Petrogenesis of an Episyenite from Iwagi Islet, Southwest Japan: Unique Li–Na Metasomatism during the Turonian

Abstract

A unique Li–Na metasomatic rock from Iwagi Islet in Southwest (SW) Japan is an episyenite that contains new Li-rich minerals, including sugilite, katayamalite, murakamiite, and ferro-ferri-holmquistite. We present petrographical, mineralogical, and geochronological data for the protoliths and episyenite. We classified the metasomatic rocks based on the mineral assemblages, from the protolith biotite granite to albitized granite, quartz albitite, hedenbergite albitite, aegirine albitite, sugilite albitite, and katayamalite albitite. The protolith of hedenbergite albitites may have been metasomatic granite that has been subjected to calcic skarnization. Albitites are formed related to fractures and shear zones that focused the fluid flow and metasomatism. Extensive albitization and formation of abundant Li minerals requires involvement of external Li-Na-Cl-rich fluids, which might be related to deep high-temperature Arima-like brines derived from dehydration of the subducted oceanic slab. Formation of the albitites began with quartz dissolution and vug formation, and record interface-coupled dissolution–reprecipitation processes in an open system. The ⁴⁰Ar/³⁹Ar age of 91.5 ± 0.3 Ma determined for the katayamalite is slightly younger than the protolith zircon U–Pb age of 93.5 ± 1.7 Ma (Turonian), reasonably explaining the timing of Li–Na metasomatism after the petrogenesis of host granites.

1. Introduction

Mineral replacement reactions occur primarily by dissolution–reprecipitation processes associated with fluid–rock interactions and metasomatism [1]. An important observation is the presence of pores, including micro-pores, in the metasomatic minerals. Experimental studies have confirmed that reactions between minerals and a fluid phase commonly involve pseudomorphic replacement via an interface-coupled dissolution–reprecipitation (ICDR) process (e.g., [2,3], and references therein).

Episyenite is a vuggy quartz-deficient rock that results from quartz dissolution, and hydrothermal leaching of quartz is one of the most common types of subsolidus alteration of granites, which is frequently associated with alkali metasomatism (e.g., [4,5,6,7]). This subsolidus leaching of quartz and vug formation enhances the permeability of the altered granites and, consequently when such altered rocks undergo mineralization (e.g., by U-, Sn–W-, and Fe–Cu–Au-bearing solutions), the vugs act as important reservoirs for fluids and sites of ore formation (e.g., [8,9,10]). Episyenitization is also important in relation to the construction of geological repositories for spent nuclear fuel, CO₂ geological storage, and ‘hot dry rock’ power generation (e.g., [7,11]).

Albitization is a common phenomenon, as Na dissolves readily in natural fluids and thus becomes mobile on a crustal scale [12]. Albitization in granites causes an overall change in bulk composition, which is dominated by the formation of Na-rich minerals in the rock and the associated removal of K by the fluid. Engvik et al. [13] suggested that an interface-coupled dissolution–reprecipitation mechanism is responsible for the coupled exchange of Ca²⁺ and Al³⁺ by Na⁺ and Si⁴⁺ during plagioclase albitization. We use the term ‘albitite’ for an albite-dominated episyenite, based on the definition of Engvik et al. [14].

Small bodies of episyenite-like rocks are distributed in the Setouchi region of SW Japan. They are known to occur in 11 areas within Cretaceous granitoids (Figure 1). The occurrence, petrography, and major element chemistry of some episyenite-like rocks have been described previously ([15], and references therein). According to Murakami [16], episyenite-like rocks commonly form small bodies, which are <100 m across and are pipe-, irregular-sheet-, or oval-like in shape. These episyenite-like rocks are all considered to be metasomatic in origin, but previous studies have not confirmed the existence of vugs resulting from quartz dissolution, which is important for the formation of episyenite. Furthermore, the petrology, geochemistry, and geochronology of these episyenite-like rocks and their relationships to regional tectonics have yet to be investigated.

Metasomatic rocks (albitites) in Iwagi Islet (No. 4 in Figure 1) constitute a distinct and rare type of Li-rich rock amongst the episyenites in the Setouchi region of SW Japan. The mineralogical diversity of these rocks is unique, and the Iwagi albitites are known as the type locality of four Li-bearing minerals: sugilite KNa₂(Fe³⁺,Mn³⁺,Al)₂Li₃Si₁₂O₃₀ [17], katayamalite KLi₃Ca₇Ti₂(SiO₃)₁₂(OH)₂ [18], murakamiite LiCa₂Si₃O₈(OH) [19], and ferro-ferri-holmquistite □Li₂(Fe²⁺₃Fe³⁺₂)Si₈O₂₂(OH)₂ [20]. Other Li-rich minerals have also been found, including zektzerite, polylithonite, and tainiolite ([21,22]; this study).

Lithium is an alkali element with an ionic radius smaller than Na, and is known to be highly mobile during fluid–rock interactions [23]. The unusual Li-bearing minerals in Iwagi Islet are considered to have formed by Li metasomatism of an igneous protolith and precipitation from a fluid enriched in Li and Na at low to moderate temperatures (i.e., during the hydrothermal stage) [24]. The murakamiite-bearing katayamalite albitite of this study contains several hundred parts per million (ppm) Li (Table 1). Such a high-Li albitite is unusual amongst the metasomatic rocks in the Setouchi region of SW Japan. The δ⁷Li values of murakamiite and Li-rich pectolite exhibit a wide range, from −9.1‰ to +0.4‰ (average = −2.9‰), and there is no obvious correlation with Li contents. These δ⁷Li values should have resulted from hydrothermal fluid–rock interactions at temperatures of 300–600 °C. The very low δ⁷Li values of down to −9.1‰ may have been due to intracrystalline Li isotopic diffusion or the involvement of deep-seated Li–Na-enriched subduction zone fluids with low δ⁷Li values [24].

In the Iwagi albitites, magmatic quartz was dissolved, and vugs were infilled by albite, alkali feldspar, sugilite, zircon, etc. Pseudomorphs provide clear evidence of dissolution–reprecipitation [1]. The key feature of a pseudomorph is the preservation of the shape of the original crystal (i.e., its volume [1]). The Iwagi albitites contain pores in quartz, feldspar, and zircon, and record mineral replacement reactions, i.e., recrystallization by dissolution–reprecipitation.

In this paper, we first examine the regional variations in the whole-rock Na, K, and Li contents of episyenite-like rocks in the Setouchi region of SW Japan.

We then focus on Iwagi Islet, and present a detailed petrographical and mineralogical description of the host granite and albitites, along with zircon U–Pb and katayamalite ⁴⁰Ar/³⁹Ar ages. We use these data to constrain the origin and nature of the Li–Na-rich metasomatizing fluid, and the petrogenesis and age of the episyenite. Finally, we present a simplified model for the origin of the episyenite, based on the concept of infiltration metasomatism. This model can explain the spatiotemporal association of the granite protolith and various albitites.

2. Geological Background

2.1. Regional Distribution and Mineralogy of Episyenite-like Rocks in the Setouchi Region

Subsolidus, quartz-depleted, hydrothermal–alkali metasomatized rocks (episyenites) that typically contain pyroxene, garnet, and amphibole occur in the Setouchi region. These rocks crop out in an E–W direction and are closely associated with Cretaceous granitoids (Figure 1) [16].

The occurrence of episyenite-like rocks is limited to Cretaceous granites, and they are not known to occur in Paleogene granites (Figure 1). The only rocks that can clearly be called episyenites are in Iwagi Islet, which highlights the rarity of this rock type. The episyenites occur mostly as small stock- or dike-like masses that grade into the surrounding granitic rocks. The textures of the episyenite-like rocks are similar to those of the surrounding granites and, in many places, their grain size, magmatic foliation, and associated aplites and pegmatites that formed during earlier magmatic to deuteric stages are preserved. The episyenites often occur along shear zones, where quartz veins are observed. Sheared textures are observed in both the episyenite-like rocks and surrounding granites.

The dominant minerals in the episyenites are plagioclase, alkali feldspar, clinopy roxene, garnet, amphibole, sugilite, katayamalite, pectolite–murakamiite, ferro-ferri-holmquistite, quartz, epidote, chlorite, and accessory quartz, titanite, allanite, carbonate minerals, Fe–Ti oxides, apatite, and zircon. Locally, small amounts of sulfide minerals, such as pyrite, chalcopyrite, galena, zincblende, and molybdenite, occur in the episyenites [16].

2.2. Occurrence of Granites and Albitites in Iwagi Islet

The granites and albitites in Iwagi Islet (34°15′47″ N, 133°9′39″ E) are the basis of this study. The geology of Iwagi Islet and the albitites in the islet were described in detail by Imaoka et al. [24].

Biotite granites are widely distributed throughout the islet (Figure 2). Based on our study, the granites can be divided into coarse-, medium-, and fine-grained types. The coarse-grained granite occurs mainly as sheets in the western, eastern, and northern parts of the islet, and at topographically lower levels, as compared with the medium-grained granite. The fine-grained granite forms tabular bodies that intrude on the upper part of medium-grained granite around Mount Sekizen-san. These granites are vertically stacked, sheet-like bodies that can be explained by a sheet-on-sheet model [25].

The albitites occur as many small, discrete bodies in the coarse-grained biotite granite (hereafter, the protolith granite is referred to as the biotite granite) around the summit of Mount Kuresaka in the eastern part of the islet. The largest body of albitite is ~30 m in diameter, and small albitite bodies (several tens of centimeters to meters in width) are generally oriented in a N–S direction and irregularly distributed over an area of 1.5 × 0.8 km².

The relationship between the albitite and surrounding biotite granite is gradational, and no crosscutting relationships can be found between the albitized granite and quartz albitite (Figure 3). The albitized granite contains aegirine-augite and ferro-ferri-holmquistite-bearing quartz veins.

As described in detail in Section 5, the albitites in Iwaki Islet can be classified based on mineralogy into (1) biotite granite (protolith), (2) albitized granite, (3) quartz albitite, (4) hedenbergite albitite, (5) aegirine albitite, (6) sugilite albitite, and (7) katayamalite albitite. These albitites and albitized granite vary in scale from a few centimeters to hundreds of meters.

3. Samples and Analytical Procedure

3.1. Samples

3.1.1. Samples for Whole-Rock and Mineralogical and Petrological Analyses

Representative rock samples sufficient to document variations in mineralogical and petrological aspects were selected after examination based on the thin section observations. To determine the whole-rock Na₂O, K₂O, and Li content variations in the albitites and granite protoliths in Iwagi Islet, and regional variations of episyenite-like rocks in the Setouchi region, we analyzed 33 samples from Iwagi Islet and 33 samples from the Setouchi region, which are listed in Tables 1 and 2.

The rock type and number of the Iwagi samples for qualitative and quantitative analyses of minerals are as follows: biotite granites (n = 8), albitized granite (6), quartz albitite (2), hedenbergite albitite (2), aegirine albitite (5), sugilite albitite (2), and katayamalite albitite (9).

3.1.2. Samples for Zircon U–Pb Dating

A protolith coarse-grained granite (sample T-69; 34°15′22″ N, 133°9′59″ E), medium-grained biotite granite (sample IW-300), and fine-grained granite (sample IW-303) from Iwagi Islet were subjected to zircon U–Pb dating. Sampling locations, other than for sample T-69, are listed in Table 1. The zircon in the analyzed sample T-69 coexists with a hydrated zircon phase (ZrSiO₄∙nH₂O).

The protolith granite consists of plagioclase (18.5 vol.%), alkali feldspar (38.4 vol.%), quartz (41.2 vol.%), biotite (1.6 vol.%), and altered minerals (0.2 vol.%). Accessory minerals include zircon, thorite/huttonite, monazite-(Ce), -(Nd), and -(Y), fluorapatite, ilmenite, fergusonite, xenotime, galena, scheelite, and wolframite.

The medium-grained biotite granite consists of plagioclase (27.9 vol.%), alkali feldspar (32.0 vol.%), quartz (38.0 vol.%), biotite (2.0 vol.%), muscovite (0.2 vol.%), and an altered mineral (0.1 vol.%). Accessory minerals include fluorite, ilmenite, zircon, thorite/huttonite, monazite-(Ce), TiO₂ mineral, xenotime, wolframite, and fluorapatite.

The fine-grained biotite granite consists of plagioclase (27.3 vol.%), alkali feldspar (33.9 vol.%), quartz (35.6 vol.%), biotite (2.8 vol.%), and muscovite (0.1 vol.%). Accessory minerals include zircon, monazite-(Ce), -(Nd), and -(Y), ilmenite, thorite/huttonite, xenotime, fluorite, cassiterite, fergusonite, fluorapatite, chalcopyrite, arsenopyrite, sphalerite, and tetrahedrite.

All the analyzed zircon grains are colorless, transparent, and euhedral. Zircon lengths and length/width ratios for the coarse-, medium-, and fine-grained granites are 120–400 μm and 1.6–3.0, 120–410 μm and 1.2–4.0, and 160–390 μm and 1.2–3.4, respectively. Representative cathodoluminescence (CL) images of the analyzed zircons and backscattered electron (BSE) images of hydrous zircons in sample T-69 are shown in Supplementary Figure S1. The CL images show that most zircon grains are characterized by fine-scale oscillatory igneous zoning that reflects variations in trace element contents (i.e., U, Th, Y, and heavy REEs). The Th/U ratios range from 0.13 to 0.69 (Supplementary Table S2), indicating an igneous origin [26]. The estimated water contents of hydrous zircons are shown in Supplementary Figure S1.

3.1.3. Sample for ⁴⁰Ar/³⁹Ar Dating

Katayamalite crystals collected from an albitite sample (IWG-168a) from Iwagi Islet were subjected to ⁴⁰Ar/³⁹Ar dating. The host rock consists of albite (79.8 vol.%), sugilite (9.8 vol.%), partially altered truscottite and other minerals, aegirine-augite (3.0 vol.%), quartz (2.7 vol.%), murakamiite–Li-rich pectolite (2.4 vol.%), katayamalite (1.9 vol.%), alkali feldspar (0.3 vol.%), and accessory zircon, britholite mineral, an unidentified Si–Th–Ca mineral, fluorapatite, and titanite.

The katayamalite occurs as single grains and aggregates. Rod-shaped (0.1–1.5 mm long) aggregates occur with oriented long axes (Supplementary Figure S2). It exhibits perfect (001) cleavage and parallel twinning, which is partly bent, and is colorless, clear, and unaltered.

3.2. Analytical Procedure

3.2.1. Whole-Rock Na₂O, K₂O and Li Analyses

Whole-rock analyses of Na₂O and K₂O contents were undertaken using X-ray fluorescence spectrometry (XRF; Rigaku ZSX Primus-II instrument) at the Center for Instrumental Analysis, Yamaguchi University, Japan. The details and accuracy of the analytical method were described by Eshima and Owada [27].

Lithium contents in the rocks were determined by inductively coupled plasma–optical emission spectroscopy (ICP–OES; Perkin–Elmer Optima 8300) at Yamaguchi Prefectural Industrial Technology Institute, Yamaguchi, Japan. Powdered rock samples (0.1 g) were digested in a mixture of hydrofluoric and perchloric acids [28], and then evaporated to dryness. The residues were dissolved by heating with 5 mL of 6 N HCl, and then diluted to 100 mL with water. The Li concentrations in the solutions were measured by ICP–OES using a wavelength of 670.706 nm. The analytical precision for all measurements was better than ±3%, as estimated from the reproducibility (2 RSD) of multiple measurements of standard solution.

3.2.2. Surface Observation and Mineral Analyses

The modal mineralogy of major minerals was determined in thin sections of 29 samples by counting a grid of >3000 points under a polarizing microscope. The abundance of accessory minerals was determined by an electron probe microprobe (EPMA; JXA-8230) at the Center for Instrumental Analysis, Yamaguchi University, whereby the total area of each mineral grain was calculated in a measured area of a polished section. The surface texture of the albitite sample was observed with a scanning electron microscope (SEM; JEOL JSM-6360OLA) at the Center for Instrumental Analysis, Yamaguchi University.

Qualitative and quantitative analyses of minerals were undertaken with the EPMA. In addition to qualitative analyses, such as back-scattered electron (BSE) imaging, quantitative analyses and compositional mapping were undertaken to identify the key compositional characteristics, textures, and mineral relationships of the studied samples. The minerals were identified only by electron-microprobe examination; no other methods, such as X-ray single-crystal structure analysis, were used.

The operating conditions for the EPMA analyses were as follows: accelerating voltage = 15 keV, beam current = 20 nA, and beam diameter = 1–5 μm. Wavelength-dispersive spectra were collected with LiF, PET, and TAP crystals to identify the interfering elements and the best wavelengths for background measurements. The ZAF method was used for data correction. The analytical errors were ±2% for major elements and ±5% for minor elements, based on the reproducibility of multiple measurements.

3.2.3. Zircon U–Pb Dating

Rocks were crushed by the selective fragmentation device using high-voltage pulsed electric discharges (selFrag-Lab) installed in the National Museum of Nature and Science (NMNS) to maximize the yield of zircon recovery. The heavy minerals were concentrated by panning and further processed with a hand magnet, and the remaining fractions were purified using heavy liquid (diiodomethane) separation. These zircons were mounted in epoxy resin together with the standard zircons TEMORA 2 (²⁰⁶Pb/²³⁸U age of 416.8 ± 1.3 Ma) [29], FC1 (²⁰⁷Pb/²⁰⁶Pb age of 1099.0 ± 0.6 Ma)[30], and OD-3 (²⁰⁶Pb/²³⁸U age of 32.8 ± 0.1 Ma) [31]. The zircons were then polished using diamond paste until the interiors of the crystal cores were exposed. Chemical zoning of the individual zircon grains was investigated by BSE and CL imaging using an SEM (JEOL JSM-6610) at the NMNS. Representative CL images of the analyzed zircons are shown in Supplementary Figure S1. Analytical sites for U–Pb dating were selected using these images by choosing euhedral regions of igneous origin. Inherited zircon cores and mineral/fluid inclusions were avoided.

Zircon U–Pb dating was undertaken with a 200 nm femtosecond laser ablation system (200FsLA; OKFs-2000K; OK Lab, Tokyo, Japan) coupled to a sector-field–inductively coupled plasma–mass spectrometer (SF–ICP–MS; ThermoFisher Scientific, Element-XR; Bremen, Germany) at the Department of Solid Earth Geochemistry, Japan Agency for Marine-Earth Science and Technology (JAMSTEC). Analysis followed the methods described by Kimura and Chang [32], Iwano et al. [31], and Kimura et al. [33]. In brief, the laser fluence on the sample surface was ~6 J cm⁻², and the samples were ablated in rotational raster mode [32] around the circumference of a circle (10 μm radius) at a velocity of 11.5 μm s⁻¹ using a laser beam diameter of ~15 μm. The repetition rate of the laser was set to 2 or 10 Hz depending on the zircon U contents. The resulted ablation craters were ~20 μm in diameter and 3 to 15 μm in depth after 60 s of ablation. The SF–ICP–MS was tuned to minimize oxide molecular yields at ThO⁺/Th < 0.1% and by maximizing the elemental sensitivity at ⁷Li, ¹¹⁵In, and ²³⁸U [32]. Time-resolved analysis was used for spot analysis, with a 20 s gas blank measurement with the laser off followed by data acquisition for ~60 s with the laser on. The gas blanks were subtracted from the zircon signals.

