Harney Peak Leucogranite, Black Hills, South Dakota, USA: Petrologic and geochemical constraints. P.I. Nabelek L/, C. Russ-Nabelek 1'3, and J.R. Denison 4.
Contrib Mineral Petrol (1992) 110:173-191
Contributions to Mineralogy and Petrology 9 Springer-Verlag1992
The generation and crystallization conditions of the Proterozoic Harney Peak Leucogranite, Black Hills, South Dakota, USA: Petrologic and geochemical constraints P.I. Nabelek L/, C. Russ-Nabelek 1'3, and J.R. Denison 4 1Department of Geological Sciences,Universityof Missouri, Columbia, MO 65211, USA 2Centre de Recherche sur la Synth~seet la Chimie des Min6reaux, CNRS 45071 Orleans Cedex 2, France 3EnvironmentalTrace Substances Research Center, Universityof Missouri Columbia, MO 65211 4Missouri University Research Reactor, Universityof Missouri, Columbia, MO 65211, USA Received April 8, 1991/Accepted August 6, 1991
Abstract. The mineralogy, petrology and geochemistry of the Proterozoic Harney Peak Granite, Black Hills, South Dakota, were examined in view of experimentally determined phase equilibria applicable to granitic systems in order to place constraints on the progenesis of peraluminous leucogranites and commonly associated rare-element pegmatites. The granite was emplaced at 3~4 kbar as multiple sills and dikes into quartz-mica schists at the culmination of a regional high-temperature, low-pressure metamorphic event. Principally along the periphery of the main pluton and in satellite intrusions, the sills segregated into granite-pegmatite couplets. The major minerals include quartz, K-feldspar, sodic plagioclase and muscovite. Biotite-{Mg No. [Molar M g O / ( M g O + F e O ) ] = 0.32-0.38} is the predominant ferromagnesian mineral in the granite's core, whereas at the periphery of the main pluton and in the satellite intrusions tourmaline (Mg No. = 0.18-0.48) is the dominant ferromagnesian phase. A1mandine-spessartine garnet is also found in the outer intrusions. There is virtually a complete overlap in the wide concentration ranges of SiO 2, CaO, MgO, FeO, Sr, Zr, W of the biotite- and tourmaline-bearing granite suites with no discernable differentiation trends on Harker diagrams, precluding the derivation of one suite from the other by differentiation following emplacement. This is consistent with the oxygen isotope compositions which are 11.5 + 0.6%0 for the biotite granites and 13.2 _ 0.8%0 for the tourmaline granites, suggesting derivation from different sources. The concentrations of TiO2 and possibly Ba are higher and of MnO and B are lower in the biotite granites. The normative Orthoclase/Albite ratio is extremely variable ranging from 0.26 to 1.65 in the biotite granites to 0.01 1.75 in the tourmaline granites. Very few sample compositions fall near the high-pressure, watersaturated haplogranite minima-eutectic trend, indicating that the granites for the most part are not minimum melts generated under conditions with aH2o = 1. Instead, most biotite granites are more potassic than the water-saturated minima and eutectics and in analogy with experimentally produced granitic melts, they are best explained Offprint requests to:
P.I. Nabelek
by melting at ~ 6kbar, aH~o < 1 and temperatures 800 ~ Such high temperatures are also indicated by oxygen isotope equilibration among the constituent minerals (Nabelek et al. 1992). Several of the tourmaline granite samples contain virtually no K-feldspar and have oxygen isotope equilibration temperatures 716-775~ Therefore, they must represent high-temperature accumulations of liquidus minerals crystallized under equilibrium conditions from melts more sodic than the watersaturated haplogranite minima or during fractionation of intruded melts into granite-pegmatite couplets accompanied by volatile-aided differentiation of the alkali elements. The indicated high temperatures, aH~o < 1, the relatively high TiO 2 and Ba concentrations and the relatively low ~?lsO values of the biotite granites suggest that they were generated by high-extent, biotite-dehydration melting of an immature Archean metasedimentary source. The ascent of the hot melts may have triggered low-extent, muscovite-dehydration melting of schists higher in the crust producing the high-B, low-Ti melts comprising the periphery of the main pluton and the satellite intrusions. Alternatively, the different granite types may be the result of melting of a vertical section of the crust in response to the ascent of a thermal pulse, with the low-~?lso biotite granites generated at a deeper, hotter region and the high~71~O tourmaline granites at a higher, cooler region of the crust. The low-Ti and high-B concentrations in the high~?lsO melts resulted in the crystallization of tourmaline rather than biotite, which promoted the observed differentiation of the melts into the granitic and pegmatitic layers found along the periphery of the main pluton and the satellite intrusions.
