Fractionation of Nb and Ta by biotite and phengite: Implications for the “missing Nb paradox”

Fractionation of Nb and Ta by biotite and phengite: Implications for the “missing Nb paradox” Aleksandr S. Stepanov and Jörg Hermann Research School of Earth Sciences, The Australian National University, Canberra, ACT 0200, Australia ABSTRACT The subchondritic Nb/Ta in both the continental crust and the depleted mantle remains enigmatic and is called the “missing Nb paradox.” We present partitioning data between biotite and granitic melt for experimental and natural samples that provide evidence that Nb is compatible in biotite and phengite. Nb can thus be enriched in the residue during partial melting of crustal rocks. Additionally, biotite and phengite in equilibrium with granitic melts preferentially incorporate Nb over Ta. Therefore incipient partial melting of biotite-rich crustal rocks produces restites with high Nb/Ta. Progressive melting of such rocks leads to the consumption of biotite and the formation of peritectic rutile or ilmenite, which retain the high-Nb/Ta signature. We suggest that such mid to lower crustal granulites could represent an important Nb-rich reservoir with high Nb/Ta. Similarly, high-Ti phengite that is present in deeply subducted sediments preferentially incorporates Nb over Ta. High-pressure incipient partial melting in the presence of residual phengite thus also produces restites with high Nb/ Ta that could be subducted to the deeper mantle. GEOLOGY, March 2013; v. 41; no. 3; p. 303–306 | hypotheses cannot explain the existence of rocks with high Nb/Ta ratios, such as the documented mantle eclogites (Rudnick et al., 2000). In this paper, we present new experimental partition coefficients and compile data from previous studies of natural rocks that highlight the potential of biotite and Ti-rich phengite to fractionate Nb from Ta during partial melting of crustal rocks. We propose that the mid to lower crustal residues after melting of K-rich rocks might represent a high-Nb/Ta reservoir that has been overlooked previously. MINERAL-MELT NB-TA FRACTIONATION Two key factors determine the redistribution of Nb and Ta during partial melting. First, the Nb mineral-melt and Ta partition coefficients (DNb or DNb, and DTamineral-melt or DTa, respectively) and modal abundances of the minerals present determine whether Nb and Ta are retained in the residue Figure 1. Partition coefficients for Nb and Ta for biotite (Bt) and phengite (Phe) from this study (TS), Nash and Crecraft (1985) (NC), and AcostaVigil et al. (2010) (AV). Ol— olivine (Dunn and Sen, 1994). For comparison, the rock-forming and accessory minerals are also shown. Amph— amphibole (Tiepolo et al., 2000; Fulmer et al., 2010; Adam and Green, 2006); Cpx—clinopyroxDNb ene (Stalder et al., 1998; Adam and Green, 2006); Grt—garnet (Stalder et al., 1998; Fulmer et al., 2010); Rt—rutile (Schmidt et al., 2004; Xiong et al., 2011); Tnt—titanite (Prowatke and Klemme, 2005); Ilm—ilmenite (Xiong et al., 2011); Phlog—phlogopite (Adam and Green, 2006). doi:10.1130/G33781.1 DNb/DTa INTRODUCTION Nb and Ta are elements regarded as geochemical twins showing very similar properties, and they display a strong affinity to Ti minerals. Therefore, it is surprising that significant fractionation of Nb and Ta are observed at a global scale. The Nb/Ta of the depleted mantle (11–16) and the continental crust (8–14) is lower than that of bulk Earth, as estimated from the Nb/ Ta values of 18–20 for chondrite meteorites (Jochum et al., 2000; Münker et al., 2003, and references therein). This apparent deficit of Nb has been coined the “missing Nb paradox”, and there have been several attempts to identify reservoirs with superchondritic Nb/Ta to solve this