Introduction

Highly fractionated silica-rich felsic igneous rocks have attracted a large amount of attention in geological literature. One of the reasons for this interest is their distinct enrichment in rare metals, including radioactive elements—Th and U (e.g., Castor and Henry 2000; Cerny et al. 2005; Cuney and Kyser 2015; Nash 2010). Unlike the granitic rocks, there is only a rather limited amount of information on their volcanic equivalents even though felsic volcanic rocks have long been considered a primary source of uranium for many types of uranium deposits (e.g., Cuney and Kyser 2015; Nash 2010; Christiansen et al. 2015). In fact, there is still debate about the origin of these silica-rich rhyolites, particularly as the rocks commonly underwent late-magmatic or post-magmatic modification (e.g., Cuney and Kyser 2015; Hildreth and Wilson 2007; Nash 2010). There are two broad categories of petrogenetic models to explain the origin of silica-rich felsic magmas. The first category invokes a derivation from basaltic parent magma by fractional crystallization or by a combined process of fractional crystallization and assimilation. The second model assumes basaltic magmas provided heat for the partial melting of crustal rocks (e.g., Clarke 1992; Winter 2001). The two processes have also been invoked for the origin of felsic magmas hosting U deposits (e.g., Cuney and Kyser 2015; Nash 2010), which are economically important and include the Streltsovskoye deposit (Transbaikalia, Russia) containing over 280,000 t of U3O8 (IAEA 2009; Cuney and Kyser 2015). The mechanisms facilitating U and Th enrichment in these rocks are poorly understood but can include both primary and secondary processes. In primary magmatic systems, these elements are typically incompatible and tend to become concentrated within highly evolved silica-rich melts, particularly during the latest stages of crystallization. Secondary uranium enrichment processes may involve an interaction of rocks with exsolved magmatic fluids or circulating heated meteoric water (e.g., Cuney and Kyser 2015; Nash 2010). In volcanic rocks, uranium mineralization is commonly concentrated in fracture-fill veins along which hot, metal-charged hydrothermal fluids have migrated.

To contribute to the debate on the origin and evolution of the rhyolitic rocks as well as to U mineralization, we investigated the Late Devonian-Early Carboniferous rhyolitic rocks of the Harvey Group in southwestern New Brunswick (Fig. 1) which hosts uranium mineralization (e.g., Gray et al. 2011; Hassan and Hale 1988; McLeod and Smith 2010; Payette and Martin 1986a, b). The purpose of the paper is to present the geochemical, isotopic and geochronological data on the volcanic rocks in order to constraint their origin, evolution and mineralization.

Fig. 1
figure 1

a The map of a part of Atlantic Canada showing the boundary of the Maritimes Basin and the location of the Harvey and Piskahegan groups (black fields) as well as b. b Simplified geological map of the southwestern part of the Fredericton sub-basin and its environs showing the location of c. Black fields—volcanic rocks of the Harvey and Piskahegan groups. c Geological sketch map of the Harvey Group and surrounding area (modified after Payette and Martin 1986b). The insert (map of Canada) shows the location of a

Full size image

Geological setting

The Late Devonian to Early Carboniferous Harvey Group is a cover sequence of the Ganderia (Gander) terrane, one of the peri-Gondwanan terranes of the Northern Appalachians. The oldest known rocks in Ganderia are the Neoproterozoic arc-related volcanic and plutonic rocks which are overlain by Cambrian arc-back-arc sequences. Ensialic arc magmatism in Ganderia continued intermittently between 640 and 455 Ma (van Staal et al. 2012). In a southwestern part of the New Brunswick, these units are covered by Silurian metasedimentary rocks. The younger magmatic activities in the New Brunswick are Upper Silurian-Lower Devonian bimodal volcanic suites and granitoid plutons. The mafic volcanic rocks have within-plate characteristics (Dostal et al. 1989). The Late Devonian to Early Carboniferous period is marked by widespread bi-modal magmatic activities, particularly around the margin of the Maritimes Basin of the Northern Appalachians. The basin is a major Late Paleozoic successor basin (Gibling et al. 2009) that was initiated after the Early to Middle Devonian Acadian Orogeny (~400 Ma).

Felsic volcanic and volcaniclastic rocks of the Harvey Group occur along and form the base of the northern margin of the Fredericton sub-basin, the southwestern portion of the Maritimes Basin, in southwestern New Brunswick (Fig. 1a), approximately 40 km southwest of Fredericton (Fig. 1b). They form a NE-trending lens-shaped body about 15 km long and 3 km wide (Fig. 1b, c). This subaerial volcanic sequence overlies Silurian metasedimentary rocks, which outcrop to the northwest and in turn is overlain by clastic sediments of the Lower Carboniferous Shin (Bonaventure) and Upper Carboniferous Cumberland groups of the Maritimes Basin (Gibling et al. 2009), which are exposed to the southeast. The Harvey Group has been correlated with the rocks of the Piskahegan Group of the Mount Pleasant caldera complex, located on the southern limb of the sub-basin (Fig. 1b). Aeromagnetic and gravity surveys also suggest that the Harvey Group represents the northwestern margin of the Mount Pleasant complex (McLeod and Smith 2010), the site of a significant polymetallic (tin, tungsten, molybdenum, zinc, indium and bismuth) deposit (e.g., Sinclair et al. 2006; Thorne et al. 2013). Both the Harvey and Piskahegan groups are considered to be the remnants of an epicontinental caldera complex that formed during the initial stages of the development of the Maritimes Basin (McCutcheon et al. 1997; McLeod and Smith 2010).

