Full Text (PDF) - American Mineralogist

American Mineralogist, Volume 91, pages 1473–1487, 2006

Fluid-mediated polymetamorphism related to Proterozoic collision of Archean Wyoming and Superior provinces in the Black Hills, South Dakota PETER I. NABELEK,1,* THEODORE C. LABOTKA,2 THOMAS HELMS,2 AND MAX WILKE1,3 2

1 Department of Geological Sciences, University of Missouri, Columbia, Missouri 65211, U.S.A. Department of Earth and Planetary Sciences, University of Tennessee, Knoxville, Tennessee 37996, U.S.A. 3 Current address: Institute für Geowissenschaften, Universität Potsdam, 14415 Potsdam, Germany

ABSTRACT Late Archean and Early Proterozoic continental margin pelites, graywackes, and quartzites in today’s Black Hills were regionally metamorphosed during the collision of the Archean Wyoming and Superior provinces beginning at ~1755 Ma. During east-west regional compression, metamorphism reached incipient garnet-biotite grade conditions at ~400 °C. Unzoned garnet (Sps44Alm44Pyr3Grs9) characterizes the mineral assemblage in graphitic metapelites. The low-temperature onset of the garnet-biotite assemblage is attributed to high Mn concentrations and low aH2O due to the presence of CH4 and CO2. During late stages of the collision, the rocks were intruded by the Harney Peak leucogranite (HPG). Vigorous fluid flow, evidenced by abundant quartz veins, metasomatism, and consumption of graphite from metapelites, occurred in the granite aureole. The lowest-grade aureole assemblage includes chlorite that overgrows the early regional foliation and new, clear, almandine rims on garnet. Higher-grade facies include staurolite, then sillimanite, and finally second-sillimanite. The prograde mineral assemblages are consistent with calculated pseudosection assemblages for the average metapelite. Beginning with the sillimanite zone, almandine garnet is mostly inclusion-free. Textures suggest that this new garnet grew after dissolution of the old garnet, for which the only remaining evidence are remnant inclusions or coarse quartz-biotite clots. The new garnet grew in response to more elevated temperatures. Garnet compositions projected onto the pseudosection indicate nearly isobaric heating between 3.5–4.5 kbar during prograde metamorphism. Occasional replacements of staurolite by muscovite and biotite are attributed to infiltration of the rocks by K-bearing magmatic fluids. Andalusite occurs as euhedral crystals in quartz veins or as late poikiloblasts along foliations planes, where it appears to have grown while Si-rich fluids passed through the rocks. Typically, it replaces earlier muscovite, suggesting that the fluids had low aH2O. The andalusite indicates decompression of the still hot, buoyant, fault-bounded block that included the HPG magma, relative to surrounding blocks in the terrane. This decompression contributed to the high-T/low-P conditions that existed in this portion of the thickened orogen and suggests that the various fault-bounded blocks in the Black Hills may have had different P-T histories. Keywords: Metamorphism, metapelites, Black Hills, low-pressure/high-temperature, garnet, pseudosection

INTRODUCTION Many collisional orogens include extensive belts of metamorphosed continental margin sediments that served as sources and paths for ascent of late-orogenic leucogranite magmas, for example in the Himalayas, the Appalachians, and the TransHudson orogen (Le Fort et al. 1987; De Yoreo et al. 1991; Brown and Solar 1998; Nabelek and Liu 2004). Pelitic rocks in collisional terranes often contain widespread andalusite-bearing assemblages, indicating regionally extensive low-pressure, high-temperature conditions. The andalusite-bearing (and other) mineral assemblages in the metamorphic rocks record processes that lead to mountain building and magmatism. The relationships among Paleoproterozoic orogenic deformation, * E-mail: [email protected] 0003-004X/06/0010–1473$05.00/DOI: 10.2138/am.2006.2120

metamorphism, and magmatism are very well exposed in the core of the Laramide-aged Black Hills uplift in western South Dakota. Here, Late Archean and Paleoproterozoic continental margin sedimentary rocks were caught up in the collision of the Archean Wyoming and Superior Provinces and served as sources for leucogranite magmas of the Harney Peak Granite. Andalusite- and cordierite-bearing assemblages localized near the granite appear to have developed in preexisting, regionally metamorphosed rocks, and have recorded magmatic and hydrothermal processes associated with emplacement and uplift of the Harney Peak Granite. Several studies have documented the orogenic processes and aspects of metamorphism in the Black Hills (e.g., Duke et al. 1990b; Redden et al. 1990; Terry and Friberg 1990; Helms and Labotka 1991; Nabelek et al. 1992, 1999, 2001; Friberg et al. 1996; Holm et al. 1997; Dahl et al. 1999, 2005a, 2005b). In

1473

1474

NABELEK ET AL.: POLYMETAMORPHISM IN THE BLACK HILLS, SOUTH DAKOTA

this paper, we synthesize the previous work with new mineral compositional data, textures, geochronological data, and phaseequilibria calculations, including pseudosections, to refine previous views of the pressure, temperature, and fluid conditions that existed during syncollisional regional and subsequent contact metamorphism by the late-orogenic Harney Peak Granite. In particular, we provide new evidence that the regional metamorphism occurred under low aH2O conditions and that the contact metamorphism was polybaric, during which the consumption and growth of some minerals was mediated by aqueous fluids. The new results lead to an improved understanding of the formation of low-pressure, high-temperature metamorphic terranes in upper-crustal sedimentary wedges that undergo deformation and burial during continental collisions.

