Spatially controlled Fe and Si isotope variations: an alternative view on the formation of the Torres del Paine pluton

Abstract

We present new Fe and Si isotope ratio data for the Torres del Paine igneous complex in southern Chile. The multi-composition pluton consists of an approximately 1 km vertical exposure of homogenous granite overlying a contemporaneous 250-m-thick mafic gabbro suite. This first-of-its-kind spatially dependent Fe and Si isotope investigation of a convergent margin-related pluton aims to understand the nature of granite and silicic igneous rock formation. Results collected by MC-ICP-MS show a trend of increasing δ⁵⁶Fe and δ³⁰Si with increasing silica content as well as a systematic increase in δ⁵⁶Fe away from the mafic base of the pluton. The marginal Torres del Paine granites have heavier Fe isotope signatures (δ⁵⁶Fe = +0.25 ± 0.02 2se) compared to granites found in the interior pluton (δ⁵⁶Fe = +0.17 ± 0.02 2se). Cerro Toro country rock values are isotopically light in both Fe and Si isotopic systems (δ⁵⁶Fe = +0.05 ± 0.02 ‰; δ³⁰Si = −0.38 ± 0.07 ‰). The variations in the Fe and Si isotopic data cannot be accounted for by local assimilation of the wall rocks, in situ fractional crystallization, late-stage fluid exsolution or some combination of these processes. Instead, we conclude that thermal diffusion or source magma variation is the most likely process producing Fe isotope ratio variations in the Torres del Paine pluton.

Appendix

The samples (0.5–1-kg specimens) were crushed and powdered using a W–C mill for bulk rock compositional and isotopic analysis. Major element compositions were determined on fused sample glasses (2:1 lithium tetraborate flux to sample ratio) using standards-based EDS X-ray analysis on the UIUC JEOL 840A scanning electron microscope. SEM–EDS analyses were automatically normalized to 100 %. Additional rock powders were obtained from Peter Michael from the University of Tulsa with major and trace element compositions published in Michael (1984, 1991). Trace element compositions were determined using laser ablation on fused sample glasses on the University of Illinois Thermo ICAP-Q ICPMS. We performed 65-µm-wide line transects on fused glasses of approximately 1-min duration using a Photon Machines 193-nm excimer laser. Ablated samples were carried with He and mixed with Ar prior to entrance into the ICAP-Q ICPMS. Dwell times on isotopes were 30 ms with ²⁹Si used as the internal standard. Data were processed using the Iolite software suite and corrected to SiO₂ values determined by EDS X-ray analysis. Accuracy is indicated by results for the USGS standards RGM-1, AGV-1, BCR-2. A total of 2–3 glass analyses of each sample and standard were performed with precision indicated by standard deviation on the analyses. Generally, we estimate ± 5 % precision on trace element concentrations. Elemental data for Torres del Paine samples are listed in Table 4.

Table 4 Bulk oxide concentrations for the Torres del Paine sample set

For Fe isotopic analysis, 62 samples were dissolved in closed Savillex beakers at 140 °C using HF and HNO₃. These were dried down, treated with concentrated HNO₃ and HCl and dried down once more. Samples were brought up in 0.6 ml 8 N HCl and put through 0.5-ml columns containing AG1-X8 anion exchange resin (Huang et al. 2011). Iron was eluted using 8 N HNO₃. Fe isotope analysis was performed on the University of Illinois Nu Plasma HR-MC-ICP-MS in high-resolution mode using a ⁵⁷Fe–⁵⁸Fe double spike technique involving separate Cr and Ni corrections. Analysis was performed using a 100 µl min⁻¹ nebulizer and a DSN-100 desolvating nebulizer. Resolution (M/ΔM) on ⁵⁶Fe was ~9000. Primary reference material IRMM-14 Fe standard was used for bracketing each analysis. The standard error on all IRMM-14 analyses was 0.02 2se showing little drift. Standard reference material BCR-2, AGV-2, RGM-2, NOD-P and UIFe were used as secondary standards. Offset and precision for standards are listed in Table 2. Precision is reported in 2σ and 2se. Per mil values were calculated using Eq. 1:

$$\updelta^{56} {\text{Fe}}_{\text{IRMM-14}} = \frac{{\left({{\text{Ratio}}\;{\text{sample}} - {\text{Ratio}}\;{\text{standard}}} \right)}}{{\left({{\text{Ratio}}\;{\text{standard}}} \right)}} \times 1000\,\permil$$

(1)

Nine samples were prepared for Si isotope analysis following the chemical procedure outlined in Zambardi and Poitrasson (2011). Sample (5 mg) and NaOH (250 mg) were weighed out into a silver crucible and fused in a furnace at 720 °C for 10 min. Fusion cakes were dissolved in H₂O and acidified to pH 1.5. Sample solutions were loaded on to AG 50 W-X12 cation exchange resin which allowed positively charged cations bond to the resin and neutral Si to pass through the columns. The columns were rinsed with H₂O to recover all of the Si. Silicon isotope analysis was performed at University of Science and Technology of China using a Neptune Plus MC-ICP-MS. Standard reference material NBS-28 was used for sample bracketing. The standard error on all NBS analyses was 0.05 2se showing little drift. Samples were run six times, and standard reference material BCR-2 was run 37 times. Offset and precision for samples and standards are listed in Table 3. Per mil values were calculated using Eq. 2.

$$\updelta^{30} {\text{Si}}_{\text{NBS - 28}} = \frac{{\left({{\text{Ratio}}\;{\text{sample}} - {\text{Ratio}}\;{\text{standard}}} \right)}}{{\left({{\text{Ratio}}\;{\text{standard}}} \right)}} \times 1000\,\permil$$

(2)

Twenty-three samples were prepared for Pb isotope analysis following methods in Gladu and Kamber (2008). Fifty milligrams of sample were dissolved in closed Savillex beakers at 140 °C using concentrated HF and HNO₃. The samples were dried down, attacked with concentrated HCl and HNO₃, brought up in 0.5 ml of 0.5 N HBr and loaded onto AG1-X8 anion exchange resin. Pb ions were eluted in 10.5 N HCl. Lead analysis was performed on the University of Illinois Nu Plasma MC-ICP-MS in low-resolution mode using admixed Tl. Analysis was performed using a 100 µl min⁻¹ nebulizer. Primary reference material SRM981 Pb isotopic standard was run every three samples with precision, and BCR-2 was used as a secondary standard and interspersed with samples. The value of these standards is listed in Table 1. Average precision on the ²⁰⁸Pb/²⁰⁴ Pb ratio was ±0.0044. Age corrections were only performed on aplites, as the corrected ages for the remainder of samples result in less than 0.08 % change.

For Sr analysis, 43 samples were dissolved in closed Savillex beakers at 140 °C using HF and HNO₃. These were dried down, treated with concentrated HNO₃ and HCl steps, brought up in 3 N HNO₃ and put through Sr Spec anion exchange resin. Sr analysis was performed on the University of Illinois Nu Plasma MC-ICP-MS in low-resolution mode. Analysis was performed using a 100 µl min⁻¹ nebulizer aspirated into a DSN-100 desolvator. SRM987 was run every three samples as a primary reference material. E&A and an in-house modern Coral solution were interspersed with samples. The measured SRM987 could differ from true by up to 0.0001 and would drift systematically through a run session; this offset from true was applied to all samples and the E&A and Coral results. The corrected ⁸⁷Sr/⁸⁶Sr values for Coral and E&A agree with known values indicating accuracy of this procedure (Table 1). Long-term reproducibility on offset-corrected E&A or Coral indicates precision of ±0.00002. Age corrections were performed on all samples.