The Late Paleozoic Maritimes Basin of the northern Appalachians covers the southern part of the Gulf of Saint Lawrence, Prince Edward Island and adjacent regions of New Brunswick and Nova Scotia (Fig. 1a). The basin contains several kilometers of middle Devonian to lower Permian strata. During the Late Devonian-Early Carboniferous the basin underwent an extension accompanied by pulses of igneous activities (Dunning et al. 2002). The Harvey as well as Piskahegan groups were emplaced during one of these pulses.

The Harvey Group (Fig. 1b, c) has been subdivided into three formations, from the bottom up: York Mills, Cherry Hill and Harvey Mountain (Beaudin et al. 1980; Payette and Martin 1986a, b). The York Mills Formation (~60 m thick) contains mostly sedimentary rocks including red sandstones, conglomerates and shale. There is, however, a volcanogenic component to the York Mills Formation, consisting of lithic and lapilli tuffs and laminated rhyolite flows and ignimbrites. The Cherry Hill Formation (~100 m thick) contains most of the U mineralization and is comprised of quartz–feldspar–phyric densely welded ash-flow tuff (“quartz-feldspar porphyry”), with poorly welded ash-flow sheets (Fig. 2) at the base and other ignimbrites and volcaniclastic sedimentary rocks. The Harvey Mountain Formation is a 75–100 m thick sequence composed of pyroclastic breccia, laminated rhyolite lava flows and ignimbrites intercalated with ash-fall tuffs.

Fig. 2
figure 2

a Weakly welded volcaniclastic rock of the Harvey Group (Cherry Hill Formation) containing lithic and mineral fragments enclosed in a fine-grained matrix composed mostly of broken shards, devitrified glass and microphenocrysts of quartz and feldspars. b Ignimbrite with laminated (flow-banding) texture defined by alternating aphanitic bands with layers containing microphenocrysts of quartz and feldspars or spherulites (Harvey Mountain Formation; plane-polarized light)

Full size image

Uranium mineralization

The Harvey Group has been the focus of uranium exploration for decades (e.g., Beaudin et al. 1980; Hassan and Hale 1988; McLeod and Smith 2010 and references therein). It contains several uranium prospects, but there has been no production. Uranium mineralization of the Harvey Group has been classified as a volcanic- and/or caldera-related type (McLeod and Johnson 2007; McLeod and Smith 2010) and is confined to felsic volcanic and intercalated volcaniclastic and clastic sedimentary rocks. The mineralization appears to be stratabound and/or structurally controlled and commonly associated with fluorite. The vein-type mineralization occurs in fracture zones and along permeable horizons within the volcanic piles (McLeod and Smith 2010). The fracture zones contain veins or lenses with pitchblende, pyrite, fluorite and secondary uranium minerals. Mineralization is also associated with chlorite-haematite and clay alteration zones and/or shear zones. Recent exploration reported values up to ~0.45 wt% U3O8 in welded ash-flow tuffs (ignimbrites) and volcaniclastic sediments (Cappella Resources Ltd. 2007).

Petrography

Field occurrences and petrography of the Harvey Group volcanic and volcaniclastic rocks were described by Beaudin et al. (1980), Gray (2010) and Payette and Martin (1986a, b). In general, the volcanic rocks of the Harvey Group are composed of quartz, plagioclase and K-feldspar set in a matrix of glass shards and very fine-grained material. Some rocks are porphyritic; the phenocrysts constitute typically 10–15 vol% of the rocks. Plagioclase is more abundant than K-feldspar; both feldspars are not primary but re-equilibrated at low temperature after forming as phenocrysts. The plagioclase is typically albite and K-feldspar contains <5 % Ab component. Phenocrysts of biotite are rare and typically altered; silicification is widespread.

The Harvey Mountain Formation is composed primarily of rhyolitic ignimbrite with prominent laminated (flow-banding) texture (Fig. 2b) defined by alternating aphanitic bands 0.5–1 mm thick with 1–10 mm thick layers composed of microphenocrysts of quartz and feldspars (<1 mm in size) or greenish gray spherulites (up to 4 mm in diameter). The dominant aphanitic bands are devitrified and can contain numerous cavities filled by quartz and frequently coated by fluorite. The lamination of aphanitic rocks is also defined by iron coating. In the Cherry Hill Formation, quartz-feldspar porphyry (densely welded ash-flow tuff) is composed of phenocrysts of euhedral quartz and both feldspars (up to 5 mm in size). The phenocrysts constitute about 15–25 % of the rock and are enclosed in a matrix of devitrified welded shards. The matrix is typically purple and aphanitic. The Cherry Hill Formation is also made up of weakly welded volcaniclastic rocks which contain lithic and mineral fragments enclosed in a fine-grained matrix (Fig. 2a) composed mostly of broken shards, devitrified glass and microphenocrysts of quartz and feldspars. These rocks show a eutaxitic texture, defined by flattened devitrified shards and pumice fragments. The York Mills Formation includes slightly welded to non-welded pyroclastic flows and laminated rhyolites. Rhyolites with devitrified matrix contain small cavities lined with quartz, feldspars, opal and fluorite. Common accessory minerals in most rocks are zircon and apatite.

Payette and Martin (1986b) inferred that although the original texture is mostly preserved, the rocks were modified by high-temperature Na-metasomatism leading to pervasive albitization followed by low-temperature K-metasomatism, accompanied by silicification. During the silicification, albite was destroyed and Na removed from the rocks as K-feldspar grew. The rocks are generally devoid of fresh primary rock-forming minerals except quartz phenocrysts.