TABLE 1. Average compositions of metasedimentary rocks* Avg. Metapelite Avg. Meta-graywacke 62.31 68.66 SiO2 0.68 0.58 TiO2 17.08 13.72 Al2O3 FeO 7.44 5.60 MnO 0.34 0.42 MgO 2.54 2.10 CaO 0.52 1.66 1.43 3.24 Na2O 4.42 2.47 K2O 0.13 0.14 P2O5 1.69 0.80 H2O * Average of analyses in Nabelek and Bartlett (2000).

GEOLOGIC SETTING AND CHRONOLOGY Much of the Proterozoic collisional terrane between the Wyoming and Superior Provinces is now buried under Phanerozoic sedimentary rocks, but it was partly exposed in the core of the Black Hills by late-Mesozoic to early-Tertiary Laramide uplift and erosion (Redden et al. 1990). In Canada, the collisional terrane between the Superior Province and smaller cratonic blocks to the northwest comprises the Trans-Hudson orogen. It contains relatively young rocks that generally yield >1800 Ma dates (Lewry and Collerson 1990). In the Black Hills, the metamorphic rocks are dominantly metaquartzites and quartz–muscovite–biotite metapelites and metagraywackes, most of which originated as turbidites (Fig. 1; Table 1). Some amphibolites, Fe-formations, and marbles also occur. The metasedimentary rocks were intruded during late stages of the Superior–Wyoming collision by the Harney Peak Granite (HPG), its satellite plutons, and numerous simple to complex pegmatites (Redden et al. 1982; Duke et al. 1990b; Norton and Redden 1990; Nabelek et al. 1992). The HPG and many of the satellite plutons were constructed by repeated intrusion of granite magma. There are also small exposures of Archean peraluminous granites and pelitic rocks near Bear Mountain along the western margin and near Little Elk Creek at the northwestern margin of the exposed Precambrian terrane (Gosselin et al. 1988). A COCORP profile across the orogen, just south of the U.S.-Canadian border, and ages of cores from the basement show that there is an intervening, buried Archean crustal block, named the Dakota block, between the two cratons (Baird et al. 1996). The Dakota block then, rather than the Superior craton, collided directly with the Wyoming craton. The rocks exposed in the Black Hills are probably part of a metasedimentary wedge that was brought onto the Wyoming and Dakota cratonic blocks during the collision. Recent chemical and isotopic dating of monazite from several key locations in the Black Hills (Dahl et al. 2005a, 2005b) supports previous evidence for a protracted history of deformation, metamorphism, and magmatism in the Black Hills (Gosselin et al. 1988; Redden et al. 1990). The oldest known rocks appear to be the late-Archean granites at Bear Mountain and Little Elk Creek with ages 2596 and 2560, respectively (McCombs et al. 2004). Although some Archean-age sedimentary rocks (now metamorphosed) crop out near these granites, most sedimentary rocks were deposited between ~2100 and 1880 Ma, as evidenced by ages of intercalated gabbro sills and felsic tuffs (Redden et al.

FIGURE 1. Geologic map of the Precambrian rocks in the Black Hills. Formation boundaries and faults (heavy solid lines) are based on the map of DeWitt et al. (1989) and local quadrangle maps. DeWitt et al. (1986) provide a description of the individual formations. Metamorphic isograds (dashed lines) are based on assemblages in Helms and Labotka (1991) and new data. The andalusite zone (thin dashed line) may have developed after the other metamorphic zones. Archean exposures at Bear Mountain (BC) and Little Elk Creek (LEC) are noted.


1990). The earliest deformation of the sedimentary sequences east of the Grand Junction fault (Fig. 1) resulted in northeasttrending F1 folds that do not show penetrative deformation (Redden et al. 1990). However, near Bear Mountain, similarly oriented deformation (D1) was penetrative, suggesting that it occurred deeper in the crust, consistent with the local occurrence of kyanite. This deformation was dated between 1785 and 1775 Ma by analysis of related monazite (Dahl et al. 2005a, 2005b) and has been attributed to north-vergent accretion of island arcs along the Cheyenne belt in southeastern Wyoming (Dahl et al. 1999). The dominant structures in the metasedimentary rocks throughout the Black Hills are NNW-trending F2 folds with steeply dipping foliation. These are thought to be related to the collision of the Wyoming and Dakota blocks. Major steeply dipping faults that superimposed contrasting metasedimentary rocks against each other have similar orientations. D2 appears to have begun at ~1755 Ma (Dahl et al. 2005a). The generally younger ages in the Black Hills compared with those from the Trans-Hudson in Canada have led to arguments for a distinct collisional event called the Black Hills orogeny or the Dakota orogeny (Chamberlain et al. 2002; Dahl et al. 2005b). Emplacement of the HPG and some satellite intrusions flattened the steeply dipping regional F2 foliation (Duke et al. 1988; Hill et al. 2004). The granite bodies themselves are undeformed except locally by subsequent emplacement of additional dikes and sills. Redden et al. (1990) obtained a concordant 1715 ± 3 Ma U–Pb age on a monazite from a thin granite sill. This is generally considered the best date for crystallization of the HPG and is consistent with a 1716 ± 3 inclusion age from stepwise heating of garnet and a 1717 ± 2 Ma chemical age on monazite from the contact aureole (Schaller et al. 1997; Dahl et al. 2005a). Yang and Pattison (2006) obtained 343 chemical ages on monazites from the HPG thermal aureole ranging from ~1800 to ~1650 Ma with a mean age of 1713 ± 39 Ma. The meaning of these ages is difficult to interpret because Yang and Pattison (2006) did not specifically correlate different monazites to the deformation events that have occurred and do not show the spatial distributions of ages in individual monazites. Their data are consistent, however, with extensive resetting of ages during magma intrusion. The extensive resetting of ages is in accord with the almost complete overprinting of regional mineral assemblages by higher-temperature assemblages in the inner contact aureole (see below). Published ages are equivocal, however, about the duration of magmatism and contact metamorphism. The many granite and pegmatite intrusions in the southern Black Hills and their growth by repeated emplacement of granite magma dikes (Duke et al. 1990b; Norton and Redden 1990) suggest a protracted period of magmatism and hence contact metamorphism. In fact, Redden et al. (1990) reported a 1728 ± 4 Ma age from two highly discordant zircons from the granite, and Krogstad and Walker (1994) obtained concordant 1704 to 1700 Ma U–Pb ages on apatite from the Tin Mountain pegmatite, located to the west of the HPG. On the one hand, the range of ages may represent uncertainty due to inheritance, discordancy, and differences in blocking temperatures of the U–Pb system in different minerals (Krogstad and Walker 1994). On the other hand, they may also indicate the duration of the leucogranite magmatism (Sirbescu