Introduction Peraluminous leucogranites constitute an important group of magmas derived by melting of the continental crust. They include many Hercynian granites of western Europe and the Himalayan leucogranite suite (Le Fort
174
et al. 1987). In North America, an important example is the Proterozoic Harney Peak Granite, Black Hills, South Dakota, the subject of this paper. Peraluminous leucogranites have many characteristics which make the understanding of their petrogenesis still somewhat controversial. The commonly observed variability in radiogenic and stable isotopes in individual plutons (Riley 1970; Walker et al. 1986; Le Fort et al. 1987; France-Lanord et al. 1988; Nabelek et al. 1992) indicates that many are generated from heterogeneous sources. Elevated STSr/86Sr and negative snd values of most leucogranites suggest that the sources had long crustal histories (Vidal et al. 1982; Miller 1985; Walker et al. 1986; France-Lanord and Le Fort 1988). In addition to the isotopic heterogeneities, the proportions of normative Q(quartz)-Ab(albite)-Or(orthoclase) are often more variable than in other granite types and the texture commonly ranges from aplitic to pegmatitic. In many instances, large rare-element pegmatite fields are associated with leucogranitic plutons (e.g., the Black Hills, the Winnipeg River district, the Yellowknife field), and while the petrogenesis of the pegmatites is still debated, extensive post-emplacement differentiation of the leucogranites has often been invoked to explain the pegmatites and the textural, mineralogic and major-element heterogeneities of the leucogranites (e.g., Cern) and Meintzer 1988; Shearer et al. 1987a; Walker et al. 1989). In order to place constraints on the conditions of partial melting and magma emplacement and the relative
The Harney Peak Granite cores the southern part of the Laramide Black Hills dome (Fig. 1). The intrusion of the granite and the related pegmatites has been dated at 1.72 Ga (Riley 1970; Redden et al. 1990). Its emplacement culminated a 2.5-1.7 Ga history of deformation and sillimanite-grade, 3-4 kbar regional metamorphism of late Archean and early Proterozoic metasedimentary rocks presently exposed around and within the granite (Redden et al. 1985; DeWitt et al. 1986; Helms and Labotka 1991a). The metasediments are
granitic metasediments rocks O biotite Ot biotite in granite tourm, in pegmatite
":42;X7
9 tourmaline
Oi3,
?]----~-46, ~7"-"-"-".71
Geology of the Harney Peak Granite
V---1
Elkhorn
3 km
importance of post-emplacement differentiation processes and source parameters on the geochemistry and mineralogy of peraluminous leucogranites, the mineralogy, petrology and geochemistry of the Harney Peak Granite are examined here in view of the available experimental phase equilibria studies on boron-flee and boron-containing granitic systems. In addition, the factors controlling the stability of the borosilicate tourmaline and biotite, both of which are common in the Harney Peak Granite as in other peraluminous leucogranites, are considered. This paper builds upon previous studies of the Harney Peak Granite, including the radiogenic isotope work of Walker et al. (1986), the stable isotope work of Nabelek et al. (1992), the morphologic and structural descriptions of Redden et al. (1985), Duke et al. (1988, 1990), the geochemical study of Shearer et al. (1987a), and additional studies of the southern Black Hills geology.
408
::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::: "Mount ain~'-:v-"-"-"-"-'C'-"-,iJ 74~ -~ "-:--:_.:_-=--=C'::.:':-."::-':"-':.'::
"."
81s- I
06,,2o ? ~28 P,728
24,
0
O 22
580 0 023
025,
580
~36 O, 677
31, ~
Calamit Peak South Dakota
Fig. 1. Geologic sketch map of the Harney Peak Granite (after DeWitt et al. 1986) showing sample locations with the dominant ferromagnesian mineral noted. The heavy lines separate low-$lsO samples in the middle of the granite from high-c~180 samples at the
periphery of the main pluton and in the satellite intrusions. Oxygen isotope equilibration temperatures among minerals in the samples are shown in italics
175 dominantly quartz-muscovite and quartz-biotite-garnet schists. The intrusion of the granite resulted in a contact aureole which overprinted the regional isograds. Near the southern part of the granite the rocks were metamorphosed to the second-sillimanite isograd. Occurrences of anatectic rocks have been found in the Willow Creak area (Shearer et al. 1987b). The Harney Peak Granite consists of a main pluton, numerous satellite intrusions, and isolated sills (Fig. 1). The main pluton and the satellite intrusions have domal structures consisting of hundreds of sills. Structural studies indicate that the domal intrusions grew laterally by the continual emplacement of new sills along the perimeter (Duke et al. 1988; Duke et al. 1990). Many sills have separated into granite-pegmatite couplets with pegmatites generally along the hanging walls of the sills (Rockhold et al. 1987; Duke et al. 1988). The layers range from < 1 m to several meters thick. The pegmatitic layers are less common (generally < 20% of the igneous rock mass) in the core of the main pluton where the rocks are fairly equigranular but become much more common along its flanks and in the satellite intrusions. In some satellite intrusions, such as the Calamity Peak or Sky Lode, many granite layers exhibit prominent millimeter- to centimeter-scale rhythmic layering of minerals which has also been called "line-rock" (Kupfer 1963; Rockhold et al. 1987; Duke et al. 1988). Fracturefilling pegmatitic dikes are also common in the satellite intrusions. The dominant minerals in the Harney Peak Granite are quartz, sodic plagioclase (albite to oligoclase), orthoclase or microcline and muscovite. The proportion of plagioclase to K-feldspar is variable, especially along the perimeter of the main pluton and in the satellite intrusions. Significant is the mode of occurrence of the ferromagnesJan minerals. Biotite is the predominant ferromagnesian mineral in the core of the main pluton whereas tourmaline predominates along the perimeter and in the satellite intrusions. The two minerals are mutually exclusive; only in very few samples are they found together. In the pegmatitic layers in the outer parts of the Harney Peak Granite, tourmaline is found with centimeter-size crystals of plagioclase and quartz usually along the foot walls of the layers. The middle and hanging walls of the layers are usually dominated by K-feldspar. Almandine-spessartine garnet is found sporadically but principally in the tourmaline granites.
