Boron isotope variations in a single monogenetic cone: La Poruña (21°53′S, 68°30′W), Central Andes, Chile
.,CarolinaMarín
La Poruña is a monogenetic volcano located within the Altiplano-Puna Volcanic Complex (21°-24°S) in the Central Andean subduction zone. Since crustal contamination of Andean magmas is ubiquitous, and because extensive geochemical data exist for La Poruña, we employ this volcano as a case study to examine the behavior of boron isotopes during crustal assimilation. We present whole-rock boron concentration and 11B/10B ratios (as δ11B values) for La Poruña lava samples that were prepared as nano-particulate pressed pellets. La Poruña B contents range from 14 to 20 μg/g and δ11B values range from −1.39 ± 0.54 ‰ (2σ) to +0.94 ± 0.30 ‰ (2σ), which overlap with the range of available whole-rock data for Central Andean lavas. Moreover, La Poruña δ11B values correlate negatively with 87Sr/86Sr ratios from the same samples. Since 87Sr/86Sr is a proxy for crustal contamination at La Poruña, the data lead us to suggest that La Poruña magmas assimilated a low-δ11B, high 87Sr/86Sr component such as Andean continental crust. Mixing models based on B and Sr isotopes support a broadly two-step magma evolution for La Poruña. In step 1, mantle-derived primary melts interacted with boron-rich slab-derived fluids with high δ11B values, which yielded subduction-modified parental magmas with ca. 3 μg/g B and relatively high δ11B values. In step 2, the high δ11B parental magmas ascended through the crust where they assimilated up to 20% crustal material, which further modified their δ11B values and 87Sr/86Sr ratios. In comparison to available regional values for B and δ11B, it appears that La Poruña and nearby volcanic centers shared a similar source and magmatic history, whereas volcanoes south of 23°S differ. We stress, however, that deconvolving the roles of various subduction and crustal inputs in the Central Andes would require further studies on individual volcanoes along the arc.
Graphical abstract
1. Introduction
Mantle hydration by slab-derived fluids plays a key role in the genesis of magmas in subduction-related tectonic environments. This process is driven by dehydration of hydrous minerals in the down-going slab, which triggers partial melting of the overlying mantle wedge (e.g. Ryan and Chauvel, 2014; Spandler and Pirard, 2013). Most igneous rocks in subduction zones have greater concentrations of boron than depleted mantle (MORB; Mid-Ocean Ridge Basalt) as well as low ratios of fluid-immobile elements to boron (e.g. Nb/B). These observations suggest that the greater boron contents of subduction zone lavas (also known as arc lavas) are due to boron transfer from slab-derived fluids to the mantle source region (e.g. Ishikawa and Tera, 1997; Rosner et al., 2003; Ryan and Chauvel, 2014). Boron isotope ratios [expressed as δ11B = ((11B/10B(sample)/11B/10B(reference))-1) × 103 ‰] are useful tracers of various inputs to the subduction factory because during subduction and dehydration 11B is preferentially fractionated into the fluid phase, which leads to the formation of boron enriched slab-derived fluids with high δ11B values (e.g. Marschall et al., 2007; Peacock and Hervig, 1999; Ryan and Chauvel, 2014). Moreover, the δ11B values of slab-derived fluids progressively evolve towards lower δ11B values due to dehydration reactions and boron loss during subduction (e.g. Hervig et al., 2002; Konrad-Schmolke and Halama, 2014; Marschall et al., 2007; Peacock and Hervig, 1999). Therefore, fluids with variable δ11B values can be produced during subduction and react with the mantle wedge to generate the large range of δ11B values seen in global arc databases (see Deegan et al., 2016 for a summary). Once magma leaves its source and migrates to the surface it may interact with rocks comprising the upper arc crust. While the upper crust has a wide range of boron isotopic compositions, the δ11B values of continental crust are often lower than arc magmas (e.g. Kasemann et al., 2000; Palmer and Swihart, 1996). Hence, boron isotopes have the potential to distinguish between different subduction-related processes such as source hydration, slab contribution, and crustal assimilation at active continental margins (e.g. de Hoog and Savov, 2018; Deegan et al., 2016; Gméling et al., 2007; Palmer, 2017; Rosner et al., 2003; Schmitt et al., 2002; Tonarini et al., 2001a, Tonarini et al., 2001b).