1475

and Nabelek 2003). Whichever the case may be, the combined metamorphic and HPG ages show that granite magmatism occurred toward the end of protracted regional metamorphism and deformation of the sedimentary rocks.

METAMORPHISM The interpretation of metamorphism in this paper is based on samples previously discussed by Helms and Labotka (1991), which came mostly from on the aureole of the HPG, and new samples that also cover the lowest regional metamorphic grades in the Black Hills far to the north of the HPG (Table 2; Fig. 1). Helms and Labotka (1991) mapped four isograds, which marked the first appearances of staurolite, andalusite, sillimanite, and sillimanite + K feldspar (Fig. 1). They determined temperatures and pressures from the compositions of garnet, biotite, and plagioclase, and the activity of H2O from the compositions of plagioclase and of muscovite, which contains a significant paragonite component. The emphasis of the present paper is on interpreting the fluid-present metamorphic history of the terrane using mineral phase equilibria within the context of textures. Therefore, we describe the salient features of pelitic schists throughout the metamorphic sequence. Biotite zone Minerals1 that occur in all pelitic assemblages include quartz, muscovite, biotite, and minor plagioclase. The biotite zone is confined to the northeastern part of the field area shown in Figure 1, mostly east of the Silver City fault, which runs through much of the terrane. Its extent is smaller than indicated on previously published sketch maps of the Black Hills because the area containing garnet is more extensive than suggested previously. Biotite-zone rocks are very fine-grained. Pelitic rocks contain the assemblage biotite + muscovite + quartz. The micas define very well the regional NNW-trending foliation. Disseminated graphite commonly occurs in centimeter-scale bands, which became sites for incipient garnet growth in the garnet zone. However, near the Silver City and the Pactola Lake faults, some outcrops are extremely graphitic, almost totally black. Graphite in those rocks may be indigenous or may have been precipitated from synmetamorphic fluids. Chlorite is rare. Lower garnet zone The distinction between the lower and upper garnet zones is subtle. Whereas the lower garnet zone has only the regional structure and mineral assemblage, the upper garnet zone also includes a contact-metamorphic overprint. Because the overprint boundary is diffuse, the boundary between the lower and upper garnet zones is poorly defined and therefore not shown in Figure 1. The most distinguishing feature of the lower garnet zone is abundant Mn-rich garnet (Stone 2005) that occurs primarily in 1

Deposit item AM-06-030, Appendix 1: Representative mineral compositions in Black Hills schists. Deposit items are available two ways: For a paper copy contact the Business Office of the Mineralogical Society of America (see inside front cover of recent issue) for price information. For an electronic copy visit the MSA web site at http://www.minsocam.org, go to the American Mineralogist Contents, find the table of contents for the specific volume/issue wanted, and then click on the deposit link there.

1476 TABLE 2.

NABELEK ET AL.: POLYMETAMORPHISM IN THE BLACK HILLS, SOUTH DAKOTA Sample locations and calculated garnet-biotite equilibration temperatures

