Permian–Carboniferous arc magmatism in southern Mexico: U–Pb dating, trace element and Hf isotopic evidence on zircons of earliest subduction beneath the western margin of Gondwana

Abstract

Undeformed felsic to mafic igneous rocks, dated by U–Pb zircon geochronology between 311 and 255 Ma, intrude different units of the Oaxacan and Acatlán metamorphic complexes in southwestern Mexico. Rare earth element concentrations on zircons from most of these magmatic rocks have a typical igneous character, with fractionated heavy rare earths and negative Eu anomalies. Only inherited Precambrian zircons are depleted in heavy rare earth elements, which suggest contemporaneous crystallization in equilibrium with metamorphic garnet during granulite facies metamorphism. Hf isotopic signatures are, however, different among these magmatic units. For example, zircons from two of these magmatic units (Cuanana pluton and Honduras batholith) have positive εHf values (+3.8–+8.5) and depleted mantle model ages (using a mean crustal value of ¹⁷⁶Lu/¹⁷⁷Hf = 0.015) (T _DMC) ranging between 756 and 1,057 Ma, whereas zircons from the rest of the magmatic units (Etla granite, Zaniza batholith, Carbonera stock and Sosola rhyolite) have negative εHf values (−1 to −14) and model ages between 1,330 and 2,160 Ma. This suggests either recycling of different crustal sources or, more likely, different extents of crustal contamination of arc-related mafic magmas in which the Oaxacan Complex acted as the main contaminant. These plutons thus represent the magmatic expression of the initial stages of eastward subduction of the Pacific plate beneath the western margin of Gondwana, and confirm the existence of a Late Carboniferous–Permian magmatic arc that extended from southern North America to Central America.

Appendix 1

Hf isotope methodology employed at LEI, CGEO, UNAM.

Hafnium isotopes were measured at LEI, CGEO, UNAM, using a Neptune Plus MC-ICPMS coupled to the Resolution M050 workstation. The Neptune Plus is equipped with 9 Faraday detectors and a bed of one 10⁻¹⁰ Ω, one 10⁻¹² Ω and eight 10⁻¹¹ Ω amplifiers. Only the latter were used for the static collection of ¹⁷²Yb, ¹⁷³Yb, ¹⁷⁵Lu, ¹⁷⁶Hf–Lu–Yb, ¹⁷⁷Hf, ¹⁷⁸Hf, ¹⁷⁹Hf and ¹⁸⁰Hf. Each ablation lasted for 40 s, employing a repetition rate of 5 Hz, a fluence of 6 J/cm² and a laser spot of 44–60 μm in diameter on the top of previous U–Pb pits. Seven hundred ml of He was used as carrier gas, to which 11 ml of N₂ was added right after the ablation cell but prior to the mixing with Ar sample gas (about 960 ml). Gas blanks were measured prior to each analysis without the laser firing, and then subtracted from the main signal during data reduction. It has been widely demonstrated that both spatial resolution and precision of in situ Hf analyses by LA–MC–ICPMS using ablation of 40–60 μm are useful and adequate for petrogenetic interpretation, even if a minimal volume of material is analyzed (Woodhead et al. 2004; Hawkesworth and Kemp 2006; Gerdes and Zeh 2009; Cecil et al. 2011). The most important factors that must be controlled to obtain geologically meaningful data are the way that mass bias and interferences are monitored and corrected. The interference corrections are crucial, because ¹⁷⁶Hf is interfered by ¹⁷⁶Lu and ¹⁷⁶Yb, both naturally present in zircon, and their content must be correctly evaluated and taken into account. As demonstrated by other authors (e.g., Woodhead et al. 2004; Gerdes and Zeh 2009), the Hf mass bias (βHf) value is different from the βYb and βLu and thus cannot be employed to correct for mass bias on these species. For Lu, the existence of only one isotope free of isobaric interferences, ¹⁷⁵Lu, does not allow to calculate a proper mass bias correction factor. In this case, ¹⁷⁵Lu was measured and the ¹⁷⁶Lu/¹⁷⁵Lu ratio = 0.02656 (Blichert-Toft et al. 1997) was employed, together with βYb, assuming that Lu fractionates as Yb. For Yb, three interference-free isotopes are available, namely ¹⁷¹Yb, ¹⁷²Yb and ¹⁷³Yb. We measured ¹⁷²Yb, ¹⁷³Yb, the two most abundant isotopes, and used the ¹⁷⁶Yb/¹⁷³Yb ratio = 0.79618 (Chu et al. 2002) in order to estimate the other isotopes of Hf. Normalizing values were ¹⁷⁹Hf/¹⁷⁷Hf = 0.7325 (Patchett and Tatsumoto 1981) and ¹⁷²Yb/¹⁷³Yb true ratio = 1.35274 (Chu et al. 2002). It has been correctly reported that such method of in situ determination of βYb works fine when the lowest Yb signal exceeds roughly 50 mV (e.g., Gerdes and Zeh 2009; Cecil et al. 2011). In our case those values were exceedingly met, due to the Neptune Plus improved interface that, once the Jet sample cones are employed, provides a total Hf signal of more than 20 V, and a ¹⁷³Yb signal averaging more than 200 mV. One significant parameter used to test the effectiveness of our interference correction on unknown zircons is to analyze well-characterized standard zircons as unknowns, to which the same correction is applied. It has been documented that natural zircons have a ¹⁷⁶Yb/¹⁷⁷Hf isotopic composition ranging between almost zero to approximately 0.25 (Belousova et al. 2010; Fisher et al. 2011), and therefore, the interference correction must effectively cover such a range. In order to simulate different concentrations of Yb, and thus of different ¹⁷⁶Yb/¹⁷⁷Hf, we used the synthetic standard zircons of Fisher et al. (2011), which are artificially doped with variable amounts of REEs, ranging between zero (Zirc 0, not doped) and 0.35 (Zirc 4). This offers an even wider range than what is usually observed in natural zircons and standard zircons, which only rarely exceed 0.1 (data re-compilation in Fisher et al. 2011; see also Fig. 9). The same synthetic standard zircons are measured interdispersed between the unknown zircons and were used to calculate a normalization factor to the value of 0.282139, which is then applied to the unknowns. In general, the absolute value of our ¹⁷⁶Yb/¹⁷⁷Hf isotopic ratio is within ±0.000025 (2 SE).

Fig. 9

¹⁷⁶Hf/¹⁷⁷Hf versus ¹⁷⁶Yb/¹⁷⁷Hf plot of MUN synthetic zircons (Fisher et al. 2011) obtained during Hf isotopic analysis with laser ablation

All the standard zircons (91500, Temora, Plešovice, R33, FC01) that we analyzed during the last 8 months indicate that the measured and corrected ¹⁷⁶Hf/¹⁷⁷Hf ratios match their respective accepted values (Table 2).

Table 2 ¹⁷⁶Hf/¹⁷⁷Hf ratios of analyzed natural standard zircons and their respective accepted values

The ¹⁷⁶Lu decay constant of 1.876 × 10⁻¹¹ (Scherer et al. 2001) as well as the chondritic Hf values of Bouvier et al. (2008) were adopted for the calculations of εHf and model ages (T _DM). The εHf and T _DM calculated employing the chondritic and depleted mantle values, respectively, are generally considered as reflecting minimum ages for the zircon’s host magma source. A second model age (β _DMC) is then calculated, using a mean crustal value of ¹⁷⁶Lu/¹⁷⁷Hf = 0.015 (Griffin et al. 2002) and projecting the initial εHf of zircons to the depleted mantle model curve.