Discovery of a New Type of Carbohydrothermal Pegmatite at Moose Creek Valley, Ice River Alkaline Complex, British Columbia – Evidence for Extensive Ti Mobilization Open Access

The British Columbia Alkaline Province (BCAP) is a northwest trending belt 150 km wide located along the boundary between the Omineca and Foreland belts of the Canadian Cordillera (Fig. 1) and includes nepheline and sodalite syenites, ijolite, carbonatites, kimberlites, lamprophyres, and associated post-magmatic and hydrothermal rocks (Pell 1994, Hoshino et al. 2017). Geological fieldwork commenced in the BCAP in the early 1800s, yet many occurrences received little to no mention in the scientific literature until the early 1900s (Pell 1994). The Ice River Alkaline Complex (IRAC), the largest within the BCAP, was one of the earliest to undergo extensive study, with the first geological investigations and recorded description by G.M. Dawson, who traced sodalite syenite pebbles from the Kicking Horse River to their source along the Ice River Valley and collected a suite of alkaline rocks for further study (Dawson 1885, 1886). The first large-scale investigation was carried out by J.A. Allan in the 1910 and 1911 field seasons during his studies of the Canadian Cordillera; he was the first to note the presence of veins containing pegmatitic aegirine, natrolite, and other zeolites, along with thick seams of sodalite (Allan 1914). The first dedicated mapping, sampling, geochemical, and crystal chemical program in the IRAC was conducted by K.L. Currie during the 1969 and 1970 field seasons and later published as a Geological Survey of Canada Bulletin (Currie 1975). In 1979, mineralogists from the Canadian Museum of Nature were reportedly the first to actively explore the IRAC for mineral samples and reported on the occurrence of the rare zeolite edingtonite and gemmy natrolite from late-stage pockets in the syenite (Grice & Gault 1981, Grice et al. 1984). Peterson (1983) later re-examined the relationship between the ijolite and syenite and discussed the evolution of the complex. Peterson (1983) also proposed late-stage liquid immiscibility between a Fe + Ti-rich and an alkaline liquid, along with silicate-carbonatite immiscibility as processes to account for the textures and assemblages seen in the complex. More recent work has focused on isotopic systematics (Parrish et al. 1987, Locock 1994), the relationship between the mafic complex and the carbonatite plug (Peterson 1983, Peterson & Currie 1994), and a petrogenetic and isotopic study of the dikes of Moose Creek Valley (Mumford 2009).

Exploration for economic deposits of rare-earth elements (REE), Nb, Ta, and fluorite has occurred within the BCAP since the 1950s. However, exploration has been limited, and few of these deposits have been sufficiently studied to provide estimates on potential reserves or to develop resources (Simandl & Clarke 2016). With >70% of current global production of REE originating from China, and >90% of global Nb coming from Brazil, disruptions in supply have become a concern for both Europe and North America (NRCan 2021). It has become imperative that Canada, which has some of the largest known reserves and resources of REE and Nb + Ta in the world, position itself as a “global supplier of choice” of the critical elements required for a green, low-carbon, high-tech future (Lasley 2021). Many of the REE prospects in Canada occur in British Columbia, including the Blue River-Upper Fir carbonatite (Ta, Nb), Aley carbonatite (Nb), Wicheeda carbonatite (REE), and the Rock Canyon Creek deposit (REE, F, Ba, Au, Ag). On the western flank of Moose Creek Valley in the IRAC, claims leased by Eagle Plains Resources Limited (2023) are being prospected for Nb and REE, along with base metals (Pb, Zn, Ag, and Cu) at the historic, metasediment-hosted Waterloo showing. Mineralization within deposits of the BCAP occurs in both primary and secondary assemblages, dominantly in veins, dikes, and sudations, the result of late-stage carbohydrothermal or immiscible fluids enriched in Na, F, Cl, Nb, LREE, Ba, and Sr derived from the parental carbonatite or syenite bodies.

The IRAC is located approximately 23 km south of the town of Field and 40 km east-southeast of the town of Golden, British Columbia, Canada, with the majority of the complex located in Yoho National Park in the west, and a small section located in Kootenay National Park to the east; only a small portion of the IRAC, straddling the northeast and northwest sides of Moose Creek Valley, is located outside of the park boundaries. Along Moose Creek Valley, mining claims leased by Eagle Plains Resources Ltd. and Chris Robak are located on the western and eastern flanks, respectively. Fieldwork and sampling are restricted to the Moose Creek Valley unless proper permits are obtained from Parks Canada for the two national parks.

In August 2023, CMN researchers were granted access to the claims in Moose Creek Valley in order to collect samples for both acquisition into the National Mineral Collection and for research purposes to study the mineralogy of late-stage syenite dikes and shed light on their paragenesis. Following in the footsteps of previous CMN mineralogists who visited the IRAC, our work in 2023 led to the discovery of a large suite of new Ti-Zr mineral species for the locality from a late-stage, carbohydrothermal pegmatite discovered on the western slope below the ridge linking Zinc Mountain and Sentry Peak. In 2024, we returned to the slope to collect further samples and to determine the lateral extent of the pegmatite. This paper is the first in a series about the IRAC and will describe the geology and mineralogy of this new occurrence, as well as discuss the paragenesis and mobilization of Ti, Zr, and Nb in late-stage carbohydrothermal alkaline systems.

The Ice River Alkaline Complex is located within the Foreland Belt on the eastern side of the British Columbia Alkaline Province (BCAP; Fig. 1), a series of folded and imbricated miogeoclinal metasedimentary rocks that underwent sub-greenschist to greenschist metamorphism during the Columbian orogeny and currently form the Main and Western Ranges of the Rocky Mountain (Gabrielse & Yorath 1991, Pell 1994). Alkaline complexes within this belt are of Mississippian-to-Devonian age and, along with the IRAC, include the Mount Mather Creek sodalite deposit, the Bearpaw Ridge sodalite syenite, the Aley, Kechika River, and Wicheeda carbonatites, and the Rock Canyon Creek REE-F-Ba deposit (Currie 1976, Pell 1994, Peterson & Currie 1994, Simandl et al. 2012, Dalsin et al. 2015, Hoshino et al. 2017, Piilonen et al. 2022). Of the 44 known strategic element deposits in British Columbia, 43 are from the Foreland Belt (Simandl et al. 2012). A common feature observed around alkaline complexes containing carbonatites in this region, including at the IRAC, is the presence of a yellow-to-brown alteration halo, visible from the air (Pell 1994, Peterson & Currie 1994, Hora & Hancock 1996).

The IRAC (Fig. 2) is a J-shaped, multiphase intrusion comprised of (1) an older, layered, feldspar-free, ultramafic to mafic complex in the west (ijolite, pyroxenite); (2) a primary carbonatite plug in the center; (3) a younger, zoned syenite body which grades from melanocratic syenite at the margins to nepheline and sodalite syenite in the core; and (4) a suite of late-stage syenite pegmatites, lamprophyres, and carbohydrothermal veins which crosscut all earlier phases, and at times each other, and intrude into the surrounding host rocks. Age dating of the IRAC over the years by various researchers has utilized every possible isotopic system, yielding a wide range of ages from (220 to 459 Ma) thought to be attributed to disturbance of the primary isotopic signatures by thermal effects and alteration by hydrothermal fluids (Locock 1994, Mumford 2009). The more recent work by Locock (1994) and Mumford (2009) gives similar emplacement ages of 356 ± 6 Ma and 359 ± 3 Ma, respectively. Mumford (2009) also dated monazite from a late-stage ferrocarbonatite dike similar to the carbohydrothermal dike observed at Mount Mather Creek by Piilonen et al. (2022) and found it to be much younger than the IRAC with an age of 165 ± 8 Ma.

The IRAC intrudes gently dipping Cambrian to Ordovician metasediments including intercalated argillaceous shales and massive limestones of the Chancellor and Ottertail formations and the overlying McKay Group (Allan 1914, Peterson & Currie 1994). Hornfels, up to several hundreds of meters away from the contact with the intrusion, is dominantly observed in the Mackay group metasediments, whereas skarns are well developed at the contact with the Ottertail Formation and rarely extend more than 50 m from the contact (Allan 1914, Currie 1975, Peterson & Currie 1994). The lack of metasomatization or fenitization of the host metasediments led Currie (1975) to suggest that late-stage post-magmatic fluids generated by the crystallizing syenite remained internal to the complex, resulting in late peralkaline segregations (veins, pockets, and dikes) within the main mass. Both the hornfels and the skarn are intruded by late-stage dikes originating from the IRAC, reflecting the fracturing and differential movement experienced by the main intrusion and the host rocks during regional deformation. The IRAC is thought to have acted as a competent unit during the Laramide Orogeny (late Cretaceous–Paleocene), during which time the complex and host rocks were thrust ∼200 km to the east (Gabrielse et al. 1991, Gabrielse & Yorath 1991). The metasediments were sheared to varying degrees, but units at the contact with the IRAC, along with late-stage dikes, are strongly deformed (Gabrielse et al. 1991, Greenwood et al. 1991). The IRAC has undergone low-grade, prehnite-pumpellyite facies regional metamorphism.