Previous geochemical studies The Harney Peak Granite has been of great interest to geologists for a long time due to its association with rare-element pegmatites. However, most studies focused on the large rare-element pegmatites and it is only recently that petrogenetic interpretations based on geochemical data of the Harney Peak Granite itself have been made. Walker et al. (1986) used primarily Nd and Sr isotope data to show that the metasediments surrounding the Harney Peak Granite could not have been its only source. For the Harney Peak Granite they obtained aNdlT~Svalues between - 9.90 and - 2.05 with an average of - 6 . 0 8 , whereas for the schists surrounding the granite they obtained an a~]15 range of - 3.03 to - 1.90. They proposed that the source must have included a long-lived Archean precursor, such as the one for the Archean Little Elk Granite found in the north1715 value of eastern Black Hills, which would have had an ~r~a - 14.19. The oxygen isotope composition corroborates the evidence for multiple sources for the Harney Peak Granite. Nabelek et al. (1992) found that the average g~sO value of the biotite-containing granites in the core of the main pluton is 11.5 + 0.6%0and in the tourmalinecontaining perimeter of the main pluton and the outer intrusions is 13.2 _ 0.8%0. A subset of the samples analyzed for stable isotopes was analyzed for major and some trace elements (Table 1) and is discussed here. The average ~3~80 values of the subset biotite and tourmaline granites are only slightly different (Fig. 2; Table 2). Inasmuch as Nabelek et al. (1992) have shown that the difference in the isotopic compositions of the biotite and tourmaline granites cannot be accounted for by magmatic differentiation or subsolidus interaction with externally derived fluids (for which there is no evidence), or assimilation, they suggested that the low-~?~80 granites
in the core of the Harney Peak Granite were generated from isotopically different source rocks than the outer, high-01so granites. It was suggested that these granites may have been generated from rocks equivalent to the country rock schists, because the average isotopic composition and variation in the outer granites is the same as that in the schist country rocks. Shearer et al. (1987a) suggested, principally on the basis of lower K20/Rb ratios ( ~ 100) and higher Li, Cs and Nb in whole rocks and lower K20/Rb ratios in K-feldspars in the outer, tourmalinerich parts of the Harney Peak Granite, that 60-80% fractional crystallization of a quartz-plagioclase-K-feldspar-muscovite assemblage from the biotite-rich granite (whole rock K20/Rb ~ 200) can explain the chemical variation between the biotite and tourmaline granites. Although on the basis of trace-element modeling they eliminated different degrees of partial melting of a source similar to the surrounding schist as a mechanism to explain the chemical variation, they did not consider the possibility that the inner and outer parts of the Harney Peak Granite were generated from multiple sources.
Analytical techniques The large variation in the grain size of the Harney Peak Granite, particularly in its outer parts, made the choice of representative samples for chemical analysis difficult. Obtaining a representative sample of the pegmatitic layers is virtually impossible and therefore no whole-rock chemical analyses of the pegmatites were made. Special care was taken with the granite samples to ensure as much as possible that the size of each sample was representative of the granite at the sample location. At many locations several samples were collected. The samples were prepared for chemical analysis by crushing them to fine powder in a SPEX ball-mill or a shatter-box. Boron (Table 1) was analyzed by Prompt-Gamma Neutron Activation Analysis at the 10 MW University of Missouri Research Reactor using a thermal neutron flux of 5 x 108 n cm-2 s-1 following the method of Hanna et al. (1981). The details of the technique used in the present study are described in Nabelek et al. (1990). The 1-o reproducibility is better than 3% at the > 10 ppm level and 30-50% near the detection limit of 0.1 ppm. For all other elements (Table 1), duplicate aliquots of the powders were analyzed on the Jarrel-Ash Inductively-Coupled Plasma Atomic-Emission Spectrometer at the University of Missouri's Environmental Trace Substances Research Center. A 250-rag sample of rock powder was fused with 600 mg of lithium metaborate flux. The glass was subsequently dissolved in 10% HNO 3 solution and then diluted to 100 ml. U.S. Geological Survey standards were used for quality control. The accuracy and precision on all the reported elements are better than 2%. Water contents were determined by ignition. The micas, tourmaline and garnet were analyzed on Cameca SX50 microprobes at the University of Tennessee and the Oregon State University using an accelerating voltage of 15 kV and beam current of 0.015 laA. The raw data were reduced using the ZAF correction scheme. Garnet analysis were normalized to 12 oxygens whereas the mica compositions were normalized to 10 oxygens plus two (OH + F - + C1 ) groups. The tourmaline analyses were normalized following Jolliff et al. (1986) to six silicons and assuming three borons. The complete mineral compositions can be obtained from the senior author upon request.
