U-Pb, Ar-Ar, and Re-Os Geochronological Constraints on Multiple Magmatic–Hydrothermal Episodes at the Lake George Mine, Central New Brunswick

The Lake George antimony mine was at one time North America’s largest producer of antimony. Despite being widely known for the antimony mineralization, the deposit also hosts a range of styles of mineralization such as multiple generations of W-Mo bearing quartz veins as well as a system of As-Au bearing quartz–carbonate veins. In situ U-Pb zircon geochronology, using LA ICP-MS, of the Lake George granodiorite yielded a weighted mean 206Pb/238U age of 419.6 ± 3.0 Ma. Step heating of phlogopite separated from the lamprophyre dykes produced a 40Ar/39Ar plateau segment date of 419.4 ± 1.4 Ma. Single molybdenite crystal analysis for Re-Os geochronology was conducted on two W-Mo-bearing quartz veins, which cross-cut altered granodiorite and altered metasedimentary rocks and yielded two dates of 415.7 ± 1.7 Ma and 416.1 ± 1.7 Ma respectively. 40Ar/39Ar geochronology of muscovite from alteration associated with Au-bearing quartz–carbonate veins yielded one representative plateau segment date of 414.1 ± 1.3 Ma. The dates produced in this study revealed that the different magmatic–hydrothermal events at the Lake George mine occurred over approximately a 10-million-year period at the end of the Silurian and the start of the Devonian following the termination of the Acadian orogeny.


Introduction
The Lake George antimony mine, located approximately 30 km west of Fredericton in west-central New Brunswick (Figure 1), was once North America's largest producer of antimony. In addition to antimony, the area surrounding the Lake George mine contains W, Mo, and Au mineralization; however, these commodities were never the focus of exploration or development.
Antimony was first discovered in the Lake George area during road construction in 1863; since then, the area has seen sporadic exploration and mining activity [1]. The Lake George Mining and Smelting company was formed in 1876 to operate the Hibbard mine. In 1869, the first shaft was sunk on the property followed by the construction of a mill and smelter by the Hibbard Antimony Company in 1880. Between 1863 and 1969, three shafts were sunk on the Hibbard, Adams, and Lawrence showings. Between 1972 and1981, 34,417 tonnes of concentrate grading 65-66% Sb was produced at the Lake George Mine [1]. Prior to 1984, approximately 0.85 Mt of ore grading approximately 4.15% antimony was mined from several occurrences. In 1990, Apocan Inc. acquired the property and reopened the mine in 1994 in response to a drastic increase in the price of antimony. Production resumed in 1996 but came to an end in December of that year with closure of the mine due to mechanical issues and falling antimony prices. Approximately 1 Mt of ore grading 4% Sb was extracted from the No. 1 and No. 2 orebodies between 1985 and 1990 [1]. As of 1997, the proven, probable, and possible ore reserves totaled 262,000 tonnes at 3.45% Sb [2]. A recent NI 43-101 report quoted an indicated mineral resource of 153,147 tonnes grading 3.73% Sb and inferred mineral resource of 718,046 tonnes grading 2.61% Sb at a cut-off grade of 1.5% Sb [3].
Previous investigations have distinguished the relative timing of magmatic and hydrothermal events within the Lake George contact-metasomatic aureole based on cross-cutting relationships [4][5][6]; however, the absolute timing of these events has not been thoroughly investigated owing to the limitations of geochronological methods and techniques previous employed. For example, previous K-Ar dating studies did not have the precision necessary to resolve the absolute timing of the various geologic events [5][6][7]. More recently, geochronological constraints were placed on the quartz-feldspar porphyry dyke (QFP) [8] and the Lake George granodiorite [9] with U-Pb zircon geochronology using thermal ionization mass spectrometry (TIMS). Despite applying a common lead correction, all analyses exhibited variable degrees of discordancy. U-Pb zircon TIMS geochronology is limited by the high cost per sample analyzed, which is why newer methods with much lower costs per analysis such as laser ablation inductively coupled plasma mass spectrometry (LA ICP-MS) have largely replaced TIMS for broad scale or reconnaissance studies. Another advantage of LA ICP-MS U-Pb geochronology over TIMS is the ability to date minerals in situ, allowing the user to selectively analyze specific domains within the zircon crystals. Modern U-Pb, Re-Os and Ar-Ar geochronological methods provide the opportunity to garner information on the absolute timing of events and quantify the duration of this multi-generational magmatic hydrothermal system.
The objective of this study is to establish the absolute timing of the various magmatic and hydrothermal events in the Lake George polymetallic system to help constrain the genesis of the multiple mineralizing events.