The measured ²⁰⁷Pb/²⁰⁶Pb and ²⁰⁶Pb/²³⁸U ratios were calibrated against TEMORA 2 values determined by isotope dilution–thermal ionization mass spectrometry (ID–TIMS) [29]. TEMORA 2 was measured before and after each of the five unknowns and used as the bracketing standard. Error propagation from the two bracketing standards and each unknown measurement was undertaken. Concordia ages and plots were obtained with Isoplot v.4 [34]. Outliers excluded from the age calculations were only selected on the basis of statistical variations and were mostly due to the ablation of mineral inclusions and inherited cores. The measured isotopic ratios and ages are quoted with 1σ errors, whereas the weighted-mean ages are presented at the 95% confidence level (Supplementary Table S1). The FC-1 and OD-3 standard zircons were also analyzed. During this study, 10 spot analyses of FC-1 and OD-3 yielded ages of 1094 ± 15 Ma (reference age = 1099.0 ± 0.6 Ma) and 32.2 ± 0.9 Ma (reference age = 32.8 ± 0.1 Ma), respectively (Supplementary Figure S3).

3.2.4. Katayamalite ⁴⁰Ar/³⁹Ar Dating

The rock sample was crushed and sieved, and the 149–300 mesh size fraction was used for the separation of katayamalite. The sieved fraction was washed in distilled water in an ultrasonic bath to remove fine particles on grain surfaces, and then dried in an oven at 80 °C. Katayamalite was separated using an isodynamic separator and heavy liquids until no impurities were observed under a microscope. The Katayamalite separates were then washed in Milli-Q^® water (>18.2 MΩ) three times. The final Katayamalite purity was >99%.

The ⁴⁰Ar/³⁹Ar dating of katayamalite was undertaken at Actlabs, Ancaster, Canada. Neutron irradiation of the sample was carried out in the TRIGA CLICIT nuclear reactor at Oregon State University, Oregon, USA, together with the FCT sanidine standard (28.201 ± 0.023 Ma, 1σ [35]). The minerals were analyzed by the step-heating technique using a 25 W CO₂ laser. The Ar isotopes were measured using an Argus-VI mass spectrometer at Oregon State University.

4. Regional Variations of Whole-Rock Alkali Content

4.1. Iwagi Albitites and Their Protolith

To determine the mass of material added and removed from the rocks, and establish whether the metasomatic processes could have been responsible for Li enrichment, we compared the chemical compositions of the protolith granite and various metasomatic samples. The Na₂O, K₂O, and Li contents of representative samples of the host biotite granites and albitites from Iwagi Islet are shown in Table 1 and Figure 4.

The albitized granite has higher Na₂O contents (3.92–7.85 wt.%) than the host biotite granites (2.97–3.76 wt.%). The Na₂O contents of quartz albitite (9.6–10.2 wt.%), hedenbergite albitite (10.6–10.8 wt.%), aegirine albitite (11.0–11.3 wt.%), sugilite albitite (10.6–10.9 wt.%), and katayamalite albitite (6.08–11.1 wt.%) are further higher than those of the host biotite granite and albitized granite (Figure 4).

The coarse-grained biotite granite (protolith of albitites) has lower K₂O contents (3.85–4.10 wt.%) than other granites (4.41–5.43 wt.%). The K₂O contents (0.82–7.38 wt.%) of the albitized granites are highly variable. The quartz albitite (0.30–0.44 wt.%), hedenbergite albitite (0.43–0.93 wt.%), aegirine albitite (0.16–0.25 wt.%), sugilite albitite (0.26–0.27 wt.%), and katayamalite albitite (0.21–4.88 wt.%) have generally lower K₂O contents than host biotite granites and albitized granites (Figure 4a).

Table 1. Whole-rock Na₂O, K₂O and Li contents of Iwagi Islet, Ehime Prefecture, SW Japan.

No.	Rock Type	Na2O (wt.%)	K2O (wt.%)	Li (ppm)	Latitude (N)	Longitude (E)
IW-41	Coarse-grained Bt granite	3.76	3.85	42.9	34°15′27.4″	133°8′40.2″
IW-44	Coarse-grained Bt granite	2.97	4.10	13.5	34°16′09.4″	133°8′15.1″
IW-47	Medium-grained Bt granite	3.07	4.50	39.7	34°15′42.7″	133°8′48.9″
IW-48	Medium-grained Bt granite	3.44	4.67	46.0	34°15′37.2″	133°8′41.2″
IW-89	Medium-grained Bt granite	3.12	5.43	10.1	34°14′43.8″	133°7′54.5″
IW-300	Medium-grained Bt granite	3.44	4.70	99.0	34°15′41.4″	133°8′49.1″
IW-303	Fine-grained Bt granite	3.58	4.72	100	34°15′29.0″	133°8′43.6″
IW-42	Fine-grained Bt granite	3.45	4.41	43.2	34°15′33.6″	133°8′42.3″
IW-03	Albitized granite	6.96	0.82	49.3	34°15′42.0″	133°9′36.6″
IW-53	Albitized granite	3.92	5.18	110	34°15′46.1″	133°9′36.3″
T-17	Albitized granite	6.34	2.17	43.0	34°15′49.1″	133°9′39.4″
T-24	Albitized granite	5.44	7.38	26.3	34°15′41.7″	133°9′38.1″
T-204	Albitized granite	7.85	3.99	87.0	34°15′50.9″	133°9′37.6″
T-206B	Albitized granite	4.68	5.55	26.0	34°15′51.3″	133°9′37.1″
IW-60	Quartz albitite	10.19	0.30	5.9	34°15′46.4″	133°9′35.9″
T-50	Quartz albitite	9.55	0.44	12.9	34°15′42.0″	133°9′38.5″
T-4	Hedenbergite albitite	10.58	0.93	17.0	34°15′40.7″	133°9′38.4″
IW-51	Hedenbergite albitite	10.75	0.43	7.4	34°15′41.2″	133°9′38.2″
W-14	Aegirine albitite	11.26	0.16	71.8	34°15′47.5″	133°9′47.2″
IW-23	Aegirine albitite	11.29	0.25	7.1	34°15′47.4″	133°9′39.2″
IW-24	Aegirine albitite	10.97	0.22	51.2	34°15′47.3″	133°9′39.1″
IW-90A	Aegirine albitite	11.15	0.24	23.1	34°15′47.1″	133°9′40.1″
IW-167	Aegirine albitite	11.06	0.23	18.9	34°15′46.9″	133°9′40.3″
IW-71	Sugilite albitite	10.62	0.27	292	34°15′47.2″	133°9′33.7″
IW-90B	Sugilite albitite	10.92	0.26	255	34°15′46.6″	133°9′40.4″
IW-2-3	Katayamalite albitite	6.08	4.88	656	34°15′47.3″	133°9′39.5″
IW-12	Katayamalite albitite	10.39	0.21	356	34°15′47.0″	133°9′40.5″
IW-168	Katayamalite albitite	10.55	0.35	441	34°15′47.4″	133°9′39.4″
IW-28	Katayamalite albitite	10.68	0.33	651	34°15′46.9″	133°9′40.3″
IW-27	Katayamalite albitite	10.38	0.35	960	34°15′47.0″	133°9′39.9″
IW-75	Katayamalite albitite	10.16	0.30	465	34°15′47.3″	133°9′40.2″
IW-122	Katayamalite albitite	11.09	0.27	393	34°15′47.3″	133°9′39.8″
IW-131	Katayamalite albitite	10.71	0.36	977	34°15′47.2″	133°9′39.7″

Bt = biotite.

Table 1. Whole-rock Na₂O, K₂O and Li contents of Iwagi Islet, Ehime Prefecture, SW Japan.

No.	Rock Type	Na2O (wt.%)	K2O (wt.%)	Li (ppm)	Latitude (N)	Longitude (E)
IW-41	Coarse-grained Bt granite	3.76	3.85	42.9	34°15′27.4″	133°8′40.2″
IW-44	Coarse-grained Bt granite	2.97	4.10	13.5	34°16′09.4″	133°8′15.1″
IW-47	Medium-grained Bt granite	3.07	4.50	39.7	34°15′42.7″	133°8′48.9″
IW-48	Medium-grained Bt granite	3.44	4.67	46.0	34°15′37.2″	133°8′41.2″
IW-89	Medium-grained Bt granite	3.12	5.43	10.1	34°14′43.8″	133°7′54.5″
IW-300	Medium-grained Bt granite	3.44	4.70	99.0	34°15′41.4″	133°8′49.1″
IW-303	Fine-grained Bt granite	3.58	4.72	100	34°15′29.0″	133°8′43.6″
IW-42	Fine-grained Bt granite	3.45	4.41	43.2	34°15′33.6″	133°8′42.3″
IW-03	Albitized granite	6.96	0.82	49.3	34°15′42.0″	133°9′36.6″
IW-53	Albitized granite	3.92	5.18	110	34°15′46.1″	133°9′36.3″
T-17	Albitized granite	6.34	2.17	43.0	34°15′49.1″	133°9′39.4″
T-24	Albitized granite	5.44	7.38	26.3	34°15′41.7″	133°9′38.1″
T-204	Albitized granite	7.85	3.99	87.0	34°15′50.9″	133°9′37.6″
T-206B	Albitized granite	4.68	5.55	26.0	34°15′51.3″	133°9′37.1″
IW-60	Quartz albitite	10.19	0.30	5.9	34°15′46.4″	133°9′35.9″
T-50	Quartz albitite	9.55	0.44	12.9	34°15′42.0″	133°9′38.5″
T-4	Hedenbergite albitite	10.58	0.93	17.0	34°15′40.7″	133°9′38.4″
IW-51	Hedenbergite albitite	10.75	0.43	7.4	34°15′41.2″	133°9′38.2″
W-14	Aegirine albitite	11.26	0.16	71.8	34°15′47.5″	133°9′47.2″
IW-23	Aegirine albitite	11.29	0.25	7.1	34°15′47.4″	133°9′39.2″
IW-24	Aegirine albitite	10.97	0.22	51.2	34°15′47.3″	133°9′39.1″
IW-90A	Aegirine albitite	11.15	0.24	23.1	34°15′47.1″	133°9′40.1″
IW-167	Aegirine albitite	11.06	0.23	18.9	34°15′46.9″	133°9′40.3″
IW-71	Sugilite albitite	10.62	0.27	292	34°15′47.2″	133°9′33.7″
IW-90B	Sugilite albitite	10.92	0.26	255	34°15′46.6″	133°9′40.4″
IW-2-3	Katayamalite albitite	6.08	4.88	656	34°15′47.3″	133°9′39.5″
IW-12	Katayamalite albitite	10.39	0.21	356	34°15′47.0″	133°9′40.5″
IW-168	Katayamalite albitite	10.55	0.35	441	34°15′47.4″	133°9′39.4″
IW-28	Katayamalite albitite	10.68	0.33	651	34°15′46.9″	133°9′40.3″
IW-27	Katayamalite albitite	10.38	0.35	960	34°15′47.0″	133°9′39.9″
IW-75	Katayamalite albitite	10.16	0.30	465	34°15′47.3″	133°9′40.2″
IW-122	Katayamalite albitite	11.09	0.27	393	34°15′47.3″	133°9′39.8″
IW-131	Katayamalite albitite	10.71	0.36	977	34°15′47.2″	133°9′39.7″

Bt = biotite.

The host granite samples have variable lithium contents (10.1–100 ppm), with the coarse-grained biotite granite containing 13.5–42.9 ppm Li, medium-grained biotite granite containing 10.1–99.0 ppm Li, and fine-grained biotite granite containing 43.2–100 ppm Li. The Li contents of the albitized granite (26.0–110.0 ppm), quartz albitite (5.2–12.9 ppm), hedenbergite albitite (7.4–17.0 ppm), and aegirine albitite (7.1–71.8 ppm) are comparable with those of the host granites. On the other hand, sugilite albitite (255–292 ppm), and katayamalite albitite (356–977 ppm) show much higher Li contents than host granites (Figure 4b).

4.2. Comparison of Different Episyenite-like Rocks in the Setouchi Region

The Na₂O contents of episyenite-like rocks in the Setouchi region, excluding the sam ples from Iwagi Islet, vary widely from 2.45 to 11.26 wt.% (Table 2). The Na₂O contents for the majority of these episyenite-like rocks are comparable with those of albitized granites of Iwagi, and their average (7.16 wt.%) is distinctly lower than the average (10.40 wt.%) of Iwagi albitites (Figure 4). The K₂O contents, again excluding the samples from Iwagi Islet, also range widely from 0.16 to 6.04 wt.%, with an average of 2.63 wt.%, much higher than that (0.57 wt.%) of albitites from Iwagi Islet (Figure 4a).

Li contents of episyenite-like rocks in the Setouchi region are relatively low and var iable (3.0–104 ppm; average = 16.4 ppm), but most are 3.0–23.9 ppm (average = 12.3 ppm), except for values of 104 and 57.2 ppm for rocks from Shodoshima and Innoshima islands, respectively. These Li contents are far lower compared with sugilite albitites and katayamalite albitites from Iwagi Islet (Figure 4b). In summary, the Iwagi albitites are characterized by distinctly higher Na₂O contents and lower K₂O contents than other episyenite-like rocks from the Setouchi region, and by the occurrence of extreme Li enrichment associated with metasomatism.

Table 2. Whole-rock Na₂O, K₂O and Li contents of episyenite-like rocks from the Setouchi region, SW Japan.

No.	Rock Type	Na2O (wt.%)	K2O (wt.%)	Li (ppm)	Locality (Numbers Correspond to Those in Figure 1)
TS-1	Cpx–Pl–Kfs–Qz albitite	10.05	0.92	14.5	1	Yamada, Taishi-cho, Osaka Prefecture
TS-2	Cpx–Pl–Kfs–Qz episyenite-like rock	7.84	2.50	11.1	1	Yamada, Taishi-cho, Osaka Prefecture
SD-1	Cpx–Pl–Kfs–Qz episyenite-like rock	7.60	3.79	3.4	2	Shodoshima Island, Kagawa Prefecture
SD-2	Cpx–Pl–Kfs–Qz episyenite-like rock	8.08	3.42	3.0	2	Shodoshima Island, Kagawa Prefecture
SD-3	Cpx–Pl–Kfs–Qz episyenite-like rock	7.18	2.48	9.9	2	Shodoshima Island, Kagawa Prefecture
SD-4	Grt–Pl–Kfs–Qz albitite	11.19	0.29	104	2	Shodoshima Island, Kagawa Prefecture
SD-5	Cpx–Grt–Pl–Kfs episyenite-like rock	8.83	1.95	10.8	2	Shodoshima Island, Kagawa Prefecture
SD-6	Hbl–Ep–Pl–Qz episyenite-like rock	7.82	2.26	13.0	2	Shodoshima Island, Kagawa Prefecture
IN-10	Cpx–Hbl–Pl–Kfs episyenite-like rock	5.45	3.69	5.2	3	Innoshima Island, Hiroshima Prefecture
IN-11	Cpx–Hbl–Pl–Kfs episyenite-like rock	2.45	2.43	57.2	3	Innoshima Island, Hiroshima Prefecture
IN-12	Cpx–Hbl–Pl–Kfs episyenite-like rock	5.84	4.30	12.1	3	Innoshima Island, Hiroshima Prefecture
NK-1	Hbl–Cpx–Pl episyenite-like rock	6.79	0.20	11.7	6	Namikata, Ehime Prefecture
NK-2	Hbl–Cpx–Pl episyenite-like rock	5.81	2.80	18.9	6	Namikata, Ehime Prefecture
NK-3	Cpx–Grn–Hbl–Pl episyenite-like rock	6.41	1.62	7.5	6	Namikata, Ehime Prefecture
NK-4	Grt–Pl–Kfs–Qz episyenite-like rock	5.68	4.15	10.6	6	Namikata, Ehime Prefecture
NK-5	Cpx–Grn–Pl–Kfs episyenite-like rock	5.63	5.38	7.6	6	Namikata, Ehime Prefecture
KR-1	Hbl–Cpx–Pl–Kfs episyenite-like rock	7.81	1.90	5.8	7	Kure, Hiroshima Prefecture
NM-1	Hbl–Cpx–Pl–Kfs episyenite-like rock	7.58	3.16	16.6	8	Nomijima Island, Hiroshima Prefecture
NM-2	Grn–Cpx–Pl–Kfs episyenite-like rock	6.58	3.19	10.6	8	Nomijima Island, Hiroshima Prefecture
MM-3	Hbl–Ep–Kfs–Pl episyenite-like rock	4.50	6.04	9.2	8	Nomijima Island, Hiroshima Prefecture
NM-4	Hbl–Cpx–Pl–Kfs episyenite-like rock	8.56	2.80	16.5	8	Nomijima Island, Hiroshima Prefecture
AI-1	Cpx–Pl–Kfs–Qz episyenite-like rock	6.01	3.50	13.7	10	Chudo, Aio, Yamaguchi Prefecture
AI-2	Cpx–Pl–Kfs–Qz episyenite-like rock	5.82	3.88	7.0	10	Chudo, Aio, Yamaguchi Prefecture
AI-3	Grt–Cpx–Hbl–Pl episyenite-like rock	5.12	3.03	11.0	10	Hazekura, Aio, Yamaguchi Prefecture
AI-4	Cpx–Hbl–Pl–Kfs–Qz albitite	11.26	0.16	11.9	10	Hazekura, Aio, Yamaguchi Prefecture
UO-1	Hbl–Cpx–Pl–Kfs episyenite-like rock	7.82	2.21	14.7	11	Utsugiono, Ube, Yamaguchi Prefecture
UO-2	Hbl–Ep–Pl–Kfs–Qz episyenite-like rock	5.38	4.14	20.5	11	Utsugiono, Ube, Yamaguchi Prefecture
UO-3	Ep–Hbl–Pl–Qz–Kfs episyenite-like rock	5.70	2.91	13.0	11	Utsugiono, Ube, Yamaguchi Prefecture
UO-4	Cpx–Pl episyenite-like rock	8.54	0.60	16.8	11	Utsugiono, Ube, Yamaguchi Prefecture
UO-5	Hbl–Pl–Qz episyenite-like rock	8.82	0.58	12.7	11	Utsugiono, Ube, Yamaguchi Prefecture
UO-6	Grt–Pl–Qz episyenite-like rock	9.72	0.17	21.9	11	Utsugiono, Ube, Yamaguchi Prefecture
UO-7	Cpx–Hbl–Pl–Qz episyenite-like rock	8.59	0.59	23.9	11	Utsugiono, Ube, Yamaguchi Prefecture
UO-8	Grt–Ep–Pl–Kfs–Qz episyenite-like rock	5.74	5.85	16.3	11	Utsugiono, Ube, Yamaguchi Prefecture

Cpx = clinopyroxene, Pl = plagioclase, Kfs = K-feldspar, Qz = quartz, Grt = garnet, Ep = epidote, and Hbl = hornblende.

Table 2. Whole-rock Na₂O, K₂O and Li contents of episyenite-like rocks from the Setouchi region, SW Japan.

