Promise and Pitfalls of Lu/Hf-Sm/Nd Garnet Geochronology

Our ability to routinely measure Lu-Hf and Sm-Nd isotopes in garnet allows broad new applications in geochronology, petrology, and tectonics. However, applications of these data can be limited by challenges in interpreting the petrologic record and preparing garnets for analysis. Here, we examine petrologic and chemical pitfalls encountered in garnet geochronology. Petrologic factors influencing trace element compositions in garnet include reactions that modify REE availability and partitioning (1,2), kinetically limited transfer of REEs to garnet (3), and bulk compositional heterogeneities (4). Interpreting the effects of these processes on Sm/Nd and Lu/Hf ages requires characterizing REE zonation prior to isotope analysis and age interpretation. Because garnet fractions are traditionally picked from crushed samples without regard to intracrystalline origins or chemistries, isochrons will represent mixtures derived to varying degrees from all periods of garnet growth. While measured zoning might generally indicate what garnet portion dominates the Lu/Hf or Sm/Nd budget, traditional mineral separation will rarely realize the chronologic potential afforded by high precision Hf and Nd isotope measurements. The potential use of alternative techniques, such as microsampling, necessitates selective digestion and/or leaching to eliminate inclusions within garnet. For Sm/Nd geochronology, H2SO4 leaching removes LREE-rich phosphates (e.g. apatite), but not silicates (e.g. epidote), precluding Sm-Nd dating of some rocks. For Lu/Hf geochronology, ubiquitous zircon microinclusions (c. 1 μm) can significantly disrupt age determinations. Microinclusions cannot be detected optically or separated physically, requiring selective chemical digestion. If complete digestion methods, such as bomb digestion, are used for garnet fractions, then "common Hf" from zircon will be contained in final solutions. These mixed analyses are of dubious utility and will fall into one of two categories based upon inclusion reaction history (5). In "best case" scenarios, zircon will be quasi-co-genetic with garnet formation and all phases will reflect the same Hf pool available during metamorphism. In this case, these "garnet" fractions may retain age information, but will yield underestimated Lu/Hf ratios that severely degrade isochron precision. However, "worst case" scenarios for these mixed-phase analyses will occur when zircon is inherited from prior events, such that garnet analyses represent two, unequilibrated Hf pools. Here, Lu/Hf isotopic analyses do not yield accurate age information on garnet growth as individual isotopic analyses reflect several, unrelated petrologic events. To avoid systematic introduction of errors of this type due to improper digestion procedures, garnet dissolutions should occur via low-pressure, hot-plate style digestion in Teflon beakers. Here, chemical conditions are unlikely to incorporate significant zircon-derived Hf within final solutions, ensuring that Lu/Hf ratios primarily reflect garnet compositions. 1. King et al., 2004, Geochem. Geophys. Geosys. 10.1029/ 2004GC000746; 2.Corrie and Kohn, 2008, J. Metam. Geol. in press; 3.Skora et al., 2006, Contr. Min. Pet. 152, 703-720; 4.Carlson, 2002, Amer.Mineral. 87, 185-204; 3. 5.Scherer et al., 2000, Geochim. Cosmochim. Acta 64, 3413-3432.