Regional Geology
In the Fredericton area, Silurian rocks of the Kingsclear Group form an approximately 80 km wide belt known as the Fredericton Trough, which lies between the Miramichi Highlands to the northwest At the Lake George mine, which lies on the southeast margin of the Pokiok Batholith, a granodiorite stock (informally referred to as the Lake George granodiorite intrudes tightly folded Early to Late Silurian turbiditic sedimentary rocks of the Burtts Corner Formation (Figures 1 and 2). Within these Silurian sedimentary sequences, deformation associated with the Acadian Orogeny produced upright, open to tight roughly north-south trending folds and accompanying northeast-trending, 20 • south-dipping axial planar cleavage within the Kingsclear Group [19]. The apex of the granodiorite stock lies approximately 500 m north of the Hibbard shaft and approximately 350 m below surface ( Figure 2). Intrusion of the granodiorite stock into the Kingsclear Group resulted in a distal biotite-grade contact metamorphic aureole that extends up to 1 km from the intrusion contact and a narrow (200-300 m from the stock) proximal cordierite-andalusite grade contact metamorphic aureole. The contact metamorphic corundum + potassium feldspar assemblage was observed [5] within the stability field of andalusite, placing an upper pressure limit on contact metamorphism at 1.75 kbar (175 MPa).

Figure 2.
Geological map of the Lake George mine, illustrating the granodiorite stock contoured at depth, the quartz-feldspar porphyry dyke, the lamprophyre dykes, and vein systems in the area (modified from [6]).
The cross-cutting relationships of the various magmatic and hydrothermal events at the Lake George mine have been described in detail [5]. A total of three intrusive events and five generations of veins are described below (Figure 3). Emplacement of the granodiorite stock is the oldest intrusive event (I-1) followed by the formation of the first generation of W-Mo bearing quartz veins (Type I). These veins have zoned calc-silicate alteration envelopes (garnet-clinopyroxene-wollastonite) and were not observed directly cross-cutting the granodiorite; however, they do cross-cut the contact metamorphic aureole [5]. Following the formation of the Type I veins, a series of subvertical roughly east-striking lamprophyre dykes (I-2) were emplaced into the sedimentary rocks. Subsequently, a northwest-striking quartz-feldspar porphyry (QFP) dyke (I-3), which cross-cuts the lamprophyre dykes, was emplaced. The second generation of quartz veining (Type II W-Mo), postdates the QFP dyke, and the Mo-bearing prehnite-quartz veinlets (Type III) cross-cut Type II veins and represent the latest W-Mo mineralizing event at the Lake George mine. The first episode of Au-Sb mineralization at the Lake George mine is a network of quartz-carbonate veins and veinlets with corresponding sericitic alteration envelopes (Type IV) containing gold-bearing pyrite and arsenopyrite, followed by the Sb-quartz veins, which contain stibnite and quartz with minor pyrite, pyrrhotite, native antimony, and arsenopyrite (Type V). The Type V veins include the Hibbard, Adams, Lawrence, and Prout veins (Figure 2), and were the sites of previous antimony mining.

Lake George Granodiorite
The Lake George granodiorite is a fine-to medium-grained and seriate to porphyritic stock that straddles the plagioclase-biotite tie line on the ACF diagram [5,22]. Petrochemical characteristics of the stock are transitional between I-and S-type granites [5]. The granodiorite cupola is not exposed at surface but has been intersected in underground drilling. The contact metamorphic aureole associated with the emplacement of the granodiorite stock predates all hydrothermal activity (i.e., Type I through Type V veins) [6].
The Hawkshaw Granite crops out 4.5 km northwest of the Lake George mine and is the nearest phase of the Pokiok Batholith ( Figure 1, [23]). The Hawkshaw Granite consists primarily of pink, equigranular to seriate biotite granite and granodiorite [24]. Both the Hawkshaw Granite and the Lake George granodiorite exhibit low ferric iron ratios indicative of ilmenite series granitic rocks. Similarities in major-and trace-elements has led some authors to suggest that the Hawkshaw Granite and Lake George granodiorite are petrogenetically linked through fractional crystallization [23]. These geochemical characteristics suggests the magma that formed the Lake George granodiorite was derived from the partial melting of arc-like lower continental crust in a volcanic arc to late orogenic environment, near the time of emplacement of the Pokiok Batholith [23].
Geochronology of the Lake George granodiorite was first undertaken using K-Ar analysis of biotite which yielded a date of 390 ± 7 Ma [7]. Previous U-Pb zircon geochronology of the Lake George granodiorite yielded an upper 207 Pb/ 206 Pb Concordia intercept date of 414 ± 2 Ma [9].

