Role of fluids on deformation in mid-crustal shear zones, Raft River Mountains, Utah

Abstract Fluids are commonly invoked as the primary cause for weakening of detachment shear zones. However, fluid-related mechanisms such as pressure-solution, reaction-enhanced ductility, reaction softening and precipitation of phyllosilicates are not fully understood. Fluid-facilitated reaction and mass transport cause rheological weakening and strain localization, eventually leading to departure from failure laws derived in laboratory experiments. This study focuses on the Miocene Raft River detachment shear zone in northwestern Utah. The shear zone is localized in the Proterozoic Elba Quartzite, which unconformably overlies the Archaean basement, and consists of an alternating sequence of quartzite and muscovite-quartzite schist. In this study, we characterize fluid-related microstructures to constrain conditions that promoted brittle failure in a plastically deforming shear zone. Thin-section analyses reveal the presence of healed microcracks, transgranular fluid inclusion planes and grain boundary fluid inclusion clusters. Healed microcracks occur in three sets, one sub-perpendicular to the mylonitic foliation, and a set of two conjugate microcracks oriented at ∼40–60° to the mylonitic foliation. Healed microfractures are filled with quartz, which has a distinct fabric, suggesting that microcracks healed while the shear zone was still at conditions favourable for quartz crystal plasticity. Transgranular fluid inclusion planes also occur in three sets, similar in orientation to the healed microfractures. Fluid inclusions commonly decorate grain and subgrain boundaries as inter- and intragranular clusters. Our results document ductile overprint of brittle microstructures, suggesting that, during exhumation, the Raft River detachment shear zone crossed the brittle–ductile transition repeatedly, providing pathways for fluids to permeate through this shear zone.


Introduction
Fluids play a critical role in controlling the chemical and physical behaviour of the crust and mantle (e.g. Caine et al. 1996;Hirschmann, 2006). Fluid-rock interaction controls rock deformation through a variety of processes including hydrolytic weakening (e.g. Griggs, 1967), pressure-solution creep (e.g. Shimizu, 1995) or metamorphic reactions (e.g. Ferry, 1994). Faults and shear zones are important structures that control fluid circulation from the upper to the lower crust (e.g. Reynolds & Lister, 1987;McCaig, 1988). Fluid circulation in the brittle upper crust is accommodated through faults and fractures that commonly increase porosity and permeability of the host rock, allowing fluids to penetrate to mid-crustal levels (e.g. McCaig, 1988;Sibson, 1992;Wehrens et al. 2016;Fig. 1). Faults also exhibit fault-valve behaviour, becoming highly permeable after seismic failure, channelling fluid discharge in the seismogenic zone (Sibson, 1990(Sibson, , 1992. This mechanism involves pre-seismic stress build-up causing increased dilatancy and drawing-in of fluid into the stressed rock, followed after seismic rupture by a pressure drop and fluid venting out of networks of microcracks (Reynolds & Lister, 1987;McCaig, 1988;Sibson, 1990). This fault-valve behaviour or 'seismic pumping' suggests that fluids may be one of the primary drivers of the seismic cycle (Cox, 2010;Miller, 2013;Zhu et al. 2020;Prando et al. 2020). In the ductile regime, fluid circulation is conceptually challenging (e.g. Connolly & Podladchikov, 2004). Recent studies have demonstrated that viscous flow can produce porosity in deforming rocks through dynamic recrystallization (Gilgannon et al. 2020) and creep cavitation processes (Fusseis et al. 2009;Menegon et al. 2015;Précigout et al. 2019;Dobe et al. 2021).
The brittle-ductile transition is therefore a critical zone to study fluid-rock interaction that occurs in faults and shear zones (e.g. Marchesini et al. 2019). Deformation at the brittle-ductile transition is cyclical, oscillating between brittle and ductile behaviour (e.g. Famin et al. 2004Famin et al. , 2005Siebenaller et al. 2013Siebenaller et al. , 2016Carter et al. 2015;Compton et al. 2017). A switch in failure mode is typically attributed to a perturbation in one of the parameters controlling deformation, such as deviatoric stress, strain rate or pore fluid pressure. Detachment shear zones associated with metamorphic core complexes evolved at the brittle-ductile transition and are rapidly exhumed (e.g. review by Whitney et al. 2013 and references therein). These shear zones are therefore a prime candidate to study in detail fluid-rock interaction and brittle-ductile deformation.
Metamorphic core complexes form during orogenic collapse of an overthickened crust (e.g. review by Whitney et al. 2013 and references therein). Extension in the cool and brittle upper crust is accommodated by faulting and fracturing, providing pathways (high porosity/permeability fractures and faults) and head gradient (steep topography) for fluid infiltration (Fig. 1). The hot, ductile lower crust flows laterally. Because of this rheologic contrast between the upper and lower crust, combined with hydrolytic weakening, pressure-solution creep and metamorphic reactions in a fluid-saturated system, strain commonly localizes at the brittle-ductile transition, leading to the development of detachment shear zones. Progressive exhumation of the detachment shear zone leads to the formation of a metamorphic core complex.
In this contribution, we investigate the record of fluid-rock interaction at the brittle-ductile transition in the detachment shear zone associated with the Miocene Raft River metamorphic core complex (e.g. Wells et al. 2000). Previous work demonstrates abundant evidence of chemical, physical and isotopic water-rock interactions in this detachment shear zone (Gottardi et al. 2011Methner et al. 2015). The goal of this study is to characterize fluidrelated microstructures to constrain conditions that promoted brittle failure in a plastically deforming system.

