Syncing fault rock clocks: Direct comparison of U-Pb carbonate and K-Ar illite fault dating methods

The timing of slip on brittle faults in Earth’s upper crust is difficult to constrain, and direct radiometric dating of fault-generated materials is the most explicit approach. Here we make a direct comparison between K-Ar dating of fault gouge clay (authigenic illite) and U-Pb dating of carbonate slickenfibers and veins from the same fault. We have dated fault generated materials from the Big Creek fault, a northwest-striking, dextral strike-slip fault system in Yukon Territory, Canadian Cordillera. Both methods yielded dates at ca. 73 Ma and ca. 60–57


DATING BRITTLE FAULTS
Brittle faults record past seismic slip events and are major fluid conduits, channeling magma, ore-generating fluids, oil, gas, and water (Haines et al., 2016). In the fault valve model, fluids mediate fault strength, acting as a catalyst for seismic slip (Sibson, 1992). Determining periods of frictional slip and fault-related fluid flow is critical for testing fault strength models. Fault histories are also important for building regional tectonic frameworks, understanding ore genesis, forming basin evolution models, and studying seismically active faults that pose geohazards and engineering challenges (e.g., Zwingmann et al., 2010). Fault timing is typi-cally constrained through cross-cutting relationships, but is more precisely achieved by direct dating of fault-generated materials such as fault gouge, slip-surface hematite, opal, pseudotachylyte, and slickenfibers (Nuriel et al., 2012(Nuriel et al., , 2019aTagami, 2012;Ault et al., 2015;Tillberg et al., 2020).
Potassium-argon (K-Ar) and/or 40 Ar/ 39 Ar dating of authigenic illite, a K-rich clay mineral that forms instantaneously during faulting, is a common method for fault dating (van der Pluijm et al., 2001;Vrolijk et al., 2018;Tsukamoto et al., 2019). However, the potential for wallrock inheritance and multiple clay-generating slip events generally results in a positive correlation between K-Ar illite age and clay grain size (e.g., Torgersen et al., 2015aTorgersen et al., , 2015b, which can be challenging to deconvolve. Carbonate U-Pb geochronology can provide direct timing constraints for geological processes including diagenesis, fluid flow, and tectonic processes (e.g., Rasbury and Cole, 2009). In situ U-Pb dating has been applied to low-U fault-related carbonate (e.g., Roberts and Walker, 2016;Nuriel et al., 2017;Parrish et al., 2018;Roberts et al., 2020). Previous indirect comparison of fault-gouge illite 40 Ar/ 39 Ar and carbonate U-Pb ages suggests concordance between the two methods (Nuriel et al., 2019b), but no direct comparison has been undertaken by applying both techniques to colocated samples from the same fault system. Here we test the hypothesis that fault-gouge illite and faultrelated carbonate provide corresponding records of fault slip through direct comparison of K-Ar fault-gouge illite and U-Pb carbonate ages from a single fault system. Our study area lies within a copper-gold porphyry district where fault slip history is critical for developing the metallogenic framework for the region.

BIG CREEK FAULT
The Big Creek fault is a >150-km-long, northwest-southeast-striking, dextral strikeslip fault network located in the Yukon Territory, Canada (Fig. 1), between and subparallel to the Denali and Tintina dextral strike-slip faults (>350 km and ∼440 km of Paleogene displacement, respectively; Lanphere, 1978;Hayward, 2015). The Big Creek fault cuts the metamorphosed Paleozoic Yukon-Tanana terrane, one of several terranes accreted to Laurentia during the early Mesozoic to form the Canadian Cordillera (Colpron et al., 2006). Intrusive units within the Yukon-Tanana terrane include the Jurassic Long Lake    (Allan et al., 2013;Colpron et al., 2016;Friend et al., 2017;Fig. 1). A suite of extrusive rocks was emplaced ca. 72-67 Ma, including the Carmacks Group volcanic rocks (Allan et al., 2013). Both intrusive Cretaceous suites are associated with Cu-Au porphyry and related deposits that are aligned and spatially associated with the Big Creek fault (Johnston, 1999), including the Nucleus and Revenue deposits, which lie within a releasing bend between the main fault strands ( Fig. 1; Allan et al., 2013).
Regional mapping and geochronology provide a framework for Big Creek fault slip ( Fig. 1; Friend et al., 2017;Allan and Friend, 2018). The ca. 77 Ma Stoddart pluton is interpreted to plug the northern fault strand, providing a minimum age for slip. The 70 Ma Seymour Creek stock is truncated by the southern fault strand, providing a maximum age for slip. Thus, there were at least three apparent slip periods: (1) structurally controlling emplacement of the Whitehorse suite ca. 112-105 Ma, (2) slip coeval with emplacement of the Casino suite and associated mineralization during 79-72 Ma, and (3) slip after 70 Ma (Allan et al., 2013;Friend et al., 2017).