Sample* Latitude Longitude TC† (4.5 kbar) H&S‡ (4.5 kbar) HP81-1 43.943 –103.387 HP95-1 core 43.798 –103.697 521 543 HP95-1 rim 43.798 –103.697 530 552 HP98-1 43.957 –103.597 HP112-1 43.900 –103.519 HP118-1 43.800 –103.567 HP121-1 43.806 –103.560 HP124-1 43.811 –103.555 HP137-1 core 43.932 –103.508 512 547 HP137-1 rim 43.932 –103.508 505 538 HP140-1-25 core 43.927 –103.498 542 550 HP140-1-25 rim 43.927 –103.498 516 534 HP140-1-26 core 43.927 –103.498 HP140-1-26 rim 43.927 –103.498 HP197-1 43.851 –103.636 552 581 HP208-1 core 43.930 –103.627 403 448 HP208-1 rim 43.930 –103.627 450 491 HP208-2 43.930 –103.627 HP213-2 43.964 –103.629 HP222-1 44.100 –103.650 384 426 HP224-6b 43.919 –103.442 540 560 HP227-1 core 44.137 –103.631 344 384 HP227-1 rim 44.137 –103.631 332 376 HP229-1 44.192 –103.760 HL2 43.781 –103.575 566 593 HL3 43.812 –103.589 578 597 HL5 43.836 –103.436 501 525 HL6 43.884 –103.445 470 536 HL8 43.866 –103.651 482 496 HL9 core 43.919 –103.698 323 380 HL9 rim 43.919 –103.698 449 492 HL12 43.966 –103.553 374 412 HL12 43.966 –103.553 300 334 HL14 core 43.943 –103.497 491 517 HL14 43.943 –103.497 458 485 HL15A 43.912 –103.493 526 578 HL17 43.939 –103.429 458 488 HL19 core 43.964 –103.606 426 474 HL19 rim 43.964 –103.606 427 468 HL20 core 43.932 –103.654 368 407 HL20 rim 43.932 –103.654 454 467 HL22 43.913 –103.512 460 487 HL23 core 43.917 –103.661 347 373 HL23 rim 43.917 –103.661 501 516 HL24 43.906 –103.652 HL25B 43.908 –103.506 525 547 HL27 43.786 –103.646 448 484 HL28 43.869 –103.598 520 544 * Samples denoted by “HP” are newly discussed; samples denoted by “HL” were previously also discussed in Helms and Labotka (1991). † Temperatures calculated using THERMOCALC. ‡ Temperatures calculated using the thermometer of Hodges and Spear (1982).

the graphite-rich bands. The garnet is full of graphite inclusions concentrated between sectors of garnet crystals (Fig. 2a). Inclusions of other minerals, including quartz and accessory phases, also occur. The garnet is mostly unzoned and has a composition of approximately Sps44Alm44Pyr3Grs9 (Fig. 2b). Ilmenite in these bands contains up to 47% pyrophanite component (Mnilmenite). The inclusion-rich, ragged-looking garnet occurs not only throughout the lower garnet zone but also persists in the staurolite zone in the Thompson Draw synform, west of the HPG. In sheared rocks, the garnet is often elongated parallel to the shear direction, which may indicate its preferential growth along a fluid pathway. Biotite was analyzed in two samples of the lower garnet zone. In one sample the XFe of biotite, defined as Fe/(Fe + Mg)·100, is 65 whereas in the second sample it is 55. TiO2 concentration is 1.6 wt% in both samples, and MnO is 0.29 and 0.38 wt%,

FIGURE 2. (a) Photomicrograph of sample 227 showing garnet in a graphitic band. (b) Composition profile of a garnet in sample 227 showing the spessartine-rich, relatively unzoned compositions of lower garnet-zone garnet.

respectively. These MnO concentrations are greater than at higher grades, except in the second-sillimanite zone. Biotite is homogeneous within individual thin sections. Upper garnet zone There are two distinguishing features of the upper garnet zone that show the contact-metamorphic overprint. The first feature is the appearance of large chlorite grains that overgrew the regional foliation (Fig. 3a). The chlorite is especially prominent in the vicinity of the staurolite isograd. Commonly, it is associated with thin quartz veins that cut through rocks. The second feature is clear, mostly inclusion-free overgrowths on garnet (Fig. 3a). The overgrowths impart strong zoning profiles on garnet where the overgrowths are significantly poorer in spessartine, with XSps ≈ 0.095 at rims of crystals, and richer in almandine (Fig. 3b). The strong zoning typically begins at the core-overgrowth boundaries. XGrs is lower in the overgrowths than in the cores, down to ~0.035.


1477

garnet (Fig. 4b). Instead, there is a small grain of a new inclusion-poor garnet. Inclusion-poor garnet crystals, even some that are several millimeters in diameter, commonly occur near clots of quartz that are rimmed by coarse biotite (Fig. 4c). These clots, probably alteration products of consumed garnets, give the schist a spotted appearance. In the staurolite zone, inclusion-rich garnets with various thicknesses of overgrowths are found in the Thompson Draw synform (Fig. 1), which were noted previously by Friberg et al. (1996). New inclusion-poor garnet grains occur mostly north of the HPG and along its western inner aureole. The overgrowths on old garnet grains are strongly zoned with compositions at rims approaching compositions of the inclusion-poor garnets, which have only minor zoning (Fig. 4d). The cores are XSps ≈ 0.16, XAlm ≈ 0.70, XPyr ≈ 0.087, XGrs ≈ 0.045. The rims are XSps ≈ 0.095, XAlm ≈ 0.77, XPyr ≈ 0.092, XGrs ≈ 0.045. The inclusion-rich grains throughout the metamorphic terrane may correspond to garnet in which Dahl and coworkers (Dahl and Frei 1998; Dahl et al. 2005a) dated monazite inclusions at ~1755 Ma, whereas the inclusion-poor garnets probably correspond to monazites that give ~1715 Ma ages. Notable is the absence of graphite in the matrix of many rocks, even those that contain inclusion-rich garnets, in contrast with graphitic matrices of rocks below the staurolite isograd. Andalusite zone

FIGURE 3. (a) Photomicrograph of sample 169 showing garnet grains with inclusion-rich cores and inclusion-poor rims. The large grain at the top of the micrograph is chlorite that grew across the preexisting foliation. (b) Compositional profile across a garnet grain in sample 213 from the upper garnet zone showing a spessartine-rich core and an almandine-rich rim. Inclusions are confined to the Mn-rich core.

