Quartz inclusions in garnet: Time capsules of early mountain building
Published:September 11, 2017
- PDF LinkChapter PDF
Kyle T. Ashley, Richard D. Law, Robert J. Bodnar, Kenneth A. Eriksson, 2017. "Quartz inclusions in garnet: Time capsules of early mountain building", Linkages and Feedbacks in Orogenic Systems, Richard D. Law, J. Ryan Thigpen, Arthur J. Merschat, Harold H. Stowell
Download citation file:
Much of the early prograde history in metamorphic rocks is lost due to overprinting at near-peak conditions or through retrograde modification during exhumation. Fortunately, inclusions encapsulated in rigid porphyroblasts may preserve a record of early burial conditions. Quartz inclusions in garnet porphyroblasts from the Strafford Dome, eastern Vermont, have homogeneous Ti concentrations ([Ti]) that differ from matrix quartz, which retains a history of Si-liberating metamorphic reactions and fluid influx. We applied growth-composition models to evaluate potential processes associated with Ti partitioning in quartz before encapsulation in garnet, including a model for constant-volume growth of quartz due to mineral dissolution-transfer processes and growth as a result of Si-liberating diagenetic and metamorphic reactions. Because these processes typically occur at low temperatures, quartz with exceedingly low [Ti] (<<1 ppm) would be formed and cannot account for the homogeneous Ti distribution at concentrations between 2.5 and 5 ppm observed in the sample. This suggests that chemical reequilibration through dynamic recrystallization must have taken place prior to encapsulation in garnet. Analysis of fluid and graphite inclusions with Raman spectroscopy in different microstructural settings allowed the characterization of fluid composition and temperature of microstructure development early in the prograde history. The findings from this study exemplify the utility of garnet hosts to shield inclusion minerals from chemical modification and recrystallization during later events. As such, they provide a window into the early stages of orogenesis and provide insights concerning the mechanisms controlling equilibration of quartz.
Our understanding of the pressure-temperature (P-T) evolution of rocks in ancient mountain belts is typically limited to the late prograde or peak metamorphism stages. This is in part due to continuous chemical and mineralogical equilibration as rocks are exposed to higher temperatures during the prograde stage. For instance, pressures and temperatures are typically determined through the exchange of divalent cations or through a set of net-transfer reactions in metapelitic assemblages above greenschist facies. With slower kinetics and lower diffusivities as temperature decreases after peak metamorphism, modification of this chemical signature may be limited, providing insight into the near-peak metamorphic conditions. Retrograde overprinting may preserve a new metamorphic paragenesis, providing evidence for the cooling path. In addition, thermochronology allows us to infer the rate of cooling that occurred. However, for early prograde metamorphism, our knowledge is limited due to the frequent participation of earlier-formed minerals in metamorphic reactions, chemical modification at elevated temperatures, and the susceptibility of minerals to recrystallization (which may modify mineral compositions). As a result of these processes, the early prograde conditions associated with metamorphism and mountain building are poorly understood.
Recent work suggests that our best opportunity for accessing the early history of a metamorphic rock may be through study of minerals encapsulated in rigid, poikiloblastic metamorphic hosts. Garnet represents such a suitable host due to: (1) its resistance to deformation and (relatively) high temperature for the onset of plasticity, (2) its largely chemically inert character with respect to quartz, and (3) its tendency to passively overgrow and encapsulate (include) a variety of minerals during its formation. The inclusion of a mineral grain in garnet shields the grain from interaction with the bulk rock and later fluids, preserving the pre-encapsulation composition in some cases (particularly in unfractured garnets, where there are no pathways to allow chemical connectivity with the matrix). Although some exceptions exist (e.g., ilmenite inclusions in garnet continue to exchange Mn, Fe, and Ti at higher temperatures; Ague and Eckert, 2012), high field strength elements (e.g., Ti, Al, V) in silicate minerals are less likely to be modified by diffusion, especially in the absence of fluids. Assessments of trace-element distributions in quartz, therefore, may provide information about processes that occurred prior to encapsulation. One such example is incorporation of Ti in quartz. The partitioning of Ti between other phases and quartz can be the result of several processes, including:
continuous growth due to solution transfer processes, which would provide a continuous record for early prograde evolution;
growth associated with quartz-producing diagenetic and metamorphic reactions;
compositions retained from a sedimentary source in the case of detrital garnet grains, which would be beneficial in sedimentary provenance sourcing and provide insight into the ancient terranes that have been eroded away;
growth associated with discrete Si-influx events, which may be linked to tectonic processes (e.g., faults acting as a conduit for fluid infiltration);
extensive recrystallization (dynamic or static) that resets [Ti] in quartz prior to inclusion in garnet; and
early brittle-fracture, with planar bands and brecciated textures present in inclusions correlating with healing of these fractures.