The Central Andes (14°-27°S) is a tectono-magmatic province formed due to subduction of the Nazca Plate below the South American Plate. The recent (<10 Ma) volcanic arc of the Central Andes was constructed mainly by eruption of andesitic-to-dacitic lavas and dacitic-to-rhyolitic ignimbrites (e.g. de Silva and Kay, 2018; Mamani et al., 2010; Trumbull et al., 2006; Wörner et al., 2018). Partial melting of the mantle wedge, driven by fluids released from the slab, is considered to be the primary process involved in generating magmas in the Central Andes (e.g. de Silva and Kay, 2018; Godoy et al., 2014; Matteini et al., 2002; Wörner et al., 2018). After magma leaves its source region, magmatic differentiation involves significant degrees of crustal assimilation due the exceptional thickness of the continental crust below the volcanic arc, as reflected by radiogenic and stable isotopes signatures of erupted lavas (e.g. Godoy et al., 2014, Godoy et al., 2017; González-Maurel et al., 2019a, González-Maurel et al., 2020; Matteini et al., 2002; Michelfelder et al., 2013; Rosner et al., 2003; van Alderwerelt et al., 2021; Wörner et al., 2018).
To date, boron isotope studies of Central Andean rocks, minerals, and melt inclusions have mainly focused on characterizing specific geochemical reservoirs (e.g. upper continental crust; Kasemann et al., 2000) and investigating large-scale regional and temporal variations along the arc (Jones et al., 2014; Rosner et al., 2003). However, detailed boron isotope studies constraining the evolution of single volcanic systems are lacking. The purpose of this work, therefore, is to focus on boron isotope variations at the scale of a single monogenetic volcano where there exist abundant contextual geochemical data. To accomplish this, we selected La Poruña volcano (21°53′S, 68°30′W; Fig. 1), which is a young (<100 ka; Wörner et al., 2000; González-Maurel et al., 2019b) volcanic system in the Central Andes that is notable for its eruption of mafic lavas that record isotopic evidence for crustal assimilation (González-Maurel et al., 2019a, González-Maurel et al., 2020). Here we present new boron elemental and isotopic data for five whole-rock lava samples from La Poruña and compare them to available data for other volcanoes in the Central Andes (Fig. 1). Although our sample set is relatively small, the samples are geochemically well characterized, which makes La Poruña a good natural laboratory to study the behavior of boron isotopes during magmatic evolution in the Central Andes.
The current volcanic front of the Central Andes is a subduction-related arc in which primary magmatism is associated with partial melting triggered by hydration of the mantle wedge by slab fluids (e.g. Godoy et al., 2014; Matteini et al., 2002; Wörner et al., 2018). This volcanic arc is constructed on thick continental crust (>60 km; Beck et al., 1996; Prezzi et al., 2009) on which the recent (<10 Ma) volcanism developed (e.g. Godoy et al., 2017; Mamani et al., 2010; Trumbull et al., 2006; Wörner et al., 2018). The Altiplano-Puna Volcanic Complex (21°-24°S, de Silva, 1989) (Fig. 1) is a tectono-magmatic province within the Central Andes where an upper crustal (4–25 km below the surface) partially molten layer, the Altiplano-Puna Magma Body (Fig. 1), has been geophysically constrained (e.g. Chmielowski et al., 1999; Spang et al., 2021; Ward et al., 2014; Zandt et al., 2003). It has previously been suggested that the petrogenesis of erupted lavas of the Altiplano-Puna Volcanic Complex involved variable degrees of contamination by partial melts of the Altiplano-Puna Magma Body as well as the thick Andean continental crust (e.g. Feeley and Davidson, 1994; Godoy et al., 2017; González-Maurel et al., 2019a, González-Maurel et al., 2020; Michelfelder et al., 2013; Rosner et al., 2003; van Alderwerelt et al., 2021; Wörner et al., 2018).