Staurolite zone The textural features of the staurolite zone are complex. At the isograd, staurolite appears to have formed by the breakdown of chlorite and seems to be in equilibrium with new, inclusionfree garnet. Commonly, the garnet is included within staurolite. Throughout the zone, staurolite grains are mostly oriented randomly and seem to overgrow the regional foliation. Textural evidence suggests that the growth of new, inclusionpoor garnet occurred by destruction of old, inclusion-rich garnet. The new garnet may have grown in a fluid-rich environment as indicated by the presence of millimeter to centimeter thick quartz veins. Figure 4a shows a partially resorbed inclusion-rich garnet crystal whose immediate neighborhood includes quartz and coarsened biotite. In the same rock is a clot with the same assemblage, including inclusions, but without an inclusion-rich

Andalusite and cordierite occur in some upper-staurolite zone rocks to the north of the HPG (Helms and Labotka 1991). The assemblage is andalusite + staurolite + garnet + biotite or andalusite + garnet + biotite. The garnet grains are usually inclusion-free and unzoned (Fig. 5). A few garnets retain their Mn-rich cores, however. Andalusite is mostly found near shear zones and faults where it grew along foliation planes. Typically it is poikiloblastic, several centimeters in diameter parallel to the foliation direction but very thin perpendicular to the foliation direction. Some grains have chiastolite crosses, but others have a dendritic shape. Biotite inclusions preserve the overall foliation in the rock and quartz inclusions are patchy, essentially undeformed. Notable is the lack of muscovite inclusions in the poikiloblasts, except for occasional coarse, randomly oriented crystals. This texture suggests that andalusite grew late during metamorphism. Blocky, euhedral andalusite crystals, locally several centimeters long, are confined to quartz veins where they grew inward from vein walls or within the veins. They commonly coexist with fluid-precipitated graphite and undeformed staurolite growing into the veins from the margins. Sillimanite zone The sillimanite zone is marked by the occurrence of fibrolitic sillimanite and the lack of staurolite except for a few remnant grains that persist beyond the sillimanite isograd. Locally, staurolite is altered to muscovite + biotite. Sillimanite is common in the northern and inner western part the HPG aureole, but is uncommon southwest of the HPG where metagraywackes are the dominant rock type. Sillimanite is generally found within the regional foliation, which in this zone was rotated and flattened by intrusion of the HPG (Hill et al. 2004). Most garnet is unzoned, inclusion-poor, and shows some rotation during deformation.

1478


FIGURE 4. Photomicrographs (a–c) show the progressive replacement of inclusion-rich garnet by regions of quartz and micas and the growth of inclusion-free garnet. (d) Compositional profile across an inclusion-free almandine garnet in sample 124 is essentially unzoned.

Most rocks are graphite-poor, except near veins where new graphite was precipitated. The outer boundary of the granite-pegmatite field roughly parallels the sillimanite isograd to the west and the andalusite isograd to the north of the HPG (Norton and Redden 1990). The correlation suggests a link between high-grade metamorphism and magmatism in this region. Second-sillimanite zone The breakdown of muscovite that occurred to the southwest of the HPG and in schist inliers within the pluton was accompanied by melting that resulted in stromatic migmatites (Shearer et al. 1987; Nabelek 1997, 1999). The migmatite zone may be underlain by granite. Melanosomes in the migmatites are dominated by coarse biotite, sillimanite, quartz, plagioclase (An16–2), tourmaline, and poikiloblastic K-feldspar (Or89–95) that has up to 1.8 wt% BaO. Leucosomes have either a tourmaline–granite mineral assemblage or are composed of several centimeter long

bands of sillimanite, instead of the expected feldspar, coexisting with quartz. Minor retrograde muscovite and tourmaline locally overgrows the sillimanite. The stabilization of sillimanite instead of feldspar in the leucosomes can be attributed to high fHF as evidenced by the presence of fluorapatite (Nabelek 1997). Except for a few localities, garnet is absent from the second-sillimanite zone. Evidence for a residue-melt equilibrium, which may have existed during partial melting, was overprinted by an approach to solid-state equilibrium among the migmatite minerals during cooling (Nabelek 1999). Summary of observations Mineral textures in the pelitic rocks suggest a contact-metamorphic overprint on regional, dominantly garnet-grade assemblages attendant on aureole deformation and fluid flow associated with granite and pegmatite emplacement. Evidence for fluid flow is recognized by: (1) production of chlorite that overgrows the regional foliation; (2) frequent alteration of garnet and staurolite


1479

and have graphite-enriched selvages; and (6) enrichment of the aureole in Li, B, Cs, and other magma-derived elements (Shearer et al. 1984, 1986; Redden and Norton 1992; Duke 1995; Wilke et al. 2002). Thus, the prograde mineral reactions in the aureole of the HPG appear to have operated on the chlorite + biotite + garnet + muscovite + quartz + plagioclase assemblage that occurs below the staurolite isograd and that records the lowest-grade thermal and fluid-flow effects of magmatism. Andalusite appears to record the last stages of metamorphism in the HPG aureole.

