Sphene (Titanite) as Both Monitor and Driver of Evolution of Felsic Magma: Miocene Volcanic Plutonic and Rocks of the Colorado River Region, NV-AZ, USA
Sphene is commonly the most abundant accessory mineral in metaluminous to weakly peraluminous igneous rocks. Its relatively large crystals preserve a wide array of zoning patterns and inclusions - notably, abundant other accessories and melt inclusions - and it is a major host for REE, U, Th, and HFSE. Thus it is a valuable repository of information about the history of the magmas from which it forms. Recent development of a Zr-in- sphene thermometer (Hayden et al CMP 155:529 2008) and of sensitive and precise in situ trace element analysis by SHRIMP-RG (Mazdab et al GSA abst 39:6:406 2007) permit more powerful exploitation of this repository. We have initiated a study of sphene in Miocene intrusive and extrusive rocks of the Colorado River extensional corridor for which extensive field, geochemical, and geochronological data provide context. Sphene is present as a late interstitial phase in some gabbros and diorites and common in quartz monzonites and granites. Among extrusive rocks, it occurs as phenocrysts in rhyolite lavas and tuffs that are products of small to giant eruptions (Peach Spring Tuff, >600 km3). Glasses that host sphene in the rhyolites are highly evolved (>76 wt% SiO2). Applying the Zr-in-sphene thermometer (TZr), SHRIMP-RG analyses indicate crystallization T between 730 and 810 C in both plutonic and volcanic rocks. This range is narrower than T estimates for zircon growth (Ti thermometry) for the same suite, which extend to somewhat lower and considerably higher values; zircons also tend to record more events and, evidently, longer histories. Ranges of REE patterns are variable and to some extent sample-specific, but all reveal common characteristics: (1) extremely high concentrations, especially for middle REE (maximum Sm in interiors 10-40x103 x chondrite); (2) deep negative Eu anomalies (Eu/Eu* ca. 0.1-0.2); (3) TZr and REE dropping toward rims - especially pronounced for MREE. Estimated Kds for REE from sphene rims and rhyolite glass or phenocryst- poor whole rocks are very high, especially for middle REE: LREE Kds ca. 50-100, MREE ~500-600 (Eu ca. 300- 400), HREE ca. 100. Late REE fractionation trends that are evident in both plutonic and volcanic sequences are clearly controlled for the most part by sphene: aplites, some leucogranites, and high-Si rhyolite whole rocks and glasses reveal extreme MREE depletion and suppressed development of Eu anomalies, a trend that is also expressed in core-to-rim REE depletion patterns in sphene crystals. Results suggest that sphene saturation in these magmas occurred in melts that were already evolved but that it had a dramatic effect on final stages of fractionation. The sphene 'fingerprint' is similar to that proposed by Glazner et al. (Geology 36:183 2008) for Sierra Nevada aplites and as they suggest it marks a late-stage process, but in contrast to their inference we demonstrate that it is evident in volcanic as well as comagmatic plutonic rocks. A better understanding of the controls of sphene saturation will lead to refined interpretation of its presence (or absence), onset of growth, and geochemical fingerprint with respect to magmatic-tectonic environments (cf. Bachmann & Bergantz JPet 49:2277 2008). We intend to address these issues further with saturation experiments and tomographic and geochemical studies of sphene and its inclusions and associated phases.
