P-T-D histories from quartz: A case study of the application of the TitaniQ thermobarometer to progressive fabric development in metapelites



[1] This study investigates the ability of quartz to record segments of the pressure-temperature-deformation (P-T-D) path in poly-deformed metamorphic terranes and the associated application of the Ti-in-quartz thermobarometer (TitaniQ). Metapelites from the Strafford Dome (eastern Vermont) were selected for this study because they record progressive fabric development during prograde metamorphism and have well constrained P-T-D histories from previous work. Results of this investigation are in agreement with previous studies and demonstrate that quartz in these samples records additional distinct intervals of the P-T-D path. Six preserved quartz equilibration events include: (1) early prograde growth during burial; (2) kyanite-in quartz-producing reactions; (3) prenappe emplacement equilibration during isobaric heating; (4) precipitation of quartz in quartz-feldspar domains during crenulation cleavage development due to solution transfer; (5) quartz-producing (chlorite-out) reactions during heating postnappe emplacement; and (6) retrograde quartz overgrowths from an influx of siliceous fluids. Modeling quartz volume allows for the identification of quartz-producing reactions in P-T space; the same may be done for Ti-phases to better constrain activities during reequilibration events. An integration of cathodoluminescence imaging, microstructural and petrographic investigation, isochemical forward stability modeling, and thermobarometry allows identification and interpretation of zoning patterns in quartz to unravel histories that would otherwise not be obtainable.

1. Introduction

[2] Quartz has been widely used to estimate temperatures of deformation in metamorphic rocks based on microstructures associated with dynamic recrystallization [Stipp et al., 2002] and the quartz c-axis fabric opening angle thermometer of Kruhl [1996]. Geochemical analysis of this common crustal phase however has only recently been exploited for thermobarometry purposes. In recent years, the solubility of Ti in quartz has been calibrated experimentally, resulting in the Ti-in-Quartz (TitaniQ) thermobarometer, and can be used to infer conditions of quartz equilibration with a precision of temperature estimates on the order of ±5°C (2σ; if pressure and the activity of TiO2 are well constrained) [Thomas et al., 2010; Wark and Watson, 2006].

[3] The applications of TitaniQ are potentially extensive and significant across a spectrum of disciplines due to the abundance and stability of quartz in continental crust [see, for example: Lang and Gilotti, 2007; Sato and Santosh, 2007; Wark et al., 2007; Wiebe et al., 2007; Holness and Sawyer, 2008; Rusk et al., 2008; Kohn and Northrup, 2009; Vazquez et al., 2009; Behr and Platt, 2011; Grujic et al., 2011; Behr and Platt, 2012; Korchinski et al., 2012; Kidder et al., 2013]. A study by Spear and Wark [2009] demonstrated that quartz, depending on textural context and metamorphic grade, (1) records temperatures of metamorphism, (2) may show metamorphic zoning and/or modification by volume diffusion, and (3) has the potential to record temperatures of fabric formation. Studies of Ti volume diffusion in quartz suggest that either high temperatures, small grain sizes and/or long times are required for grains to fully reequilibrate via volume diffusion during metamorphism [Cherniak, 2010; Cherniak et al., 2004, 2007] and Spear et al. [2012] have modeled Ti volume diffusion in quartz to infer metamorphic time scales.

[4] This study assesses the application of the TitaniQ thermobarometer to constraining pressure-temperature-deformation (P-T-D) histories in metamorphic tectonites that record progressive deformation during crenulation cleavage development. The Strafford Dome in eastern Vermont was selected for this study because the petrological and structural histories of the rocks have been well constrained (Figure 1) [Menard and Spear, 1994; T. Menard, Rensselaer Polytechnic Institute, unpublished thesis, 1991], providing exceptional context for the samples being analyzed. The metapelitic rocks of the Strafford Dome record evidence of multiple stages of fabric development associated with prograde metamorphism (up to peak kyanite/staurolite-grade) during the Acadian Orogeny (397–350 Ma) [Spear and Harrison, 1989]. The presence of quartz in distinct microstructural domains such as inclusion suites that define internal foliations in garnet porphyroblasts, garnet pressure shadows, and in matrix foliation microlithons facilitate the comparison of relationships between garnet growth, deformation and quartz recrystallization/(re)equilibration. The metapelitic samples also tightly confine TiO2 activities to c. 1.0 [Ghent and Stout, 1984].

Figure 1.

(a) Geologic map of the Strafford Dome including: metamorphic isograds, sample locations where P-T paths were constructed by T. Menard (unpublished thesis, 1991) and Menard and Spear [1994], strike and dip of bedding and foliations (S1 and S2) outlining dome structure (data from Howard [1969]), and samples analyzed for this study. Inset figure is an overview map showing location of geologic map (shaded area) with Eastern Vermont and Merrimack tectonic belts emphasized (modified from Spear et al. [2002]). (b) Cross section (A–A′, shown in a) along the traverse of samples collected in this study modified after Howard [1969]. Generalized isobars are overlain on the cross section (from T. Menard (unpublished thesis, 1991)). Samples from this study are grouped with nearby samples with P-T histories from T. Menard (unpublished thesis, 1991) and Menard and Spear [1994], and labeled “A” (eastern flank nearing Monroe Fault), “B” (near core of dome), and “C” (northeast flank of dome, parallel to dome trends and isograds). Inset P-T plot shows generalized P-T paths for the metamorphic belts, displaying the typical Barrovian and Buchan/Abukuma facies series for the Eastern Vermont and Merrimack belts, respectively. (c) Compiled P-T histories for various samples across the Strafford Dome [Menard and Spear, 1994]. Note how peak conditions decrease east of the dome toward the Monroe Fault.

2. Background

2.1. Regional Geology of the Strafford Dome

[5] The Strafford Dome is the northern-most of a series of N–S-trending domes in eastern Vermont (Figure 1) [Doll et al., 1961; Menard and Spear, 1994]. The Monroe Fault to the east and the Richardson Memorial Contact to the west separate the generalized anticlinorium structure from New Hampshire and pre-Silurian rocks, respectively [Hatch, 1988; Menard and Spear, 1994; Spear et al., 2002]. Lithologies that define the Strafford Dome principally comprise the Silurian to Early Devonian Waits River and Gile Mountain Formations, which collectively make up the Connecticut Valley Trough. The older Waits River Formation consists of micaceous limestones, pelitic schists, and the Standing Pond amphibolites [Doll, 1944; Fisher and Karabinos, 1980]. This formation is a calcareous flysch deposit with dm- to 10 m scaled interleaving of pelitic and carbonate layers. The Standing Pond amphibolite is a thin (10–100 m) unit of basic metavolcanic material [Evans et al., 2002]. The Gile Mountain Formation has similar lithologies as the Waits River Formation, although it has a paucity of limestones and has a profusion of quartz-mica schists and amphibolites [Doll, 1944], with pelites and psammites being the main constituents.

[6] The Gile Mountain Formation preserves sedimentary bedding locally [Fisher and Karabinos, 1980; Menard and Spear, 1994; Woodland, 1977], with bedding-parallel micaceous schistosity (S1) [Menard and Spear, 1994; White and Jahns, 1950; T. Menard, unpublished thesis, 1991]. Previous workers have attributed development of crenulation cleavage (S2 schistosity) and large-scale recumbent folds across the dome [White and Jahns, 1950] to progressive deformation associated with emplacement of a nappe [Woodland, 1977]. Map patterns and structural data from the core of the dome reveal the doming of S2 [Doll, 1944; Menard and Spear, 1994; White and Eric, 1944; Woodland, 1977]. Widely spaced or kink banding (S3) locally deforms S2 [Fisher and Karabinos, 1980; Woodland, 1977], but the relationship between S3 and metamorphism is not well constrained.

