Mechanism of metamorphic zircon growth in a granulite-facies ...

American Mineralogist, Volume 95, pages 1796–1806, 2010

Mechanism of metamorphic zircon growth in a granulite-facies quartzite, Adirondack Highlands, Grenville Province, New York William H. Peck,1,* M.E. Bickford,2 James M. McLelland,1 Ashley N. Nagle,1,† and Gretchen J. Swarr1 1 Department of Geology, Colgate University, Hamilton, New York 13346, U.S.A. Department of Earth Sciences, 204 Heroy Geology Laboratory, Syracuse University, Syracuse, New York 13244, U.S.A.

2

Abstract Granulite-facies quartzites from the Adirondack Highlands (Grenville Province, New York) contain accessory zircon with ubiquitous metamorphic overgrowths. Detrital zircon cores are resorbed, preserve various internal zoning styles and inclusions, and have ages of 1.3 to 2.7 Ga. These ages constrain the timing of deposition of the protolith sandstone and suggest that the Adirondack Highlands were proximal or part of Laurentia during deposition. Metamorphic overgrowths formed in the quartzites during the Shawinigan orogeny (1.20–1.14 Ga). The average volume of overgrowths in eight samples ranges from 62–87%, with a positive correlation between zircon growth and feldspar content (melt productivity during metamorphism). Melt volumes and Zr solubility were too small to allow the overgrowths to have formed in one dissolution-precipitation event. Crystal-size distributions suggest zircon coarsening by the dissolution of small crystals and Zr transfer via a partial melt, and thus zircon overgrowths date anatexis. These results have implications for provenance studies, as dissolution of small zircon grains and growth of new zircon could bias age spectra of metasedimentary rocks. Keywords: Metamorphic zircon, quartzite, U-Pb geochronology, Ostwald ripening, Adirondack Highlands, Grenville Province

Introduction Large zircon overgrowths are very common in high-grade metaigneous and metasedimentary rocks, but the mechanism of their formation is rarely obvious (e.g., Corfu et al. 2003). This makes measured U-Pb ages from “metamorphic” overgrowths difficult to correlate with the rock’s burial and thermal history. In some rocks the onset of partial melting correlates with overgrowth formation (e.g., Roberts and Finger 1997), but in other rocks breakdown of Zr-rich phases such as garnet or ilmenite appear to control new zircon growth (e.g., Fraser et al. 1997; Bingen et al. 2001; Degeling et al. 2001). Aqueous fluids can also promote zircon development (e.g., Rubin et al. 1989) or recrystallize igneous zircon (e.g., Pidgeon et al. 1998). In this study, we examined the mechanisms of metamorphic zircon growth and investigated the geochronology of granulitefacies quartzites and psammites from the Adirondack Highlands, New York. These quartzites were observed in a previous investigation to contain large zircon overgrowths (Peck et al. 2003). The advantage of using quartz-rich rocks to study metamorphic zircon growth is that, other than protolith zircon, they do not contain abundant minerals that can be a source of zirconium, and melt productivity in these rocks should be low. These characteristics allow the relative importance to zircon growth of mineral breakdown, aqueous fluids, and melt to be evaluated. Understanding zircon dissolution and growth also has important implications * E-mail: [email protected] † Present address: Department of Geological Sciences, Brown University, Providence, Rhode Island 02912, U.S.A. 0003-004X/10/1112–1796$05.00/DOI: 10.2138/am.2010.3547

for understanding zircon age distributions in provenance studies. In the quartzites studied, crystal-size distributions, examination of core-rim textures, U-Pb geochronology by SHRIMP, and whole-rock chemistry point to small amounts of melt being the fundamental control on zircon dissolution and metamorphic zircon growth in these rocks, even though textural evidence for melting has been annealed. These results have important implications for the interpretation of zircon overgrowths in high-grade quartzofeldspathic rocks.

Geologic setting The granulite-facies Irving Pond quartzite is located in the core of the Canada Lake antiform, a large east-plunging recumbent isocline in the Mesoproterozoic Adirondack Highlands (McLelland and Isachsen 1986; Peck and Valley 2004). The quartzite has an exposed thickness of >3000 m and crops out for ~100 km along strike (McLelland et al. 1979). The Irving Pond quartzite consists of decimeter-scale layers of quartz-rich, quartzofeldspathic, and pelitic lithologies. Quartz-feldspar oxygen isotope fractionations from this unit correlate with metamorphic temperatures of 734 ± 38 °C (n = 15; Peck and Valley 2004), and cation geothermometry from nearby migmatitic pelites yields temperatures of ca. 790 °C or greater and pressures of 7–9 kbar (Storm and Spear 2005). The depositional age and age of metamorphism of the Irving Pond quartzite are now well constrained. Nappe formation in the Adirondacks affects the 1.10 Ga Hawkeye granite suite (McLelland et al. 2001) and, therefore, deformed the Irving Pond quartzite during the ca. 1.10–1.03 Ga Ottawan phase of the