No.	Rock Type	Na2O (wt.%)	K2O (wt.%)	Li (ppm)	Locality (Numbers Correspond to Those in Figure 1)
TS-1	Cpx–Pl–Kfs–Qz albitite	10.05	0.92	14.5	1	Yamada, Taishi-cho, Osaka Prefecture
TS-2	Cpx–Pl–Kfs–Qz episyenite-like rock	7.84	2.50	11.1	1	Yamada, Taishi-cho, Osaka Prefecture
SD-1	Cpx–Pl–Kfs–Qz episyenite-like rock	7.60	3.79	3.4	2	Shodoshima Island, Kagawa Prefecture
SD-2	Cpx–Pl–Kfs–Qz episyenite-like rock	8.08	3.42	3.0	2	Shodoshima Island, Kagawa Prefecture
SD-3	Cpx–Pl–Kfs–Qz episyenite-like rock	7.18	2.48	9.9	2	Shodoshima Island, Kagawa Prefecture
SD-4	Grt–Pl–Kfs–Qz albitite	11.19	0.29	104	2	Shodoshima Island, Kagawa Prefecture
SD-5	Cpx–Grt–Pl–Kfs episyenite-like rock	8.83	1.95	10.8	2	Shodoshima Island, Kagawa Prefecture
SD-6	Hbl–Ep–Pl–Qz episyenite-like rock	7.82	2.26	13.0	2	Shodoshima Island, Kagawa Prefecture
IN-10	Cpx–Hbl–Pl–Kfs episyenite-like rock	5.45	3.69	5.2	3	Innoshima Island, Hiroshima Prefecture
IN-11	Cpx–Hbl–Pl–Kfs episyenite-like rock	2.45	2.43	57.2	3	Innoshima Island, Hiroshima Prefecture
IN-12	Cpx–Hbl–Pl–Kfs episyenite-like rock	5.84	4.30	12.1	3	Innoshima Island, Hiroshima Prefecture
NK-1	Hbl–Cpx–Pl episyenite-like rock	6.79	0.20	11.7	6	Namikata, Ehime Prefecture
NK-2	Hbl–Cpx–Pl episyenite-like rock	5.81	2.80	18.9	6	Namikata, Ehime Prefecture
NK-3	Cpx–Grn–Hbl–Pl episyenite-like rock	6.41	1.62	7.5	6	Namikata, Ehime Prefecture
NK-4	Grt–Pl–Kfs–Qz episyenite-like rock	5.68	4.15	10.6	6	Namikata, Ehime Prefecture
NK-5	Cpx–Grn–Pl–Kfs episyenite-like rock	5.63	5.38	7.6	6	Namikata, Ehime Prefecture
KR-1	Hbl–Cpx–Pl–Kfs episyenite-like rock	7.81	1.90	5.8	7	Kure, Hiroshima Prefecture
NM-1	Hbl–Cpx–Pl–Kfs episyenite-like rock	7.58	3.16	16.6	8	Nomijima Island, Hiroshima Prefecture
NM-2	Grn–Cpx–Pl–Kfs episyenite-like rock	6.58	3.19	10.6	8	Nomijima Island, Hiroshima Prefecture
MM-3	Hbl–Ep–Kfs–Pl episyenite-like rock	4.50	6.04	9.2	8	Nomijima Island, Hiroshima Prefecture
NM-4	Hbl–Cpx–Pl–Kfs episyenite-like rock	8.56	2.80	16.5	8	Nomijima Island, Hiroshima Prefecture
AI-1	Cpx–Pl–Kfs–Qz episyenite-like rock	6.01	3.50	13.7	10	Chudo, Aio, Yamaguchi Prefecture
AI-2	Cpx–Pl–Kfs–Qz episyenite-like rock	5.82	3.88	7.0	10	Chudo, Aio, Yamaguchi Prefecture
AI-3	Grt–Cpx–Hbl–Pl episyenite-like rock	5.12	3.03	11.0	10	Hazekura, Aio, Yamaguchi Prefecture
AI-4	Cpx–Hbl–Pl–Kfs–Qz albitite	11.26	0.16	11.9	10	Hazekura, Aio, Yamaguchi Prefecture
UO-1	Hbl–Cpx–Pl–Kfs episyenite-like rock	7.82	2.21	14.7	11	Utsugiono, Ube, Yamaguchi Prefecture
UO-2	Hbl–Ep–Pl–Kfs–Qz episyenite-like rock	5.38	4.14	20.5	11	Utsugiono, Ube, Yamaguchi Prefecture
UO-3	Ep–Hbl–Pl–Qz–Kfs episyenite-like rock	5.70	2.91	13.0	11	Utsugiono, Ube, Yamaguchi Prefecture
UO-4	Cpx–Pl episyenite-like rock	8.54	0.60	16.8	11	Utsugiono, Ube, Yamaguchi Prefecture
UO-5	Hbl–Pl–Qz episyenite-like rock	8.82	0.58	12.7	11	Utsugiono, Ube, Yamaguchi Prefecture
UO-6	Grt–Pl–Qz episyenite-like rock	9.72	0.17	21.9	11	Utsugiono, Ube, Yamaguchi Prefecture
UO-7	Cpx–Hbl–Pl–Qz episyenite-like rock	8.59	0.59	23.9	11	Utsugiono, Ube, Yamaguchi Prefecture
UO-8	Grt–Ep–Pl–Kfs–Qz episyenite-like rock	5.74	5.85	16.3	11	Utsugiono, Ube, Yamaguchi Prefecture

Cpx = clinopyroxene, Pl = plagioclase, Kfs = K-feldspar, Qz = quartz, Grt = garnet, Ep = epidote, and Hbl = hornblende.

5. Petrographic and Mineralogical Descriptions

5.1. Overview of Protolith Granite and Albitite

The biotite granite is metasomatically altered to sodic episyenite (albitite). The episy enite was formed by (a) albitization of K-feldspar, (b) vug formation by dissolution of magmatic quartz, and (c) vug infilling by metasomatic minerals such as albite and sugilite. The episyenites are characterized by variable leaching of magmatic quartz and a porous texture (Figure 5) and are recognizable in the field by the whitening of feldspar (albitization), as compared with the unaltered granite.

In Iwagi Islet, quartz-depleted vuggy rocks are commonly observed and, in many samples, vugs with polygonal shapes are infilled with sugilite, aegirine–aegirine-augite, pectolite, calcite, and Fe–Mn-rich minerals, with quartz locally remaining at the rims of the vugs. Newly formed, euhedral albite crystals project into the vugs. The albitites are petrographically and geochemically heterogeneous.

We recognized centimeter-scale zoning in the albitites, and the following facies in the metasomatic rocks of Iwagi Islet from the protolith to the inner zone of metasomatism: (1) original biotite granite, (2) albitized granite, (3) quartz albitite, (4) hedenbergite albitite, (5) aegirine albitite, (6) sugilite albitite, and (7) katayamalite albitite (Figure 6).

The Iwagi albitites possess various accessory mineral assemblages that record episy enitization. We now outline the paragenesis of the accessory minerals. The details of the phase assemblages and micro-textural characteristics are described below. The lithologies and modal mineralogies of the rock types are summarized in Figure 6, Figure 7 and Figure 8.

Quartz comprises 36–45 vol.% (average = 40.1 vol.%) of the protolith biotite granite, 8–40 vol.% (24.4 vol.%) of the albitized granite, 11–13 vol.% (11.8 vol.%) of the quartz albitite, 0.1–1.4 vol.% (0.8 vol.%) of the hedenbergite albitite, 0–4.0 vol.% (1.1 vol.%) of the aegirine albitite, 0.1–0.9 vol.% (0.5 vol.%) of the sugilite albitite, and 0–2.9 vol.% (1.0 vol.%) of the katayamalite albitite (Supplementary Table S2; Figure 8). Thus, some of the albitite granites show a marked decrease in quartz content relative to the protolith biotite granite, and quartz decreases to an average of 11.8 vol.% in the quartz albitite, and further decreases to <1.1 vol.% in the various albitites (Supplementary Table S2).

Alkali feldspar contents are 19–38 vol.% (average = 31.5 vol.%) in the biotite granite, 4–59 vol.% (33.4 vol.%) in the albitized granite, 0.5–3.1 vol.% (1.8 vol.%) in the quartz albitite, 0–5.0 vol.% (2.5 vol.%) in the hedenbergite albitite, 0–0.2 vol.% (0.2 vol.%) in the aegirine albitite, 0–0.6 vol.% (0.3 vol.%) in the sugilite albitite, and 0–0.5 vol.% (0.2 vol.%) in the katayamalite albitite. Alkali feldspar contents in the albitized granite either increase or decrease significantly from those of biotite granite (Figure 7). The various types of albitites have low alkali feldspar contents of ≤2.5 vol.% (Supplementary Table S2).

Plagioclase comprises 19–36 vol.% (average = 25.6 vol.%) of the biotite granite, 31–55 vol.% (40.4 vol.%) of the albitized granite, 83–87 vol.% (84.7 vol.%) of the quartz albitite, 90–93 vol.% (91.7 vol.%) of the hedenbergite albitite, 93–100 vol.% (97.4 vol.%) of the aegirine albitite, 86–99 vol.% (92.8 vol.%) of the sugilite albitite, and 70–95 vol.% (84.5 vol.%) of the katayamalite albitite. The plagioclase content in the albitite granite either increases or decreases significantly from those of biotite granite and increases to an average of 91.7 vol.% in the quartz albitite. Other albitites all contain an average of >84.5 vol.% plagioclase (Supplementary Table S2 and Figure 7).

The biotite granite contains 1.6–3.8 vol.% (2.6 vol.%) biotite, and the albitized granite contains 0.1–2.3 vol.% (1.0 vol.%) biotite and 0–2.3 vol.% (0.4 vol.%) clinopyroxene. Clinopyroxene contents are 1.3–1.5 vol.% (average = 1.4 vol.%) in the quartz albitite, 1.0–2.2 vol.% (1.6 vol.%) in the hedenbergite albitite, 0–2.2 vol.% (0.6 vol.%) in the aegirine albitite, 0.1–3.7 vol.% (1.9 vol.%) in the sugilite albitite, and 1.2–18.2 vol.% (5.5 vol.%) in the katayamalite albitite (Supplementary Table S2).

In addition, the sugilite albitite and katayamalite albitite contain sugilite, katayamalite, pectolite–murakamiite, altered minerals, and unidentified Si–Fe–Ca minerals (Supplementary Table S3; Figure 8)

5.2. Description of Individual Rocks

5.2.1. Biotite Granite Protolith

The coarse-grained biotite granite represents the protolith of albitites in the Iwagi Islet (Figure 2) and has not been affected by albitization. The granite has a hypidiomorphic and porphyritic texture (Figure 9 and Figure 10a) and contains occasional pegmatites and mafic microgranular enclaves. It crops out in relatively low-lying areas near the coasts of the islet (Figure 2). The biotite granite hosting the albitites is homogeneous, and the main minerals are plagioclase (19–27 vol.%; oligoclase–albite), alkali feldspar (30–34 vol.%; mainly microcline), quartz (43–46 vol.%), and biotite (2–4 vol.%; Supplementary Table S2). The granite appears to be unaltered to the naked eye, but under the microscope both the plagioclase and biotite are partly altered. Plagioclase is often altered to sericite in the central parts of grains, and epidote occurs in plagioclase. Biotite is partly altered to chlorite, epidote, and titanite, and zircon and quartz are porous; thus, the host granite has experienced some late-stage hydrothermal alteration.

In the deformed rocks, quartz exhibits undulatory extinction, and large grains consist of smaller grains with interlocking grain boundaries. Plagioclase exhibits kinked twin lamellae and occurs as recrystallized aggregates in some samples. The biotite cleavage is often deformed.

Quartz (~4 mm) contains many micro-pores (several microns in size) and is anhedral. The alkali feldspar exhibits a wide variety of microstructures and varies from coarsely perthitic to pure K-feldspar, replacing plagioclase, and it has a microcline texture. It contains numerous inclusions, resulting in a dusty appearance in thin sections. Plagioclase (~3 mm) is euhedral–subhedral and exhibits albite twinning. Biotite (~2 mm) is the main mafic phase, and occurs commonly as small clusters that contain ilmenite and other accessory minerals. It is subhedral–euhedral and exhibits light yellow to brown pleochroism. Biotite, quartz, and alkali feldspar occur as symplectic intergrowths (Figure 10b).

Accessory minerals include zircon, muscovite, thorite/huttonite, ilmenite, xenotime-(Y), barite, fluorapatite, fluorite, monazite-(Ce), monazite-(Nd), fergusonite, cassiterite, wolframite, sphalerite, pyrite, and galena (Figure 8). This granite is considered to be part of the ilmenite series [37] based on its assemblage of accessory minerals and lack of magnetite.

Zircon (0.04–0.08 mm) is euhedral–subhedral and occurs as inclusions in biotite (Figure 9). It contains numerous tiny thorite/huttonite inclusions (Figure 11a). Thorite mantles uranothorite, and zircon surrounds thorite (Figure 11b). Zircon is partially replaced by monazite-(Ce) (Figure 11c). Some zircons in sample T-69 contain hydrated zircon phase (ZrSiO₄·nH₂O). Thorite/huttonite (2–87 µm) occurs as anhedral grains in biotite and quartz and is often associated with zircon. Muscovite (40 µm) occurs in the cores of plagioclase grains, and is often associated with biotite.

Allanite (200–500 µm) occurs as euhedral–subhedral crystals in biotite, and exhibits oscillatory zoning (Figure 12a). Ilmenite (~100 µm) is subhedral–anhedral and occurs along the cleavage in biotite (Figure 12b). Xenotime-(Y) (22–30 µm) occurs as anhedral crystals in plagioclase and biotite (Figure 12c). Monazite-(Ce) (6–17 µm) occurs only in biotite as anhedral grains, while monazite-(Nd) (2–4 µm across) occurs in alkali feldspar, plagioclase, and quartz as anhedral crystals. Monazite occurs as euhedral–anhedral crystal aggregates and/or isolated single grains and is often associated with zircon.

Fluorapatite (6–23 µm) occurs as anhedral grains in quartz and biotite. It is locally associated with monazite-(Ce) and xenotime-(Y) (Figure 13a). Fergusonite occurs as euhedral–subhedral crystals in plagioclase (Figure 13b). Cassiterite occurs as anhedral grains in plagioclase (Figure 13c).

Fluorite (7–16 µm) occurs as euhedral–anhedral grains in alkali feldspar, quartz, and plagioclase. Barite (15–200 µm) occurs along cleavage planes in biotite. The W-bearing minerals, such as scheelite (4–21 µm) and wolframite (14 µm), occur as anhedral grains in alkali feldspar. Subhedral sphalerite (4–16 µm) occurs in K-feldspar. Euhedral pyrite (6–12 µm) and anhedral galena (2–9 µm) with thorite/huttonite occur in quartz.

5.2.2. Albitized Granite

The albitized granite is transitional between the host biotite granite and quartz albitite (Figure 3) and is characterized by the local occurrence of minerals that provide evidence of metasomatism, i.e., acicular ferro-ferri-holmquistite (Figure 14a), aegirine, aegirine-augite (Figure 14b), hedenbergite, rare arapovite and katayamalite, and abundant recrystallized albite).

We identified grayish transparent quartz, light brown alkali feldspar, white plagio clase, and black aegirine-augite or aegirine in the albitized granite. The albitized granite is similar to the biotite granite in terms of texture but has different mineralogy and chemical composition. The albitization has bleached the original pinkish-gray granite and turned it white.

The main minerals are albite with cloudy oligoclase cores, microcline, quartz, ae girine-augite, and biotite. Fine flakes of biotite are sporadically included in aegirine-augite. A dense network of irregular veins and veinlets of quartz is commonly observed. Alkali feldspar is replaced by albite pseudomorphs, with occasional remnants still present in the newly formed albite. The original normal zoning of the plagioclase is well preserved, with core compositions of An = 0.3–22 mol.%, and rim compositions of An = 0.1–1.3 mol.%. The calcic plagioclase cores contain inclusions of celadonitic muscovite and K-feldspar.

Quartz (0.15–5 mm) is anhedral and locally exhibits wavy extinction. The original plagioclase (0.1–2 mm) is euhedral–subhedral, with albite twinning. The albite aggregates are recrystallized, and a patchwork of albite has replaced K-feldspar (Figure 15a). Alkali feldspar (0.2–3 mm) is subhedral–anhedral and perthitic. K-feldspar remaining in the albitites is partly replaced by other minerals (Figure 15b).

Biotite (0.15–3 mm) is euhedral–anhedral, and often occurs as symplectic inter growths with K-feldspar and quartz (Figure 16). This texture is also observed in the granites that have not undergone metasomatism. Biotite is replaced by recrystallized albite, and secondary albite (tens of microns in size) occurs around and within biotite (Figure 17a). The biotite is partially replaced by ferro-ferri-holmquistite (Figure 17b). Aegirine (0.02–0.5 mm) is subhedral–anhedral and exhibits yellow to green pleochroism.

Accessory minerals include zircon, thorite/huttonite, titanite, ferro-ferri-holmquistite (Figure 14a), muscovite, polylithonite, galena, fluorapatite, fluorite, TiO₂ mineral, xenotime-(Y), baddeleyite, fluorbritholite-(Ce), fluorcalciobritholite, monazite-(Ce) and -(Nd), turkestanite, arapovite, cassiterite, titanite, calcite, sphalerite, and arsenopyrite (Figure 8). Zircon (2–115 µm) is euhedral–anhedral and included in alkali feldspar, plagioclase, quartz, and biotite. Zircon is closely associated with thorite/huttonite (Figure 18a). Some zircons show distinct zoning and contain minute grains of thorite/huttonite. Baddeleyite (3–150 µm) occurs in fractures in quartz and some K-feldspar (Figure 18b,c).

Ferro-ferri-holmquistite (~270 µm long) occurs as aggregates of acicular crystals and exhibits light inky blue interference colors. Anhedral thorite/huttonite (3–90 µm) occurs in quartz and plagioclase and is often included in zircon. Titanite (3–25 µm) occurs in quartz, plagioclase, and alkali feldspar as euhedral–anhedral grains. It often contains Ti-rich veins. Muscovite (10–75 µm) occurs in the cores of plagioclase and is closely associated with biotite.

Polylithionite is rare and occurs in K-feldspar (Figure 19a). Fluorapatite (3–18 µm) occurs in alkali feldspar and quartz as euhedral–anhedral crystals, and often contains inclusions of fluorbritholite-(Ce) (Figure 19b) and REE-rich minerals. Fluorbritholite-(Ce) (3–46 µm) is euhedral–anhedral and occurs in plagioclase and quartz (Figure 19b). Subhedral–anhedral monazite-(Ce) (2–80 µm) occurs in quartz and plagioclase. It is heterogeneous and occurs as aggregates in biotite (Figure 19c).

Fluorite (2–80 µm) occurs in plagioclase and alkali feldspar as subhedral–anhedral crystals. TiO₂ mineral (3–35 µm) occurs in biotite as anhedral crystals, and in plagioclase and alkali feldspar. Xenotime-(Y) (3–20 µm) occurs in biotite as anhedral crystals in close association with zircon and monazite-(Ce). Calcite (~15 µm) is interstitial to quartz and alkali feldspar.

Katayamalite (5–40 µm) occurs rarely as euhedral–anhedral crystals in quartz in a few samples (Figure 20a). Rounded hedenbergite occurs in quartz (Figure 20b). Euhedral turkestanite (14–30 µm; Figure 20c) and anhedral arapovite (5 µm; Figure 20d) occur in quartz and albite, respectively.

Euhedral cassiterite (1–6 µm) occurs in alkali feldspar. Anhedral ilmenite (3–42 µm) occurs mainly along cleavage planes in biotite, and in quartz and alkali feldspar. Galena (1.5–16 µm) occurs as euhedral–anhedral crystals in alkali feldspar and plagioclase. Sphalerite (4 µm) and arsenopyrite (3–8 µm) occur in alkali feldspar.