La Poruña (21°53′S, 68°30′W) is a monogenetic volcanic structure located in the volcanic front of the Altiplano-Puna Volcanic Complex (Fig. 1). La Poruña consists of a scoria cone from which at least two lava flows erupted at ca. 100 ka (González-Maurel et al., 2019b; Marín et al., 2020) (Fig. 2). A detailed description of the lithofacies associated with the cone and the lava flows of La Poruña suggests different eruptive styles, which varied from Strombolian to effusive (Marín et al., 2020). Petrographic characteristics of samples from the scoria cone and the lava flows are similar (based on point counting and automated SEM mineralogical analysis; González-Maurel et al., 2019b and Marín et al., 2020). The main mineralogical assemblage at La Poruña includes plagioclase + orthopyroxene + clinopyroxene ± olivine ± amphibole phenocrysts (>0.5 mm) at ca. 5–40 vol% set in a groundmass of ca. 30–50 vol% consisting of glass and minor amounts of plagioclase + orthopyroxene + clinopyroxene ± olivine (González-Maurel et al., 2019b; Marín et al., 2020). The vesicle content of the erupted products varies from <5 vol% (Lithofacies Lava Flow Central; Fig. 2) to ca. 40 vol% (Lithofacies Cone and Lava Flow Peripheral; Fig. 2) (González-Maurel et al., 2019b; Marín et al., 2020).
Fig. 2. Satellite image (Google Earth™) showing the different lithofacies of La Poruña (after Marín et al., 2020). Locations of samples selected for boron isotope analysis in this study are also shown (samples are color-coded to Fig. 3).
Volcanism at La Poruña was likely supplied by a relatively primitive sub-arc parental magma that underwent moderate degrees of crustal contamination (<28%) en route to the surface (Godoy et al., 2017; González-Maurel et al., 2020). The mechanism of magma-crust interaction at La Poruña has been suggested to have been via “assimilation during turbulent ascent” (Godoy et al., 2020; González-Maurel et al., 2019b). In this model, assimilation of crustal material was more significant in the presence of hot, mafic magma and the degree of crustal assimilation decreased with increasing magmatic differentiation due to the changing thermal conditions over time (cf. Huppert and Sparks, 1985). As a consequence, the less evolved samples record the highest degree of crustal assimilation, i.e. samples with relatively low SiO2 have high 87Sr/86Sr ratios (see González-Maurel et al., 2019b; Godoy et al., 2020). The 87Sr/86Sr ratio of the La Poruña samples is therefore used as a proxy for crustal contamination in this paper, as outlined in Fig. 3.
Fig. 3. 87Sr/86Sr vs SiO2 (% m/m) diagram showing (a) the composition of samples from La Poruña (data from González-Maurel et al., 2019a, González-Maurel et al., 2019b), and (b) the proposed crustal contamination model. Note that at La Poruña, 87Sr/86Sr is used as a proxy for crustal contamination, where relatively high 87Sr/86Sr ratios in lava samples indicate that the corresponding magma assimilated continental crust with high 87Sr/86Sr ratios. Stars indicate the samples selected for boron isotope analysis in this study (for sample locations see Fig. 2). Error bars are smaller than symbol size.
3. Analytical methods
3.1. Sample selection
Five samples from La Poruña were selected for whole-rock boron elemental and isotopic analysis. All samples are fresh and show no obvious signs of alteration. One sample is from the cone (POR 15 03), one is from the central flow (POR 15 05), and three are from the peripheral flow (POR 05; POR 14 01; POR 17 04) (Fig. 2). The selected samples encompass the full range of 87Sr/86Sr ratios for La Poruña (Fig. 3). Detailed petrographic descriptions of the samples are provided by González-Maurel et al. (2019b) and Marín et al. (2020). Geochemical data including major and trace elements and 87Sr/86Sr ratios were previously published in González-Maurel et al., 2019a, González-Maurel et al., 2019b and are compiled in Supplementary Material Table SM1. The SiO2 composition of the selected samples varies from 56.6 to 60.5% m/m, while 87Sr/86Sr ratios vary from 0.706176 to 0.706640 (Table 1).
Table 1. Geochemical characteristics of the analyzed samples.
