INTRODUCTION

1 The Lake George mine is located approximately 40 km west of Fredericton, New Brunswick (Fig. 1a) and was formerly the largest producer of antimony in North America. Interest in the area has recently been revived with the discovery of gold anomalies in drill cores (Morrissy 1991; Lentz et al. 2002a). These anomalies are thought to be associated with the nearby Lake George granodiorite stock (Seal et al. 1988), which is situated 500 m northwest of the Hibbard Shaft and 350 m below surface. A strongly fractured contact aureole around the stock likely provided conduits for mineralizing fluids bearing Sb, W, Mo, Au, and base metals. Tightly folded calcareous turbiditic rocks, metamorphosed to greenschist facies (Seal et al. 1987, 1988; Caron 1996) of the Silurian Kingsclear Group form the country rock around the Lake George granodiorite, which lies approximately 350 m below the surface along the southeastern margin of the Pokiok Batholith. A contact aureole surrounding the intrusion is indicated by the presence of biotite and cordierite porphyroblasts in the pelitic rocks (Seal et al. 1987; 1988; Caron 1996).

2 Prior to the intrusion of the granodiorite stock, steeply dipping east-trending lamprophyre dykes were discordantly emplaced into the Silurian Kingsclear Group. The focus of the present study is a northwest-trending quartz-feldspar porphyry (QFP) dyke which also intrudes the metasedimentary rocks, although its relationship to the granodioritic stock is uncertian (Morrissy and Ruitenberg 1980; Lentz et al. 2002b). The QFP dyke has a north-northwest orientation (Procyshyn and Morrissy 1990), dips steeply to the west (80°), and is 3 to 5 m wide (Seal et al. 1988). The dyke bifurcates with a separation of less than 25 m between the branches, which merge along strike and at depth. The QFP dyke crosscuts narrow east-west lamprophyre dykes (Lentz et al. 2002b) but is inferred to be intruded by the Lake George quartz-hornblende-plagioclasephyric granodiorite stock at depth (Fig. 1).

3 The Pokiok Batholith (Fig. 1a) has been subdivided into five units (Ruitenberg and Fyffe 1982; Lutes 1987; Whalen 1993, Whalen et al. 1996). The Hartfield tonalite is the oldest unit (U-Pb titanite, 415 ± 1 Ma), followed by the Skiff Lake granite (U-Pb zircon, 409 ± 2 Ma) and the Hawkshaw granite (U-Pb titanite, 411 ± 1 Ma) (Bevier and Whalen 1990a, b; Whalen 1993). The Lake George granodiorite stock (U-Pb zircon 412 +5/-4 Ma; McLeod et al. 2003) is chemically and petrologically similar to the Hawkshaw granite of the Pokiok Batholith, although there are some mineralogical differences (Yang et al. 2002a, b). The Allandale granite is the youngest unit (U-Pb monazite, 402 ± 1 Ma) of the Pokiok Batholith. It is possible that the Hartfield, Skiff Lake, and Hawkshaw units represent a sequential fractionation and emplacement series, although the age of the Allandale granite is too young to be related to the main intrusive episodes of the batholith.

4 The purpose of this study is to examine the QFP dyke in the area and to determine its relationship to the magmatic activity in the region and timing relative to the two episodes of mineralization (W-Mo-Au and Sb-Au). Previous reports suggested that the dyke predates the granodioritic intrusion (Morrissy and Ruitenberg 1980), which is consistent with it being locally crosscut and altered by both the Hibbard Sb vein and the intrusion-related gold-bearing quartz-scheelite-molybdenite veinlets (Seal et al. 1988), although crosscutting relationships between the dyke and the stock were not observed in this study. As the granodiorite stock is linked to the exocontact W-Mo-Au mineralization (Lentz et al. 2002a), information on the QFP dyke could enhance exploration efforts for Au (W-Mo) in the area, because (1) it could indicate the presence of larger granodiorite intrusions at depth and of magmatic systems of similar petrologic association, (2) intrinsic metal abundances may reveal the presence of a favourable magmatic hydrothermal system, and (3) aspects related to dyke emplacement indirectly provide evidence of the role of volatiles in emplacement that indirectly controls the distribution of exocontact Au-W-Mo mineralization in the region.