Lamprophyre Dykes
A series of east-striking, steeply south-dipping lamprophyre dykes at the Lake George mine were emplaced into the Kingsclear Group metasedimentary rocks. These fine-to medium-grained hornblende-phyric dykes are basaltic trachyandesitic in composition and have ultra-potassic geochemical characteristics [25]. The geochemistry and mineralogy of these dykes are characteristic of shoshonitic lamprophyres of calc-alkaline affinity. K-Ar geochronology of actinolitic hornblende from the dyke yielded a date of 382 ± 7 Ma [6]. The same method applied to phlogopite from these dykes yielded a date of 415 ± 8 Ma [6].

QFP Dyke
A quartz-feldspar porphyry (QFP) dyke at the Lake George mine strikes north-northwest and dips steeply to the west, is approximately 3-5 m thick and contains quartz and two feldspars as phenocrysts within an aphanitic groundmass. Normative mineralogy calculations from whole-rock analysis suggests that the dyke plots at the boundary between the granite and granodiorite on a QAP diagram [26]. On trace element geotectonic discrimination diagrams their composition plots in the volcanic arc granitoid field [27]. The size and texture of the phenocrysts and groundmass in the QFP dyke are relatively consistent from depth to the surface, with the exception of more brecciated margins near surface. The groundmass is recrystallized and variably hydrothermally altered which was interpreted to reflect auto-metasomatic alteration by exsolved deuteric fluids [8].
Uranium-lead zircon geochronology conducted through isotope dilution-thermal ionization mass spectrometry (ID-TIMS) on the QFP dyke yielded an upper intercept date of 420.8 + 5.9/−4.0 Ma [8]. Five single and double zircon fractions were analyzed and corrected for common Pb. The four zircon analyses used to produce the age were 94.0-99.4% concordant and the MSWD of this age was 2.9.

Sampling
Samples were collected with the aims of understanding the origins of the gold mineralization as well as the temporal relations between as many magmatic-hydrothermal events as possible using more modern and robust analytical methods. The samples used in this study were collected from drill core that is housed on the mine site or were part of a catalogue of historic samples housed at the University of New Brunswick (UNB) in Fredericton. Samples of the Lake George granodiorite were taken from drill core which was least altered in appearance ( Figure 4). The samples of lamprophyre dyke came from the preexisting sample catalogue as well as from least altered samples of drill core ( Figure 5).
Type II W-Mo-bearing quartz veins were sampled from drill core and were submitted for molybdenite extraction and analysis. Molybdenite which was sampled and separated from these veins was done with special attention to choosing molybdenite that was not in contact with any other sulphides. Sample KT-11-03 ( Figure 6a) was collected from a scheelite-and molybdenite-bearing quartz vein in drill hole 81-27 from a depth of 2438.8 to 2424.4 feet (743.4-739.0 m) that cut across altered Lake George granodiorite. The sample consists of sericitized coarse-grained granodiorite that contains molybdenite associated with quartz. The irregular shape of the 3 cm long molybdenite mass suggests that it may have been remobilized. The second sample (KT-11-04; Figure 6b    The samples of gold-bearing quartz-carbonate veins were taken from drill core with historic gold assays [28,29], from large bulk sample piles brought to surface during mining, as well as historic samples originating from the gold room [1]. Samples of gold-bearing quartz-carbonate veins were taken and muscovite from the green sericitic alteration was separated and analyzed using step heating 40 Ar/ 39 Ar geochronology. Gold in these samples is refractory and is hosted within sulphides (arsenopyrite and arsenian pyrite) found within the green sericitic alteration envelopes surrounding the veins. Sample LG-80-33-1848 ( Figure 7) was taken from drill core housed at the mine site and contains a quartz-carbonate vein with associated sericitic alteration surrounding the vein from which coarse-grained (up to 1 mm diameter) muscovite was separated. Sample 03-LG-1 ( Figure 8) was a grab sample taken from a rock pile at surface that was extracted for bulk rock assay from the gold zone underground at the 850 level [29]. Muscovite from sample 03-LG-1 was too small to be picked by hand and therefore was analyzed as whole-rock chips (180-250 µm diameter) which contained >95% muscovite and quartz with the remainder being made up of chlorite, ankerite, and sulphides. Sample LG-78-4-143.5 ( Figure 9) was taken from drill core and contains muscovite from the sericitic alteration surrounding the gold-bearing quartz-carbonate vein hosted within altered sedimentary rocks. Sample LG-85-3 ( Figure 10) was part of the catalogue at UNB, which was originally taken underground from the gold zone in the 1150 level on the west side. Muscovite from sample LG-85-3 and LG-78-28-143.5 ft were also separated, but the crystals from these samples were relatively small (>50 µm).    LG-85-3. This sample was taken from underground at the Lake George mine from the gold zone in the 1150 level on the west end.