2.2. Metamorphism and Paragenesis

[7] Metamorphism of the rocks of the Strafford Dome was studied in detail by Menard and Spear [1994] and is associated with the Devonian Acadian Orogeny (397–350 Ma) [Spear and Harrison, 1989]. The paragenesis of muscovite + biotite + quartz ± ilmenite constitutes the majority of the biotite grade metamorphic zone. Higher metamorphic-grade parageneses consist of garnet + biotite + muscovite + quartz ± plagioclase ± tourmaline (garnet-grade metamorphism) and kyanite + staurolite + garnet + plagioclase + biotite + quartz (staurolite/kyanite-grade metamorphism); the highest grade rocks are found in the core of the dome. Chlorite + K-feldspar ± biotite ± calcite ± sericite are found locally replacing prograde and peak assemblages.

[8] Figure 1c illustrates P-T conditions for the Strafford Dome compiled from T. Menard (unpublished thesis, 1991) and Menard and Spear [1994]. Peak P-T conditions range from ca. 450°C, 5 kbar near the Monroe Fault to the east to ca. 600°C, 10 kbar near the core of the dome. P-T paths for low-grade rocks are ill-constrained but higher grade rocks experienced an episode of near isothermal loading, attributed to nappe emplacement and associated with progressive crenulation cleavage development, followed by isobaric heating to peak conditions. Retrograde paths are poorly preserved suggesting both rapid exhumation and limited access of retrograde fluids.

3. Methods

[9] Samples used in this study are from the garnet and staurolite/kyanite grades (Figure 1) and include both oriented samples (“SD” in sample name) as well as unoriented samples from the Menard thesis collection (“TM” in sample name) (T. Menard, unpublished thesis, 1991). The biotite grade rocks in this region are the focus of a subsequent study due to there being a less developed P-T framework from previous studies and additional complications when analyzing these low-temperature rocks. Microstructural analysis was conducted to identify quartz in different fabric domains such as matrix foliation, inclusions defining foliations within garnet, and pressure shadows. These areas were examined by cathodoluminescence (CL) imaging (see below) and target areas for application of the TitaniQ thermobarometer were microdrilled from the thin sections and mounted in 1″ epoxy rounds.

[10] Cathodoluminescence (CL) imaging and spectroscopy were used in the 415 nm λ-specific mode to characterize Ti distribution within a single crystal and between neighboring grains [Rusk et al., 2006; Spear and Wark, 2009; Wark and Spear, 2005]. The Gatan MonoCL attached to the Cameca SX-100 electron microprobe at Rensselaer Polytechnic Institute (RPI) was used, with a 10 nA beam current and 15 kV accelerating potential maintained during analysis. Dwell times of c. 0.5–1.0 ms·pixel−1 were held constant without any pronounced streaking (see, Spear and Wark [2009] for details). Back-scattered electron (BSE) imaging for textural analysis and phase identification was done with a 20 nA, 15 kV incident electron beam. X-ray elemental mapping of garnet and surrounding matrix phases was done with beam currents of 200 nA and 15 kV. Garnet maps for the elements Al, Ca, Fe, Mg, Mn, and Ti were collected, with tiling required due to large porphyroblast size (several mm in diameter) resulting in scan times between 4 and 6 h. Low-resolution thin section maps were conducted for the elements Ti, Fe, Ca, and Mg to locate and distinguish Ti-stoichiometric essential accessory phases (e.g., ilmenite, sphene, rutile), which are required for constraining Ti activity.

[11] The electron microprobe was also utilized to conduct quantitative spot analyses on garnet porphyroblasts to construct chemical profiles and calibrate X-ray maps, and on garnet-biotite and garnet-ilmenite pairs for geothermometry [Docka, 1984; Feenstra and Engi, 1998; Ferry and Spear, 1978; Martin et al., 2010; Pownceby et al., 1987, 1991; Tracy, 1982; Tracy et al., 1976; Woodsworth, 1977]. Garnet was analyzed for Si, Ti, Al, Cr, V, Mg, Ca, Mn, Zn, and Fe using a 15 kV accelerating potential and 20 nA current. Data was collected for 5–7 spots to assure consistency and the results averaged. Analyses were taken on regions of elemental highs and lows as revealed by X-ray maps. Ilmenite inclusions within garnet were analyzed for Na, Al, K, Ca, Ti, Mn, Fe, Cr, Mg, Si, and F using 20 s dwell times, 15 kV accelerating potential and 20 nA current.

[12] Quartz grains and domains in different structural contexts were first imaged by CL using a blue filter to reveal Ti zoning. Based on the results of these analyses, representative grains of different zoning patterns were analyzed by SIMS to quantify [Ti].

[13] Titanium concentrations were measured via secondary ionization mass spectrometry (SIMS) spot analyses utilizing the Cameca IMS 1280 at the Northeast National Ion Microprobe Facility at the Woods Hole Oceanographic Institute. All analysis spot locations were guided by CL imagery of target domains, where representative grains with different zoning patterns from various domain and structural contexts were analyzed. An O- primary beam under ultrahigh vacuum (<1 × 10−7 Pa) was used to sputter atoms from the top several atomic layers of the sample surface. An incident ion beam with an impact energy of c. 23 kV acceleration potential and 4.05 nA current was used. A 240 s presputter was used during sample analyses to remove surface contamination, followed by 10 analytical cycles analyzing 30Si, 40Ca, and 48Ti with dwell times of 5 s, 5 s, and 10 s, respectively; 48Ti was corrected for mass interference of 48Ca based on measured 40Ca intensities. Ion beam spot size varied slightly with focusing, and was typically 10–15 μm in diameter. Titanium concentrations were calculated based on a calibration curve of [Ti] versus 48Ti/30Si constructed from analyses of TitaniQ standards QTiP-7, QTiP-14, QTiP-38, and QTiP-39 [Thomas et al., 2010]. If zoning is present in quartz crystals, analyses from the same zone were averaged since analyses often were statistically identical and weighted averaging of these domains would reduce statistical uncertainty for the zone being interpreted.

[14] X-ray fluorescence bulk rock analysis of sample 09SD08A was collected to allow for forward stability modeling and modeling of volume percent quartz and Ti-oxides in P-T space (system MnO-Na2O-CaO-K2O-FeO-MgO-Al2O3-SiO2-H2O-TiO2, or MnNCKFMASHT), computed with the program Perple_X [Connolly, 2009]. These models are used in interpreting quartz and Ti-phase production/consumption to make interpretations regarding Ti growth or consumption of quartz in association with metamorphic reactions to explain zoning patterns, and to qualitatively assess activities of TiO2 during the rock's evolution. A representative 5 g sample of whole rock from sample 09SD08A was prepared for bulk rock chemical analysis, with all weathered edges removed. The sample was powdered in an aluminum ball mill and passed through 80 μm mesh sieve. The powders were sent to Franklin and Marshal's XRF lab and measured for major element chemistry with titration (for ferric iron estimation) and loss on ignition. Solution models utilized in the computed pseudosections include: ideal solution model for the ilmenite-geikielite-pyrophanite; omphacitic-pyroxene [Diener and Powell, 2010] (modification of the Green et al. [2007], omphacite); feldspar [Fuhrman and Lindsley, 1988]; chloritoid, hydrous cordierite, staurolite, garnet, and chlorite [Holland and Powell, 2011]; white mica [Auzanneau et al., 2010; Coggon and Holland, 2002]; Ti-Fe3+-biotite [Tajmanová et al., 2009] (extended to Mn solution after Tinkham et al. [2001]). In the models, quartz, rutile and sphene were considered as pure phases and water was considered to be in excess.

[15] Results of this study are presented for sample groupings based on proximity and metamorphic grade (Figures 1a and 1b), and are integrated with the previously established P-T-D context of T. Menard (unpublished thesis, 1991) and Menard and Spear [1994]. Reported errors are calculated to include all sources of analytical uncertainty to 1σ (supporting information Appendix A1), and TiO2 activities were assumed to be 0.97 (as per Ghent and Stout [1984]). The thermobarometric calibration published by Thomas et al. [2010] was used in this study for P-T estimation. Reasons for selecting this choice over the recently published calibration by Huang and Audétat [2012] will be discussed later in this manuscript (section 5.3). A data repository of all analyses collected can be found in supporting information Appendix B.