Improved Methods for U/Pb Geochronology of Baddeleyite by LA-ICPMS
The most popular mineral for uranium-lead geochronology is zircon (ZrSiO4) but for many rocks of interest, particularly those with basic to ultrabasic compositions, zircon is not present and baddeleyite (ZrO2) is the main alternative. Microbeam analyses by laser ablation-inductively coupled plasma mass spectrometry (LA- ICPMS) are increasingly used for applications in the Earth sciences where high precision data are not required. However, while zircon geochronology by LA-ICPMS has expanded rapidly in use over the past decade, baddeleyite geochronology has lagged far behind. This has been due in part to the lack of well- characterized baddeleyite standard reference materials for microbeam analyses, but even more importantly, because of difficulties caused by severe and variable laser-induced Pb/U fractionation during ablation of spots in the mineral. We have compared the extent and consistency of time-dependent fractionation of Pb/U ratios in baddeleyite as a function of ablation conditions. All experiments were made using a Finnigan ELEMENT-XR magnetic sector ICPMS coupled to a GeoLas 193 nm ArF excimer laser ablation system. The baddeleyites studied included ca. 380 Ma Kovdor (Russia), ca. 1100 Ma Forest Center (Minnesota), and ca. 2060 Ma Phalaborwa (South Africa). Using a 40-micron spot, 3 J/cm2, and 5 Hz, the 206Pb/238U ratios in the baddeleyites show very large increases of approximately 80% over 150 pulses (30 sec) of ablation. In comparison 206Pb/238U ratios in zircon under the same ablation conditions increase by only approximately 15%. The spot to spot variability in the size of the Pb/U fractionation is also much greater in baddeleyite (approximately 10%RSD) than in zircon (approximately 2 to 3%RSD). This leads to imprecise and inaccurate ages for baddeleyite when using spot analyses even with matrix matched calibration standards. In contrast, by ablating with line scans, made with a 20 micron spot, and moved over the sample surface at 1 micron/sec at 5 Hz and 2 to 5 J/cm2, Pb/U fractionation in baddeleyite is reduced significantly. Over 150 pulses (30 sec) of ablation, 206Pb/238U ratios increase by less than 15% with less than 5% RSD variation from line scan to line scan. Using Forest Center baddeleyite as the calibration standard, we find U-Pb concordia ages of 377+/-8 Ma for Kovdor, and 2053+/-12 Ma for Phalaborwa.
Accessory Mineral Geochronology and Trace Element Fingerprinting of Metamorphic Processes
Zircon and monazite are the most versatile tools for geochronological studies in magmatic, metamorphic and sedimentary rocks. New trace element techniques are now used to link growth and modification to pressure and temperature evolution of magmatic and metamorphic rocks and to coexisting minerals. Studies on the distribution of rare earth elements (REE) have mainly focused on garnet and zircon relationships and which distribution patterns constitute equilibrium under different metamorphic conditions. But a more detailed understanding of the growth and modification processes of accessory phases is needed to provide better constraints for genetic models and multiple method datasets (U-Pb, REE, trace element thermometry, imaging) is essential. We use this approach on examples from high-temperature low-pressure granulites of Rogaland (SW Norway) and UHT medium pressure granulites from the Labwor Hills (Uganda) to illustrate the influence of major and accessory mineral reactions on the trace element signature of zircon and monazite. Relatively flat zircon HREE patterns, often associated with coexisting garnet, can also be found in orthopyroxene-bearing, garnet free assemblages. The zircon-opx distribution patterns are similar to zircon-garnet pairs from UHT leucosomes and granulites. Some garnet-bearing granulites are characterized by zircon overgrowths with xenotime inclusions and elevated Y- and HREE-contents, interpreted to reflect garnet breakdown at high T. Zircon and monazite in Labwor Hills metasedimentary granulites both have modified domains. Monazite has low Th rims that yield erroneous high chemical ages, interpreted to be due to loss of Th, while remnants of radiogenic Pb remain during the recrystallization process. Zircon in contrast has high Th rims and domains along sealed cracks that are characterized by anomalously high Th/U ratios (not previously associated with metamorphic growth or modification) and unusual REE patterns. These features are interpreted to result from fluid-influenced recrystallization that lead to coeval leaching of Th from monazite and precipitation of high Th zircon. These results can also help reevaluate the interpretation of detrital zircon, i.e. the practice of invoking metamorphic zircon growth based only on low Th/U or granulite conditions on flat HREE patterns should be abandoned.
