Crystal chemistry of uedaite-(Ce) from the Karasugawa mine, Fukushima, Japan, and the Heftetjern pegmatite, Norway

Abstract

The crystal structure of uedaite-(Ce), ideal formula ^A1Mn^2+A2Ce^M1-2Al₂^M3Fe²⁺(SiO₄)(Si₂O₇)O(OH), was studied using single crystals from the Karasugawa mine, Fukushima, Japan (hereafter, KG), and the Heftetjern pegmatite, Norway (HJ), by means of electron microprobe and X-ray diffraction. Both the specimens have similar composition, and are rich in MnO (8.1-8.5 wt%) and FeO (13.8-14.8 wt%). The Fe²⁺/total Fe ratios were estimated as 0.93. Structure refinements converted to R₁ = 1.70 and 1.55% for KG and HJ, respectively. The unit-cell parameters are a = 8.8200(2), b = 5.69346(12), c = 9.9745(3) Å, β = 114.016(3)°, and V = 457.52(2) ų for KG, and a = 8.8212(3), b = 5.69723(11), c = 9.9821(3) Å, β = 114.083(2)°, and V = 458.00(2) ų for HJ. The cation assignments of the A and M sites are as follows: ^A1Mn²⁺_0.71Fe²⁺_0.17Ca_0.12, ^A2Ce_0.40Nd_0.27La_0.09Pr_0.06Sm_0.06Y_0.05Gd_0.03Na_0.01Ca_0.03, ^M1Al_0.85Fe³⁺_0.09Fe²⁺_0.05Ti_0.01, ^M2Al_1.00, and ^M3Fe²⁺_0.91Mg_0.02Al_0.07 for KG, and ^A1Mn²⁺_0.70Fe²⁺_0.16Ca_0.14, ^A2Ce_0.39Nd_0.28La_0.09Sm_0.07Pr_0.06Y_0.06Gd_0.04Na_0.01, ^M1Al_0.83Fe³⁺_0.09Fe²⁺_0.07Ti_0.01, ^M2Al_1.00, and ^M3Fe²⁺_0.90Mg_0.02Al_0.08 for HJ. The unit-cell volume of uedaite is the smallest among allanite-group minerals, resulting from the smallest combination of dominant cations at the A1 and M3 sites, Mn²⁺ and Fe²⁺. In the uedaite structure, the substitution of Ca for Mn²⁺ induces a topological change at the A1 site, as reported in previous studies, and the structure also shows an elongation of the A1-O5 bond. This elongation results from the edge-sharing coordination polyhedra of Fe²⁺-dominated M3 and Al-dominated M1 sites, being slightly tilted to maintain bonding with adjacent SiO₄ tetrahedra. Moreover, a positive correlation between the β angle and the O4-M3-O8 angle was observed. This correlation is suggested to be a general feature across the entire epidote supergroup minerals.

INTRODUCTION

Epidote-supergroup minerals occur in various rock types and geological environments. Among them, the allanite-group minerals are important rock-forming minerals that serve as reservoirs for light rare earth elements (REE; e.g., Gieré and Sorensen, 2004). The general formula of the epidote-supergroup minerals is A₂M₃[T₂O₇][TO₄](O, F)(OH, O). In the crystal structure, the A sites are subdivided into A1 and A2, while the M sites are subdivided into M1, M2, and M3. The volumes of the octahedral sites decrease in the order M3 > M1 > M2. The tetrahedral T sites are primarily filled with Si. Only a few species contain fluorine or lack a hydroxyl (OH) group. The monoclinic crystal structure consists of Si₂O₇ and SiO₄ units linked by two types of chains running parallel to the b-axis, formed from edge-sharing octahedral sites (Ito et al., 1954; Dollase, 1968). One chain consists of M2 octahedra (single-chain), while the other forms a zigzag chain of central M1 octahedra with M3 octahedra attached on alternating sides along its length. The M sites are primarily occupied by trivalent cations, such as Al³⁺, Fe³⁺, and Mn³⁺. The M2 site exhibits a strong preference for Al, while the occupancies of the M1 and M3 sites depend on the competing ions. Divalent cations often occupy the M sites, especially the M3 site, with the heterovalent substitution observed in the allanite group and REE-rich specimens: Ca²⁺(A2) + Me³⁺(M3) ↔ REE³⁺(A2) + Me²⁺(M3) (Armbruster et al., 2006). The key cation defining the root name are at the A1 and M3 sites in the epidote-supergroup minerals. Understanding the crystal chemistry of the allanite group, which has diverse compositions reflecting its formation environment and exhibits various cation substitutions across multiple polyhedra, is expected to provide deeper insights into the mechanisms of elemental concentration, solubility limits, and the relationship between cation behavior and structural variation. However, despite extensive studies being reported (e.g., Gieré and Sorensen, 2004, and references therein), the crystal chemistry of allanite remains only partially understood.

Uedaite and androsite are known as the ^A1Mn²⁺-analogue series belonging to the allanite group. Uedaite-(Ce) was approved as a new mineral species in 2006, and its occurrence and crystal chemistry have been reported by Miyawaki et al. (2008). Uedaite and androsite series are defined by the formulas, ^A1Mn^2+A2REE^3+M1Me^3+M2Al^M3Fe²⁺(SiO₄)(Si₂O₇)O(OH), and ^A1Mn^2+A2REE^3+M1Me^3+M2Al^M3Mn²⁺(SiO₄)(Si₂O₇)O(OH), respectively, which are characterized by the substitution of Fe²⁺ ↔ Mn²⁺ at the M3 site. Uedaite can also be described as an analogue of allanite with Mn²⁺ at the Al site replacing Ca. In terms of chemical properties, therefore, the crystal structure of uedaite is expected to have intermediate features between that of androsite and allanite. However, Nagashima et al. (2015, 2026) highlighted the unique crystal structure of uedaite compared to other ^A1Mn²⁺-rich epidote-supergroup minerals. However, the only verified crystal structure of uedaite is that reported by Miyawaki et al. (2008), making it challenging to establish its uniqueness. In this study, uedaite from two different localities was examined to investigate its crystal-chemical properties and clarify the uniqueness of its crystal structure.