4. Results

4.1. Eastern Flank of the Dome: Garnet Grade

4.1.1. Petrography and Microstructures

[16] Sample TM623 is a crenulated garnet schist with the assemblage garnet + quartz + feldspar + biotite +muscovite + ilmenite + magnetite + rutile ± apatite  ± zircon ± opaques (Table 1; Figures 1a and 1b). In outcrop, the dominant foliation, defined by compositional banding and inferred to be the regional S1 schistosity, dips moderately–steeply to the east-southeast; crenulation lineations plunge shallowly to the south. In thin section, the S1 foliation is defined by compositional layering, preferred orientation of micas and ilmenite that is parallel to the compositional layering, and by grain size variation (Figure 2a). Crenulation cleavages (S2) tend to be parallel or anastomosing and asymmetric and are defined by pressure solution seams in the micaceous domains. Rare kink-banding (S3) is present and nearly perpendicular to the S2 crenulation cleavage throughout the sample. Quartz grains in S1 microlithons are flattened parallel to S1; quartz in hinges of crenulated S1 folations is more equant. Garnet porphyroblasts lack pressure shadows and include the S2 crenulation cleavage (Table 2). Large, randomly oriented biotite porphyroblasts also include the crenulation cleavage. A quartz-rich domain in a crenulation cleavage hinge zone, where quartz grains predominantly display polygonal textures (i.e., grain boundaries meet at 120° triple junctions) and minor undulose extinction, was selected for further analysis. The quartz in this domain comprises a microlithon defining the S1 foliation that has been crenulated during S2 development. We denote crenulated S1 foliations in S2 microlithons as S1/S2 in subsequent sections.

Table 1. Sample Locations and Mineralogy
Sample NameUTM CoordinatesaMineralogyQuartz AbundanceZoneRock Type
  1. a

    UTM Zone: 18T.

TM6234852176.0283997.53Garnet + quartz + plagioclase + biotite + muscovite + ilmenite ± magnetite ± rutile ± zircon ± apatitedomainsGarnetGarnet schist
09SD08A4854879.0293209.30Garnet + quartz + biotite + muscovite + plagioclase ± titanite ± ilmenite ± rutile ± zircon ± opaquesabundantStauroliteGarnet schist
TM4554854879.0293209.30Garnet + plagioclase + muscovite + opaquesdomainsStauroliteGarnet schist; retrograde mylonitization
TM7834865448.0284272.90Kyanite + garnet + quartz + plagioclase + plagioclase + biotite + muscovite + tourmaline + ilmenite ± opaques ± zirconabundantStauroliteGarnet-kyanite schist
10SD03B24865448.0284272.90Garnet + amphibole + quartz + plagioclase + biotite + muscovite + ilmenite ± opaques ± zirconabundantStauroliteGarnet-amphibole schist
10SD03D4865448.0284272.90Kyanite + garnet + quartz + plagioclase + biotite + muscovite + ilmenite ± opaques ± zirconabundantStauroliteGarnet-kyanite schist
Figure 2.

Cross-polarized, transmitted light photomicrographs for samples from the Strafford Dome. (a) Montage photomicrograph along the length of the thin section for sample TM623. The S1 foliation is defined by quartz and mica domains, which has been crenulated in response to nappe emplacement resulting in an S2 crenulation cleavage. Both garnet and large randomly oriented biotite porphyroblasts include the S2 foliation. (b) Garnet porphyroblasts with linear inclusion trails (Si) of S1 that are rotated relative to the external foliation (Se) at an oblique angle (ϕ up to 35°; sample 09SD08A). Blue, dashed line shows the approximate grain boundary location for the garnets. Pressure shadows developed in response to inhomogeneous strain due to the rigidity of the garnet porphyroblasts. Sample is cut perpendicular to foliation and parallel to the lineation. (c) Crenulated quartz vein with dynamic subgrain rotation recrystallization, resulting in ribbon grains with associated subgrains (forming an oblique foliation; sample TM455). Box shows location of CL image from Figure 3d.

Table 2. Microstructure Summary of Samples
Sample NameQtz. Recryst. MechanismaIncluded Foliation in Grt (Si)Grt Growth Relative to Foliation DevelopmentbQuartz Occurrence
  1. a

    SGR: subgrain rotation recrystallization.

  2. b

    ⊃: syn-tectonic; P: porphyroblast growth.

TM623StaticS1; S2S2<PMatrix, inclusion, vein (posttectonic)
09SD08AStaticS1S1<P<S2Matrix, inclusion, vein (posttectonic)
TM455SGRS2S2<PMatrix, inclusion, vein (pretectonic)
TM783StaticS1/S2; S2S2⊃PMatrix, inclusion
10SD03B2StaticS1/S2; S2S2⊃PMatrix, inclusion
10SD03DStaticS2S2⊃<PMatrix, inclusion, vein (posttectonic)

4.1.2. Laboratory Results

[17] CL imaging of the matrix quartz from sample TM623 reveals quartz grains with homogeneous dark cores with brighter rims that range from being very narrow (<5 μm) to wider and patchy in appearance (Figure 3a). Transitions from dark cores to brighter rims appear fairly discrete. Linear bands of higher CL intensity are locally present. Five TitaniQ analyses yield [Ti] from 2.2 ± 0.2 to 3.4 ± 0.4 ppm for the dark cores and 5.6 ± 0.2 ppm for a bright rim (Figures 3a and 4).

Figure 3.

Summary blue λ CL images of quartz in various microstructural settings, displaying different Ti-distribution patterns. Ti spot analyses where shown indicate [Ti] in ppm. Image widths are c. 800–900 μm. Qtz = quartz; bt = biotite; feld = feldspar. (a) Image of sample TM623 showing hinge zone of S1/S2 crenulation cleavage with zoned quartz. (b) Quartz from a garnet pressure shadow in S2 foliation displaying significantly larger bright rims than typically observed in inclusions (sample 09SD08A). Some grains are homogeneous and also display bright cores. (c) Quartz vein concordant with S2 foliation displays homogeneous Ti distribution (sample 09SD08A). (d) Deformed quartz vein exhibiting subgrain rotation recrystallization (sample TM455). The dynamically recrystallized grains (darker) and ribbon subgrains (brighter) form an oblique foliation.

Figure 4.

Summary diagram of Ti analyses collected with respect to sample and microstructural location. Concentrations are in ppm. See supporting information Appendix B for complete data tables corresponding to all data shown in the figure.

[18] Samples that define the location of the garnet-isograd in proximity to TM623 constrain peak pressures c. 6 kbar, with peak temperatures in the range of 500–550°C [Menard and Spear, 1994; T. Menard, unpublished thesis, 1991]. We attempted to confirm Tmax for this sample by measuring garnet-biotite pairs (using larger biotite porphyroblasts in contact with garnet, both of which overgrew the S2 crenulation cleavage) for Fe-Mg exchange geothermometry (Table 3). Four calibrations were selected for this study: Thompson [1976], Ferry and Spear [1978], and two calibrations from Bhattacharya et al. [1992] considering mixing parameters for the pyrope-almandine asymmetric regular solution presented by Ganguly and Saxena [1984] and Hackler and Wood [1989]. The range of calculated temperatures for the four calibrations is small (523–531°C; assuming P = 6 kbar) with an uncertainty of around ±25°C.

Table 3. Garnet-Biotite Chemical Analyses for TM623a
 Garnet (n = 12)Biotite (n = 3)
OxideWt %cpufbWt %cpuf
  1. a

    Cation per unit formula takes into consideration Al in the M1 site mixing solution models with total tetrahedral and octahedral sum of 6.9 with 11 oxygens for biotite.

  2. b

    cpuf = cations per unit formula.

  3. c

    Determined by two-site mixing and cation sum of 8 for 12 oxygens.

  4. d

    n.a. = not analyzed.

  5. e

    n.d. = not determined.