Analytical methods

Whole-rock major elements (Table 1, ESM for Appendix 1) were analyzed by X-ray fluorescence at the Department of Earth and Environmental Sciences of University of Ottawa, Ontario (Canada) using a Philips PW2400 spectrometer. Trace elements were determined using lithium metaborate–tetraborate fusion at the Activation Laboratories Ltd. in Ancaster, Ontario, by an inductively coupled plasma mass spectrometer (ICP–MS). Based on replicate analyses, the precision is generally between 2 and 10 % for trace elements and 2 % for major elements. The detection limits and information on the trace element analyses are available at the Activation Laboratories web site (www.actlabs.com).

Table 1 Major and trace element analyses of the rocks of the Harvey Group
Full size table

Nd-isotope ratios of the rhyolitic rocks (Table 2) were determined at the Atlantic Universities Regional Facility at the Department of Earth Sciences of Memorial University of Newfoundland (St. John’s, Newfoundland, Canada). Concentration data are standard ICP–MS analyses and are precise to ±5 % (2σ). Ratios of 147Sm/144Nd were measured directly (without the separation) by ICP–MS with an estimated precision of ±0.5 % (2σ). The isotopic ratios of 143Nd/144Nd were determined using a thermal ionization mass spectrometer operated in a static mode. Measured 143Nd/144Nd values were normalized to a 146Nd/144Nd ratio of 0.7219 calibrated to the JNdi-1 standard (143Nd/144Nd = 0.512115, Tanaka et al. 2000). An average value for 143Nd/144Nd = 0.511849 resulted from replicate analyses of the LaJolla standard, which was determined repeatedly throughout the study. The 2 σ values for all samples are less than or equal to 0.000008 for 143Nd/144Nd and are given in Table 2. εNd (t) values were calculated with respect to CHUR using a present-day 143Nd/144Nd ratio of 0.512638 and 147Sm/144Nd ratio of 0.196593 and subsequently age-corrected. TDM1 model age was calculated using a linear evolution for a mantle and having a present-day Epsilon value of +10. TDM2 model age was calculated according to the model of DePaolo (1988).

Table 2 Sm-Nd isotopic data for the rhyolitic rocks of the Harvey Group
Full size table

Zircons were separated from a quartz-feldspar porphyry sample from the Cherry Hill Formation (NB-17) for 206Pb/238U dating. The separated zircons were mounted in epoxy and photographed in backscattered and cathodoluminescence imagery. Zircon U‐Pb isotopic analyses were performed by laser ablation ICP–MS at the National Taiwan University in Taipei, Taiwan. The full set‐up and methods are described by Chiu et al. (2009). The laser ablation (LA) was preformed using a He gas carrier to improve material transport efficiency. Standard blanks were measured for ∼1 min and calibration was performed using GJ‐1 zircon standard, Harvard reference zircon 91,500 and Australian Mud Tank carbonatite zircon. Data processing was completed using GLITTER 4.0 software for the U–Th–Pb isotope ratios and common lead (Andersen 2002). Isoplot v. 3.0 was used to plot the Concordia diagram and to calculate the weighted mean U–Pb age (Ludwig 2001).

In situ zircon Hf isotope analyses (Table 3) of the sample NB-17 were carried out using a New Wave UP 213 LA microprobe, attached to a Nu Plasma multi-collector ICP–MS, coupled with a fixed detector array of 12 Faraday cups and 3 ion counters at the Institute of Earth Sciences, Academia Sinica in Taipei. Instrumental conditions, data acquisition, and analytical procedures were similar to Griffin et al. (2000, 2006). All results were calibrated to zircon standards, Mud Tank (Woodhead and Hergt 2005) and 91,500 (Wiedenbeck et al. 1995), with typical within run precision (2σ) of ±0.000030 on the 176Hf/177Hf analysis.

Table 3 Results of zircon U/Pb geochronology and Lu–Hf isotopes for the Cherry Hill rhyolite
Full size table

The εHf(t) values are calculated using chondritic ratios of 176Hf/177Hf (0.282785) and 176Lu/177Hf (0.0336) as derived by Bouvier et al. (2008). The 176Lu decay constant of 1.865 × 10−11 year−1 reported by Scherer et al. (2001) is used in the calculation.

Geochronology

206Pb/238U LA–ICP–MS geochronological dating of zircons from the Cherry Hill sample revealed a bimodal distribution of ages centered on a dominant 361 ± 7 Ma (MSWD = 1.7) and subordinate 403 ± 11 Ma (MSWD = 2.1; Fig. 3, Table 3; ESM for Appendix 2). The ~360 Ma age is interpreted as the age of the emplacement of the Harvey Group volcanic rocks. These data are consistent with the age of the Piskahegan Group and the Mount Pleasant complex (e.g., Anderson 1992; McCutcheon et al. 1997; Thorne et al. 2013; Tucker et al. 1998) and the stratigraphic history of the basin. Older, early Devonian age (403 Ma) recorded by zircon likely represents the age of the contaminant of the parental magma (see below).

Fig. 3
figure 3

Concordia plot for LA-ICP-MS zircon U–Pb dating results for rhyolite from Cherry Hill Formation of the Harvey Group. a Grouping of ~400 Ma zircons and b grouping of ~360 Ma zircons