Analytical Methods
Samples of Lake George granodiorite and lamprophyre dyke were collected for in situ laser ablation inductively coupled plasma-mass spectrometry (LA ICP-MS) U-Pb geochronology in order to establish absolute timing of these magmatic events. To determine the age of emplacement of Type II W-Mo veins, the veins were sampled and molybdenite separates prepared for Re-Os geochronology. Whole rock samples of lamprophyre dyke and gold-bearing quartz-carbonate veins were submitted for 40 Ar/ 39 Ar analysis using step heating.

LA ICP-MS U-Pb Methodology
All U-Pb geochronology was conducted at the University of New Brunswick isotope geochemistry lab using in situ LA ICP-MS. The system in use at UNB is comprised of a Resonetics M-50 193 nm ArF excimer laser coupled with an Agilent 7700x quadrupole ICP-MS. The LA ICP-MS system at UNB is equipped with a variety of features that enhance their ability to date minerals with common lead, most of which are focused upon improving the sensitivity of the ICP-MS. These features include mixing of N 2 downstream from the ablation cell which enters through a Laurin Technic PTY PEEK Y-connector and squid smoothing device before reaching the ICP-MS, mercury traps which are in line with all gas lines entering the ablation cell (keeps 204 Hg < 150 cps), along with a secondary external rotary pump [30]. The primary standard used for external calibration of the data from these analyses was the NIST 610 doped glass standard. The matrix matched consistency standard used when ablating zircon was the Temora 1 standard [31]. Zircon was ablated using a fluence of 4.5 J/cm 2 , a frequency of 4.5 Hz and a spot size of 36 µm. When titanite was ablated, the matrix-matched consistency standard was the Khan Mine standard [32]. Titanite was ablated using a fluence of 6 J/cm 2 , a frequency of 4.5 Hz and a spot size of 45 µm. Data reduction was performed using IOLITE [33] and plotted using Isoplot [34].

Re-Os Methodology
To establish an absolute age for the timing of the W-Mo stockwork veins, two molybdenite-bearing samples were collected for Re-Os geochronological analysis. Drill core samples containing molybdenite were submitted to the University of Alberta Radiogenic Isotope Facility. Each sample was subjected to metal-free crushing after which the molybdenite was extracted by gravity and magnetic concentration methods. Methods used to analyze the molybdenite are described in detail by other authors [35,36]. The 187 Re and 187 Os concentrations in the molybdenite were determined by isotope dilution mass spectrometry using Carius-tube digestion, solvent extraction, anion exchange chromatography, and negative thermal ionization mass spectrometry techniques. A mixed double spike containing known amounts of isotopically enriched 187 Re, 190 Os, and 188 Os analysis was used. Isotopic analysis was performed using a ThermoScientific Triton mass spectrometer with Faraday collectors. Total procedural blanks for Re and Os are less than 3 picograms and 2 picograms, respectively, which are insignificant for the Re and Os concentrations in molybdenite. The Chinese molybdenite powder HLP-5 [37] is routinely analyzed as a Re-Os age standard. For this sample over a period of two years, an average Re-Os date of 221.5 ± 0.4 Ma (1 SD uncertainty, n = 8) is obtained. This Re-Os age date is indistinguishable from the established age of 221.0 ± 1.0 Ma [37]. 40 Ar/ 39 Ar geochronology was conducted at the Queen's University Ar-Ar geochronology lab.