imbalance. Rutile-bearing mantle eclogites display high Nb content and often have high Nb/ Ta and Nb/La (Rudnick et al., 2000). Based on these observations, it has been suggested that partial melting of eclogites in the presence of rutile might be a viable process to fractionate Nb from Ta. However, rutile is typically the major host for Nb and Ta in eclogites and preferentially incorporates Ta over Nb (Schmidt et al., 2004; Prowatke and Klemme, 2005; Xiong et al., 2011). Therefore melting in the presence of rutile should result in restites with low Nb/Ta, opposite to what is necessary for the explanation of the Nb paradox. Metasomatic modification of mantle eclogites by fluids has been proposed as a process to generate eclogites with high Nb/Ta, though the origin, nature, and composition of such fluids is unknown (Aulbach et al., 2008). Alternative explanations invoked more radical solutions. It has been suggested that the missing Nb may reside in the core of Earth (Wade and Wood, 2001), that Hadean crust with high Nb/ Ta was subducted to the mantle (Nebel et al., 2010), or that Earth has a nonchondritic Nb/Ta (Campbell and O’Neill, 2012). However, these or are enriched in the partial melt. The very mineral-melt low DNb and DTamineral-melt for mantle minerals such as olivine, garnet, and clinopyroxene (Fig. 1) explain why Nb and Ta are enriched in the crust. On the other hand, Ti phases such as mineral-melt rutile, ilmenite, and titanite have high DNb , and a relatively small amount of such a phase is sufficient to make Nb and Ta compatible in the residue (Fig. 1). The second factor relates to the ability of a mineral to fractionate Nb from mineral-melt Ta, which is expressed as DNb /DTamineral-melt. mineral-melt All the Ti phases that have high DNb and DTamineral-melt, such as rutile, titanite, and ilmenite, preferentially incorporate Ta over Nb, resulting in restites that have a lower or equal Nb/Ta than the protolith and coexisting partial melts with low Nb content and high Nb/Ta (Fig. 1). Amphibole was considered as the only rock-forming mineral that has been shown to have DNb > DTa at certain conditions and compositions (Tiepolo et al., 2000). However, the fractionation of Nb from Ta is rather small and Nb is mostly incompatible in amphibole (DNb < 1 for the majority of experiments; only one experiment by Tiepolo et al. [2000] reported DNb = 1.63). Thus it is unlikely that partial melting in the presence of amphibole alone produces restites with superchondritic Nb/Ta and higher Nb content than the protolith. We have analyzed Nb and Ta partitioning between biotite and melt formed at 2.5 GPa, 750 °C and 800 °C, and between high-Ti phengite and melt at 4.5 GPa, 900 °C and 1000 °C (Table 1), in experiments documented in detail by the previous studies of Hermann and Spandler (2008) and Hermann and Rubatto (2009). A short description of Experiment C-1505, which | Published online 25 January 2013 © GEOLOGY 2013 Geological Society 2013 of| www.gsapubs.org America. For permission to copy, contact Copyright Permissions, GSA, or editing@geosociety.org. | March 303 TABLE 1. Nb AND Ta CONCENTRATIONS (IN ppm) AND Bt-MELT AND Phe-MELT PARTITION COEFFICIENTS IN OUR HIGH-PRESSURE EXPERIMENTS (TS), AND IN RHYODACITE/RHYOLITE FROM NASH AND CRECRAFT (1985) (NC) AND ACOSTA-VIGIL ET AL. (2010) (AV) Ref Sample Mineral Temperature (°C) Pressure (GPa) Nb (melt) Ta (melt) Nb (mica) Ta (mica) TiO2 (wt%) DNb DTa DNb/DTa TS TS TS TS NC NC NC AV AV AV C-2446 C-1846 C-1505 1563a 4 8 20 HO-50A HO-33A HO-54 Bt Bt Phe Phe Bt Bt Bt Bt Bt Bt 750 800 900 1000 2.5 2.5 4.5 4.5 775 865 700–850 700–850 700–850 0.6 0.6 0.6 38 26 145 63 19 13 19 9.5 9.1 13 27 21 103 57 1.4 1.8 1.6 1.3 1.1 1.3 19 50 23 84 68 118 93 83 84 64 4.3 19 5.7 26 1 .6 3.5 2.5 4.8 