Major and trace-element geochemistry T h e difference b e t w e e n the c3180 values of the i n n e r a n d o u t e r parts of the H a r n e y P e a k G r a n i t e p r o v i d e s a basis for s e p a r a t i n g the H a r n e y P e a k G r a n i t e s a m p l e s i n t o low~?lSo a n d h i g h - 0 t s o suites, respectively ( N a b e l e k et al. 1992; Fig. 2). This d i v i s i o n also effectively separates the b i o t i t e - b e a r i n g g r a n i t e suite from the t o u r m a l i n e - b e a r i n g
176
Table 1. Major element (wt %), trace-element concentrations (ppm), and temperatures
Sample
SiO 2
Low-0~80 granites 6B 73.4 7B 73.3 13A 74.4 13C 74.4 14A 72.9 20A 74.1 21C 75.1 22 75.2 23A 73.0 24B 73.1 25 74.8 26 72.3 27B 72.7 30A 69.8 31A 73.7 32 73.3 44A 72.7
TiO 2
A120 3
FeO
MnO
MgO
CaO
Na20
0.09 0.06 0.02 0.02 0.01 0.03 0.16 0.05 0.03 0.05 0.06 0.09 0.08 0.16 0.06 0.14 0.04
14.3 14.5 15.2 14.7 15.6 15.0 13.5 13.6 15.2 15.1 13.9 14.9 14.3 15.8 15.2 14.2 14.9
0.68 0.44 0.37 0.52 0.60 0.21 1.52 0.52 0.46 0.52 0.78 0.66 0.72 1.46 0.75 1.20 0.62
0.01 0.00 0.01 0.02 0.01 0.00 0.02 0.01 0.03 0.01 0.03 0.01 0.01 0.03 0.04 0.02 0.03
0. l 7 0.08 0.08 0.08 0.06 0.07 0.41 0.09 0.10 0.13 0.13 0.17 0.19 0.37 0.12 0.23 0.09
0.52 0.47 0.55 0.67 0.96 0.76 0.76 0.49 0.62 0.67 0.75 0.56 0.56 0.75 0.89 0.92 0.85
2.60 2.89 4.24 5.37 5.99 4.06 2.81 3.05 2.84 3.00 3.01 2.87 2.87 3.73 4.71 3.57 3.67
0.00 0.00 0.02 0.02 0.03 0.01 0.03 0.01 0.02 0.02 0.01 0.01 0.02 0.01 0.01 0.02 0.04 0.01 0.02
14.5 15.2 15.0 15.1 15.2 14.9 15.5 15.2 15.5 15.3 13.9 16.3 15.0 15.4 14.4 16.1 15.8 14.8 16.0
0.16 0.37 0.82 0.55 0.40 0.22 0.68 0.14 0,59 0.97 0.57 0.98 1.58 0.73 0.24 1.31 1.63 0.73 0.51
0.03 0.05 0.08 0.04 0.01 0.01 0.12 0.00 0.11 0.12 0.04 0.02 0.25 0.04 0.10 0.13 0.11 0.06 0.03
0.02 0.04 0.10 0.10 0.07 0.03 0.12 0.04 0.09 0.23 0.09 0.13 0.13 0.11 0.05 0.15 0.45 0.07 0.17
0.21 0.28 0.48 0.24 0.81 1.35 1.01 1.60 0.40 0.55 0.27 0.40 0.90 0.49 0.26 0.57 0.76 0.61 0.49
4.52 5.34 4.13 2.79 4.06 5.99 3.81 5.49 3.63 3.50 3.28 7.00 3.93 4.73 5.29 7.31 5.50 6.40 5.48
High-#180 granites 2A 3 10B 10D 16C 17 18A 19 34A 35A 38A 39A 43A 45B 46A 49A CPI-1 CP5-1B CP12-1A
73.8 73.9 74.6 72.5 74.6 75.4 72.9 75.3 71.9 71.6 73.9 72.2 72.6 74.2 77.0 72.4 72.5 75.3 73.1
granite suite, although some high-~?~aO granite samples (16C, 18A) contain biotite rather than tourmaline or both (43A). However, the pegmatites coexisting with the latter three samples do contain tourmaline. Some high-c?lSO granites contain practically no ferromagnesian mineral; the bulk of the Mg and Fe in the samples is located in the muscovite. The pertinent major and trace-element compositions of the granites are represented by Harker diagrams in Fig. 3. In both suites there is a large variation in virtually all the elements analyzed, emphasizing the heterogeneous nature of the H a r n e y Peak Granite. Moreover, there are no discernible trends of these elements with silica which could potentially indicate fractional crystallization. In order to determine whether the average concentration of a given oxide is different in the low-c?