SAMPLES

Uedaite-(Ce) from the Karasugawa mine in Fukushima, Japan, has been reported by Harada et al. (2020). The Karasugawa mine is located within the hornblende-biotite granodiorite, which belongs to the Late Cretaceous granitoids of the Abukuma terrain (e.g., Kubo et al., 2003; Ishihara and Chappell, 2008). Small pegmatite deposits are found in granitoids of the Abukuma Mountains, with approximately 300 small and medium-sized mines recorded (Kubo et al., 2003). Since Wakita and Nagashima (1968) reported that the Karasugawa mine was a pegmatite, it can be assumed to be one of the small pegmatites within the Abukuma granitoids mentioned above. Hasegawa (1957, 1960) confirmed Mn-bearing allanite-(Ce) from such pegmatites in the Abukuma Massif. In particular, his specimens from Ohari and Shiozawa showing Mn-analogues of allanite-(Ce) composition are considered to be uedaite-(Ce). It is assumed that our specimen from the Karasugawa mine and his samples formed through similar processes and under comparable conditions. In our study, the specimens provided by Mr. S. Harada were collected from the mine dumps. From this locality, the occurrence of xenotime-(Y) and a fergusonite-like mineral in microcline has been reported by Wakita and Nagashima (1968), and the presence of bastnäsite and clinofergusonite-(Y), formally ‘β-fergusonite-(Y)’, has also been documented by Sakurai et al. (1969) and Kato (1973), respectively. Recently, Harada et al. (2020) re-examined the specimens from this locality and reported quartz, potassium feldspar, Fe-tourmaline, and muscovite as the main components. Along with the previously reported xenotime, clinofergusonite, and bastnäsite, they also identified uedaite-(Ce), hydroxybastonäsite-(Ce), monazite-(Ce), zircon, anatase, brookite, Y-bearing almandine, a yttrium silicate with thalénite-like composition, and a thorium silicate with thorite-like composition. The crystals of uedaite-(Ce) range from black to grayish black in color with a vitreous luster, and are embedded in quartz. The thin prismatic crystals are usually 1.5 mm wide and about 30 mm long.

The occurrence of uedaite-(Ce) from the Heftetjern pegmatite in Norway was first reported by Kristiansen (2017, 2018). He noted that uedaite-(Ce) cannot be distinguished from the more common allanite-(Ce) at this locality without chemical analysis. According to Steffenssen et al. (2020), the Heftetjern pegmatite is one of the over 300 large pegmatites in the Tørdal district. These are located within a 3 km-wide, NE-SW-trending 10 km belt that lies within the volcanosedimentary supracrustal sequence of the Nissedal Outlier (1200-1300 Ma). Both the outlier and the older basement are intruded by the Tørdal-Treungen granite, which is regarded as the source of the pegmatite swarms (Bergstøl and Juve, 1988). Bergstøl and Juve (1988) reported that the main minerals of the Heftetjern pegmatite are amazonite, feldspar, microcline, albite-oligoclase, cleavelandite, quartz, and mica. Details of the mineral assemblage of the Heftetjern pegmatite were summarized by Kristiansen (2009). This pegmatite is distinguished by local enrichment of Sc and Sn, and therefore, some Sc-analogue minerals (kristiansenite, heflikite, heftetjernite, oftedalite, cascandite, bazzite, scandiobabingtonite, and thortveitite), cassiterite, and Sc-Sn-bearing ixiolite have been reported as accessory minerals alongside cerium- or yttrium-rich minerals, such as gadolinite, allanite, and monazite (e.g., Bergstøl and Juve, 1988; Segalstad and Eggleston, 1993; Raade and Kristiansen, 2000; Steffenssen et al., 2020). The present uedaite-(Ce) specimens, provided by Dr. Kristiansen, are grayish-black prismatic crystals associated with small amounts of biotite and spessartine in a quartz-feldspar matrix.

Hereafter, the uedaite-(Ce) from the Karasugawa mine and that from the Heftetjern pegmatite are referred to as ‘KG’ and ‘HJ’, respectively.

EXPERIMENTAL

Chemical analysis

Chemical analyses of the resin-mounted crystals were conducted using an electron microprobe analyzer (JEOL JXA-8230) installed at the Centre for Instrumental Analysis, Yamaguchi University, Japan. Afterward, crystals with the specific chemical composition from each locality were extracted from the resin-mounts for the collection of X-ray single-crystal diffraction data. Operating conditions were: an accelerating voltage of 15 kV, a beam current of 20 nA, and a beam diameter of 1-5 µm. Wavelength-dispersive X-ray spectra were measured using LiF, PET, and TAP monochromators to identify interfering elements and determine the best wavelengths for background measurements. The abundances of Si, Ti, Al, Cr, V, Fe, Mn, Ni, Mg, Ca, Sr, Ba, Na, K, P, F, Cl, Y, La, Ce, Pr, Nd, Sm, and Gd were measured. The peak and background positions of each element were carefully verified to prevent overlap. Several elements not listed in Table 1 were below the detection limit. Although the presence of other REE, such as Eu, Dy, Ho, and Er, was carefully confirmed in a preliminary analysis, they were all found to be below the detection limit. The ZAF correction was applied to all elements. The H₂O was not measured directly due to insufficient sample mass, but single-crystal XRD refinement confirmed the position of hydrogen. The slightly low total wt% may be due to the effects of cracks hidden beneath the beam spot.

Table 1. Chemical compositions of uedaite-(Ce)