In a vuggy space, rounded K-feldspar grains occur in TiO₂-rich veins (TiO₂ = 76–82 wt.%, Al₂O₃ = 3.4–4.5 wt.%, SiO₂ = 3.8–4.4 wt.%, and Nb₂O₅ = 0.1–0.6 wt.%), which interconnect along interstitial grains in the manner of fluids in porous media (Figure 21). This indicates that a TiO₂-rich interstitial fluid existed between grain boundaries, and thus may represent an initial stage of episyenitization and the migration path of fluids containing high-field-strength elements (HFSEs).

Titanite in the vugs is decomposed and replaced by a mesh-like structure rich in Ti, Al, and Si (Figure 22). Hematitization (i.e., oxidation) has resulted in the brick-red color of the altered rocks.

Many small veinlets that are several microns wide are developed in the albitized granite. These veinlets occur along crystal boundaries (Figure 23a) or polygonal fractures (Figure 23b). They are rich in rare earth elements (REEs) such as Ce, Y, and Sm, as well as F, Si, Mn, and U (Figure 23). In addition, Zr-rich veins occur locally in the minerals.

5.2.3. Quartz Albitite

Quartz albitite is characterized by the occurrence of >80 vol.% recrystallized albite (Supplementary Table S2; Figure 8). This rock occurs as a marginal face around albitized granite (Figure 3). The albitization is more intense than in the albitized granite. Residual quartz (10.8–12.8 vol.%) and alkali feldspar (0.5–3.1 vol.%) are present. There is a small amount of K-feldspar (0.5–3.1 vol.%) and quartz (10.8–12.8 vol.%). Aegirine and aegirine–augite contents are 1.3–1.5 vol.%. Biotite occurs rarely as aggregates. Pore spaces contain opal (SiO₂·nH₂O).

Increasing albitization is accompanied by a decrease in the modal quartz plus alkali feldspar content and an increase in the recrystallized albite content (Supplementary Table S2; Figure 7 and Figure 8). The quartz albitite is defined as having ~10 vol.% modal quartz. The quartz albitite contains visible white feldspar, gray transparent quartz, reddish brown altered minerals, and trace amounts of black aegirine.

Under the microscope, quartz (0.3–1 mm) is anhedral and often exhibits wavy extinc tion. It contains minute (~0.1 mm) euhedral aegirine. Recrystallized albite (0.25–2 mm) is euhedral–subhedral and exhibits albite twinning. Original plagioclase (0.25–2 mm) is often kink-banded and broken. Alkali feldspar (0.05–0.4 mm) is anhedral. Aegirine (~0.01 mm) occurs as aggregates of anhedral crystals associated with reddish brown altered minerals, and as isolated euhedral–subhedral grains.

Accessory minerals include katayamalite, zircon, fluorapatite, titanite, xenotime-(Y), monazite-(Ce), monazite-(Nd), thorite/huttonite, wollastonite, baddeleyite, zektzerite, gittinsite, miserite, ferro-ferri-holmquistite, fluorbritholite-(Ce), fluorcalciobritholite, galena, arsenopyrite, pyrite, chalcopyrite, and altered minerals (Figure 8). Katayamalite is extremely rare and occurs as independent euhedral small crystals in quartz.

Zircon (~71 µm) is subhedral–anhedral and occurs in albite and quartz in close asso ciation with anhedral aggregates of titanite. It is locally rich in Hf. Porous zircon is ubiquitous in the Iwagi albitites (Figure 24a).

Fluorapatite (3–234 µm) occurs in alkali feldspar and albite as euhedral–anhedral crystals. It contains REE-rich minerals, such as monazite, and occurs as aggregates. Titanite (8–407 µm) occurs as independent anhedral crystals in albite and quartz, and often includes zircon. Heterogeneous xenotime-(Y) (17–240 µm) occurs in quartz and albite as anhedral crystals. Anhedral monazite-(Ce) (1–56 µm) occurs in albite. Vermicular monazite-(Nd) (2–56 µm) occurs in quartz and is often enclosed by fluorapatite. It is decomposed and has a heterogeneous appearance (Figure 24b).

Fluorbritholite-(Ce) and fluorcalciobritholite (3–12 µm) occur in quartz as euhedral–subhedral crystals. Thorite/huttonite (4.5–200 µm) occurs in quartz and albite. Wollastonite (7–26 µm) and zektzerite (6 µm) occur in quartz as anhedral grains. Baddeleyite (2–18 µm) and gittinsite (10–32 µm) are present in albite as anhedral grains (Figure 24c). Miserite (3–9 µm) occurs in quartz as euhedral–subhedral crystals. Acicular ferro-ferri-holmquistite (~55 µm long) occurs in quartz and albite. Euhedral galena (10 µm) occurs in albite. Anhedral arsenopyrite (4–15 µm) occurs in quartz. Pyrite (2–5 µm) and chalcopyrite (1.5 µm) occur in albite as anhedral grains.

A characteristic intergrowth of elongated zircon crystals was observed in a quartz albitite vug (Figure 25). These vugs are not observed in the other albitites.

The quartz albitite also contains a vug that is filled with fluorapatite, monazite, and K-feldspar (Figure 26).

Veins with a width of several microns to several millimeters occur in all the albitites (Figure 27). These contain mainly Si, Al, Mn, Na, P, Ba, and REEs, which were injected along the veinlets. Some veinlets contain >50 wt.% SiO₂, MnO, and Ce₂O₃, and some contain 1–13 wt.% FeO, 5–10 wt.% P₂O₅, 6–9 wt.% BaO, and 1–2.5 wt.% F.

5.2.4. Hedenbergite Albitite

The hedenbergite albitite is defined by the appearance of hedenbergite, andradite, wollastonite, titanite, kristiansenite, and magnetite, and is dominated by albite (90–93 vol.%) and contains minor quartz (0.1–1.4 vol.%) and alkali feldspar (0–5.0 vol.%; Supplementary Table S2). The albitite contains visible white plagioclase, light brown alkali feldspar, black hedenbergite, dull brown altered minerals, and small quartz grains (Figure 6).

Albite (0.2–3 mm) is subhedral–anhedral and exhibits albite and cross-hatched twin ning. Aggregates of numerous, small, clear granular crystals occur at the grain boundaries of large broken and deformed albite grains. A polygonal albite vug is shown in Figure 28a. The albites around the vug and inner euhedral albite are pure albites (Ab > 99 mol.%). The “fluffy” mineral inside the vug is an unidentified Si–Fe–Al mineral. Alkali- feldspar (0.3–0.9 mm) is anhedral and occurs in quartz and albite.

Hedenbergite (10–100 µm) occurs as anhedral aggregates and isolated euhedral–anhedral grains (Figure 28b), with the former often being associated with reddish altered minerals. Quartz (0.3–1 mm) is anhedral.

Accessory minerals include zircon, titanite, xenotime-(Y), fluorapatite, monazite-(Ce), -(Nd), and -(La), andradite, wollastonite, fluorite, magnetite (Figure 28c), thorite/huttonite, ferro-ferri-holmquistite, kristiansenite, gittinsite, turkestanite, and altered minerals, such as an unknown Si–Fe mineral (Figure 8).

Zircon (5.7–73 µm) occurs in albite and quartz as subhedral–anhedral grains. Titanite (3–120 µm) is anhedral and associated with hedenbergite and altered Si–Fe minerals. Titanite also occurs as aggregates in close association with zircon in quartz and albite at the grain boundaries of hedenbergite and altered minerals. Some titanite is Sn-rich (Figure 29). Xenotime-(Y) (5–78 µm) occurs in albite as anhedral crystals. Fluorapatite (3–44 µm) occurs in quartz, albite, and K-feldspar as euhedral–anhedral crystals. Monazite-(Ce) (15–52 µm) and monazite-(Nd) (8–23 µm) occur as anhedral crystals in albite and K-feldspar, respectively. Monazite-(La) is only found in this rock type. Andradite (18–120 µm) is closely associated with titanite (Figure 30).

Wollastonite (8–16 µm) is replaced by pectolite and quartz (Figure 31). Fluorite (~12 µm) occurs in K-feldspar as anhedral crystals. Magnetite (2–25 µm) occurs in albite as euhedral grains (Figure 28c). Thorite/huttonite (~70 µm) is heterogeneous and occurs in albite. Ferro-ferri-holmquistite (~14 µm long) is acicular and occurs in albite. Kristiansenite (5–15 µm) occurs in the interstices between quartz, albite, and K-feldspar, and has heterogeneous Sn contents (Figure 32). Gittinsite (2 µm) is closely associated with titanite. Turkestanite (2–5 µm) occurs in albite. A vug filled with quartz and titanite is shown in Figure 33.

5.2.5. Aegirine Albitite

The aegirine albitite is characterized by aegirine, aegirine-augite, pectolite–murakamiite, tainiolite, and dalyite, and a high albite content (93.0–99.7 vol.%; average = 97.4 vol.%; Supplementary Table S2; Figure 8). White to pale pink albite and black aegirine are evident with the naked eye (Figure 34). This aegirine albitite (a) shows a gradual transition to sugilite albitite (b) and katayamalite albitites (c) on the scale of a hand specimen (Figure 34).

Some of the aegirine albitites contain aggregates of large aegirine crystals up to 9 mm in size. The major minerals include albite, quartz, aegirine, alkali feldspar, and pectolite–murakamiite. Albite (~6 mm) is subhedral–anhedral, and aggregates of numerous small granular crystals occur at the grain boundaries of large broken and deformed albite grains. The albite exhibits cross-hatched twinning (Figure 35). Quartz (0.5–1.7 mm) occurs as cylindrical crystals associated with aegirine and an unknown Si–Fe–Ca mineral. Opal with a string-like or rounded irregular shape occurs in a vug (Figure 36a). Rectangular quartz also occurs in aegirine-augite (Figure 36b).

Aegirine (0.01–0.3 mm) occurs as anhedral aggregates and isolated euhedral–anhe dral grains. It locally exhibits zoning (Figure 37), with Mg- and Ca-rich cores (aegirine-augite) and Na-rich rims (aegirine).

Aegirine is often associated with an unidentified Si–Fe–Ca mineral. Small amounts of alkali feldspar occur as anhedral crystals in quartz and albite. Pectolite (0.04–1.0 mm) occurs as anhedral crystals in albite. Accessory minerals include zircon, fluorapatite, zektzerite, calcite, tainiolite, wollastonite, fluorbritholite-(Ce), monazite-(Ce), dalyite, an unknown Si–Th–Ca mineral, turkestanite, truscottite, and sphalerite (Figure 8).

Zircon (3–40 µm) is euhedral–anhedral. An unknown Si–Fe–Ca mineral (4–17 µm) is anhedral, and associated with aegirine and quartz. These are commonly associated with quartz. Fluorapatite (2–90 µm) occurs as euhedral–anhedral crystals in albite and is often associated with alkali feldspar.

Zektzerite (10–65 µm) is an extremely rare mineral and its occurrence on Iwagi Islet [21] is the only report of this mineral in Japan. It occurs as independent grains enclosed by albite and can also include zircon grains (Figure 38a).

Tainiolite (90 µm) occurs as anhedral crystals in quartz and albite (Figure 38b). Wol lastonite (23–150 µm) occurs as subhedral–anhedral crystals in albite. Fluorbritholite-(Ce) (2–22 µm) occurs as anhedral grains at the boundaries between quartz and K-feldspar. Monazite-(Ce) (4–11 µm) occurs in albite and is often associated with zircon. Dalyite is also a very rare Zr-bearing mineral and, in Japan, was first described from Iwagi Islet by Imaoka et al. [21]. It occurs as an anhedral colorless grain (14 × 29 μm) enclosed by albite (Figure 38c). Calcite (12–45 µm) occurs in albite with K-feldspar, and often encloses pectolite.

An unknown Si–Th–Ca mineral (3.8–17 µm) is heterogeneous in appearance and oc curs as anhedral crystals in quartz. Turkestanite (5 µm) occurs as euhedral–anhedral crystals in albite. Sphalerite (5 µm) occurs as subhedral crystals with K-feldspar in albite. Truscottite (5–70 µm) occurs as petal-like crystals (Figure 39).

A fine, mesh-like titanium material and small amounts of aegirine are only observed in a vug (Figure 40).

5.2.6. Sugilite Albitite

The sugilite albitite (Figure 41a) is defined by sugilite and murakamiite, and is dominated by albite (>92.8 vol.%; Supplementary Table S2; Figure 8). White and transparent albite, olive green sugilite, and black aegirine are visible to the naked eye. Other major minerals are quartz and K-feldspar. Albite (0.2–5 mm) is subhedral–anhedral, and cross-hatched-twinning is often present. Some crystals are deformed (Figure 41b).

Sugilite (~0.1–4 mm) is anhedral and closely associated with aegirine and pectolite–murakamiite. Aegirine (0.05–0.4 mm) occurs as anhedral aggregates and isolated euhedral–anhedral crystals. Figure 42a shows tiny euhedral aegirine crystals. Aegirine is closely associated with sugilite and pectolite–murakamiite. Very small amounts of K-feldspar occur at the grain boundaries of albite. Small amounts of quartz (0.1–1.1 mm) occur as anhedral grains. Pectolite–murakamiite (0.05–0.65 mm) is subhedral–anhedral and is associated with quartz.

Accessory minerals include zircon, fluorapatite, thorite/huttonite, turkestanite, fluorbritholite-(Ce), fluorcalciobritholite, truscottite, an unknown Si–Th–Ca mineral, and galena (Figure 8). Zircon (6–60 µm) occurs as subhedral–anhedral grains in albite. Fluorapatite (2–710 µm) occurs as isolated grains in albite (Figure 42b) and as aggregates. Thorite/huttonite (2–100 µm) is heterogeneous in size and occurs in albite and quartz. Turkestanite (3–11 µm) occurs in albite (Figure 42c) and aegirine-augite. Fluorbritholite-(Ce) (1–12 µm) occurs in K-feldspar and albite. Fluorcalciobritholite (3 µm) occurs with aegirine in albite. Truscottite (5–80 µm) occurs as petal-shaped crystals.

An unknown Si–Th–Ca mineral (15 µm) occurs as anhedral crystals in albite. Galena (3 µm) occurs in albite with K-feldspar. The mafic minerals are replaced by argillitic minerals.

Quartz exhibits partial to complete dissolution (e.g., [4,6]) and the voids are filled by calcite, sugilite, aegirine, pectolite, and opalized quartz. Figure 42d shows a vug filled with botryoidal opal (SiO₂∙nH₂O). Figure 43 and Figure 44 show a vug with opalized quartz at its edges as shown in pink on the Si color map. Sugilite is widely distributed in the vug, and aegirine is also present (Figure 44). The black areas around the opalized quartz are areas where no minerals were reprecipitated after quartz dissolution (Figure 43).

5.2.7. Katayamalite Albitite

The katayamalite albitite is characterized by ≥0.1 vol.% katayamalite (Supplementary Table S2). It is heterogeneous, especially in the color index even at the scale of a hand specimen (Figure 6 and Figure 34), and sometimes exhibits aggregate (see Supplementary Figure S2). It is also heterogranular and dominated by fine- to medium-grained albite. It is typified by a lack of quartz (<2.9 vol.%) and alkali feldspar (<0.5 vol.%), and consists mainly of albite (70.2–94.7 vol.%), aegirine (2.2–18.2 vol.%), pectolite–murakamiite (0.6–4.8 vol.%), and lesser sugilite (~9.6 vol.%) and katayamalite (0.1–2.4 vol.%; Supplementary Table S2), and rare K-feldspar. A melanocratic area that contains up to 18 vol.% aegirine–aegirine-augite is distributed irregularly in the leucocratic albite-rich area.

Albite is a major constituent of the katayamalite albitite and makes up ≤94.7 vol.% of the most leucocratic rocks. It occurs as subhedral–anhedral crystals of 0.1–5 mm in size. The following three types of occurrence were noted by Murakami and Matsunaga [38]: (1) large anhedral crystals with parallel, fine lamellae, which exhibit polysynthetic albite and cross-hatched twinning (Figure 45a,b) and are considered to have formed by the complete replacement of K-feldspar by albite [39,40], (2) large subhedral crystals that are untwinned or show albite twinning (Figure 45a,b), and (Figure 45c) aggregates of numerous, small, clear, granular neoblasts at the boundaries of large albite grains (Figure 45a,c).

Deformation microstructures are evident in the larger primary albite grains (Figure 46). Undulatory extinction and deformation twins are the most common microstructures, along with curved and bent twin lamellae.

Pectolite–murakamiite is colorless in thin sections and occurs in mosaic aggregates or as isolated anhedral crystals (0.05–1.7 mm) with distinct {100} and {001} cleavages. Murakamiite is indistinguishable from pectolite to the naked eye and under the microscope, because the chemical composition changes gradually from Li-dominant (murakamiite) to Na-dominant (pectolite). Pectolite–murakamiite is often replaced by calcite, and numerous calcite veinlets are evident in Figure 47. Mosaic aggregates consist of pectolite–murakamiite, sugilite, aegirine, albite, and katayamalite (Figure 47).

The katayamalite (0.1–2 mm long) is colorless and clear without dusty inclusion in the thin section, and exhibits deep bluish interference colors. It is usually euhedral with perfect (001) cleavage and twinning. It exhibits bright blue–white luminescence under a short wavelength (253.7 nm) [41].

Sugilite is almost colorless under the microscope, but alters to a yellowish-brown, dusty mineral. Minute grains of pectolite, albite, and aegirine are sporadically present in the sugilite. Aegirine-augite appears to have formed along cracks in albite (Figure 48a) and is pleochroic olive green to yellow. K-feldspar and quartz occur as small relict minerals.

Accessory minerals include zircon, fluorapatite, calcite (Figure 48b), fluorbritholite-(Ce), fluorcalciobritholite, zektzerite, titanite, turkestanite, monazite, thorite/huttonite, wollastonite, truscottite, and chalcopyrite (Figure 8). Zircon (3–45 µm) is subhedral–anhedral and occurs in albite.

The zircon shows well-developed zoning. Fluorapatite (2–150 µm) is euhedral–anhedral and occurs in albite and aegirine-augite, and is frequently associated with britholite group minerals. Calcite (3–325 µm) occurs as veins and pools (Figure 48b) in albite and pectolite, and can include pectolite. It also occurs as interstitial grains between the reprecipitated silica, aegirine-augite, and katayamalite. Calcite is modally abundant in the katayamalite albitites as compared with the other albitites (Figure 8).

Fluorbritholite-(Ce) (2–31 µm) and fluorcalciobritholite (4–20 µm) are anhedral, and the former occurs in pectolite and albite, while the latter occurs in aegirine-augite. Titanite (12–20 µm) is present in katayamalite. Turkestanite (4–10 µm) occurs as euhedral–anhedral crystals in albite. Anhedral monazite (4 µm) occurs in albite. Anhedral thorite/huttonite (3–14 µm) occurs in albite and associated aegirine. Acicular to fibrous aggregates of wollastonite (~330 μm long) occur in pectolite–murakamiite (Figure 49a). Figure 49b shows pectolite–murakamiite replaced by calcite. Figure 49c shows calcite veins in pectolite–murakamiite. Truscottite (~100 µm) occurs as petal-shaped crystals. Anhedral chalcopyrite (5.5 µm) occurs in fractures in albite.

6. Mineral Chemistry

As described in the previous section, qualitative and quantitative EPMA analysis confirmed that the samples from Iwagi Islet contain a variety of minerals, including silicate, phosphate, halide, oxide, sulfide, carbonate, sulfate, and tungsten minerals. In this section, we document the chemical compositions of some of the silicate, phosphate, and oxide minerals. Their ideal chemical formulae were taken from the RRUFF website [36].