Full size image

Geochemistry

Major elements

The Harvey Group volcanic rocks have variable major element compositions (Table 1) but compared to typical granites and rhyolites (e.g., Clarke 1992; Chayes 1985), they have high contents of SiO2 (76–87 wt%; LOI-free) and alkalis with K2O/Na2O > 1 but are low in TiO2 (~<0.1 wt%), MgO (<0.3 wt%), CaO (<1 wt%) and P2O5 (<0.03 wt%). On the quartz–alkali feldspar–plagioclase (QAP) diagram (Fig. 4), the Harvey rocks are scattered although most of them fall into the field of rhyolite, away from an average of the North American granitoid rocks (Chayes 1985). Their primary chemical compositions were modified by secondary processes as indicated by high values of SiO2 (in some samples >80 wt%), which are not representative of magmatic rocks. In order to evaluate the effects of various post-magmatic processes, Payette and Martin (1986a,b) and Gray et al. (2011) determined the compositions of the melt inclusions in quartz phenocrysts in these rocks (ESM for Appendix 3 and 4). A comparison of the whole-rock and the melt inclusion compositions shows some differences in the concentrations of the major elements. The melt inclusions have higher contents particularly of Al, Na and K than the whole rocks. On the QAP diagram (Fig. 4), whole rocks display a trend from the melt inclusions toward the Q apex. The whole rocks are higher in silica leading to a net dilution of other major elements. This is consistent with the observed silicification of the rocks, which was accompanied by K-metasomatism resulting in an increase of K2O/Na2O but a drop of CaO and CaO/Al2O3 ratios at about 77 wt% SiO2 (Fig. 5). The gradual decrease of Al2O3, alkalis, FeO*, and TiO2 with increasing SiO2 in the rhyolites is also consistent with fractionation of feldspars and minor amounts of biotite and Fe–Ti oxides although these trends, in part, also reflect silicification of the rocks. Low abundances of MgO, P2O5 and TiO2 imply that both the rocks and melt inclusions are highly fractionated. The melt inclusions are peraluminous with aluminum saturation indices (A/CNK = molar Al2O3/[CaO + Na2O + K2O]) > 1. Thus, it can be assumed that the peraluminous characteristics of the volcanic rocks represent the original magmatic features despite of their alteration. Compositionally, both the rhyolites and melt inclusions share many similarities with evolved rhyolites and granites (including some felsic rocks from the Mount Pleasant complex) (e.g., Taylor 1992; Dostal and Chatterjee 1995; Linnen and Cuney 2005; Yang et al. 2003).

Fig. 4
figure 4

Quartz–alkali feldspar–plagioclase (QAP) classification diagram used to characterize the modal composition of the volcanic rocks (modified after Le Maitre et al. 1989). The plot illustrates the composition of the Harvey Group felsic volcanic rocks. The compositions of the melt inclusions of the Harvey and Piskahegan groups (Gray et al. 2011; melt-Harvey and melt-Piskahegan) and the Harvey Group (Payette and Martin 1986a, 1986b; melt-Harvey P&M) as well as the average of the North American granitoid rocks (NA granite) of Chayes (1985) are shown for comparison

Full size image
Fig. 5
figure 5

Variation diagrams of major elements and their ratios for the Harvey rhyolitic rocks using SiO2 as a differentiation index. Melt inclusions from quartz phenocrysts of the Harvey and Piskahegan rhyolitic rocks (Gray et al. 2011: melt-Harvey; melt-Piskahegan and Payette and Martin 1986a, 1986b: melt—Harvey P&M) are also shown for comparison

Full size image

Payette and Martin (1986a,b) reported a F-bearing liquidus mineral (britholite) in the melt inclusions and inferred that the rhyolites are F-rich. Both Payette and Martin (1986 a,b) and Gray et al. (2011) reported the elevated F concentrations in the melt inclusions (>0.2 wt%). In addition, fluorite, a secondary mineral, is abundant in the matrix and the vugs (e.g. Payette and Martin 1986a, b).

Trace elements

A comparison of chondrite-normalized rare earth element (REE) and primitive mantle-normalized trace element patterns of the whole rocks and glass of the melt inclusions (Figs. 6 and 7) shows that the abundances of high-field strength elements (HFSE), REE, Y and Th were not significantly affected by post-magmatic processes with the exception of a silica dilution effect. Relative to typical granitic and rhyolitic rocks (e.g., Clarke 1992; Cerny et al. 2005; Linnen and Cuney 2005), the rhyolites and melt inclusions are enriched in Rb, Th, U, Nb, Ta, Y and heavy REE (HREE).

Fig. 6
figure 6

Chondrite-normalized rare earth element patterns of the Harvey rhyolitic rocks. Melt inclusions from quartz phenocrysts of the Harvey (Cherry Hill Formation) and Piskahegan rhyolitic rocks (Gray et al. 2011) as well as the topaz-bearing rhyolite from the Spor Mountain, Utah (Christiansen et al. 1984), F-rich granites from Mount Pleasant granitic suite (Yang et al. 2003) and Li-F-rich leucogranite from the Davis Lake pluton, Nova Scotia (Dostal and Chatterjee 1995; Dostal et al. 2004) are also shown for comparison. Normalizing values are after Sun and McDonough (1989). a Harvey Mountain Formation; b Cherry Hill Formation; c York Mills Formation; d melt inclusions from Harvey rhyolites (Cherry Hill Formation); e melt inclusions from Piskahegan rhyolites; f Spor Mountain (sample SM-31); DLP-Davis Lake pluton (sample 79-100), MPG-Mount Pleasant granite (PRL-95-1-1771)

Full size image
Fig. 7
figure 7

Primitive mantle- normalized incompatible element patterns of the Harvey rhyolitic rocks. Melt inclusions from quartz phenocrysts of the Harvey (Cherry Hill Formation) rhyolitic rocks (Gray et al. 2011) are also shown for comparison. Normalizing values are after Sun and McDonough (1989). a Harvey Mountain Formation; b Cherry Hill Formation; c York Mills Formation; d melt inclusions of the Harvey rhyolites (Cherry Hill Formation)