Here, we present quartz growth models for Ti partitioning during diagenetic/metamorphic evolution, and we characterize the processes associated with the Ti distribution within quartz inclusions. Model results were evaluated with respect to sample 09SD08A from the Strafford Dome, eastern Vermont, which was described and characterized by Ashley et al. (2013). This sample was selected because the pressure-temperature (P-T) history of the area is well constrained by previous studies (Menard and Spear, 1994), including the linkage of heating profiles to quartz-producing metamorphic reactions (Ashley et al., 2013) and near-peak geospeedometry results on 09SD08A from diffusion studies in quartz (Spear et al., 2012) and garnet (Spear, 2014). Also, the sample shows similarities in quartz inclusion Ti chemistry and duration of metamorphism to Barrovian rocks from the Moine thrust sheet in northern Scotland (Ashley et al., 2015b), suggesting the possibility that results obtained here may be applicable in other similar environments. In addition to the growth models, fluid inclusions in quartz were evaluated to infer fluid controls on the localized environment at the time of quartz formation prior to inclusion in garnet. Graphite inclusions in quartz in different microstructural settings were analyzed to infer temperatures of deformation evolution through Raman spectroscopy of carbonaceous material (RSCM) thermometry. Rutile needles were used to make qualitative inferences about original protolith quartz trace-element composition (i.e., compositions required to exsolve the rutile needles observed in thin section). Implications from these findings are discussed here to further enhance our understanding of the early prograde evolution in convergent orogens and to assess the importance of mineral shielding for inhibiting chemical reequilibration of earlier-formed mineral grains.
Sample 09SD08A is a semipelitic garnet schist from the Devonian-age Gile Mountain Formation in east-central Vermont. This is a flysch deposit with interleaving pelitic and psammitic layers, with minor calcareous units (Doll, 1944; Fisher and Karabinos, 1980). The rock reached staurolite-kyanite grade and contains the assemblage garnet + quartz + biotite + muscovite + plagioclase + accessories. An S0-parallel slaty mica schistosity (S1) formed during an early ductile event (White and Jahns, 1950). D2 deformation associated with large recumbent folding across the Strafford Dome crenulated and folded S0 and S1, producing an S2 schistosity (White and Jahns, 1950). D2 may have involved nappe emplacement and is associated with a near-isothermal pressure increase of 1–2 kbar (Woodland, 1977; Menard and Spear, 1994). Interkinematic garnets (i.e., grown between deformation events; for details of terminology, see Passchier and Trouw, 2005) contain linear trails of inclusion quartz, with a preferred alignment along the early S1 slaty cleavage, oriented at an oblique angle to, and truncated by, the matrix foliation (φ up to 35°; Figs. 1A and 1B). The matrix contains a spaced foliation defined by compositional layering and mica preferred orientation (Ashley et al., 2013). Strain shadows on the fringes of garnet porphyroblasts consist of quartz, feldspar, and mica. Late-stage, coarse-grained quartz veining is concordant with the S2 foliation.
Quartz and ilmenite inclusions are abundant in garnet. Quartz inclusions contain randomly oriented rutile needles (Fig. 1G), whereas matrix quartz contains none. Cathodoluminescence (CL) imaging and high-resolution trace-element analysis of inclusion quartz by Ashley et al. (2013) revealed homogeneous compositions of 2.5–5 ppm Ti. CL intensity was proportional to [Ti]; therefore, intensity distribution was used to characterize Ti distribution (Spear and Wark, 2009). Linear bands of higher Ti crosscut inclusions (an artifact of diffusion from the rutile needles; Figs. 2A and 2D), with a diffusive gain of Ti around grain boundaries as a result of post-encapsulation heating (where garnet typically contains much more Ti relative to quartz, with hundreds to thousands of parts per million Ti common; Spear and Wark, 2009; Spear et al., 2012). Spear et al. (2012) utilized these diffusion profiles to constrain the duration of near-peak metamorphism at ~1.6 m.y. (time interval from encapsulation to peak temperature and cooling to 460 °C). In comparison, matrix quartz contains more complex zoning patterns, with an increase in Ti toward the rim attributed to growth through Si-liberating metamorphic reactions during heating, and sharp, low-Ti overgrowths resulting from Si-influx during retrogression (at a lower temperature; Ashley et al., 2013).
Thermodynamic calculations were made using a gridded minimization approach in the program Perple_X (Connolly, 2009) in the system MnO-Na2O-CaO-K2O-FeO-MgO-Al2O3-SiO2-H2O-TiO2 (MnNCKFMASHT). The 2004 update (ds5.5) of the Holland and Powell (1998) thermochemical database was implemented to model equilibrium phase stability across a P-T window from 100 °C to 650 °C and 0.5 to 10 kbar (Fig. 3). In addition, calculations incorporated the thermodynamic data for the low-grade metamorphic phases added to the data set by Massonne and Willner (2008); of these, ferro-, magnesio-, and manganese-stilpnomelane are stable solutions to the models presented in this study (unless otherwise noted, these end members are simplified when referred to in the text by the “stilpnomelane” group name). Solution models considered are summarized in Table 1. Quartz, rutile, and titanite were regarded as pure phases, with a pure H2O fluid phase assumed to be in excess. Mineral abbreviations follow Whitney and Evans (2010), with the exception of “wm” for white mica.