6.1. Silicate Minerals

6.1.1. Alkali Feldspar

Alkali feldspar (KAlSi₃O₈–NaAlSi₃O₈) occurs in all the rock types. In the host biotite granite and albitized granite, alkali feldspar mainly occurs as perthite. Perthitic albite is large in size and abundant, with variable Ab compositions. Thus, it is difficult to determine the bulk chemical composition of the perthitic alkali feldspars by EPMA, and we determined the composition of K-feldspar in the part excluding the perthites.

As albitization proceeded, the perthitic albite was replaced and the modal amount of alkali feldspar decreased (Supplementary Table S2). Orthoclase components of the K-feldspars are 90.9–98.2 mol.% for biotite granite, 92.0–98.3 mol% for albitized granite, 90.1–98.2 mol.% for the quartz albitites, 92.7–97.9 mol.% for the hedenbergite albitites, 94.4–99.7 mol.% for the aegirine albitites, 90.2–98.0 mol.% for the sugilite albitites, and 88.8–100.0 mol.% for the katayamalite albitites. Except for some crystals, the Or contents are >90 mol.% (Supplementary Table S3).

6.1.2. Plagioclase

Plagioclase (NaAlSi₃O₈–CaAlSi₂O₈) occurs in all the studied rock types and increases in abundance with the degree of albitization. Plagioclase in the biotite granites and albitized granites has a relatively wide compositional range from albite to oligoclase (Ab = 79.0–98.8 mol.% for biotite granite; 77.5–99.7 mol.% for albitized granite; Supplementary Table S3). In contrast, the plagioclase in the albitites has nearly end-member Ab compositions (Supplementary Table S3): Ab = 99.3–99.5 mol.% for the quartz albitites; 98.7–99.6 mol.% for the hedenbergite albitites; 99.1–99.7 mol.% for the aegirine albitites; 99.1–99.5 mol.% for the sugilite albitites; and 98.6–99.6 mol.% for the katayamalite albitites.

6.1.3. Biotite

Biotite, which has an ideal formula of K(Mg,Fe)₃AlSi₃O₁₀(F,OH)₂, occurs in the biotite granite and some of the albitized granites. The biotite data are listed in Supplementary Table S3. The Mg/(Fe + Mg) ratios are low (0.028–0.033), and the TiO₂ contents of biotite in the unaltered granite are 0.99–1.85 wt.%. The biotite has K₂O contents of 8.55–9.42 wt.%.

6.1.4. Clinopyroxene

There are three types of clinopyroxene in the samples from Iwagi Islet. The former two are aegirine to aegirine-augite (sodic pyroxene), which is formed by Na metasomatism and occurs widely in the albitized granite and all types of albitites. The latter is hedenbergite, which occurs in the hedenbergite albitite and is also found in the albitized granite.

The analytical results were also plotted on the Q–J diagram with the vertical axis represent ing Ca + Mg + Fe²⁺ and the horizontal axis representing twice the value of Na (Figure 50) after Morimoto et al. [42]. Normal pyroxenes plot in the area defined by the lines Q + J = 2 and Q + J = 1.5. The Iwagi pyroxenes show a very wide compositional variation on the Q + J = 2 line, and are classified as Ca–Mg–Fe pyroxenes (Quad), Ca–Na pyroxenes, and Na pyroxenes.

The clinopyroxene data were also plotted on a Q (Wo + En + Fs)–Jd–Ae ternary diagram, shown in Figure 51, and were classified as aegirine-augite, aegirine, and hedenbergite. The aegirine-augite and aegirine compositional data are listed in Supplementary Table S3.

The Ca–Na and Na pyroxenes in the albitized granite are almost all aegirine-augite (aegirine component, NaFe³⁺Si₂O₆, Ae = ~52.6 mol.%), but a small amount is aegirine (Figure 51). The quartz albitite contains mainly aegirine-augite (Ae = ~50.1 mol.%) and some aegirine (Ae = ~91.2 mol.%). The Ca-Na pyroxenes in the hedenbergite albitite has an average Ae content of 38.9 mol.% (Supplementary Table S3), which is lower than that in the aegirine albitite, sugilite albitite, and katayamalite albitite.

The Ca–Na and Na pyroxenes in the aegirine albitite, sugilite albitite, and katayama lite albitite vary from aegirine-augite (Ae = 24.0 mol.%~) to aegirine (Ae = ~94.2 mol.%). As such, the compositions of the clinopyroxene vary widely from aegirine-augite to pure aegirine, and those in the albitized granite, quartz albitite, and hedenbergite albitite have lower Ae contents than those in the aegirine albitite, sugilite albitite, and katayamalite albitite (Supplementary Table S3).

The chemical compositions of hedenbergite in the albitized granite are Wo = 39.7–49.2 mol.%, En = 1.5–6.3 mol.%, and Fs = 49.7–51.3 mol.%, and those in the hedenbergite albitite are Wo = 47.7–49.5 mol.%, En = 2.8–9.5 mol.%, and Fs = 41.9–49.0 mol.%.

6.1.5. Sugilite

Sugilite is a Na–K–Fe³⁺ silicate with a milarite-type structure that was first reported by Murakami et al. [17]. Sugilite has an ideal formula of Na₂K(Fe³⁺,Mn³⁺,Al)₂Li₃Si₁₂O₃₀ [43], and the crystal structure was determined by Kato et al. [44]. Although the Iwagi sugilite is Mn-free, sugilite in Mn deposits is Mn-bearing [45,46,47], and manganoan sugilite containing up to 7.04 wt.% MnO has been reported from the Furumiya mine in Japan [48].

The Iwagi sugilite is Al₂O₃-poor as compared with the Hoskin and Woods mines [47]. The Iwagi sugilite is Na₂O- and K₂O-poor as compared with the Wessels, Hoskins, and Woods mines. The Iwagi sugilite has Fe > Al > Mn, and that from the Woods and Hoskins mines has Al > Fe [47]. The Mg/Fe ratio has the lowest value of 0.0013. The LiO₂ content of separated sugilite reported by Murakami et al. [17] is 2.88 wt.%. Similar to katayamalite, the sugilite is rich in the HFSEs, such as TiO₂ (0.7–1.1 wt.%), but TiO₂ contents are lower by one to two orders of magnitude as compared with the katayamalite.

Sugilite has also been reported in bedded seams of massive material interlayered with fine-grained aegirine and Mn oxides in the Wessels mine, South Africa [45].

6.1.6. Katayamalite

Katayamalite, which is a Ca–Li–Ti silicate with an ideal formula of KLi ₃Ca₇Ti₂(SiO₃)₁₂(OH)₂, was first reported by Murakami et al. [18]. The structure was originally reported to be triclinic [49], similar to that of baratovite. How ever, Baur and Kessner [50] showed that baratovite and katayamalite are both monoclinic and structurally identical. Andrade et al. [51] redetermined the structure of katayamalite from the type locality at Iwagi Islet as being monoclinic and assigned it as an OH-dominant analog of the baratovite-group. Katayamalite (baratovite) is a very rare mineral and the only other known occurrence is in metasomatic quartz–albite–aegirine rock in the Dara-i-Pioz alkaline igneous complex and moraine rocks of the Dara-i-Pioz glacier in northern Tajikistan [52,53,54]. At this locality, baratovite is associated with miserite, ekanite, sugilite, and titanite.

Katayamalite occurs in some samples of the albitized granite and quartz albitite, in addition to the katayamalite albitite. The katayamalite is often associated with aegirine, pectolite, and sugilite (Figure 47). The katayamalite is rich in TiO₂ (10.99 wt.%), CaO (28.25 wt.%), and Li₂O (3.25 wt.%).

6.1.7. Pectolite–Murakamiite

The pectolite (NaCa₂Si₃O₈[OH])–murakamiite (LiCa₂Si₃O₈[OH]) series involves a solid solution with murakamiite, which is a Li analog of pectolite that was discovered in a metasomatic albitite at Iwagi Islet by Imaoka et al. [19]. Pectolite–murakamiite series minerals form during the advanced stages of albitization, occur in the aegirine albitite, sugilite albitite and katayamalite albitite, and are often associated with sugilite, aegirine, and fluorapatite (Figure 45 and Figure 47).

When the cation total is calculated using O = 9, the number of Na cations is 0.46, and Li is 0.55, i.e., substitution by Li of Na occurs. This substitution is also consistent with the lattice parameters. The murakamiite has Li/(Li + Na) > 0.5, and we propose that this murakamiite is a Li analog of pectolite [19]. The murakamiite has a wide range of Li/(Li + Na) ratios from 0.52 to 0.60, indicating it has a highly variable composition. The EPMA and LA–LIBS results are consistent with those obtained by LA–ICP–MS [19]

Murakamiite is associated with pectolite with Li/(Li + Na) ratios of 0.46–0.50. Therefore, murakamiite may also be a major sink for Li, as well as katayamalite, sugilite, ferro-ferri-holmquistite, and tainiolite.

6.1.8. Ferro-Ferri-Holmquistite

Ferro-ferri-holmquistite has an ideal formula of □Li₂(Fe²⁺₃Fe³⁺₂)Si₈O₂₂(OH)₂, and was first reported by Nagashima et al. [20]. It typically occurs in the low-grade albitized rocks, such as the albitized granite, quartz albitite, and quartz veins near the albitites, and in a few samples of hedenbergite albitite. Li₂O contents are up to 3.64 wt.%. FeO and Fe₂O₃ contents are high (12.39 and 26.23 wt.%, respectively), and Al₂O₃ (3.04 wt.%) and MgO (0.99 wt.%) contents are low.

6.1.9. Zircon

Zircon (ZrSiO₄) is present in all the studied rocks. Representative analyses of the zircons are listed in Supplementary Table S3. HfO₂ contents of the zircons are high in the granites (1.1–10.2 wt.%). Zircons in the albitites generally have low HfO₂ contents of 1–3 wt.%. The measured contents of other elements are ~0.88 wt.% Y₂O₃, ~0.16 wt.% TiO₂, ~0.78 wt.% ThO₂, and ~0.60 wt.% UO₂.

6.1.10. Thorite/Huttonite

Thorite and huttonite have the same chemical formula ThSiO₄, but different structures (i.e., they are dimorphs). Thorite is tetragonal and huttonite is monoclinic. We have not determined the X-ray properties of the thorite/huttonite, and thus cannot ascertain if it is thorite or huttonite. The thorite/huttonite occurs in the host rocks and weakly Na-metasomatized rocks (i.e., biotite granite, albitized granite, quartz albitite and a few samples of the hedenbergite, sugilite, and katayamalite albitites), but is not present in the aegirine albitite.

Representative analyses of the thorite-like minerals are listed in Supplementary Table S3. ThO₂ contents vary widely from 48.7 to 66.7 wt.%, and UO₂ contents vary widely from 12.4 to 23.9 wt.%. The thorite-like mineral is particularly rich in UO₂ (e.g., sample T-69: ~23.91 wt.%) and is often called uranothorite (Figure 10b). SiO₂ contents vary from 15.9 to 17.4 wt.% and P₂O₅, Y₂O₃, REE₂O₃, ZrO₂, and F contents are low. The low totals are probably due to alteration and amorphization by metamictization, along with the variable H₂O contents.

6.1.11. Titanite

Titanite (CaTiSiO₅) occurs in the albitized granite and quartz, hedenbergite, and katayamalite albitites, and is not present in the host biotite granite (Figure 8). Therefore, titanite is a characteristic mineral of albitization. Representative analyses of the titanite are listed in Supplementary Table S3. The TiO₂ contents are high in the katayamalite albitite (37.8 wt.%), and low in the albitized granite (26.7–34.0 wt.%), quartz albitite (23.6–28.0 wt.%), and hedenbergite albitite (30.3–35.0 wt.%). In the albitized granite, titanite is enriched in Nb₂O₅ (~12.5 wt.%). Titanite intergrown with zircon (analyses 250 and 692) in the quartz albitite is rich in ZrO₂ (~6.8 wt.%). The Sn-rich titanite (SnO₂ = ~2.7 wt.%) in the hedenbergite albitite is associated with kristiansenite, indicating late-stage formation and the concentration and transportation of Sn during hydrothermal processes. The measured contents of other elements are 0.6–0.8 wt.% FeO, 0.1–2.1 wt.% F, ~2.0 wt.% Na₂O, ~0.8 wt.% K₂O, and ~0.7 wt.% Y₂O₃.

6.1.12. Andradite

Andradite has an ideal formula of Ca₃Fe³⁺₂(SiO₄)₃ and was only recognized in the hedenbergite albitite. A representative analysis of the andradite is listed in Supplementary Table S3. The compositional variation is limited.

6.1.13. Tainiolite

Tainiolite (KLiMg₂Si₄O₁₀F₂) is only rarely found in the aegirine albitite [22]. Representative analyses are listed in Supplementary Table S3. Li₂O contents were estimated based on ideal stoichiometry. The MgO contents range from 19.76 to 20.47 wt.%, and K₂O contents from 11.73 to 12.11 wt.%.

6.1.14. Zektzerite

Zektzerite (NaZrLiSi₆O₁₅) was found in some quartz, aegirine, and katayamalite albitite samples [22]. Representative analyses of the zektzerite are listed in Supplementary Table S3, with the Li₂O value (2.82 wt.%) from Dunn et al. [55]. The chemical composition is as follows: SiO₂ = ~66 wt.%, Na₂O = 5–6 wt.%, ZrO₂ = ~22 wt.%, and trace amounts of HfO₂, Al₂O₃, and FeO.

6.1.15. Gittinsite

Gittinsite (CaZrSi₂O₇) occurs in a few samples of the quartz and hedenbergite albitites (Figure 8 and Figure 24c). Representative analyses of the gittinsite are listed in Supplementary Table S3. No large compositional variations were observed. This mineral has been reported to occur in a peralkaline granite in western Mongolia [56].

6.1.16. Turkestanite

Turkestanite has an ideal formula of (K,◻)(Ca,Na)₂ThSi₈O₂₀·nH₂O, and is a rare Th- and REE-bearing silicate. It occurs as rounded, isolated, small grains within albite and aegirine grains in only a few samples of albitized granite and hedenbergite, aegirine, sugilite, and katayamalite albitites.

Representative analyses of the turkestanite are listed in Supplementary Table S3. Measured values for SiO₂ = 54.3–56.7 wt.%, ThO₂ = 15.1–24.2 wt.%, UO_{2 =} 1.1–5.7 wt.%, CaO = 7.3–12.0 wt.%, Na₂O = 0.6–3.4 wt.%, and K₂O = 4.4–5.2 wt.%. The (REE + Y)₂O₃ contents range from 1.5 to 9.1 wt.%. The empirical structural formula shows that the A site is occupied by Th (0.486–0.766 apfu), U (0.033–0.185 apfu), REE + Y (0.060–0.479 apfu), and minor Fe³⁺ (~0.015 apfu), while the B site contains residual Ca (1.084–1.807 apfu), Na (0.144–0.905 apfu), and REEs (0.000–0.078 apfu). The C site is occupied partially by K (0.796–0.939 apfu) and occasionally by Na (up to 0.069 apfu), leaving a small vacancy of up to 0.226 apfu.

6.1.17. Arapovite

Arapovite has an ideal formula of (K_1−x◻_x)(Ca,Na)₂U⁴⁺Si₈O₂₀ [x ≈ 0.5]. It was only found in one sample of albitized granite (Figure 42). A representative analysis of the arapovite is listed in Supplementary Table S3. It has high UO₂ and ThO₂ contents of 14.14 and 11.58 wt.%, respectively. Measured values for SiO₂, CaO, Na₂O, K₂O, and (REE + Y)₂O₃ are 55.2, 10.3, 1.3, 2.8, and 2.6 wt.%, respectively. On the basis of 8 Si apfu, the following empirical structural formula was obtained: (K_0.52□_0.48)_Σ1.0(Ca_1.60Na_0.37Ba_0.01)_Σ1.980 (U⁴⁺_0.46Th_0.38Zr_0.01Y_0.06La_0.02Ce_0.04Nd_0.03Sm_0.01Pr_0.01)_Σ1.02Si₈O₂₀, with an ideal arapovite formula.

6.1.18. Wollastonite

Wollastonite (CaSiO₃) is present in the quartz albitite and some of the hedenbergite, aegirine, and katayamalite albitites. It occurs as anhedral, isolated grains in quartz and albite (Figure 8 and Figure 49). Representative analyses of the wollastonite are listed in Supplementary Table S3. Wollastonite exhibits little variation in composition within a sample or between different samples: SiO₂ = 50.3–51.7 wt.%, CaO = 46.4–47.5 wt.%, mean FeO = 0.32 wt.% (n = 26), mean MnO = 0.50 wt.%, and MgO < 0.04 wt.%.

6.1.19. Kristiansenite

Kristiansenite has an ideal formula of Ca₄Sc₂Sn₂(Si₂O₇)₂(Si₂O₆OH)₂. It was only recognized in the hedenbergite albitite, in which it occurs as anhedral grains in albite and quartz (Figure 32). Representative analyses of the kristiansenite are listed in Supplementary Table S3. The kristiansenite yields the following compositional range: SiO₂ = 37.5–42.8 wt.%, FeO = 2.85–5.20 wt.%, CaO = 16.5–18.9 wt.%, ZrO₂ = 4.37–8.32 wt.%, SnO₂ = 14.11–19.10 wt.%, and Sc₂O₃ = 5.64–8.66 wt.%.

6.1.20. Truscottite

Truscottite is a calcium silicate mineral and has an ideal formula of Ca₁₄Si₂₄O₅₈(OH)₈·2H₂O. It occurs in the aegirine albitite, sugilite albitite, and katayamalite albitite. The chemical composition of the truscottite is listed in Supplementary Table S3, showing SiO₂ = 58.1 wt.% and CaO = 29.9 wt.%. The low TiO₂ content of 0.05 wt.% indicates limited replacement of Si by Ti. The low contents of Al₂O₃, FeO, MgO, Na₂O, and K₂O (0.12, 0.50, 0.13, 0.31, and 0.10 wt.%, respectively) indicate that the replacement of Ca by Al, Fe, Mg, Na, and K is limited.

6.1.21. Miserite

Miserite is a rare K–Ca–Y rare earth silicate with an ideal formula of K_1.5−x(Ca,Y,REE)₅(Si₆O₁₅)(Si₂O₇)(OH,F)₂yH₂O. It occurs in the albitized granite and quartz albitite. A representative analysis is listed in Supplementary Table S3. Major elements are as follows: SiO₂ = 53.66 wt.%, CaO = 29.80 wt.%, K₂O = 6.99 wt.%, REE₂O₃ = 4.42 wt.%, F = 5.72 wt.%.

6.1.22. Dalyite

Dalyite (K₂ZrSi₆O₁₅) occurs in some of the aegirine al bitites [21]. The analyses show it to be K₂ZrSi₆O₁₅, and the major element contents vary as follows: SiO₂ = 61.24–62.00 wt.%, K₂O = 15.67–16.65 wt.%, and ZrO₂ = 21.02–21.31 wt.%. There is a significant substitution of Hf for Zr, and an insignificant replacement of K by Ba and Na (Supplementary Table S3).

6.2. PhosphateMineral

6.2.1. Fluorapatite

Fluorapatite has an ideal formula of Ca₅(PO₄)₃F and was identified in all the rock types, with its modal abundance increasing with increasing albitization (Figure 8). Representative analyses of the fluorapatite are listed in Supplementary Table S3. Most samples contain fluorapatite (≥2 wt.% F, up to 4.8 wt.%) with Cl contents at or below EMPA detection limits. In addition to CaO, P₂O₅, and F, REEs and minor amounts of SiO₂ (2.95 wt.%) and Al₂O₃ (0.81 wt.%) are present in the fluorapatite. The fluorapatite in the aegirine albitite has a relatively high Sr content of up to 2.31 wt.%.

