Half a million years of magmatic history recorded in a K-feldspar megacryst of the Tuolumne Intrusive Complex, California, USA

K-feldspars reach megacrystic size ( > 3 cm) relative to their groundmass in many granitoid plutons and some volcanic rocks. However, the nature of the growth environment and the time scales for megacrystic growth remain poorly constrained. Chemical abrasion–isotope dilution–thermal ionization mass spectrometry with trace-element analysis (CA-ID-TIMS-TEA) U-Pb geochronology was carried out on zircon inclusions from the core and rim of one K-feldspar megacryst sampled from the interior of the Tuolumne Intrusive Complex (TIC), California, USA. Combined with new zircon ages from the groundmass, these data can test if K-feldspar megacrysts are igneous and capable of recycling and transport in the magmatic system or whether they formed by textural coarsening in low-melt-fraction or subsolidus conditions. The zircon ages reveal that the megacryst core is 0.5 m.y. older than the rim, which itself is older than the groundmass. Core ages match zircon dates from the TIC’s porphyritic Half Dome unit, and rim and groundmass ages overlap with the younger Cathedral Peak unit. Trace elements of the zircons from the megacryst core and rim are similar and less evolved than the groundmass zircons. The core-to-rim age progression of zircon inclusions is inconsistent with subsolidus K-feldspar coarsening, but instead indicates that megacrysts in the TIC grew in an igneous environment over at least 0.5 m.y., and that growth likely occurred spanning two or more intrusive episodes. This supports models of an increasingly maturing magmatic system, where crystal recycling from older into younger magma batches is common.

Zircon inclusions within megacrysts and their host granodiorite can be useful for reconstructing the megacryst growth history because they are excellent U-Pb chronometers, their magmatic growth ceases once they are enclosed and armored by K-feldspar crystals (Barboni and Schoene, 2014), and they can record geochemical information about the liquid from which they crystallized.The magmatic origin models listed above predict younger zircon ages from megacryst core to rim over either short durations (hypothesis 1) or longer, more protracted durations (hypotheses 2 and 3).Zircon trace elements from the megacryst core and rim and the host granodiorite could help to differentiate if magma mixing and moderate melt percentages were involved (hypothesis 2), or if zircons grew in a rigid, highly evolved mush at very low melt percentages (hypothesis 3).A subsolidus origin (hypothesis 4), on the other hand, predicts no relationship between zircon age and position within the megacryst, unless the zircons are also metasomatic.To explore these possibilities, we obtained 36 zircon dates and associated geochemical compositions using chemical abrasion-isotope dilution-thermal ionization mass spectrometry and trace-element analysis (CA-ID-TIMS-TEA) from the core, rim, and surrounding groundmass of one K-feldspar megacryst from the Tuolumne Intrusive Complex (TIC) in the Sierra Nevada, California (United States).Our results suggest a magmatic growth history over 0.5 m.y.related to recycling of K-feldspar from different magmatic units.
The 92.8-88.8Ma Half Dome granodiorite (Paterson et al., 2016) is subdivided into the outer, equigranular Half Dome (eHD) and the inner, porphyritic Half Dome (pHD) granodiorites (Bateman and Chappell, 1979).While the eHD contains sub-to euhedral Kfeldspar crystals <1 cm, the pHD contains characteristic euhedral K-feldspar up to 3 cm long.These phenocrysts commonly contain abundant, several millimeter or smaller inclusions of euhedral plagioclase, hornblende, biotite, titanite, apatite, Fe-oxides, and zircon (Fig. 1C; Burgess and Miller, 2008;Moore and Sisson, 2008), which are typically crystallographically aligned along growth zones (Vernon and Paterson, 2008;Memeti et al., 2014).The 88.8-85.1 Ma Cathedral Peak granodiorite (CP; Paterson et al., 2016) contains megacrystic K-feldspars up to 15 cm long, particularly along the pHD contact (Fig. 1C; Bateman and Chappell, 1979).Some of these CP megacrysts, including the 7 cm megacryst (sample  analyzed in this study, exhibit mineral inclusion-rich cores (just like those found in the pHD) and mineral inclusion-poor rims (Fig. 2).LFO-74 was collected from a pHD-CP hybrid unit in the southeastern TIC, where all units grade into each other over tens to hundreds of meters (Fig. 1B).

METHODS
Zircon inclusions were separated from LFO-74, which was cut perpendicular to the crystallographic b axis.Both halves were cut into core, mantle, and rim sections defined by mineral inclusions (core) or lack thereof (rim; Fig. 2).The mantle (between the core and rim) was excluded from analyses to focus on retrieving the minimum duration of megacryst growth.Zircons were separated from the groundmass of the same sample with megacrysts removed.The euhedral, oscillatory-zoned zircons were dated using U-Pb CA-ID-TIMS geochronology following methods slightly modified from Mattinson (2005) and Barboni and Schoene (2014).The trace-element composition of each dated zircon was measured using inductively coupled plasma-mass spectrometry (ICP-MS) following the methods of Schoene et al. (2010).Analytical work was done at Princeton University (New Jersey, USA), and detailed methods, trace-element data, and concordia plots are presented in the GSA Data Repository 1 .All subsequent discussion uses 206 Pb/ 238 U dates, as this isotopic system provides the most precise dates for rocks of Cretaceous age.