A dynamic titania activity (aTiO2) was considered for the modeled P-T window, following the protocol of Ashley and Law (2015), which is summarized here. The chemical potential of TiO2 was calculated from the thermodynamic output of the isochemical forward stability models. These system chemical potentials were evaluated against that of a standard state (i.e., saturated in TiO2); the standard state chemical potential was calculated by saturating the bulk rock chemistry in TiO2, which results in the stabilization of rutile across the entirety of the modeled P-T space. The difference between the system and standard state chemical potentials can be used to determine the affinity toward rutile (i.e., aTiO2). This approach allows for the calculation of aTiO2 in rutile-absent systems. For pelite compositions, the modeling indicates that although large portions of P-T space have titania activities ~1.0, some portions are substantially lower than this, resulting in large errors when projecting Ti isopleths relative to an assumed fixed activity (Ashley and Law, 2015). Ti concentrations in quartz were calculated at different pressures and temperatures using the solubility equation of Thomas et al. (2010). This calibration was preferred over others (e.g., Wark and Watson, 2006; Huang and Audétat, 2012) because: (1) the equation includes a pressure-dependent solubility term, and (2) previous work has correlated Ti concentrations in metapelitic rocks to quartz growth during Si-liberating metamorphic reactions (Ashley et al., 2013), with the Thomas et al. (2010) solution providing the best fit to previous pressure and temperature estimates for the rocks. However, it should be noted that these calibrations cannot be applied to dynamically recrystallized quartz, because dislocation arrays and subgrain boundaries sweeping through recrystallizing quartz crystals may result in local reequilibration with an intergranular medium rather than with the bulk rock (Ashley et al., 2014).
Chemical zoning in quartz formed during prograde metamorphism was investigated through of the construction of three different growth-composition models. The analytical solutions are well suited for evaluating the effects of various processes on the partitioning behavior of Ti into quartz prior to encapsulation in garnet. For all calculations, quartz growth was considered to follow a steady-state geotherm (one-layer model) derived from Fourier’s law (Spear, 1993) with average bulk constants for a standard pelitic crust (thermal conductivity, k = 2.25 W m–1 K–1; heat production from radioactive decay, A = 0.75 µW m–3; heat flow, Q* = 30 mW m–2; crust thickness, D = 35 km; values from Spear, 1993; Figs. 3 and 4). We recognize that assuming a steady-state burial path is an oversimplification of the process; however, it is considered here for three reasons: (1) The relative burial trend (i.e., the P-T trajectory) is more important than the exact path taken; i.e., a higher geothermal gradient should produce a compositional growth zoning similar to that determined in this study (albeit shifted toward higher Ti concentrations). (2) A perturbed system will evolve toward a steady state, provided time scales are sufficiently long. (3) Many of the metamorphic reactions modeled are not pressure sensitive at higher temperatures (>350 °C), where deviation between geotherm projections would be most significant (i.e., higher-T field boundaries are very steep and are nearly isothermal; Fig. 3), making the exact path taken not of great significance when considering burial processes. Likewise, Ti substitution in quartz is nearly pressure independent. In addition, it is important to mention that all models investigated here follow ideal end-member solutions, whereas in nature, processes would be coupled and more complex than described here. The starting model temperature was 134 °C, and the maximum temperature was that at which garnet growth starts (500 °C; Menard and Spear, 1994; Ashley et al., 2013). Illustrative representations of these modeling results are presented as quadrant cross sections of spherical grains, where volume growth is transformed to radial growth to more meaningfully reflect the crystal cross section (analogous to that imaged in a thin section cut through a quartz crystal; Fig. 5).
The first model investigated involved a constant-volume rate of quartz growth, with a fixed titania activity of 1.0 (Fig. 5A). This model is equivalent to solution-transfer processes leading to a constant rate of precipitation of silica on quartz grain boundaries during burial. Model two (Fig. 5B) also assumed a constant-volume rate of quartz growth; however, titania activity was considered to be dynamic and to change according to the stable Ti-buffering phase (Fig. 3). For the third model (Fig. 5C), quartz growth occurred during Si-liberating diagenetic and metamorphic reactions with a dynamic titanium activity considered. To determine the volume percent of quartz precipitation along the profile established in Figure 4, the “werami” subsidiary program to the Perple_X package was implemented. Modeling results were evaluated against inclusion and matrix quartz chemistries observed in sample 09SD08A to infer equilibration conditions.