Full size image

According to immobile trace elements, the rocks are rhyolites (Fig. 8) and have within-plate characteristics (Fig. 9) in agreement with their emplacement during the early stages of basin extension. The abundances of several HFSE and REE such as Zr, Hf, Nb and Yb and the element ratios such as Yb/Nb (Fig. 10) in the rhyolites decrease with increasing SiO2, while the ratios of other elements such as including La/Yb (Fig. 10) do not change noticeably. The decrease of Yb/Nb ratio among others indicates that some of these trends are due to fractional crystallization. Compared to fractionated peralkaline granitic/rhyolitic rocks (e.g., Clarke 1992), they are enriched in Rb but depleted in Zr (Fig. 10). The rhyolites have also the low Zr/Y ratios (<4). The distinct enrichment of Rb is also reflected in low K/Rb (80–110) and high Rb/Sr (~1–30) ratios compared to typical upper crustal values (283 and 0. 26, respectively, Rudnick and Gao 2003) but are similar to other anorogenic felsic rocks on the Rb/Sr versus K/Rb diagram (Fig. 11). The graph also shows that the distinct variations of the Harvey volcanics in the K–Rb–Sr space are mainly due to Sr, suggesting that Rb was probably not modified during the post-magmatic processes. Unlike the volcanic rocks, the melt inclusions are high in U (~20 ppm; Fig. 10).

Fig. 8
figure 8

Zr/TiO2 versus Nb/Y classification diagram of Winchester and Floyd (1977) for rhyolitic rocks of the Harvey Group. Melt inclusions from quartz phenocrysts of the Harvey rhyolitic rocks (Gray et al. 2011) are also shown for comparison. Symbols Harvey Mountain Fm (cross); Cherry Hill Fm (plus); York Mills Fm (square box); melt inclusions of Harvey Group (triangle)

Full size image
Fig. 9
figure 9

a Rb versus (Nb + Y) and b Nb versus Y (in ppm) tectonomagmatic discrimination diagrams of Pearce et al. (1984) for rhyolitic rocks of the Harvey Group. Melt inclusions from quartz phenocrysts of the Harvey and Piskahegan rhyolitic rocks (Gray et al. 2011) are also shown for comparison. Fields ORG ocean ridge granites, VAG volcanic arc granites, syn-COLG syn-collisional granites, WPG within-plate granites

Full size image
Fig. 10
figure 10

Variation diagrams of trace elements and their ratios for the Harvey rhyolitic rocks using SiO2 as a differentiation index. Melt inclusions from quartz phenocrysts of the Harvey and Piskahegan rhyolitic rocks (Gray et al. 2011: melt-Harvey; melt-Piskahegan) are also shown for comparison. Note that the melt inclusions of Harvey and Piskahegan groups have different concentrations of Zr, Yb and U as well as the different La/Yb ratios

Full size image
Fig. 11
figure 11

Rb/Sr versus K/Rb plot for the Harvey rhyolitic rocks. The melt inclusions as well as most Harvey rocks have Rb/Sr ratios >1. The boundary (at Rb/Sr = 1) that separates the fields of orogenic and anorogenic granites and rhyolites as well as the fields for various granitic rocks are after Abdel-Rahman (2006). The field for the Bokan (A-type) granite is after Dostal et al. (2014). The abundances of Sr are dominant variables. Symbols: Harvey Mountain Fm (cross); Cherry Hill Fm (plus); York Mills Fm (square box); melt inclusions of Harvey Group (triangle)

Full size image

The chondrite-normalized REE patterns of the rhyolitic rocks (Fig. 6) are relatively flat with (La/Yb)n ~1 (0.1–3) accompanied by distinct negative Eu anomalies (Eu/Eu* ~0.005–0.03). This type of pattern is rare among felsic igneous rocks but is associated with highly fractionated F-rich leucogranites and rhyolites (Fig. 6) (e.g., Irber 1999; Christiansen et al. 1984, 1986; Dostal et al. 2004; Taylor 1992; Yang et al. 2003). Bau (1996) named such patterns as “tetrad REE patterns”. The mantle-normalized trace element plots of the Harvey rocks also show a significant depletion in Sr, Ba, Eu and Ti, which reflects the highly evolved nature of the rocks (Fig. 7). Other characteristics of the patterns are relative enrichments in Rb, Ta and Hf accompanied by a depletion of Nb relative to Ta and Zr relative to Hf. Relative to typical crustal value of ~37 (e.g., Rudnick and Gao 2003; Shaw et al. 1976), the rocks have low Zr/Hf (~20) ratio. In general, these patterns are similar to those of metal-rich leucogranites and rhyolites (e.g., Cerny et al. 2005; Linnen and Cuney 2005; Dostal 2016 and references therein).

Isotopes

Nd isotopic ratios for the Harvey rocks are given in Table 2 where they are aged-corrected to 360 Ma. The εNd(t) values (+0.6 to −1.0) for the Harvey rhyolitic rocks are similar to the values of felsic rocks of the Mount Pleasant caldera (Anderson 1992; McCutcheon et al. 1997) and Devono-Carboniferous topaz-bearing granitic plutons of Whalen et al. (1996) from this part of the New Brunswick (Fig. 12). The ages and tectonic settings of the rocks of the Mount Pleasant complex are comparable to those of the Harvey Group. All of these rocks also plot within the range of Devono-Carboniferous felsic rocks of the Maritimes Basin in New Brunswick and Nova Scotia (Keppie et al. 1997; Pe-Piper and Piper 1998). The ɛHf(t) zircon values yielded an average value of 7.0 ± 1.7 (MSWD = 3.4) for the age of 403 Ma (range 4.7–9.5) and a value of 6.3 ± 2.6 (MSWD = 13) for the age of 360 Ma (2.8–9.9; Table 3). The values for 403 Ma are assumed to represent a mid-crustal contaminant of the Harvey magma.