MINERAL EQUILIBRIA Stability fields of assemblages

FIGURE 5. (a) Compositional profile across an essentially unzoned garnet grain in a sample from the andalusite zone. (b) A back-scattered electron image of the garnet shown in a. The garnet is inclusion-poor and rounded. (c) AFM projection of the andalusite-zone mineral assemblage showing the composition of the average metapelite (Table 1), which was used to calculate the pseudosection in Figure 6.

to assemblages containing micas; (3) apparent fluid-mediated consumption of old garnet and production of new garnet; (4) disappearance of most graphite from the matrix of many rocks in the aureole; (5) abundant quartz veins that cut across foliation

It is useful to examine stability ranges of the mineral assemblages using the pseudosection approach because it relates mineral assemblages and mineral compositions in a given bulkrock composition directly to P-T conditions of metamorphism (Powell et al. 1998). The high Mn concentrations in Black Hills metapelites require consideration of this element in construction of pseudosections as Mn strongly impacts the stability range of garnet (Hodges and Spear 1982; Spear and Cheney 1989; Mahar et al. 1997). We constructed a P-T pseudosection for the average composition of Black Hills metapelite (Table 1; Fig. 5c) in the system MnO–Na2O–CaO–K2O–FeO–MgO–Al2O3–SiO2–H2O. The pseudosection in Figure 6 was calculated with the program THERMOCALC (v. 3.23) using the Holland and Powell (1998) thermodynamic database, updated in November 2003 to include the aluminosilicate triple point at 550 °C and 4.5 kbar from Pattison (1992). This location of the triple point effectively enlarges the stability field of andalusite compared with the triple point at 500 °C and 3.8 kbar (Holdaway 1971). Water was assumed to be in excess because of the field and petrographic evidence for its presence in the HPG aureole. It is noted, however, that there was a significant carbonic component in the fluids, especially during the regional M1 metamorphism (Huff 2004), which affected the mineral assemblages (discussed below). K-feldspar was assumed to be absent at temperatures lower that the maximum stability of muscovite, although detrital grains occur in some metagraywackes. Quartz, muscovite, and plagioclase occur in all subsolidus assemblages, but have liquidi above the solidus. Boundaries between some fields contain low-variance assemblages, which, however, are too narrow to show in the figure. Compared with Mn-poor systems (Spear and Cheney 1989; Mahar et al. 1997; Powell et al. 1998; Tinkham et al. 2003; Wei et al. 2004), most notable is the expanded stability field of garnet to 590 °C in the second-sillimanite zone. This pattern of increasing temperature toward the southern part of the HPG has been noted previously (Helms and Labotka 1991; Friberg et al. 1996). It suggests a thermal imprint by the HPG. However, the reason for the maximum being centered on its southern part is not clear. One possibility is that the second-sillimanite zone represents part of the roof of the pluton through which hot fluids were escaping from the magma, whereas elsewhere, the

FIGURE 10. Distribution of peak temperatures recorded by garnet rims and coexisting biotite. The solid gray lines are contours of 480, 520, and 590 °C. The dashed lines are the garnet, staurolite, sillimanite, and sillimanite + K feldspar isograds.

metamorphic rocks are mostly along the sides of the pluton or far above deeply buried granite bodies. There is gravity evidence for the existence of a granite body just under the second sillimanite zone, called the Park Dome (Duke et al. 1990b). Although there is a consistent regional distribution of apparent garnet–biotite equilibration temperatures, they are lower than calculated temperatures of the mineral assemblages or those indicated by garnet compositions in the pseudosection (Fig. 8). For example, for the staurolite zone, the Hodges and Spear (1982) thermometer gives ~490–580 °C temperatures, whereas the mineral assemblage and garnet rim compositions alone suggest 540–600 °C (Figs. 6 and 8). For the second-sillimanite zone, the thermometer gives 590–600 °C, whereas the solidus is at ~650 °C. Thus, the thermometer temperatures are ~50 °C lower, and the garnet-biotite equilibration temperatures from the thermodynamic database are still lower. One possibility for this discrepancy is retrograde reequilibration of biotite with Fe-bearing fluids whereas, because of low diffusivities of its components, garnet preserves conditions that existed during its growth.


Prograde reactions These textural and pseudosection analyses of mineral assemblages suggest that regional metamorphism at the current erosion level of the southern Black Hills reached garnet–biotite conditions under low aH2O at 3–4 kbar, on which thermal and fluid-flow effects associated with magma emplacement were superimposed. From the lowest-grade, contact-metamorphic assemblage to staurolite-grade conditions, the reaction was: chl + grt + ms → st + bt + qtz + H2O

(1)

This reaction is continuous over a very narrow temperature interval. The subsequent breakdown of staurolite, defining the sillimanite isograd, presumably occurred by the reaction st + ms + qtz → sil + grt + bt + H2O

(2)

Several samples from the sillimanite isograd contain both reactant and product minerals. In the field, the breakdown of muscovite at the second-sillimanite isograd appears to be related to melt production by the reaction: ms + pl + q (±H2O) → melt + sil ± kfs

(3)

Although muscovite occurs in leucosomes, it is clearly a retrograde alteration product of sillimanite that is common in the migmatites. There is no evidence for prograde breakdown of muscovite to kfs + sil + H2O before the onset of migmatization. There is also no evidence for coexistence of muscovite with granite melt that is suggested by the broad thermal stability field in which muscovite coexists with melt at >3.5 kbar (Fig. 6), or by pseudosections of White et al. (2001) for their “average” metapelite. A possible explanation for this discrepancy is that the reaction occurred under low aH2O conditions, either because a fluid was not present (least likely), because the fluid was not pure H2O, or because as introduced water drove the melting reaction, it dissolved in the melt so that the system remained effectively water-absent. Under water-absent conditions, the melt-present stability field of muscovite narrows considerably, essentially to a discontinuous breakdown of muscovite upon melting (Patiño-Douce and Harris 1998).