DYKE PETROLOGY

5 The rock-forming minerals in the QFP dyke at Lake George were first examined in polished drill-hole cores samples (Fig. 2) to observe variations in texture and mineralogy. [Note: Sample numbers indicate diamond drill hole number and depth. For simplicity, the m (for metres) is omitted from each sample number.] The metasedimentary host rocks in contact with the dyke have preserved no macroscopic evidence of contact hornfelsing associated with dyke emplacement. Samples from the host rocks contain 30 to 60 % anhedral quartz and subhedral and anhedral feldspar phenocrysts ranging from 5 to 12 mm in length (Fig. 2a). The dyke mineralogy generally consists of porphyritic and fine-grained quartz, plagioclase, and orthoclase with accessory biotite (< 2 mm in length), chlorite, muscovite, pyrite, pyrrhotite (< 3 %), chalcopyrite, calcite, titanite, and rutile. Ilmenite, zircon, monazite, and apatite are present but recognizable only at the sub-microscopic scale through SEM-EDS analysis. The samples differ in their degree of sericitization. Based on observed abundance of sericite and chlorite, sample DDH 81-25-19.2 is the least altered and sample DDH 79-6-195.5 is most altered. Sample DDH 81-21-181.1 (Fig. 2b) has a prominent fracture running the length of the core with strong sericitization. Sample DDH 81-26-63.4 (Fig. 2c) is interesting in that it contains a fragment that has a dark homogeneous microcrystalline matrix, morphologically and compositionally unlike the sedimentary host rocks (see Scratch et al. 1984). Sample DDH 79-6-210 has dark rounded xenoliths up to 10 mm in size (10 to 15 percent) (Fig. 2d) (see also DDH 81-26-63.4). These xenoliths resemble autobrecciated altered intrusive rocks, consistent with their similar compositions (see below).

6 Rounded quartz occurs generally in two size populations (see Fig. 3a); phenocrysts and phenoclasts (ave 0.7 mm) with a groundmass consisting of grains <0.1 mm in size. They measure up to 6 mm, although the average is about 0.7 mm. The larger grains commonly contain melt inclusions and secondary fluid inclusions, and appear to have been a nucleation point for crystal growth (see Fig. 3b).

Display large image of Figure 1

Display large image of Figure 2

Display large image of Figure 3

7 Feldspars are present in high proportions in every sample, but typically have undergone extensive alteration to sericite. Plagioclase crystals show various degrees of sericitization, but some display relict zoning and twinning (Fig. 4; Table 1). The zoning in most of the plagioclase crystals is irregular and the edges of some grains are rounded, which is evidence for periods of both crystal growth and resorption, and (or) possible abrasion. Plagioclase has inclusions of potassium feldspar, apatite, zircon, chlorite, Fe sulphides, and rutile. Granophyric (more wormy) to graphic (more cuneiform) intergrowths between feldspar and quartz are evident in three samples (DDH 79-6-174.3, DDH 81-26-63.4, and DDH83-1-37.5; Fig. 5). Chlorite appears to be a local alteration product of reddish-brown biotite (Fig. 6). Primary biotite crystals (4.4 wt. % TiO2 by SEM-EDS analysis) contain inclusions of zircon, apatite, pyrite, albite, interlayered muscovite, and locally trellis-like rutile needles.

8 Sample magnetic susceptibility ranges from 1.1 to 13.5 × 10-4 (S.I. units) (see Table 2), typical of transitional reduced to oxidized magnetite-type magma (see Ishihara 1981), although most samples contain magnetic pyrrhotite rather than magnetite. Pyrrhotite and chalcopyrite are present in most samples, although probably a high temperature alteration feature, as they coexist with alteration assemblages (as discussed later) and with biotite phenocrysts.

Display large image of Figure 4

Table 1. SEM-EDS analyses of transects of zoned plagioclase grains from the QFP dyke at the Lake George mine

Display large image of Table 1

9 Euhedral titanite occurs as crystals less than 0.5 mm across, consistent with a relatively oxidized igneous system, although inconsistent with the reduced nature of the Lake George granodiorite. Apatite measures less than 0.1 mm in length and zircon is typically usually less than 100 microns, both as inclusions in biotite and plagioclase. As zircon and apatite apparently began to crystallize early, they may not be reliable thermometric indicators of dyke emplacement, but rather of maximum magma temperatures (Harrison and Watson 1984; Hanchar and Watson 2004).

10 Chlorite is present along fractures associated with calcite and, in some instances, contains intergrown bands of calcite. Narrow calcite veins (less than 0.2 mm wide) occupy micro-fractures, suggesting late paragenesis. Individual calcite grains contain pyrite inclusions. Some also contain minor pyrrhotite (<2 %) and chalcopyrite, typically tens of microns to 0.3 mm in size. Muscovite is commonly interstitial to larger plagioclase grains replacing K-feldspar phenocrysts; muscovite rosettes are common in altered parts of the dyke.

Display large image of Figure 5

Display large image of Figure 6

11

GEOCHRONOLOGY

12 Five single or double zircon fractions were analyzed (Table 3). On a concordia diagram (Fig. 7), they plot in a sub-concordant position (94 to 99.4 % concordant) with the exception of grains Zr3, which are only 74 % concordant with an apparent 207Pb/206Pb date of 757 Ma (see Table 3). The four remaining grains define a well-constrained upper intercept date of 420.8 +5.9/-4.0 Ma (MSWD=2.9) with a lower intercept date of 59 +85/-87 Ma. The date of 421 Ma is considered to be the age of intrusion of the QFP dyke. This age is older than the Lake George granodioritic stock (412 +5/-4 Ma; McLeod et al. 2003), although the ages are the same within error. The apparent date of 757 Ma from the Zr3 grains may be attributed to the presence of xenocrystic zircon incorporated from the country rock during dyke emplacement.