6.2.2. Monazite-(Ce), and -(Nd)

There are two types of monazite (LREE,Th,U,Ca)(P,Si)O₄ in Iwagi Islet: monazite-(Ce) and -(Nd). Monazite-(Ce) occurs in the biotite and albitized granites, and the hedenbergite albitite and aegirine albitite, while monazite-(Nd) occurs in the biotite granite and quartz albitite (Figure 8). Representative analyses of the monazite-(Ce) and -(Nd) are listed in Supplementary Table S3, respectively. The monazite-(Ce) contains 51.3–68.0 wt.% REE₂O₃ (average = 62.6 wt.%; n = 36) and 1.64–20.7 wt.% ThO₂ (average = 6.8 wt.%). The monazite-(Nd) in the quartz albitite contains 57.1–64.1 wt.% REE₂O₃ (average = 56.2 wt.%; n = 12) and 0.0–15.5 wt.% ThO₂ (average = 2.9 wt.%).

6.2.3. Xenotime-(Y)

Xenotime-(Y) (YPO₄) is a rare earth phosphate mineral that is present in the host bio tite granite, quartz albitite and hedenbergite albitite, and a few samples of aegirine albitite. Representative analyses of the xenotime-(Y) are listed in Supplementary Table S3. REE₂O₃ (excluding Y₂O₃) contents of analyses IW-303 and IW-3 are 23.9 and 18.4 wt.%, respectively, and the xenotime-(Y) has high contents of heavy REEs.

6.2.4. Fluorbritholite-(Ce) and Fluorcalciobritholite

The britholite-group minerals are relatively common accessory minerals that contain the REEs. These include britholite-(Ce) (Ce,Ca)₅(SiO₄)₃OH, britholite-(Y) (Y,Ca)₅(SiO₄)₃OH, fluorbritholite-(Ce) (Ce,Ca)₅(SiO₄)₃F, fluorbritholite-(Y) (Y,Ca)₅(SiO₄)₃F, and fluorcalciobritholite (Ca,REE)₅(SiO₄,PO₄)₃F [57]. Two types of britholite, fluorbritholite-(Ce) and fluorcalciobritholite, were found in the studied samples. Abundant fluorbritholite-(Ce) was found in the aegirine, sugilite, and katayamalite albitites, and small amounts of britholite in the albitized granite and quartz albitite. Fluorcalciobritholite occurs in the sugilite albite and katayamalite albitite, and in a few samples of the albitized granite. Representative analyses of the britholite are listed in Supplementary Table S3. Fluorbritholite-(Ce) contains 26.2 wt.% Ce₂O₃ and 1.9 wt.% F. Fluorcalciobritholite contains 38.1 wt.% REE₂O₃, 19.3 wt.% CaO, and 2.7 wt.% F. From the former analysis, the empirical formula based on Si + P = 3 apfu is (Ce_1.27Y_0.10La_0.56Pr_0.23Nd_0.56Sm_0.10Eu_0.02Ca_2.01Al_0.02Ba_0.01Sr_0.05Hf_0.01Th_0.02U_0.05)_5.01(Si_2.59P_0.41)_3.00O₁₂F_0.81 which is close to the ideal fluorbritholite-(Ce) formula. From the latter analysis, the em pirical formula based on Si + P = 3 apfu is (Ca_2.634Y_1.00La_0.24Ce_0.52Pr_0.08Nd_0.20Sm_0.05Eu_0.01Ti_0.01Fe³⁺_0.01Th_0.29U_0.05)_5.09(Si_2.63P_0.37)_3.00O₁₂F_1.10 which is close to the ideal fluorcalciobritholite formula.

6.3. Oxide Minerals

6.3.1. Baddeleyite

Baddeleyite (ZrO₂) is found in albitized granite and quartz albitite (Figure 8) [21]. Representative analyses of baddeleyite are listed in Supplementary Table S3. Baddeleyite is almost stoichiometric ZrO₂ with SiO₂ of ~3.51 wt.% and small amounts of Hf (1.22 wt.% HfO₂) and Nb (2.21 wt.% Nb₂O₅) substituting for Zr. The Y content is high, up to 16.88 wt.%. The actinide (U + Th) content of the baddeleyite from this study is very low (<0.52 wt.%), similar to those from contact aureole of the Stubenberg granite, Austria [58].

6.3.2. Fergusonite

Fergusonite (YNbO₄) rarely occurs in the biotite granite and katayamalite albitite. Representative analyses of the fergusonite are listed in Supplementary Table S3. Yttrium and Nb contents are 18.9–27.4 and 41.7–45.3 wt.% respectively.

6.3.3. Ilmenite

Ilmenite (FeTiO₃) occurs only in the host biotite granite and some of the albitized granites. Representative analyses of the ilmenite are listed in Supplementary Table S3. Manganese enrichment of up to 29.6 wt.% MnO in the ilmenite occurs in the biotite granite. The Nb₂O₅ contents of ilmenite vary from 0.57 to 5.12 wt.%.

7. U–Pb and ⁴⁰Ar/³⁹Ar Age

In this section, we first clarify the U–Pb age of zircon in order to know the age of the emplacement and crystallization of protolith granite, and the activity of granite that plays a role as a circulation engine for metasomatic fluid. Next, in order to clarify the period of unique Li–Na metasomatism that produced albitite on Iwagi Islet, we will reveal the ⁴⁰Ar/³⁹Ar age of katayamalite. Finally, we clarify the correlation between the timing of the two events.

7.1. Zircon U–Pb Ages

The zircon U–Pb age data are listed in Supplementary Table S4. Normal concordia plots are shown in Figure 52. The 17 grain analyses of the protolith coarse-grained granite all plot on the concordia (Figure 52a). The resultant ²³⁸U–²⁰⁶Pb age is 91.5 ± 1.9 Ma (95% confidence; MSWD = 0.14). For the medium-grained granite, 20 zircon grains were analyzed and 17 analyses yielded a tight age cluster with a mean age of 93.5 ± 1.7 Ma (MSWD = 0.05; Figure 52b). For the fine-grained granite, 20 grains were analyzed and 16 concordant ages yielded a tight age cluster of 93.3 ± 1.6 Ma (MSWD = 2.2; Figure 52c). All the U–Pb ages are interpreted to represent the crystallization ages of magmatic zircons, and there are no significant age differences between the samples. A total of 50 concordant ages, excluding 3 statistical outliers, yielded an average age of 92.8 ± 1.0 Ma (MSWD = 1.6; Figure 52d).

7.2. Katayamalite ⁴⁰Ar/³⁹Ar Age

The ⁴⁰Ar/³⁹Ar data are listed in Supplementary Table S5. Figure 53 shows the age spec tra, which yielded a plateau age of 90.91 ± 0.23 Ma from five steps (relative ³⁹Ar abundances = 3.0%, 3.5%, 4.0%, 4.6%, and 5.2%; Supplementary Table S5). These fractions comprise 71.49% of the total released ³⁹Ar. The total fusion age is 90.32 ± 0.17 Ma, which is slightly younger than the plateau age. This is due to Ar loss in the low-temperature fraction.

Figure 54 shows a normal isochron diagram that yields an age of 91.45 ± 0.26 Ma with an initial ⁴⁰Ar/³⁶Ar ratio of 196.91 ± 35.80. Figure 55 is an inverse isochron diagram that yields an age of 91.46 ± 0.26 Ma with an initial ⁴⁰Ar/³⁶Ar ratio of 195.45 ± 35.64. These isochron ages are slightly older than the plateau age, because the initial ⁴⁰Ar/³⁶Ar ratios obtained from the isochron diagrams are lower than the atmospheric Ar isotopic ratio of 295.5.

8. Discussion

In the following section, we integrate the geological, petrological, mineralogical, and geochronological results to determine the petrological evolution of the protolith granite and albitites on Iwagi Islet. The discussion is focused on the evidence for infiltration Li–Na metasomatism and the nature of the fluid.

8.1. Li–Na Metasomatism

The geology and occurrence of albitites, and Li isotope data for murakamiite on Iwagi Islet have been described in detail by Imaoka et al. [24]. The albitites occur as numerous small, discrete bodies in albitized granite, which gradually grade into a biotite granite around the summit of Mount Kuresaka in the eastern part of the islet. The largest body of albitite is 30 m in diameter, and small albitite bodies that are up to several tens of centimeters or meters wide are irregularly distributed over an area of 300 × 200 m². There are no clear intrusive margins between the albitites and surrounding albitized granite, and the boundaries are gradational with diffuse zones of quartz albitite between the two rock types (Figure 3).

The protolith biotite granite and albitized granite are visibly distinguishable, as the former is mostly pink and the latter is white in color, although the primary textures may be partially preserved. Whitening of feldspars or granitoid outcrops is one of the notable megascopic indicators for strong and pervasive albitization (e.g., [5,8,59]).

An important observation is the presence of pores, including micro-pores, in the metasomatic minerals. Vug formation occurred via dissolution of magmatic quartz, and vugs were infilled by metasomatic minerals, such as albite and sugilite (e.g., Figure 28a and Figure 43). The episyenites are characterized by variable leaching of magmatic quartz and are porous. These occurrences of albitites in Iwagi Islet indicate they are not magmatic in origin, such as Li-rich granites, but are metasomatic in origin.

The Iwagi albitites consist mainly of albite, while the Setouchi episyenites consist mainly of oligoclase [16]. Moreover, Li-rich minerals, such as sugilite, katayamalite, murakamiite, and ferro-ferri-holmquistite, are not found in the Setouchi episyenite-like rocks, apart from the Iwagi albitites. The albitites from Iwagi Islet were compared with those from the Setouchi region. The average Na₂O content in the Setouchi region is lower than that of albitites from Iwagi Islet. The average K₂O content of albitites from Iwagi Islet is lower than that of episyenite-like rocks in the Setouchi region (Figure 4a).

The average Li content of the studied quartz albitites is 9.4 ppm, that of hedenbergite albitites is 12.2 ppm, that of aegirine albitites is 25.1 ppm, that of sugilite albitites is 274 ppm, and that of katayamalite albitites is 576 ppm. Lithium contents of episyenite-like rocks in the Setouchi region are low and vary from 3.0 to 104 ppm (average = 16.4 ppm; Figure 4b). The Li contents of the sugilite and katayamalite albitites from Iwagi Islet are much higher (~960 ppm) than those of other episyenite-like rocks in the Setouchi region (Table 1 and Table 2). In summary, the Li contents of episyenite-rocks are anomalously high on Iwagi Islet.

The Li in albitite has two possible sources. One is the pegmatites that are rich in Li minerals, such as lepidolite, petalite, and spodumene. The second is Li-rich fluids. The first possibility can be excluded because there are no Li-rich pegmatites in Iwagi Islet. The second possibility can be assessed from Li isotope ratios (δ⁷Li), because δ⁷Li values may be a useful tracer for determining the source of Li.

Imaoka et al. [24] found that murakamiite–pectolite in Iwagi Islet shows low δ⁷Li values (−9.1 to +0.4‰), which requires involvement of fluids with very low δ⁷Li values (−6.3 to +3.2‰), assuming equilibration at plausible temperatures of crystallization between 300 and 600 °C. The occurrence of such low δ⁷Li values in crustal fluid is rare, but some deep crustal fluids are reported to show comparable low δ⁷Li values, such as earthquake swarm-related deep fluids beneath the Ontake volcano, central Japan (−5‰ to +2‰ [60]), and Arima-type brines upwelling along the major faults in SW Japan (+1.4 to +3.1‰ [61]). Interestingly, deep crustal fluids from Ontake exhibit significant Li enrichment, with high Li/Na mass ratios of up to 2.0 × 10⁻³ (cf. 1.5 × 10⁻⁵ in seawater). Arima-type deep-seated brines exhibit further Li enrichment (Li/Na = 2.6 × 10⁻³ in Arima and 7.3 × 10⁻³ in Kashio; [62,63]). These Li/Na ratios are consistent with the values for murakamiite-bearing albitite (6.4 × 10⁻³). Thus, in terms of Li isotope and Li/Na ratios, parental fluids of the Iwagi murakamiiite–pectolite can be originated from the above-mentioned types of deep-seated fluids. The observed Li, O, and H isotope and elemental characteristics suggest that Arima-type fluids are derived from dehydration of the subducted Philippine Sea Plate slab [61,62]. As such, Li–Na-enriched, low-δ⁷Li Arima-like subduction zone fluids might have migrated upward along a shear zone after granite emplacement and caused the Li–Na metasomatism that formed the zoned albitites containing murakamiite−pectolite.

One problem exists on the origin of the Arima-type fluid from the Philippine Sea Plate because it did not exist in Turonian age (~91.5 Ma) beneath the SW Japan arc. Onset of subduction of the Philippine Sea Plate occurred at about 15 Ma [64], and subduction of the young Shikoku Basin of the Philippine Sea Plate resulted in shallow dehydration, forming Arima-type deep-seated slab fluids [65]. Instead, during the Turonian, the Izanagi Plate was subducting beneath SW Japan [66], which caused a widespread magma flare-up in the Inner Zone of SW Japan, including Iwagi Islet. The Cretaceous peak magmatic stage appeared from 95 to 90 Ma [67]. This coincides with the age of magmatism and metasomatism on Iwagi Islet. Therefore, a similar high-temperature slab could have existed in the Turonian age related to the magmatic flare-up along the active continental margin and the Arima-like brines.

8.2. Structural Controls on Li–Na Metasomatism: Implication for Fluid Pathways

The Li–Na metasomatism is observed in a restricted area of Iwagi Islet. The charac teristic textures in the Iwagi albitites are mosaic aggregates of aegirine (Figure 14b) and cross-hatched twinning of albite (Figure 35 and Figure 45) [22,24]. The occurrence of such textures is scarce in normal igneous rocks, but not uncommon in metasomatic and recrystallized rocks. The cross-hatched twinning of albite (Figure 35 and Figure 45) is considered to form by the complete replacement of K-feldspar by albite [40,68,69]. The occurrence of microcline relics in some albite crystals confirms this interpretation.

These textures, as well as conspicuous strain-induced textures, such as bent lamellae in albite (Figure 41), curved cleavage, and healed microstructures, are regarded to be the products of dynamic recrystallization (i.e., deformation-related recrystallization processes [70]) caused by deformation of a solid intrusion and/or regional tectonic stress. Deformation results in the nucleation of smaller crystals of albite at grain boundaries. Increasing temperature results in a lower rate of nucleation of albite grains and increases the rate at which the grains are consumed by grain boundary migration [71].

Examples of structurally controlled regional metasomatism reported so far are asso ciated with wide metamorphic temperature–pressure conditions and tectonic settings. Permeability is key to metasomatism and may be intrinsic or the result of deformation processes, such as (micro-) fracturing, faulting, foliation and shear development, and hydro-fracturing [72]. Partial recrystallization fabrics with intracrystalline patterns of dynamic recrystallization are present in the albite and aegirine grains with strong undulose extinction and sub-grain development. Fracturing and bending of mineral grains (Figure 41), curved cleavage, coarse twinning, bent twin lamellae, and recrystallized aggregates of aegirine (Figure 14b) and albite suggest the rocks formed in the presence of stress and heat, which was possibly associated with faulting. This suggests that the studied rocks are alkaline metasomatites related to fault zones. Fractures and shear zones are important loci for fluid flow and metasomatism [73,74].

Veins are common features in rocks and useful structures for determining paleo-stress, strain, pressure, temperature, and fluid composition and origin, as well as fluid pathways. According to Bon et al. [75], there are four vein–fluid transport mechanisms: (1) diffusion of dissolved material through a stagnant pore fluid, (2) flow of fluid with dissolved material through pores, (3) flow of fluid with dissolved material through fractures; and (4) movement of fractures and their contained fluid and dissolved material (i.e., mobile hydro-fractures).

Deformation-induced fracturing of rocks (Figure 46) and the resulting veins and vein lets (Figure 23, Figure 27 and Figure 48b) along crystal boundaries (Figure 23a) or existing polygonal fractures could have facilitated fluid circulation in the present case. The occurrence of metasomatic rocks (episyenite-like rocks) in the Setouchi region along the E–W direction (Figure 1) suggests this was the locus for large-scale fluid movement in the crust. These physical conditions would have resulted in mass transfer and the associated formation of metasomatic mineral assemblages (Table 1).

8.3. Mineral and Elemental Behavior

All albitized rocks present in this study exhibit various degrees of episyenitization. The lithology, mineralogy, and textures have been strongly influenced by the evolution of episyenitization. The mineral assemblages and their modes differ significantly according to rock type (Figure 8). In this section, we systematically examine the mineralogical variations from the host biotite granite to various albitite types.

Approximately 50 mineral species were identified by microscopic and EPMA obser vations of the albitites and their presumed protolith granite (Figure 8). These include major, accessory, primary, secondary, stable, and metastable minerals. Replacement textures are common, indicating disequilibrium and/or local equilibrium during crystallization, and the observed textural complexity indicates overprinting due to infiltration metasomatism. In addition to quartz dissolution and alkali metasomatism, the episyenites may record a complex history of alteration and hydrothermal mineral growth [76].

8.3.1. Albitization

Albitization is a common process, during which hydrothermal fluids convert plagio clase and/or K-feldspar into nearly pure albite, and it is caused by the infiltration of an ascending hydrothermal fluid that has high Na/K and Na/Ca ratios [77]. Albitization was a ubiquitous process in Iwagi Islet (Figure 15), where alkali feldspar was replaced by albite. Furthermore, as is evident from Figure 7 and Figure 8, and Supplementary Table S3, the oligoclase in the original granite and albitized granites has been converted to albite.

Engvik et al. [13] proposed the following reaction for the transition from oligoclase (An₁₂) to albite (An₀):

Na_0.88Ca_0.12[Al_1.12Si_2.88O₈] + 0.24(Na⁺)_aq + 0.48H₄SiO₄ = 1.12 Na[AlSi₃O₈] + 0.12(Ca²⁺)_aq + 0.96H₂O

(1)

Silica and Na⁺ are required for the reaction 1, and Ca is released into the solution. The SiO₂-rich fluid after quartz dissolution facilitates the precipitation of albite, and Ca²⁺ liberated by the reaction might be fixed by calcite and fluorite. In contrast, the replacement of K-feldspar by albite observed in albitized granite is attributable to alkali ion exchange between fluid and feldspar, as follows (reaction 2 of Moore and Liou [69]):

KAlSi₃O₈ + Na⁺ = NaAlSi₃O₈ + K⁺

(2)

The K-feldspar (Or_93.1–96.0) in the albitized granites has a composition similar to that in the original granite (Or_93.1–97.0), suggesting that it represents the residual igneous K-feldspar and subsequently converted to albite with increasing the albitization.

Fluid–crust interactions during metasomatic and/or hydrothermal processes control the rheology, porosity, and elemental redistribution within Earth’s crust [78]. Halogens and most prominently Cl, are the most abundant ligands in crustal fluids [79]. It is widely accepted that Cl-bearing solutions can flow easily along grain boundaries and through porous metasomatic minerals, e.g., Duan et al. [80]. This is facilitated by their reactivity, inducing interface-coupled dissolution–reprecipitation (ICDR) reactions, which is a process that links fluid-induced mineral replacement reactions, porosity development, and fluid transport.