Oxide (wt%)	Karasugawa mine (KG)			Heftetjern, Norway (HJ)
	Ave	Std	Range	Ave	Std	Range
	n = 16			n = 11
SiO2	29.72	0.14	29.43-30.04	29.92	0.24	29.57-30.34
TiO2	0.08	0.05	0.00-0.15	0.07	0.06	0.00-0.14
Al2O3	15.85	0.13	15.54-16.03	15.97	0.13	15.77-16.11
V2O31)	0.02	0.03	0.00-0.07	0.02	0.02	0.00-0.06
FeO1)	14.31	0.24	13.84-14.67	14.53	0.19	14.23-14.80
MnO1)	8.22	0.12	8.05-8.43	8.26	0.14	8.11-8.52
NiO	0.04	0.04	0.00-0.13	0.02	0.02	0.00-0.07
MgO	0.15	0.01	0.13-0.17	0.16	0.02	0.13-0.19
CaO	1.36	0.03	1.32-1.40	1.27	0.08	1.10-1.42
Na2O	0.07	0.02	0.04-0.12	0.05	0.03	0.00-0.10
Y2O3	0.99	0.07	0.89-1.10	1.15	0.08	0.99-1.25
La2O3	2.37	0.08	2.24-2.51	2.36	0.11	2.24-2.56
Ce2O3	10.62	0.15	10.27-10.89	10.58	0.23	10.20-10.89
Pr2O3	1.61	0.15	1.35-1.85	1.68	0.12	1.45-1.92
Nd2O3	7.43	0.23	7.07-8.00	7.66	0.18	7.24-7.83
Sm2O3	1.85	0.14	1.54-2.08	1.93	0.18	1.52-2.14
Gd2O3	0.99	0.18	0.65-1.36	1.08	0.16	0.83-1.28
F	0.24	0.11	0.06-0.50	0.27	0.10	0.08-0.43
-O=F	0.10			0.12
Total	95.82			96.86
Σcations = 8
Si	3.02	0.01	3.00-3.04	3.01	0.02	2.98-3.03
Ti	0.01	0.00	0.00-0.01	0.01	0.00	0.00-0.01
Al	1.90	0.01	1.86-1.92	1.90	0.01	1.87-1.91
V3+	0.00	0.00	0.00-0.01	0.00	0.00	0.00-0.01
Fe2)	1.22	0.02	1.18-1.25	1.22	0.02	1.20-1.25
Mn2+	0.71	0.01	0.69-0.73	0.70	0.01	0.69-0.73
Ni	0.00	0.00	0.00-0.01	0.00	0.00	0.00-0.01
Mg	0.02	0.00	0.02-0.03	0.02	0.00	0.02-0.03
Ca	0.15	0.00	0.14-0.15	0.14	0.01	0.12-0.15
Na	0.01	0.00	0.01-0.02	0.01	0.01	0.00-0.02
Y	0.05	0.00	0.05-0.06	0.06	0.00	0.05-0.07
La	0.09	0.00	0.08-0.09	0.09	0.00	0.08-0.10
Ce	0.40	0.01	0.38-0.40	0.39	0.01	0.38-0.40
Pr	0.06	0.01	0.05-0.07	0.06	0.00	0.05-0.07
Nd	0.27	0.01	0.26-0.29	0.28	0.01	0.26-0.28
Sm	0.06	0.00	0.05-0.07	0.07	0.01	0.05-0.07
Gd	0.03	0.01	0.02-0.05	0.04	0.01	0.03-0.04
Total	8.00			8.00
F−	0.08	0.04	0.02-0.16	0.09	0.03	0.03-0.14

¹⁾ V, Fe, and Mn as V₂O₃, FeO, and MnO.

²⁾ Based on a total positive charge of 25 to maintain charge balance, the oxidation state of Fe was estimated as 1.13Fe²⁺ + 0.09Fe³⁺ for both specimens.

Single-crystal structure analysis

X-ray diffraction data for the single crystals were collected using a Rigaku XtaLAB Synergy-R/DW diffractometer equipped with a HyPix-6000HE detector, installed at Yamaguchi University, Japan. The crystals [0.06 × 0.09 × 0.13 mm³ for KG and 0.04 × 0.08 × 0.11 mm³ for HJ] were mounted on a glass fiber. Intensity data were collected at room temperature using MoKα radiation (λ = 0.71073 Å). Preliminary unit-cell parameters and an orientation matrix were derived from six sets of frames and refined during the integration process of the intensity data. CrysAlis^Pro (Agilent, 2014) was used to process the diffraction data and to correct empirical absorption. The reflection statistics and systematic absences were consistent with space groups P2₁ and P2₁/m. Subsequent attempts to solve the structure revealed that the observed average structure is centrosymmetric, indicating that P2₁/m is the correct space group. Structural refinement was performed using SHELXL-2019/3 (Sheldrick, 2015). Scattering factors for neutral atoms were employed. To obtain the site-scattering values for each cation site, the site occupancies were refined for Mn and Ca at A1, Ce and Ca at A2, and Al and Fe at M1 and M3. Although the site occupancy at M2 for both samples was refined with Al and Fe scattering factors during the preliminary refinement, as well as for the other octahedral sites, it was fixed at 1.0 Al in the final refinement because it was fully occupied by Al within the standard deviation. Moreover, the direct-occupancy refinement at the T sites converged to full occupancy, with Si scattering factors within the standard deviation. Then, their occupancies were fixed at 1.0 Si, as supported by the microprobe results. Final cation assignments were determined using the average chemical compositions of each measured crystal. Hydrogen positions of the hydroxyl groups were also derived from difference-Fourier synthesis and refined with U_iso fixed at 0.05 Å and a restraint of O-H = 0.980(1) Å (Franks, 1973).

RESULTS

Chemical compositions

The chemical compositions of the uedaite-(Ce) are given in Table 1, where the total number of cations, except H, was normalized to eight. Both uedaite samples are chemically homogeneous (Fig. 1) and have comparable compositions (Table 1). They are rich in Mn and Fe, as the typical of uedaite. Of the 1.22 Fe atoms per formula unit (apfu) in both specimens (Table 1), all Fe assigned to the A site must be ferrous. In addition, Fe at the M3 site is also ferrous based on the substitution scheme of allanite-group minerals. The Fe²⁺/Fe³⁺ ratio was calculated assuming a total positive charge of 25 to maintain charge neutrality. As a result, Fe²⁺/(Fe²⁺ + Fe³⁺) is 0.93 in both KG and HJ, thus the total 1.22 Fe apfu = 1.13Fe²⁺ + 0.09Fe³⁺. Although the predominant REE in both specimens is Ce, Nd is the second dominant. The empirical formulas based on the average chemical data are ^A(Ca_0.15Mn²⁺_0.71Fe²⁺_0.17Na_0.01Ce_0.40Nd_0.27La_0.09Pr_0.06Sm_0.06Y_0.05Gd_0.03)_Σ2^M(Al_1.90Fe²⁺_0.96Fe³⁺_0.09Mg_0.02Ti_0.01)_Σ2.98^TSi_3.02O₁₂(OH)_0.92F_0.08 for KG (n = 16), and ^A(Ca_0.14Mn²⁺_0.70Fe²⁺_0.16Na_0.01Ce_0.39Nd_0.28La_0.09Pr_0.06Sm_0.07Y_0.06Gd_0.04)_Σ2^M(Al_1.90Fe²⁺_0.97Fe³⁺_0.09Mg_0.02Ti_0.01)_Σ2.99^TSi_3.01O₁₂(OH)_0.91F_0.09 for HJ (n = 11). Among the epidote-supergroup minerals, which occur in a variety of environments and have a wide range of chemical compositions, it is very rare for significant amounts of Fe²⁺ to be distributed in the A site. This characteristic may be unique to uedaite, as it was also observed in the type specimen from Shodoshima, Kagawa, Japan, studied by Miyawaki et al. (2008).