40 Ar/ 39 Ar Methodology
The samples were first crushed, sieved, and then washed. Mineral separates were purified by hand picking. For fine-grained rocks, whole rock fragments in the 180-250 um diameter range were used. After preparation, the sample along with a flux monitor was placed in an irradiation capsule and irradiated with fast neutrons at the McMaster Nuclear Reactor in Hamilton Ontario. The samples were irradiated at the McMaster reactor in position 5C or 8D without Cd shielding. Hb3Gr hornblende GA1550 biotite (1081.0 Ma and 99.738 Ma, respectively; referenced to FC at 28.294 Ma, [38]) were used as flux monitors. After irradiation, the samples were step heated using a 30W New Wave Research MIR 10-30 CO 2 laser in a stainless-steel chamber attached to an ultrahigh-vacuum purification system. Argon isotope analysis was done using an MAP-216 mass spectrometer [39]. The 40 Ar/ 39 Ar dates calculated for these samples use the 40 K decay constant and flux monitor ages [38] and all errors are quoted at 2-sigma. Age spectra were generated using ISOPLOT 3.71 [34]. Because the samples are very radiogenic, no meaningful inverse isochron dates or initial 40 Ar/ 36 Ar ratios could be defined.

Lake George Granodiorite
A total of 14 spot analyses were conducted on six zircon grains ( Figure 11) from the medium-grained granodiorite sample (sample LG-80-33-2017). The Temora 1 zircon standard has an accepted age of 417 Ma and the analyses produced a concordant age of 419.5 ± 2.9 Ma (Figure 12). Of the 14 analyses from six zircons, 11 yielded near-concordant results (Table 1). From those 11 near-concordant analyses, five formed a roughly concordant population without requiring a common lead correction (see bold analyses, Table 1) which yielded a weighted mean 206 Pb/ 238 U date of 419.6 ± 3.0 Ma ( Figure 13).     Table 1). Data point error ellipses are reported at 2σ.

Lamprophyre Dykes
In situ LA ICP-MS U-Pb geochronology was conducted on titanite crystals ( Figure 14) from two samples (78-21 and 80-12, see Figure 5). The results of the analyses plot as three populations. In order to produce dates for each of these populations, a 204 Pb correction was applied. The oldest population has five data points and yielded a 204 Pb corrected concordia date of 476 ± 11 Ma (Table 2, Figure 15). The next younger population, comprised of 18 analyses, yielded a 204 Pb corrected weighted mean 206 Pb/ 238 U date of 435 ± 6 Ma (Table 3, Figure 16). The youngest population consisted of four analyses and yielded a 204 Pb corrected concordia date of 412 ± 9 Ma (Table 4, Figure 17). The titanite data was standardized using the Khan Mine titanite, which has an accepted age of 520 ± 2 Ma [40,41].   Table 2). Data point error ellipses are reported at 2σ.  Figure 16. This population of eighteen analyses from a lamprophyre dyke is the second oldest U-Pb date produced using titanite from a lamprophyre dyke. The date listed above is a 204 Pb corrected weighted mean 206 Pb/ 238 U date (data shown in Table 3). Error bars are to 2σ.  Table 4).