4.2 3.5 2.15 2.40 1.18 2.08 4.54 4.28 4.46 5.27 5.59 4.39 0.50 1.96 0.15 1.32 3 .5 8 9.08 4.89 8.71 9.21 4.85 0.16 0.91 0.06 0.45 1.18 1.91 1.56 3.64 3.93 2.64 3.2 2.2 2.8 2.9 3.0 4.8 3.1 2.4 2.3 1.8 Note: Nb and Ta concentrations from Acosta-Vigil et al. (2010) are corrected for new NIST values from Jochum et al. (2011). has not been published before, is given in the Appendix. We analyzed biotite, phengite, and glass from these four experimental runs by laser ablation–inductively coupled plasma–mass spectrometry (LA-ICP-MS) using the analytical procedure presented in the Appendix. Both biotite and phengite are found to preferentially incorpomica-melt rate Nb over Ta, such that DNb /DTamica-melt > 2 (Fig. 1). Our experimental data are in agreement with biotite-melt partitioning from dacite and Bt-melt rhyolite, which display DNb /DTaBt-melt of 1.8– Bt-melt 4.8, and high DNb of 3.6–9 (Nash and Crecraft, 1985; Acosta-Vigil et al., 2010) (Table 1). mica-melt DNb increases with increasing temperature Bt-melt (Table 1). The increase of DNb from 0.5 to 9 correlates with an increase from 2.1 to 5.6 wt% Bt-melt TiO2 in biotite. In contrast, DNb / DTaBt-melt is not correlated with TiO2 content of mica. While it is well known that biotite can accommodate high TiO2 content (4–5 wt%), phengite has not previously been considered an important reservoir for Ti and Nb. However, at ultrahigh-pressure (UHP) conditions, phengite is stable to high temperatures and, in our experiments, contains 2 wt% TiO2 when formed at 4.5 GPa and 1000 °C (Table 1). Our results provide evidence that in such Ti-rich micas, Nb is compatible during partial melting. While there are only a few studies that report mica-melt partitioning data for both Nb and Ta, there are additional studies that support that Nb is compatible in biotite in felsic magmas. Ewart and Griffin (1994) report Bt-melt DNb = 4.6 and 9.1 for low-Si and high-Si rhyolite, respectively. An interesting observation is that Nb is incompatible, and no Nb-Ta fractionation is observed, in phlogopite coexisting with a basaltic melt (Fig. 1) (Adam and Green, 2006). This suggests that Nb and Ta more readily enter the less polymerized mafic melts with respect to felsic melts as previously reported in the detailed study of Nb and Ta partitioning between amphibole and melt (Tiepolo et al., 2000). Therefore the sum of data demonstrates mica-melt that the scatter of DNb and DTamica-melt is related to a variety of parameters such as pressure, temperature, mica, and melt composition. The most 304 favorable conditions for Nb and Ta fractionation occur during high-temperature partial melting of crustal rocks. In the following section, we compile data from key localities of crustal anatexis to investigate whether intracrustal Nb and Ta fractionation does happen during partial melting of mica-bearing rocks. NB-TA FRACTIONATION DURING LOWER CRUSTAL MELTING Crustal enclaves from the El Hoyazo dacite, Spain, represent an excellent natural laboratory to study incipient partial melting at mid to lower levels of the continental crust (0.6 GPa, 700– 850 °C; Acosta-Vigil et al., 2010). The granulite facies xenoliths were rapidly cooled by the volcanic eruption, preserving the products of partial melting as quenched glasses. Acosta-Vigil et al. (2010) presented a complete data set of A trace element compositions of major and accessory minerals, melts, and melt inclusions in El Hoyazo enclaves and their host magma. The melt inclusions have significantly lower Nb and Nb/Ta than the bulk rock composition (Fig. 2A), indicating that Nb was compatible during partial melting and remained in the residue. Accessory ilmenite has high Nb content of 627–655 ppm, but its Nb/Ta of 10–12 is close to the bulk rock composition. Biotite