| Sample | Latitude | Longitude | SiO2⁎, a | Nb (μg/g)a | 87Sr/86Sra | 2σ (10−6)a | B (μg/g) | δ11B (‰) | 2σ (‰)b | δ11B (‰) | 2σ (‰) | δ11B (‰) | 2σ (‰) | δ11B (‰) | 2σ (‰) | δ11B (‰) | 2σ (‰) |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Sample mean | δ11B values of individual spot analyses | ||||||||||||||||
| POR 14 01 | 21° 53′ 27” | 68° 30′ 39” | 57.05 | 6.71 | 0.706640 | 8.10 | 17 | −1.39 | 0.54 | −0.77 | 0.21 | −1.35 | 0.15 | −1.53 | 0.17 | −1.91 | 0.17 |
| POR 15 03 | 21° 53′ 32” | 68° 29′ 58” | 57.93 | 6.09 | 0.706353 | 12.6 | 14 | −0.14 | 0.30 | +0.03 | 0.24 | −0.02 | 0.23 | −0.35 | 0.13 | −0.23 | 0.10 |
| POR 15 05 | 21° 55′ 32” | 68° 34′ 2” | 60.18 | 6.23 | 0.706184 | 9.62 | 17 | +0.38 | 0.43 | 0.70 | 0.17 | +0.23 | 0.17 | +0.21 | 0.18 | – | – |
| POR 17 04 | 21° 54′ 55” | 68° 32′ 25” | 60.86 | 6.20 | 0.706176 | 11.0 | 20 | +0.94 | 0.30 | +1.03 | 0.15 | +0.67 | 0.11 | +1.02 | 0.14 | +1.06 | 0.12 |
| POR-05 | 21° 53′ 42” | 68° 30′ 31” | 58.70 | 6.55 | 0.706441 | 10.0 | 14 | −0.12 | 0.34 | +0.21 | 0.22 | −0.12 | 0.21 | −0.26 | 0.21 | −0.30 | 0.16 |
| Reference materials analyzed as unknowns | |||||||||||||||||
| JR-1 | This study | 130 | +3.01 | 0.26 | +2.85 | 0.05 | +3.06 | 0.07 | +3.07 | 0.08 | +3.05 | 0.08 | |||||
| Gangjian et al., 2013 | +2.80 | 0.20 | |||||||||||||||
| JR-2 | This study | 140 | +2.70 | 0.32 | +2.55 | 0.08 | +2.78 | 0.07 | +2.77 | 0.06 | – | – | |||||
| Gangjian et al., 2013 | +2.70 | 0.20 | |||||||||||||||
| JB-3 | This study | 20 | +6.02 | 0.26 | +5.93 | 0.16 | +5.99 | 0.11 | +6.02 | 0.16 | +6.15 | 0.18 | |||||
| Gangjian et al., 2013 | +5.60 | 0.10 | |||||||||||||||
| JA-2 | This study | 20 | −9.65 | 0.26 | −9.54 | 0.12 | −9.63 | 0.17 | −9.71 | 0.16 | −9.73 | 0.11 | |||||
| Gangjian et al., 2013 | −9.30 | 0.40 | |||||||||||||||
| Additional reference materials | |||||||||||||||||
| JB-2 | Liu et al., 2018 | 27 | +7.12 | 0.2 | |||||||||||||
- ⁎
SiO2 recalculated to 100% (m/m) water free.
- a
Data after González-Maurel et al., 2019a, González-Maurel et al., 2019b.
- b
Full error propagation considering measurements of the samples and reference material (JB-2), error on the value used for the reference material, and error on the value used for NIST951 (11B/10B = 4.04362 ± 0.001 2σ; Liu et al., 2018).
3.2. Analysis of boron concentrations
Approximately 3 g of homogeneous sample powders were made into nano-particulate pressed pellets (“nano-pellets”) following established methods (Garbe-Schönberg and Müller, 2014) at Kiel University using a FRITSCH Pulverisette 7 PL and slightly modified milling and tablet-pressing protocols. The nano-pellets were analyzed for multi-trace elemental abundances – including boron – at the ICPMS-Lab, Institute of Geosciences, Kiel University, Germany using an AGILENT 8900 ICP-MS/MS instrument coupled to a COHERENT 193 nm GeoLas HD laser ablation unit (for details see Garbe-Schönberg and Müller, 2014). Average analytical precision for all elements analyzed with five spots per pellet was 2.6% RSD (1SD) and 2.8% RSD for boron while average measurement uncertainty over the length of the session as determined from replicate measurements of the samples was 2.6% RSD for all trace elements but increased to 9.5% RSD for boron. Boron elemental data for all individual spot measurements on nano-pellets of the La Poruña samples and the reference materials are provided in Supplementary Material Table SM2.