Regional geology
2.a. The Raft River metamorphic core complex The Raft River Mountains are located in northwestern Utah's portion of the Great Basin province and are a component of the larger Raft River-Albion-Grouse Creek metamorphic core complex (Fig. 2;Compton, 1975). The Raft River Mountains are composed of Archaean through Permian aged amphibolite-to greenschist-facies rocks. Following several episodes of contraction during the Mesozoic to early Cenozoic Sevier orogeny (Wells, 1997;Hoisch et al. 2002;Harris et al. 2007), the Raft River experienced two pulses of extension during Cenozoic time: an initial Eocene to early Oligocene (~42-37 Ma) extension along the W-rooted Middle Mountain shear zone (Saltzer & Hodges, 1988;Wells et al. 2000), and Oligo-Miocene extension (~25 to 15 Ma) along the top-to-the-E Raft River detachment shear zone (Wells et al. 2000;Wells, 2001;Gottardi et al. 2011;Gottardi & Teyssier, 2013).

Methodology
The samples investigated in this study were collected on a vertical transect across the Raft River detachment shear zone at Clear Creek Canyon. Six representative samples of the quartzite mylonite were analysed (Fig. 2). The detachment shear zone also hosts numerous quartz veins that are either transposed parallel to the foliation (samples RR09-Ky06 and RR10-02) or at a high angle to the foliation (sample RR10-09). Petrographic analysis was conducted on standard thin-sections (30 μm thick) prepared from eight quartzite mylonite samples collected across the Raft River detachment shear zone (Fig. 2c). Thin-sections were prepared from billets cut perpendicular to the mylonitic foliation and parallel to the lineation. Whole thin-section scans were acquired with a digital film scanner to gather plain-and crossed-polarized images in order to investigate the orientation of healed microfractures. Microstructural analysis was then conducted using a Nikon Eclipse LV100 microscope using 5×, 10×, 20× and 50× objectives and a 100× oil immersion objective to observe fluid inclusions. Using a Leitz Wetzlar 3-axis universal stage mounted to a Zeiss optical microscope, the orientation of intracrystalline fluid inclusion planes and their host quartz crystal c-axis were determined using 150 μm thick thin-sections of the same eight samples. The software ImageJ (Schneider et al. 2012) was used for measuring fluid inclusion plane orientation and fluid inclusion size. The software Orient (Vollmer, 2015) was used to plot stereographic projections of orientation data, such as fluid inclusion planes.