4.2. Core of the Strafford Dome: Kyanite-Staurolite Grade

4.2.1. Petrography and Microstructures

[19] Sample 09SD08A (Figure 2b) is a garnet schist from the western flank of the dome's core with the assemblage garnet + quartz + feldspar + biotite +  muscovite + rutile ± ilmenite ± apatite ± sphene ± zircon ± opaques (Table 1). In outcrop, WNW-dipping S2 schistosity dominates and open crenulation cleavage (S1/S2) is locally apparent. The S2 foliation is a spaced foliation defined by compositional layering and the preferred orientation of micas and opaques. Locally, crenulations of S1 are observed in S2 microlithons. There is a relatively small volume of S2 cleavage domains present (<5–10%) that consists of biotite + muscovite + ilmenite. Quartz grains locally display a shape-preferred orientation (long axes parallel to S2) where bound above and below by biotite grains in cleavage domains. Quartz-rich S2 microlithon domains contain grains that often display more equant grain shapes and foam textures. The thin section contains several garnets 3–4 mm in diameter with linear inclusion trails (primarily quartz and ilmenite) that define an internal foliation (S1) oblique to the matrix foliation (ϕ up to 35° is typical; Table 2; Figure 2b). Some garnets have flattened quartz inclusions that define the included S1 foliation. In other garnets, there is no apparent shape-preferred orientation of quartz inclusions. Garnets typically have strain shadows consisting of quartz, feldspar and mica. Coarse-grained quartz veins cut through the sample and are concordant with the S2 foliation. Representative spots from a garnet inclusion suite, a garnet strain shadow, the matrix (quartzo-feldspathic domain), and a quartz vein were selected for further analysis.

[20] TM455 (Figure 2c) is mylonitized garnet schist that contains asymmetric C′-type shear banding, with the foliation defined by compositional layering and preferred orientation of micas. Two types of garnet are present within the sample: numerous small (sub-mm) inclusion-free grains (in micaceous domains) and larger (up to 2 mm) porphyroblasts that preserve a crenulation cleavage (crenulated S1 foliations) defined by quartz and ilmenite inclusions present in more quartz-rich domains (Table 2). Mica fish are present and, locally, some of the larger garnets are porphyroclastic and display chlorite and biotite tails. Pressure solution seams are present in micaceous domains (composed of muscovite + biotite + ilmenite). Highly strained quartz veins exhibit extensive subgrain rotation recrystallization. Subgrains adjacent to ribbon quartz define an oblique foliation at a low angle relative to the ribbon grain (Figure 2c). A large garnet with an included foliation and a sample of the recrystallized quartz vein (including ribbon and subgrains) were selected for further analysis.

4.2.2. Laboratory Results

[21] X-ray mapping of garnet in 09SD08A reveals a slight decrease in Mn from core to rim, with a very thin (<10 μm) bright rim in Mn that surrounds the grain (supporting information Appendix C). Fe and Mg increase slightly from core to rim, while Ca exhibits an unusual blocky zonation. CL images of quartz inclusions defining the internal, S1 foliation from the core of the garnet reveal them to have dark cores ([Ti] of 2.6–3.6 ppm) with thin (<5 μm), bright rims (Figures 4 and 5c); the rims were too narrow to analyze with SIMS. Brighter patches occur within dark grains, sometimes adjacent to the rims, and yield [Ti] in the range of 5.3–7.1 ppm. Bright, linear features are locally observed in CL images of included grains; an analysis from a high density area of these bright bands yielded [Ti] of 8.4 ± 0.4 ppm.

Figure 5.

Forward stability model for 09SD08A in the system MnNCKFMASHT. Representative bulk rock chemistry is (determined from X-ray fluorescence analysis, in weight percent oxides): 77.22% SiO2, 0.80% TiO2, 13.23% Al2O3, 4.56% FeOtot, 0.06% MnO, 2.45% MgO, 1.31% CaO, 2.02% Na2O, 2.52% K2O. Water is considered to be in excess. White-colored fields are invariant (eight phases stable), with increasing intensive degrees of freedom toward dark blue. Chl, chlorite; mica, white mica; zo, zoisite; law, lawsonite; sph, sphene; ab, albite; plag, plagioclase; gt, garnet; ru, rutile; feld, feldspar; bio, biotite; ky, kyanite; ilm, ilmenite; sill, sillimanite; hCrd, hydrouscordierite; and, andalusite; acti, actinolite; mic, microcline; ctd, chloritoid; st, staurolite.

[22] Quartz grains in three distinct matrix fabric domains of 09SD08A were also analyzed: strain shadows on garnet porphyroblasts, quartzo-feldspathic microlithons that define the S2 foliation, and an undeformed quartz vein. Grains for the quartz-rich S2 domains contain larger, ribbon quartz grains and finer grained quartz. In CL, the finer-grained crystals typically show dark cores that grade to brighter mantles (from 4.8–9.2 ppm Ti; Figure 6d) surrounded by dark rims of variable thickness. Some grains are homogeneous in CL and completely dark. Locally, a sharp contact between the bright mantles and dark rims is very apparent; where analyzed, a dark rim yields [Ti] of 6.7 ± 0.3 ppm and exhibits polygonal texture with the neighboring grains. The larger quartz grain in the matrix (S2 foliaton) is more homogeneous in CL (7.2–7.4 ppm Ti; Figure 6d). Quartz grains in the strain shadows (Figure 3b) have dark cores, bright mantles that are much more extensive than in the matrix S2 quartzo-feldspathic domains, and thin dark rims. The bright mantles yield [Ti] up to 9.8 ± 0.3 ppm, with similar low [Ti] cores (c. 3.7 ± 0.4 ppm). Quartz from the vein in the sample is homogeneous in CL (3.4 ± 0.2 ppm Ti; Figure 3c).

Figure 6.

Integrated modeling, zoning patterns, and Ti analyses for sample 09SD08A. (a) Previously constrained P-T history [Menard and Spear, 1994] is overlain on the modeled volume percent of quartz. Warmer colors (e.g., red) represent more volume of quartz in the rock than cooler colors. (b) Volume of rutile allows for better constraint on activity-constraining phases in P-T space. (c) CL image showing quartz inclusions defining S1 foliation in garnet with narrow, bright rims (resulting from Ti volume diffusion from garnet into quartz; cf Spear and Wark [2009]). (d) Matrix quartz (S2 foliation) that display typical dark core and bright rims, but with an additional dark rim that forms a sharp contact with the other zoning present in the grain. Secondary growth resulting from a quartz-producing reaction is the probable cause of the rim. Electron beam damage caused by BSE imaging is the cause of the false-bright region outlined by the dashed line. Dark striations in this bright region are artifacts of polishing after the BSE imaging. Low Ti cores in inclusion quartz preserves early prograde growth, which must have occurred c. 350°C based on isopleth and the general intersection with a steady state geotherm (1). Menard and Spear [1994] identified a c. 2 kbar pressure increase due to nappe emplacement at c. 500°C, with peak-P at 8.7 kbar (2). An isobaric heating path following peak-P must have occurred and can be constrained through the Ti isopleths from gradationally zoned matrix quartz (3). The polygon for this region represents the Ti increase observed through this gradation to determine the heating path experienced by the rock. A retrogressed rim resulting from an influx of Si-fluids (since no-quartz producing reactions would occur during retrogression) would occur c. 400–450°C, depending on the depth of the rocks during fluid infiltration (4).

[23] X-ray mapping of a large garnet porphyroblast in sample TM455 showed the garnet to be fairly chemically homogeneous, except for a distinct rim that contains a significant increase in Ca and a decrease in Mn. CL images of quartz inclusions within the garnet display grains with patchy zoning (dark regions as low as 2.8 ± 0.3 ppm Ti; brighter regions up to 5.1 ± 0.3 ppm Ti; Figure 4). Ti analysis of the recrystallized quartz vein revealed Ti concentrations of c. 4.8 ± 0.4 ppm for the ribbon grains, with lower [Ti] for the dynamically recrystallized grains (c. 3.6 ± 0.5 ppm; Figure 3d).