GEOCHEMISTRY

13 Whole-rock major-, trace-, and rare-earth-element (REEs) analyses of the QFP dyke are presented in Table 2. The Lake George granodiorite and dyke fall just within the low SiO2 suite of Devonian intrusions described by Whalen et al. (1996). ACNK [molar Al2O3/(CaO+K2O+Na2O)] averages 1.09 for the least-altered samples. Normative mineralogy calculated from the whole-rock analysis and plotted on a QAP diagram (LeMaitre et al. 1989) places seven of the ten samples in the granodiorite field, and two in the granite field. The normative average of these nine samples is quartz 32.3 %, corundum 2.66 %, orthoclase 18.95 %, albite 22.32 %, anorthite 12.2 %, hypersthene 7.06 %, magnetite 1.38 %, ilmenite 1.03 %, and apatite 0.28 %. The tenth sample, DDH 79-6-210, is unusual in that it plots as in the quartz-rich granitoid field and has much lower CaO and Na2O contents (0.41 and 0.22 %, respectively), suggesting that the sample has been strongly altered; it contains altered xenoliths of similar composition (see also sample 81-26-63.4).

Table 2. Major-, minor- and trace-element analyses of the QFP dyke at the Lake George mine.

Display large image of Table 2

14

15 The Nb versus Y and the Rb versus Y+Nb discrimination diagrams (Figs. 9a, b) (Pearce et al. 1984) show that the dyke has a volcanic-arc, I-type granite affinity, similar to the Lake George granodiorite stock (Yang et al. 2002a). Normalized to primitive mantle and chondrite compositions (Fig. 10a, b), the dyke shows negative Ba (ave. 404 ppm), Nb (ave. 16 ppm), TiO2 (ave. 0.54 wt. %), and Sr (ave. 186 ppm) anomalies.

16 Multi-element (spider) and chondrite-normalized REE patterns for the QFP dyke and proximal plutonic units (Fig. 10a, b) are similar with notable negative Ba, Nb, P, and weak Ti anomalies. The spider diagrams also show that sample DDH-81-21-181.1 (fractured) and sample DDH 81-26-63.4 (dark glassy matrix) are similar in composition to the other samples despite textural differences. The QFP dyke most closely resembles the Lake George granodiorite and the Hawkshaw granite (Fig. 11a, b). The La/Yb ratios for these units are also similar (13.0, 14.4, 15.2, respectively), whereas the ratios for the Hartfield tonalite (16.5) and the Skiff Lake granite (14.6) are relatively higher. The lower LREE and HREE abundances in the QFP dyke, granodiorite stock, and Hawkshaw granite, relative to more primitive phases of the Pokiok Batholith, suggest low temperature fractionation process (e.g., Miller and Mittlefehldt 1982).

17 The Rb, Ba, and Sr ternary diagram (Fig. 12) shows the differentiation trend in the Pokiok Batholith. Based on whole-rock compositions, Sr content should decrease, whereas Ba increases within a differentiating intermediate to felsic system (el Bouseily and el Sokkary 1975). The Rb-Ba-Sr diagram shows that the Hartfield tonalite is the most primitive (oldest) and the Allandale granite is the most differentiated unit of the batholith, although the latter unit likely represents a separate magmatic event, based on its much younger age. However, this diagram displays some scatter for the Lake George granodiorite and QFP dyke samples, attributed to alteration.

18 Although zircon saturation thermometry was not used due to the possibility of inheritance, it is notable that the Hartfield tonalite and the Skiff Lake and Hawkshaw granite units exhibit decreasing Zr and TiO2 (Fig. 13), consistent with fractional crystallization. This variation is consistent with the saturation behaviour of zircon in crystallizing granitic magmas (Hanchar and Watson 2004), and the decreasing REE contents noted above.

19 The average S content determining using pressed pellet XRF analysis is 3250 ± 1450 ppm; compared to the results from the fused bead method, the data indicate that S is dominantly tied up in sulphide minerals. The average δ34S for the bulk S contents of the QFP dykes is 1.0 ± 1.1 ‰ (Table 2); this narrow range near typical mantle values in a reduced I-type intrusion (Ohmoto and Goldhaber 1997) is consistent with a high-T magmatic origin for the sulphide minerals, with negligible fractionation evident (see detailed discussion in Yang 2005).