Raman Spectroscopic Characterization of Fluid Inclusions
Unpolarized Raman spectra were collected using the JY Horiba LabRam HR (800 mm) microprobe at Virginia Tech, with a 600 lines/mm grating. Excitation was by a green 514.57 nm (100 mW) solid-state argon laser focused through a 100× objective, with collection by an electronically cooled open electrode charge coupled device. Spectra ranges of ~370–2125 cm–1 (centered on 1300 cm–1) and ~2500–3900 cm–1 (centered on 3225 cm–1) were measured over three accumulations at 10–30 s. Times were adjusted to maximize peak intensities while avoiding detector saturation. The presence of CO2 in fluid inclusions was evident by the Fermi diad at 1285 cm–1 and 1388 cm–1, with H2O identification through the broad band from ~3100 to 3700 cm–1. Fluid inclusions from multiple locations within sample 09SD08A were analyzed. Isolated fluid inclusions and fluid inclusion planes along healed fractures in quartz inclusions in garnet were analyzed for composition and mineral content. Secondary fluid inclusions were identified as planar features within the quartz inclusions and were interpreted to result from the healing of microcracks (the formation of which must have preceded garnet growth and quartz encapsulation; e.g., Simmons and Richter, 1976; Kranz, 1983; Lespinasse, 1999; van den Kerkhof and Hein, 2001; Goldstein, 2003). These results were compared to fluid inclusions in matrix quartz to better constrain the evolution of fluids during progressive metamorphism and to assess the ability of garnet to shield fluid and mineral inclusions contained in inclusion quartz from chemical reequilibration.
RSCM Material Thermometry
Graphite commonly forms in metasedimentary rocks during diagenesis and metamorphism of organic material. During the conversion from amorphous carbonaceous material to well-crystallized graphite with increasing metamorphic grade, the Raman spectrum changes systematically (Beyssac et al., 2002). Beyssac et al. (2002) derived a linear expression to correlate the degree of graphitization to temperature. A peak area ratio (R2) is used, where R2 = D1/(G + D1 + D2) (where D1, G, and D2 represent the areas under the D1, G, and D2 peaks, respectively) and T (°C) = -455 × R2 + 641. D1–D3 are defect bands at 1350 cm–1, 1620 cm–1, and 1500 cm–1, respectively, and G is a graphite band at 1580 cm–1. D2 and D3 form shoulders on the G band, with D3 neglected because it is difficult to resolve as a result of relatively low intensity and broadness, and overlap with the G band. A peak intensity ratio (R1) equals D1/G. Estimated uncertainty in temperature with the peak area technique is ±50 °C (Beyssac et al., 2002; Rahl et al., 2005). Completely encapsulated graphite inclusions in quartz were analyzed in this study to avoid complications associated with measuring graphite exposed to the thin section surface, where a vectorized high stress induced through mechanical polishing may cause surface flaking and deflection of crystal orientation.
Analytical conditions that were used for fluid inclusion characterization were also used for graphite analyses. A six-degree polynomial baseline correction was implemented, with peak fitting through a Gaussian function. Graphite was analyzed from two microstructural settings in quartz inclusions in garnet: as isolated inclusions within the quartz crystals, and from planar inclusion trails along healed fractures in quartz. Analysis of these two populations was conducted to correlate temperatures with quartz growth/recrystallization and fracture healing, respectively.
Phase Stability, Quartz Growth, and Titania Activities
At the start of the modeled profile, 19.1 modal % stilbite and 6.5 modal % stilpnomelane are stable, in addition to 7.1% chlorite (Fig. 6A). By ~190 °C, all stilbite and stilpnomelane transforms to laumontite and illite (white mica). Further heating produces prehnite (epidote; ~250 °C), with zoisite and minor amounts of actinolite stable above ~305 °C; increases in modal abundances of chlorite, albite, and quartz balance the reactions during progressive heating through the anchizone and into the epizone. A significant amount of quartz growth is expected during these reactions (>15%; Figs. 3 and 4B). Minor (<1 vol%) quartz resorption is observed at ~450 °C associated with An-rich plagioclase above the peristerite gap (Fig. 4B).
Biotite growth initiates at ~370 °C (4.5 kbar), with garnet-in reactions occurring at ~500 °C (6.8 kbar; Fig. 3). Garnet growth results in rapid consumption of chlorite, which by ~530 °C reacts out and results in silica release, which was interpreted by Ashley et al. (2013) to lead to precipitation of matrix quartz during heating.
Rutile is stable at low temperatures (<200 °C); however, it occurs in very low abundance (<0.5 modal %) and is likely the result of diagenetic growth through dissolution of Ti-rich detrital minerals such as ilmenite, titanite, and/or recycled rutile (e.g., Valentine and Commeau, 1990). This results in titania activities of 1.0 at low temperatures (Fig. 4C). At higher temperatures (>200 °C), titanite is present, which results in a reduction of aTiO2, until rutile again becomes stable near 500 °C (Fig. 6B).