Fig. 12
figure 12

ɛNd(t) versus time plot comparing new Sm–Nd isotopic data of the Harvey Group rhyolites with a compilation of Sm–Nd isotopic data for the Siluro-Devonian basalts of the Gander terrane (Dostal et al. 2016), felsic rocks of the Mount Pleasant caldera (Anderson 1992), Devonian “boundary” (topaz-bearing) granitic plutons (Whalen et al. 1996) and Devono-Carboniferous felsic rocks of the Maritimes Basin in Nova Scotia and New Brunswick (Keppie et al. 1997; Pe-Piper and Piper 1998). Shaded area is the Avalonian basement and subcontinental lithospheric mantle (after Keppie et al. 2012; Murphy et al. 2011). The field for Mesoproterozoic rocks is from Murphy et al. (2008)

Full size image

Discussion

Petrogenesis

The volcanic rocks of the Harvey Group were emplaced contemporaneously with the formation of the Mount Pleasant caldera complex (~360–370 Ma: Anderson 1992; McCutcheon et al. 1997; Thorne et al. 2013, including emplacement of volcanic rocks of the Piskahegan Group dated at 363.8 ± 2.2 Ma; Tucker et al. 1998). This contemporaneity supports an inference of McLeod and Smith (2010), among others, that the Harvey Group is a northwestern extension of the Mount Pleasant caldera (Fig. 1).

The Harvey volcanic rocks (and their melt inclusions) are peraluminous rhyolites that were emplaced in a within-plate environment (Fig. 9), which is consistent with their geological setting characterized by extension-related tectonics and rifting associated with the opening of the Maritimes Basin. The low contents of Ti, Mg and P in the volcanic rocks imply that they are highly evolved. Assuming that the melt inclusions represent the parent magma, the variation trends of major and trace elements suggest that the Harvey volcanic rocks were produced by fractional crystallization of feldspars and minor amounts of biotite, Fe–Ti oxides and accessory minerals. However, the rocks underwent post-magmatic hydrothermal alterations which partially modified their compositions. The alteration included silicification and K-metasomatism accompanied by changes in K2O/Na2O and CaO/Al2O3 ratios at about 77 wt% SiO2 (Fig. 5).

Several models have been invoked for the origin of within-plate rhyolitic rocks such as those of the melt inclusions. They include (1) partial melting of a crustal source (e.g., Sisson et al. 2005), (2) combined assimilation–fractional crystallization (AFC) of basaltic magmas (e.g., Marks and Markl 2001) and (3) extensive fractional crystallization of mafic or intermediate magma (e.g., Turner et al. 1992).

A number of partial melting, fractional crystallization and assimilation–fractional crystallization processes for the origin of the melt inclusions of the Harvey Group have been evaluated using geochemical model calculations. The most robust results were those obtained by fractional crystallization of the andesitic rocks of the Mount Pleasant caldera, in particular of South Oromocto andesites (Anderson 1992; McCutcheon et al. 1997). The andesites were interpreted to be the result of AFC processes involving the crustal contamination of rift-related tholeiitic basalts and of a change from non-oxic to oxidizing conditions accompanied by a relatively high content of magmatic water. As the andesitic rocks contain hornblende indicative of mid-crust crystallization and show a calc-alkaline fractionation trend that is typical of high P(tot) and PH2O conditions, the contamination had to take place at a deeper crustal level (e.g., Osborn 1959; Zimmer et al. 2010). Although the contaminant was a middle crust source, such a rock was not a granulite. Granulites are typically depleted in U (Dostal and Capedri 1978). The contaminant was probably fluid-rich phlogopite/biotite-bearing mid-crustal metasedimentary rock. The 403 Ma inherited zircon age obtained from the Harvey rhyolite probably reflects the age of such a contaminant. The Piskahegan rhyolites are probably correlative of the Harvey rocks although they appear to be less evolved.

Model calculations

The petrological evolution of mafic liquids can be modeled using MELTS software (Ghiorso and Sack 1995; Smith and Asimow 2005). MELTS is calibrated for a wide range of silicate bulk compositions that encompass nearly the entire spectrum of igneous rocks. The software permits the adjustment of magmatic conditions (i.e., relative oxidation state, pressure and water content) of the parental composition. The adjustment of the magmatic conditions allows the user to test a variety of hypotheses that are not necessary related to the parental composition but can provide better constraints on possible liquid evolution.

Fractional crystallization modeling was carried out using a starting composition of South Oromocto andesite from the Mount Pleasant caldera complex (Table 4) in attempt to replicate the petrogenesis of the Cherry Hill Formation and Piskahegan Group melt inclusions (ESM for Appendix 3). The relative oxidation state (i.e., ƒO2 = FMQ 0) and initial water contents (i.e., 2 wt%) used in the model are typical of basaltic andesites, but there is limited geological evidence to constrain the pressure of fractionation; therefore, we attempted modeling at a variety of pressures (Moore and Carmichael 1998; Benjamin et al. 2007). Rare earth elements (i.e., La, Yb) were also included in the model to monitor their behavior during the liquid evolution, in particular to determine whether the flat chondrite-normalized patterns are related to magma evolution or to other processes.