is an abundant phase in the enclaves and has high Nb content of 63–81 ppm and Nb/Ta significantly higher than the bulk rock. The colinearity between biotite, bulk rock, and partial melt indicates that residual biotite rather than ilmenite was the phase controlling the Nb-Ta redistribution. Indeed, mass balance calculations (Acosta-Vigil et al., 2010) demonstrated that biotite is the main host of Nb and Ta in these rocks. B Figure 2. Nb-Ta fractionation during crustal anatexis. A: Compositions of phases in the enclaves from El Hoyazo, Spain (Acosta-Vigil et al., 2010). Low Nb/Ta of melts and melt inclusions are due to high Nb content and high Nb/Ta of biotite. Although ilmenite is present, it did not exert a strong influence on the Nb/Ta of the melt. Grt—garnet; Pl—plagioclase; Bt— biotite; Ilm—ilmenite. B: Bulk rock concentrations in metasedimentary rocks of the IvreaVerbano zone, Italy (Bea and Montero, 1999). Restites of granulite facies metapelites have higher Nb concentrations and higher Nb/Ta, whereas leucosomes have lower Nb concentrations and Nb/Ta than the amphibolite grade metasediments. The Nb/Ta of global oceanic subducted sediments (GLOSS; Plank and Langmuir, 1998), which is a good proxy for average upper crust, and of carbonaceous chondrites (Jochum et al., 2000), which is a proxy for global Earth, are given for comparison. www.gsapubs.org | March 2013 | GEOLOGY The further progress of partial melting can be investigated in the classical lower crustal section of the Ivrea-Verbano zone (IVZ) in Italy (0.6–0.8 GPa, 700–900 °C; Bea and Montero, 1999). Whole rock compositions of amphibolite facies metasediments (protoliths) have typical crustal Nb/Ta of 11–13 (Bea and Montero, 1999). In contrast, granulite facies equivalents (restites) have higher Nb/Ta of 18–20 (Fig. 2B). Leucosomes in amphibolite and granulite facies rocks mostly have low Nb/Ta of 3–6 and low Nb content of <10 ppm. The increase of Nb in granulite facies metasediments demonstrates that Nb was compatible during melting. As in El Hoyazo, biotite is the main host for Nb and Ta in medium-grade rocks in the IVZ (Luvizotto and Zack, 2009), and the observed relative enrichment of Nb over Ta can be explained by the presence of residual biotite during incipient partial melting. As melting progresses, the amount of biotite decreases from 40%–44% in amphibolite facies rocks to 0%–5% in granulites (Bea and Montero, 1999). During the gradual breakdown of biotite, peritectic rutile and more melt were formed. IMPLICATIONS FOR THE “MISSING NB PARADOX” Data from experiments and natural rocks provide clear evidence that biotite and Ti-rich phengite are able to fractionate Nb from Ta. We propose a two-stage process to explain the formation of reservoirs that have high Nb/Ta. Incipient partial melting in biotite-rich rocks produces restite with elevated Nb content and high Nb/Ta (Fig. 2). This process is most effective in rocks with small amounts of Ti phases, where biotite is the main host for Nb and Ta such as in the presented examples. Therefore, partial melting of rocks with elevated K and relatively low Ti content such as metapelites, graywackes, or granodiorites will produce the greatest Nb-Ta fractionation. Progressive melting of Ti-rich biotite results in the formation of peritectic rutile and ilmenite. At conditions relevant Rt-melt to crustal melting DNb /DTaRt-melt is close to unity or slightly lower (Xiong et al., 2011). Therefore the presence of residual rutile (or ilmenite) results in the preservation or slight lowering of the Nb/Ta of the restite. However, the effect is not big enough to counteract the previous fractionation imposed by biotite, as shown in the example of the IVZ granulites (Fig. 2B). More importantly, the presence