The investigated pegmatite was discovered by C. Robak, who initially found natrolite- and anatase-bearing boulders in the talus slope 200 m below the actual exposure. Following the talus trail, the in situ occurrence was discovered. The pegmatite itself is emplaced into the Unit 12 syenite as described by Currie (1975) at 2258 m elevation in a steep couloir on the western side of the ridge between Sentry Peak and Zinc Mountain (Fig. 2b). It outcrops horizontally over a short N–S strike of 2.5 m and is a consistent 12 cm thick, pinching out slightly at the southern point (Fig. 3). In 2024, during further exploration 200–400 m further north along the ridge from the initial discovery, additional outcrops of the pegmatite were discovered at the same altitude. The top and bottom contacts of the pegmatite at both locations are altered extensively, with increased alteration at the bottom contact, extending up to 20 cm away from the border (Fig. 4). The original mineralogy of the host syenite (alkali feldspar, Ti-bearing augite, ilmenite, nepheline, sodalite) has been completely replaced by a secondary assemblage of aegirine, biotite, calcite, nontronite, titanite, and a suite of trace minerals including apatite, baryte, sphalerite, and thorite; only trace vestiges of primary augite and ilmenite remain. In thin section, the altered syenite is cloudy, patchy, and many of the remnant augite grains are amorphous or have very mottled, irregular extinction. Replacement minerals, in particular nontronite, are extremely fine-grained, adding to the mottled appearance observed optically. Only secondary biotite occurs as medium-grained, subhedral crystals replacing ilmenite. Within the host syenite, fracture-fill veins parallel to the lower pegmatite contact are common, most often containing fibrous aegirine crystallizing perpendicular to the margins (1–1.5 cm wide), often with cores of coarse-grained calcite (up to 3 cm wide).

The texture of the pegmatite is heterogeneous, with both massive pegmatitic and vuggy zones along with minor brecciation and fracture-filling. The pegmatitic zones at the margins contain massive prismatic natrolite crystals up to 10 cm in length (Fig. 3b), blocky calcite, and euhedral titanite (up to 6 cm). Much of the pegmatite (75%) is vuggy, with cavities containing euhedral, water-clear natrolite, fibrous aegirine, and a later suite of Ti-Zr-REE minerals (Fig. 5a, b). A total of 27 mineral species have been described from the pegmatite so far (Table 1), including an early assemblage of aegirine, analcime, calcite, natrolite, phlogopite, and titanite, with later replacement minerals including allanite, anatase, ancylite-(Ce), apatite, baryte, biotite, brookite, calcite, catapleiite, gibbsite, gonnardite, henrymeyerite, ilmenite, lorenzenite, lucasite-(Ce), natrolite, nordstrandite, sodalite, srilankite, thorite, titanite, vinogradovite, wairakite, wadeite, and zircon. Many of these species are new finds for the IRAC, including anatase, brookite, gibbsite, gonnardite, henrymeyerite, nordstrandite, thorite, vinogradovite, wadeite, and wairakite.

As the pegmatite is dominated by natrolite and aegirine, and is in close proximity to the zeolite-rich syenite zone described by Currie (1975) which tapers out along the ridge between Zinc Mountain and Sentry Peak, it would be a reasonable preliminary conclusion to include it as part of Unit 15. The altered zeolite-rich syenite occurs as a sheet-like body in the upper, central part of the IRAC on the eastern side of Zinc Mountain (see Fig. 2), near the contacts with mafic and country rocks, and is dominated by alkali feldspar, biotite, carbonate, minor aegirine, magnetite, hematite, and localized, late-stage, fine-grained granular and radiating sheaves of natrolite, and granular analcime (Currie 1975). Open pockets and seams, described as pneumatolytic in origin by Locock (1994), containing euhedral, gemmy, water-clear natrolite crystals, the rare Ba zeolite edingtonite and analcime, along with trace aegirine, ancylite-(Ce), catapleiite, ilmenite, magnetite, pyrite, and galena (Grice & Gault 1981, Grice et al. 1984, Locock 1994) have also been found. The contacts between Unit 15 and the sodalite and nepheline syenites are gradational, suggesting localized, diffuse hydrothermal alteration, not a distinct intrusive unit. In contrast, the Moose Creek pegmatite has sharp contacts with the host syenite. It also has a more extensive, distinct assemblage dominated by Na, Ca, Fe, Ti, Zr, and Ce minerals and, notably, a lack of feldspar, sodalite, and nepheline. It is therefore not included in Unit 15, but as part of the late-stage pegmatite assemblage.

Representative crystals for all minerals in the assemblage were mounted in epoxy and polished. Chemical analyses were done with a JEOL Superprobe 8230 operating in wavelength-dispersive mode using Probe for EMPA Extreme Edition software (https://www.probesoftware.com/) at the University of Ottawa. Analytical conditions for the various minerals are reported in Table 2. Time-dependent intensities were obtained for volatiles and other beam-sensitive elements to assess and correct for migration under the electron beam. Raw intensities were converted to concentrations using the default φρZ corrections of the Probe for EMPA software package (Armstrong 1988). Elemental interferences were corrected using empirical overlap corrections. Full EMP analyses for all minerals can be found in Tables 3–9 and 11.

Cathodoluminescence (CL) and back-scattered electron (BSE) imaging of polished grain mounts was undertaken using a JEOL 6610Lv scanning electron microscope equipped with a monochromatic Gatan miniCL detector with operating conditions of 15 kV and a beam current of approximately 1 nA.

Powder X-ray diffraction (pXRD) data were collected using a Bruker D8 discover-MR A25 equipped with Dectris Eiger2 R 500K detector. The instrument uses an Incoatec Cu microfocus source (IµS) operating at 50 kV and 1 mA. The sample is mounted on a 250 µm spherical powder pin. A statistical approach (Rowe 2009) is used to calibrate the image correction parameters (sample-to-detector distance and X-Y beam center coordinates). Further data correction is achieved with a residual error correction curve obtained with an NBS Si 640a standard and tested with an annealed fluorite standard. The resulting 2θ correction curve is applied to an XY data file (with a 0.005° step size) of the diffractogram to fix the instrumentation error, which results in a 2θ error of below ±0.001° for the whole diffraction pattern for diffraction peaks and features. Sample exposure time during data collection was 300 s over a range of 70° 2θ.

The modal mineralogy of the pegmatite is heterogeneous and changes depending on the proximity to the contacts with the host syenite. Estimates of the mineral contents are difficult to determine, although natrolite is the dominant mineral (∼50%), with lesser aegirine (30%), TiO₂ polymorphs (5–8%), titanite (3–5%), calcite (3%), and analcime/wairakite (2%). All other minerals occur in minor amounts yet play a crucial role in understanding the paragenesis of the pegmatite, the composition of the silico-carbohydrothermal fluids and the mobilization of Ti in the system. Table 1 lists all the minerals identified in the pegmatite. Of particular importance is the absence of fluorite and sodalite, present in all the syenites and syenite pegmatites described thus far from the IRAC.

Aegirine, NaFe³⁺Si₂O₆, is a ubiquitous mineral in all the late-stage syenites and pegmatites at the IRAC (Currie 1975), as well as a secondary phase in the ultramafic rocks and the silicocarbonatite (Peterson 1983). Within the main intrusion, clinopyroxene grades from augite and hedenbergite in the early mafic phases and finally to Ti-augite and aegirine in the leucocratic, sodalite-nepheline, and altered syenites (Units 13–15) with increasing fractionation and alkalinity index [AI = Al−(K+Na)] (Currie 1975, Peterson & Currie 1994). In the Moose Creek Valley pegmatite, aegirine, along with natrolite, is a major primary and late-stage rock-forming phase and occurs throughout the paragenetic sequence. At the contact with the host rock, aegirine occurs with natrolite as a parallel growth of acicular, dark green crystals 1 cm wide, with crystals oriented perpendicular to the pegmatite margin (Fig. 4). Coarser, dark green prismatic crystals up to 1.5 long and 2 mm wide are found associated with phlogopite and natrolite in the more massive parts of the pegmatite; aegirine is a common inclusion within natrolite. However, the dominant habit of aegirine within the pegmatite is as fibrous, colorless to light green sprays (1–3 mm long), balls, and mats throughout the matrix and within the vugs (Fig. 6), a prime example of secondary aegirine common to many alkaline pegmatites (Piilonen et al. 1998, 2013). All aegirine crystals show patchy compositional zoning (Fig. 7). Aegirine compositions range from Ae₈₆Di₄Hd₁₀ to Ae₉₆Di₄Hd₁ (Table 3). Elevated Ti (0.05–0.12 apfu) and ^[6]Al (0.06–0.18 apfu) contents are noteworthy in the Moose Creek Valley aegirine. Aegirine has an average (48 analyses) composition of (Na_0.96Ca_0.01)_Σ0.97(Fe³⁺_0.68Al_0.11Fe²⁺_0.09Ti_0.09Mg_0.05)_Σ1.02(Si_1.99Al_0.01)_Σ2.00O₆.

Analcime and wairakite. Analcime, Na₂(Al₂Si₄O₆)·2H₂O, and the topologically identical Ca-analogue wairakite, Ca(Al₂Si₄O₆)·2H₂O, are common zeolite phases in low-temperature, water-rich assemblages associated with burial metamorphism of sedimentary rocks, alteration of mafic volcanic rocks, and hydrothermal systems in active and fossil geothermal areas. It also occurs as both a primary and post-magmatic phase in alkaline syenites, phonolites, and associated pegmatites, often associated with natrolite.

Analcime has been described from late syenites and in Unit 15 in the IRAC by Grice & Gault (1981) as white, opaque masses up to 8 × 7 cm with lamellar striations due to twinning. Similar subhedral to euhedral, white, opaque analcime up to 15 cm in width with the characteristic striations was found in the talus slopes below Zinc Mountain in 2023. Within the Moose Creek Valley pegmatite, analcime also occurs as subhedral to euhedral, white trapezohedra up to 3 cm (avg. 0.5 cm). Very fine-grained, white, sugary wairakite is often found in the cores of analcime, as well as along the contacts with natrolite and the pegmatite margin.