RESULTS
Zircon dates become progressively younger from the megacryst core, where 8 of 11 zircons group around 89.0 Ma, to the rim, where 10 of 12 zircons group at 88.45 Ma (Figs. 2 and 3).The groundmass of the megacryst is younger than the rim, with 10 of 13 zircons clustering at ca. 88.25 Ma.The age difference between the youngest core zircon and youngest rim zircon, 88.86 ± 0.06 Ma and 88.29 ± 0.15 Ma, respectively, is ∼0.5 m.y.The older zircon grains found in the core and rim are earlier-grown zircons likely remobilized and reincorporated from older TIC crystal mushes (antecrysts) or the surrounding host rock (xenocrysts) during new magma intrusions (Miller et al., 2007;Paterson et al., 2016).A significantly older grain (99.81 ± 0.23 Ma) dated in the rim is outside the TIC age range and is interpreted to be a xenocryst (Fig. 2).Trace elements from the core and rim zircons show the same concentrations and variations, while groundmass zircons cluster at more evolved compositions (Fig. 3; Fig. DR3 in the Data Repository).

Igneous Origin for Megacrystic K-Feldspar
The U-Pb zircon age results from this study show a younging trend from the core to the rim of the megacryst.The zircons from the groundmass of the megacryst are even younger (Figs. 2 and 3), favoring igneous growth.Euhedral, oscillatory-zoned zircon morphology (Fig. DR4) and the coexistence with other magmatic mineral inclusions (Moore and Sisson, 2008;Vernon and Paterson, 2008) support the interpretation that the zircons are of magmatic origin.These results are incompatible with a subsolidus, metasomatic textural-coarsening origin (hypothesis 4; Johnson and Glazner, 2010;Glazner et al., 2017).
This result is consistent with petrographic, physical, and chemical evidence that supports a magmatic origin for K-feldspar megacrysts: Igneous megacrysts are euhedral and display Carlsbad twinning; they show oscillatory zoning of slow-diffusing trace elements; and they contain crystallographically aligned, euhedral mineral inclusions with hypersolidus temperatures of crystallization (Moore and Sisson, 2008;Vernon and Paterson, 2008;Vernon, 2010;Memeti et al., 2014).They are also locally aligned parallel to magmatic foliations and lineations, accumulate in clusters, or become entrained into enclaves and magmatic structures formed by flow sorting (Zák et al., 2007;Vernon and Paterson, 2008;Paterson, 2009;Paterson et al., 2018), all of which argue for Kfeldspar growth in a melt-present environment.Holness et al. (2018) showed that, although K-feldspar is one of the last phases to saturate, as predicted by experiments, the magma may still contain 50%-60% melt at this stage, leaving enough space for megacrystic growth.A magmatic environment has also been suggested based on megacrystic growth in a melt with changing isotopic composition (Hobden et al., 1999;Gagnevin et al., 2005a;Davidson et al., 2008).Erupted volcanic K-feldspar megacrysts (<5 cm) with glass inclusions also favor growth in a melt-rich environment (Clavero et al., 2004).Thus, evidence from these other studies and our new U-Pb zircon ages provide ample evidence that sample LFO-74 and many other investigated plutonic and volcanic megacrysts are not only of igneous origin, but they grew in crystal mushes capable of movement and sometimes eruption.
The 0.5 m.y.age difference between the megacryst core and rim zircons supports the idea of continuous, prolonged growth in a longlived magma body, or pulsed, rapid growth in multiple magma bodies, and/or a combination of these two end members, rather than fast, short-lived growth of the whole crystal (hypothesis 1).The K-feldspar core ages match the pHD bulk sample zircon ages, and the Kfeldspar rim and groundmass zircon ages agree with the CP bulk sample zircon ages from previous studies (Memeti et al., 2010;Paterson et al., 2016).The megacryst-groundmass bulkrock zircon age comparisons suggest the megacryst started growth in the pHD magma and was transferred into the CP magma by magma mixing or mechanical plucking, where continued growth resulted in the final size (Fig. 1C).This interpretation is consistent with the significantly higher mineral inclusion density in the K-feldspar cores, typical of pHD K-feldspars, compared to the inclusion-poor rims, typical of CP K-feldspars (Figs. 1C and 1D).The gradational contacts and hybrid units in this area of the TIC (Fig. 1B) also support the idea of interaction between adjacent units, and they make megacryst growth in different magma batches likely (Fig. 1D).Our interpretations of megacryst growth due to crystal recycling are supported by isotope data and mineral traceelement analyses from other studies that show that K-feldspars crystallized in different magma compositions over time (Troll and Schmincke, 2002;Słaby and Götze, 2004;Ginibre et al., 2004Ginibre et al., , 2007;;Gagnevin et al., 2005aGagnevin et al., , 2005b;;Zellmer and Clavero, 2006;Słaby et al., 2007aSłaby et al., , 2007bSłaby et al., , 2017)).The trace-element data from the analyzed zircons act as geochemical recorders of magma composition from which the zircons, and thus the K-feldspar grew.Trace elements from the core and rim zircon populations display similar means and large variances, even though these zircons are distinctly different in age.This is consistent with observations made using wholerock isotopes and geochemistry showing that the interior units of the TIC become increasingly more similar in their compositions (Memeti et al., 2014).This trend is, in part, due to widespread mixing/mingling and recycling processes (Paterson et al., 2008(Paterson et al., , 2016)).Furthermore, although the similar pHD and CP magma compositions do not allow for the distinction between pHD-and CP-derived zircons based on chemistry alone, the core-to-rim age difference, the similarities with pHD and CP whole-rock zircon dates, and the characteristic pHD and CP mineral inclusion patterns are significant and consistent with pHD K-feldspar phenocryst transfer into the CP.The slightly younger groundmass zircons have more evolved (low Zr/Hf) and homogeneous trace-element compositions (Fig. 3) that record the crystallization of melt that had previously crystallized a significant volume of zircon (Claiborne et al., 2006).The groundmass zircons are collectively compositionally more evolved than the chemically heterogeneous zircons from the megacryst.This suggests that the megacrysts could not have grown at low melt percentages as suggested by hypothesis 3. Furthermore, time scales of textural coarsening in igneous systems are unrealistically large to grow even millimeter-size minerals (Gualda, 2019).

