yields typical d11B values of +14‰ to +16‰ (Spivack et al. 1987; Ishikawa and Nakamura 1993; Palmer and Swihart 1996).
The boron concentration of fresh MORB is <1 ppm and has a d11B value of −7‰, but when circulating seawater interacts with ocean crust at <100 °C, then the boron is taken up into secondary minerals. Again, there is large variability in the boron content and d11B of altered oceanic crust, but compilations of ocean cores and ophiolite sections give average boron contents of ~5 ppm and d11B values of +3‰ for the upper oceanic crust (Smith et al. 1995). The lower, gabbroic part of the oceanic crust is less altered (Vils et al. 2009), but the upper mantle may be altered to serpentinite by circulating seawater, particularly at slow-spreading mid-ocean ridges. This geological setting leads to typical boron concentrations of 20–90 ppm and d11B values of between +10‰ and +15‰ (Boschi et al. 2008).
Subduction Zone Processes
The Forearc Zone
The initial stages of subduction commonly involve the formation of an accretionary wedge of sediment and portions of oceanic crust being scraped off the subducting plate (Fig. 1B). At shallow depths, fluids are expelled along fractures and faults in the wedge (Martin et al. 1996). Boron concentrations in these fluids can reach >10 times seawater levels, with d11B values similar to the adsorbed fraction of boron in clays (You et al. 1995).
Not all fluids in the forearc return boron directly to the oceans. Consider serpentinite: it is formed during the hydration of the overlying mantle wedge by water driven off the subducting slab. Because serpentinite is buoyant, it may ascend into the accretionary prism and the forearc mélange to form boron-rich serpentinite seamounts and muds (Benton et al. 2001; Savov et al. 2007). However, some serpentinite generated in the forearc is dragged deeper into the mantle (Straub and Layne 2002). The d11B of serpentinite minerals and the fluids from which they were derived are similar to the exchangeable component of sediments, but other chemical and isotopic signatures indicate that this adsorbed fraction cannot be the major boron source. Rather, much of the boron within serpentinites that have formed in the subduction zone (as distinct from serpentinites formed prior to slab subduction) is derived from boron that was structurally bound in sediments and in altered oceanic crust. This structurally bound boron is then extracted by the breakdown of hydrous minerals during increasing pressures and temperatures (Benton et al. 2001; Pabst et al. 2012). The d11B of fluids released from the altered oceanic crust shows a progressive change during subduction (Peacock and Hervig 1999) – a decrease from +25‰ at ~100 °C to +5‰ at 500 °C, with the d11B of the slab restite showing a corresponding decrease from −3‰ to −10‰ (Pabst et al. 2012).
The Arc Magma Production Zone
Arc volcanic rocks contain elevated boron levels and distinct d11B values relative to the mantle. These values reflect boron input from the subducted slab (Fig. 1C). For example, in the Izu arc (Western Pacific Ocean), the amount and isotopic composition of slab-derived boron varies according to the distance of the volcanic centre from the subduction front and the depth to the slab (Ishikawa and Nakamura 1994). Thus, volcanoes closest to the slab have higher d11B values and B/Nb ratios (+7‰ and ~40, respectively) than those that sit further back from the forearc and where the depth to the slab is greater (+1‰ and ~2, respectively).
Similar patterns are observed in other arcs, with a common feature that the d11B of arc rocks extend to values (up to +18‰) that are too high to be derived from quantitative extraction of boron from sediments and/or altered oceanic crust. Instead, the likely source of boron with elevated d11B is from serpentinite dehydration. There are two possible sources of serpentinite-derived boron. First, serpentinite that formed in the subduction zone and is then dragged down on top of the subducting slab; second, serpentinite that formed prior to subduction and that is located deeper in the subducting lithosphere (Fig. 1B). Three-dimensional modelling suggests that even in NE Japan, where 130 Ma oceanic crust is being subducted, the temperature at the top of the slab beneath the volcanic front is likely ~800 oC (Morishige and van Keken 2014). Because antigorite undergoes virtually complete dehydration and release of boron in water-rich melts and/or silica-rich aqueous fluids at ~700 °C (Harvey et al. 2014), it is, therefore, likely that any serpentinite-derived boron within arc rocks is derived from the base of the subducted lithosphere (Fig. 1C) where temperatures may be as low as 475 °C beneath the volcanic front (Stern 2002). Importantly, boron released from the slab into the melt generation zone is quantitatively transferred to the arc rocks because boron is highly incompatible during mantle melting.