FAULT-GENERATED MATERIALS
The Big Creek fault crops out in a few meterscale fractured outcrops, where slickenlines and fault planes follow the regional northwest-south-east trend (Fig. 1D). We sampled freshly exposed fault gouge from three newly dug trenches: Seymour Creek (SC), Mechanic Creek (MC), and Happy Creek (HC; Figs. 1 and 2; precise loca-tions are listed in the Supplemental Material 1 ). SC gouge was sampled from a 20-cm-wide, clayrich seam oriented 097°/75° (strike/dip) within fractured and altered Stoddart pluton monzonite (ca. 77 Ma; Allan and Friend, 2018). MC gouge was sampled from a 150°-trending, 5-m-wide subvertical gouge zone bounded by silica-altered wall rock of 107 ± 1 Ma Whitehorse suite granodiorite (Friend et al., 2017). HC gouge was sampled from a 20-cm-wide gouge zone oriented ∼173/74° and hosted in Paleozoic layered schist.
The 6-km-long mineralized trend that hosts the Revenue-Nucleus Au-Ag-Cu-Mo deposits is contained mainly within Whitehorse suite granodiorite and surrounding Yukon-Tanana terrane (Fig. 1C). The deposits are interpreted to be high-level expressions of a porphyry system (Friend et al., 2017). Mineralization is thought to be related to magmatic fluids associated with the Late Cretaceous intrusive bodies. Steeply inclined hydrothermal breccia pipes, e.g., the Blue Sky porphyry breccia (BSPB), are interpreted to be emplaced along the margins of rotated blocks within the Big 1 Supplemental Material. Full analytical materials, gouge and detailed carbonate sample descriptions and locations, summary of additional U-Pb geochronology plots and interpretation, structural data, and Data Set S1 (K-Ar results), Data Set S2 (X-ray diffraction results), Data Set S3 (U-Pb carbonate results), and Data Set S4 (U-Pb standard reproducibility). Please visit https://doi .org/10.1130/GEOL.S.12730811 to access the supplemental material, and contact editing@geosociety.org with any questions. Creek fault releasing bend that are structurally-controlled by approximately north-southand northwest-southeast-striking, subvertical faults (Figs. 1C, 1E). Samples (all quoted as core depth in meters) collected from the BSPB drill core include carbonate slickenfibers from core RVD18-19 (314 m) and mineralized carbonate veins oriented ∼239°/73° from core RVD18-21 (542 m and 548 m) ( Fig. 1E; ∼90° to faults in the same core). Veins are characterized as (1) ∼1-mm-thick blocky carbonate with pyrite margins, and (2) ∼1-10-mm-thick vuggy carbonate with chalcopyrite, pyrite, and barite inclusions. Approximately north-south-striking fault gouge was collected from core RVD18-21 (557) m ( Fig. 2; see the Supplemental Material).

METHODS AND RESULTS
Fault gouge samples were separated via ultra-centrifugation to isolate clay-sized fractions 2.0-0.6, 0.6-0.2, and <0.2 µm, which were then characterized using X-ray diffraction and imaged using a scanning electron microscope (Fig. 2). The analytical method used for K-Ar dating is described in Zwingmann and Mancktelow (2004). Ages are reported at 2σ (see the Supplemental Material and Data Set S1 for the full methods and results). Gouge samples contain ∼25% (HC) to ∼100% (MC) illite. No other K-bearing phases are present except ∼2% K-feldspar in the 2-0.6 µm fraction of HC Data Set S2). Gouge ages range from ca. 79 Ma to 44 Ma, with common ages for MC and HC gouge, and for SC and BSPB gouge ( Fig. 2; Table 1).
Carbonate slickenfibers and veins were characterized under plane-polarized light. Suitable domains with high U (∼<1-10 ppm) and low Pb were identified with spot analyses. U-Pb isotopic analysis was conducted using laser ablation-inductively coupled plasma-mass spectrometry (LA-ICP-MS), as described in Parrish et al. (2018;Supplemental Material and Data Set S4). Analyses of seven carbonate polished sections yielded U-Pb ages of ca. 73 and 60-57 Ma (Table 1; Fig. 2; Data Set S3).