4.3. Northeast of the Strafford Dome, Along Strike: Kyanite-Staurolite Grade

4.3.1. Petrography and Microstructures

[24] Samples TM783, 10SD03B2, and 10SD03D (Figure 1) are kyanite-garnet schists with the assemblage garnet + kyanite + quartz + plagioclase + biotite + muscovite + ilmenite ± amphibole ± tourmaline ± rutile ± opaques ± apatite ± zircon (Table 1). Crenulation cleavage is apparent in outcrop and both S1 and S2 foliations dip moderately to the NE; S3 kink-banding is locally apparent. As in outcrop, asymmetric crenulation cleavage (S1/S2) is apparent in thin section. There are gradational transitions between the S2 cleavage domains (composed of biotite, muscovite, kyanite; c. 40% mica domains) and microlithons, with smooth, parallel-anastomosing cleavage domains present. Large biotite porphyroblasts statically overgrew the matrix to include the crenulation cleavage. Undulose extinction in quartz is common, and quartz domains in both the matrix and included foliations often display polygonal textures. Some quartz grains are locally elongated parallel to the foliation they define; in the matrix this apparent shape-preferred orientation is limited to grains neighboring micas in S2 cleavage domains. The samples contain numerous large (up to c. 12 mm) garnet porphyroblasts. Large garnet porpyroblasts have inclusion suites defining an S1/S2 crenulation cleavage in the core that grade outward into spiral S2 inclusion trails. Smaller porphyroblasts have linear inclusion trails that extend continuously into the external S2 foliation (see TM783 in Figure 7 and 10SD3D in supporting information Appendix C; Table 2). Ilmenite grains in the sample are limited to inclusions in the cores of garnet porphyroblasts (S1/S2 crenulations). Coarse-grained quartz veins are largely concordant with the S2 foliation. Two large garnet porphyroblasts with spiral inclusion trails and one small garnet porphyroblast with linear inclusion trails were selected for further analysis.

Figure 7.

(a) Syn-tectonic “snowball” garnet from sample TM783, with (b–f) inset CL images of quartz inclusions. Boxes illustrate the location the CL images were captured from. Figure 7b shows linear bright features are commonly present in quartz inclusions and correspond to planar defects. Figures 7c–7f show false-color CL images of inclusions in which quantitative spot analyses were conducted, including matrix S2 foliaion (Figure 7c), rim (S2 foliation) (Figures 7d and 7e), and core (S1/S2 crenulation cleavage) (Figure 7f) grains. Ti spot analyses are shown with measured [Ti] indicated in ppm. Note the CL mirror was offset from center in these images resulting in erroneous grading in CL intensity.

4.3.2. Laboratory Results

[25] X-ray mapping of the garnet porphyroblasts show a notable decrease in Mn and an increase in Mg from core to rim for TM783 and 10SD03B2 (e.g., TM783 in Figure 8; both garnets have spiral inclusion trails); 10SD03D (linear inclusion trail) is chemically homogeneous except for a slight variation in Mn that appears to form an annulus, the zoning of which crosscuts the included foliation (supporting information Appendix C). CL images were collected from inclusions throughout the garnet porphyroblasts and surrounding matrix grains. While concentric zoning in quartz grains (e.g., darker cores and brighter rims) is locally apparent in these samples, quartz zoning is more variable in terms of the prevalence of features such as patchy zoning and bright banding.

Figure 8.

X-ray element maps of syn-tectonic garnet from sample TM783 showing a decrease in Mn and an increase in Mg from core to rim (attributed to increasing P during nappe emplacement). Profile plot (below) is from a to a′. Secondary growth results in increased Ca (decreased Mn) around the rim.

[26] In TM783 (snowball garnet), the lowest [Ti] obtained (3.7 ppm) is associated with quartz inclusions defining S1/S2 crenulations in the core of the garnet. There, some zoning from darker cores to brighter rims is apparent, but core [Ti] varies from 3.9–5.2 ppm. The highest [Ti] (7.0–7.8 ppm) are obtained from the garnet-matrix interface (i.e., where grains are only partially included in the garnet and therefore communicate with the matrix grain boundary network) and matrix domain where CL images reveal zoned grains with dark cores and bright rims neighboring relatively homogenous bright grains (e.g., Figure 7e). Quartz from the S2 linear inclusion trail (near top of porphyroblast, where box “d” is located) yielded intermediate [Ti] (4.0–5.5 ppm) relative to garnet core and matrix domains. In comparison, 10SD03B2, which has similar garnet chemical zoning and microstructures to TM783, has Ti of 4.8–5.8 ppm associated with dark quartz cores (in CL) defining S1/S2 crenulations in the core of the garnet. More brightly zoned regions reach 7.8 ± 0.6 ppm Ti near the garnet's core and up to 9.7 ± 0.5 ppm near the rim. Garnet-ilmenite thermometry was conducted on TM783 (Table 4) resulting in Tmax estimates of 521 ± 35°C (2σ) [Martin et al., 2010, calibration].

Table 4. Ilmenite-Garnet Exchange Thermometer Analyses for TM783a
Trav. No.XMnbXFeIIcXMgdnXMnbXFeIIcnFinal (lnKD)T (°C)2σ (°C)
  1. a

    Note: Other cations present in garnet's X-site (e.g., Zn) were excluded from mole fraction calculations for these samples due to the negligible partitioning of these elements.

  2. b

    XMn = Mn/(Mg+Ca+Mn+Fe2+).

  3. c

    XFeII = Fe2+/(Mg+Ca+Mn+Fe2+).

  4. d

    XMg = Mg/(Mg+Ca+Mn+Fe2+).


[27] CL imaging of quartz grains associated with S2 in sample 10SD03D locally reveals slightly more systematic zoning. Quartz inclusions in garnet in 10SD03D have low [Ti] (3.0–3.8 ppm) dark cores and bright rims with [Ti] up to 8.6 ± 0.3 ppm (Figure 9). Quartz in the matrix or near the garnet-matrix interface shows similar zoning, but the bright rims present only reach 6.6 ± 0.3 ppm where analyzed.

Figure 9.

(a) Photomicrograph of a garnet porphyroblast from sample 10SD03D with linear S2 foliation inclusion trails that extends continuously with external foliation, S2. (b and c) Matrix quartz near the rim of the garnet that show quartz that grade from low-Ti cores to high-Ti rims (similar to that observed in 09SD08A). The thin section was cut perpendicular of the dominant S2 foliation and parallel to the dip direction (049°/29° = dip azimuth, dip. Garnet growth was late syn to postkinematic with respect to S2 foliation development, with inclusions being primarily quartz and ilmenite.

Figure 10.

Summary P-T data for groupings A–C designated in Figure 1b using the TitaniQ thermobarometer calibration by Thomas et al. [2010], with comparison P-T paths from T. Menard (unpublished thesis, 1991) and Menard and Spear [1994]. Note the striking similarities between this study and previous studies in Figures 10b and 10c. (a) TM623 structurally lies between TM637 and TM549; P-T conditions are expected to reflect this structural position relative to the isograds. Due to the lack of definitive rims/mantles in quartz of different [Ti], interpreting the analyses is cautioned (see discussions for more detail). However, garnet-biotite exchange thermometry constrained peak-T to c. 525°C. This reflects expected Tmax for the sample. (b) With 09SD08A collected from the same outcrop as TM458a, the generated P-T history from TitaniQ thermobarometry should overlap the results from previous studies. Fabric development of S2 occurred at peak P resulting from nappe emplacement and clearly overlays the previously constrained histories. Polygon graded from blue to red with increasing temperature represented the heating path constrained by graded matrix quartz (Figure 6). Analyses collected for TM455 are excluded because dynamically recrystallized grains and quartz veins were measured, which may contain [Ti] not representative of temperatures of deformation (see discussion for more details). (c) This sample has the advantage of constraining three points between peak P associated with burial and fabric development, better confining the thermobarometric histories of these rocks. Garnet-ilmenite exchange thermometry (this study) constrained peak temperatures at c. 520°C, which overlaps with expected Tmax from TM445.

5. Discussion

5.1. TitaniQ Results Relative to P-T-D Path Constraints

[28] In order to determine where [Ti] isopleths lie in P-T space, the activity of TiO2 must be known or assumed. In this study, a TiO2 activity of the growth medium of 0.97 was assumed for all samples based on the reaction studies for typical metapelites by Ghent and Stout [1984]. An activity of c. 1.0 is supported by using the AX program (v. 2 for windows platform) [Holland, 2008] on the electron microprobe chemical analyses of the ilmenite traverses in TM783. The program calculates activities of mineral endmembers using microprobe data for ilmenite assuming ideal two-site mixing for three oxygens for the Ti-Fe-Mn-Mg solid solution between ilmenite-hematite-pyrophanite-geikielite, consistently returning ilmenite activities >0.90 [Ghent and Stout, 1984].