The elevated d11B in arc rocks suggests that dehydration fluids are a major boron source, but other isotopes and trace elements (e.g. 10Be and Th) indicate that melting sediments supply an increasing portion of incompatible elements as the slab is subducted deeper. Indeed, the lowest d11B values in the Izu arc (+1‰) lie furthest from the forearc and are more similar to sediments. They also have B/Nb ratios (~2.5) that are closer to sediments (~1) than aqueous fluids (>100) and the mantle (0.1) (Ishikawa and Nakamura 1994). The extent to which the sediment signal comes from melting versus extraction of fluid mobile elements by deeply sourced water is subject to debate (Stern 2002). Although enrichment of fluid-immobile species, such as 10Be and Th, in arc rocks requires sediment melting, most thermal models predict that subducted sediments are not heated sufficiently to melt phengite, the likely main boron host in the deeply subducted slab (Domanik and Holloway 1996).
Recycling into the Deep Mantle
The majority of boron entering the subduction zone is ultimately derived from seawater (either as sediment pore water or incorporated into hydrous minerals) and from continental crust (as clastic sediments). Much of this boron is recycled into seawater during dewatering of the slab during early subduction, or is recycled into arc crust in the form of arc volcanic and plutonic rocks (Moran et al. 1992). For any crustal and/or seawater boron to be recycled into the deep mantle it must be hosted in phases that are stable at greater pressures and temperatures than those within the arc magma generation zone. This either requires boron to be incorporated into minerals that are stable under upper mantle conditions and/or that minerals in crustally derived subducted sediments are not broken down during descent into the mantle.
While most of the boron in subducted altered oceanic crust and the overlying sediments is lost to fluid phases during dehydration and passes into the sub-arc mantle before the slab descends into the deep mantle, there are boron-bearing minerals that may survive beyond the volcanic front and that may be subducted into the deep mantle. Of the secondary boron-bearing minerals formed in the subducted slab, the most stable is phengite, which can persist up to temperatures of >1,000 °C and pressures up to 100 kbar (Domanik and Holloway 1996). Continued preferential extraction of high d11B fluids leads to phengite having d11B values as light as −18‰, but, importantly, it may still retain up to 50 ppm boron (Pabst et al. 2012; Halama et al. 2014). Thus, there is the potential for boron that has ultimately been derived from the continental crust and seawater (but with a much lighter d11B value) to be recycled into the deep mantle.
Subduction Zone Boron Budgets
Impact on Seawater Boron Isotopes
Attempts to incorporate subduction zone processes into global seawater boron isotope budgets have relied on data from only a few sites, but the simplest interpretation of these data is that most of the boron expelled back into seawater at subduction zones has a d11B of +13‰, compared to +40‰ in seawater. A plausible mass-balance can be achieved between the amount and isotopic composition of boron entering the subduction zone in pore waters and adsorbed to sediments, and that which is expelled back into seawater in forearc fluids (You et al. 1995).
The amount of boron expelled back into seawater at convergent margins depends on the amount of sediment entering the subduction zone. Higher subduction rates may result in increased return of labile boron to the oceans and lower seawater d11B values, although the long residence time of boron in the oceans (~10–20 My) mitigates against large and rapid changes in seawater d11B values.
Reconstruction of seawater d11B values over the past 50 My (Raitzsch and Honisch 2013) suggests that there has been an ~3‰ increase in seawater d11B since the Late Eocene, superimposed on shorter-term oscillations of up to 2‰ (Fig. 2A). It is interesting to note, therefore, that over this period there appears to be a first-order correlation between time-averaged subduction rates (Fig. 2B) (Chen et al. 2015) and the reconstructed seawater d11B, with peaks in the latter coinciding approximately with lower subduction rates. This apparent coincidence is not proof of a causal link between subduction rates and seawater d11B, but the existence of subduction-rate reconstructions extending back ~180 My (Engebretson et al. 1992) provides a testable hypothesis if reconstructions of seawater d11B values can be reliably extended further back in time.
Recycling into the Mantle
There is a strong correlation between the budgets of fluid mobile elements, such as boron, entering subduction zones and the content of these elements in arc rocks (Plank and Langmuir 1993). However, mass-balance uncertainties are too large to constrain mismatches in inputs and outputs in subduction zones. Nevertheless, the question as to whether boron is recycled into the deep mantle may be addressed by searching for d11B signatures of subduction in ocean island basalts that are unequivocally derived from the deep mantle and that have other isotope signatures (e.g. in Sr, Nd or Pb) that suggest a contribution from recycled subducted slabs (White 2015).
The first requirement is, however, knowledge of the d11B value of the primitive mantle (i.e. the composition of the mantle before extraction of any crust). This has been a difficult problem to solve because the boron concentrations of mantle-derived rocks are low and easily perturbed by post-eruptive alteration of primary signatures by circulating fluids and high-level assimilation of crustal material, including previously altered mantle-derived rocks. Recently, Marschall et al. (2017) undertook an ion-probe study of the boron isotope systematics of MORB and, together with consideration of other indices of alteration and assimilation, were able to determine that the d11B value of the mantle from which MORB is derived (depleted MORB mantle) is −7‰ ± 1‰, with a boron concentration of <0.1 ppm. Comparison with other stable isotope systems (White 2015) suggests this also likely represents the d11B of primitive mantle. Because MORB probably derives from mantle previously depleted in elements enriched in the crust (White 2015) and because boron is highly incompatible, the boron concentration of the primitive mantle is probably higher than that of depleted MORB mantle.