1796

Peck et al.: Metamorphic zircon growth in a granulite-facies quartzite

Grenvillian orogeny (Rivers 2008). The quartzite unit forms the core of the antiform, and is surrounded and crosscut by the 1.25 Ga Canada Lake charnockite (McLelland and Chiarenzelli 1990). A SHRIMP U/Pb study (Heumann et al. 2006) shows that partial melting in pelites from the Canada Lake antiform occurred during the 1.18–1.16 Ga contractional phase of the Shawinigan orogeny (McLelland et al. 2010), which may have caused metamorphic mineral growth in the quartzite as well. Detrital zircon SHRIMP U/Pb ages from Adirondack metapelites (Heumann et al. 2006; Bickford et al. 2007) fall into an interval of ca. 1350–1220 Ma, suggesting that deposition of protoliths was in a restricted basin not receiving sediment from the Superior Province. Prior to this investigation, the only age constraint on provenance of Adirondack quartzites is a bulk U-Pb zircon age of ca. 2.6 Ga from the Swede Pond quartzite (Fig. 1; Silver 1965). Here we demonstrate that the Irving Pond quartzite contains 2.7 to 2.6 Ga detrital zircon grains consistent with a Superior Province provenance. Both of these conclusions are consistent with a plate tectonic model developed by McLelland et al. (2010).

Materials and methods Sampling strategy Eight fresh, unaltered samples, selected from previous studies of the Irving Pond quartzite (Peck et al. 2003; Peck and Valley 2004) to represent a variety of

1797

lithologic types along 58 km of strike of this unit, were chosen for a systematic study of zircon textures. Seven out of the eight samples were from a traverse of ~10 km parallel to the E-W axis of Irving Pond antiform (Fig. 1). For U-Pb geochronology, a representative, ca. 100 kg sample of glassy quartzite sample was collected from locality 1 (Fig. 1).

Analytical methods Zircon grains were analyzed for U-Pb geochronology with the sensitive high-resolution ion microprobe (SHRIMP II) in the J.C. Roddick Ion Microprobe Laboratory at the Geological Survey of Canada (GSC). Methods are described by Stern (1997). For geochronology a glassy quartzite sample was crushed and ground, and zircon grains were separated using Wilfley Table, heavy liquid, and magnetic methods. Unaltered zircon grains lacking visible cracks were handpicked under a microscope to be representative of the zircon population. These zircon grains were cast in a 1 inch round epoxy mount along with fragments of the GSC laboratory standard zircon (6266 zircon: 206Pb/238U isotope dilution age = 559 Ma; 910 ppm U; 206Pb/238U ratio 0.09059). The zircon mount was polished with diamond compound to reveal equatorial cross-sections, and zircon grains were imaged using backscattered electrons and cathodoluminescence to identify inherited cores and overgrowths for analysis, and to avoid cracks and alteration. Zircon grains were analyzed using an O− primary beam focused into an elliptical spot with an average radius of 20–30 µm. Instrument conditions for analyses beginning with 7866a or 7866b are described in McLelland et al. (2004), and conditions for the remainder are described in Heumann et al. (2006). Data from each analytical session was calibrated for U/Pb bias and measured 204Pb counts were used to correct for common Pb. Data reduction was accomplished using an in-house program at the GSC and final corrected ratios and ages are presented with 1σ analytical uncertainties in Table 1. Samples for textural analysis were analyzed for whole-rock major elements by X-ray fluorescence and for Zr content by ICP-OES at XRAL Laboratories/SGS Canada (Table 2).

Figure 1. Geology of the Adirondack Highlands and the study area. (a) Location map of the Grenville Province. (b) The Adirondack Highlands, bound on the northwest by the Carthage-Colton myolonite zone (CC) and elsewhere by Paleozoic rocks. Anorthosites are stippled. (c) Geology of the Canada Lake antiform (McLelland and Isachsen 1986). Black = quartzite and pelitic metasediments. Stipple = charnockite, metatonalite, and metagranitic rocks. Diagonal ruled = pyroxene-hornblende-quartz-plagioclase gneiss. White = garnet-biotite-quartz-plagioclase gneiss. Heavy dashed = boundary of Precambrian exposure. Sample Location 1 (02AK74 and 81, 97ADK4, and the Geochronology sample) is the Town of Caroga Highway Department sand and gravel depository adjacent to Rt 29A-10 at the east end of Canada Lake. Location 2 = 02AK8, Location 3 = 02AK25 and 36, Location 4 = 02AK39.