5.1.1. TM623 (Garnet Grade)

[29] Figure 9a presents results for TM623 along with P-T paths established for nearby samples by T. Menard (unpublished thesis, 1991) and Menard and Spear [1994]. Two isopleths are shown for quartz grains in the hinge of the S1/S2 crenulation cleavage: 2.2 ppm represents the lowest [Ti] from a dark core (in CL images) of a zoned quartz grain and 5.6 ppm represents the highest [Ti] obtained from a bright rim. Neither isopleth intersects with the Tmax estimate from garnet-biotite thermometry (Figure 10a), or with the polygons representing P-T path constraints for nearby samples. Microstructural evidence suggests that the garnet and biotite both postdate crenulation cleavage development because the crenulation cleavage is included in both minerals, with no evidence for tightening or modification of crenulations in the matrix relative to inclusion suites. Biotite porphyroblasts also have random orientations, supporting static recrystallization. Therefore, no analyses of quartz appear to represent peak P-T conditions, but are instead likely recording intervals of the early prograde path. Sluggish Ti diffusivities and the lack of quartz-producing reactions (none are observed in the pseudosection model for this sample) account for the lack of quartz equilibration during peak metamorphism in this sample. This may also be attributed the lack of deformation near peak T, resulting in no dissolution-precipitation creep to drive new quartz growth. We also note that CL images show relatively sharp boundaries between dark and bright zones, suggesting no substantial modification of these zones via volume or pipe diffusion.

[30] While the dark cores (CL) in quartz observed in this sample (and in virtually all samples analyzed in this study) could be either relics of early prograde (low T) quartz formation during burial or inherited from the protoliths, our preferred interpretation is that these cores most likely preserve early prograde conditions since the protoliths are pelites that would have received detrital quartz from igneous, metamorphic, or recycled sedimentary sources and thus would have variable [Ti] (C. K. McWilliams, Indiana University, unpublished thesis, 2008). The cores and rims of zoned quartz grains likely formed via precipitation during the two main stages of fabric development: (1) S1 spaced foliation defined by quartz-feldspar (Q-F) and mica (M) domains, and (2) and S2 crenulation cleavage. Precipitation of quartz in Q-F domains and hinge zones of S1/S2 crenulation cleavage is likely the result of pressure-solution attending foliation development and/or production of quartz via metamorphic reactions [c.f. Williams et al., 2001].

5.1.2. 09SD08A & TM455 (Kyanite-Staurolite Grade)

[31] For sample 09SD08A, garnet growth was intertectonic with respect to burial and nappe emplacement, with S1 preserved as the internal foliation. Pressure shadows formed on the grain as a result of inhomogeneous strain (contemporaneously with crenulation cleavage development) due to the rigidity of the porphyroblast relative to the groundmass during S2 foliation development; garnet rotation occurred during this fabric development and causes the included S1 foliation to become oblique with the external S2 foliation. Lack of abrupt or steep compositional gradients of major element zoning in garnet suggests garnet growth was associated with only a small segment of the P-T path. A bright garnet rim (Mn X-ray maps) is attributed to garnet resorption by prograde reactions releasing Mn that diffuses back into the garnet [e.g., Kohn and Spear, 2000].

[32] Figure 9b presents a summary of the results obtained for 09SD08A integrated with existing P-T data. Four isopleths are shown: 2.6 ppm represents minimum [Ti] of dark cores of zoned quartz in the garnet-matrix interface (also the interface between quartz of the included S1 and external S2 where quartz grains are not fully enclosed in garnet and foliation associations are not clear cut); 7.8 ppm represents brighter mantle regions (around dark cores) grading to 9.8 outward; and 6.7 ppm is associated with the surrounding dark rims in the garnet pressure shadow and matrix S2 foliation (Figure 3c). Relative ages inferred from the textural relationships are: 2.6 ppm is the oldest isopleth; 7.8 and 9.8 ppm are intermediate in age; and 6.7 ppm is the youngest. T. Menard (unpublished thesis, 1991) estimated peak burial (prenappe emplacement) conditions to be 7.3 kbar and 500°C for nearby sample TM458a (Figures 1a and 1c); minor heating was interpreted to follow before a c. 1.5 kbar isothermal pressure increase occurred (up to 8.7 kbar), resulting from nappe emplacement. Peak T of 600°C was determined for a sample to the southeast, in the core of the dome (TM732) (T. Menard, unpublished thesis, 1991).

[33] A pseudosection model for sample 09SD08A can be found in Figure 5. Contouring quartz volume in P-T space (Figure 6a) indicates a dominant quartz-producing reaction due to a chlorite-out, biotite-producing reaction. This reaction requires balancing by removal of K from muscovite, which drives the chemistry to the muscovite-paragonite-margarite solvus, resulting in the stability of two white micas. Plagioclase is also required for balancing, and the result is the production of quartz. A small region of staurolite-production is observed, which produces quartz in the process. However, due to the Al-undersaturation in this sample, Al is largely incorporated by plagioclase, and the stability of staurolite is very reserved in P-T space. Ilmenite and rutile are stable around this reaction, suggesting the reaction would be occurring in a TiO2-saturated system (aTiO2 is ∼1.0).

[34] Similar to TM623, the oldest isopleths associated with the lowest [Ti] do not intersect polygons representing P-T path constraints for nearby samples. The low Ti cores of zoned quartz found as S1 inclusions in or near matrix S2 domains likely represent recrystallization of quartz during the early prograde path, perhaps during burial. The thin bright rims present on quartz inclusions are probably attributed to volume diffusion of Ti from the garnet, postinclusion [Spear and Wark, 2009] and potentially provides insight into the time scales of prograde metamorphism postgarnet growth to cooling below 450°C [Spear et al., 2012].

[35] The 7.8 and 9.8 ppm isopleths more firmly overlap with existing constraints. The associated gradient of [Ti] is observed (Figure 5d) increasing toward the rim of quartz in pressure shadows. The isopleths plot in a region in P-T space that corresponds to quartz production (determined through isochemical forward stability modeling), suggesting that growth due to quartz-producing reactions is a probable explanation. The models (Figure 5) also suggest that the metamorphic reaction may be a chlorite-out, quartz-producing reaction occurring at peak pressure. This strongly suggests new quartz was being precipitated on matrix quartz during isobaric heating postnappe emplacement due to metamorphic reactions, starting at c. 500°C (assuming 8.7 ± 0.5 kbar from Menard and Spear [1994]) and increasing to peak T (540°C). We infer that the lower-Ti (6.7 ppm) overgrowths must have occurred during the retrograde path because there is no evidence for another event to drive an increase in pressure needed to account for a decrease in [Ti] as part of the prograde path. Pseudosection analysis suggests that quartz should be consumed by retrograde metamorphic reactions (Figure 6a), therefore an influx in fluids saturated with Si must have occurred during exhumation, probably at c. 400–450°C or lower, depending on the depth of the fluid influx. Dissolution-precipitation creep involving Ti mobilization during retrogression is also possible, although considering the short metamorphic time scales suggested by Spear et al. [2012], this mechanism may not be sufficient for the amount of material transport observed in such short time scales.

[36] The interpreted influx of siliceous fluids needs to be considered since open systems would obviously have an effect on the models. It is unlikely that this additional silica would have any effect on our models since the volume added is minimal (observed rims in CL) and the rock is very siliceous (c. 72% SiO2 by weight) to begin with (quartz is already considered in excess). The addition of other components in this fluid is a possibility, however since quartz veins present in the sample contain no evidence of deformation (most likely forming as a result of this fluid flux) and are devoid of any other phases, it is unlikely that any other components were added to the system at this time in any significant concentrations. These veins were avoided when preparing samples for bulk rock chemical analyses.