Results
We first present a microstructural analysis of the quartzite mylonite, then focus on the secondary fluid inclusions found in six quartzite mylonite samples collected across the detachment shear zone, and two quartz vein samples collected near the top of the detachment shear zone (Fig. 2c). Our results show that fluid inclusions are found in three types of microstructures: (1) healed microcracks, (2) transgranular fluid inclusion planes, and (3) grain boundary fluid inclusion clusters.

4.a. Microstructural analysis
The quartzite is characterized by two quartz grain populations: coarse elongate (>500 μm long) grains or 'relict' grains, and finer recrystallized grains (20-100 μm) (Fig. 3a-c). The relict grains have moderate to high aspect ratios (1:2 to 1:5) and define the  macroscopic fabric. They typically exhibit strong undulose extinction, deformation bands and deformation lamellae (Fig. 3d). Recrystallized grains are found at the boundaries of relict grains (Fig. 3b, c). They are equant or blocky to slightly elongate and have a direct relation to subgrains present in large grains.
Recrystallized grains locally form a low-angle oblique secondary foliation at 30-35°to the macroscopic foliation, indicating a top-to-the-E sense of shear, consistent with other kinematic criteria (Compton, 1980, Sullivan, 2008Gottardi & Teyssier, 2013). Although not common, shear bands tend to form at a shallow angle (~10-20°) to the mylonitic foliation. Shear bands are typically associated with grain-size reduction. Relict grains in the vicinity of shear bands show an asymmetric shape and muscovite tails indicative of a top-to-the-E sense of shear.
Muscovite grains define the mylonitic foliation and typically pin relict quartz grains (Fig. 3c). Muscovite tends to form thin folia that rarely exceed a thickness of 50 μm. Muscovite grains also form two populations: large folia defining the foliation (Fig. 3c) and occasionally forming shear bands (Fig. 3d), and small, rhombshaped grains commonly found pinning quartz grains (Fig. 3b).

4.b. Fluid inclusions related to deformation fabrics
The distribution of fluid inclusions was observed in thin-sections oriented perpendicular to the foliation and parallel to the lineation (XZ). The nomenclature for describing fluid inclusions is defined by the physical and chemical properties of each inclusion (Kranz, 1983;Roedder, 1984;Lespinasse, 1999;Lespinasse et al. 2005).

R Gottardi and B Hughes
Fluid inclusions that define planar trails (fluid inclusion planes) are known as secondary fluid inclusions (Fig. 6). We observe secondary fluid inclusions that cross multiple grains, referred to as transgranular, and single grains, referred to as intragranular.
Grain boundary fluid inclusions are also common in the studied samples. Although not observed in this study, the nomenclature for individual isolated fluid inclusions that formed coevally with either vein formation or host mineral crystallization are called primary fluid inclusions.

4.b.1. Healed microfractures
Healed microfractures are common structures found in all studied samples (Fig. 4). These through-going fractures cut across the entire thin-section (Fig. 4a). The fractures cut across the mylonitic foliation, and often offset muscovite grains (Fig. 4b). The width of the healed microfractures ranges from~20 to 100 μm (Fig. 4b). The microfractures are filled with quartz, which has a fabric, not as strong as the fabric in adjacent quartz grains in the mylonite. Two adjacent microfractures sometimes exhibit stepover patterns. Fluid inclusions are abundant in the vicinity of the healed microfractures, typically occurring in planes parallel to the microfracture (Fig. 4c). The tip of the microfractures is commonly accompanied by horsetail structures that fan out both upward and downward (Fig. 4d). Fluid inclusions are particularly abundant in the horsetail fractures and follow the offshoot fractures (Fig. 4d). The dip of the healed microfracture was measured with respect to the mylonitic foliation in the XZ plane of each thin-section (Fig. 5a). Three sets have been identified: a conjugate set that is oriented at~30°± 10°and~130°± 10°clockwise from the mylonitic foliation, and a set that is sub-perpendicular to the mylonitic foliation (Fig. 5).