1483

The proportion of muscovite to biotite in the replacements is about the same as in this reaction. This reaction, assuming pure Fe end-members of staurolite and biotite, is represented on a log(aH+/K+) vs. log(aH2O) diagram in Figure 12. It has the form log aH+/K+ = 1.210 log aH2O + log Keq

(5)

The equilibrium constant was calculated using thermodynamic parameters of fluid species in the Slop98 thermodynamic database that is used as input to the program SUPCRT92 (Johnson et al. 1991). The equilibrium constant has only a weak dependence on pressure and increases with decrease in temperature. It does not change appreciably if aqueous SiO2 instead of quartz is assumed to be a reactant. Thus, the reaction is predicted to be driven away from staurolite stability by infiltration of K-bearing, aqueous fluids, such as those that have emanated from the HPG and nearby pegmatites (Sirbescu and Nabelek 2003). The garnet compositions and sequence of prograde reactions suggest a trend of increasing temperature at pressures above the apex of andalusite stability. Locally, some andalusite that occurs in the HPG aureole may have formed by reaction analogous to Reaction 2, as suggested by Helms and Labotka (1991). However, the most common occurrence of andalusite, as dendritic poikiloblasts that grew along foliation planes and as euhedral crystals in quartz veins, suggests that its growth was related to fluids that carried silica and picked up Al as they flowed through the metapelites. At the conditions of metamorphism, the solubility of Al in metamorphic fluids can be significant, even when in equilibrium with the low-solubility corundum (Walther 2001). The solubility especially increases with decrease in pH. The fact that foliated muscovite is usually absent from the andalusite poikiloblasts suggests the reaction: 2 ms + 2 H+ → 3 and + 3 q + 2 K+ + 3 H2O

(6)

Ion-exchange reactions Two mineral replacement features are common in the HPG aureole. They demonstrate a shifting composition of fluids in the aureole during metamorphism. The first is replacement of staurolite by muscovite + biotite at some locations (Fig. 11). This sample has S3 (syn-granite emplacement) foliation wrapping around an inclusion-poor garnet, which appears to be stable, and the replaced staurolite. Biotite in the pseudomorph is quite a bit coarser than in the foliated matrix. Clearly, the S3 foliation preceded alteration of staurolite. Alteration of staurolite suggests the following replacement reaction: 0.290 st + 1.911 q + 1.210 H2O + K+ → 0.806 ms + 0.194 bt + H+ (4)

FIGURE 11. Muscovite and biotite replacing staurolite in the upperright portion of the photograph. Note that matrix biotite is wrapped around the replaced staurolite whereas the replacing muscovite and biotite are not foliated.

1484


estimated for migmatites in the second sillimanite zone also are shown, with aH+/K+ determined from apatite and muscovite compositions (Nabelek 1997). Low-aH2O conditions could have been achieved by incorporation of carbonic components into fluids, for which there is ample evidence in the Black Hills (Duke et al. 1990a; Huff 2004). Flow of low-aH2O fluids along shear zones, faults, and foliation planes would drive reaction 6 toward andalusite + quartz, which dominate the poikiloblastic andalusite domains. Relatively elevated aH+/K+ conditions can result from release of F from muscovite during its breakdown. Muscovite below the second sillimanite isograd typically contains ~0.1 wt% F and 9.5 wt% K2O, which is equivalent to aH+/K+ of ~1 × 10–2. At the metamorphic conditions, F should be fully associated with H+ (Zhu and Sverjensky 1991), so that F is an indicator of the minimum H+ activity. It is notable that reaction 6 occurs on the staurolite side of reaction 4. Indeed, the assemblage staurolite + andalusite + quartz + graphite is seen in some veins in the HPG aureole. The occurrence of late andalusite in these rocks indicates that the P-T paths of the rocks passed through the andalusite stability field during retrogression while magma-induced fluid flow was still active. Cordierite that occurs in some more magnesian rocks in the HPG aureole also indicates a low-pressure portion of the metamorphic history (Helms and Labotka 1991). These rocks are not plotted in the pseudosection, however, because it is not appropriate for the compositions of these rocks.

DISCUSSION Role of fluids in metamorphism

F IGURE 12. (a and b) Diagrams showing the dependence of sillimanite, muscovite, and K-feldspar stability on log(aH+/K+) and log(aH2O) at two sets of P-T conditions. The relevant ion-exchange reactions are described in the text and are marked by numbers. Superimposed is the reaction between staurolite and muscovite + biotite. Estimated fluid composition in migmatites in the second-sillimanite zone (Nabelek 1997) is shown in part b. Introduction of a K-bearing fluid results in the breakdown of staurolite to biotite + muscovite in a. A decrease in aH2O because of introduction of a carbonic fluid species results in breakdown of muscovite to andalusite in b.