Relatively few ocean island basalt (OIB) suites have been analysed for boron isotopes and much of the wide spread in d11B values for individual sample suites has been ascribed to contamination of the magma source by high-level assimilation of altered rocks, rather than by variations in the mantle source. Thus, light d11B values (as low as −14‰) in some Icelandic melt inclusions have been attributed to the interaction of lavas with geothermal fluids (Brounce et al. 2012). In contrast, high-level crustal assimilation of altered oceanic crust is thought to be responsible for the high d11B values (up to +12‰) in western Azores (Portugal) ocean island basalts. The lower values (−7.4‰ to −3.3‰) observed in the eastern Azores are considered to more closely reflect mantle source compositions (Genske et al. 2014).
A compilation of ocean island basalt data (excluding those most obviously affected by high-level alteration/assimilation) is presented in Figure 3. Also shown are mixing trajectories between two potential subduction components: phengite and clastic sediments. Note that fractional crystallization may yield higher B concentrations, with no change in d11B values. Many of the low d11B OIB data could be construed to lie on a trend from mantle d11B values to the phengite d11B field (Fig. 3). In contrast, data from Hawaii (USA) (Tanaka and Nakamura 2005), the eastern Azores (Genske et al. 2014) and intraplate volcanic rocks from NE China (Li et al. 2016) trend to higher d11B values than the primitive mantle. This may reflect mixing of a primitive mantle source with deeply subducted and recycled sediments, as previously suggested for the Hawaiian data (Tanaka and Nakamura 2005). Interestingly, potential mixing lines with recycled sediments (Fig. 3) form a better fit to the data if the sediment endmember has d11B values that fall to the lower end of the range observed in clastic sediments. This would support the suggestion above that any subducted and recycled boron is likely to have a lighter d11B value than material entering the subduction system. Further studies are required to determine whether the OIB data truly reflect mantle heterogeneities, but separating the potential role of high-level assimilation/alteration processes from variations in the original mantle source will be difficult.
This review illustrates that boron more than lives up to its description of being “light and lively” in the upper portions of subduction zones. When it comes to deeper sections of the subducted slab, however, there are indications that there may be a more furtive side to its character, with some boron resolutely clinging on to refractory phases and returning to the deep mantle. This may be where boron behaviour in subduction zones has most promise of contributing to wider advances in Earth science. But there are several questions that must be resolved before this potential can be realised. While much information has been obtained from rocks metamorphosed under subduction zone conditions, overprinting by later fluids and the mobility of boron during retrograde metamorphism does complicate interpretations. Experimental studies of boron partition coefficients and of isotope fractionation between fluids and minerals and the amounts of these minerals formed under subduction zone conditions would help resolve these uncertainties. At present, there is little boron data available for ocean island basalts, particularly compared to other isotope systems. But advances in boron-isotope geochemistry will likely require a combination of micro-analytical techniques and a better evaluation of potential contamination of mantle signatures by high-level crustal assimilation and alteration.
Thanks to Gavin Foster, Tom Gernon and Rex Taylor for their comments on the manuscript. Cees-Jan De Hoog and Jeff Ryan also provided very constructive reviews. Thanks also to the wider boron isotope community that has grown exponentially since I joined a handful of pioneers 30 years ago.