3.3. Analysis of boron isotopes
The same nano-pellets on which elemental concentrations were determined in Kiel, were analyzed for δ11B compositions at the MC-ICP-MS Facility, Department of Geological Sciences, University of Cape Town, South Africa using a Nu Instruments NuPlasma HR coupled to an ASI RESOlution SE 193 nm laser ablation unit. To maximize boron sensitivity, the NuPlasma HR was operated at 6 kV and fitted with a light-element, high-abundance skimmer cone. A laser spot size of 220 μm was used and scanned at 6 μm.sec−1 along a 450 μm line, after rapid pre-ablation cleaning with a 240 μm spot. Ablation was conducted under helium (400 ml.min−1) and nitrogen (9 ml.min−1), blended with argon prior to injection into the plasma. A laser energy setting of 1.2 kV was used, which yielded an energy density of ±4.3 J.cm−2. All analyses of unknown samples were bracketed with analyses of a JB-2 nano-pellet for which a δ11B value of 7.12 ± 0.20 ‰ was assigned, enabling the referencing of all unknown sample δ11B data to a11B/10B ratio value of 4.04362 ± 0.001 for NIST SRM951 (Liu et al., 2018). The final δ11B value of each unknown sample is the average of 4 successive JB-2-bracketed analyses with the 2σ values based on the reproducibility of these analyses and the propagated uncertainty of the reference values used for JB-2 and NIST SRM951. Results for reference materials with published boron isotope compositions and prepared as nano-pellets are reported in Table 1.
4. Results
The analyzed samples have boron contents ranging from 14 to 20 μg/g and δ11B values ranging from −1.39 ± 0.54 to +0.94 ± 0.30 ‰ (2σ uncertainty; see Table 1). The new data reported here are plotted in Fig. 4 along with literature data for Central Andean whole-rocks (Rosner et al., 2003; Schmitt et al., 2002) and melt inclusions (Jones et al., 2014; Schmitt et al., 2002). The La Poruña data overlap with available whole-rock data, and in some cases, they appear to “fill a gap” in the arc-wide dataset of Rosner et al. (2003) (e.g. Fig. 4b-e). The La Poruña samples cluster at relatively low Nb and low B contents compared to the regional data (Fig. 4a) and their boron content increases with decreasing Nb/B, similarly to the literature whole-rock data (Fig. 4b). Notably, the Nb/B ratio of La Poruña samples is significantly lower than MORB (see arrow in Fig. 4b), which is typical for arc sample suites and indicates involvement of boron-rich, slab-derived fluids in petrogenesis (e.g. Ishikawa and Tera, 1997).
With respect to δ11B values, the La Poruña samples plot close to the δ11B versus Nb/B regression line constructed for a suite of Central Andean whole-rock lava samples by Rosner et al. (2003) (Fig. 4c). However, looking at the La Poruña samples in isolation shows that there is a rather weak correlation between δ11B and Nb/B (R2 = 0.24; not shown). On at a plot of δ11B versus B, the La Poruña samples again overlap with the regional trend and plot close to samples from the near-by volcanoes Licancabur and Sairecabur (Fig. 4d). Finally, on a plot of δ11B versus 87Sr/86Sr, we can see that the La Poruña samples plot within the space defined by the three geochemical end-members of Altered Ocean Crust (AOC), Mid-Ocean Ridge Basalt (MORB), and Andean continental crust (Fig. 4e). On closer inspection, a strong positive correlation between δ11B and 87Sr/86Sr is revealed (R2 = 0.91; Fig. 4f). This relationship will be examined further in the Discussion section below.
5. Discussion
5.1. Overview of B contents and δ11B values for La Poruña
Compared to the available whole-rock boron concentration data for the Central Andes, La Poruña has relatively low boron contents at 16 μg/g on average (Fig. 4a). There is a strong correlation between B and Nb/B among our samples (R2 = 0.93; Fig. 4b). Since the Nb/B ratio is a proxy for fluid transfer in subduction zones, this relationship supports a role for slab-derived fluids at La Poruña, similar to what has been proposed for other arc lava suites (e.g. de Hoog and Savov, 2018; Ishikawa and Tera, 1997; Tonarini et al., 2001a). There is a weak positive correlation between B concentration and SiO2 content (R2 = 0.48; not shown). Although there is scatter in the data and the dataset is small, the overall increase in B with SiO2 is consistent with B behaving as an incompatible element during differentiation (see also Schmitt et al., 2002).