Duan et al. [80] performed experiments designed to quantify the effects of chloride and fluoride on the fluid-mediated albitization of perthite (intergrowth of albite and microcline) at 600 °C and 2 kbar. They found that the albitization rate depends on the Cl/(Cl + F) ratio, with the amounts of albite formed reaching a maximum when the Cl/(Cl + F) ratio in solution is between 0.5 and 0.99 (i.e., Cl/F = 1–100). Subducting slab-derived Arima-type brines, which might be involved in metasomatism in Iwagi Islet (see 8.1.), show high Na/K, Na/Ca, and Cl/(Cl + F) mol. ratios (9.7, 12, and >0.99, respectively [62]). Thus, interactions with such deep fluids favor extensive albitization, such as that observed in Iwagi Islet.

8.3.2. Zirconium Minerals: Zircon Replacement Reactions

The principal Zr-bearing minerals chemically identified in the Iwagi albitites include zircon (ZrSiO₄), dalyite (K₂ZrSi₆O₁₅), zektzerite (LiNaZrSi₆O₁₅), baddeleyite (ZrO₂), and gittinsite (CaZrSi₂O₇). Amongst these, zircon is the most ubiquitous in all the samples. The most important observations regarding the petrography and chemistry of the Zr-bearing minerals in the albitites are as follows: (1) Porous zircon is commonly observed in the albitites, but not in some of the host biotite granite where the zircons have well-developed crystal shapes and exhibit oscillatory zoning indicative of a magmatic origin. On the other hand, some of the host granites have porous zircons and contain abundant Th- and U-rich thorite/huttonite inclusions (Figure 11a). (2) Mantles of zektzerite on porous zircon occur in the aegirine albitites. (3) The dalyite occurs in small pores in the aegirine albitites, and its composition is close to the ideal formula, with a low Na content (<0.015 to 0.03 apfu). Observation (1) suggests the early crystallized zircon in the ‘fresh’ host granite became unstable and porous due to its instability in alkaline–peralkaline solutions and fluids. Th- and U-rich thorite/huttonite inclusions have been reported in zircons modified by fluid-assisted coupled dissolution–reprecipitation [81,82,83]. Observation (2) implies the zircons were readily dissolved during albitization and that the chemical compositions of the Li–Na-rich fluids were suitable to induce resorption of early zircon. From observation (2), we conclude that zektzerite is a metasomatic rather than magmatic mineral. Imaoka et al. [21] proposed the following metasomatic reaction for the formation of zektzerite in an open system:

ZrSiO₄ + 5SiO₂ + Li⁺ + Na⁺ + 2OH⁻ = LiNaZrSi₆O₁₅ + H₂O
 zircon              zektzerite      

(3)

where the silica, Na⁺, and Li⁺ in solution are required for the reaction.

The observed dissolution of magmatic quartz (Figure 36a) favors the above reaction, as SiO₂ (quartz) is supplied to the solution. The occurrence of anhedral dalyite in a small pore (observation 3) indicates it formed as a result of metasomatism by the interstitial solution trapped in the pore. The formation of dalyite in the albitites can be explained by the following reaction:

ZrSiO₄ + 5SiO₂ + 2K⁺ + 2OH⁻ = K₂ZrSi₆O₁₅ + H₂O
    zircon           dalyite        

(4)

where the silica and K⁺ are required for the reaction.

Excess silica can be supplied by dissolution of magmatic quartz (Figure 36a). Because alkali feldspar (19.3–35.6 vol.%) and quartz (35.9–45.1 vol.%) originally contained in the host granite dramatically decrease to 0.2 and 0.1 vol.%, respectively, in the albitites (Supplementary Table S2), a part of K⁺ and SiO₂ lost through these reactions and Li⁺ and Na⁺ supplied by hydrothermal fluids could facilitate the growth of zektzerite and dalyite.

Imaoka et al. [24] inferred that Li isotopic fractionations associated with murakamiite and Li-rich pectolite crystallization on Iwagi Islet could be attributed to metasomatic hydrothermal fluid–rock interactions at 300–600 °C. Dalyite has been synthesized hydrothermally at 340 °C and 600 bar [84], which would suggest the metasomatic formation of dalyite occurred at a relatively low temperature during the paragenetic sequence of events that formed the albitite. Under such low-temperature conditions, dalyite of potassic group 2 of Jeffery et al. [85] is stable, and very limited substitution of Na for K can occur. Thus, dalyite in metasomatic rocks, such as the Iwagi albitites, shows more limited miscibility than that of the same mineral in igneous rocks.

With regards to dalyite: (1) in Iwagi Islet, it occurs only in the metasomatic albitites, and not in the host granite, (2) dalyite on the Azores is inferred to be the last mineral to have crystallized [86], as is the case for our samples, (3) Furnes et al. [87] suggested that dalyite in the ultrapotassic Sunnfjord dike was either metasomatic in origin or, if magmatic in origin, had been altered by metasomatism, and (4) Na contents are often below the detection limits [85]. Taking these observations and our data into account, we conclude that the dalyite in the Iwagi albitites is metasomatic in origin.

In summary, the Iwagi albitites record Zr-bearing mineral replacement reactions that occurred via dissolution–reprecipitation, which changed the original igneous mineralogy of the ‘fresh’ host granite during infiltration metasomatism.

8.3.3. Thorium-Bearing Minerals

Thorium-bearing minerals chemically identified in Iwagi Islet include thorite/hut tonite (ThSiO₄), turkestanite (K,◻)(Ca,Na)₂ThSi₈O₂₀·nH₂O), and arapovite ((K_1−x◻_x)(Ca,Na)₂U⁴⁺Si₈O₂₀ [x ≈ 0.5]). Amongst these minerals, thorite/huttonite occurs only in the host biotite granite. The thorite/huttonite is abundant in the host biotite granite and is thus the most important Th carrier. Thorite/huttonite is often included in euhedral zircon as tiny crystals (<1 µm; Figure 11a) and forms the core (several tens of microns) mantled by zircon (Figure 11b). This indicates that thorite/huttonite and zircon are primary igneous minerals, and thorite/huttonite (ThSiO₄) is isomorphous with zircon (ZrSiO₄), allowing Zr ↔ Th substitution.

In some albitized granites, turkestanite and arapovite occur, but thorite/huttonite occurs in the albitized granite and quartz albitites. In the hedenbergite albitites, turkestanite again appears. In the advanced stages of albitization, from the aegirine albitite to sugilite albitite to katayamalite albitite, unknown Si–Th–Ca minerals are prevalent. Thorite/huttonite is also rarely present in rocks that undergo such an advanced stage of albitizations. Based on the paragenetic succession, some Th in the host granite was incorporated by arapovite and turkestanite (Supplementary Table S3) in the low-grade albitization, and then unknown Si–Th–Ca minerals became the main hosts for Th during high-grade albitization.

8.3.4. Calcium Minerals

Accessory minerals containing Ca as the main element include fluorapatite (Ca₅(PO₄]₃F), fluorite (CaF₂), aegirine (NaFe³⁺Si₂O₆) –aegirine-augite ([Ca,Na][Fe³⁺,Mg,Fe²⁺][Si₂O₆]), hedenbergite (CaFe²⁺Si₂O₆), calcite (CaCO₃), wollastonite (CaSiO₃), andradite (Ca₃Fe³⁺₂[SiO₄]₃), kristiansenite (Ca₄Sc₂Sn₂Si₂O₇]₂[Si₂O₆OH]₂), turkestanite (K,◻)(Ca,Na)₂ThSi₈O₂₀·nH₂O), miserite (K_1.5−x[Ca,Y,REE]₅[Si₆O₁₅][Si₂O₇] (OH,F]₂yH₂O), gittinsite (CaZrSi₂O₇), and truscottite (Ca₁₄Si₂₄O₅₈(OH)₈·2H₂O). Fluorapatite is found in all rock types (Figure 7), and it tends to increase in abundance with albitization. Fluorite occurs in the host biotite granite, albitized granite, and hedenbergite albitite; therefore, it disappears during relatively low-grade albitization.

Aegirine–aegirine-augite occurs in some albitized granite and all albitites, except for protolith granite (Figure 8); thus, it characterized albitization in the study area. For clinopyroxene, the aegirine component is higher in the aegirine, sugilite, and katayamalite albitites (Supplementary Table S3). The Na was supplied by the fluid to form aegirine. Furthermore, aegirine-augite is surrounded by aegirine (Figure 37). These results suggest that as the Na content of the whole-rock increased with albitization, Na-rich aegirine crystallized which was different from the aegirine-augite formed during the relatively low-grade albitization (Figure 51).

Hedenbergite characterizes hedenbergite albitite, and is also found in albitized granite. The latter has a slightly larger compositional range, but no large differences in chemical composition (Figure 51).

Although wollastonite is rare in the low-grade albitites, it occurs in large quantities in the katayamalite albitites in close association with pectolite–murakamiite and aegirine-augite. Wollastonite has very little variation in composition within a sample or between different samples (Supplementary Table S3). Calcite occurs between grains of wollastonite, pectolite–murakamiite, and aegirine-augite. Calcite occurs mainly in the aegirine albitite and katayamalite albitite and may occur in veins within pectolite–murakamiite (Figure 47 and Figure 49c) or as inclusions in pectolite–murakamiite (Figure 49b).

Wollastonite is replaced by pectolite and quartz in the hedenbergite albitites (Figure 31). Infiltration and Na metasomatism are further confirmed by this occurrence. The formation of pectolite at the expense of wollastonite is indicative of brine infiltration. The growth of pectolite on wollastonite is suggestive of a pectolite-forming reaction similar to that proposed by Heinrich [88] for meta-cherts in the contact aureole of the Bafa del Diente intrusion (Mexico):

3CaSiO₃ + HCl + NaCl_aq = NaCa₂Si₃O₈(OH) + CaCl_2aq.
wollastonite          pectolite       

(5)

The hedenbergite albitites have low Li contents (7.4–17.0 ppm; Table 1), and if the murakamiite component is present, the following reaction is relevant:

3CaSiO₃ + HCl + (Na,Li)Cl_aq + Li⁺ = (Na, Li)Ca₂Si₃O₈(OH) + CaCl_2aq.
Wollastonite            pectolite–murakamiite     

(6)

The widespread growth of pectolite on wollastonite has been documented in calc-silicate lenses in migmatites from northeast Sardinia, Italy [89]. This reaction is favored by the infiltration of Na-rich metasomatic fluids [88].

The accessory mineral contents (Figure 8) show that the katayamalite albitites con tain abundant calcite. The hydrothermal fluid that caused albitization was most CO₂-rich during the formation of the katayamalite albitites. Gittinsite (CaZrSi₂O₇) occurs as interspatial space of recrystallized albitite (Figure 24c); thus, it crystallizes from low-temperature hydrothermal fluids. Truscottite also occurs as an interstitial mineral of Ca-rich minerals such as aegirine (Figure 39) and katayamalite, so, it crystallizes at relatively lower temperatures.

We identified tiny turkestanite, with only a few specimens of albitized granite, hedenbergite albitite, aegirine albitite, sugilite albitite and katayamalite albitite (Figure 42c, Supplementary Table S3). Vilalva and Vlach [90] also identified turkestanite in association with other Ca-bearing mineral phases such as aegirine-augite, fluorite and britholite-group mineral in the evolved peralkaline granites from the Morro Redondo Complex, south Brazil, and suggested it precipitated during post-magmatic stages in the presence of residual HFSE-rich fluids carrying Ca.

8.3.5. Lithium-Bearing Minerals

The Li-bearing minerals include sugilite (Na₂K[Fe³⁺,Mn³⁺,Al]₂Li₃Si₁₂O₃₀), katayama lite (KLi₃Ca₇Ti₂[SiO₃]₁₂[OH]₂), pectolite–murakamiite (NaCa₂Si₃O₈[OH]–LiCa₂Si₃O₈[OH]), ferro-ferri-holmquistite (□Li₂(Fe²⁺₃Fe³⁺₂)Si₈O₂₂(OH)₂), zektzerite (NaZrLiSi₆O₁₅), polylithonite (KLi₂AlSi₄O₁₀F₂), and tainiolite (KLiMg₂Si₄O₁₀F₂). The host biotite granite contains no Li minerals. In the albitized granite, the biotite partly dissolved and suggesting Li was incorporated by ferro-ferri-holmquistite, and polylithonite. The ferro-ferri-holmquistite soon disappeared as albitization progressed. Taeniolite occurs only in the aegirine albitites (Figure 8). Zektzerite appears in the quartz albitites and increases in content as albitization progresses.

The sugilite and katayamalite albitites that formed during the high-grade albitiza tion have higher Li contents than the other albitites (Table 1, Figure 4). Based on the modal abundances (Supplementary Table S2, Figure 8) and Li contents of each mineral described in the previous section, it is conceivable that in addition to the zektzerite, Li occurs mainly in sugilite, katayamalite, and pectolite–murakamiite.

Second occurrences of sugilites in the Wessels mine, South Africa are found at varia ble distances away from a bostonite (i.e., alkali micro-syenite) dike, with the closest being ~100 m from the dike. Based on this occurrence, the sugilites are thought to have formed from infiltration metasomatism by highly alkaline hydrothermal fluids at 400–600 °C and <1 kbar [46].

8.3.6. Phosphorus-Bearing Minerals

The granites and albitites in Iwagi Islet contain monazite ([REE,Th]PO₄), apatite (Ca₅[PO₄]₃[OH,F,Cl]), xenotime-(Y) (YPO₄), and britholite-group minerals. Monazite is a common accessory mineral in most igneous and metamorphic rocks [91]. Mineral–fluid interactions lead to the replacement of monazite by fluorapatite in igneous and metamorphic rocks [92,93]. The decomposition products depend on the fluid composition [94]. The most common products are REE-rich apatite and thorite/huttonite (ThSiO₄ phases) [95].

In the granites of Iwagi Islet, monazite is stable and abundant (Figure 11c). Monazite is replaced by fluorapatite in the quartz albitite (Figure 24b). In the albitized granite, fluorapatite is replaced by fluorbritholite-(Ce) (Figure 19b), and minute fluorbritholite-(Ce) grains occur at the crystal rims of fluorapatite in the katayamalite albitites.

From these occurrences, it is clear that monazite was replaced by fluorapatite, and then fluorapatite was replaced by fluorbritholite-(Ce) during metasomatism. According to Zirner et al. [96], monazite undergoes the coupled substitution of Na⁺ + REE³⁺ = 2Ca²⁺ under the influence of Na-containing fluids, which crystallizes fluorapatite. This suggests that Na-rich hydrothermal solution led to albitization, and the REEs in monazite were replaced by Ca, which produces fluorapatite.

Both britholite-group minerals and apatite have the same hexagonal crystal structure, and it is known that the Ca²⁺ + P⁵⁺ = REE³⁺ + Si⁴⁺ substitution occurs [96]. Monazite, fluorapatite, and britholite-group minerals underwent this substitution due to the effect of the fluid that caused albitite formation, which crystallized in the order monazite → fluorapatite → britholite group minerals.

Monazite is abundant or common (Figure 8) in biotite granite, albitized granite, quartz albitite, and hedenbergite albitites. Monazite-(Ce) and monazite-(Nd) contains up to 68 wt.% and 58 wt.% REE₂O₃, respectively (Supplementary Table S3). On the other hand, monazite is absent or rare in the sugilite and katayamalite albitites, and britholite- group minerals are abundant or common. Thus, britholite-group minerals such as fluorbritholite-(Ce) (REE₂O₃ = 58 wt.%; Supplementary Table S3) and fluorcalciobritholite (REE₂O₃ = 38 wt.%; Supplementary Table S3), are the second major repository of REEs instead of monazite.

8.3.7. Titanium-Bearing Minerals

The Ti-bearing minerals chemically identified in Iwagi Islet include ilmenite (FeTiO₃), titanite (CaTiSiO₅), TiO₂ mineral, and katayamalite (KLi₃Ca₇Ti₂[SiO₃]₁₂[OH]₂). Titanite, ilmenite, and TiO₂ mineral occur mainly in biotite or along fractures in the granite, and thus when biotite disappears due to albitization, so do ilmenite, TiO₂ mineral, and titanite. Titanite is also common in the low-grade albitized rocks such as the albitized granites, quartz albitites, hedenbergite albitites, and in some cases, present in katayamalite albitites. As such, in the host granite, Ti occurs mainly in ilmenite and TiO₂ mineral and as albitization progressed the titanite was formed from Ti in the hydrothermal solutions that caused albitization. It is likely that katayamalite, which contains ~10 wt.% TiO₂, was also produced.

Titanite is a typical accessory mineral in alkaline silicate rocks and, in some cases, is also an important carrier of HFSEs and REEs. Titanite with elevated Zr contents of up to 6.80 wt.% (Supplementary Table S3) was found in a quartz albitite vug (Figure 25a), which is generally rare in nature [97] and mostly restricted to late derivatives of alkali-rich silica-undersaturated rocks, such as nepheline and soda syenites or the products of their metasomatic alteration in the Dara-i-Pioz Complex, northern Tajikistan [53]. Titanite containing up to 17.7 wt.% ZrO₂ has been reported from hydrothermally altered alkaline igneous rocks (teschenite) of Early Cretaceous age in the Outer Western Carpathians, Czech Republic [98]. This titanite contains up to 19.6 wt.% Nb₂O₅ and 1.1 wt.% REE₂O₃, along with 1.82 wt.% Na₂O (≤0.08 apfu). The mineral assemblage of the leucocratic dikes and patches hosted by the teschenites was formed by magmatic fractional crystallization, high-temperature hydrothermal autometamorphic overprinting, and low-temperature hydrothermal alteration. Kropáč et al. [98] concluded that precipitation of Zr–Nb-rich titanite by the hydrothermal auto-metamorphic overprint occurred at 390–510 °C.

The studied titanite exhibits intergrowths with zircon, and both crystals are elongate or platy (Figure 25). Urueña et al. [99] documented metamorphic titanite + zircon pseudomorphic intergrowths formed by the metamorphic breakdown of igneous zirconolite in syenodiorite and syenite in Sveconorwegian Province, Sweden. We cannot explain this by the breakdown of zirconolite, because we did not observe zirconolite, and rutile is also a product. Further research is needed on the titanite and zircon intergrowths.

Elongate microcrysts of titanite with high contents of Zr and Hf (11.4–15.3 wt.% ZrO₂; up to 0.5 wt.% HfO₂) were documented in silicocarbonatites from the Afrikanda alkaline ultramafic complex, Kola Peninsula, Russia [100]. Fluid-induced metasomatism controlled the local major and trace element budget. The titanite enriched in Nb, Zr, and Hf occurs near the primary carriers of high-field-strength elements (HFSEs).

The quartz albitite vugs (Figure 24) and albitized granite vugs contain Ti-rich veins (Figure 21) and residual titanite (Figure 22). The hedenbergite albitite vugs are filled by quartz and titanite (Figure 33), and an aegirine albitite vug contains a fine mesh-like Ti material (Figure 40). As such, these vugs are filled with HFSE-rich minerals, indicating the vugs played an important role in the channeling, migration, and budget of the HFSEs.

