3.4. DISCUSSION
3.4.1. Calcium Isotopic Record of Marine Carbonates
Calcium isotope measurements of carbonates are generally light by up to several per mil (De La Rocha & DePaolo, 2000; DePaolo, 2004; Fantle & DePaolo, 2005; Farkas et al., 2007; Griffith et al., 2008; Heuser et al., 2005; Kasemann et al., 2005; Watkins et al., 2017; Zhu & MacDougall, 1998) when compared to unmelted/non‐metasomatized peridotites (Kang et al., 2017), komatiites (Amsellem et al., 2019), and most igneous rocks (e.g., Antonelli & Simon, 2020; Chen et al. 2019; Schiller et al., 2016; Simon & DePaolo, 2010), as shown in Fig. 3.3. It is possible that the marine carbonate record may have evolved over time towards higher δ44Ca values (Farkas et al., 2007). The notion that marine carbonates had light calcium isotope in the distant past has been challenged recently by Blattler and Higgins (2017), who report that, on average, Precambrian carbonates (n=505) are indistinguishable from BSE (± 0.24‰). All carbonate sediments undergo an extended diagenetic evolution after deposition. However, the effects of recrystallization involving pore fluids on the calcium isotopic composition in modern carbonates are minor compared to the magnitude of their light calcium isotope compositions (Fantle & DePaolo, 2005; 2007). Whether Precambrian carbonates were deposited with light calcium isotope compositions and subsequently modified to heavy calcium isotope compositions remains to be seen, but some modification might be expected due to isotopic exchange with heavy seawater and/or relatively heavy Ca‐bearing fluids derived from a silicate mantle or crustal reservoir (i.e., John et al., 2012). Nevertheless, detailed work of Fantle and DePaolo (2005) and Griffith et al. (2008), followed up by the compilation of data from over 70 studies included in Fantle and Tipper (2014), shows that over at least the last ~20–30 Ma there has been a significant difference between “lighter” carbonate and “normal” silicate calcium reservoirs.
Figure 3.4 Ba/Th‐87Sr/86Sr (a), 143Nd/144Nd‐87Sr/86Sr (b), and 206Pb/204Pb‐87Sr/86Sr (c) diagrams for Central American volcanic arc lavas. All samples from the volcanic front (VF) have geochemical signatures interpreted to be elevated above values for back‐arc lavas (BA) by the addition of a sedimentary subducted component, e.g., enriched in 87Sr/86Sr. Back‐arc lavas including YO1 remain within the mantle field, reflecting mixtures of MORB‐like depleted mantle and HIMU, enriched mantle (data and illustrative mixing curves from Carr et al., 1990; Feigenson & Carr, 1986; Feigenson et al., 2004; Patino et al., 1997, 2000). Orange and gray squares represent hemipelagic and carbonate DSDP 495 sediment compositions, respectively.