1798


Table 1. Sensitive high-resolution ion microprobe (SHRIMP II) analytical data from Irving Pond quartzite zircons, Adirondack Highlands, New York Apparent ages (Ma) 204 206 207 Spot U Th/ Pb/ Pb/ ±206Pb/ Pb/ ±207Pb/ Disc. 206 238 238 206 206 Spot name location (ppm) U Pb f(206)204 U U Pb Pb (%) 7866-3.1 core 107 0.56 0.000144 0.002500 1403 20 1475 31 4.9 7866-8.1 core 278 1.41 0.000049 0.000840 1944 22 1955 10 0.6 7866-12.1 core 500 0.32 0.000010 0.000170 1532 17 1599 40 4.2 7866-18.1 core/rim 304 0.22 0.000010 0.000170 1350 13 1498 12 9.9 7866-30.1 core 88 0.33 0.000357 0.006180 1434 23 1355 78 –5.8 7866-33.1 core 457 0.58 0.000015 0.000260 1348 14 1354 12 0.5 7866-37.1 core 343 0.37 0.000046 0.000800 1757 61 1787 79 1.7 7866-41.1 core 681 0.39 0.000020 0.000340 1318 15 1350 17 2.4 7866-44.1 core 607 0.32 0.000047 0.000820 1360 15 1367 26 0.5 7866-47.1 core 278 0.35 0.000010 0.000170 1376 16 1408 14 2.3 7866-51.1 core 456 0.73 0.000019 0.000340 1832 21 1927 51 4.9 7866-60.1 core 216 0.74 0.000033 0.000560 1327 14 1330 30 0.2 7866-63.1 core 178 0.25 0.000014 0.000230 1526 17 1820 21 16.1 7866-74.1 core 231 0.78 0.000007 0.000120 1906 70 1951 75 2.3 7866-75.1 met rim 630 0.04 0.000034 0.000590 1052 11 1064 15 1.1 7866-77.1 met rim 632 0.04 0.000022 0.000380 1125 11 1120 17 –0.5 7866-82.1 core 217 0.52 0.000025 0.000440 1338 27 1336 51 –0.2 7866-86.1 core 272 0.95 0.000031 0.000540 1368 14 1415 25 3.4 7866-93.1 core 108 0.87 0.000165 0.002860 1691 31 1708 58 1.0 7866-98.1 core 276 0.47 0.000045 0.000780 1670 21 1803 14 7.3 7866-98.2 core/rim 598 0.00 0.000017 0.000300 1161 12 1202 32 3.4 7866-99.1 core 207 0.82 0.000016 0.000280 1572 17 1684 15 6.6 7866-101.1 core 213 1.02 0.000002 0.000040 2507 72 2685 58 6.6 7866-116.1 core 131 0.80 0.000010 0.000170 1585 25 1635 26 3.1 7866-133.1 core 553 0.62 0.000008 0.000140 1431 15 1432 14 0.1 7866-136.1 core 784 0.11 0.000010 0.000180 1376 21 1389 27 0.9 7866-142.1 core 448 0.37 0.000010 0.000170 1536 15 1665 15 7.7 7866-149.1 core 197 0.62 0.000085 0.001480 1363 14 1379 39 1.2 7866-150.1 met rim 688 0.02 0.000028 0.000490 1147 12 1141 14 –0.6 7866-153.1 met rim 107 0.00 0.000126 0.002180 1121 17 1223 57 8.3 7866-154.1 core 182 0.78 0.000034 0.000600 1619 22 1631 18 0.7 7866a-18.1 core 533 0.43 0.000002 0.000030 1650 20 1710 14 3.5 7866a-21.1 core 132 0.30 0.000010 0.000170 1298 18 1362 42 4.8 7866a-25.1 core 121 0.43 0.000053 0.000920 1274 17 1286 40 0.9 7866a-34.1 met grain 717 0.15 0.000004 0.000070 1144 11 1133 17 –1 7866a-42.1 core 129 0.52 0.000010 0.000170 1392 41 1356 66 –2.7 7866a-43.1 core 497 0.42 0.000073 0.001260 1316 17 1296 17 –1.6 7866a-45.1 met rim 449 0.19 0.000010 0.000170 1140 12 1190 15 4.2 7866a-45.2 core 482 0.56 0.000060 0.001030 1433 18 1430 21 –0.2 7866a-49.1 core 227 0.47 0.000003 0.000040 1311 46 1386 38 5.4 7866a-6.1 core 230 0.40 0.000005 0.000090 1265 22 1415 36 10.6 7866b-1.1 core 464 0.40 0.000010 0.000170 2568 40 2719 23 5.6 7866b-9.1 core 88 1.19 0.000028 0.000480 2584 35 2647 17 2.4 7866b-16.1 met rim 837 0.03 0.000037 0.000650 1137 11 1175 20 3.3 7866b-25.1 met rim 630 0.03 0.000010 0.000170 1143 11 1163 13 1.8 Notes: Uncertainties reported at 1σ (absolute) and are calculated by numerical propagation of all known sources of error (see Stern 1997). f(206)204 refers to mole fraction of total 206Pb that is due to common Pb, calculated using the 204Pb-method; common Pb composition used is the surface blank. Discordance relative to origin = 100*[1 – (206Pb/238U age)/(207Pb/206Pb age)].