Table 4 MELTS model composition and parameters
Full size table

The liquid evolution results (black circles), at 10 °C intervals, are shown in Fig. 13 and represent the starting composition under the magmatic conditions outlined in Table 4. The models were broadly able to produce compositions similar to the melt inclusions at a pressure of 0.6 GPa (i.e., 6 kbar). The modeled crystallization sequence is as follows: orthopyroxene (1170–1120 °C), diopside-augite (1110 °C), pigeonite (1110–860 °C), Ti-magnetite (1070–920 °C), plagioclase (1130–720 °C), apatite (1030–820 °C), ilmenite (920–720 °C), alkali feldspar (920–720 °C) and garnet (850–720 °C). The proportions of the minerals, compositions and their temperature range are shown in Table 5. Interestingly the model indicates that garnet will crystallize at 850 °C although garnet is not observed in magmatic rocks of the Mount Pleasant complex. Consequently we ran a “garnet-free” model with identical parameters as in Table 4 but suppressed garnet crystallization. The garnet-suppressed model is shown as the white circles (at 10 °C intervals) in Fig. 14, and the results are summarized in Table 6. The main difference between the two models is that pigeonite crystallizes to 720 °C, whereas in the garnet model it stops crystallizing at 860 °C. Although both models show very good agreement with nearly all major elements, CaO, P2O5 and K2O are not well replicated.

Fig. 13
figure 13

Major and trace elemental results of MELTS modeling showing the liquid evolution curves of the starting composition (Table 4). The black circles represent liquid compositions at 10 °C intervals. The white circles represent liquid compositions at 10 °C intervals in the garnet suppression model. The gray circles in panel f represent the liquid compositions in the monazite fractionation model. Additional data from Gray (2010)

Full size image
Table 5 Proportion and composition of crystallized minerals
Full size table
Fig. 14
figure 14

MELTS modeling for the evolution of CaO and P2O5. The overlapping black and white dots represent the liquid evolution compositions of the garnet and garnet-suppressed models based on the starting composition in Table 4. The gray dots are the ‘additional apatite’ fractionation compositions. The modeled composition was adjusted to maintain P2O5 at 0.01 wt%. Composition of the apatite from Deer et al. (1992): CaO = 55.84 wt%; P2O5 = 42.05 wt%. Additional data from Gray (2010)

Full size image
Table 6 Proportion and composition of crystallized minerals in the garnet-suppressed model
Full size table

The CaO and P2O5 may be due to lower amounts (i.e., <1 %) of apatite fractionation in the model. If the model is adjusted to incorporate more apatite (i.e., ~1 % to ~2 %) fractionation, then the CaO content drops closer (i.e., ~0.9 wt%) to the melt inclusion compositions (i.e., ≤ 0.7 wt%) (Fig. 14). Additional apatite fractionation could explain the composition of the melt inclusions from the Piskahegan Group as they have much higher LREE concentration than the Cherry Hill melt inclusions (Fig. 6; ESM for Appendix 3). However, the melt inclusions from the Cherry Hill Formation have much lower La than the Piskahegan melt inclusions but have similar P2O5 content (i.e., <0.05 wt%). A possible explanation is that the Cherry Hill melt inclusions experienced monazite fractionation rather than apatite fractionation as the partition coefficients (Kd) of La in apatite (<25) and La in monazite (~100 or higher) are significantly different (Fujimaki 1986; Stepanov et al. 2012). Using the mineral proportions outlined (substituting monazite for apatite) in Table 6 and Kd values of La and Yb in clinopyroxene (KdLa = 0.014; KdYb = 0.09), plagioclase (KdLa = 0.013; KdYb = 0.02), alkali feldspar (KdLa = 0.08; KdYb = 0.03) and monazite (KdLa = 119; KdYb = 13) in basaltic andesite, we demonstrate in Fig. 13f that a possible monazite fractionation path could explain the depleted LREE patterns of the Cherry Hill inclusions. If monazite is the principal phosphate mineral to fractionate to produce the Cherry Hill inclusion compositions, then an additional mineral must be responsible for the depletion of CaO. The most likely candidate to reduce CaO without affecting P2O5, other major elements or the HREE and Y is fluorite as it is known to be present within the Harvey volcanic rocks.

The discrepancy between the K2O content is problematic as the melt inclusions have ~6 wt%, whereas the model predicts ~3.5 at ~75 wt% SiO2. The reason for the difference between the modeled and measured K2O is uncertain, but it could be due to subsequent assimilation of potassic siliceous fluids into the magma system during fractionation or limitations of the modeling software but most likely the major issue is the fact that the starting compositions of the model is restricted largely due to the limited published data on the South Oromocto andesites (Anderson 1992; McCutcheon et al. 1997; Gray 2010). Furthermore, the South Oromocto andesites have experienced post-emplacement alteration to varying degrees (Anderson 1992; McCutcheon et al. 1997). Although the models show general agreement between the liquid evolution curves and the bulk composition of the melt inclusions, there are a number of uncertainties (i.e., exact parental magma composition) that prevent a more definitive result. We cannot rule out the possible influence of F-rich fluids in the system and that the Cherry Hill Formation and Piskahegan Group melt inclusions, although from the same complex, represent different magma batches or experienced slight differences (e.g., crustal assimilation, influence of hydromagmatic fluids) in their petrogenesis that cannot readily be modeled.

Nd isotopes and REE distributions in the rhyolites

Similarities of ɛNd(t) in the rhyolites suggest that the Nd isotopic system was not disturbed by hydrothermal fluids. The ɛ(t) values of the Harvey rhyolitic rocks show a good agreement with the data of Anderson (1992) and McCutcheon et al. (1997) for the felsic rocks of the Mount Pleasant complex. An exception is one sample (NB-11), which has a high negative value (−4.6) probably reflecting upper crustal contamination as also suggested by chemical composition. The ɛNd(t) values (~0) of both Harvey and the Mount Pleasant caldera negate a significant contribution from substantially older continental crust or sedimentary rocks derived from it. The rocks plot into the field of Avalonian and peri-Rodinian lithosphere (Fig. 12) and their ɛNd(t) values are consistent with derivation from a subcontinental lithospheric mantle with a contribution from a middle crust. The slightly lower ɛHf zircon values for the Harvey (360 Ma) magma than for the 403 Ma zircon indicate a possible addition of a contaminant with higher ɛHf (t), but the analytical precision does not permit a conclusive answer.