[37] TM455 is different from all other samples analyzed as part of this study due to the mylonitic fabric. We infer that TM455 shares a similar prograde path as 09SD08A. Quartz in garnet inclusions preserve patchy zoning (dark cores and bright rims; [Ti] ∼2.8 ppm and 5.5 ppm, respectively) and associated [Ti] as observed and determined locally in 09SD08A. While ribbon quartz grains preserve similar [Ti] to bright patches, dynamically recrystallized grains in the recrystallized quartz vein have lower [Ti] (∼3.6 ppm). A decrease in [Ti] in the recrystallized grains could be associated with deformation associated with increasing pressure or decreasing temperature, however interpreting in this context should be cautioned since recrystallization mechanisms associated with regimes below grain boundary migration have shown to be unsuccessful at efficiently reequilibrating the subgrains with respect to Ti [Grujic et al., 2011]. Based on the chlorite and biotite wings on garnet porphyroclasts, deformation is interpreted to accompany retrograde metamorphism and exhumation, and therefore associated with lower temperatures and pressures.

[38] The constrained retrograde conditions for 09SD08A are remarkably similar to that suggested by the dynamic recrystallization microstructures (subgrain rotation recrystallization) exhibited in deformed quartz veins in sample TM455 [Stipp et al., 2002]. This may suggest a coeval relationship between timing of fluid influx and mylonitization occurring during the metamorphic retrograde path. Further reconnaissance for shear sense indicators and quartz fabric opening angle thermometry [Kruhl, 1996, 1998] could further support this hypothesis.

5.1.3. TM783, 10SD03B2, 10SD03D (Kyanite-Staurolite Grade)

[39] Figure 9c presents a summary of the results obtained for samples TM783 and 10SD03D integrated with existing P-T data. Four isopleths are shown: 3.7 ppm represents the dark core of a zoned quartz inclusion from the S1/S2 crenulation cleavage included in the core of a garnet porphyroblast in TM783; 5.5 ppm represents an average of three analyses from TM783 of relatively unzoned quartz in a linear (with S2) inclusion trail in garnet; 7.8 ppm comes from a bright rim of zoned quartz in the S2 matrix of TM783; and 8.6 ppm represents a bright rim region of a zoned quartz inclusion in garnet inclusion trail linear with S2 in sample 10SD03D. The 3.7 ppm isopleth from sample TM783 is the oldest based on textural relationships; the relative timing of the 5.5 and 7.8 ppm isopleths cannot be directly inferred without making assumptions about the origin of the quartz.

[40] As a comparison to samples analyzed here, we compare our results to a nearby sample (TM445) to the northeast of location B (just off the location map shown in Figure 1a). Menard and Spear [1994] constrained conditions associated with S1 development for TM445 at 5.6 kbar and 475°C, and their projected P-T path suggests peak-metamorphic conditions c. 6.7 kbar and 525°C (TM445). In medium to high grade samples (e.g., TM783, 10SD03B2 and 10SD03D), regardless of microstructural context, low-Ti cores are preserved in quartz grains similar to those observed in the lower grade rocks to the southeast (i.e., TM623). Inclusions of quartz in garnet in these samples, however, have bright CL rims that are wider (more extensive) than those observed in many other samples (e.g., sample TM783; Figures 7c and 6d) and similar in width to the rims on some medium grade matrix S2 quartz (e.g., sample 09SD08A; Figure 3d). Furthermore, some medium to high grade samples have quartz inclusions in garnet that have very thin rims that result from volume diffusion of Ti into the inclusion from garnet [c.f. Spear and Wark, 2009; Spear et al., 2012](e.g., sample 09SD08A; Figure 3b). Because quartz inclusions in garnet with wide CL rims are only observed inside of syntectonic garnet, an event that produced these high Ti rims must have occurred prior to garnet growth. In as much as kyanite in sample TM783 is crenulated by nappe-stage cleavage development, the kyanite and quartz producing reaction [c.f. Spear and Wark, 2009]:

display math

is the most likely source of the high-Ti rims on this sample.

[41] Isopleths from low-Ti cores in quartz inclusions do not intersect these P-T constraints and are interpreted the same as other samples: as early prograde quartz growth. Brecciated textures occasionally observed in inclusions in garnet (in CL; e.g., Figure 7d) are inferred to be the result of either late-stage fracturing (where evident by garnet alteration along fractures) or high strain features that form during crenulation; the garnet acted as a crucible to prohibit recrystallization and strain release.

[42] We infer that our TitaniQ analyses may constrain temperatures associated with kyanite- and quartz-producing reactions at 435 ± 9°C (5.5 ppm Ti at 5.7 kbar estimated by T. Menard (unpublished thesis, 1991); TM783 S2 inclusion quartz rims). Subsequently, T increased to 460 ± 9°C (8.6 ppm Ti at 5.7 kbar; 10SD03D S2 inclusion quartz rims). Minimal Fe/(Fe+Mg) increase during garnet growth suggests slight T increase during nappe emplacement. Nappe emplacement increased P by c. 1 kbar [Menard and Spear, 1994], with peak P conditions and S2 fabric development constrained to 474 ± 9°C at 6.7 kbar (7.8 ppm Ti; TM783 matrix S2 quartz rims). Note that this temperature is based on a single rim analysis; further rim analyses were difficult due to the small size of the rims and the potential for mixing between the two domains (core and rim). The formation of a crenulation cleavage must liberate and redistribute silica, at the very least removing it from the mica domains. This silica presumably gets precipitated in quartz domains and results in the observed Ti overgrowths. In samples where this is not shown, silica could possibly leave the rock entirely. This allows for the refinement of P-T histories in association with metamorphic reactions occurring or fabric development.

5.2. Summary of findings for the Strafford Dome

[43] The analyses from the Strafford Dome highlight the potential power of employing the TitaniQ thermobarometer in P-T-D studies of fabric development in metapelites. Findings from across the Strafford Dome appear to record several stages of the P-T-D paths, some of which are in agreement with those previously published by Menard and Spear [1994]. Six equilibration or quartz-forming events are preserved in the samples: (1) early prograde (dark CL cores, Ti = 2–4 ppm) attributed to precipitation of quartz in microlithons during S1 cleavage development; (2) kyanite-in quartz-producing reactions, determined through assessment of pseudosection modeling (e.g., bright CL rims on quartz inclusions, inclusions (inferred to be inherited) in S2 foliations, and dark cores of S2 matrix foliation, TM783, [Ti] = 5.5 ppm); (3) prenappe emplacement equilibration during isobaric heating (rims on quartz inclusions in 10SD03D, Ti = 8.6 ppm); (4) fabric development of S2 schistosity during progressive crenulation cleavage development (and associated dissolution in cleavage domains and precipitation in microlithons domains); (5) quartz-producing (chlorite-out) reaction during isobaric heating following nappe emplacement (graded quartz mantles up to 9.8 ppm, 09SD08A); and (6) precipitation (from siliceous fluid influx) and/or recrystallization during retrograde metamorphism and deformation (e.g., lower Ti overgrowths, 09SD08A matrix grains, Ti = 6.7 ppm).

[44] Our interpretations of the relative timing of (re)equilibration/(re)crystallization of quartz in the rocks history is usually coincident with quartz-producing metamorphic reactions or deformation, rather than a continuum of (re)equilibration with changing P or T due to burial or exhumation. Interestingly, no domains analyzed in matrix quartz or quartz inclusions were inferred to be associated with peak metamorphism (at peak T). This may be a result of the rate of the change in P (essentially instantaneous compared to the heating resulting from loading), no quartz-producing reactions at peak T, the apparent lack of deformation at peak T, and the short interval for peak metamorphism [Spear et al., 2012] which will not allow for reequilibration via volume diffusion.

[45] The formation of a crenulation cleavage must liberate and redistribute silica, at the very least removing it from the M domains. The dissolved quartz presumably gets precipitated in pressure shadows of garnet, fold hinges, or in microlithons and results in the observed Ti overgrowths observed in this study. If deformation is occurring with increasing T, zoning of quartz with increasing Ti is expected. In samples where this is not observed, silica could possibly leave the rock entirely, precipitating as veins elsewhere.

[46] The bright banding could be the result of dislocation substructure. Randomly distributed dislocations will migrate through the crystal lattice via dislocation glide and climb into lower energy arrays to form subgrain boundaries, and eventually, with high enough misorientations, will form new grain boundaries and recrystallized grains. The recrystallization process can enhance Ti diffusion by creating pathways for both pipe diffusion (through subgrain boundaries) and grain boundary diffusion (through newly formed grain boundaries).