Textural analysis methods Thin sections of samples for textural analysis were examined using a scanning electron microscope and a petrographic microscope. Feldspar-quartz-quartz dihedral angles were measured for 50 triple-junctions from each sample (25 from Qtz-rich 97ADK4) in thin sections after the methods of Stickels and Hucke (1964) (Table 3). Sub-samples of ca. 10–15 g were crushed using a small jaw crusher and separated using methylene iodide (SG = 3.33). This method avoided excessive breakage of zircon crystals as is common when large volumes of samples are processed using a disk grinder. Only whole crystals were examined as part of textural analysis. Zircon crystals in the heavy mineral split were separated and digitally imaged as loose crystals to determine crystal-size distributions, and images were processed using the public-domain NIH Image program developed at the U.S. National Institutes of Health (available at http://rsb.info.nih.gov/nih-image/) (e.g., Bindeman 2003). Areas, lengths, and widths of best-fit ellipses were measured for 204–330 crystals from each sample for a total of 2022 crystals (Table 3). Zircon crystals for scanning electron microscope imaging were sieved to separate crystals with different widths, hand-picked under a binocular microscope, and cast in three different epoxy 1 inch round mounts. This procedure allowed the three mounts to be ground and polished to the appropriate zircon half-width,

so the zircon images examined represent close to the true lengths and radii of the crystals. For the eight samples, 476 zircon crystals were imaged using backscattered electrons and cathodoluminescence (Fig. 2). Of the examined zircon crystals, 460 have identifiable cores surrounded by later rims that truncate zoning (51–60 zircon crystals from each sample). Crystal outlines and cores were traced, and images were processed using the ImageJ software package (Sheffield 2007). Areas, lengths, and widths of best-fit ellipses were measured using ImageJ. Radii were calculated as if crystal outlines and cores were circular, and volumes were then calculated for spherical geometries (after Nemchin et al. 2001) (Table 4). Performing this calculation using a tetragonal formula for zircon crystals and cores after Bindeman (2003) yields calculated % rims that are on average different by only 1 vol%. Average grain-lengths measured from epoxy mounts are 11 µm shorter than grain separate measurements, but do not show systematically lower areas, indicating that the exact centers were only slightly missed in mount preparation and grinding.

Zircon texture analysis Sample descriptions The sample suite was selected to have a large range of quartz:feldspar ratios (ca. 1:3 to 100:1), ranging from psammites


1799

Table 2. Whole-rock geochemical data of Irving Pond quartzites, Adirondack Highlands, New York Sample SiO2 TiO2 Al2O3 02AK25 93.67 0.15 3.39 02AK31 84.30 0.09 8.65 02AK39 72.21 0.61 14.48 02AK74 79.12 0.65 8.84 02AK81 74.39 0.54 12.97 527-8 77.19 0.33 11.83 97ADK2 96.39 0.29 1.42 97ADK4 84.57 0.10 6.88 Notes: Oxide compositions are in weight percent.

Fe2O3 0.97 2.46 4.13 2.52 3.70 3.37 0.40 1.96

MnO 0.01 0.01 0.13 0.06 0.11 0.10 0.00 0.10

MgO 0.32 0.19 0.61 1.86 0.97 0.56 0.20 1.27

CaO 0.14 0.29 1.20 0.75 1.09 2.32 0.03 0.38

Na2O 0.76 1.77 1.95 1.15 1.64 2.52 0.05 0.97

K2O 0.51 3.85 3.31 2.71 2.19 1.68 0.63 0.55

P2O5 0.01 0.02 0.02 0.02 0.04 0.01 0.01 0.02

Zr (ppm) 230 110 320 260 300 210 310 165

Total 99.95 101.64 98.68 97.71 97.67 99.93 99.45 96.82

Figure 2. Cathodoluminescence images of representative core/rim relations in zircon from the Irving Pond quartzite. Images A–I are representative images of zircons measured for textural analysis. Images J–L are representative images of zircons analyzed for geochronology, with 207 Pb/206Pb ages of spots in Ma and 1σ uncertainties.