Natrolite at the IRAC has been described from carbonatites as an alteration product after nepheline and alkali feldspar in the syenites, interstitially and in late-stage segregations in the ultramafic rocks with andradite, pectolite, and titanite, in the altered zeolite-rich syenite (Unit 15), and from open, pheumatolytic pockets and seams associated with aegirine, analcime, ancylite-(Ce), calcite, catapleiite, edingtonite, and ilmenite (Currie 1975, Grice & Gault 1981, Peterson 1983, Grice et al. 1984, Locock 1994). Almost pure endmember, white to colorless, transparent to gemmy, often doubly terminated, prismatic natrolite crystals up to 16 × 3 × 3 cm long were found in calcite pockets within nepheline syenite by Grice & Gault (1981).

In the Moose Creek Valley pegmatite, natrolite is the main rock-forming mineral and occurs in all stages of crystallization. It occurs as massive to splintery, flattened, prismatic white to colorless crystals up to 25 cm long in the lower zone, often growing parallel to the pegmatite margin, associated with massive analcime, yellow, euhedral titanite, acicular aegirine, and coarse-grained calcite. In the more porous zones of the pegmatite, natrolite occurs as euhedral, white to colorless, transparent to gemmy prismatic crystals with complex terminations, often doubly terminated, as was observed in other parts of the IRAC by Grice & Gault (1981).

Calcite is a common primary and secondary phase throughout the IRAC, occurring in both the ultramafic and syenitic rocks and their associated residua, as well as being the dominant phase in the primary carbonatite and the carbohydrothermal dikes. Unlike many alkaline complexes which boast a wide range of REE and fluorcarbonate minerals, only ancylite-(Ce), barytocalcite, cebaite-(Ce), and strontianite have been described from the IRAC. With the exception of ancylite-(Ce), calcite is the only other carbonate mineral in the Moose Creek Valley pegmatite. As with the Ti minerals, it occurs in all parageneses. Calcite, along with natrolite, titanite, and aegirine, is among the first minerals to crystallize at the lower margins of the pegmatite where it occurs as euhedral, coarse-grained (up to 4 cm wide), pinkish-white to beige rhombs within natrolite.

Within the unaltered syenite units at the IRAC, biotite is the dominant mica, with an average composition of (K_0.90Na_0.13Ca_0.02)_Σ1.05(Fe²⁺_1.52Mg_1.08Ti_0.07)_Σ2.67(Al_1.18Si_2.75)_Σ3.93O₁₀(OH)₂ (Currie 1975). However, within the altered syenite adjacent to the pegmatite and within the pegmatite itself, phlogopite is the dominant mica species. It occurs as subhedral, reddish-brown, fine- to medium-grained crystals replacing ilmenite in the host syenite adjacent to the pegmatite. Margins between the ilmenite and phlogopite are often embayed, reflecting the replacement process. Inclusions of thorite (10–20 μm) are common within the phlogopite. In the pegmatite itself, phlogopite is found predominantly at the border adjacent to the host syenite as blackish-green, bronze to brown, subhedral crystals with ragged margins, associated with corroded aegirine. Minor very fine-grained phlogopite occurs in the vugs associated with sugary blue anatase, ancylite-(Ce), iridescent, doubly terminated ilmenite (0.25–0.75 mm), and powdery yellow balls of thorite (2–10 μm). The average composition (Table 4) of phlogopite from the Moose Creek Valley pegmatite is (K_0.90Na_0.01Ca_0.01)_Σ0.92(Mg_2.10Fe²⁺_0.78Ti_0.07Mn_0.01)_Σ2.97(Al_1.05Si_2.95)_Σ4.00O₁₀(OH)₂, significantly more Mg-rich than biotite from the main syenite units but similar in composition to phlogopite from other late-stage pegmatites (syenite, lamprophyre, and carbonatite) described by Mumford (2009).

Titanium mineralization in the Moose Creek Valley pegmatite is represented by a large suite of both primary and secondary Ti silicate and oxide minerals including titanite, anatase, brookite, ilmenite, lorenzenite, vinogradovite, lucasite-(Ce), and srilankite. Titanite is the primary Ti mineral within the pegmatite, while all other phases, including the TiO₂ polymorphs, are secondary, the result of either in situ replacement of primary titanite or later autometasomatism by residual fluids. Rutile, the most common TiO₂ polymorph, is absent.

Titanite, CaTi[SiO₄]O, is a ubiquitous accessory mineral in all the rock units at the IRAC with the exception of the central carbonatite (Unit 10), the nepheline-sodalite syenite (Unit 14), and the altered zeolite-rich syenite (Unit 15). Along with perovskite in early mafic units and the carbonatites (up to 10% by volume), it is the dominant Ti mineral in the complex. It is always associated with a Ti-bearing augite, apatite, and magnetite, with modal abundances up to 9% in the jacupirangite and ijolite and up to 5% in the syenite (Currie 1975). The modal abundance and habit of titanite evolves systematically from the earlier ultramafic to later syenitic units (Peterson 1983). In the pyroxenite and mela-ijolite, titanite forms thin rims on earlier perovskite, and, less commonly, euhedral, opaque, dark brown crystals 2–10 mm in length which are commonly twinned, whereas in ijolite, these dark brown crystals are further overgrown by a transparent, colorless to light yellow, untwinned titanite, resulting in anhedral, globular masses (Currie 1975). In the urtite and syenite, titanite occurs as transparent to translucent, light honey yellow, subhedral to euhedral, wedge-shaped crystals (Currie 1975, Peterson 1983).

Within the Moose Creek Valley pegmatite, titanite occurs as both a primary and secondary phase. It is one of the first minerals to crystallize in the paragenetic sequence along with aegirine, calcite, and natrolite and is present dominantly along the margins of the pegmatite in the more massive assemblage rather than in vugs. Primary titanite occurs as transparent to translucent, honey yellow to dark yellow, sharp, euhedral, wedge-shaped, often twinned, crystals which range in size from 0.5 cm to 6.5 cm (Fig. 8). In back-scattered electron images (BSEI), primary titanite displays complex compositional zoning with zoneless or minor concentrically zoned cores, which is overprinted by undulatory, irregular, and patchy zoning and fracturing toward the rims (Fig. 9). Primary, unaltered titanite is almost endmember (Table 5); F was analyzed for but is below detection limit. In general, cores of primary titanite are slightly enriched in Ti, Na, and Mn relative to rims, late-stage fractures, and altered patches, which are enriched in Ca, Fe, and Al (Fig. 10). Many of the primary titanite crystals have been partially or completely altered to anatase + calcite, occurring as pseudomorphs (Fig. 11).

Secondary titanite occurs as anhedral, fine-grained, opaque, light yellow balls coating altered anatase in vugs and as very fine-grained, granular casts of the original titanite, associated with late-stage, fibrous aegirine, drusy calcite, and ancylite-(Ce) (Fig. 12). Compositionally it is distinct from the primary titanite, containing consistently higher abundances of Fe (up to 0.08 apfu or 2.67 wt.% FeO) with an average composition of (Ca_0.96Na_0.03)_Σ0.99(Ti_0.89Fe_0.07Al_0.02)_Σ0.98Si_1.06O₅ (Fig. 10).

Anatase is the dominant TiO₂ polymorph within the Moose Creek Valley pegmatite, with brookite occurring as a trace mineral. Anatase occurs in four different habits and stages in the pegmatite: (Ana1) as a pseudomorph phase with calcite after titanite; (Ana2) as massive anhedral (1–5 mm) to euhedral (up to 1 cm wide) often etched and striated, grains acting as fracture-filling between more pegmatitic sprays of natrolite and nests of fibrous aegirine and throughout vuggy zones associated with aegirine, ancylite-(Ce), and late-stage Na-Ti silicates; (Ana3) as a secondary growth of equant, sky blue to light blue, euhedral crystals (0.01–0.05 mm) growing on top of Type 2 anatase, intergrown with late-stage Na-Ti silicates; and (Ana4) as a very fine-grained, sugary, sky blue coating on earlier anatase associated with aegirine, ancylite-(Ce), baryte, biotite, ilmenite, and thorite.

The pseudomorphs retain the typical wedge-shape cross section of the original titanite and range in size from 0.5 to 6.5 cm in width, with replacement of the titanite by anatase (±brookite) and calcite ranging from partial to complete (Fig. 11). The majority of the pseudomorphs show little to no volume loss through development of porosity or vugs; however, some have undergone late-stage autometasomatism and alteration of the anatase, resulting in open cavities and late crystallization of Ti, Zr, and REE minerals, including brookite and a second generation of euhedral, very fine-grained (0.05–0.01 mm), sky blue anatase and beige to yellow, very fine-grained rosettes of secondary titanite (Fig. 13).

Anatase within the pseudomorphs and in the bulk of the pegmatite occurs as massive anhedral to euhedral, fine to coarse-grained, dark blue-black to blue, opaque to translucent crystals with a luster that ranges from metallic to adamantine. It is associated with very fine-grained, interstitial, white, yellow or beige calcite and often crosscut by a second generation of natrolite. Within the pseudomorphs, anatase is typically irregularly intergrown with calcite and is embayed by secondary titanite and aegirine. Back-scattered electron imaging (BSEI) reveals minor, subtle compositional zoning in anatase, with bright areas enriched in Fe and/or Nb.