Turning to the boron isotope data, La Poruña δ11B values vary from −1.39 to +0.94 ‰ (Fig. 4, Table 1). These values agree well with boron isotope data from other volcanoes in the region (Fig. 4c-d) and they vary systematically with 87Sr/86Sr ratios obtained on the same samples (Fig. 4e-f). We will now explore potential reasons for these variations. As described above, boron isotope variations in volcanic arcs are often attributed to variable inputs of isotopically heterogeneous slab-derived fluids to the mantle wedge, which in turn may be related to the distance from the arc trench (e.g. Tonarini et al., 2001a; Rosner et al., 2003; Palmer, 2017; Agostini et al., 2021). Since La Poruña is a young, monogenetic volcanic complex, its distance from the arc front has remained fixed during its growth and evolution (González-Maurel et al., 2019b). Therefore, the observed variations in δ11B values at La Poruña cannot be related to the geometry of the volcanic arc. The variations in δ11B values at La Poruña may instead be a result of fractional crystallization. However, this hypothesis can be ruled out on the basis of the similar mineral content of the different lithofacies at La Poruña (González-Maurel et al., 2019b; Marín et al., 2020). Moreover, at La Poruña there is a change in δ11B values of >2‰ over a relatively narrow SiO2 range of ca. 4 % m/m, which is likely not due to fractional crystallization since boron isotopes are not expected to fractionate significantly during crystallization at high temperatures (see the fractionation models in Kaliwoda et al., 2011). There are also different δ11B values among samples with similar volume % of vesicles, which leads us to rule out isotopic fractionation due to degassing (cf. Schmitt et al., 2002). In the section below, we will now consider crustal assimilation as a process that may have modified boron isotope ratios at La Poruña.
5.2. Crustal assimilation and mixing models
In Section 5.1 above we ruled out volcano-arc geometry, fractional crystallization, and degassing as factors controlling the variability in δ11B values at La Poruña. Given that crustal assimilation is well documented both for the Central Andes at large and for La Poruña in particular (e.g. with respect to Sr and O isotopes; González-Maurel et al., 2019a, González-Maurel et al., 2019b, González-Maurel et al., 2020), we now turn our focus to the question of whether the boron isotope variations at La Poruña reflect interaction between magma and its host rocks during crustal storage and ascent. Strontium isotopes are a good proxy for crustal contamination and co-vary significantly with δ11B in our samples (Fig. 4f), whereby the most contaminated sample (i.e. the one with the highest 87Sr/86Sr ratio) has the lowest δ11B value. This observation suggests that La Poruña parental magma assimilated a contaminant with high 87Sr/86Sr and low δ11B values, such as the Andean continental crust (87Sr/86Sr = 0.71428, Ort et al., 1996 and δ11B = −8.9 ‰, Kasemann et al., 2000). A similar scenario has previously been suggested for the Central Andes, as outlined below.
To explain regional variations in δ11B values in the Central Andes, Rosner et al. (2003) suggested a two-step petrogenetic model. In the first step, primary mantle-like melts with relatively low δ11B values and low 87Sr/86Sr ratios were mixed with 11B-enriched AOC-derived slab fluids. In the models of Rosner et al. (2003), addition of small amounts (ca. 1.5%) of AOC-derived slab fluid to a mantle-derived melt yielded parental arc magmas with δ11B values of ca. -2.8 to +6.4 ‰, depending on whether isotope fractionation across the arc was considered or not (i.e. from volcanic front to back arc). In the second stage of evolution, these subduction-modified parental arc melts were mixed with Andean continental crust, the latter of which is characterized by low δ11B values of around −8.9 ‰ (Kasemann et al., 2000). These two stages of mixing were suggested by Rosner et al. (2003) as a way to explain the B contents and δ11B values of their suite of Central Andean lavas. In Fig. 5 we present similar B
Sr isotope mixing models, but with additional mixing curves that intersect the data for La Poruña. In order to satisfactorily explain the La Poruña data, we found that a subduction-modified parental arc magma with relatively high δ11B values is required (i.e. around +5.1 to +6.4 ‰; see Fig. 5). We therefore suggest that at La Poruña the primary mantle source interacted with high δ11B slab-derived fluids generated in the volcanic front and that there was little or no input of slab-derived fluids generated in the back arc, as these would have relatively low δ11B values (see Rosner et al., 2003 and Fig. 5b). In summary, our models indicate that La Poruña magmas record up to 20% addition of Andean continental crust (δ11B = −8.9 ‰) to a high δ11B subduction-modified arc parental melt (δ11B = +5.1 to +6.4‰) that contained ca. 3 μg/g boron. The amount of assimilation required to explain the data would vary depending on the exact end-members chosen (e.g. uncertainties exist around the exact nature of the slab-derived fluids involved and it may be an over-simplification to represent the Andean continental crust as a fixed point when it naturally has a range of values). Nonetheless, we offer broad constraints on the processes controlling δ11B values at La Poruña that are consistent with the models presented in Rosner et al. (2003) and, notably, with the results of mixing models using oxygen isotopes where up to 28% crustal assimilation was suggested for this volcano (González-Maurel et al., 2020).