8.3.8. Sulfide and Other Non-Silicate Minerals

Some of the minerals described disappeared in the low-grade albitization. The W-bearing minerals such as scheelite (CaWO₄) and wolframite ([Fe,Mn]WO₄), Nb oxide mineral fergusonite (YNbO₄), and barite (BaSO₄) occur only in the host biotite granite (Figure 8), and thus disappeared during the low-grade albitization. Sulfide minerals such as sphalerite ([Zn,Fe]S), galena (PbS), arsenopyrite (FeAsS), pyrite (FeS₂), chalcopyrite (CuFeS₂), and tetrahedrite ([Cu,Fe]₁₂Sb₄S₁₃) occur mainly in the host biotite granite, albitized granite, quartz albitite, and rarely in the aegirine and sugilite albitites. Importantly, sulfide minerals are not observed in the hedenbergite and katayamalite albitites (Figure 8). This indicates that the sulfide minerals disappeared in the low-grade albitization, and the liberated elements were transported outside the system. Sulfide minerals dissolve readily in an acidic Cl-bearing solution at ambient temperatures [101,102,103].

8.4. Vug Diversity and Formation

Rounded grains of alkali feldspars and Ti-rich veins connecting the vugs (Figure 21), and residual titanite and Ti-rich veins, as well as Si and Al materials (Figure 22), occur in the albitized granite (Figure 21). Vugs consisting of alkali feldspar are restricted to the albitized granites. Vugs in quartz albitites contain intergrowths of elongated zircon and Zr-rich titanite, elongated zircon and quartz (Figure 25), and fluorapatite, monazite, and K-feldspar (Figure 26). In the hedenbergite albitites, euhedral albite and fluffy crystal project into the vugs (Figure 28), and quartz and titanite (Figure 33) occur in the vugs. In the aegirine albitites, the original magmatic quartz has been dissolved, and opal with string-like or rounded, irregular shapes and opalized quartz (SiO₂∙nH₂O) have formed in the vugs (Figure 36). A Ti-rich fine-meshed material was also observed in the vugs (Figure 40). In the sugilite albitites, botryoidal opal (SiO₂∙nH₂O) was observed in a vug (Figure 42d), and sugilite, aegirine, and opalized quartz infilled the vug (Figure 43). Secondary quartz is usually present in the cavities.

Dissolution of quartz is one of the most important alteration processes because it pro vides enhanced porosity for the circulation of hydrothermal fluids (i.e., an increase in the fluid/rock ratio). Although numerous studies have investigated the origins of episyenites worldwide, few have addressed the basic mechanism of quartz dissolution. The prevailing view [4,76] is that dissolution occurs as a result of isobaric cooling in the retrograde solubility field of quartz. The solubility maximum in the retrograde solubility field of quartz for pure water (i.e., where the solubility of quartz increases with decreasing temperature) extends from 340 °C at vapor pressure to 520 °C at almost 900 bars. Addition of NaCl increases the solubility of quartz and shifts the solubility maximum of the retrograde solubility region towards higher temperatures [11]. In summary, it is apparent that retrograde dissolution of quartz during isobaric cooling is only possible at fluid pressures of <900 bar by aqueous fluids of low to moderate salinity [11].

8.5. Origin of Hedenbergite Albitites and Possible Involvement of Skarnization

The hedenbergite albitite is unique in terms of its constituent minerals. In particular, it contains hedenbergite, wollastonite, andradite, titanite, monazite-(La), kristiansenite (a Ca–Sc–Sn mineral), and magnetite. It lacks sulfide minerals. In the hedenbergite albitites, titanite and andradite are closely associated (Figure 30). Compared with titanites in the albitized granites, those in the hedenbergite albitite are rich in SnO₂ (~2.69 wt.%; Supplementary Table S3).

Considering these unique characteristics, this rock type may have had a protolith different from other types of albitites. Such characteristic mineral assemblage of the hedenbergite albitites (hedenbergite, wollastonite, and andradite) looks like some kind of skarn. Skarns are rocks that form by replacement of carbonate-bearing rocks during regional or contact metamorphism and metasomatism. Skarns may form by metamorphic recrystallization of impure carbonate protoliths, and by infiltration metasomatism by magmatic-hydrothermal fluids [104]. Skarns can be subdivided by protoliths, e.g., if the protolith is of sedimentary origin, it can be referred to as an exoskarn and if the protolith is igneous, it can be called an endoskarn [105]. Classification can be made based on the protolith by observing the skarn’s dominant composition and the resulting alteration assemblage. Calcic skarns are characterized by mineral assemblages containing garnet, clinopyroxene, and wollastonite [105]. Thus, it is possible to assume that the protolith of the hedenbergite albitite in Iwagi Islet may have been metasomatic granite subjected to calcic skarnization. Furthermore, the replacement of wollastonite by pectolite (Figure 31) appears to be consistent with the pre-existence of wollastonite before Li–Na metasomatism. The prominent crystallization of wollastonite in the katayamalite albitite may also be understood in the same framework.

The hedenbergite albitite occurs as a zoned metasomatitic rock of granitic origin at the southeastern end of the albitite outcrop. This implies that the protolith of the hedenbergite albitite is an endogenous skarn that formed by reaction of the biotite granite with the carbonate-rich fluid (or melt) infiltrated through a shear zone. Such a fluid (or melt) might be produced by intrusion of granitic magma to limestone [106]. There are no known limestone outcrops in Iwagi Islet, but limestone bodies are distributed on the islands around the islet. For example, there is a large limestone/marble body on Yuge Island to the east of Iwagi Islet, and skarn minerals such as hedenbergite, wollastonite, and andradite occur between the limestone and granite. The replacement of wollastonite by pectolite indicates that the zoned skarnization preceded the Li–Na metasomatism and represents an earlier stage of metasomatic events that occurred in the biotite granite of Iwagi Islet.

8.6. Implications of the Zircon U–Pb and Katayamalite ⁴⁰Ar/³⁹Ar Ages

Zircon U–Pb dating and katayamalite ⁴⁰Ar/³⁹Ar dating were undertaken to determine the timing of emplacement and crystallization of the host granite, and the age of the Li–Na metasomatic event that produced the albitites. Because of the high closure temperature of the U–Pb system in zircon, zircon U–Pb dating is considered to date the magmatic ages of granitoids. The zircon U–Pb age of the protolith coarse-grained granite is 91.5 ± 1.9 Ma (Figure 52a). In addition, the medium- and fine-grained granites have zircon U–Pb ages of 93.5 ± 1.7 Ma (Figure 52b) and 93.3 ± 1.6 Ma (Figure 52c), respectively. All three U–Pb ages are interpreted to represent the crystallization ages of magmatic zircons.

Katayamalite is a Li–Na metasomatic mineral in Iwagi Islet. The katayamalite is a ring silicate and contains substantial K₂O of ~3.5 wt.%. As Jackson et al. [107] discovered that the rare gas, such as Ar is trapped in an unoccupied ring site of cordierite, the katayamalite is expected to be structurally favorable to the retention of radiogenic argon. Thus, the katayamalite is suitable for ⁴⁰Ar/³⁹Ar dating of metasomatic events. As far as we are aware, this is the first example of katayamalite dating.

The Iwagi katayamalite yielded a well-defined ⁴⁰Ar/³⁹Ar plateau age of 90.91 ± 0.23 Ma (Figure 53). Normal and inverse isochron ages and initial ⁴⁰Ar/³⁶Ar ratios are 91.45 ± 0.26 Ma and 196.91 ± 35.80, and 91.46 ± 0.26 Ma and 195.45 ± 35.64, respectively (Figure 54 and Figure 55). These isochron ages are slightly older than the plateau age. This is due to the initial ⁴⁰Ar/³⁶Ar ratios obtained from the isochron diagrams being lower than the atmospheric Ar isotopic ratio of 295.5. This lower ratio could be due to katayamalite crystallizing in a ³⁶Ar-enriched environment. Therefore, the isochron ages are considered to represent the time of mineral formation. Therefore, the isochron age of 91.5 Ma (Turonian) reflects the timing of albitization by the Li–Na metasomatic fluids.

The coarse-grained granite, the protolith of albitites, shows a zircon U–Pb age that is indistinguishable from the ⁴⁰Ar/³⁹Ar age of katayamalite. This indicates that the emplacement of granite and Li–Na metasomatism occurred almost simultaneously. On the other hand, the age of the medium-grained granite appears to be slightly older than that of the katayamalite, exceeding the analytical error. The mineral compositions of the three types of granite (coarse-, medium-, and fine-grained granites) distributed on Iwagi Islet are exactly the same. These granites are vertically stacked sheet-like bodies [24], and their emplacement can be explained by a sheet-on-sheet model [25]. The current elevation of the boundary between the medium-grained granite and the coarse-grained granite is about 100 m, and the elevation of Mt. Kuresaka, where the albitite is distributed, is 74.7 m. Considering the time of formation (Turonian) of the albitite distributed around Mt. Kuresaka and the subsequent uplift and erosion, it is highly likely that the medium-grained granite and albitite were in contact with each other. Given the average age values, the katayamalite can be two million years younger than the medium-grained granite. In this case, the coarse-grained granite may have solidified in relatively later stage within the granite body of Iwagi Islet, and the Li–Na metasomatism occurred soon after the solidification.

Alternatively, considering the presence of hydrated zircon phase (ZrSiO₄·nH₂O) in sample T-69, it is possible that the apparently younger age of the coarse-grained granite compared with the fine-grained and medium-grained granites may represent the timing of such zircon alteration associated with metasomatism. Regarding hydrous zircons, Tani et al. [108] reported the age of magmatic zircons (11.27 ± 0.13 Ma) in gabbro of the oceanic core complex in the Philippine Sea and the age of those modified by fluid-assisted coupled dissolution–reprecipitation (10.79 ± 0.29 Ma). The latter is ~0.3–0.5 m.y. younger than the former, and they suggested that it may represent the timing of fluid infiltration after the solidification of the gabbro.

It is noted that the initial ⁴⁰Ar/³⁶Ar ratios of the Iwagi katayamalite had extremely low values of 195–197. It is well known that the Archaean atmospheric argon had a lower ⁴⁰Ar/³⁶Ar ratio than the present value, and the ratio changed to the present value due to the degassing from mantle having an extremely high ⁴⁰Ar/³⁶Ar ratio and from the crust having the radiogenic argon ⁴⁰Ar (cf. Pujol et al. [109]). Kaneoka [110] pointed out that the mass-dependent isotopic fractionation of noble gases was common in volcanic materials. Matsumoto et al. [111] analyzed the ⁴⁰Ar/³⁶Ar and ³⁸Ar/³⁶Ar ratios of historical lavas from Japanese Islands and revealed that the lavas recorded the mass fractionation to light isotope enrichment. Mass fractionation has been expressed with nonlinear numerical analysis by Aston [112], Kaneoka [113], and Ryu et al. [114]. This mass fractionation occurs when light mass isotopes move faster than heavier ones during the transport of gas. The residual gas is enriched in heavier isotopes in comparison with the isotope ratio of the original gas. This means that the diffusive gas is enriched in the light isotopes. Ryu et al. [114] calculated the ³⁸Ar/³⁶Ar and ⁴⁰Ar/³⁶Ar ratios in the diffusive gas during the transport of original gas with atmospheric argon isotopes of ³⁸Ar/³⁶Ar = 0.187 and ⁴⁰Ar/³⁶Ar = 295.5. The ⁴⁰Ar/³⁶Ar ratio reached ca. 280 in the diffusive gas in which the ³⁶Ar isotope was most enriched. The historical lavas analyzed by Matsumoto et al. [111] yielded an ⁴⁰Ar/³⁶Ar ratio from ca. 280 to 295.5. This suggests that the mass fractionation in the volcanic rocks took place during one gas transport. For now, this is the only way that the mass fractionation provides the ³⁶Ar-enrichment. The Iwagi katayamalite ⁴⁰Ar/³⁹Ar analyses revealed that the initial ⁴⁰Ar/³⁶Ar ratios were 195–197, as described before. This value is extremely lower than 280 when the ³⁶Ar isotope is most enriched, suggesting it is hard to obtain during one gas transport. The Iwagi katayamalite crystallized during the albitization of granite by the Li–Na metasomatic fluids as described before. The multiple gas transport might take place during the metasomatism to form the Iwagi katayamalite. However, its detailed mechanism is unknown. As far as we are aware, this is the first example to carry out the ⁴⁰Ar/³⁹Ar analyses of the katayamalite formed during the Li–Na metasomatism of granites. Further analytical results are awaited before a detailed mechanism can be discussed.

8.7. Albitite Formation Model

Based on the results of this study, we constructed a simple model for the petrogenesis of the Iwagi albitites (Figure 56). The fine-grained and medium-grained granites, and the coarse-grained biotite granite (protolith of albitites), were crystallized at 93.5 to 91.5 Ma. Some-times, during the cooling stage of the granite body, carbonate-rich fluids (or melt) were infiltrated into the biotite granite along the shear zone, resulting in formation of the zoned, small-scale calcic endoskarn around Mt. Kuresaka. Such fluids might originate from the interaction between the granite and a local roof-rock limestone, and circulate through the shear zone, driven by the heat of the granite.

Soon after the skarnization, extensive Li–Na metasomatism occurred at 91.5 Ma. The metasomatism progressed through the interaction with external Li–Na–Cl-rich fluids infiltrated along the shear zones. Minerals such as quartz in the host biotite granite were leached by high-Na fluids (i.e., episyenitization), resulting in bleaching and whitening of the metasomatized granite and abundant pore formation in it. The increased porosity allowed the fluids to migrate to the reaction interface, which facilitated rapid diffusion of elements associated with metasomatism [2]. Albite crystallized, replacing K-feldspar and igneous plagioclase and filling the pores, producing albitites. Aegirine-augite also extensively crystallized. As such, albitization was enhanced by the elevated permeability. At relatively high temperatures (600–400 °C), Na–Li–Cl-rich fluids interacted with the central part of the albitized zone, resulting in progressive trace element redistribution. Pores were filled with Li-rich minerals, such as sugilite, murakamiite, and katayamalite. These processes produced the zones of katayamalite albitite and sugilite albitite.

As the leading fluid front progressed beyond the zones of katayamalite albitite and sugilite albitite, the agirine albitite zone composed mainly of albite and aegirine-augite formed. The aegirine albitite shows the highest modal abundance of albite in the Iwagi albitites. Abrupt decreases of Li-rich minerals in this zone might be due to the decreased Li abundance in the metasomatizing fluids resulted from consumption within the zones of katayamalite albitite and sugilite albitite. In the albitized rocks formed by further progress of the leading fluid front, preservation of some minerals and textures of the protolith granite became evident, although the crystallization of albite and aegirine-augite still continued. The quartz albitite zone typically contained 12 vol.% quartz and 2 vol.% alkali feldspar derived from the protolith granite. The albitized granite zone marked the outermost edge of the metasomatic zoning, where typically 24 vol.% quartz, 33 vol.% alkali feldspar and 1 vol.% biotite derived from granite remained, and characterized the appearance of recrystallized albite, aegirine-augite, and ferri-ferro-holmquistite.

Albitized regions also became larger by progressive metasomatism via small fractures. The presence of REE-rich (Ce-rich) veins (Figure 23 and Figure 27) indicates mass transport by channelized flow. Near and around fractures at the leading front, abundances of quartz and alkali feldspar decreased to ~10 vol.% and ~0.5 vol.%, respectively. Zircon was also dissolved by the metasomatizing fluids, and porous zircons formed (Figure 24a).

As such, the albititized zones were arranged in the following order away from the pathway of the metasomatizing fluids toward the leading fluid front: katayamalite albitite, sugilite albitite, aegirine albitite, quartz albitite, and albitized granite. At the southeastern end of the albitite outcrop, the hedenbergite albitite zone formed because there the protolith granite had undergone calcic skarnization along the shear zone. Such metasomatic zoning occurred on a centimeter (Figure 34) to several tens of meter scale over an area of 500 × 250 m² near Mount Kuresaka in Iwagi Islet.

As described earlier, external fluids associated with skarnization that preceded albitization could be supplied from the contact between the local granite and the roof-rock limestone, and this hypothesis is supported by the occurrence of calcic skarn around the Iwagi Islet. However, the occurrence of extensive Li metasomatism combined with extreme albitization is peculiar to the Iwagi Islet and cannot be seen in other localities of episyenite-like rocks in the Setouchi region (Figure 4). Such a strong Li–Na metasomatism requires the involvement of external fluids highly enriched in Li and Na. Based on Li isotope analyses of murakamiite and Li-rich pectolite, Imaoka et al. [24] proposed that such high-Li–Na fluids could be Arima-type fluids derived from dehydration of the subducted oceanic slab (see Section 8.1). If this is the case, it is conceivable that the fluid pathway (shear zones) large enough to allow the deep Arima-type fluids to migrate upward was developed only in the Iwagi region. This might be controlled by the tectonics of this region, which requires further study on this topic in the future.

9. Conclusions

We undertook a petrographical, mineralogical, and geochronological study of the protolith biotite granite and albitites on Iwagi Islet, SW Japan. Our main conclusions are as follows:

(1) The Li contents of albitites (episyenites) in Iwagi Islet are anomalously high relative to episyenite-like rocks in the Setouchi region of SW Japan. Iwagi Islet is the type locality of four new Li minerals: sugilite, katayamalite, murakamiite, and ferro-ferri-holmquistite, all of which are metasomatic in origin.

(2) Various albitites occur in Iwagi Islet. The rocks around Mount Kuresaka can be classified into the protolith biotite granite, albitized granite, quartz albitite, hedenbergite albitite, aegirine albitite, sugilite albitite, and katayamalite albitite. Recrystallized albite, aegirine-augite, and ferro-ferri-holmquistite in addition to original igneous minerals occur in the albitized granite. Quartz albitite contains residual quartz (≤10 vol.%) and newly crystallized aegirine-augite. The hedenbergite albitite, aegirine albitite, sugilite albitite, and katayamalite albitite are characterized by the appearance of minerals, as indicated by their respective names.

(3) Fracturing, curved cleavage, coarse twinning, bent lamellae, and recrystallization characterize most of the rock-forming minerals, which suggests that the rocks formed in the presence of stress and heat, possibly associated with faulting. This suggests that the studied rocks are alkaline metasomatites related to fractures and shear zones that focused the fluid flow and metasomatism. Extensive albitization and formation of abundant Li minerals require the involvement of external Li–Na–Cl-rich fluids, which might be related to the Arima-type deep-seated slab fluids.

(4) The formation of albitites on Iwagi Islet began with the decomposition and dissolution of quartz in the host rock. Mineral resorption and overgrowths occurred due to fluid-induced metasomatism of primary igneous minerals. These replacement textures record the destabilization of minerals. The Iwagi granite and albitites document interface-coupled dissolution–reprecipitation processes in an open system.

(5) Inferred from characteristic mineral assemblage of the hedenbergite albitites (hedenbergite, wollastonite, and andradite), its protolith may have been metasomatic granite that has been subjected to calcic skarnization. Since hedenbergite albitite occurs as a zone at the southeastern end of the albitite outcrop, it is likely an endogenous skarn formed when carbonate-rich fluids (or melts) of limestone origin intruded along shear zones, prior to the Li–Na metasomatism.

(6) The ⁴⁰Ar/³⁹Ar age of katayamalite was 91.50 ± 0.26 Ma (Turonian). On the other hand, the U–Pb age of zircon was 91.5 ± 1.9 Ma for the coarse-grained granite, 93.5 ± 1.7 Ma for the medium-grained granite, and 92.8 ± 1.0 Ma for the whole granite. Geologically, the medium-grained granite could also be a protolith during albitite formation. The average zircon age of the coarse-grained granite is indistinguishable from that of katayamalite, but the medium-grained granite and the whole granite show slightly older ages exceeding analytical errors. This suggests that the Li–Na metasomatism occurred in a relatively later stage during the petrogenesis of the whole granitic body in Iwagi Islet.