CONCLUSIONS
CA-ID-TIMS-TEA U-Pb zircon ages from the core and rim of a K-feldspar megacryst and its surrounding groundmass from the pHD-CP hybrid unit in the Tuolumne Intrusive Complex, California, show younging trends from core to rim, which, along with field, petrographic, and geochemical evidence, support a magmatic origin for TIC K-feldspar megacrysts.The 0.5 m.y.age difference between megacryst core and rim zircon ages suggests that Kfeldspar megacrysts had the capability to grow over at least that long-either continuously or during punctuated intervals-in a magmatic system undergoing magma mixing and crystal recycling.The 0.5 m.y.age difference and the similarities with whole-rock zircon ages from the pHD and CP units as well as the traceelement characteristics compared to those in the groundmass zircons indicate growth of the K-feldspar into its megacrystic size by growth in different, interacting magma bodies, rather than through fast growth or in situ textural coarsening in a crystal-rich, closed-system magma body.

ACKNOWLEDGMENTS
This research was funded by U.S. National Science Foundation grants EAR-1550935 and EAR-1624854 awarded to Memeti. Louis Oppenheim provided the K-feldspar megacryst for this study.We thank Scott Paterson and Cal Barnes for fruitful discussions on the Tuolumne Intrusive Complex and beyond, and Melanie Barboni for her initial guidance in megacryst zircon separation.Thanks go to Jonathan Miller, Stephen Daly, and an anonymous reviewer for their very thoughtful reviews, which helped to clarify the manuscript.Barboni, M., and Schoene, B.,

1Figure 1 .
Figure 1.Sample location of megacryst LFO-74.(A) Inset map of the Tuolumne Intrusive Complex (TIC) and its location in California, USA.SF-San Francisco, LA-Los Angeles.(B) Location of LFO-74 in the southeast lobe of the TIC where contacts are gradational and hybrid units separate main TIC units.Map was adapted from Huber et al. (1989) by Louis Oppenheim (May 2019, personal commun.).(C) Increasing K-feldspar size and distribution of mafic mineral inclusions for Cathedral Peak granodiorite (CP), porphyritic Half Dome (pHD), and equigranular Half Dome (eHD) granodiorites.(D) Cartoon of the K-feldspar recycling hypothesis indicating inclusion-rich pHD K-feldspars are recycled into the CP unit, where they grow inclusion-poor rims.

Figure
Figure 2. Chemical abrasion-isotope dilutionthermal ionization mass spectrometry (CA-ID-TIMS) U-Pb zircon ages from core, rim, and groundmass of K-feldspar megacryst LFO-74 (from the southeastern Tuolumne Intrusive Complex, California, USA).Identification of K-feldspar core and rim based on mafic mineral density is shown in inset photo.Horizontal dashed lines represent single-zircon bulk sample age ranges from Paterson et al. (2016) used for comparison with K-feldspar megacryst zircon inclusion ages.One older zircon from the rim (99.8 ± 0.23 Ma) is not shown on plot.Uncertainties are 2σ.Abbreviations used are the same as in Figure 1.
Figure 3. U-Pb age probability density plot and Zr/Hf ratios for zircons from core, rim, and groundmass of megacryst LFO-74 (collected from the southeastern Tuolumne Intrusive Complex, California, USA).Only zircons with ages between 88.2 Ma and 89.4 Ma are shown.Zr/Hf values of core and rim zircons show overlapping concentrations.Groundmass values group separately from core and rim values at more-evolved compositions, suggesting growth from a more-evolved, zircon-fractionated composition during final crystallization.Probability density plot shows distinct age differences between three zircon separates.