Chemical Zoning during Quartz Growth
The first two growth-composition models assumed a constant volume rate of quartz growth through processes such as solution transfer. Both models predict increasing Ti from quartz core to rim, corresponding to increasing solubility of Ti in quartz during heating, with peak concentrations of ~10 ppm Ti immediately preceding inclusion of the quartz into the garnet host (Figs. 5A and 5B). When dynamic titania activity is considered, the absence of ilmenite or rutile stability in portions of P-T space results in a reduction of Ti in quartz, an artifact of titania buffering with activities <1.0 (Fig. 4C). These effects are most notable at radial fractions 0.5–0.7 in Figure 5B, with continued growth along the heating path resulting in lower Ti concentrations due to the reduced TiO2 activity. In addition, [Ti] > 1.0 ppm is restricted to the radial fractions >0.85 (Fig. 5B), compared to <0.8 required for model 1, assuming aTiO2 = 1.0 (Fig. 5A).
The third growth model considers silica production through diagenetic and metamorphic reactions and shows a significantly different zonation pattern. The majority of growth occurs at T < 300 °C, and the resultant quartz contains very low concentrations of Ti (~95 radial % of quartz contains <50 ppb Ti; Figs. 4D and 5C). Late-stage quartz growth through Si-liberating metamorphic reactions precipitates rims with higher [Ti], up to ~10 ppm, similar to the constant-growth models. The same Ti depletion zone resulting from reduced titania activity observed in Figure 5B is seen in Figures 4D and 5C; however, it is offset toward the rim due to variable volume growth rates during burial.
Portions of the modeled quartz within the compositional range measured in sample 09SD08A (2.5–5 ppm Ti) are illustrated by dashed lines in Figure 5. In each scenario, this zone is small, ~12% of the radial fraction for constant growth at aTiO2 = 1.0, decreasing to <2% of the radial fraction when growth through reactions is considered.
Fluid Inclusion Characterization
Twelve fluid and six mineral inclusions from quartz inclusions in garnet and from matrix quartz were analyzed to understand the mineral and fluid distribution in these microstructural domains. Quartz inclusions in garnet contained isolated inclusions of high-density (liquid) CO2, graphite, rutile, and titanite (Fig. 7B). CO2 inclusions typically had an equant or negative crystal shape and only rarely contained two phases (liquid and vapor) at room temperature. This indicates that the density of the CO2 in the inclusion is greater than the density of liquid CO2 in equilibrium with vapor at room temperature, or >0.75 g cm–3. Analysis of the fluids showed sharp, high-intensity bands, which are characteristic of liquid CO2. Likewise, the fluid inclusion planes in quartz inclusions in garnet contained high-density CO2 and graphite inclusions (Fig. 7C). Fluid inclusions were less abundant in matrix quartz (arranged randomly) and, where analyzed, only consisted of liquid CO2 (Fig. 7D). H2O was not detected in any of the inclusions analyzed, as evidenced by the absence of the broad H2O band at ~3100–3700 cm–1 (Fig. 7). We note, however, that Lamadrid et al. (2014) documented that small amounts of H2O may be present but not detected during analysis of CO2-rich inclusions at room temperature, but may be revealed during analysis at higher temperature. All inclusions observed were small (a few micrometers in diameter); however, the largest documented (~8 µm) fluid inclusions were present as isolated inclusions in quartz inclusions in garnet.
Temperature Constraints from Raman Spectroscopy
Analyses of isolated graphite inclusions in quartz resulted in smaller G/D1 peak intensity ratios and more pronounced D2 shoulders on the graphite band than observed for graphite inclusions along healed fractures in quartz. This trend is generally indicative of lower temperatures and less well-crystallized graphite (Kouketsu et al., 2014, their figure 2). Peak positions, intensities, widths, and areas obtained from Raman analysis are listed in Table 2 (Fig. 8). The resultant peak intensity ratios (R1) for the graphite are 0.69 and 0.36, and the peak area ratios (R2) are 0.52 and 0.35 for the isolated inclusions and inclusions along healed fractures, respectively. This equates to temperatures of 408 ± 50 °C for isolated inclusions and 488 ± 50 °C for inclusions along fluid inclusion planes.