Brookite is an accessory phase associated with anatase. In contrast to anatase, brookite is green-brown to red-brown in color, transparent to translucent, with a vitreous to submetallic luster. It occurs as either anhedral to subhedral, blocky crystals up to 1.5 mm in width embedded in natrolite, coexisting and often intergrown with anatase, or as isolated, euhedral, elongate to slightly stubby, complex bipyramidal crystals, on silky nests of lorenzenite and associated with Type 3 anatase (Fig. 13b).

Compositionally, both anatase and brookite are close to their ideal formula of TiO₂ with only minor variations observed in Fe, Nb, and Al contents (Table 6).

Ilmenite is found in both the altered host rock and as a very late-stage mineral in the pegmatite within the vug assemblages associated with aegirine, ancylite-(Ce), baryte, lucasite-(Ce), thorite, and secondary titanite, and as inclusions in calcite with anatase and lorenzenite (Fig. 14). It occurs as euhedral, prismatic, doubly terminated, iridescent black crystals averaging 0.8 mm long (Fig. 15). The average composition is (Fe²⁺_0.87Mn_0.12)_Σ0.99TiO₃ with Mn contents up to 0.35 apfu (Table 6).

Lorenzenite, Na₂Ti₂Si₂O₉, and vinogradovite, Na₄Ti₄(Si₂O₆)₂[(Si,Al)₄O₁₀]O₄·(H₂O,Na,K)₃, are important titanosilicate indicator minerals in both miaskitic and agpaitic alkaline rocks. Both minerals typically occur in late-stage or hydrothermal assemblages in pegmatites, miarolitic cavities, or veins associated with syenites from the important alkaline complexes worldwide, including Mont Saint-Hilaire, Québec, Canada (Horváth et al. 2019); the Larvik Plutonic Complex, Norway (Larsen et al. 1992, Larsen 2010, Andersen et al. 2013); the Khibina and Lovozero massifs, Russia (Pekov 2000, Arzamastsev et al. 2008); and Narsarssuk and Ilímaussaq, Greenland (Flink 1901, Karup-Møller 1986). Lorenzenite and vinogradovite are often found in close association and have been reported as homoaxial pseudomorphs of vinogradovite after lorenzenite from Lovozero in Russia (Pekov 2000). In the Moose Creek Valley pegmatite, lorenzenite is the dominant species, with vinogradovite less common. Lorenzenite occurs as very fine-grained (up to 0.2 mm long), fibrous, silky, white sprays and mats in cavities associated with aegirine, anatase, brookite, wadeite, and rarely with ancylite-(Ce) (Fig. 16). Vinogradovite occurs as bladed to fibrous, pearly, white to pale gray, anhedral to subhedral crystals up to 0.5 mm in length associated with lorenzenite, aegirine, anatase, brookite, and wadeite. In the Moose Creek Valley pegmatite, lorenzenite is commonly replaced by vinogradovite epitaxially (Fig. 17), similar to what is observed at Lovozero (Pekov 2000).

Compositionally, lorenzenite is almost pure endmember, with only minor concentrations of K (0.00–0.02 apfu), Ca (0.01 apfu), Hf (0.00–0.06 apfu), Fe²⁺ (0.01–0.02 apfu) and Al (up to 0.05 apfu). The average (33 analyses) formula is (Na_1.90Ca_0.01)_Σ1.91(Ti_2.00Hf_0.01Fe²⁺_0.01)_Σ2.03Si₂O₉ (Table 7) Vinogradovite has a wider range in compositions with K = 0.36–0.55 apfu, Nb = 0.00–0.02 apfu, Fe²⁺ = 0.04–0.13 apfu, Mg = 0.00–0.03 apfu, and Al = 1.27–1.46 apfu. Potassium contents exceed what is required to fill the Naϕ7 polyhedra site and therefore excess K (0.44 apfu) is partitioned into the channel OW(2) site with H₂O (Rastsvetaeva et al. 1968, Kalsbeek & Rønsbo 1992). The average formula of IRAC vinogradovite (22 analyses) is (Na_3.97K_0.41Ca_0.01)_∑4.00(Ti_3.85Fe²⁺_0.08Mg_0.01)_∑3.96(Si_6.66Al_1.34)_∑8.00·[(H₂O)_2.56K_0.44]_∑3.00 (Table 7).

A number of extremely rare Ti oxide minerals have been found within the Moose Creek Valley pegmatite, including henrymeyerite, Ba(Ti₇Fe²⁺)O₁₆, a member of the priderite group; lucasite-(Ce), CeTi₂(O,OH)₆; and srilankite, ZrTi₂O₆, a member of the samarskite group. The IRAC is the first known Canadian locality for lucasite-(Ce) and srilankite. It is also the first known locality in the Americas for lucasite-(Ce).

In the Moose Creek Valley pegmatite at the IRAC, both henrymeyerite (Fig. 18) and srilankite (Fig. 19) are late-stage phases that occur as 50–100 μm, anhedral grains associated anatase, brookite, calcite, and ilmenite in the titanite pseudomorphs. The average composition of henrymeyerite is Ba_1.04(Ti_6.94Fe²⁺_1.10V_0.03)O₁₆, and (Zr_0.99Si_0.01)_Σ2.00(Ti_1.96Ca_0.04Fe²⁺_0.01)_Σ2.01O₆ for srilankite, both almost ideal endmember compositions (Table 6).

Lucasite-(Ce) at the IRAC was found in a single assemblage dominated by pegmatitic calcite (6 cm wide), massive sprays of splintery natrolite (up to 10 cm long), with euhedral analcime being replaced by wairakite near the lower pegmatite margin. It occurs as euhedral, tabular orange crystals (0.02 × 0.1 mm) in radiating, fan-like clusters up to 0.4 mm wide alone or growing epitaxially on prismatic, doubly terminated ilmenite in vugs on natrolite, often associated with light green ancylite-(Ce) (Fig. 20). Very fine-grained, granular brown balls of thorite often occur growing on the terminations of the lucasite-(Ce) crystals.

Analysis of the IRAC lucasite-(Ce) by EMPA results in consistently low totals (87–93 wt.%; Table 8). This is potentially a result of the fine grain size and extreme thinness of the crystals, but may also suggest the possibility of the presence of OH⁻, as suggested by Nickel et al. (1987). Calculation of empirical bond valences for the type lucasite-(Ce) indicated a mixture of trivalent and tetravalent cations in the 8-coordinated Ce site (3.74 v.u.), a low valence sum at the O(2) site (1.62 v.u.), which, along with the low analytical totals (98.38 wt.%), led Nickel et al. (1987) to conclude that there is OH⁻ substitution for O in the structure. As such, lucasite-(Ce) formulae were calculated on the basis of 2 Ti pfu and 6 O, with H₂O calculated by difference. All REE were assumed to be trivalent. Compared to the type material, IRAC lucasite-(Ce) is enriched in Pb (0.02–0.14 apfu), Th (0.14–0.25 apfu), Nd (0.14–0.18 apfu), Fe (0.07–0.20 apfu), Ca (0.11–0.17 apfu), and Si (0.36–0.47 apfu) and depleted in La (0.03–0.04 apfu). The ∑REE ranges from 0.42 to 0.53 apfu, with the remainder of the site occupied by Th, Pb, and Ca. The average composition is (Ce_0.22Th_0.18Nd_0.16Ca_0.14Pb_0.07Pr_0.04La_0.03Sm_0.02K_0.01)_Σ0.88(Ti_1.37Si_0.44Fe_0.11Al_0.07Zr_0.01Mn_0.01)_Σ2.00O_4.27(OH)_1.73. Further work, including single-crystal X-ray diffraction, on the crystal chemistry of lucasite-(Ce) is underway to better understand the anionic and cationic distribution in the structure.

Ancylite-group minerals are common in syenites, carbonatites, and carbohydrothermal veins as late-stage minerals associated with Ba-REE carbonates, REE fluorocarbonates, strontianite, and baryte (Knudsen 1991, Ngwenya 1994, Wall & Mariano 1995, Cooper 1996, Zaitsev et al. 1998, Al Ani & Sarapää 2013, Dalsin et al. 2015). Among alkaline deposits worldwide, ancylite-group mineral compositions indicate significant flexibility in terms of Ca, Sr, and REE contents (Flink 1898, Orlandi et al. 1990, Pekov et al. 1997, Zaitsev et al. 1998, Reguir & Mitchell 2000, Petersen et al. 2001, Piilonen et al. 2022, Wang et al. 2023).

Ancylite-(Ce) in the Moose Creek Valley pegmatite is an extremely late phase in the crystallization sequence. It is the only carbonate phase other than calcite in the pegmatite, and, along with lucasite-(Ce), the only REE-dominant mineral in the assemblage. It occurs as euhedral, stubby, twinned dipyramidal crystals, singly and in clusters, which vary from translucent and gemmy to frosted and opaque, with a wide range of colors from colorless to pale violet, cream yellow, light green, yellow-green, yellow, brown, and light pink to dark pinkish-orange. Crystals range in size from 50 μm to 1.5 mm and are associated with fibrous aegirine, etched anatase, ilmenite, lorenzenite, and thorite (Fig. 21). Prismatic sprays of aegirine and anhedral thorite grains (1–90 μm) are common inclusions within some samples, dominantly those which display green, yellow, or pink-orange coloration. In back-scattered electron images, those grains containing inclusions have extremely diffuse, mottled zonation suggesting a degree of alteration. Samples which are generally inclusion-free show primary, concentric, and sector zonation, the result of Ca ↔ Sr and Ce ↔ La ↔ Nd substitutions.