of rutile or ilmenite in this second stage of melting provides a mechanism to retain Nb and Ta in the restite and thus preserve the signature acquired during the first stage of melting. The second stage of melting occurs at high temperatures, where the monazite solubility is greatly enhanced (Stepanov et al., 2012). If temperatures exceed 850–900 °C, restites will be characterized by low Th, U, and light rare earth element (LREE) contents GEOLOGY | March 2013 | www.gsapubs.org because of dissolution of monazite, but will not be significantly depleted in Nb because of peritectic rutile formation during biotite breakdown, resulting in rocks with superchondritic Nb/La and Nb/Ta. Therefore, restitic lower crust that underwent significant biotite melting could be a first potential high-Nb/Ta reservoir needed to explain the “missing Nb paradox”. Present estimates of the lower crust with a subchondritic Nb/Ta of 8.3 (Rudnick and Fountain, 1995) are mainly based on a lower crust dominated by mafic rock types as deduced from xenoliths. On the other hand, when lower crustal sections are exposed such as in the Alps, restitic granulites that derive from pelitic protoliths are common. Mineralogically, these restites are rich in garnet and aluminosilicate, resulting in densities that can exceed mantle values (Hermann et al., 1997). Hence, delamination of such restites from the base of the crust (Hacker et al., 2011) is a feasible process that would produce mantle heterogeneity with high-Nb/Ta and high-Nb/La characteristics. In El Hoyazo enclaves and IVZ granulites, Nb-Ta fractionations occurred in Phanerozoic metasedimentary rocks that already displayed subchondritic Nb/Ta ratios. Therefore, it is important to evaluate whether Nb-Ta fractionation by anatexis of crustal rocks is a modern process or was already operating during Proterozoic/Archean times. Recent, detailed analyses of Nb/Ta in Eoarchean and Mesoarchean tonalite-trondhjemite-granite (TTG) associations displayed highly variable Nb/Ta ratios of 7–27 (Hoffmann et al., 2011). Even larger fractionations were observed in Archean migmatitic TTGs where Nb/Ta ratios range from 14 to 42 with a strongly superchondritic average Nb/Ta of 29. Biotite is an important phase in metamorphosed tonalites. We speculate that residual biotite might play an important role in the observed intracrustal fractionation of Nb and Ta (Hoffmann et al., 2011) in TTGs, which experienced partial melting and melt extraction. Therefore intracrustal differentiation of Nb and Ta might have been active for a long time, and middle to lower crustal residues from old continental crust that derive from melting of biotite-bearing protoliths could represent another important missing high-Nb/Ta reservoir. Our new high-pressure experimental data show that high-Ti phengite is also able to fractionate Nb from Ta, and thus it is worth evaluating whether melting in subduction zones is able to produce restites with high Nb/Ta. Subducted K-rich sediments have 20%–40% phengite with <0.8 wt% bulk rock TiO2. As phengite is able to incorporate more than 2 wt% TiO2 at UHP conditions (Hermann and Spandler, 2008), only very small amounts of rutile are present prior to melting. During incipient melting, Nb will be preferentially retained in phengite and the restite will acquire high Nb/Ta ratios. The rutile-bearing residue produced after melting of phengite-rich metasediments represents a further potential high-Nb/Ta reservoir that can be subducted to deeper parts of the mantle. Occasionally, subduction-related magmas have high Nb/Ta ratios, high Nb, and high K. These magmas have previously been explained by complicated fractionation of Nb and Ta by amphiboles in the mantle during multiple stages of melting (Stolz et al., 