4.b.2. Transgranular fluid inclusion planes
Fluid inclusion planes cross multiple grains (transgranular) and single grains and subgrain boundaries (intergranular). Transgranular fluid inclusion planes tend to group in multiple parallel sets rather than be isolated (Fig. 6a, b). The orientation of these transgranular fluid inclusion planes is somewhat similar to the healed microcracks: in boudinaged vein samples collected near the top of the detachment shear zone, transgranular fluid inclusion planes are oriented sub-perpendicular to the mylonitic foliation (Fig. 6a). In the quartzite mylonite samples, the transgranular fluid inclusion planes occur in two conjugate orientations, dipping at~40°± 10°and 135°± 10°measured clockwise from the mylonitic foliation (Figs 5b, 6b), an orientation similar to the conjugate sets of healed microfractures. These transgranular fluid inclusion planes typically cut across multiple elongate relict quartz grains that define the mylonitic foliation and can be hundreds of microns long. Conjugate sets of fluid inclusions commonly cross-cut each other (Fig. 6b). Transgranular planes typically preserve a high number (100-1000s) of fluid inclusions. Fluid inclusion planes are predominantly found to be populated by small circular 1-3 μm single-phase inclusions within mylonite (Figs 7, 8). Many of these inclusions have dark rims that could indicate possible trapped high-density fluids (Fig. 8a). Isolated inclusions are multi-phased at room temperature and are typically characterized by a vapour and liquid component (Fig. 8b). Larger inclusions commonly exhibit three phases (Fig. 8c). These triphase inclusions are composed of the typical vapour and liquid components but exhibit a euhedral solid crystal overlying the top section of the inclusion (Fig. 8c).

4.b.3. Grain boundary fluid inclusions
Grain boundary fluid inclusions and clusters of intragranular fluid inclusions are common in all studied quartzite mylonite samples, rare in the boudinaged vein samples. They are common around recrystallized quartz grains and typically outline subgrain boundaries in large relict grains (Fig. 6c). They form dense clusters at grain boundaries. When inclusions are trapped within the interior of a relic, they form clusters that are sub-planar, oriented subperpendicular to transgranular fluid inclusion planes. Lastly, fluid inclusions are particularly abundant in quartz deformation lamellae (Fig. 6d). Deformation lamellae are found in 20-50 % of quartz grains in the quartzite mylonite. They are very well developed in large relict quartz grains that are elongated at a high angle to foliation but occur also in smaller recrystallized grains. Deformation lamellae are commonly decorated by fluid inclusions. Deformation lamellae and fluid inclusion planes usually have a similar orientation.

4.c. Quartz c-axis orientation
The orientations of intracrystalline fluid inclusion planes and their host quartz crystal c-axes were measured using a universal stage mounted on an optical microscope. Results gathered with the universal stage were plotted on equal-area, lower-hemisphere stereographic projections ( Fig. 9; Table 1). For both the quartzite mylonite and quartz vein samples, quartz c-axes cluster in the centre of the stereonet, forming a point distribution. The trend

R Gottardi and B Hughes
and plunge of quartz c-axes ranges from 351°to 14°and 63°to 74°, with an average of 4°± 10°and 69°± 2°. These results indicate that the measured quartz grains are oriented such that their c-axes are contained within the foliation plane, and oriented perpendicular to the lineation. This orientation is consistent with quartz dynamic recrystallization by prism <a> slip. The orientation of the fluid inclusion planes within the measured quartz grains is also very consistent across all samples. The strike and dip of the fluid inclusion planes range from 48°to 98°and 10°to 17°, with an average of 82°± 12°and 26°± 2°. These results indicate that the fluid inclusion planes are oriented subparallel to the basal crystallographic planes of quartz (Fig. 9), perpendicular to the foliation and parallel to the lineation.