Lead Isotopic Composition of Melt Inclusions From the Bishop Tuff, Long Valley, California
The 760,000 year old Bishop Tuff of the Long Valley Caldera in California is a classic large-volume, high-silica rhyolite ash flow tuff. Its major element composition is remarkably homogeneous (75.5 to 77.6 wt percent SiO2) but trace element concentrations are strongly zoned from early to late in the eruptive sequence. The origin of the magmatic system has been variously ascribed to magma mixing of crust and mantle melt components but the nature of these components and their distribution within the plumbing system are poorly defined. We have analysed lead isotopes in 29 melt inclusions in quartz, and plagioclase phenocrysts, from a single sample of the early plinian phase of the Bishop Tuff collected from the Chalfant Quarry near Bishop, CA, using a 193 nm excimer laser ablation (LA)-multiple ion counter (MIC)-inductively coupled plasma-mass spectrometry (ICPMS). The MIC array on the NEPTUNE ICPMS instrument allowed, for the first time, simultaneous collection of all four lead isotopes (204Pb, 206Pb, 207Pb, 208Pb) at total lead concentrations levels of only 25 to 30 ppm, in high-silica rhyolite melt inclusions. The melt inclusions were polished and exposed at the surface of the host mineral and ablated with spot sizes of either 30 or 40 microns. Measurements of lead isotopes in the plagioclase phenocrysts were collected under similar conditions. The melt inclusion and plagioclase data define lead isotope mixing arrays between the upper mantle and old upper crust sources of Kramers and Tolstikhin (1997). Preservation of lead isotope heterogeneities within the single crystals of quartz and plagioclase grains suggests that melts from discrete mantle and upper crustal sources coexisted at early stages of Bishop Tuff magmatic evolution and only became homogenized during or just prior to eruption.
Construction and Evolution of the Mount St. Helens Magmatic System During the Swift Creek Eruptive Stage (16-9 ka) Revealed by Zircon
U-series geochronology and trace element analyses of zircon record the evolution and construction of the sub- volcanic magmatic system of Mount St. Helens during its Swift Creek eruptive stage (16-9 ka). This timeframe was characterized by episodic eruptions of relatively cool, wet, and evolved lavas and tephras followed by emplacement of hotter, drier, and less evolved eruptive products. These fluctuations between magma types potentially represent the first evidence of well-developed magmatic cycles within the Mount St. Helens plumbing system (Clynne et al., in press). We compare the geochronology and geochemistry of zircons from Swift Creek rock samples to those from samples that span the rest of the eruptive history (Claiborne et al., 2008). U-Th age spectra demonstrate that zircon within Swift Creek rocks predominantly crystallized between 20 and 80 ka (∼70% of analyses), with crystallization peaks at ∼30 and 55 ka. Minor populations of ages are also present at ∼105, 160, 210, and 250 ka (∼25% of analyses). Most crystallization ages range from tens to hundreds of thousands of years before eruption, indicating that Swift Creek magmas extracted crystals from previous episodes of crystallization. However, some zircon analyses (<10%) yield U-Th ages within error of eruption age, potentially allowing us to track Swift Creek magmatic evolution from its incipient stages to eruption. Both of these observations are consistent with previously documented zircon populations from samples spanning the eruptive history of the volcano (Claiborne et al., 2008). Although some age populations observed in zircons from early (300-250 ka) and late (160-35 ka) Ape Canyon eruptive stage rock samples reappear in zircons from Swift Creek samples, the major populations from these earlier eruptive episodes are absent or sparse. Conversely, zircons from the Cougar stage (28-18 ka) exhibit comparable age peaks to Swift Creek zircons. These observations suggest that Swift Creek and Cougar magmas sampled similar crystal reservoirs and had little, if any, interaction with reservoirs affected by Ape Canyon magmatism. Preliminary application of the Ti-in-zircon thermometer (Ferry and Watson, 2007) yields zircon crystallization temperatures below the eruption temperatures of the final host magmas for reasonable SiO2 and TiO2 activity values. Each rock sample also exhibits wide ranges in zircon trace element concentrations (Hf, U, Th, and REEs) that strongly overlap with those of all other samples. Similar compositional ranges are observed between chemical zones of individual zircons. These findings are consistent with previous geochemical results and, together with U-Th age spectra, point to repeated intrusion of new magma batches, cooling, and production of crystal reservoirs beneath the volcano. They suggest that Swift Creek zircons have grown from magmas of variable composition that repeatedly recycle crystals from existing reservoirs. Furthermore, several potential fractionation indicators suggest that older zircon populations grew from more evolved melts than later populations. In comparison to the 105, 160, 210, and 250 ka crystal populations, the 30 and 55 ka populations have lower Hf, higher Th/U, and lower Yb/Gd. This decreasing fractionation with time suggests that the onset of well-developed magmatic cycles at Mount St. Helens during the Swift Creek stage coincides with input of less evolved melt into the magmatic system.