Evaluating the Validity of Models for Predicting Quartz Production through Diagenesis
One of the growth models considered here examines quartz precipitated as a result of Si-liberating diagenetic and metamorphic reactions (Fig. 5C). Considering the volume of quartz produced through these processes, diagenesis is more important than metamorphism because significant silica production occurs during the early prograde history. Previously, van de Kamp (2008) estimated the amount of silica released through the transformation of smectite to illite to muscovite, assuming conservation of alumina. This work was based on the fact that smectite clays alter to less siliceous illite (and at higher temperatures, muscovite) during burial and heating (e.g., Hower et al., 1976; Lynch et al., 1997). Nearly 95% of the transformation from smectite to illite occurs at <200 °C, with an equivalent amount of illite being converted to muscovite by 300 °C (Merriman and Frey, 1999). The findings of van de Kamp (2008) suggest that the alteration of sedimentary muds to shales releases ~14–20 wt% SiO2 during heating from surface temperatures to 200 °C, with an additional 18–28 wt% silica evolved by 500 °C. Our simulations suggest that >12 vol% (~16.5 wt%) of quartz is produced during heating from 134 °C to 300 °C (Fig. 4B), with little additional silica produced through metamorphic reactions at higher temperatures (until the ~3 vol% chlorite-out, silica-producing reaction is reached at ~525 °C; discussed by Ashley et al., 2013). Thus, these results are similar to estimates of the amount of silica produced during the smectite to muscovite transformation in the same temperature window (~15 wt%; van de Kamp, 2008, their figure 1). Likewise, zeolite-group minerals are expected to liberate significant amounts of silica during diagenesis. Heulandite group zeolites are high in silica, containing Si/Al ratios of 4.74–5.19 (Tsolis-Katagas and Katagas, 1990), which are greater than those observed for smectite (Si/Al ratios of ~3; van de Kamp, 2008). The higher silica release predicted in our thermodynamic models relative to a pure smectite evolution by van de Kamp (2008) is likely due to the stability of stilbite (zeolite) at low temperatures, which contains 27 cations per formula unit (cpfu) of Si, decreasing to 4 cpfu Si in the transformation to laumonite (zeolite) at ~190 °C. The actual amount of silica produced is thus sensitive to the starting mineral assemblage, which can be reasonably constrained through pseudosection modeling of the representative bulk rock composition. The large amount of silica production predicted by the thermodynamic models is consistent with the findings by van de Kamp (2008) and suggests that the majority of quartz growth should occur at low temperatures (<300 °C). This is consistent with observations of natural samples. These calculations assume silica remains in the system (closed system) and is not liberated from the rock, which forms the basis of the isochemical thermodynamic models. Leached silica from shales (i.e., removed from the rock) would inhibit continued quartz growth, in which case other growth processes would have to be active.
Early Prograde Growth Processes
None of the three numerical growth models considered in this paper (Fig. 5) produced Ti-zoning characteristics similar to those observed in quartz inclusions in garnet grains from sample 09SD08A (Fig. 9). Whereas the analyzed profiles have homogeneous Ti concentrations (±a few ppm Ti), growth through solution transfer and diagenetic/metamorphic reactions would result in strongly zoned crystals that span over three orders of magnitude in Ti concentration (Fig. 9). Ti variations of this magnitude would be easily resolved, owing to the sensitivity of CL imaging, and therefore solution-transfer and reaction-producing events cannot be the primary mechanism responsible for the observed Ti distribution in quartz. Rather, the homogenized, low-Ti characteristics of the inclusion quartz are identical to quartz that has been dynamically recrystallized through subgrain rotation (Grujic et al., 2011; Ashley et al., 2013; Kidder et al., 2013). These recent studies have shown that regardless of the original (undeformed) quartz porphyroclast composition, dynamically recrystallized grains have [Ti] <10 ppm. This appears to be a commonly observed consequence of dynamic recrystallization through subgrain rotation, and it speaks to the efficiency of this mechanism in liberating Ti from the quartz lattice (Ashley et al., 2014). In addition, remnant porphyroclasts have been observed with partially recrystallized subgrains that contained higher [Ti] than the recrystallized grains (Grujic et al., 2011; Ashley et al., 2013). We interpret the patches of higher Ti concentration in the inclusion quartz (Figs. 2A and 2D) as being an artifact of this heterogeneous and incomplete recrystallization process.
The presence of 120° triple junctions between inclusion quartz grains (Figs. 2A and 2D) also supports the interpretation that reequilibration of Ti in quartz inclusions represents a pre-encapsulation recrystallization event. The temperatures and/or duration of heating that occurred before encapsulation of quartz in garnet must have been sufficient to promote grain boundary area reduction, as indicated by observed 120° triple junctions between inclusion boundaries. Fracturing and healing of fractures must have occurred post-recrystallization but prior to encapsulation because: (1) healed fractures crosscut quartz grain boundaries and regions of quartz with reset Ti chemistries, and (2) encapsulation would inhibit fracture healing due to the isolation from fluids present in the matrix. Brittle fracturing that postdates recrystallization at higher temperatures could be the result of a localized rapid increase in strain rate, shifting the rock from the plastic to the brittle regime (Knipe, 1989, 1990). However, such interpretations applied to our sample are speculative and cannot be confirmed due to the limitations on the data that may be retrieved from inclusion quartz (e.g., grain-size analysis for paleostress and paleostrain rate estimations). Diamond and Tarantola (2015) described fluid inclusions in quartz deformed by weak ductile shearing, where the development of a plane of neonate fluid inclusions around relict inclusions provided evidence for differential stress magnitudes and predeformation fluid properties. However, these fluid inclusion morphologies/structures were not observed in our sample, and therefore our sample did not experience similar conditions during deformation.