Benton LD, Ryan JG, Tera F (2001) Boron isotope systematics of slab fluids as inferred from a serpentinite seamount, Mariana forearc. Earth and Planetary Science Letters 187: 273-282
Boschi C, Dini A, Früh-Green GL, Kelley DS (2008) Isotopic and element exchange during serpentinization and metasomatism at the Atlantic Massif (MAR 30oN): insights from B and Sr isotope data. Geochimica et Cosmochimica Acta 72: 1801-1823
Brounce M, Feineman M, LaFemina P, Gurenko A (2012) Insights into crustal assimilation by Icelandic basalts from boron isotopes in melt inclusions from the 1783–1784 Lakagígar eruption. Geochimica et Cosmochimica Acta 94: 164-180
Chen J, Kravchinsky VA, Liu X (2015) The 13 million year Cenozoic pulse of the Earth. Earth and Planetary Science Letters 431: 256-263
Domanik KJ, Holloway JR (1996) The stability and composition of phengitic muscovite and associated phases from 5.5 to 11 GPa: implications for deeply subducted sediments. Geochimica et Cosmochimica Acta 60: 4133-4150
Engebretson DC, Kelley KP, Cashman HJ, Richards MA (1992) 180 million years of subduction. GSA Today 2: 93-95, 100
Genske FS and 6 coauthors (2014) Lithium and boron isotope systematics in lavas from the Azores islands reveal crustal assimilation. Chemical Geology 373: 27-36
Halama R, Konrad-Schmolke M, Sudo M, Marschall HR, Wiedenbeck M (2014) Effects of fluid–rock interaction on 40Ar/39Ar geochronology in high-pressure rocks (Sesia-Lanzo Zone, Western Alps). Geochimica et Cosmochimica Acta 126: 475-494
Harvey J and 6 coauthors (2014) 11B-rich fluids in subduction zones: the role of antigorite dehydration in subducting slabs and boron isotope heterogeneity in the mantle. Chemical Geology 376: 20-30
Ishikawa T, Nakamura E (1993) Boron isotope systematics of marine sediments. Earth and Planetary Science Letters 117: 567-580
Ishikawa T, Nakamura E (1994) Origin of the slab component in arc lavas from across-arc variation of B and Pb isotopes. Nature 370: 205-208
Li H-Y, Zhou Z, Ryan JG, Wei G-J, Xu Y-G (2016) Boron isotopes reveal multiple metasomatic events in the mantle beneath the eastern North China Craton. Geochimica et Cosmochimica Acta 194: 77-90
Marschall HR and 5 coauthors (2017) The boron and lithium isotopic composition of mid-ocean ridge basalts and the mantle. Geochimica et Cosmochimica Acta 207: 102-138
Martin JB, Kastner M, Henry P, Le Pichon X, Lallement S (1996) Chemical and isotopic evidence for sources of fluids in a mud volcano field seaward of the Barbados accretionary wedge. Journal of Geophysical Research: Solid Earth 101: 20325-20345
Moran AE, Sisson VB, Leeman WP (1992) Boron depletion during progressive metamorphism: implications for subduction processes. Earth and Planetary Science Letters 111: 331-349
Morishige M, van Keken PE (2014) Along-arc variation in the 3-D thermal structure around the junction between the Japan and Kurile arcs. Geochemistry, Geophysics, Geosystems 15: 2225-2240
Pabst S and 6 coauthors (2012) The fate of subducted oceanic slabs in the shallow mantle: insights from boron isotopes and light element composition of metasomatized blueschists from the Mariana forearc. Lithos 132-133: 162-179
Palmer MR, Swihart GH (1996) Boron isotope geochemistry: an overview. Reviews in Mineralogy 33: 709-744
Peacock SM, Hervig RL (1999) Boron isotopic composition of subduction-zone metamorphic rocks. Chemical Geology 160: 281-290
Plank T, Langmuir CH (1993) Tracing trace elements from sediment input to volcanic output at subduction zones. Nature 362: 739-743
Raitzsch M, Hönisch B (2013) Cenozoic boron isotope variations in benthic foraminifers. Geology 41: 591-594
Savov IP, Ryan JG, D’Antonio M, Fryer P (2007) Shallow slab fluid release across and along the Mariana arc-basin system: insights from geochemistry of serpentinized peridotites from the Mariana fore arc. Journal of Geophysical Research: Solid Earth 112, doi: 10.1029/2006JB004749
Smith HJ, Spivack AJ, Staudigel H, Hart SR (1995) The boron isotopic composition of altered oceanic crust. Chemical Geology 126: 119-135
Spivack AJ, Palmer MR, Edmond JM (1987) The sedimentary cycle of the boron isotopes. Geochimica et Cosmochimica Acta 51: 1939-1949
Stern RJ (2002) Subduction zones. Reviews of Geophysics 40: 3-1 – 3-38, doi: 10.1029/2001RG000108
Straub SM, Layne GD (2002) The systematics of boron isotopes in Izu arc front volcanic rocks. Earth and Planetary Science Letters 198: 25-39
Tanaka R, Nakamura E (2005) Boron isotopic constraints on the source of Hawaiian shield lavas. Geochimica et Cosmochimica Acta 69: 3385-3399
Vils F, Tonarini S, Kalt A, Seitz HM (2009) Boron, lithium and strontium isotopes as tracers of seawater–serpentinite interaction at Mid-Atlantic ridge, ODP Leg 209. Earth and Planetary Science Letters 286: 414-425
White WM (2015) Probing the Earth’s deep interior through geochemistry. Geochemical Perspectives 4: 95-251
You C-F, Chan LH, Spivack AJ, Gieskes JM (1995) Lithium, boron, and their isotopes in sediments and pore waters of Ocean Drilling Program Site 808, Nankai Trough: implications for fluid expulsion in accretionary prisms. Geology 23: 37-40