Fig. 5. Mixing models for BSr isotopes. These models are largely based on Rosner et al. (2003) and include “static” models (a), where no slab-fluid isotopic fractionation takes place and “dynamic” models (b), where the δ11B value of the slab-derived fluid changes as a function of distance to the trench. The models involve two steps. In Step 1, a mantle-derived melt is mixed with < 2% of an AOC-derived slab fluid (Step 1 mixing curves shown after Rosner et al., 2003). This yields a subduction-modified parental arc melt with a δ11B value of +5.1 ‰ in the static model and with δ11B values between −2.8 ‰ (for addition of back arc fluids; BA) and + 6.4 ‰ (for addition of volcanic front fluids; VF) in the dynamic model. In Step 2, the subduction-modified parental melt is mixed with Andean continental crust. Note that in the dynamic model, a parental melt formed by mixing with relatively low δ11B values (i.e. back arc) cannot reproduce the data from La Poruña. Therefore (and for clarity), Step 2 mixing curves are not shown for a BA-modified parental melt in panel (b). Values for mantle wedge, AOC fluids, and Andean crust are from Table 3 in Rosner et al. (2003). Mixing curves are color-coded to the B content of the parent arc magma (0.1 to 10 μg/g). The La Poruña data can be explained by up to 20% mixing of continental crust into a subduction-modified, parental magma with relatively high initial δ11B values (+5.1 to +6.4 ‰) and with B contents of around 2 to 3 μg/g.
5.3. Regional perspective
Boron enrichment in volcanic arcs is generally thought to be due to addition of boron to the mantle melting region via slab-derived fluids (e.g. de Hoog and Savov, 2018; Ishikawa and Tera, 1997; Tonarini et al., 2001a). At La Poruña, an initially mantle-derived melt with very low B content (e.g. B = 0.07 μg/g for MORB; Marschall et al., 2017) was most likely modified by addition of small amounts of slab-derived fluids such that its boron content increased to around 2 to 3 μg/g, as indicated by our mixing models (Fig. 5). Subsequent crystallization would have led to an overall increase in B with SiO2 as B is expected to behave as an incompatible element during differentiation (see also Schmitt et al., 2002).
Comparing our data for La Poruña (this study) to data for other Central Andean volcanoes (Rosner et al., 2003), we can see that lavas from Sairecabur and Licancabur volcanoes have lower δ11B values than those of La Poruña, which we propose is due to higher degrees of assimilation of the low δ11B Andean continental crust at those localities (Fig. 6, Fig. 7). Indeed, greater degrees of crustal contamination towards the center of the Altiplano-Puna Magma Body have previously been suggested based on 87Sr/86Sr ratios and other geochemical characteristics of coeval volcanism within the Altiplano-Puna Volcanic Complex (e.g. Michelfelder et al., 2013; Godoy et al., 2017; González-Maurel et al., 2019a; Fig. 7). The broad-stroke differences in the B contents and δ11B values of the group of volcanoes comprising La Poruña, Sairecabur, and Licancabur compared to the group comprising Lascar, Cerro Overo and Rincón (Fig. 6) imply variations whose underlying cause might be related to large-scale changes in subduction thermal conditions, slab fluid composition, and/or mantle composition (e.g. Comte et al., 2016; Contreras-Reyes et al., 2021; Gao et al., 2021; Godoy et al., 2019; Jones et al., 2014; Kay et al., 1994; Rosner et al., 2003). It is currently difficult to distinguish between these possibilities, especially when comparing data from low spatial resolution studies where only one or two data points per volcano are available. We stress here that detailed study of the boron isotopic variations at individual volcanic complexes in different parts of the Central Andes would be needed in order to clarify the sources of regional heterogeneities and to fully assess the relative roles of different subduction and crustal inputs along the arc.