Re-Os Molybdenite (W-Mo Veins)
Single molybdenite crystal Re-Os isotopic analysis of sample KT-11-03 ( Figure 6a) yielded a date of 415.7 ± 1.7 Ma. Analysis of molybdenite from sample KT-11-04 ( Figure 6b) yielded a date of 416.1 ± 1.7 Ma. The isotopic data and model ages are listed below in Table 5. The analytical data for phlogopite from sample 80-12 is listed in Table 6 and the age spectrum is plotted in Figure 18. The 40 Ar/ 39 Ar data did not yield a reliable plateau date but the last step yielded a high temperature plateau segment date of 419.4 ± 1.4 Ma (34.1% of the 39 Ar released).  40 Ar/ 39 Ar geochronology was conducted on very fine-to coarse-grained hydrothermal muscovite from the alteration envelopes that mantle gold-bearing quartz-carbonate veins (Type IV). The 40 Ar/ 39 Ar isotopic data from the analyses of gold-bearing samples LG-80-33-1848 ft, 03-LG-1, LG-78-4-143.5 ft, and LG-85-3 are reported in Tables 7-10, respectively, and the age spectra are shown in Figures 19-22. Plots of the cumulative percentage of 39 Ar released versus the Ca/K ratios were also generated (Figure 23a-d) to assess purity. The fine grain size of muscovite in some samples (as small as 1-10 µm) may have increased the potential for recoil loss of 39 Ar during irradiation to have occurred [42]. This would have the effect of increasing the 40 Ar/ 39 Ar ratio resulting in erroneously old dates. A summary of the up to date geochronology of all magmatic-hydrothermal events at the Lake George Mine is presented in Figure 24.         Analysis of the muscovite separate from sample LG-80-33-1848 produced a 13 step (out of 13) 40 Ar/ 39 Ar plateau date of 410.3 ± 1.1 (2σ) Ma (100% of the 39 Ar) in Isoplot ( Figure 19). On the graph of cumulative percentage of 39 Ar released versus the Ca/K ratio (Figure 23a), all of the calculated Ca/K values are <0.05, except the first step confirming sample purity.
Using analyses from sample 03-LG-1, an age spectrum plot was created ( Figure 20) and a plateau date (steps from 2 to 12) of 418.3 ± 1.6 (2σ) Ma (MSWD of 1.1, 85.9% of the 39 Ar released; Figure 20) was calculated. The graph of Ca/K versus cumulative percentage of 39 Ar released (Figure 23b) shows that this plateau date was created using steps with Ca/K ratios as high as 2.06. A plateau segment date (steps from 6 to 10) of 415.9 ± 2.9 Ma (2σ, MSWD of 0.3 and 27.3% of the 39 Ar) was also calculated. The Ca/K ratios for the steps used to calculate this plateau segment date range from 0.09 to 0.18 (Figure 23b).
The age spectrum plot from sample LG-78-4-143.5 shows a progressive decrease in age for each step from low temperature to high temperature ( Figure 21). A 40 Ar/ 39 Ar plateau segment date (steps from 2 to 6) of 422.4 ± 1.7 (2σ) Ma (MSWD of 1.3 and 67.8% of the 39 Ar) was also calculated. On the graph of cumulative percentage of 39 Ar released versus the Ca/K ratio (Figure 23c), the steps used to calculate the plateau segment date had Ca/K ratios of 0.01. The high temperature plateau segment defined by steps from 7 to 10 has a date of 418.4 ± 2.6 Ma (MSWD of 0.23 and 26.0% of the 39 Ar). The steps used to calculate the high temperature plateau segment date had Ca/K ratios ranging from 0.01 to 0.04.
From the step heating data for sample LG-85-3, an age spectrum plot was generated ( Figure 22, Table 10) and a 40 Ar/ 39 Ar high-temperature plateau segment date (steps from 7 to 8) of 414.1 ± 1.3 (2σ) Ma (MSWD of 1.8, 67.3% of the 39 Ar released) was calculated. The age spectrum plot shows a large range in ages through the low-to mid-temperatures steps, which correspond to steps with high calculated Ca/K ratios (from 0.55 to 2.27; Figure 23d).

Lake George Granodiorite
The results of the in situ LA ICP-MS U-Pb zircon geochronology of the Lake George granodiorite ( Figure 6) produced a Late Silurian date of 419.6 ± 3.0 Ma (Figure 13). This suggests that the Lake George granodiorite is not genetically related to the Hawkshaw Granite as previously hypothesized [23] and is the oldest intrusive phase dated in the Pokiok Batholith. Previous U-Pb TIMS zircon geochronology on this intrusion yielded an upper intercept ( 206 Pb/ 207 Pb) date of 414 ± 2 Ma [9], calculated using three zircons to which a common lead correction was applied with each zircon exhibiting varying degrees of discordancy. The date produced from the current study is a weighted mean 206 Pb/ 238 U date from 5 near-concordant analyses, none of which required a common lead correction. This updated date is in agreement with the geochronology conducted in this and other studies and is corroborated by the relative timing determined by the cross-cutting relationships documented by previous authors [5]. For these reasons, this newly calculated date (419.6 ± 3.0 Ma) is believed to be a more accurate estimate of the true age of emplacement of the Lake George granodiorite.