[47] In this study, the microstructural context of quartz with respect to garnet was important for deciphering the relative timing of quartz-producing events relative to garnet growth and deformation. The integration of TitaniQ data with microstructural context, garnet compositional data, and pseudosection modeling could be a very powerful tool in defining distinct segments of P-T-D histories. In particular, using garnet isopleths to determine P-T conditions during porphyrobalst growth could be informative. In these samples, however, it is probable that modification of garnet chemistry may have occurred (inferred from blurred chemical zoning, rather than euhedral growth zoning patterns; see Mn element map in Figure 8). Therefore, garnet was used more for the microstructural context it places quartz in for better interpretations of [Ti], rather than for thermobarometry purposes.

5.3. Calibration Selection

[48] A recent TitaniQ calibration by Huang and Audétat [2012] was considered in addition to the Thomas et al. [2010] calibration for employed in this study. However, when the Huang and Audétat [2012] calibration was applied to the samples presented in this study, calculated temperatures were c. 100°C higher than that calculated by the calibration of Thomas et al. [2010] (Figure 11). These increased temperature estimates are higher than that previously recorded by other thermobarometric techniques for the Strafford Dome (T. Menard, unpublished thesis, 1991). Also, when comparing garnet growth to the garnet-in reaction line, the P-T path derived from the Huang and Audétat [2012] calibration would require greater overstepping of the garnet-in reaction than for the P-T path derived from the Thomas et al. [2010] calibration (Figure 11). For this reason, we employed the Thomas et al. [2010] calibration in this study.

Figure 11.

Comparison of Ti isopleth locations in P-T space when using the calibrations by Thomas et al. [2010] (used in this study) and Huang and Audétat [2012]. Figures 11b and 11c showed the striking similarities between the calculated P-T conditions for various deformation events when using the Thomas et al. [2010] calibration compared to previous studies. The Huang and Audétat [2012] calibration, however, results in temperatures c. 100°C higher and temperatures higher than experienced by the Strafford Dome (determined through previous studies and exchange thermometers that record Tmax). The garnet-in reaction line (orange dashed line) for sample 09SD08A is included for reference.

5.4. Implications for Thermobarometry of Polydeformed Terranes

[49] In metamorphic tectonites, dynamic recrystallization of quartz is common with the quartz chemically reequilibrating Ti during deformation [Kohn and Northrup, 2009]. Grujic et al. [2011], however, showed that complete resetting of [Ti] during dynamic recrystallization is only achieved through grain boundary migration above c. 540°C, due to the sluggish volume diffusion of Ti and short duration of contact metamorphism. In other words, in the absence of grain boundary migration effectively resetting [Ti] as the grain boundaries sweep through quartz domains, reequilibration requires Ti diffusion which is inefficient (for volume diffusion) at the temperatures at which bulging and subgrain rotation occur. Rather, pipe diffusion would be more effective at chemical redistribution during subgrain rotation recrystallization. In sample TM455, where subgrain rotation recrystallization is present in a deformed quartz vein, dynamically recrystallized grains have lower [Ti] than the relict ribbon quartz grains, suggesting pressure increase or temperature decrease during recrystallization (similar to that observed by Kohn and Northrup [2009]); however these interpretations are strongly cautioned since the mechanism for Ti reequilibration in dynamically recrystallized grains would probably be different than that for new quartz growth, and no correlation can be directly made between [Ti] and intensive (P or T) parameter perturbations.

[50] The preservation of dark quartz cores, as seen in CL, with low [Ti] raises an interesting point in this study that most quartz grains did not fully reequilibrate with respect to Ti, despite deformation. Rather, reequilibrated rims were observed (in the matrix and in inclusions) and quartz typically displayed polygonized rims due to static recrystallization or by quartz redistribution resulting from (a) a prograde quartz-producing reaction, (b) deformation that mobilizes SiO2, or (c) Si-charged fluid influx. One aspect that may influence this result is that in the Strafford Dome rocks analyzed, quartz is not typically the interconnected “weak” phase; mica is. Deformation in these rocks was likely concentrated in the mica-rich domains, and isolated quartz grains/domains underwent limited crystal-plastic deformation. Focused deformation in mica domains and lack of interconnectivity between quartz grains would inhibit reequilibration via dynamic recrystallization (e.g., grain boundary migration) at the recorded temperatures. Also, crenulation cleavage development, which results in SiO2 remobilization, cuts across compositional layering and quartz-rich domains.

[51] Due to the sluggish volumetric diffusion rates of Ti in quartz in an environment where dynamic recrystallization is not facilitating Ti redistribution, insight into numerous points in the rock's P-T-D history can be gained when coupled with observations regarding the various microstructural contexts in which quartz is found. This is a result of the quartz chemistry not “resetting” or homogenizing completely at elevated temperatures. Even in cases where grain boundary area reduction is the dominant mechanism for reducing internal free energy of the crystal lattice (such as in the samples investigated here), Ti zoning is effectively preserved. In this study, bright patches that appear internal to the quartz grains yield similar [Ti] to rims. Thus, we infer that patchy textures are an artifact of the thin section representing a 2-D slice of a 3-D grain. The examination of blurred versus sharp contacts between internal [Ti] domains within quartz speaks to the duration of metamorphic stages recorded in these samples for example, as modeled by Spear et al. [2012].

5.5. Remaining Challenges for TitaniQ Thermobarometry

[52] To accurately calculate temperatures, assumptions about component activities must be made. A study by Ghent and Stout [1984] suggests the typical aTiO2 for metapelites is 0.97, which was the value used in this study. However, to have this activity be effective in the (re)equilibration of quartz, a Ti-bearing phase must be stable and active in the metamorphism. This can be confirmed by contouring the volume percent of these phases in P-T space from the isochemical phase equilibrium models produced. For the calculated temperatures in this study, either ilmenite or rutile is present and changing in abundance as the result of ongoing metamorphism and activities of c. 1.0 can be assumed.

[53] Another complication that must be handled for accurate thermometry calculations is constraints on pressures during equilibration. For this study, a previous P-T framework was used to constrain pressures along the P-T-D history of the rocks. Because there is no preserved pressure information on early prograde growth of quartz, approximate temperatures may be constrained by the intersection with modeled steady state geotherms. Overestimating pressure (due to nature not behaving in steady state on geological time scales typically observed) could be minimized by incorporating a nonsteady state geotherm model (such as applied in Behr and Platt [2011]). Constraints on TiO2 activity of the growth medium at these early prograde conditions is also challenging due to the lack of Ti-phases present. The pressure overestimation and the activity underestimation, therefore, are counteracting and may reduce the total deviation from actual temperatures during that event when using the TitaniQ thermobarometer and allow for a good approximation. Similar complications arise for retrograde equilibration, such as that observed by the resultant overgrowth rims in 09SD08A matrix grains resulting from Si-influx during retrogression.

[54] Several [Ti] measurements on quartz veins were collected, and although basic interpretations about vein formation relative to deformation can be made, the confidence in interpreting the [Ti] measurements is low. Further studies on this should focus on the ability to constrain Ti activities for these veins and the ability to accurately calculate P or T if no rutile is present in the vein.


[55] Much appreciation is given to Keith Klepeis, Robert Badger, and Arne Bomblies for reviews that contributed to the thesis work this manuscript was derived from. Personal communications with Richard Law, Mark Caddick, and Don Stahr were insightful and appreciated. Assistance with figure development and refinement by Lizzie Hasting greatly improved the quality of the illustrations in this manuscript. This manuscript benefited from thorough, constructive reviews by Whitney Behr and an anonymous reviewer. Thanks are extended to Nobu Shimizu for assistance in data collection with the SIMS at Woods Hole Oceanographic Institute. This material is based on work supported by the National Science Foundation under grants EAR-0948529 (awarded to L. Webb) and EAR-0948530 (awarded to F. Spear and J. Thomas). In addition, the Vermont Geological Society is gratefully acknowledged for receipt of a student research grant (awarded to K. Ashley) in support of this work.