All formulae have been calculated on the basis of 2 cations + 2 CO₃ as per the ideal ancylite composition, REESr(CO₃)₂(OH)·H₂O. However, when formulae are calculated on this basis, a wide range of Sr/REE ratios is observed in natural samples, including those from Moose Creek Valley. Powder X-ray diffraction was used to eliminate the possibility of other REE carbonate impurities or admixtures (i.e., bastnäsite) that would complicate calculation of the ancylite formula.

In order to account for these variations, the idealized ancylite-group formula, determined by Dal Negro et al. (1975), is given as REE_x(Sr, Ca)_2−x(CO₃)₂(OH)_x·(2−x)·H₂O, controlled by a coupled substitution of Sr²⁺ + H₂O ↔ REE³⁺ + OH⁻. On this basis, samples from Moose Creek Valley have x values that range from 0.93 to 1.34, with (Sr+Ca+Th+Ba+Cs) values from 0.65 to 1.07. Only ancylite from the Pivot Creek ferrocarbonatite have higher x values of 1.48, suggesting an upper limit x ∼ 1.5 for the REE ↔ (Sr + Ca) solid solution. The degree of REE substitution within the ancylite group is a significant issue and has implications for ordering within the structure.

Although a wide range of colors of ancylite are observed within the Moose Creek Valley pegmatite, compositionally (Table 9) all are ancylite-(Ce); color cannot be correlated to chemistry for any of the samples. Samples show limited solid solution with calcioancylite, with only minor Ca (0.03–0.26 apfu Ca), with minor Ba (0.00–0.05 apfu) and Th (0.00–0.08 apfu). Cerium, La, and Nd contents span a wide range with Ce = 0.41–0.70 apfu (avg. 0.53 apfu), La = 0.11–0.56 apfu (avg. 0.25 apfu), and Nd = 0.08–0.41 apfu (avg. 0.24 apfu). The average composition is (Ce_0.53La_0.25Nd_0.24Pr_0.06Sm_0.01Gd_0.01)_Σ1.11(Sr_0.69Ca_0.16Ba_0.02Th_0.02Cs_0.01)_Σ0.89(CO₃)₂(OH)·H₂O.

Two compositional populations are observed (Fig. 22): (1) analyses with Sr/Ca > 5, ∑REE > 1.05, and La > Nd, represented by cores of zoned crystals and those which are generally inclusion-free, and (2) analyses with Sr/Ca < 5, ∑REE < 1.05, and Nd > La, represented by rims of zoned crystals and those samples containing abundant aegirine and thorite inclusions. The cores and inclusion-free crystals display a much wider range in composition than the rims. A plot of Nd versus Nd/La (Fig. 23) reveals a negative trend with a change in slope at Nd/La ∼ 1 and La = 0.25 apfu, the division between core and rim compositions. Although some of the primary zonation patterns are overprinted by patchy diffusion patterns, the general trend from core to rim is that of decreasing ∑REE and Sr/Ca with increasing (Ca+Sr+Th+Ba+Cs) and Nd/La. Similar Nd-rich ancylite-group minerals have been found in ferrocarbonatite at Pivot Creek, New Zealand (Cooper et al. 2015), and albitite at Ilímaussaq, Greenland (Petersen et al. 2001). Figure 24 shows the compositional range of Ce-La-Nd for ancylite from various alkaline localities. Ancylite-(Ce) from Moose Creek Valley are enriched in Nd relative to those from a syenite-dominant alkaline complex such as Mont Saint-Hilaire, Québec, but do not show the extreme Nd-enrichment observed at Haast River, New Zealand, or at Ilímaussaq, Greenland. Most ancylite-group mineral analyses from the literature, whether from a syenite or carbonatite complex, tend toward La > Nd. Figure 25 shows the relationship between Sr/Ca and Nd/Ce for Moose Creek Valley ancylite-(Ce) compared to those from the literature.

Fluorine contents are below detection limit, unlike ancylite from many alkaline complexes which have substantial F contents including from Mont Saint-Hilaire, Québec, with up to 0.52 apfu F (Piilonen, unpublished data); Mount Mather Creek, British Columbia, with up to 0.24 apfu F (Piilonen et al. 2022); 0.14–0.42 apfu F in ancylite-group minerals from carbonatites and hydrothermalites at Khibina, Russia (Zaitsev et al. 1998, Belovitskaya et al. 2002); 0.22 apfu in ancylite-(Ce) from quartz-calcite-chlorite veins from the Opava Mountains, Poland (Janeczek et al. 2020); and 0.15 apfu F in ancylite-(La) from the Ilímaussaq complex, Greenland (Petersen et al. 2001).

Nordstrandite and gibbsite are two of four naturally occurring Al(OH)₃ polymorphs, along with doyleite and bayerite. They are low-temperature alteration products, found in bauxite and laterite deposits, in dolomitic oil shales, and as late-stage minerals in pegmatites and miarolitic cavities in peralkaline syenites (Chao & Baker 1982, Chao et al. 1985, Kovács-Pálffy et al. 2008). Of all the Al(OH)₃ polymorphs, nordstrandite is the rarest in nature, however it is the most common polymorph found in SiO₂-undersaturated alkaline intrusions worldwide, including at Mount Mather Creek, British Columbia (Piilonen et al. 2022); Saint-Hilaire and the Saint-Amable sill, Québec (Chao & Baker 1982, Horváth 2010); the Princess Sodalite Mine, Bancroft, Ontario (Sabina 1982); the Cerro Sapo deposit, Ayopaya province, Brazil (L. Hováth, pers. commun.); Narssârssuk, Greenland (Petersen et al. 1976); the Larvik Plutonic Complex, Norway (Larsen 2010); and the Khibina, Kovdor, and Lovozero massifs, Russia (Pekov 2000, Yakovenchuk et al. 2005).

At Moose Creek Valley, nordstrandite and gibbsite are extremely late-stage products of the breakdown of analcime and natrolite. They occur as very fine-grained, white, anhedral, powdery masses associated with gonnardite, wairakite, and trace sodalite.

Wadeite, K₂ZrSi₃O₉, and catapleiite, Na₂Zr(Si₃O₉)·2H₂O, are both members of a group of cyclosilicates whose structures can be described as having three-membered [Si₃O₉]⁶⁻ rings linked to isolated M⁴⁺O₆ octahedra (where M = Zr, Ti, Sn). Large alkali and alkaline-earth cations occupy channels formed by the pinwheel-like rings. In wadeite (P6 or P6₃/m), K is positioned along the threefold axis within the rings, whereas in catapleiite, the smaller cation Na is shifted off the threefold axis, lowering the overall symmetry to monoclinic or orthorhombic.

Wadeite in the Moose Creek Valley pegmatite occurs in the late-stage vug assemblage associated with secondary aegirine, ancylite-(Ce), anatase, lorenzenite, vinogradovite, and secondary titanite. It occurs as colorless, pale beige, pale pink, or white hexagonal plates with beveled edges, singly or in rosettes, often speared by aegirine (Fig. 26). The rims are generally transparent whereas cores tend toward being cloudy and translucent. Crystals range in size from 0.1 to 0.9 mm in diameter. In BSEI, catapleiite, Na₂Zr(Si₃O₉)·2H₂O, was observed replacing wadeite along fractures starting at the rim and in some cores, similar to what has been observed in nepheline syenites in the Saima alkaline complex, China (Wu et al. 2015).

The powder X-ray diffraction (PXRD) pattern for the wadeite at Moose Creek Valley matches that of other wadeite samples (Table 10), however the EMPA data gave low totals (∼91 wt.%), and only half the K₂O (10 wt.%) required by the ideal wadeite formula (24 wt.%) (Table 11). Further examination of the material by both PXRD and single-crystal X-ray diffraction (SCXRD) indicate the mineral to be wadeite, but with a mixed large cation site. Work is ongoing to elucidate the structure of the IRAC wadeite.

The formulae for both wadeite and catapleiite were calculated on the basis of 3 Si pfu (Table 11). The low totals for wadeite, plus the fact that the mineral dessicates dramatically under the electron beam, indicate the presence of H₂O or OH⁻ in the K site, or H₂O in the channels similar to catapleiite. The average composition of wadeite is either (K_0.81Ca_0.09Na_0.06□_1.04)_Σ2.00(Zr_1.02Ti_0.01Al_0.01Fe_0.01)_Σ1.05Si₃O₉·nH₂O, or [K_0.81Ca_0.09Na_0.06(H₂O, OH⁻)_1.04]_Σ2.00(Zr_1.02Ti_0.01Al_0.01Fe_0.01)_Σ1.05Si₃O₉. The average composition of catapleiite is (Na_1.36K_0.11Ca_0.02□_0.51)_Σ2.00(Zr_1.06Ti_0.01Fe_0.01)_Σ1.08Si₃O₉·2H₂O. Detailed single-crystal XRD, Rietveld analysis, and Raman/FTIR spectroscopy is currently in progress to determine the structure and anionic composition of the K-deficient wadeite.

Mixed silicate-carbonate carbohydrothermal dikes, sudations, and breccias have been noted across the British Columbia Alkaline Province including at the Aley carbonatite, the Wicheeda Lake alkaline complex, the Kechika River area, the Ice River Alkaline Complex, and Mount Mather Creek (Pell 1994, Piilonen et al. 2022). These late-stage bodies result from fluids exsolved from primary rocks within the complexes (ijolite, syenite, and carbonatite) facilitating the concentration and/or remobilization of high field strength elements (HFSE), REE, Ba, Sr, Na, and C.