Lamprophyre Dyke
Understanding the absolute timing of lamprophyre dyke emplacement is critical as lamprophyres have been implicated in enriching melts in large ion lithophile elements (LILE), light rare earth elements (LREE) as well as gold [43,44]. The results of in situ analysis of titanite using LA ICP-MS U-Pb geochronology of a lamprophyre dyke produced three populations (Figures 15-17). The date produced from the youngest of the three populations of titanite (412 ± 9 Ma, Figure 17) is the only U-Pb titanite date produced, which would not be xenocrystic as the two other populations yielded dates that are too old when compared to cross-cutting relationships established underground [6]. Titanite from both populations also underwent variable degrees of lead loss. Although this date (412 ± 9 Ma) is within error of many of the other dates produced in this study, the known cross cutting relationships [6] place emplacement of this dyke shortly after that of the Lake George granodiorite, which make this date too young to be representative of the age of lamprophyre dyke emplacement. This date also lacks the precision necessary to resolve the absolute timing between the various magmatic phases.
The 40 Ar/ 39 Ar geochronology of phlogopite from the lamprophyre dyke (sample 80-12), specifically the high temperature plateau segment date (419.4 ± 1.4 Ma), is most likely the best estimate of the age of emplacement of the lamprophyre dyke. Partial chloritization of the phlogopite is evident based upon the hump from low-moderate temperature steps and significant amount of 39 Ar released at low temperature, both of which are evidence that recoil loss and/or redistribution of 39 Ar during irradiation occurred [45]. Several authors have documented the capacity of chloritization in biotite to enhance the recoil redistribution and loss of 39 Ar in irradiated samples [45,46]. In metamorphic terranes, the effects of very minor chloritization in biotite are even more effective at resetting the Ar-clock than temperature [47]. Auto-metasomatism is common in lamprophyric dyke systems [48], consequently there is considerable potential for primary phlogopite to be altered to chlorite during and directly following dyke emplacement. For these reasons, the 419.4 ± 1.4 Ma date should be considered a maximum age for emplacement of the lamprophyre dyke and is the most representative of the true age of dyke emplacement produced from this study.

LG-80-33-1848
Sample LG-80-33-1848 produced a plateau segment date of 410.3 ± 1.1 Ma, which is the youngest of all the dates produced from the Type IV gold-bearing quartz-carbonate veins. The 40 Ar/ 39 Ar age spectrum ( Figure 19) displays a progressive increase in the dates, ranging from 370 to 412 Ma with increasing step temperature suggestive of a diffusion-loss profile, which is often seen in larger crystals [43]. The plateau segment date of sample LG-80-33-1848 overlaps well with the age of the Hawkshaw Granite obtained from U-Pb of titanite (411.2 ± 2.0, [20]). The interpretation of this date is that it represents the age this sample reached the closure temperature for argon retention in muscovite (~400 ± 50 • C, [49]) meaning it is a cooling age, possibly due to the larger muscovite crystals or because it was the only sample crossing cutting the Lake George granodiorite.

03-LG-1
The plateau segment date of 418.3 ± 1.6 Ma calculated for sample 03-LG-1 was calculated using steps with Ca/K values as high as 2.06 which are much greater than the Ca/K ratios predicted for muscovite. The high temperature plateau segment date (415.9 ± 2.9 Ma) was produced from steps with minimum Ca/K values (steps from 7 to 10) and therefore is considered to be the best estimate of the age of muscovite formation in this sample. The age spectrum diagram has a low-moderate temperature hump and a large portion of the 39 Ar released at low temperature, an indication that recoil loss of 39 Ar occurred during irradiation. The recoil loss of 39 Ar is likely due to the very small grain size of the muscovite as well as being analyzed as whole-rock chips. If recoil loss of 39 Ar occurred, then the 40 Ar/ 39 Ar ratio would increase, which would ultimately produce an erroneously old date. For these reasons the dates produced from this sample are not likely the most representative for the age of the Type IV gold-bearing quartz-carbonate veins.

7.3.3.
LG-78-4-143.5 The oldest date obtained from 40 Ar/ 39 Ar geochronology of the Type IV veins was from sample LG-78-4-143.5 (422.4 ± 1.7 Ma). The 40 Ar/ 39 Ar plateau diagram displays a progressive decrease in the age with each step (see Figure 21). This progressive decrease in age along with a large percentage of the 39 Ar released at low temperatures, suggests recoil loss of 39 Ar has occurred. The fine grain size of this muscovite results in an increased chance of recoil loss of 39 Ar, thus increasing the 40 Ar/ 39 Ar ratio, which would ultimately result in an erroneously old age. The high temperature plateau segment date from this sample (418.4 ± 2.6 Ma) is also problematic as the Ca/K ratios for the steps used to produce that date are much higher (>0.2) than those expected from muscovite and therefore the date would not be representative of the muscovite associated with the Type IV gold-bearing quartz-carbonate veins.