Figure 1. Back-scattered electron image and X-ray maps showing the distribution of Mn, Fe, and Ce of the uedaite specimen from the Karasugawa mine (a) and Heftetjern (b).

Crystal structure refinements and determination of site occupancies

Crystallographic data and refinement parameters are summarized in Table 2. The refined atomic positions and thermal displacement parameters are listed in Supplementary Table S1 (Supplementary Tables S1-S3 are available online from https://doi.org/10.2465/jmps.260125). Interatomic distances, selected angles, and distortions of octahedral sites are present in Table 3. The crystal structure of uedaite-(Ce) is shown in Figure 2. The unit-cell volume of uedaite is the smallest among all previously reported allanite-group minerals, and in particular, the c-dimension of the KG specimen, 9.9745(3) Å, is likely the shortest among all reported epidote-supergroup minerals.

Table 2. Data collection and details of structure refinement

Sample locality		Karasugawa mine	Heftetjern, Norway
Sample code		KG	HJ
Crystal size (mm3)		0.06 × 0.09 × 0.13	0.04 × 0.08 × 0.11
Space group		Monoclinic P21/m
Unit-cell dimensions	a (Å)	8.8200(2)	8.8212(3)
	b (Å)	5.69346(12)	5.69723(11)
	c (Å)	9.9745(3)	9.9821(3)
	β (°)	114.016(3)	114.083(3)
	V (Å3)	457.52(7)	458.00(7)
Z		2
Dcalc (g/cm3)		4.34	4.34
Radiation		MoKα (λ = 0.71073 Å)
Monochromator		VariMax optics
Diffractometer		Rigaku XtaLAB Synergy-R/DW with HyPix-6000HE
Scan type		ω scan
Absorption correction		CrysAlisPro (Agilent, 2014)
Absorption coefficient μ (mm−1)		8.49	8.48
θmin-θmax(°)		2.2-40.2	2.2-40.2
Collected reflections		15561	15664
Unique reflections		3084	3087
Criterion for observed reflections
Rint (%), Rσ (%)		1.90, 1.23	1.83, 1.31
Index ranges		−16 ≤ h ≤ 16, −10 ≤ k ≤ 10, −18 ≤ l ≤ 18	−16 ≤ h ≤ 16, −10 ≤ k ≤ 10, −18 ≤ l ≤ 18
Refinement on F2 using		SHELXL-2019/3 (Sheldrick 2015)
R1 (%), wR2 (%)		1.70, 4.57	1.55, 4.06
No. of parameters		124	124
Weighting scheme1)		w = 1/[σ2(Fo2) + (0.0238P)2 + 0.29P]	w = 1/[σ2(Fo2) + (0.0209P)2 + 0.31P]
Δρmax (e Å−3)		0.84 (1.04 Å from H10)	2.08 (0.62 Å from M3)
Δρmin (e Å−3)		−2.80 (0.47 Å from A2)	−1.05 (0.47 Å from A2)

¹⁾ The function of the weighting scheme is w = 1/[σ²(F_o²) + (a·P)² + b·P], where P = [Max(F_o², 0) + 2F_c²]/3, and the parameters a and b are chosen to minimize the differences in the variances for reflections in different ranges of intensity and diffraction angle.

Table 3. Selected interatomic distances (Å), bond angles (°), and distortion parameters¹⁾ for the octahedrally coordinated sites

			Karasugawa, KG	Heftetjern, HJ				Karasugawa, KG	Heftetjern, HJ
A1-	O1	×2	2.258(1)	2.2599(9)	M1-	O1	×2	1.9979(9)	1.9971(8)
	O3	×2	2.2326(9)	2.2348(8)		O4	×2	1.8470(8)	1.8468(7)
	O5		2.646(2)	2.647(2)		O5	×2	1.9825(8)	1.9834(8)
	O7		2.258(1)	2.260(1)		VI<M1-O>		1.943	1.943
	VI<A1-O>		2.314	2.316	DI (oct)			0.033	0.033
	O6		3.0282(11)	3.028(1)	<λ oct>			1.007	1.007
	O9		3.0728(12)	3.0693(12)	σθ (oct)2			16.05	16.35
	O9	×2	3.1128(6)	3.1141(6)	O1-M1-O4			88.41(4)	88.42(4)
δ[(A1-O6)-(A1-O5)]			0.382	0.381	O1-M1-O5			90.26(4)	90.30(4)
					O4-M1-O5			96.44(4)	96.51(4)
A2-	O2	×2	2.5857(9)	2.5854(8)
	O2′	×2	2.4512(9)	2.4495(8)
	O3	×2	2.8330(9)	2.8354(9)	M2-	O3	×2	1.8835(8)	1.8833(8)
	O7		2.319(1)	2.319(1)		O6	×2	1.8966(7)	1.8982(7)
	O10		2.570(1)	2.566(1)		O10	×2	1.8979(7)	1.8989(7)
	O8	×2	2.9549(4)	2.9578(4)		VI<M2-O>		1.8926	1.894
	X<A2-O>		2.654	2.654	DI (oct)			0.003	0.004
					<λ oct>			1.006	1.006
Si1-	O1	×2	1.6425(9)	1.6431(8)	σθ (oct)2			19.95	19.89
	O7		1.602(1)	1.601(1)	O3-M2-O6			89.68(5)	89.67(4)
	O9		1.637(1)	1.639(1)	O3-M2-O10			91.07(4)	91.04(4)
	IV<Si1-O>		1.631	1.631	O6-M2-O10			97.32(3)	97.32(3)
O1-Si1-O1′			113.46(7)	113.48(6)
O1-Si1-O7			112.40(4)	112.41(4)
O1-Si1-O9			106.97(4)	106.95(4)	M3-	O1	×2	2.2922(9)	2.2946(9)
O7-Si1-O9			103.84(7)	103.85(7)		O2	×2	2.1893(9)	2.1912(8)
						O4		2.007(1)	2.006(1)
Si2-	O3	×2	1.6368(9)	1.6366(8)		O8		2.000(1)	1.997(1)
	O8		1.602(1)	1.601(1)		VI<M3-O>		2.162	2.162
	O9		1.633(1)	1.633(1)	DI (oct)			0.049	0.050
	IV<Si2-O>		1.627	1.627	<λ oct>			1.058	1.057
O3-Si2-O3′			113.10(6)	113.07(6)	σθ (oct)2			182.91	181.44
O3-Si2-O8			110.04(4)	110.05(4)	O1-M3-O1′			80.00(5)	79.98(4)
O3-Si2-O9			107.76(4)	107.76(4)	O1-M3-O2			92.49(3)	92.53(3)
O8-Si2-O9			107.95(7)	107.99(7)	O1-M3-O4			76.90(3)	76.86(3)
					O1-M3-O8			115.42(4)	115.27(3)
Si3-	O2	×2	1.6257(9)	1.6273(8)	O2-M3-O2′			92.19(5)	92.15(4)
	O5		1.652(1)	1.653(1)	O2-M3-O4			89.77(3)	89.87(3)
	O6		1.643(1)	1.643(1)	O2-M3-O8			78.62(3)	78.69(3)
	IV<Si3-O>		1.637	1.638
O2-Si3-O2′			102.66(7)	102.65(6)	O10…O4			2.893(2)	2.898(2)
O2-Si3-O5			113.61(4)	113.61(4)	Si1-O9-Si2			134.89(9)	134.94(9)
O2-Si3-O6			111.67(4)	111.64(4)
O5-Si3-O6			103.92(7)	103.96(6)