~ sO and high-c?lso granites, given the chemical heterogeneities of the granite suites it is useful to perform a statistical t-test (LeMaitre 1982). Given two populations of samples, the t-test determines whether the means of the populations are different at a given confidence level. There is virtually a complete overlap in SiO2, CaO, M g O , FeO, Sr, Zr, and W
concentrations, and at the 95% confidence level, the average concentrations in the two granite suites are statistically indistinguishable (Table 2). O n the other hand, the statistical analysis shows that at the 95% confidence level, TiO 2 is on the average lower while B and M n are higher in the high-c?lsO granites. The M g No. [molar M g O / ( M g O + FeO)], is higher in the low-c?lso suite at the 94% confidence level. The distribution of Ba is somewhat equivocal. Shearer et al. (1987a) suggested that Ba concentration is higher in the biotite granites than the tourmaline granites. O u r data shows this to be true only at the 86% confidence level. A c o m m o n characteristic of leucogranites is their high alumina saturation index [ASI; m o l a r A1203/(K20 + N a 2 0 + CaO)]. In the low-c?lao granites, the ASI ranges from 1.10 to 1.31 and in the high-c?laO granites from 1.08 to 1.42, emphasizing the peraluminous composition of both rock suites. While there is a considerable overlap in both suites of rocks, at the 95% confidence level the average ASI is slightly higher in the high-~?tsO granites.
177 Table 1. (continued) K20
P205
H20
Sum
B
Sr
Ba
W
Zr
T(~
6.56 6.55 3.60 2.55 2.21 4.06 3.96 5.36 5.91 5.70 4.86 6.61 6.77 6.26 2.86 4.86 5.55
0.09 0.20 0.09 0.12 0.12 0.24 0.18 0.17 0.32 0.14 0.12 0.22 0.20 0.07 0.12 0.00 0.00
0.54 0.53 0.73 0.34 0.31 0.77 0.96 0.54 0.74 0.93 0.64 0.85 0.60 0.91 0.65 0.57 0.56
99.0 99.0 99.2 98.8 98.8 99.2 99.4 99.0 99.2 99.4 99.1 99.3 99.1 99.3 99.2 99.0 99.0
9 8 8 9 4 12 6 5 6 6 7 7 7 5 5 10 7
88 28 13 13 14 34 35 42 32 33 39 54 32 40 14 75 45
322 77 14 10 8 125 69 68 55 84 71 116 52 108 16 686 95
61 74 70 98 93 105 97 61 96 57 76 105 83 98 88 58 70
42 22 9 6 28 21 63 44 17 42 39 23 24 18 26 113 76
520 • 38
4.91 3.02 3.01 6.98 3.15 0.43 4.03 0.58 5.75 6.01 6.38 1.03 3.99 2.39 0.92 0.11 0.80 0.51 3.21
0.30 0.30 0.19 0.15 0.11 0.16 0.19 0.20 0.44 0.14 0.00 0.28 0.07 0.29 0.18 0.22 n.d. n.d. n.d.
0.60 0.84 0.86 0.35 0.74 0.50 0.70 0.58 0.63 0.42 0.44 0.68 0.51 0.94 0.89 0.39 n.d. n.d. n.d.
99.1 99.3 99.3 98.8 99.2 99.0 99.1 99.1 99.0 98.8 98.9 99.0 99.0 99.4 99.4 98.7 97.6 98.5 99.0
15 16 38 385 22 31 n.d. 11 484 1,124 892 1,785 6 133 20 528 3,478 1,211 1,397
24 4 23 34 36 47 70 31 62 72 17 21 35 9 21 35 n.d. n.d. n.d.
11 2 35 155 58 12 116 17 12 236 56 12 39 9 17 18 n.d. n.d. n.d.
77 82 102 85 52 82 93 90 97 78 69 85 95 68 107 79 n.d. n.d. n.d.
11 19 18 18 17 0 21 43 22 39 20 89 59 23 23 28 n.d. n.d. n.d.