5.a. Deformation mechanisms
The crystal-plastic and brittle microstructural record of the quartzite mylonite from the Raft River detachment shear zone provides insight into fluid circulation during exhumation of the shear zone. Microstructures indicate that quartz deformed primarily by dislocation creep, subgrain rotation and minor grain boundary migration recrystallization ( Fig. 3; Regime II of Hirth & Tullis, 1992), suggesting deformation temperatures of 450-500°C (Stipp et al. 2002). This temperature range is probably related to the latest stages of crystal-plastic deformation of quartz, because the majority of relict quartz grains exhibit microstructures such as relict quartz grains with undulose extinction and deformation lamellae, that are all evidence of high-stress/high-strain rate deformation conditions (Fig. 3d), indicating that deformation was occurring close to the brittle-ductile transition. In particular, deformation lamellae have been interpreted to reflect plastic high-stress deformation in alloys and metals (Drury, 1993), and short-term plastic and eventually brittle deformation related to coseismic loading in rocks (Trepmann & Stöckhert, 2003;Trepmann & Seybold, 2019). Deformation lamellae were preserved owing to a lack of recovery during deformation, suggesting that recovery in the quartzite was not efficient enough to keep up with strain rate and offset strain hardening (e.g. Gottardi & Teyssier, 2013). These observations suggest that during the latest stages of exhumation, the detachment system evolved close to the brittle-ductile transition, switching back and forth between dislocation creep and glide-controlled exponential creep, depending on the temperature/strain-rate/fluid conditions (Gottardi & Teyssier, 2013).
The mylonitic fabric is overprinted by several brittle structures: healed microfractures and transgranular fluid inclusions planes (Figs 4, 6). Both of these structures occur in similar orientations: one is sub-perpendicular to the mylonitic foliation, and a conjugate set occurs oriented at~35°± 10°and~135°± 10°clockwise from the mylonitic foliation (Fig. 10). These two different orientations may be caused by a minor switch in the stress field during exhumation of the detachment shear zone. Healed microfractures and transgranular fluid inclusion planes are consistent in orientation across all studied samples (Fig. 5). The orientation of these structures is consistent with the E-directed shear kinematics preserved in plastically deformed footwall mylonite, such as type-II S-C fabrics, asymmetric tails around feldspar porphyroclasts, shear bands and asymmetric quartz c-axis cross-girdles (Wells, 1997;Sullivan, 2008;Gottardi & Teyssier, 2013).

5.b. Condition of fluid entrapment
The oxygen and hydrogen stable isotope geochemistry of quartz, muscovite and fluid inclusions from the quartzite mylonite suggests that meteoric fluid infiltration occurred during the Miocene extension and exhumation of the Raft River detachment shear zone (Gottardi et al. 2011Methner et al. 2015). Oxygen isotope geothermometry of quartz and syn-kinematic muscovite indicates that Miocene deformation by dynamic recrystallization of quartz and muscovite occurred at a temperature of 350-485°C (Gottardi et al. 2011. The hydrogen isotope signatures of fluid inclusions (−94 ‰ to −82 ‰) and fabricforming recrystallized muscovite (−120 ‰ to −90 ‰) demonstrate that meteoric fluids infiltrated the shear zone during the Miocene deformation event (Gottardi et al. 2011. The microstructural observations indicate that trapping conditions of the fluid likely occurred during coeval brittle and plastic deformation of quartz, but before complete cooling of the footwall rocks. The fact the quartz grains filling the microfractures have a preferred orientation and show evidence of subgrain rotation recrystallization (Fig. 4) suggests that the temperature was still above~400°C when the microfractures healed. Altogether, these microstructural observations suggest that brittle-ductile processes that allowed fluid flow and the trapping of fluid inclusions must have occurred between the latest stage of crystal-plastic deformation and before the complete exhumation and cooling of the footwall rocks.