Other relevant reactions involve the stability of K-feldspar in the fluid-bearing system: 2 ksp + 2 H+ → and + 5 q + 2 K+ + H2O

(7)

3 ksp + 2 H+→ ms + 6 q + 2 K+

(8)

These reactions are shown in Figure 12 and have forms analogous to Equation 5. Equilibrium constants for these reactions also were calculated using the relevant thermodynamic parameters in the Slop98 database. The diagram shows that andalusite is stable relative to muscovite and K-feldspar at relative high aH+/K+ and low aH2O at 500 °C and 3 kbar, conditions at which andalusite is stable relative to sillimanite or kyanite. For reference, conditions

This study makes it quite evident that flow of compositionally variable fluids must be considered in studies of high-T–low-P metamorphism that is induced by magma intrusion. The fluids can affect the stability fields of mineral assemblages beyond the simple controls by pressure and temperature. Our analysis of mineral assemblages and compositions in the Precambrian core of the Black Hills gives evidence for superposition of reactions induced by intrusion of the late-orogenic HPG on rocks that were already deformed and metamorphosed during the early part of collision of the Wyoming and Superior cratons. The fluid-flow system that was generated by magma intrusion had a substantial effect on the mineral assemblages. The early regional metamorphism of graphitic metapelites began at ~1755 Ma. It occurred in the presence of fluids with a large carbonic component that was responsible for the broad coexistence of garnet and biotite and the lack of chlorite even at temperatures as low as ~400 °C. Garnet was additionally stabilized by high concentrations of Mn in graphitic metasedimentary layers in which it began to grow (Stone 2005). The late-orogenic metamorphism in the aureole of the HPG occurred in the presence of fluids in which aH2O varied such that much graphite was destroyed from matrixes of many rocks but also precipitated in vein selvages. The assemblage containing garnet + biotite + chlorite is the lowest-grade contact-metamorphic assemblage that was superposed on the regional garnet + biotite assemblage. Although the higher-grade assemblages in the HPG aureole are largely the product of heat flow from the granite, low-aH2O fluids that flowed along foliation planes and along shear-zones and


faults were probably responsible for the late production of poikiloblastic andalusite from muscovite that is seen in some parts of the HPG aureole. Conversely, the occurrences of staurolite replacements by muscovite + biotite resulted from an influx of K-bearing fluids emanating from the magma. Fluids also appear to have mediated the destruction of older, regional garnets and the production of inclusion-poor new garnets at the more elevated contact metamorphic temperatures. The clots of quartz with biotite rims that mark places where garnets once occurred give some schists a spotted appearance. Dahl et al. (2005b) attributed the production of such clots and growth of new garnet to a two-stage process, in which the spots were first generated by a retrograde ion-exchange reaction involving the breakdown of garnet during influx of H2O, K+, and Fe2+, followed by a prograde production of new garnet by reaction of biotite with quartz and either paragonite or H+. They did not account for Mn in their proposed reactions and did not identify a source for the ion-bearing retrograde fluids. We argue instead that the destruction of the old garnet and the production of the new garnet simply was related to the higher temperatures in the contact aureole where the old spessartine-rich garnet became unstable in favor of the more almandine-rich garnet as required by the new equilibrium conditions (Fig. 8). Fluids were the catalyst for the recrystallization. Metamorphism and uplift of the Harney Peak Granite block All evidence from the mineral assemblages, textures, and garnet compositions indicates that regional metamorphism along the north-south axis of the Black Hills sampled for this study occurred at pressures of 3–4.5 kbar. This pressure range contrasts with the marginal Archean exposures at Bear Mountain and Little Elk Creek that had a history of pressures at ~7.5 kbar prior to their juxtaposition against the lower pressure axis of the Black Hills (Terry and Friberg 1990). However, there are several lines of evidence that contact metamorphism in the block (see Fig. 1) that includes the HPG and is bounded by the Keystone-Empire Mine faults on the east side and the Hill City fault on the west side, the “HPG block,” was polybaric. First, the late andalusite suggests that the HPG aureole was undergoing decompression while still hot. Second, whereas the F2 regional foliation within the HPG block was flattened probably because of magma intrusion underneath, leading to numerous recumbent folds, outside this block the foliation has remained steep (Ratté and Wayland 1969; Hill et al. 2004). The difference in foliation orientation is especially sharp across the Keystone fault. Yet, outside of the HPG block, contact metamorphism is still quite obvious. If magma intrusion had occurred at the current level relative to the bounding blocks, then one would expect that its compressional effects would have reached across the bounding faults together with heat and fluid flow. One possible explanation is that the HPG block was initially heated and deformed by intrusion of magma at a depth greater than its current position relative to its bounding blocks. The block was then uplifted relative to the neighboring crustal blocks because it was more buoyant due to the presence of melt, and because it was hot, it caused contact metamorphism on the bounding blocks. Movement of the major faults in the Black Hills after intrusion of melts and during metamorphism

1485

is evident in precipitation of graphite along some of these faults and prevalence of secondary fluid inclusions within pegmatites that occur near the faults (Sirbescu and Nabelek 2003). Others (e.g. Redden et al. 1990) have previously referred to the granite and its proximal area as the “Harney Peak dome,” but we prefer the term “block” if it was uplifted as a more-or-less coherent block along faults instead of by folding. It is not clear, however, how deep the HPG block was at the time of granite emplacement. Friberg et al. (1996) noted that metamorphosed mafic inliers within the HPG itself preserve elevated P-T conditions of ~6.3 kbar and 775 °C, though they did not provide data. These conditions also suggest a fairly high-pressure origin for the HPG block. It is quite possible that the extensive resetting of mineral assemblages in the HPG block wiped out most evidence for its high-pressure history. The Thompson Draw synform (see Fig. 1) has been interpreted as a downwarping of relatively colder rocks by Friberg et al. (1996) based on the occurrence of old, inclusion-filled garnets and by Holm et al. (1997) based on 40Ar/39Ar ages of micas. Micas in the synform gave 1600–1655 Ma cooling ages when in the HPG and BM regions give