Fig. 6. Global Multi-Resolution Topography image for the Central Andes showing regional variations in (a) boron content (μg/g) and (b) δ11B values (‰) in whole-rock lava samples from several volcanoes. Dashed lines represent the depth of the Wadatti-Benioff zones (after Cahill and Isacks, 1992). Data are from this study and Rosner et al. (2003). For analytical uncertainties see Methods. Background map made with GeoMapApp (www.geomapapp.org; Ryan et al., 2009).
Fig. 7. a) Whole-rock δ11B values of La Poruña (this work) compared to lavas (Rosner et al., 2003) and melt inclusions (averages from Jones et al., 2014) from the Central Andes. Also shown are the ranges of δ11B values for Central Andean ignimbrites (Schmitt et al., 2002), Central Andean basement rocks (Kasemann et al., 2000), hydrated mantle wedge (Straub and Layne, 2002), MORB (Marschall et al., 2017), and serpentinite (Scambelluri and Tonarini, 2012). b) Schematic model for the Central Andes (between 21°-24°S) showing a variably subduction-modified sub-arc mantle and a thick continental crust wherein lies the Altiplano-Puna Magma Body (APMB). The B and B-isotope data for La Poruña (this study) suggest a subduction-modified parental arc magma with relatively high δ11B values (+5.1 to +6.4 ‰) and with B contents of around 2 to 3 μg/g (see box 1). This parental magma then ascended through the arc crust where its δ11B values were modified by crustal assimilation involving low δ11B continental crust (see box 2). The magma plumbing system shown for La Poruña is based on Gonzalez-Maurel et al. (2019b).
6. Concluding remarks
The availability of geochemically well-characterized lava samples from La Poruña monogenetic scoria cone allowed for detailed investigation regarding how δ11B values are modified as a consequence of crustal assimilation. La Poruña B contents range from 14 to 20 μg/g and δ11B values range from −1.39 ± 0.54 ‰ (2σ) to +0.94 ± 0.30 ‰ (2σ), which overlap with the range of available whole-rock data for Central Andean lavas. A striking feature of our data is that La Poruña δ11B values correlate negatively with 87Sr/86Sr isotopic ratios obtained on the same samples (R2 = 0.91). Since 87Sr/86Sr is a strong proxy for crustal contamination at La Poruña, the trend towards relatively low δ11B values with increasing 87Sr/86Sr (i.e. with increasing degree of crustal contamination) suggests that the B isotope compositions of La Poruña magmas were controlled by assimilation of a crustal component with low δ11B values and high 87Sr/86Sr ratios, such as the local Andean continental crust. Mixing models based on B and Sr isotopes support a broadly two-step magma evolution for La Poruña (summarized in Fig. 7). In step 1, mantle-derived primary melts interacted with boron-rich slab-derived fluids with high δ11B values, which yielded subduction-modified parental magmas with ca. 3 μg/g B and relatively high δ11B values. In step 2, the high δ11B parental magmas ascended through the crust where they assimilated up to 20% crustal material, which further modified their δ11B values and 87Sr/86Sr ratios. In comparison to available regional values for B and δ11B, it appears that La Poruña and nearby volcanic centers shared a similar source and magmatic history, while volcanoes south of 23°S differ. Deconvolving the roles of various subduction and crustal inputs in the Central Andes would require further studies on individual volcanoes along the arc.
Declaration of Competing Interest
Benigno Godoy reports financial support was provided by the National Fund For Scientific Technological and Technological Innovation Development and the National Commission for Scientific and Technological Research Funds for Research Centers in Priority Areas. Frances M. Deegan reports financial support was provided by the Swedish Research Council and Uppsala University. Petrus le Roux reports financial support was provided by the National Research Foundation of South Africa.
Acknowledgments
We thank Margaret Hartley for her thorough and thoughtful review that helped to improve this work. We also thank an anonymous reviewer for their input and Greg Shellnut for editorial handling. This work is based on research supported in part by funding from ANID “FONDECYT Iniciación Project” n°11200013 to B.G., the Swedish Research Council (Vetenskapsrådet) and the Section for Natural Resources and Sustainable Development (NRHU) at the Department of Earth Sciences, Uppsala University to F.M.D., and the National Research Foundation of South Africa (Grant number: 129318) to P.L.R.
Comments
Post a Comment