¹⁾ DI(oct) is the bond-length distortion parameter (Baur, 1974). <λ_oct> and σ_θ(oct)² are the polyhedral volume distortion and angle variance, respectively (Robinson et al., 1971).

Figure 2. Crystal structure of uedaite-(Ce) using the data obtained from the Karasugawa crystal projected down [010] (a) and the topological relationship among the A1, M1, and M3 octahedra (b), drawn with VESTA 3 (Momma and Izumi, 2011). Dashed lines indicate H···O bond.

The number of electrons at each cation site, derived from the refined site occupancies and cation assignments for the A and M sites, is listed in Table 4. The cation assignments at the A1, A2, M1, M2, and M3 sites were determined after the previous structural studies of epidote-supergroup minerals (e.g., Armbruster et al., 2006). The procedure is as follows: (1) Na and REE, such as La, Ce, and Nd, were assigned to the A2 site, and the deficiency was filled with the Ca ions. (2) After all the remaining Ca is assigned to the A1 site, the Mn²⁺ ions are assigned. Since Ca and Mn²⁺ did not completely fill the A2 site, the deficiency was filled with Fe²⁺, the largest among the remaining divalent cations. (3) Only Al ion was assigned to the smallest M2 octahedral site based on the results of structural analysis. (4) The cations at the M1 and M3 sites were assigned based on the observed number of electrons, calculated from the refined site occupancy value and their ionic radii. The divalent octahedral cations, Mg²⁺ and remaining Fe²⁺, are favorable to distribute the largest M3 octahedral site due to the large ionic radii compared to trivalent cations. The refined Al occupancy at the M3 site is interpreted as the sum of Al and Mg. For instance, regarding the M3 site of KG, the refined Al occupancy, 0.099, can estimate 0.02 Mg apfu at that site. Therefore, the cation assignment for the M3 site turns out to be Fe²⁺_0.91Mg_0.02Al_0.07. Subsequently, the remaining Fe²⁺ ions [0.05 apfu = 1.13 apfu − (0.17 at A1 + 0.91 at M3)], Fe³⁺, Ti, and Al are assigned to the M1 site based on the result of chemical analysis. As a result of the above scheme, the structural formulae can be represented as ^A1(Mn²⁺_0.71Fe²⁺_0.17Ca_0.12)_Σ1^A2(Ce_0.40Nd_0.27La_0.09Pr_0.06Sm_0.06Y_0.05Gd_0.03Na_0.01Ca_0.03)_Σ1^M1(Al_0.85Fe³⁺_0.09Fe²⁺_0.05Ti_0.01)_Σ1^M2Al^M3(Fe²⁺_0.91Mg_0.02Al_0.07)_Σ1Si_3.00O₁₂(OH) for KG and ^A1(Mn²⁺_0.70Fe²⁺_0.16Ca_0.14)_Σ1^A2(Ce_0.39Nd_0.28La_0.09Sm_0.07Pr_0.06Y_0.06Gd_0.04Na_0.01)_Σ1^M1(Al_0.83Fe³⁺_0.09Fe²⁺_0.07Ti_0.01)_Σ1^M2Al^M3(Fe²⁺_0.90Mg_0.02Al_0.08)_Σ1Si_3.00O₁₂(OH) for HJ.

Table 4. Refined site occupancies and cation assignments with numbers of electrons of the A and M sites¹⁾

Site	Refined site occupancy	Observed no. of e−	Cation assignment	Estimated no. of e−
Karasugawa (KG)
A1	Mn0.940(8)Ca0.06	24.70	Mn2+0.71Fe2+0.17Ca0.12	24.57
A2	Ce0.960(2)Ca0.040	56.48	(Ce0.40Nd0.27La0.09Pr0.06Sm0.06Y0.05Gd0.03)Σ0.96Na0.01Ca0.03	56.37
M1	Al0.911(3)Fe0.089	14.16	Al0.85Fe3+0.09Fe2+0.05Ti0.01	14.91
M2	Al1.0	13	Al1.0	13
M3	Fe0.901(4)Al0.099	24.71	Fe2+0.91Mg0.02Al0.07	24.81
Heftetjern (HJ)
A1	Mn0.967(7)Ca0.033	24.84	Mn2+0.70Fe2+0.16Ca0.14	24.46
A2	Ce0.9498(19)Ca0.0502(19)	56.09	(Ce0.39Nd0.28La0.09Sm0.07Pr0.06Y0.06Gd0.04)Σ0.99Na0.01	57.44
M1	Al0.905(3)Fe0.095	14.24	Al0.83Fe3+0.09Fe2+0.07Ti0.01	15.17
M2	Al1.0	13	Al1.0	13
M3	Fe0.897(4)Al0.103	24.66	Fe2+0.90Mg0.02Al0.08	24.68

¹⁾ The chemical compositions are fixed by EPMA data. The oxidation state of Fe at the M1 site is estimated based on the charge-balance calculation.