The flat chondrite-normalized REE patterns with distinct negative Eu anomalies of the felsic rocks are relatively rare and controversial. They are mostly restricted to highly fractionated granites. These “tetrad-like” rare earth patterns have been attributed to fluid–rock or fluid–melt interactions during late-magmatic stages (e.g., Bau 1996, Irber 1999; Jahn et al. 2001; Monecke et al. 2002). Veksler et al. (2005) inferred that these patterns are related to the partitioning of REE between silicate melts and coexisting fluoride melts. Alternatively, Yurimoto et al. (1990), Duc-Tin and Keppler (2015) and others argued that the shape is the result of crystallization and removal of accessory phosphate phases, most notably monazite and apatite.

The scanning electron microscope observation indicates that the bulk of REE occurs in accessory phases, like in many other felsic rocks. As the melt inclusions and whole rocks have rather similar REE patterns, the magma was modified by a fluid–melt interaction or by a separation of accessory minerals prior to incorporation into quartz crystals. The model calculations show that crystallization of accessory phases from an andesitic parent can account for the flat REE patterns of the Harvey rocks. However, the presence of fluorite in the Harvey sequence and the elevated concentrations of F in the melt inclusions suggest that these highly fractionated melts formed at the top of felsic magma chamber where volatile elements (F, Cl, water) also become concentrated and could have interacted (e.g., Nash 2010) and produced the “tetrad” REE patterns (Veksler et al. 2005). It appears that either or both processes (fractional crystallization of accessory phosphates and fluid–melt interaction) could be involved in the generation of the “tetrad-like” rare earth patterns.

The release of magma and fluids from the top of the chamber may produce an ash flow with composition of the Harvey rhyolites. The magma chamber was probably at a mid-crustal depth because felsic magma tends to erupt as ash-flow tuffs and form a caldera from such reservoirs (John 2001). The zoned pyroclastic sequences indicate that the magma chamber of the Mount Pleasant caldera was zoned (McCutcheon et al. 1997). Zoned magma chambers are typically more fertile (i.e., containing ore-forming fluids) than large homogeneous magma chambers (Rytuba 1994).

Uranium mineralization

The Harvey Group, which can be considered to be a part of the Mount Pleasant volcanic caldera complex (McCutcheon et al. 1997; McLeod and Smith 2010), is a promising site for uranium mineralization and represents a potentially important uranium source for a deposit. The group is composed predominantly of highly fractionated peraluminous rhyolitic rocks (ignimbrites, pyroclastics, and lavas). Volcanic rocks are mainly aphyric, and most U was probably hosted by volcanic glass (e.g., Dostal and Capedri 1975; Dostal et al. 1976; Zielinski 1978, 1981), which has been strongly devitrified by hydrothermal alteration. Glass devitrification facilitates the release of uranium from the rhyolites and is easily leachable by oxidizing hydrothermal fluids. Gray et al. (2011) reported the U contents of silicate melt inclusions trapped in quartz phenocrysts of rhyolite of ~20 ppm. This value is comparable to the U contents of the U-rich Macusani volcanic glass (Pichavant et al. 1988; Nash 2010) and volcanic glass in Transbaikalia, which is considered to represent a U source for the Streltsovskoye deposit (Cuney and Kyser 2015; Chabiron et al. 2003). A comparison of the high U contents and low Th/U ratios of the melt inclusions (~2.3; Gray et al. 2011) with those of the rhyolitic whole rocks of the Harvey Group (Fig. 10) indicates that a significant amount of the initial U from the rhyolite has been lost during the alteration processes. Thus, the Harvey rocks represent a potential source of U via devitrification of the glass by oxidizing fluids. This process could have produced strata-bound vein-type mineralization hosted in volcanic rocks and associated sediments.

Conclusions

The 360 Ma old volcanic rocks of the Harvey Group are highly fractionated peraluminous within-plate F-rich rhyolites. Compared to typical peraluminous rhyolites or granites (e.g., Clarke 1992), the rocks, including the melt inclusions, have higher Rb, Th, U, Y and heavy REE, but lower contents of CaO, MgO, P2O5, TiO2, Ba, Sr and Eu and have uncommon distributions of REE and HFSE. The rhyolites were formed by fractional crystallization of an intermediate magma similar to andesites of the Mount Pleasant caldera complex. The highly evolved Harvey melts enriched in several lithophile elements formed at the top of felsic magma chamber where also volatile elements (F, Cl, water) became concentrated. Although topaz has not been found in the Harvey volcanics, their major and trace element compositions resemble topaz-bearing rhyolites which are known to host U mineralization. The volcanic rocks of the Piskahegan and Harvey groups have similar Nd isotopic characteristics and could have been formed in the same magma chamber. However, unlike the Harvey rocks, the Piskahegan rhyolites were not generated at its apex. The Harvey volcanic rocks had the potential to develop a significant mineralization of U. The erupted U-rich glassy rhyolites may be favorable U source rock amendable to leaching by post-magmatic hydrothermal and meteoric water. The bulk of U in these rocks resides in glass and accessory minerals. During the secondary processes, glass was devitrified and its U was removed either by hydrothermal or meteoric water and deposited in mineralized zones. The Harvey part of the caldera complex composed of highly fractionated pyroclastic rhyolitic material represents a promising setting for the genesis of U deposits related to volcanic rocks.