Within the IRAC, syenite, lamprophyre, and carbohydrothermal dikes occur both internally and externally to the main complex units. Syenite dikes range in width from 10 cm to >5 m, whereas lamprophyre dikes, often with chill margins, rarely exceed 2 m in width (Currie 1975, Mumford 2009). Carbohydrothermal dikes are less common, varying from 10 cm to >1.25 m in width. Their lateral exposure within the IRAC varies from a few meters to more than 200 m, with many terminating or obscured by talus slopes.

On the eastern side of the IRAC, syenite dikes run subparallel to the margins of the complex, trending NNW–NW, consistent with the regional tectonic fabric, and are concentrated toward the roof of the intrusion, cross-cutting the layered mafic subcomplex, the syenite subcomplex, and the Ottertail Formation (Currie 1975, Mumford 2009). Compositional similarities between the syenite and lamprophyre dikes and the main complex units suggest they were derived by periodic tapping of the same fractionating magma chamber (Mumford 2009).

Carbohydrothermal dikes are concentrated in the layered mafic rocks and are typically fine-grained to pegmatitic (Mumford 2009). Mumford (2009) classified these dikes into three types based on dominant mineralogy: (1) a phlogopite-monazite-alkali feldspar-bearing ferrocarbonatite, (2) a pegmatitic calciocarbonatite, and (3) a phlogopite-alkali feldspar-bearing calciocarbonatite. Age dating and Sm-Nd and Rb-Sr isotopic ratios for the ferrocarbonatite dike are distinctly different than those for the calciocarbonatite, syenite, and lamprophyre suite, leading Mumford (2009) to suggest a separate source and injection event, possibly linked to the formation of the silico-carbohydrothermal breccia at Mount Mather Creek (Piilonen et al. 2022).

Isotopic and trace element ratios for the syenite, lamprophyre, and calciocarbonatite dikes are virtually identical to those for the ultramafic, syenite, and carbonatite rocks of the main complex. This similarity suggests that the entire complex, including the late-stage dikes, formed via fractional crystallization of an alkali basalt magma derived from a depleted mantle (Locock 1994, Mumford 2009). Local mineralogical variations and unique assemblages are attributed to crustal contamination by host sediments and late-stage metasomatism.

The Moose Creek Valley pegmatite is mineralogically unique and not represented in the dike suites described by either Currie (1975) or Mumford (2009). Table 1 lists the paragenetic sequence and complete list of identified minerals. Based on textural relationships and chemical composition of the phases, there are four stages of mineralization: (1) fracturing and pegmatite emplacement via exsolved silico-carbohydrothermal fluids enriched in H₂O and CO₂; (2) primary magmatic pegmatite assemblage dominated by natrolite, analcime, titanite, calcite, and aegirine; (3) autometasomatism of the primary assemblage to form a secondary assemblage of anatase, calcite, aegirine, natrolite, and brookite; (4) destabilization of anatase and remobilization of Ti and Zr to form late-stage titano- and zirconosilicates; and (5) late-stage Ca-autometasomatism to form a true carbohydrothermal assemblage:

1. pegmatite emplacement

2. natrolite + analcime + aegirine + calcite + titanite ± phlogopite

3. anatase + brookite + calcite + wairakite

4. natrolite + aegirine + vinogradovite + lorenzenite + catapleiite + wadeite + srilankite

5. secondary titanite + calcite + ilmenite + nordstrandite + gibbsite + ancylite-(Ce) + baryte + henrymeyerite + lucasite-(Ce) + thorite

The primary assemblage (2) of pegmatitic natrolite, analcime, aegirine, calcite, and titanite indicates an initial fluid enriched in Na-Ca-Fe-Ti, CO₂, and H₂O. Additionally, it must have been able to scavenge and remobilize Ti from the earlier rock units to produce a pegmatite containing titanite and TiO₂ polymorphs as part of the rock-forming assemblage. Further, composition of the bulk assemblage remains consistently dominated by the elements Na-Ca-Fe-Ti-CO₂-H₂O throughout the proposed crystallization sequence, suggesting that either (1) the entire sequence resulted from the crystallization of a single initial fluid crystallizing as a closed system or (2) fluids that were introduced subsequently did not vary significantly in composition. The observed systematic decrease in the volume of crystallizing Ti-bearing minerals, with titanite and TiO₂ polymorphs in steps (2) through (4) being the most abundant, however, suggest the compositional evolution of a single fluid in a close system (i.e., scenario 1). Thus, observed alteration of previously crystallized material likely results from autometasomatic processes.

In most geologic settings, Ti and the other HFSE like Zr, Y, and REE, are considered immobile. Therefore, they are often used as indicators for classifying rocks, tracing their petrogenesis, or quantifying fluid-rock interactions (Salvi et al. 2000). However, under certain conditions, it has been shown that HFSE can be mobile and can be transported and concentrated by hydrothermal fluids. This occurs during processes such as alkali-metasomatism, fenitization and carbonatization in alkaline complexes (Salvi & Williams-Jones 1990, Salvi et al. 2000, Linnen et al. 2014, Kozlov et al. 2018), albitization or weathering in granites (Cathelineau 1987, Du et al. 2012), weathering of sediments by hydrocarbon-rich fluids (Liu et al. 2019), formation of skarns (Gieré 1986, Gieré & Williams 1992), fluids exsolved during anatexis of subducted continental crust (Gao et al. 2007, Chen et al. 2022), and in alteration zones within metallogenic deposits (Schandl & Gorton 1991).

In most alkaline systems, the transport of HFSE is attributed to complexing with mixed ligands (CO₃²⁻, HCO₃⁻, F⁻, OH⁻, HS⁻, and HPO₄²⁻) within an exsolved hydrothermal fluid or brine (Salvi & Williams-Jones 1990, Aja et al. 1995, Salvi et al. 2000, Piilonen et al. 2006, Ryzhenko et al. 2006, Migdisov et al. 2011, Linnen et al. 2014, Kozlov et al. 2018). These brines are responsible for scavenging HFSE from primary phases, including pyroxenes, biotite, amphibole, titanite, rutile, and zircon, and redistributing them as secondary phases. Fluid inclusions from the Tamazeght alkaline complex in Morocco have high salinity (25 wt.% NaCl equivalent), are Ca-rich, and contain several HFSE daughter minerals, including parasite, zircon, a Ti-Nb silicate, and Mn hydroxide, all evidence for extensive mobilization of HFSE during hydrothermal alteration (Salvi et al. 2000). In most cases, interaction with a Na-Ca-rich host rock or second exsolved fluid is required to cause rapid cooling of the brine, saturation of the brine with Ca resulting in deposition of fluorite, and subsequent formation of secondary HFSE minerals. Notably absent from the Moose Creek Valley pegmatite are fluorite, fluorapatite, and REE fluorocarbonates—minerals containing essential F typically found in hydrothermally altered alkaline pegmatites and carbohydrothermal dikes. In addition, F is absent in minerals such as ancylite-(Ce) and titanite. This suggests that the system as a whole is devoid of F and that Ti mobilization via F-complexation did not occur. Instead, other ligands, particularly OH⁻, must have been responsible (Salvi et al. 2000). Studies on the solubility of rutile under supercritical, hydrothermal conditions (500 °C/1 kbar) by Ryzhenko et al. (2006) have shown that rutile solubility, and the possibility for Ti mobility, increases linearly with NaOH concentration, supporting the case for Ti-OH complexation. Ayers & Watson (1993) also found the formation of Ti(OH)₄ complexes, resolution from the dissolution of rutile, under supercritical conditions (800–1200 °C/1–3 kbar) to be a viable method of Ti transport.

Currie (1975) proposed that the altered zeolite syenite formed from the mixing of a hydrous, alkali-rich fluid exsolved from the main carbonatite, with a residual, H₂O-undersaturated syenitic magma. It is likely that the Moose Creek Valley pegmatite has a similar origin. However, despite mineralogical similarities to Unit 15, the Moose Creek Valley pegmatite is unique, with a more extensive, distinct assemblage dominated by Na, Ca, Fe, Ti, Zr, and Ce minerals. It notably lacks feldspar, sodalite, nepheline magnetite, hematite, and sulfides, common in Unit 15. The pegmatite also exhibits sharp contacts with the host syenite, contrasting with Unit 15’s gradational contacts with surrounding sodalite and nepheline syenite, and contains pockets and seams indicative of late-stage, localized hydrothermal activity. While Currie (1975) suggested the presence of a syenitic magma for Unit 15, the Na-Ca-Ti-Fe-CO₂-H₂O-dominant primary mineral assemblage in the Moose Creek Valley pegmatite indicates an interaction between exsolved SiO₂-undersaturated, alkali-rich fluids from the ultramafic layered subcomplex and CO₂-rich carbohydrothermal fluids from the carbonatite.

The layered ultramafic subcomplex rocks and the early syenites in the IRAC share a common geochemical signature—enrichment in Ti. As noted earlier, titanite is a ubiquitous accessory mineral in all the rock units at the IRAC except the central carbonatite (Unit 10), the nepheline-sodalite syenite (Unit 14), and the altered zeolite-rich syenite (Unit 15). Alongside perovskite in early mafic units and the carbonatite (up to 10% by volume), titanite is the dominant Ti mineral in the complex. Additionally, many of the primary rock-forming minerals are enriched in Ti, including augite (up to 5 wt.%), phlogopite/biotite (3 wt.%), magnetite (18 wt.%), and titanian andradite (19 wt.%). The Ti garnet schorlomite, Ca₃Ti₂(SiO₄)(Fe³⁺O₄)₂, is locally abundant in the ultramafic melteigite.