3.4.2. Calcium Isotopic Record of Mantle‐Derived Rocks
Most mantle‐derived rocks and peridotites studies report calcium isotope compositions similar to BSE (e.g., Amsellem et al., 2019; Ionov et al., 2019; Kang et al., 2017; Simon & DePaolo, 2010). There are reports of mantle rocks including peridotites that have distinctly heavy δ44Ca isotopic compositions, e.g., Amini et al. (2009), Huang et al. (2010), Kang et al. (2017), and Lu et al. (2019), and a few that are distinctly light, e.g., Amsellem et al. (2019) and Zhao et al. (2017). In order to investigate this variability, Huang et al. (2010) measured mineral separates from mantle rocks and found that δ44Ca in orthopyroxenes (opx) are significantly heavier than their associated clinopyroxenes (cpx) by 0.36–0.75‰. The magnitude and sign of the measured differences are generally consistent with first principles equilibrium intermineral isotope fractionation calculations that fundamentally depend on calcium concentration (i.e., Ca—O bond strength) and temperature (Antonelli et al., 2019a; Wang et al., 2017). In detail, calcium isotope fractionation between opx and cpx (Δ44/40Caopx‐cpx) less than ~0.26‰ and greater than ~0.60‰ is likely related, at least in part, to disequilibrium calcium isotopes effects such as metasomatic metamorphism (Zhao et al., 2017). Disequilibrium effects have also been reported for volcanic settings during rapid crystal growth (Antonelli et al., 2019b) and between opx and cpx and with other minerals during granulite facies and ultrahigh‐temperature metamorphism (Antonelli et al., 2019a). In these cases, calcium concentration likely plays an important role, i.e., it is lower by ~1/32 in opx compared to cpx, and may be more easily affected by isotope fractionation governed by diffusive loss (or gain) of calcium. This would explain why cpx tends to have compositions closer to BSE but why opx compositions can vary wildly – opx as high as δ44Ca ~6 has been found in mafic granulite samples (Antonelli et al 2019a). Interestingly, a recent investigation using calcium isotope signatures of carbonatite and silicate metasomatism and melt percolation found little evidence for calcium isotopic heterogeneity and concluded that metasomatism tends to decrease δ44Ca values of metasomatized mantle materials, but that its effects are usually limited (≤0.3‰) (Ionov et al., 2019).
3.4.3. Calcium Isotopes Exhibit no Evidence for Carbonate Sediment Recycling at Subduction Zones
In the studied Central American arc magmas, I found no evidence for calcium isotopic heterogeneity and thus no evidence for carbonate recycling or any isotopic fractionation related to subduction. This is the case despite the fact that I selected rocks that have both little to no geochemical evidence for sediment subduction, i.e., YO1 has MORB‐like trace element signatures and depleted mantle (DM) radiogenic isotope compositions, and rocks with strong trace element and radiogenic isotope signatures for carbonate sediment subduction (Fig. 3.2).
To date, resolvable radiogenic calcium isotopic signatures have not been observed in any oceanic or arc basalts (Huang et al., 2011; Marshall & DePaolo, 1989; Simon et al., 2009). This might not be surprising given the work of Caro et al. (2010) who, despite finding well‐defined excesses of 40Ca in some river waters draining into the ocean, report that no discernable effects of 40K decay, to within their reported analytical precision (~0.4 epsilon units, 2σ), exist in marine carbonate samples ranging in age from Archean to recent.
There have been recent studies of mantle‐derived rocks that find little evidence that recycling of carbonates affects the calcium isotope values of the mantle on a global or regional scale (Antonelli et al. 2019a; Ionov et al., 2019). However, other calcium isotope studies of primitive igneous rocks report evidence for recycling, e.g., Banergee and Chakrabarti (2019), Chen et al. (2018), Huang et al. (2011), Kang et al. (2016, 2017), and Liu et al. (2017). My results are significant since the trace element and radiogenic isotope signatures (e.g., high Ba/La, Ba/Th, 87Sr/86Sr, 206Pb/204Pb; see Figs. 3.2 and 3.4) of Central American lavas suggest a significant contribution from subducted carbonates (Patino et al., 2000; Sadofsky et al., 2008). The geochemical decoupling reported herein contrasts with the signatures reported for the ocean island basalts studied by Huang et al. (2011). In the Huang et al. (2011) study, stable mass‐dependent calcium isotope signatures vary and correlate with other geochemical parameters (i.e., Sr/Nb and 87Sr/86Sr) used to support the interpretation that Hawaiian lavas represent recycling of ancient calcium bearing surface materials.
All samples from the volcanic front (VF) are interpreted to be elevated in their Pb and Sr radiogenic isotopes above values for back‐arc lavas (BA) by the addition of a sedimentary subducted component (see Fig. 3.4; Carr et al., 1990; Feigenson & Carr, 1986; Feigenson et al., 2004; Patino et al. 1997, 2000). Back‐arc lavas including YO1 remain within the mantle field, reflecting mixtures of MORB‐like depleted mantle (DM) and enriched mantle (HIMU). The potential sedimentary contribution