647 • 657 • 661 + 728 •
60 49 79 42
848 _ 41 580 _+ 17 580 • 47 561 • 81 677 • 53 809 • 2 815 • 76
662 • 42 408 4- 22 709 _+ 195 696 • 111 688 _+ 113 584 _+ 23 747 • 22 781 • 71 716 • 7 775 • 14
n.d., not determined The variation in the normative O r / A b ratio, and therefore the K 2 0 / N a 2 0 ratio, is very large in both granite suites (Table 2; Fig. 4). The average ratios of the low- and high-~? 180 granites significantly differ, being 1.07 and 0.58, respectively. However, the average ratio of the high-~?~80 granites must be interpreted with caution. First, pegmatitic layers comprise a large percentage of the mass of the perimeter of the main pluton and the satellite intrusions. The layers are generally more potassic than the coexisting granites (Rockhold et al. 1987; D u k e et al. 1991) and therefore, the overall composition of the high-c~80 suite is less sodic than suggested by present sampling. Second, the average O r / A b ratio of the high-~ 180 granites is skewed by seven samples lying next to the Q - A b join which have virtually no K-feldspar. These samples were associated in the field with pegmatitic layers and as will be shown below, they can be considered as accumulations of early-crystallizing phases. W i t h o u t these samples, the O r / A b ratios of the tourmaline and biotite granite suites become statistically indistinguishable and the overall variation is similar.
The distribution of tourmaline, and hence of B, in the high-c3180 granite suite is not correlated with N a 2 0 / K 2 0 as it is, for example, in the H i m a l a y a n BadrinathG a n g o t r i and M a n a s l u plutons where tourmaline generally is confined to the most sodic granites (Scaillet et al. 1990; F r a n c e - L a n o r d and Le F o r t 1988). At H a r n e y Peak, only samples with more than ~ 100 p p m B contain tourmaline. In other samples the bulk of B is found in muscovite (Rockhold et al. 1987).
Mineral compositions The compositions of minerals give important information about the phase equilibria and crystallization conditions of a cooling magma. F o r the purposes of this study we analyzed the major- and minor-element composition of biotite, tourmaline, muscovite and garnet. The results agree largely with those of Shearer et al. (1987a). O u r analyses indicated the biotite composition to be rather invariable with M g No. ranging between 0.32 and 0.38.
178
4t
avg. = 11.8%o
3
~2 &
1
~
10.0
11.0
12.0 3 1 8 0 (WR)
13.0
14.0
12.0 3 1 8 0 (WR)
13.0
14.0
7 6
B avg. = 13.1%o
~5
~4 ~3 2 1
10.0
11.0
Fig. 2A, B. Histograms of ~180 values of granites discussed in this paper: A biotite granites from the central parts of the Harney Peak Granite; B dominantly tourmaline granites from the periphery of the main pluton and satellite intrusions
The content of A1w ranges from 0.39 to 0.48 and TiP2 from 2.1 to 2.9wt % (Fig. 5A). In the high-~?~80 sample 43A which contains both tourmaline and biotite, the TiP2 concentration is lower at 1.2 wt %. Inasmuch as there is
no other titanium-bearing phase in the samples, biotite contains the bulk of titanium. The Mg No. of tourmaline is more variable ranging between 0.18 and 0.48 (Fig. 5B). There is no apparent systematic geographic correlation. However, the average Mg No. is approximately the same as that of biotite, most likely reflecting the similar Mg Nos. of the low- and high~3180 granites. Only some tourmalines have variable Mg/Fe but no systematic zoning is apparent. The elbaite component (calculated without Li which was not analyzed) varies between 21 and 26%. Garnet is generally confined to the most Mn-rich highc3180 granites. It is zoned in only some samples with the rims more Mn-rich. The compositions of garnet vary between PyrTAlmvsSps15 and PyriAlm4,~Sps55 (Fig. 5C). Muscovite is found in all Harney Peak samples. The proportion of paragonite is 3.6-7.5% and of celadonite 7.9-13.1%, assuming all Fe to be ferrous. The muscovite composition is more variable and on the average more Fe rich in the high-~?lSO granites (Fig. 5D, E). The average concentration of M n O is higher and of T i P 2 lower in the muscovite of the high-USO granites. These minor-element concentrations appear to reflect the relative amounts in the whole rocks. We did not analyze feldspars in this study. However, the analyses of Shearer et al. (1987a) and Rockhold et al. (1987) suggest that in the core of the Harney Peak Granite, the plagioclase is somewhat more calcic (up to An21 ) and more variable in composition than in the satellite intrusions. For example, in the Calamity Peak satellite intrusion, the composition is Ab91 _97Anl 8Or1-2 and is the same in the granitic and pegmatitic layers (Rockhold et al. 1987). The plagioclase is unzoned. Potassium feldspar occurs both as microcline and orthoclase. In the
Table 2. Results of t-tests
Element
Average Low-c~180
Average High-c~180
SD
SD
t-value"
t?J80 for all samples in Nabelek et al. (1992) 11.5 c3~80 in samples discussed here 11.8 SiO~ 73.4 TiP 2 0.068 AI20 3 14.69 FeO 0.71 MnO 0.017 MgO 0.15 CaP 0.69 Na2 O 3.61 KzO 4.96 P2Os 0.14
0.6 0.6 1.3 0.047 0.65 0.36 0.012 0.10 0.16 0.98 1.50 0.08
13.2 13.1 73.7 0,016 15.21 0.69 0.071 0.12 0.62 4.85 3.01 0.20
0.8 0.4 1.5 0.010 0.60 0.44 0.061 0.10 0.38 1.28 2.24 0A0
5.67 6.86 0.53 4.39 2.48 0.10 3.80 1.08 0.80 3.30 3.09 1.81
ASI Mg/(Mg + Fe) K20/Na20
0.07 0.06 0.68
1.24 0.22 0.78
0.09 0.07 0.72
2.38 1.99 3.22
B Sr Ba Zr W
1.18 0.27 1.53 7 37 116 36 82
"Critical t-value for 95% confidence level is 2.03
2 21 163 27 17
643 34 50 30 84
907 20 65 20 14
2.97 0.46 1.54 0.72 0.38
179 2.0
.