5.c. Model for fluid circulation
The microstructural record preserved in the quartzite mylonite of the Raft River detachment shear zone correlates to rapid exhumation of hot mid-to lower crustal rocks that can drive fluid circulation and heat advection into the upper crust (Morrison & Anderson, 1998;Gottardi et al. 2011;Siebenaller et al. 2013Siebenaller et al. , 2016 (Figs 4, 6). These results suggest that brittle fractures, such as normal faults and healed microcracks, are important structures for fluid transfer in metamorphic core complexes and extending crustal rocks in general (e.g. López et al. 1994;Sibson, 2000). Based on our microstructural observation, we propose the following mechanism for meteoric fluid infiltration in the deforming shear zone. Abundant evidence of high-stress plastic deformation in the detachment shear zone suggests that the response of the quartzite mylonite to stress hardening is seismic failure (Fig. 11). When the detachment shear zone is still at depth, a~400-500°C temperature allows quartz deformation by dislocation creep processes such as subgrain rotation (Figs 3,  11a). However, owing to high differential stress, dislocation creep cannot keep up with either a fast strain rate or strain hardening, leading to high-stress plasticity features such as undulose extinction and deformation lamellae (Fig. 6d). Eventually, stress buildup leads to embrittlement and brittle failure. Brittle failure leads to the formation of faults, fractures and microcracks, which allow for surface fluids to be pumped into the shear zone ( Fig. 11b; Sibson, 1977;McCaig, 1988;Lister & Davis, 1989;Carter et al. 2015;Siebenaller et al. 2016;Compton et al. 2017;Prando et al. 2020;Menegon & Fagereng, 2021). After rupture, during post-seismic relaxation, stress decays and the detachment shear zone re-enters the dislocation creep regime. Microfractures heal, as indicated by the fabric preserved in the quartz fill, thereby trapping fluid inclusions (Figs 4, 6, 11c). The intragranular fluid inclusion planes and subgrain boundary fluid inclusions likely represent the remnants of some former brittle fluid inclusion plane reworked by the cyclical ductile shearing. Additional stress build-up likely expelled large fluid inclusions from the fluid inclusion planes and broke them down into smaller inclusions (>1 μm). Crystal-plastic deformation by subgrain rotation and grain boundary migration eventually redistributes these small inclusions as grain and subgrain boundary fluid inclusion clusters (Fig. 11). These tight clusters of small fluid inclusions (>1 μm) typically indicate that explosion of larger inclusions was likely caused by isobaric heating or decompression forces during tectonic processes (Vityk & Bodnar, 1995). This cycle likely repeated several times during the exhumation of the detachment shear zone.

Conclusions
Microstructural analysis of samples of quartzite mylonite from the Raft River detachment shear zone provides insight into meteoric fluid infiltration near the brittle-ductile transition during progressive exhumation of the shear zone. Our microstructural observation reveals the presence of several fluid-related microstructures: healed microcracks, transgranular fluid inclusion planes and grain boundary fluid inclusion clusters. The presence of fluid-related microstructures, in combination with highstress/high-strain rate microstructures such as deformation lamellae and undulose extinction, suggest that the detachment shear zone evolved close to the brittle-ductile transition. Healed microcracks and transgranular fluid inclusion planes occur in three sets, one sub-perpendicular to the mylonitic foliation, and a conjugate set oriented at~40-60°to the mylonitic foliation. Healed microfractures are filled with quartz, which has a distinct fabric, not as strong as the mylonitic fabric, suggesting that microcracks healed while the shear zone was still at conditions favourable for quartz crystal plasticity. Fluid inclusions also commonly decorate grain and subgrain boundaries as inter-and intragranular clusters. Our results show ductile overprint of brittle microstructures, which suggests that, during exhumation, the detachment shear zone may have crossed the brittle-ductile transition repeatedly, providing opportunities for fluid to permeate the detachment shear zone.