20 Au (<2 to 46, average 20 ppb), W (<1 to 3 ppm), and Mo (<1 to 11, average 7 ppm) contents in the QFP dyke are substantially lower than averages for the Lake George granodiorite (32 ppb, 9, and 7 ppm, respectively; Yang et al. 2002a). The average Cu abundance (112 ± 42 ppm) in the QFP dyke is notably higher than in the Lake George granodiorite (average 41 ± 38 ppm). S abundance co-varies with both Au and Cu (Fig. 14), supporting a deuteric origin for the S and metal enrichment; these data are consistent with the observations of Stirling (1979) of S enrichment in the most quenched contact zones of a similar dyke. Average Sb (80 ± 193 ppm) and As (124 ± 215 ppm) contents are significantly higher in the dyke than in the Lake George granodiorite, indicative of alteration, but they are unrelated to S, Cu, and Au abundance based on their low correlation coefficients (<0.3). However, all of these values are above crustal averages (Wedepohl 1995) and point to enrichment associated with magmatic volatiles related to emplacement of the QFP dyke.

GEOTHERMOMETRY

21 In terms of magma geothermometry, apatite saturation temperature (AST) calculated using P2O5 abundance and compositional corrections was estimated for the Lake George granodiorite to be 902 ± 14°C. This temperature is interpreted to represent a melt temperature with apatite as a liquidus phase (Yang et al. 2002b). Similarly, apatite also occurs as a liquidus phase in the QFP dyke (this study) and in units of the Pokiok Batholith (i.e., Hartfield, Skiff Lake, Hawkshaw, and Allandale granitoids) (Lutes 1987; Whalen 1993; Yang et al. 2002a, b). Thus, we calculated liquidus temperatures for these granitoids using AST (Harrison and Watson 1984) information which may help constrain their genetic relations. Average AST values for the QFP dyke, Hartfield, Skiff Lake, Hawkshaw, and Allandale granitoids are 902°, 909°, 955°, 958°, and 982°C, respectively. The similarity in AST between the dyke and the units of the batholith suggests that they may be genetically related, consistent with compositional similarities. AST decreases from the Hawkshaw granite (958°C) to the QFP dyke and stock (902°C), consistent with fractional crystallization (cf. Yang et al. 2002a).

Table 3. U-Pb data for the QFP dyke (DDH 81-21) at the Lake George antimony mine, New Brunswick.

Display large image of Table 3

Display large image of Figure 7

Display large image of Figure 8

Display large image of Figure 9

DISCUSSION

CONCLUSIONS

32 Although seemingly simple, the QFP dyke has numerous features that help to constrain the emplacement of the magma and understand the role of such subvolcanic bodies in forming the mineralization in the Lake George area.

• 1) The QFP dyke yielded an age of 420.8 +5.9/-4.0 Ma by U-Pb zircon geochronology. Therefore, it is likely genetically related to units of similar age in the Pokiok Batholith, and possibly the Lake George granodiorite stock.
• 2) The QFP dyke has petrochemical features of a reduced I-type granodiorite, like the Lake George granodiorite stock. Therefore, it is probably a fractionation product of a deeper level unit of the Pokiok Batholith.
• 3) The presence of two feldspar phenocryst phases supports the existence of a deep level (> 3 kb, 300 MPa) magma chamber feeding this dyke. Oscillatory zoning in rounded plagioclase phenocrysts/phenoclasts reveals a dynamic system, which caused melting/resorption. This system was likely caused by several episodes of magma mixing or the introduction of fresh magma pulses from the source.
• 4) Amphibole, which is present in most of the other Pokiok granitoid units, reacted with the fractionated melt to form biotite prior to dyke emplacement. This change explains the similar REE patterns and abundances throughout the Pokiok Batholith, even in the absence of amphibole.
• 5) Autobrecciation along the margins of the dyke, sub-rounded fragments, subhedral to rounded and fragmented plagioclase and quartz phenocrysts (phenoclasts), and the autometasomatic effects of deuteric fluids support fluidization (vapour-magma emulsion) as a key emplacement mechanism for the dyke, i.e., analogous to tuffisite. Thermal modelling coupled with integration of the estimated magma properties affecting rheology of this particular system is consistent with these observations.
• 6) The presence of pyrrhotite (> 10 times sulphur saturation levels), which accounts for the range in magnetic susceptibilities, with chalcopyrite (with minor gold) in the QFP dyke was probably produced by high-T subsolidus reactions of deuteric fluids with the quenched walls of the dyke as it was emplaced; this alteration is manifested as alteration in the microcrystal-line groundmass. The sulphur isotopic signature (δ34S = 1.0 ± 1.0 ‰) and S abundance (3250 ± 1450 ppm) indicate that magmatic volatiles were directly involved during dyke emplacement (autometasomatic) and by inference that Cu and Au were associated with these volatile components.

APPENDIX: ANALYTICAL METHODS