LG-85-3
Sample LG-85-3 produced a high temperature plateau segment date of 414.1 ± 1.3 Ma, which is within error of the high temperature plateau segment date from sample 03-LG-1. These two dates overlap within error of each other, despite sample 03-LG-1 being analyzed as whole-rock chips and sample LG-85-3 as a muscovite separate. A potential explanation for the slightly older age from sample 03-LG-1 would be the much finer grain size of muscovite in that sample resulted in some recoil loss of 39 Ar during irradiation. This recoil loss of 39 Ar would have the effect of increasing the 40 Ar/ 39 Ar of the sample and thus produce a slightly older age. The fine-grained nature of both samples and the presence of other unidentified potassium-bearing phases complicated the interpretations of the age spectrums.
Based upon cross-cutting relationships documented by previous research [6], the age of the Type IV gold-bearing veinlets must be younger than all other magmatic-hydrothermal events at the Lake George Mine, other than antimony-bearing quartz veins for which the deposit was previously mined. The dates obtained from the Type II quartz veins using single crystal Re-Os of molybdenite (approximately 416.1 ± 1.7 Ma) should be slightly older than the age of the Type IV gold-bearing quartz-carbonate veinlets. The high temperature plateau segment date produced from sample LG-85-3 is likely the best estimate of the age of emplacement of the Type IV gold-bearing quartz-carbonate veins.

Conclusions
Although the relative timing of the various magmatic-hydrothermal events at the Lake George mine was previously documented, the absolute timing of these events using geochronology conducted in the current study gives insight into the duration of this multi-episodic magmatic-hydrothermal to later overprinting hydrothermal systems and provides a framework for the exploration of similar mineralization along the eastern margin of the Pokiok Batholith.

1.
LA ICP-MS analyses of zircon from the Lake George granodiorite produced a weighted mean 206 Pb/ 238 U date of 419.6 ± 3.0 Ma from 6 near-concordant zircon analyses. This age suggests that the Lake George Granodiorite is one of the oldest intrusions in the Pokiok Batholith and cannot be related to the nearby Hawkshaw Granite as previously proposed [23]. The Hartfield tonalite (415 ± 1 Ma, [20]) is the intrusion closest in age to the Lake George granodiorite. More recent work on the Hartfield tonalite suggests its age does overlap with the Lake George granodiorite (417 ± 2 Ma from zircon, [50]).

2.
LA ICP-MS analyses of titanite from the lamprophyre dyke produced an age that lacks the precision necessary to establish the absolute timing of its emplacement. 40 Ar/ 39 Ar geochronology of phlogopite from the lamprophyre dyke produced a high temperature plateau segment date of 419.4 ± 1.4 Ma. This date is in agreement with the observed cross-cutting relationships and is a reasonable preliminary age, likely representative of the true timing of emplacement of these calc-alkaline lamprophyre dykes.

3.
Analysis of two single molybdenite crystals using Re-Os geochronology from the Type II W-Mo veins yielded dates of 415.7 ± 1.7 Ma and 416.1 ± 1.7 Ma. These dates are in agreement with previously established cross-cutting relationships [6] and the geochronology of other magmatic hydrothermal events and are likely representative of the true age of emplacement of the Type II W-Mo veins. 4. 40 Ar/ 39 Ar geochronology of hydrothermal muscovite from the sericitic alteration envelopes that surround the Type IV gold-bearing quartz-carbonate veins was severely complicated by the very fine grain sizes of the muscovite analyzed as well as the contribution of 39 Ar from mineral phases other than muscovite. The sample that generated a date most representative of the age of formation of the Type IV gold-bearing quartz-carbonate veins is sample LG-85-3, which yielded 40 Ar/ 39 Ar plateau segment date of 414.1 ± 1.3 Ma. This date is in agreement with the known cross-cutting relationships [6] as well as the other geochronology produced in this study.

5.
We conclude that all mineralization occurred during a period between approximately 417 and 413 Ma, which followed the emplacement of the Lake George granodiorite, the lamprophyre dykes, and the QFP dykes (see Figure 24). To this point, this is the oldest mineralization identified, which is associated with the emplacement of the Pokiok Batholith.

Conflicts of Interest:
The authors declare no conflict of interest.