.
.
0 Iow-~180 1.5
9 high-a180
1.0 '
u
. granites
o
9149
~
mlilmOlffl~ i[~1 on
0.068
,0000
,
, 70
F .
,
, 72
,
,
0 ;
;
70
72
,
,
'
.
78
.
.
9
;
":
9
' ,
9149
1.3
' 70
68
I
o6:)o~
' 72
70
72
.
.
.
'~176
,
.
,
.
,
,
t
;
68
70
72
,
,
ii 0 dII
' 74
' 76
0
~ ~~
0.1
78
00 "68
0.3[
t
0
0.2~
o
78
' "
68
,
9
0
,
o ~
o
70
~~ 0 1 ~ ~ ~ 72 74 76
70
y9
__m
1"
;
68
70
72
,
,
78 .
3.0
9
~
2.5
~
2.0
~
0.0 78
68
Si02
Oo
;
,
II.
' 70
,
,
"
1
9
;
,
n
76 9
,
78 .
o
9
' 72
J l
) ;
74 9
111 -O~t~
"
0,
,
,
o
~ 1 0
m 76
o
-
;
,
,
~,~
'
~
, 74
o
I
9
,
I
78
oo
I
,
, 72
o,
" 76
.
[
lOO ~-
~
;
74 .
9
t
'
olO0~
Si02 Fig. 3. Silica variation (Harker) diagrams.
1
9
m9
I.
0.1
.
9
~ImI 9
'~176176
"I I ~
9
[
~*
m,~ 9
,
" 0 ~ml~)" ;
9
76
o
0 ;
9
o
74
'o'
9 9
8 0 [-
0.4[
ore~
1.0
.
~'~. 0 . 2
~_oO0 0o
" 68
76 '
o
oo o
20I '
,
0
.601
74 '
1.4
0
0.3
I
(3[~000~
1.1
,
0 O0 ~
.
1
,
0
0 10
,
76
9 9
'
,
.~. s, 0.05 r
,
. "! .
,0or 8
I 74
.
1000 I
10[
,
~". 0 . 1 5
~Oool~,
0.5
~1
0.20
granites
,
IIDI
I 74
,
I 76
9 ,
78
Si02
ASI, A1203/(CaO + N a 2 0 + K20)molar; Mg No., M g O / ( M g O + FeO)molar.
granitic layers, the composition is O r 9 6 _ 8 7 A b 4 _ 1 2 , whereas in the pegmatitic layers it can be as sodic as Orv4Ab2s (Rockhold et al. 1987; Shearer et al. 1987a; Duke et al. 1991).
calculated modes are given in Table 3. The residual sums of squares for all the analyses are low, indicating very good regressions. Intensive parameters
Mineral modes
Pressure and temperature
The availability of whole-rock and mineral chemical analyses permits the calculation of mineral modes. The concentration of each oxide, Ai, in the whole rock is given by: k
A i = >] XyBi,j
(1)
j=l
where Xj is the modal proportion of mineral j in the rock and B/d is the concentration of oxide i in mineral j. The proportion of each mineral j can be found using the method of least squares (Bryan et al. 1969). Zoning and exsolution occurs in the feldspars; therefore the proportions of the three individual end-members were determined. Likewise, apatite was assumed to be the hydroxy end-member. Otherwise, muscovite, biotite, and garnet compositions from sample 6B and tourmaline composition from sample 10D were used for all the low-0~So granites, whereas for all the high-01so granites, muscovite, tourmaline, and garnet compositions from sample 34A and biotite composition from sample 43A were used. The
The Harney Peak Granite lacks mineral assemblages which could be used to estimate the emplacement pressure. However, due to the presence of spodumene rather than petalite in several zoned Black Hills pegmatites, Norton and Redden (1990) suggested that the minimum pressure was 3.25 kbar if the pegmatites crystallized at 550 ~ This pressure is in accord with 2.5-3.7 kbar determined from mineral assemblages in the metamorphosed country rocks (Helms and Labotka 1991a). Nabelek et al. (1992) determined oxygen isotope equilibration temperatures among minerals for most samples discussed in this paper (Table 1; Fig. 1). The temperatures were determined by linear regression of c?lsO values of at least three mineral separates from each sample following the method of Javoy et al. (1970). In the low-~?lsO samples, the temperatures range from > 800 ~ to 520~580~ in closely spaced samples in the core of the pluton. Many high-c~lSO samples considered by Nabelek et al. (1992) and in this study, including those with little potassium, record temperatures in excess of 700 ~
180
Fugacity of fluorine and chlorine
\
~-
=
-
[]
~ ~ [i--i
_
,~
.....