The mineralogy and geochemistry of the ultramafic layered subcomplex suggests decreasing fO₂, aH₂O, and aCO₂ and increasing Fe with progressive fractional crystallization (Currie 1975). Exsolved fluids from the ultramafic rocks would therefore be enriched in OH⁻, CO₃²⁻, and alkalis and depleted in Fe and Ti. Peterson (1983) suggested that liquid immiscibility between a Fe-Ti-rich liquid and a Fe-Ti-poor, alkali-rich liquid is responsible for the presence of localized, rare, oscillatory intergrowths of Ti-andradite alternating with Ti-poor andradite-grossular, pectolite, natrolite, and cancrinite in the ijolite units. This further supports evidence for extensive exsolved volatile- and alkali-rich fluids arising from the ultramafic subcomplex. Traveling along the extensive NNW–NW fracture system within the complex, the exsolved fluids would have had ample opportunity to mobilize Ti from the main complex rocks, particularly from primary titanite, Ti-bearing mafic rock-forming minerals and perovskite. Titanite readily undergoes hydrothermal alteration in the presence of CO₃²⁻ and F, forming secondary TiO₂ polymorphs, calcite, and quartz (Hollabaugh et al. 1989). If the initial fluid was enriched in OH⁻ relative to CO₃²⁻, Ti from the ultramafic rocks would remain in solution as a Ti-OH complex until the fluid experienced cooling or mixed with another fluid, initiating crystallization.

At Moose Creek Valley, the fluid responsible for the Ti-dominant mineral assemblages including the rare species henrymeyerite, srilankite, and lucasite-(Ce), likely resulted from the mixing of exsolved fluids from the carbonatite and the ultramafic layered subcomplex. In the pegmatite, the relative activities of OH⁻ and CO₃²⁻ (aCO₃²⁻/aOH⁻), the Na/Ca ratio, and decreasing T dictated the mineral assemblage during the crystallization history of the pegmatite—essentially resulting in alternating cycles of Na- and Ca-autometasomatism (Table 12).

During the initial stages, we predict that the aCO₃²⁻/aOH⁻ and Na/Ca ratios were slightly less than 1, resulting in a primary assemblage of hydrous, Si-deficient zeolites (analcime and natrolite), titanite, calcite, and aegirine. Decreasing temperature and increasing relative aCO₃²⁻ resulted in the alteration of primary titanite to TiO₂ polymorphs (Fig. 11), calcite, and H₄SiO₄ and increasing the relative aCa within the system. Increased aCa also resulted in the localized transformation of analcime to sugary, fine-grained wairakite. The presence or absence of analcime, wairakite, or any of the other zeolite phases is controlled by a complex combination of parameters which include temperature, pH, cation concentrations, aSiO₄⁴⁻, composition of the groundwater or fluid, porosity and permeability of the host rock, and water:rock ratios, as well as composition of the protolith or host rock (Browne 1978, Chipera & Apps 2001, Weisenberger & Selbekk 2009, Spürgin et al. 2019). At Moose Creek Valley, all these factors may have played a role.

Anatase, the low T TiO₂ polymorph, is stable in acidic conditions at temperatures less than 50 °C (Hollabaugh et al. 1989). With increasing pH and at moderate alkalinity, anatase becomes stable at temperatures between 200 and 400 °C (Keesman 1966, Lerz 1968). With increased aNa and higher alkalinities, anatase becomes unstable and will form either Na-bearing titanites or Na-titanosilicates (Fig. 27) (Keesman 1966, Lerz 1968).

Khomyakov (1995) alkalinity modulus rating, which is based on the typomorphism of rare-metal minerals with the general formula AxMySipOq, where A = Na, K and M = Ti, Zr, Nb, Be, Al, can be used to categorize nepheline syenites and their derivatives into four types based on their alkalinity. The alkalinity modulus is calculated based on the atomic proportions such that K_alk = (x*100)/(x+y+p). Minerals with K_alk < 15%, such as ilmenite and allanite, are representative of miaskitic assemblages, where as low agpaitic rocks are characterized by minerals with K_alk = 15–25%, including eudialyte, låvenite, titanite, and pyrochlore. Vinogradovite, with K_alk = 25%, is representative of only moderately agpaitic environments, whereas lorenzenite, with a K_alk = 33%, is indicative of highly agpaitic environments. Within the Moose Creek Valley pegmatite, anatase in the pseudomorphs has been destabilized due to Na-autometasomatism (NaOH), resulting in a late-stage assemblage of lorenzenite, vinogradovite, and srilankite, along with the Zr minerals catapleiite and wadeite. The epitaxial relationship of vinogradovite after lorenzenite (Figs. 16, 17) in the pegmatite also supports decreasing Na and Ti during this stage of the paragenesis. The presence of these late-stage minerals testifies to the unique Ti-rich nature of the fluid throughout all stages of crystallization and the extensive mobilization of Ti and Zr by those fluids.

Further Ca-autometasomatism in the last stages of the paragenesis may have resulted in the formation of secondary, fine-grained titanite (Fig. 12), calcite, gibbsite, and nordstrandite, followed by a carbohydrothermal assemblage of ancylite-(Ce), baryte, henrymeyerite, srilankite, ilmenite, lucasite-(Ce), and thorite.

The presence of the Al(OH)₃ polymorphs nordstrandite and gibbsite in the late stages of crystallization supports a SiO₂-undersaturated alkaline melt enriched in OH⁻ and suggests remobilization of Al from earlier phases (analcime or natrolite). Nordstrandite is typically the last mineral to form in the paragenetic sequence in these complexes. Studies of the alumina–water system have been limited to ambient pressures and temperatures in aqueous systems which mimic conditions in soils (Barnhisel & Rich 1965, Violante & Huang 1993) and as such little is known about the stability and conditions of formation of the Al(OH)₃ polymorphs in alkaline igneous environments. While there is still debate about the exact conditions under which the Al(OH)₃ polymorphs form in geological conditions, and relationships in the Al–O–H system in general, it is generally agreed that nordstrandite is the high pH polymorph (7.5–9), is unstable at high SiO₄⁴⁻ activities, and requires an intermediate rate of dissolution and nucleation of Al-OH polymers (Violante & Huang 1993, Kawano & Tomita 1996, Ramesh 2012). Furthermore, evidence suggests the presence of acidic fluids with high aF⁻ impedes nucleation of Al-OH polymers and growth of nordstrandite (Peskleway et al. 2005). This supports the observation that F is almost absent at Moose Creek Valley and the conclusion that OH⁻ and CO₃²⁻ were the main ligands within the crystallizing melt.

The presence of minerals such as ancylite-(Ce), calcite, and the rare Ti minerals henrymeyerite, srilankite, and lucasite-(Ce) is indicative of crystallization from a carbohydrothermal fluid enriched in REE, Ti, Ca, Ba, and Sr (Zaitsev et al. 1998). Ancylite-(Ce) has only previously been noted at the IRAC in natrolite pods within Unit 15 (Grice & Gault 1981); no other REE-bearing minerals have been noted in the IRAC. Recent studies have shown REE to be highly soluble and easily transported long distances in alkali (Na)-rich carbonatitic fluids, with anionic ligand species (Cl⁻, F⁻, or CO₃²⁻) being less important (Anenburg et al. 2020). In contrast, REE have low mobility in alkali-free carbonatitic melts and are hosted in insoluble REE carbonate and phosphate phases within the carbonatite body itself, rather than in external dikes. Ancylite is commonly used as an indicator of carbohydrothermal activity, crystallizing at temperatures <350 °C (Wall & Zaitsev 2004). The lack of F in ancylite-(Ce) from Moose Creek Valley reflects the paucity of F within the carbohydrothermal fluid, suggesting other ligands are the dominant species along with CO₃²⁻ for element transfer.

Henrymeyerite, srilankite, and lucasite-(Ce) are also considered indicator minerals for carbohydrothermal fluid activity. Henrymeyerite was originally discovered in carbonatite veins at the Kovdor ultramafic alkaline complex, Kola Peninsula, Russia (Mitchell et al. 2000). It has since been found at other alkaline complexes, dominantly in carbonatite or carbohydrothermal veins, including the Khibiny massif, Kola Peninsula, Russia (Mikhailova et al. 2006); Gordon Butte, Crazy Mountains, Montana, USA (Chakhmouradian & Mitchell 2002); the Seblyavr massif, Russia (Sorokhtina 2000); Šebkovice, Czech Republic (Krmíček et al. 2011); and the Ilímaussaq complex, South Greenland (Cegiełka et al. 2022). It has also been found in eastern Antarctica in lherzolite and harzburgite xenoliths which have undergone extensive carbonate metasomatism (Kogarko et al. 2007). As with srilankite, it appears to be an indicator mineral for carbohydrothermal fluid activity associated with deep-seated, mafic rocks under reduced conditions. At Kovdor and Seblyavr, temperatures of formation for henrymeyerite have been estimated at 160–300 °C (Kogarko et al. 2007, Cegiełka et al. 2022).