Bond-valence sums were calculated using the electrostatic strength function of Brown and Altermatt (1985) and the bond-valence parameters of Gagné and Hawthorne (2015). The results are given in Table 5. The calculated bond-valence sums and refined hydrogen positions indicate that one hydroxyl group is located at O10 (donor, 1.31 valence unit, v.u.) with O4 acting as acceptor (1.65 v.u.).

Table 5. Bond-valence analysis (v.u.) weighted on the cation assignments of Table 4

Site	KG	HJ
A1	1.81	1.83
A2	2.70	2.73
M1	2.86	2.88
M2	3.10	3.10
M3	1.99	1.99
Si1	3.93	3.92
Si2	3.98	3.98
Si3	3.88	3.86
O1	1.90	1.90
O2	1.97	1.97
O3	1.96	1.97
O4	1.65	1.65
O5	1.89	1.91
O6	2.02	2.02
O7	1.91	1.92
O8	1.74	1.74
O9	2.07	2.06
O10	1.31	1.31

DISCUSSION

Allanite group minerals are characterized by REE³⁺ at the A2 sites and their species are distinguished by the dominant cations at the A1 and M3 sites. The M3 site is primarily occupied by divalent cations (Me²⁺), such as Fe²⁺ in allanite and uedaite, Mn²⁺ in akasakaite and androsite, and Mg²⁺ in dissakisite. Ca is the predominant cation at the A1 site in allanite and akasakaite, whereas Mn²⁺ dominates at the A1 site in uedaite and androsite. Namely, uedaite can be a Mn-analogue of allanite, substituted with Mn²⁺ for Ca in the A1 site, and also an Fe-analogues of androsite, substituted with Fe²⁺ for Mn²⁺ in the M3 site. In many cases, structural variations in the epidote supergroup minerals have been studied primarily within individual series due to their wide range of chemical compositions and substitutional relationships. In the allanite group, the relationship between cation substitution and structural variation has typically been discussed separately for Mn²⁺-free/poor and Mn²⁺-rich specimens. However, uedaite is a species that can help address these distinct issues separately.

Effect of Ca ↔ Mn substitution in the A1 site

The topological change at the A1 site caused by the substitution of Mn²⁺ for Ca has been repeatedly confirmed in studies (Bonazzi et al., 1996; Nagashima et al., 2010, 2013, 2015; Biagioni et al., 2019). With increasing the occupancy of Mn²⁺ at the A1 site, the oxygen atoms at the O3 and O1 sites move closer to the cation at the A1 site, while the O atoms at the O6 and O9 sites (7^th-9^th neighbor O atoms) shift away from the cation at the A1 site. In contrast, the positions of the O5 (6^th neighbor) and O7 sites remain unaffected by the Mn²⁺ content at the A1 site. Bonazzi et al. (1996) first suggested the positive correlation between δ[(A1-O6)-(A1-O5)] and occupancy of Mn²⁺ at the A1 site, and Biagioni et al. (2019) proposed the regression equation, δ[(A1-O6)-(A1-O5)] = 0.32(1) + 0.28(3) × Mn²⁺ at the A1 site (Fig. 3). However, the observed value of δ[(A1-O6)-(A1-O5)] for the uedaite specimens in this study (∼ 0.38 Å) and Miyawaki et al. (2008) (0.36 Å) are much shorter than the expected values (∼ 0.52 Å for this study) using this equation (Supplementary Table S2). These shorter δ[(A1-O6)-(A1-O5)] values in the uedaite specimens are derived from the lengthened A1-O5 distance relative to the other ^A1Mn²⁺-rich member of the allanite group minerals. However, this difference should not be interpreted as an anomaly unique to uedaite.

Figure 3. The variation of δ[(A1-O5)-(A1-O6)] (Å) as a function of Mn²⁺ content at the A1 site. The solid line is the regression line suggested by Biagioni et al. (2019): δ[(A1-O5)-(A1-O6)] (Å) = 0.32(1) + 0.28 × Mn²⁺ at A1 (apfu). The relationship between uedaite and ^M1Al-dominant allanite is represented by the break line. The data is summarized in Table S2. Data sources are as follows: This study and Miyawaki et al. (2008) for uedaite-(Ce); Dollase (1971), Kartashov et al. (2002), Hoshino et al. (2005, 2010), Orlandi and Pasero (2006), Hoshino et al. (2010), Nagashima et al. (2011), Kolitsch et al. (2012), Nagashima et al. (2013), Škoda et al. (2012), and Pieczka et al. (2025) for allanite; Nagashima et al. (2015) and Biagioni et al. (2019), and Nagashima et al. (2026) for akasakaite; Cenki-Tok et al. (2006), Nagashima et al. (2015), and Števko et al. (2023) for androsite; Peacor and Dunn (1988), Sokolova et al. (1991), Pautov et al. (1993), and Ragu et al. (2025) for dollaseite group minerals.

The variation in the δ[(A1-O6)-(A1-O5)] value simply seems to correlate with the occupancy of Mn²⁺ in the A1 site, among epidote supergroup minerals having different cation distributions at other sites, such as androsite, akasakaite, allanite, and piemontite in the data set (Supplementary Tables S2 and S3). It is, however, clear not to solely reflect the effect of substituting Mn²⁺ for Ca in the A1 site. Therefore, the proposed linear correlation seems to reflect a combined effect involving both the A and M sites. The comparison between ^M1Al-dominant allanites and uedaite, where Fe²⁺ and Al dominate the M3 and M1 sites, respectively, in this study, is appropriate for examining the effect of Ca ↔ Mn substitution in the A1 site. As a result, the contribution of the substitution at the A1 site is represented by the equation, δ[(A1-O6)-(A1-O5)] = 0.301(6) + 0.11(1) × Mn²⁺ at the A1 site (break line in Fig. 3). However, it should be noted that the A1-O5 elongation and the newly proposed equation are only applicable when the Mn substitutes for Ca at the A1 site in the ^M3Fe²⁺-dominated species, uedaite and allanite. It remains to be further investigated whether a similar equation can be derived for the akasakaite-androsite series, where Mn²⁺ dominates at the M3 site. No significant differences in the Ca/Mn ratio at the A1 site were observed between these species (Fig. 3), and the variety of cation distributions at sites other than the A1 site makes it challenging to isolate the pure effect of Ca ↔ Mn²⁺ substitution at the A1 site.