1996; Koenig and Schuth, 2011). Our study suggests that phengite-bearing restites derived from incipient melting of metasediments at high pressure will have high Nb/Ta. The progressive melting of such sediments at deeper levels of subduction will produce a slab melt with high K, high Nb, and high Nb/Ta that might contribute to the composition of rare arc magmas. This represents an alternative way to produce arc rocks with high Nb/Ta. Phengitebearing rocks are not restricted to subducted sediments. Subduction of altered oceanic crust or alkali-basalts produces eclogites with variable amounts of phengite. These rocks generally also contain high TiO2 and thus rutile. It remains to be tested whether high-Ti phengite in these rocks is able to impose a Nb-Ta fractionation during incipient partial melting and whether this is a viable process to explain mafic eclogites within the mantle that show high Nb/Ta and high Nb/La (Rudnick et al., 2000). CONCLUSIONS The compatible behavior of Nb and the preferential incorporation of Nb over Ta in biotite during partial melting provide a mechanism to create mid to lower crustal rocks with high Nb/Ta. Similarly, high-Nb/Ta restites might be produced during partial melting of subducted sediments due to the similar behavior of these elements in phengite. Therefore we suggest that intracrustal differentiation might be an important process to produce Nb-rich rocks with superchondritic Nb/Ta that represent one of the missing reservoirs to balance the subchondritic Nb/ Ta of the upper crust and the depleted mantle. APPENDIX Experiment C-1505 used the same starting material as reported in Hermann and Rubatto (2009), with 6.2 wt% of H2O and additionally 9% of a carbonate mix consisting of 20% calcite and 80% dolomite. The experiment was run at 4.5 GPa and 900 °C for 117 h. This run produced large phengite flakes that are several hundred micrometers in size. Phengite coexists with garnet, clinopyroxene, coesite, glass, and accessory Mg-calcite; kyanite and rutile were also observed. Nb and Ta concentrations were analyzed by LAICP-MS at the Research School of Earth Sciences, The Australian National University, using a pulsed 193 nm Ar-F excimer laser with 100 mJ source energy at a repetition rate of 5 Hz, coupled to an Agilent 7500 quadrupole ICP mass spectrometer (Eggins et al., 1998). Laser sampling was performed in a He-ArH2 atmosphere using a spot diameter of 22 µm. Data acquisition was performed by peak hopping in pulse counting mode, acquiring individual intensity data for each element during each mass spectrometer sweep. 305 A total of 60 s, comprising a gas background of 20–25 s and 30–35 s signal, were acquired for each analysis. Nb and Ta concentrations were calculated with NIST standard reference material (SRM) 612 values (Jochum et al., 2011) as the external standard and SiO2 or CaO as the internal standard. Details of the analytical methods are from Hermann and Rubatto (2009). Nb and Ta concentrations from the work by AcostaVigil et al. (2010) reported in Table 1 were corrected for new NIST values. ACKNOWLEDGMENTS We thank Oliver Nebel, Tanya Ewing, and Daniela Rubatto for constructive comments on an early version; M. Tiepolo and M. Barth for journal reviews; and W. Collins for editorial handling, which helped to improve the paper. This work was financially supported by the Australian Research Council. REFERENCES CITED Acosta-Vigil, A., Buick, I., Hermann, J., Cesare, B., Rubatto, D., London, D., and Morgan, G.B., 2010, Mechanisms of crustal anatexis: A geochemical study of partially melted metapelitic enclaves and host dacite, SE Spain: Journal of Petrology, v. 51, p. 785–821, doi:10.1093 /petrology/egp095. Adam, J., and Green, T., 2006, Trace element partitioning between