Srilankite was originally discovered in placer deposits in Rakwana, Sri Lanka, as anhedral inclusions in zircon and baddeleyite associated with geikielite, spinel, and perovskite (Willgallis et al. 1983). It is generally associated with metamorphic and hydrothermally altered mafic and ultramafic rocks and has been found as inclusions in pyrope-almandine and rutile from lamprophyre, Tobuk-Khatysyr field, Aldan Shield, Yakutia (Biller et al. 2018); in pyrope from mafic diatremes at Garnet Ridge, Arizona (Wang et al. 1999); in gabbro cross-cutting serpentinized peridotite in the West Indian mid-ocean ridge (Morishita et al. 2004); as a replacement product after ilmenite, rutile, and högbomite from anorthosites in Chiapas, Mexico (Cisneros de León et al. 2016); as inclusions in rutile in a hornfels associated with nepheline syenite at Mt. Kaskanyunchor, Khibina massif, Russia (Kostrovitskiy et al. 1993); from granulites in the Caledonides, western Norway (Bingen et al. 2001); in the Russian Urals as inclusions in rutile in a corundum-bearing spinel-sapphirine hornblendite in the Ilmeny-Vishnevogorsk Complex, Chelyabinsk oblast (Korinevsky & Blinov 2016); and as inclusions in ilmenite and rutile from the Mindyak lherzolite massif in the Main Uralian Fault Zone (Gottman et al. 2018). All occurrences of srilankite are suggestive of a genetic link with deep-seated, mafic, SiO₂-depleted magmas coupled with metasomatism and late-stage enrichment in Zr, Ti, Nb, and U (Korinevsky & Blinov 2016, Gottman et al. 2018). At Mt. Kaskanyunchor, the association of srilankite with armalcolite, (Mg,Fe²⁺)Ti₂O₅, a mineral stable only at extremely reduced conditions, constrains conditions of formation to those of low oxygen fugacity. Srilankite has been successfully synthesized experimentally in the TiO₂–ZrO₂ system at both high P-T conditions (1440 °C and 28 kbar), mimicking those in eclogite and granulite facies (Troitzsch & Ellis 2004), and at low P-T (150–900 °C and 1–2 kbar) in hydrothermal conditions (Willgallis et al. 1987).

Henrymeyerite was originally discovered in carbonatite veins at the Kovdor ultramafic alkaline complex, Kola Peninsula, Russia (Mitchell et al. 2000). It has since been found at other alkaline complexes, dominantly in carbonatite or carbohydrothermal veins, including the Khibiny massif, Kola Peninsula, Russia (Mikhailova et al. 2006); Gordon Butte, Crazy Mountains, Montana, USA (Chakhmouradian & Mitchell 2002); at the Seblyavr massif, Russia (Sorokhtina 2000); at Šebkovice, Czech Republic (Krmíček et al. 2011); and at the Ilímaussaq complex, South Greenland (Cegiełka et al. 2022). It has also been found in eastern Antarctica in lherzolite and harzburgite xenoliths which have undergone extensive carbonate metasomatism (Kogarko et al. 2007). As with srilankite, it appears to be an indicator mineral for carbohydrothermal fluid activity associated with deep-seated, mafic rocks under reduced conditions. At Kovdor and Seblyavr, temperatures of formation for henrymeyerite have been estimated at 160–300 °C (Kogarko et al. 2007, Cegiełka et al. 2022).

Lucasite-(Ce) was originally found in heavy mineral separates derived from an olivine lamproite tuff at Argyle, Kimberely region, Western Australia (Nickel et al. 1987). It has since been identified from late-stage hydrothermal and carbohydrothermal assemblages at the Koashva Open Pit, Khibiny Massif, Russia (Pekov & Nikolaev 2013), replacing ilmenite, magnetite, and zirconolite in apatite-magnetite and olivine phoscorite xenoliths in the Mount Bonga carbonatite complex, Angola (Amores-Casals et al. 2019); as needles in titanite from an ultramafic lamprophyre dike at Vinoren, Norway (Zozulya et al. 2020); and as a late-stage mineral after burbankite and loparite-(Ce) in syenite pegmatites which have undergone alteration by carbohydrothermal fluids at Santa Cruz de Tenerife province, Canary Islands, Spain (Dill & Rüsenberg 2023).

Temperatures of formation for the pegmatite can be estimated on the basis of each paragenetic assemblage and the stability temperatures for many of the dominant mineral species, including anatase (200 and 400 °C), ancylite-(Ce) (<350 °C), analcime/wairakite (200–350 °C), and the rare Ti minerals henrymeyerite, srilankite, and lucasite-(Ce) (160–300 °C). For the main Ice River Alkaline Complex rocks, Currie (1975) estimated that the ultramafic layered subcomplex formed between 1000 and 900 °C, with P = 2–3.5 kbar and fO₂ above the quartz-fayalite–magnetite buffer but never exceeding the hematite–magnetite buffer. The syenite subcomplex is estimated to have crystallized between 900 and 600 °C, P = 2.7 kbar, with fO₂ similar to that of the ultramafic rocks. In both cases, the fO₂ of the exsolved residuum would have been higher, likely near the hematite–magnetite buffer.

Anatase and brookite are often described as “metastable,” as they readily transform to rutile at temperatures greater than 600 °C (Hanaor & Sorrell 2011). In fact, thermodynamic studies of the TiO₂ polymorphs suggest that negative pressures would be required for anatase to be more stable than rutile at high temperatures (Smith et al. 2009, Hanaor & Sorrell 2011). Smith et al. (2009) also note that, due to the very small difference in ΔG between the two phases, bulk rutile is only marginally more stable with respect to anatase and that the presence of small quantities of chemical impurities could affect relative stabilities. Although in nature, anatase and brookite are common phases in magmatic and hydrothermal environments. Conversely, they have also been described from low-temperature (50 °C) sedimentary sequences where late-stage diagenetic processes have resulted in alteration of detrital Ti-bearing minerals (biotite). In this setting, complexation and mobilization of Ti is facilitated by organic ligands to form anatase and brookite nanocrystals. Smith et al. (2009), on the basis of thermodynamic calculations, suggested that anatase is only stable in the presence of an aqueous fluid at low temperatures, precipitating as nanocrystals before further coarsening, and may be a reliable indicator of such conditions in both terrestrial and extraterrestrial environments. The only relevant experimental study dealing with the stability of TiO₂ polymorphs in hydrothermal alkaline environments was done by Keesman (1966), which supports a low-temperature stability field for both anatase and brookite (150–400 °C; Fig. 27). The transition between anatase and brookite is controlled by the Na/Na+Ti ratio: at higher Na/Na+Ti (>25%), brookite is stable, whereas at high values (>50%), Na-titanate begins to form. It is therefore not surprising that anatase and brookite at Moose Creek Valley are crystallizing during a Ca-autometasomatism phase when Na/Ca is decreased. Estimates for the formation of anatase and brookite within the Moose Creek Valley pegmatite are 200–300 °C. Temperatures of formation (200–350 °C) for analcime and wairakite (Liou 1971), also part of the early pegmatite assemblage, are in agreement with those for the TiO₂ polymorphs. Further cooling of the pegmatite fluid would result in large-ion incompatible elements such as Ba, Sr, and LREE, excluded from earlier crystallizing silicate phases (analcime, natrolite, aegirine), preferentially partitioning into the CO₂-rich fraction, resulting in co-precipitation of minerals such as ancylite-(Ce), baryte, and thorite. Ancylite-(Ce) is only stable in carbohydrothermal environments at temperatures <350 °C (Wall & Zaitsev 2004). Figure 27 indicates the proposed T stability field for the Moose Creek Valley pegmatite.

A unique carbohydrothermal pegmatite has been discovered on the western slope of Moose Creek Valley in the Ice River Alkaline Complex, British Columbia, Canada. The mineral assemblage is dominated by Ti minerals including titanite, anatase, brookite, and a variety of titanosilicates and rare Ti oxides. It is the first time that lucasite-(Ce), CeTi₂(O,OH)₆, and srilankite, ZrTi₂O₆, have been found in Canada. The Moose Creek Valley pegmatite formed from the mixing of exsolved, SiO₂-undersaturated, alkali-Ti-Fe-rich carbohydrothermal fluids from the ultramafic layered subcomplex and the primary carbonatite. Titanium from primary minerals in the main complex rocks (titanite, Ti-bearing mafic minerals, perovskite) was scavenged and remobilized via Ti-OH⁻ complexes in the exsolved fluids, resulting in a unique mineral assemblage dominated by Ti oxides and silicates. Relative fluctuations in aOH⁻ and aCO₃²⁻ and the Na:Ca ratio in the cooling fluid resulted in alternating stages of Na- and Ca-autometasomatism within the pegmatite and five distinct paragenetic assemblages. The pegmatite appears to have acted as a closed system with no evidence of crustal contamination, extensive interactions with the wall rock, or the input of additional fluids or melts. All assemblages support crystallization in a SiO₂-undersaturated, alkali- and volatile-rich environment at high pH (7–9) and temperatures ranging from 400 to 150 °C.

The authors are extremely grateful to Eagle Plains Resources Ltd. for allowing us access to their mining claims on the western slope of Moose Creek Valley and to collect samples for this research in the 2023 and 2024 field seasons. A big thank you to Alpine Helicopters (Golden) for getting us to and from Moose Creek Valley safely. The comments and suggestions by Editor J.G. Shellnutt, reviewer E. Kozlov, as well as an anonymous reviewer helped to improve this manuscript and are greatly appreciated. This research was funded by grants from the Canadian Museum of Nature to both PCP and AJL.