Fault Gouge
The MC and HC gouges, sampled close to the main fault trends, have very different host-rock ages (mid-Cretaceous and Paleozoic, respectively), yet have near-identical K-Ar results, with common ages of ca. 73 Ma and ca. 50 Ma (Fig. 2). BSPB and SC gouges, sampled from small fault zones situated within the fault bend blocks hosted by the same mid-Cretaceous granite as MC, yielded common ages of ca. 58 Ma and ca. 45 Ma (Fig. 2). Based on (1) a lack of primary igneous muscovite in wall rocks, (2) absence of other K-bearing phases in the dated fractions (except for minor traces of K-feldspar in the 2-0.6 µm fraction of HC gouge, which could explain its slightly older age of ca. 79 Ma), (3) clear disassociation between host-rock age and gouge illite K-Ar dates, and (4) illite morphology (Fig. 2), the dated illite samples are interpreted to reflect multiple fault-generated, authigenic illite ages with virtually no inherited component. The 2-0.6 µm fraction of SC gouge yielded a ca. 65 Ma age that was not reproduced in other size fractions or gouge samples. This lack of reproducibility suggests it represents a mixed age between older and ca. 58 Ma illite. For the above reasons, attempts to deconvolve the illite age results using the Illite Age Analysis method (Pevear, 1999), which assumes that different size fractions represent end-member mixtures of detrital grains and a single fault-gouge population, would be inappropriate for this study (e.g., Torgersen et al., 2015a).
All four samples yielded Eocene K-Ar ages for the <0.2 µm size fraction. This fine fraction could represent either late slip events, or partial thermal or fluid-related resetting that only affected the finest-grained illite. Thermal resetting of illite is possible above 260 ± 30 °C (Hunziker et al., 1986). Zircon (U-Th)/He ages from the study area are 79-70 Ma, indicating that the region was near or below ∼180 °C, the nominal closure temperature for He in zircon, during all dated slip events (Bineli Betsi et al., 2012). Furthermore, apatite (U-Th)/He ages of 57-39 Ma show that the region was at ∼60 °C or lower during the Eocene (Bineli Betsi et al., 2012). This rules out thermal resetting and suggests that the finest size fraction represents either two periods of illite growth during Eocene fault slip as implied in Figure 3, or growth during low-temperature fluid flow, or combined faultfluid events.

Sample
Grain size (µm)    Fig. 2). The Cretaceous U-Pb age is recorded in a carbonate vein, while both veins and slickenfibers record Paleocene ages. Thus, mineralized carbonate veining is associated with both periods of fault slip, consistent with mobilization of metals in carbonic fluids during syn-to post-mineralization faulting. The Cretaceous carbonate veins are not oriented parallel to the interpreted regional σ 1 (approximately north-south; Fig. 1E). Veins likely formed during crack-seal events along preferentially oriented (northeast-southwest) extensional fractures (Fig. 1E) parallel to a local σ 1 within rotated blocks in the fault bend (Fig. 3). These veins are parallel to regional northeasttrending faults ( Fig. 1F; Sánchez et al., 2014). Subsequent <60 Ma fault slip and related fracturing and fluid flow reactivated preexisting (ca. 73 Ma) fractures during movement on the Big Creek fault, consistent with K-Ar illite results.

FAULT SLIP ON THE BIG CREEK FAULT
Fault slip periods at ca. 73 Ma and ca. 60-57 Ma interpreted from both the K-Ar illite and U-Pb carbonate results support Big Creek fault field relations indicating slip at 77-70 Ma and <70 Ma (Friend et al., 2017). Fault-gouge illite and carbonate record late-to post-granite emplacement fault movement at ca. 73 Ma (Fig. 3). Fault-gouge illite and carbonate slickenfibers and veins show subsequent reactivation at ca. 60-57 Ma along north-striking fault planes in the releasing bend at shallower crustal levels (Bineli Betsi et al., 2012;Fig. 3). This is compatible with minor offsets of ca. 72-67 Ma Carmacks Group volcanic rocks along the fault trace to the northwest (Colpron et al., 2016). Paleogene slip likely occurred throughout the northwest-southeast-trending dextral strikeslip fault corridor, coeval with the Tintina and Denali faults (Lanphere, 1978;Hayward, 2015;Sánchez et al., 2014), and also coincides with fault-related magmatism and mineralization in northern British Columbia (Ootes et al., 2017) and in the Canadian Rocky Mountains (Pană and van de Pluijm, 2015), suggesting the existence of a regionally extensive fault-controlled fluidflow event. Subsequent slip at ca. 50 Ma (MC and HC gouge) and ca. 45 Ma (BSPB and SC gouge) can be inferred as continued reactivation as the Tintina and Denali faults continued to accommodate major displacements.

FAULTS AND FLUIDS
Our results provide, for the first time, an independent check on K-Ar illite and U-Pb carbonate direct fault dating methods. These methods provide complementary data sets, particularly useful toward resolving interpretational complexities arising from a fault history comprising multiple discrete faulting episodes. In the context of interpretation of K-Ar fault-gouge illite ages, our data corroborate previous work showing that different size fractions can record different slip periods (Torgersen et al., 2015a).
Because the age precision of our methods spans a few million years, the individual dates of fault-generated materials may span multiple seismic events during extended periods of fault activity. Furthermore, the study area records a complex faulting and fluid-flow history of which our data may represent only a partial history. However, the correspondence between field relationships and mineralized carbonate vein and fault-gouge illite ages contributes to understanding the interaction between faulting and fluid-flow periods. Although the carbonate veins formed coevally with seismic, gougegenerating fault slip periods, we do not have evidence for carbonate veins forming at other times in the absence of seismic slip. Within the constraints of our data set, the fault system may therefore have been impermeable during aseismic intervals. This implies empirical support for the fault-valve model, in which fluids are a major control on fault strength, and indicates that seismic pumping may have facilitated intervals of fluid flow (and hence mineralization) in this fault system (Sibson et al., 1975;Sibson, 1992). We anticipate that the ability to date multiple fault and fluid-related materials formed during fault slip events combined with chemical and stable-isotope methods will significantly advance future studies of fault-related fluid-rock systems.