mica- and amphibole-bearing garnet lherzolite and hydrous basanitic melt: 1. Experimental results and the investigation of controls on partitioning behaviour: Contributions to Mineralogy and Petrology, v. 152, p. 1–17, doi:10.1007/s00410-006-0085-4. Aulbach, S., O’Reilly, S.Y., Griffin, W.L., and Pearson, N.J., 2008, Subcontinental lithospheric mantle origin of high niobium/tantalum ratios in eclogites: Nature Geoscience, v. 1, p. 468– 472, doi:10.1038/ngeo226. Bea, F., and Montero, P., 1999, Behavior of accessory phases and redistribution of Zr, REE, Y, Th, and U during metamorphism and partial melting of metapelites in the lower crust: An example from the Kinzigite Formation of IvreaVerbano, NW Italy: Geochimica et Cosmochimica Acta, v. 63, p. 1133–1153, doi:10.1016 /S0016-7037(98)00292-0. Campbell, I., and O’Neill, H., 2012, Evidence against a chondritic Earth: Nature, v. 483, p. 553–558. Dunn, T., and Sen, C., 1994, Mineral/matrix partition coefficients for orthopyroxene, plagioclase, and olivine in basaltic to andesitic systems—A combined analytical and experimental study: Geochimica et Cosmochimica Acta, v. 58, p. 717– 733, doi:10.1016/0016-7037(94)90501-0. Eggins, S., Rudnick, R., and McDonough, W., 1998, The composition of peridotites and their minerals: A laser-ablation ICP-MS study: Earth and Planetary Science Letters, v. 154, p. 53–71, doi:10.1016/S0012-821X(97)00195-7. Ewart, A., and Griffin, W.L., 1994, Application of proton-microprobe data to trace-element partitioning in volcanic rocks: Chemical Geology, v. 117, p. 251–284. Fulmer, E.C., Nebel, O., and van Westrenen, W., 2010, High-precision high field strength element partitioning between garnet, amphibole and alkaline melt from Kakanui, New Zealand: 306 Geochimica et Cosmochimica Acta, v. 74, p. 2741–2759, doi:10.1016/j.gca.2010.02.020. Hacker, B.R., Kelemen, P.B., and Behn, M.D., 2011, Differentiation of the continental crust by relamination: Earth and Planetary Science Letters, v. 307, p. 501–516, doi:10.1016/j.epsl.2011 .05.024. Hermann, J., and Rubatto, D., 2009, Accessory phase control on the trace element signature of sediment melts in subduction zones: Chemical Geology, v. 265, p. 512–526, doi:10.1016/j.chemgeo .2009.05.018. Hermann, J., and Spandler, C.J., 2008, Sediment melts at sub-arc depths: An experimental study: Journal of Petrology, v. 49, p. 717–740, doi:10.1093/petrology/egm073. Hermann, J., Müntener, O., Trommsdorff, V., Hansmann, W., and Piccardo, G.B., 1997, Fossil crust-to-mantle transition, Val Malenco (Italian Alps): Journal of Geophysical Research (Solid Earth), v. 102, no. B9, p. 20,123–20,132. Hoffmann, J.E., Münker, C., Naeraa, T., Rosing, M.T., Herwartz, D., Garbe-Schöenberg, D., and Svahnberg, H., 2011, Mechanisms of Archean crust formation inferred from high-precision HFSE systematics in TTGs: Geochimica et Cosmochimica Acta, v. 75, p. 4157–4178, doi:10.1016/j.gca.2011.04.027. Jochum, K.P., Stolz, A., and McOrist, G., 2000, Niobium and tantalum in carbonaceous chondrites: Constraints on the solar system and primitive mantle niobium/tantalum, zirconium/ niobium, and niobium/uranium ratios: Meteoritics & Planetary Science, v. 35, p. 229–235, doi:10.1111/j.1945-5100.2000.tb01771.x. Jochum, K.P., Weis, U., Stoll, B., Kuzmin, D., Yang, Q., Raczek, I., Jacob, D.E., Stracke, A., Birbaum, K., Frick, D.A., Günther, D., and Enzweiler, J., 2011, Determination of reference values for NIST SRM 610–617 glasses following ISO guidelines: Geostandards and Geoanalytical Research, v. 35, p. 397–429, doi:10.1111/j.1751-908X.2011.00120.x. Koenig, S., and Schuth, S., 2011, Deep melting of old subducted oceanic crust recorded by superchondritic Nb/Ta in modern island arc lavas: Earth and Planetary Science Letters, v. 301, p. 