1800


Table 3. Average grain measurements of whole zircon crystals from Irving Pond quartzites, Adirondack Highlands, New York Average Average Average Elongation Sample length (µm) ±1σ width (µm) ±1σ elongation ±1σ vs. length correlation 02AK25 93 25 62 1 1.5 0.4 0.04 02AK31 106 19 66 1 1.6 0.4 0.46 02AK39 89 19 54 1 1.5 0.3 0.00 02AK74 94 23 63 1 1.5 0.3 0.25 02AK81 92 23 56 1 1.7 0.4 0.29 527-8 106 27 67 13 1.6 0.3 0.44 97ADK2 101 20 70 12 1.5 0.3 0.32 97ADK4 97 25 62 13 1.6 0.3 0.29 Notes: Elongation is width/length. Elongation vs. length correlation is the R2 value.

Table 4. Average measurements of zircon cores and rims from Irving Pond quartzites, Adirondack Highlands, New York Average core measurements in µm Sample vol% rims Major axis Minor axis Elongation 02AK25 62 64 38 1.8 02AK31 66 70 40 1.8 02AK39 87 51 25 2.2 02AK74 73 61 34 1.9 02AK81 77 65 31 2.3 527-8 67 82 42 2.1 97ADK2 67 70 38 2.0 97ADK4 70 72 35 2.2 Notes: Volume percent rim measurements average ±25% (1σ, uncertainty in the mean is ±4%). Elongation is width/length.

to quartzites. Samples range from ~20 to ~100% xenoblastic quartz, with plagioclase, K-feldspar, opaques, and garnet. Some samples contain up to 10–15% biotite, sillimanite, altered pyroxene, and rare late white mica. Grains are slightly elongate, grain-lengths average ca. 0.25–1.0 mm, and quartz is up to 2–3 mm in quartz-rich samples. In samples with modally abundant quartz, feldspar is evenly distributed and isolated between quartz grains in the rock. Quartz-poor (< ~50%) samples show moderately developed, centimeter-scale, plagioclase and/or Kfeldspar-rich layering. Samples do not show obvious evidence for melt segregation (either from external melt or partial melting), and quartz-quartz-feldspar dihedral angles average 98 to 115° (Table 3), consistent with high-temperature textural equilibrium (e.g., Hiraga et al. 2002). Whole-rock SiO2 ranges from 72–96 wt%, and shows excellent negative correlations with Al2O3 (R2 = 0.97), very good negative correlation with Fe2O3, MnO, and Na2O (R2 = 0.64–0.75), good negative correlation with TiO2, CaO, K2O, and P2O5 (R2 = 0.40–0.54), and poor negative correlation with MgO (R2 = 0.18). All of these trends are consistent with sedimentary mixtures of feldspar, ferromagnesian silicates, and oxides with variable amounts of quartz. Whole-rock chemistry of samples spans sandstone compositions from lithic arenite, arkose, and subarkose to sublithicarenite and quartz arenite using SiO2/Al2O3, Na2O/K2O, and Fe2O3/K2O as discriminants (after Herron 1988). Sample 97ADK2, the most quartz-rich sample, is unusual geochemically compared to the other samples. For all other samples Na2O/K2O = 0.42–1.77, whereas 97ADK2 has Na2O/K2O = 0.08. This may just indicate the K-feldspar-rich nature of subsample used for geochemistry. This quartz-rich rock plots as an outlier in plots of whole-rock SiO2 contents displaying the relationship between whole-rock chemistry and zircon texture, and was collected 48 km east of where the other samples were collected.

External morphologies of zircon crystals and crystal-size distributions All of the examined samples contain zircon populations with broadly similar characteristics (Table 3). Zircon grains are subrounded prisms with irregular surfaces, visible prismatic faces, and poorly or undeveloped pyramidal faces. Average zircon lengths are 92–106 µm, widths are 56–70 µm, and elongation (width/length) ranges from 1.5–1.7. Of the 2022 zircon crystals measured, very small zircon crystals are rare (0.3% were 71% average zircon rim volumes, open symbols are