The difference in trace-element distribution within inclusion quartz relative to matrix quartz (Fig. 2) is an important observation because it gives evidence that the garnet host acts as an effective shield in preventing further chemical modification of inclusion quartz (apart from Ti diffusion from the garnet host), isolating the included quartz from growth processes that were active postentrapment, as seen with matrix quartz. This shielding also prevents further recrystallization (dynamic or static) from occurring, as is evident from isolated graphite inclusions and trace-element chemistries of inclusion quartz that suggest temperatures of recrystallization of ~400 °C, consistent with subgrain rotation recrystallization (Stipp et al., 2002) and Ti-in-quartz temperature estimates for the quartz inclusion compositions (~3.8 ppm Ti). Further dynamic recrystallization could not have taken place, because the fluid inclusion plane would extend to inclusion quartz grain boundaries and would be disrupted or removed during the recrystallization process. Temperatures for fracture healing (determined by RSCM on graphite inclusions in the fluid inclusion plane) occurred at ~485 °C, before the encapsulation in garnet took place, with no fluid inclusion plane containing graphite recording higher temperatures. Recognition of included graphite as isolated grains and in fluid inclusion planes in quartz inclusions is also important because it suggests that graphite inclusion prevents further crystallographic restructuring. The heating of graphite results in a continuous disordered-ordered structural transformation that provides the foundation of the RSCM thermometer; despite continued heating to ~550 °C, the graphite inclusions preserve the degree of graphitization associated with encapsulation in quartz, rather than homogenizing to the peak metamorphic temperatures as would be expected for unbounded matrix graphite. The limitations of mineral inclusions to volumetrically “relax” (e.g., Gillet et al., 1984; Ashley et al., 2015a) may inhibit further graphitization during heating and allow the graphite to be used as an early prograde entrapment thermometer.
Fluid Evolution during Progressive Metamorphism, with Implications for Thermodynamic Modeling
The widespread presence of high-density, liquid CO2 fluid inclusions with no observed H2O is significant for several reasons. First, these fluid inclusions occur in quartz that grew throughout the prograde history (both inclusion and matrix quartz). Carbonaceous material is not present in the matrix, suggesting that the fluid component of the rock contained, to some extent, CO2 throughout the duration of the prograde history, resulting from the maturation of biogenic carbon (Huff and Nabelek, 2007). The presence of CO2-rich fluid inclusions in quartz inclusions that formed during the prograde history has important implications concerning the origin of CO2-rich fluid inclusions in medium-to high-grade metamorphic rocks. Often the isochores for the fluid inclusions do not project through the peak metamorphic conditions. This, in turn, has led some workers to suggest that the fluid inclusions have reequilibrated volumetrically along the retrograde path (Sterner and Bodnar, 1989; Vityk et al., 2000), or that the fluid inclusions originally contained an H2O-CO2 mixed fluid and later experienced postentrapment loss of H2O (Hollister, 1990; Bakker and Jansen, 1991), whereas still other workers have suggested that the CO2-rich fluid inclusions were trapped late in the metamorphic history during retrogression (Lamb et al., 1987; Lamb, 1990). The presence of CO2-rich fluid inclusions in quartz inclusions that formed and were encapsulated in garnet along the prograde path suggests that, at least in some cases, CO2-rich fluids were present during prograde and/or peak metamorphism.
Thermodynamic models are commonly applied in metamorphic petrology to constrain stability fields of mineral assemblages, which can be used to infer P-T conditions experienced by the rock. These models require knowledge of the fluid composition throughout the metamorphic history. When carbonaceous material is not observed in the matrix, a pure H2O fluid phase is typically assumed. The presence of CO2 could have an impact on the predicted mineral stability fields, typically resulting in extension of these fields to higher temperatures for decarbonation reactions (e.g., Will et al., 1990; Evans et al., 2010) and driving dehydration reactions (and thereby extending the stability of many assemblages) to lower temperatures with the presence of C in a COH fluid phase (e.g., Connolly and Cesare, 1993). This translation of phase boundaries is most notable with the minor addition of CO2 to the fluid at lower temperatures (e.g., 400 °C), because aCO2 increases asymmetrically and rapidly at XCO2 < 0.2 in the CO2-H2O system (Sterner and Pitzer, 1994; Holland and Powell, 2003). In this scenario, temperatures reported for metamorphic reactions may be overestimated. The scarcity of CO2 fluid inclusions in matrix quartz relative to inclusion quartz may suggest a significant degassing of fluids from the rock during progressive burial, as suggested by Yardley and Bodnar (2014). This is expected, as metamorphic decarbonation would continue to higher temperatures as more carbon is lost from the system. The removal of CO2 may minimize the impact CO2 has on modeled phase stability, leaving peak P-T estimation through this method the most reliable part of the entire metamorphic history.