265–274, doi:10.1016/j.epsl.2010.11.007. Luvizotto, G., and Zack, T., 2009, Nb and Zr behavior in rutile during high-grade metamorphism and retrogression: An example from the Ivrea-Verbano zone: Chemical Geology, v. 261, p. 303– 317, doi:10.1016/j.chemgeo.2008.07.023. Münker, C., Pfänder, J., Weyer, S., Büchl, A., Kleine, T., and Mezger, K., 2003, Evolution of planetary cores and the earth-moon system from Nb/Ta systematics: Science, v. 301, p. 84–87, doi:10.1126/science.1084662. Nash, W., and Crecraft, H., 1985, Partition coefficients for trace elements in silicic magmas: Geochimica et Cosmochimica Acta, v. 49, p. 2309–2322, doi:10.1016/0016-7037(85)90231-5. Nebel, O., van Westrenen, W., Vroon, P.Z., Wille, M., and Raith, M.M., 2010, Deep mantle storage of the Earth’s missing niobium in late-stage residual melts from a magma ocean: Geochimica et Cosmochimica Acta, v. 74, p. 4392–4404, doi:10.1016/j.gca.2010.04.061. Plank, T., and Langmuir, C., 1998, The chemical composition of subducting sediment and its consequences for the crust and mantle: Chemical Geology, v. 145, p. 325–394, doi:10.1016 /S0009-2541(97)00150-2. Prowatke, S., and Klemme, S., 2005, Effect of melt composition on the partitioning of trace elements between titanite and silicate melt: Geochimica et Cosmochimica Acta, v. 69, p. 695– 709, doi:10.1016/j.gca.2004.06.037. Rudnick, R., and Fountain, D., 1995, Nature and composition of the continental crust: A lower crustal perspective: Reviews of Geophysics, v. 33, no. 3, p. 267–309, doi:10.1029/95RG01302. Rudnick, R., Barth, M., Horn, I., and McDonough, W., 2000, Rutile-bearing refractory eclogites: Missing link between continents and depleted mantle: Science, v. 287, p. 278–281, doi:10.1126/science.287.5451.278. Schmidt, M., Dardon, A., Chazot, G., and Vannucci, R., 2004, The dependence of Nb and Ta rutilemelt partitioning on melt composition and Nb/Ta fractionation during subduction processes: Earth and Planetary Science Letters, v. 226, p. 415– 432, doi:10.1016/j.epsl.2004.08.010. Stalder, R., Foley, S., Brey, G., and Horn, I., 1998, Mineral aqueous fluid partitioning of trace elements at 900–1200 °C and 3.0–5.7 GPa: New experimental data for garnet, clinopyroxene, and rutile, and implications for mantle metasomatism: Geochimica et Cosmochimica Acta, v. 62, p. 1781– 1801, doi:10.1016/S0016-7037(98)00101-X. Stepanov, A.S., Hermann, J., Rubatto, D., and Rapp, R.P., 2012, Experimental study of monazite/melt partitioning with implications for the REE, Th and U geochemistry of crustal rocks: Chemical Geology, v. 300–301, p. 200–220, doi:10.1016/j .chemgeo.2012.01.007. Stolz, A., Jochum, K., Spettel, B., and Hofmann, A., 1996, Fluid- and melt-related enrichment in the subarc mantle: Evidence from Nb/Ta variations in island-arc basalts: Geology, v. 24, p. 587– 590, doi:10.1130/0091-7613(1996)024<0587: FAMREI>2.3.CO;2. Tiepolo, M., Vannucci, R., Oberti, R., Foley, S., Bottazzi, P., and Zanetti, A., 2000, Nb and Ta incorporation and fractionation in titanian pargasite and kaersutite: Crystal-chemical constraints and implications for natural systems: Earth and Planetary Science Letters, v. 176, p. 185–201, doi:10.1016/S0012-821X(00)00004-2. Wade, J., and Wood, B., 2001, The Earth’s ‘missing’ niobium may be in the core: Nature, v. 409, p. 75–78, doi:10.1038/35051064. Xiong, X., Keppler, H., Audetat, A., Ni, H., Sun, W., and Li, Y., 2011, Partitioning of Nb and Ta between rutile and felsic melt and the fractionation of Nb/Ta during partial melting of hydrous metabasalt: Geochimica et Cosmochimica Acta, v. 75, p. 1673–1692, doi:10.1016/j .gca.2010.06.039. Manuscript received 24 June 2012 Revised manuscript received 19 September 2012 Manuscript accepted 24 September 2012 Printed in USA www.gsapubs.org | March 2013 | GEOLOGY