s
_
"-, f-1
_
,
",,
D-r
e~
.u
~.
=
/
-
.
:. i,.
o -
~ , ,
" 800 ~ under
188 water-undersaturated conditions at pressures not much more than 5-6 kbar. The relatively high-Ti concentrations, but lack of a titanium phase, suggest the involvement of biotite during melting. Given these constraints, the generation of the 0180 melts probably involved the biotite dehydration-melting reaction: biotitess + aluminosilicate + plagioclas% + quartz --* garnet~ + liquid _+ K-feldspar
This continuous reaction has a steep dP/dT slope and for biotite Mg No. of 0.56 begins at approximately 760 ~ at pressures between 6 and 8 kbar (Fig. 10; Le Breton and Thompson 1988), Vielzeuf and Holloway (1988) reported biotite disappearance at > 850 ~ at 10 kbar during fluidabsent melting of metapelite with > 50% melt produced. For a system with an aluminosilicate, Vielzeuf and Montel (1991) found that the reaction biotit% + plagioclase~ + quartz ~ gametes + orthopyroxene + liquid + K-feldspar
5
[2]. ~4
L g~
[3]
occurs at approximately 70~ higher than reaction [2]. However, given the high normative corundum content of the granites it is likely that an aluminosilicate was present during melting. The exact source for the low-0*sO granites is not known but it could have been a deeper sedimentary sequence equivalent to the garnet-biotite schists, amphibolites, and marble which are found in the core of the Harney Peak Granite. The metasediments appear to have a sedimentary provenance different from the schists into which the satellite intrusions were emplaced (J. Redden, personal communication). These rocks may be equivalent to the base of the lower Proterozoic/Archean section found near the Bear Mountain Granite to the west of Harney Peak (J. Redden, personal communication). It is recalled that the Nd isotopic data show that at least one source component for the Harney Peak Granite must have been of Archean age (Walker et al. 1986). The phase equilibria and geochemistry of the highc~lSO suite require melting conditions of > 700 ~ variable water activities and possibly pressures < 5 kbar. The occurrence of migmatized schists within the Harney Peak Granite and in the Willow Creek locality some distance from the main pluton (Shearer et al. 1987b) attests to local melting even at the present level of exposure. The low concentrations of TiO2 in the high-~sO granites suggest lack of extensive melting of biotite. On the other hand, the relatively high concentrations of Rb and Li (Shearer et al. 1987a) indicate that generation of these granites involved incongruent dehydration melting of muscovite in the source rocks via reaction muscovit% + plagioclas% + quartz --* sillimanite + liquid ( • biotit% • K-feldspar) [4] (Thompson 1982). Shearer et al. (1987b) found evidence for this reaction in the migmatized schists at Willow Creek. The position of this reaction in P-Tspace is shown in Fig. 10. Below 5kbar, this reaction occurs below 700~ However, some F substitution in muscovite and Ca substitution in feldspar is likely to displace the reaction to a higher temperature. Vielzeuf and Holloway (1988) found that during melting of a metapelite at 10 kbar,
500
600
700 temperature (~
800
900
Fig. 10. Pressure-temperature diagram showing the water-saturated granite melting reactions: melting reaction involving muscovite (Thompson (1982); melting reaction involving biotite in the presence of an aluminosilicate mineral (Le Breton and Thompson 1988); melting reaction without the presence of an aluminosilicate mineral (Vielzeuf and Montel 1991) Superimposed are the approximate melting regimes for the low- and high-O180 Harney Peak melts as well as the geothermal gradients in the Harney Peak area before melting and after emplacement of the melts
reaction 4 occurs at approximately 750~ and that the proportion of melt is small. The difference in the reactants and products in reactions 2 and 3 as opposed to reaction 4 explains not only the difference between the TiO 2 concentrations in the two granites suites but the concentrations of other minor and trace elements, assuming similar concentrations in the parent metasediments. As shown by the reactions, garnet was likely to have been in the residue of the low