New Perspectives on Metamorphic Evolution
The results of this study allow us to infer a more complete metamorphic history for sample 09SD08A than was previously possible. Here, we combine these results with previous findings to provide a chronology of events and processes that occurred throughout the P-T evolution of the sample (for a graphical summary of this evolution, see Fig. 10).
The protolith contained high-Ti quartz (evident by the presence of exsolved rutile needles) that was eroded and transported to form a flysch deposit in the foreland basin of the developing orogen (Doll, 1944; Fisher and Karabinos, 1980). During early stages of burial, clay minerals were transformed into micas through the anchizone and epizone, releasing silica in the process. At ~410 °C, Ti in the original quartz reequilibrated to very low concentrations (2.5–5 ppm) during subgrain rotation recrystallization of quartz. During recrystallization, isolated inclusions of CO2, graphite, and titanite were encapsulated in the quartz. Grain boundaries were progressively annealed into low-energy configurations during heating, promoting triple-junction “foam” textures between neighboring quartz grains. Brittle fracturing followed, potentially associated with a localized and instantaneous increase in strain rate. Healing of fractures in the quartz occurred at ~490 °C, trapping inclusions of graphite and CO2. During these early stages of heating, water was largely sequestered into hydrous silicate phases.
Fracture healing was immediately followed by the overstepping of the garnet isograd, with garnet growth encapsulating quartz grains that had a preferred alignment along the S1 slaty cleavage. Once included, Ti began to diffuse from garnet into the grain boundaries of inclusion quartz. Garnet shielded and disconnected inclusion quartz from chemical communication with the matrix and prevented further dynamic recrystallization. Nappe emplacement followed garnet growth, resulting in an isothermal pressure increase of ~2 kbar (Menard and Spear, 1994). Garnets were rotated during development of the S2 schistosity.
Immediately following nappe emplacement, the chlorite-out reaction was overstepped, and matrix quartz was overgrown with released silica, incorporating higher concentrations of Ti during continued heating (Ashley et al., 2013). Matrix quartz contains few CO2 inclusions, probably because of removal of carbon from the rock through metamorphic decarbonation processes. Peak temperatures of ~550 °C were reached (Menard and Spear, 1994) before rapid cooling to 460 °C occurred. Rapid cooling is required to impede further diffusion of Ti from the surrounding garnet host into the inclusion quartz. These Ti diffusion profiles are short (<6 µm) and call for near-peak metamorphism to occur in a short amount of geologic time (~1.6 m.y.), suggesting pulsed-style metamorphism (Spear et al., 2012). At cooler temperatures (<450 °C), late-stage Si-charged fluid influx occurred, resulting in rim overgrowths in matrix quartz having lower-Ti concentrations (Ashley et al., 2013).
This study presents growth-composition models for evaluating mechanisms of Ti equilibration in quartz during early prograde metamorphic evolution. These models show that subsequent dynamic recrystallization resets any prior chemistry, and they suggest that rocks where garnet growth occurs at lower temperatures would be more informative for probing earlier into the prograde history. This would be particularly true for high-Mn rocks, in which earlier garnet growth may be promoted. The evaluation of fluid and mineral inclusions provides insight into temperatures of deformation processes and the evolution and composition of the fluid component of the rock. Investigations in metamorphic petrology are continuing to migrate away from the classical limitations of inferring generic paths to peak pressures and temperatures. Here, we show the potential to deconvolve complex histories with respect to pressure, temperature, metamorphic timing and duration, deformation, fluid flux, and reactions if the assumption of a steady-state geotherm is valid. With the investigation of quartz inclusions in garnet, we are able to infer environmental controls from the early stages of orogenesis, the signatures of which are typically overprinted and lost during progressive heating and deformation.
We thank Brian Romans for thoughtful discussions and direction during the early stages of this work. Reviews by Djordje Grujic and by an anonymous second reviewer, and editorial suggestions from Harold Stowell greatly improved the quality of this manuscript. We are grateful for assistance by Charles Farley with use of the Raman spectrometer at Virginia Tech for fluid inclusion characterization. This work is supported by the National Science Foundation under grant EAR-1220345 (awarded to R.D. Law).
Figures & Tables
Linkages and Feedbacks in Orogenic Systems
CONTAINS OPEN ACCESS
- crystal growth
- crystal zoning
- fluid inclusions
- framework silicates
- garnet group
- Gile Mountain Formation
- metamorphic rocks
- mineral inclusions
- native elements
- P-T conditions
- prograde metamorphism
- Raman spectra
- silica minerals
- United States
- east-central Vermont
- Strafford Dome