In the crystal structure of uedaite, the A1-O5 elongation results from a slight tilt of the edge-sharing M3 and M1 octahedra, which can be interpreted as the octahedra adopting a geometric arrangement suited for maintaining bonding with adjacent SiO₄ tetrahedra (Fig. 2). This tilt shifts the O5 site away from A1, lengthening the A1-O5 bond, and moves O1 closer to A1, shortening the A1-O1 bond (Fig. 2b). These shifts are caused by the Fe²⁺-dominated M3 and Al-dominated M1 sites at the edge-sharing being slightly tilted (or rotated) to maintain bonding with adjacent SiO₄ tetrahedra. This phenomenon is not observed in the akasakaite-androsite series, which is characterized by the Mn²⁺-dominated M3 site, even when the M1 site is predominantly occupied by Al. Therefore, the tilt of edge-shared M3 and M1 octahedra to maintain bonding with adjacent SiO₄ tetrahedra can be explained by the fact that the M3 site becomes more compact when filled with Fe²⁺ ions, which have a smaller ionic radius than Mn²⁺ ions. However, when examining the A1-O5 bond, the effect of cation substitution on the M1 site, which directly bonds to O5, cannot be avoided.

For instance, dollaseite-(Ce) and khristovite-(Ce) also have comparable δ[(A1-O6)-(A1-O5)] values (∼ 0.51 Å; Peacor and Dunn, 1988; Sokolova et al., 1991; Pautov et al., 1993) despite the absence of Mn²⁺ at the A1 site (Fig. 3 and Table S2). This is due to the short A1-O5 bond, which results from M1-O5 elongation, mainly caused by Mg²⁺ at the M1 site. It indicates that if the M1 site is sufficiently large, increasing Mn²⁺ in the A1 site has a limited effect on δ[(A1-O6)-(A1-O5)]. Conversely, when the M1 site is small, an increase in Mn²⁺ in the A1 site significantly influences δ[(A1-O6)-(A1-O5)]. This cannot be inferred from the crystal structure, which varies with the interaction of various substitutions. Therefore, the relationship between occupancy of Mn²⁺ at the A1 site and the value of δ[(A1-O6)-(A1-O5)] is practical for specimens rich in Mn²⁺ at the A1 and M3 sites, as confirmed by numerous previous studies, but does not apply to uedaite.

Variation of the beta angle

The variation in the β angle has previously been attributed to the cation substitution at specific sites. For example, a decrease in the β angle with increasing REE³⁺ at the A2 site has repeatedly been reported (e.g., Gieré and Sorensen, 2004). In addition, Nagashima et al. (2015) noted that increasing Mn²⁺ at the A1 site also results in a decrease in the β angle. Considering these relationships, Nagashima et al. (2026) confirmed the correlation between Σ(Mn + REE + Sr) at the A sites and the β angle. However, the specific structural changes that cause the β-angle to vary with increasing Mn²⁺ and REE concentrations remain unclear. In this study, we examine the O4-M3-O8 angle, which varies with cation substitution, and show that the variation in the β angle can be explained through a unified topological approach.

In the epidote supergroup minerals, a positive correlation is observed between the β angle and the O4-M3-O8 angle (Fig. 4), and is supported by the topological configuration (Fig. 2b). Substituting trivalent or divalent cations for Al at the M3 site lengthens the M3-Oi bonds. Notably, in species where REE³⁺ predominantly occupies the A2 site, Me²⁺ substitution expands the M3 site, accompanied by deformation and an increase in M3-Oi bond lengths. An increase in the M3-O4 and M3-O8 bonds by Me²⁺ substitution leads to a decrease in the O4-M3-O8 angle. In this study, the O4-M3-O8 angle was measured along the direction of the interior angle of uedaite (Fig. 2b). The O4-M3-O8 angles of present specimens of uedaites are 163.12(7)° for KG and 163.39(7)° for HJ (Table S2), while those of the clinozoisite-epidote series are bent in the opposite direction and ranges from 181 to 185° (Table S3) in a direction corresponding to the interior angle of uedaite. So, the decrease in the β angle associated with an increase in the occupancy of Mn²⁺ at the A1 site noted by Nagashima et al. (2015), particularly that in androsite, does not reflect an increase in the occupancy of Mn²⁺ at the A1 site but rather results from Mn²⁺ substitution at the M3 site.

Figure 4. Variation of the O4-M3-O8 angle (°) versus cell-dimension β (°). The data is summarized in Tables S2 and S3. Symbols without error bars represent standard deviations (1σ) smaller than the symbol size. Data sources not represented in the caption of Figure 3 are as follows: Rouse and Peacor (1993), Lavina et al. (2006), and Hoshino et al. (2008) for dissakisite, Dollase (1968, 1971), Gabe et al. (1973), Stergiou et al. (1987), Kvick et al. (1988), Comodi and Zanazzi (1997), Nagashima et al. (2007), and Gatta et al. (2010) for clinozoisite (Czo)-epidote (Ep) series, Giuli et al. (1999) for synthetic Czo-Ep series, Miyajima et al. (2003) for niigataite, Dollase (1971) for hancokite, Dollase (1969), Catti et al. (1988, 1989), Ferraris et al. (1989), Bonazzi et al. (1992), Bermanec et al. (1994), Bonazzi et al. (1996), Nagashima et al. (2010) for Czo-piemontite (Pm) series, Bonazzi et al. (1990) for piemontite-(Sr), Armbruster et al. (2002) for tweddillite, Pieczka et al. (2024) for heflikite, Nagashima et al. (2019) for Ca₂Al₂(V, Al)Si₃O₁₂(OH), and Nagashima et al. (2024) for Ca₂Al₂(Sc, Al)Si₃O₁₂(OH).

This perspective remains valid when considering uedaite, which contains the predominant amount of Mn²⁺ at the A1 site and REE at the A2 site. Based on previous interpretations, one might expect the β angle of uedaite to be as small as that of androsite. However, in reality, the β angle of uedaite is not that small. It suggests that the β angle is controlled by a simple topological variation rather than by cation substitutions at specific sites.