Synthetic zircon doped with Hf-Lu and Yb as potential reference materials for in situ microanalysis
Preliminary results are reported for a pilot study initiated to test the potential use of synthetic zircon (ZrSiO4) doped with hafnium (Hf) and varying amounts of ytterbium (Yb) and lutetium (Lu), as a Hf isotopic reference material for analysis of natural zircon crystals. The micron-scale Hf isotopic homogeneity in these (500-1000 microns) synthetic zircon crystals occurs at both the intra- and inter-grain scale levels. As such, these materials show high potential as standards for in situ microanalytical mass spectrometric techniques like secondary ion mass spectrometry (SIMS) and laser ablation-multicollector-inductively coupled plasma mass spectrometry (LA-MC-ICPMS) of zircon. The isotopic homogeneity of these synthetic zircon crystals can be utilized to monitor the complex, and sometimes large corrections necessary for in situ microanalytical techniques, where a chemical separation to remove the significant isobaric inferences is not performed. The major obstacles in doing accurate in situ measurement of 176Hf/177Hf in zircon are the corrections for instrumental mass bias and the unavoidable 176Yb and 176Lu isobaric interference on 176Hf. To monitor the effectiveness of these corrections, a set of Hf + Yb +Lu doped synthetic zircon crystals were grown with 176Hf representing from 96% to 60% of the total 176 amu signal (176Yb/177Hf = 0.01-0.22, 176Lu/177Hf= 0.0004-0.0055), covering the range expected in natural zircon samples. Due to intrinsic chemical zoning in the large flux-grown synthetic zircon crystals, the Hf elemental concentrations vary by ~15% from core (mean 12000 ppm) to rim (mean 10000 ppm) and 176Yb/177Hf varies up to ~30% in some grains. This variation in Hf and Yb concentration allows a full range of possible zircon compositions to be monitored in an analytical session, with a single homogeneous Hf isotopic composition, thereby closely approximating the full range of natural zircon crystals that may be encountered in Hf tracer isotope studies of zircon.
Linking Magmatism and Mineralization using In-Situ Nd- and Sr-isotope Systematics of Vein-Hosted Apatite at the Morila Au Mine, Mali
At the Morila Au Mine, Mali, auriferous arsenopyrite, scheelite, and alloys of Au-Sb and Bi-Te occur in polymineralic veins with timing relationships that match those of nearby calc-alkaline plutons. Although an intrusion-related origin is implicated spatially and temporally, robust isotopic links between alteration and ore deposition and the adjacent 'smoking gun' pluton have yet to be established. Previous attempts to fingerprint the mineralizing fluids at Morila using S-isotopes for sulfides in host metasediments (R. Quick, unpublished data) yielded an equivocal range of values between +2 to +6 permil (relative to CDT) reflecting mixing and isotopic re-equilibration during the growth of Aspy at the expense of pyrrhotite within a tortuous micro-porosity network. As an alternative, radiogenic isotope systematics for mineralized veins that contain accessory minerals amenable to in-situ isotope measurements can be used to help minimize potential homogenization of isotope signatures. At Morila, LREE- and Sr-bearing apatite occurs in calc-alkaline intrusions and mineralized polymineralic veins. This allows us to compare the isotopic signature of mineralized veins to that of nearby plutonic rocks. Three distinct vein assemblages were targeted for in-situ work: 1) scheelite-bearing veins hosted by granodiorite stocks, 2) Au-Loellingite-Aspy-bearing veins in the upper footwall of the deposit, and 3) late-stage peraluminous granitic veins distal to the ore body. Apatite was analyzed by LA-MC-ICPMS in standard thin sections thereby providing complete control on the textural context of each grain. The results, calculated at t = 2095 Ma, yield a narrow range of epsilon-Nd values between -1 to +2 but a wide range of initial 87Sr/86Sr between 0.7015 to >0.7030. Whereas the Nd-isotope data overlaps with the full range of intrusive rocks at Morila and with the host metasediments, the bulk of the Sr-isotope data for the veins matches the isotopic signature of late-stage granites. Integrated with macro- and microscopic evidence for AFC processes prior to granite emplacement, this in-situ Nd- and Sr-isotope dataset strengthens the genetic links between magmatism and mineralization and highlights the importance of magma hybridization in intrusion-related Au deposits.