Because structural variations in the epidote supergroup minerals have been studied primarily within individual series, such as the clinozoisite-epidote series, the clinozoisite-piemontite series, or the allanite group, there are very few indicators of structural variation across the entire epidote supergroup. Therefore, the structural variation, applicable to the entire epidote supergroup, is significant because it can enhance understanding of the relationships among the various series and groups within this supergroup.

ACKNOWLEDGMENTS

We thank Mr. S. Harada and Dr. R. Kristiansen for providing the studied uedaite specimens, and Mr. Y. Morifuku (Center for Instrumental Analysis, Yamaguchi University) for his technical assistance. We also thank the editor, Dr. M. Hamada, Dr. R. Miyawaki and an anonymous reviewer for their comments. This study results from the utilization of research equipment shared in the MEXT Project for Promoting Public Utilization of Advanced Research Infrastructure (Program for supporting the construction of core facilities), Grant Nos. JPMXS0440400024 and JPMXS0440400025, and it is also supported by the Core Clusters for Research Initiative of Yamaguchi University. One of the authors, M.N., gratefully acknowledges financial support from the Grants-in-Aid for Scientific Research from the Japan Society for the Promotion of Science (No. JP23K03551).

SUPPLEMENTARY MATERIALS

Supplementary Tables S1-S3 and CIF files are available online from https://doi.org/10.2465/jmps.260125.

REFERENCES

• Agilent (2014) CrysAlis PRO. Agilent Technologies Ltd, Yarnton, Oxfordshire, England.
•  Armbruster,  T.,  Gnos,  E.,  Dixon,  R.,  Gutzmer,  J., et al (2002) Manganvesuvianite and tweddillite, two new Mn³⁺-silicate minerals from the Kalahari manganese fields, South Africa. Mineralogical Magazine, 66, 137-150.
•  Armbruster,  T.,  Bonazzi,  P.,  Akasaka,  M.,  Bermanec,  V., et al (2006) Recommended nomenclature of epidote-group minerals. European Journal of Mineralogy, 18, 551-567.
•  Baur,  H. (1974) The geometry of polyhedral distortions. Predictive relationships for the phosphate group. Acta Crystallographica, B30, 1195-1215.
•  Bergstøl,  S. and  Juve,  G. (1988) Scandian ixiolite, pyrochlore and bazzite in granite pegmatite in terdal, Telemark, Norway. A contribution to the mineralogy and geochemistry of scandium and tin. Mineralogy and Petrology, 38, 229-243.
•  Bermanec,  V.,  Armbruster,  T.,  Oberhänsli,  R. and  Zebec,  V. (1994) Crystal chemistry of Pb- and REE-rich piemontite from Nezilovo, Macedonia. Schweizerische Mineralogische und Petrographische Mitteilungen, 74, 321-328.
•  Biagioni,  C.,  Bonazzi,  P.,  Pasero,  M.,  Zaccarini,  F., et al (2019) Manganiakasakaite-(La) and ferriakasakaite-(Ce), two new epidote supergroup minerals from Piedmont, Italy. Minerals, 9, 353.
•  Bonazzi,  P.,  Menchetti,  S. and  Palenzona,  A. (1990) Strontiopiemontite, a new member of the epidote group from Val Graveglia, Liguria, Italy. European Journal of Mineralogy, 2, 519-523.
•  Bonazzi,  P.,  Garbarino,  C. and  Menchetti,  S. (1992) Crystal chemistry of piemontites: REE-bearing piemontite from Monte Brugiana, Alpi Apuane, Italy. European Journal of Mineralogy, 4, 23-33.
•  Bonazzi,  P.,  Menchetti,  S. and  Reinecke,  T. (1996) Solid solution between piemontite and androsite-(La), a new mineral of the epidote group from Andros Island, Greece. American Mineralogist, 81, 735-742.
•  Brown,  I.D. and  Altermatt,  D. (1985) Bond-valence parameters obtained from a systematic analysis of the inorganic crystal structure database. Acta Crystallographica, B41, 244-247.
•  Catti,  M.,  Ferraris,  G. and  Ivaldi,  G. (1988) Thermal behavior of the crystal structure of strontian piemontite. American Mineralogist, 73, 1370-1376.
•  Catti,  M.,  Ferraris,  G. and  Ivaldi,  G. (1989) On the crystal chemistry of strontian piemontite with some remarks on the nomenclature of the epidote group. Neues Jahrbuch für Mineralogie. Monatshefte, H.8, 357-366.
•  Cenki-Tok,  B.,  Ragu,  A.,  Armbruster,  T.,  Chopin,  S. and  Medenbach,  O. (2006) New Mn- and rare-earth-rich epidote-group minerals in metacherts: manganiandrosite-(Ce) and vanadoandrosite-(Ce). European Journal of Mineralogy, 18, 569-582.
•  Comodi,  P. and  Zanazzi,  P.F. (1997) The pressure behavior of clinozoisite and zoisite: An X-ray diffraction study. American Mineralogist, 82, 61-68.
•  Dollase,  W.A. (1968) Refinement and comparison of the structures of zoisite and clinozoisite. American Mineralogist, 53, 1882-1898.
•  Dollase,  W.A. (1969) Crystal structure and cation ordering of piemontite. American Mineralogist, 54, 710-717.
•  Dollase,  W.A. (1971) Refinement of the crystal structures of epidote, allanite and hancockite. American Mineralogist, 56, 447-464.
•  Ferraris,  G.,  Ivaldi,  G.,  Fuess,  H. and  Gregson,  D. (1989) Manganese/iron distribution in a strontian piemontite by neutron diffraction. Zeitschrift für Kristallographie, 187, 145-151.
• Franks, F., Ed. (1973) Water: A comprehensive treatise, vol. 2, pp. 684, Plenum, New York.
•  Gabe,  E.J.,  Portheine,  J.C. and  Whitlow,  S.H. (1973) A reinvestigation of the epidote structure: Confirmation of the iron location. American Mineralogist, 58, 218-223.
•  Gagné,  O.C. and  Hawthorne,  F.C. (2015) Comprehensive derivation of bond-valence parameters for ion pairs involving oxygen. Acta Crystallographica, B71, 562-578.
•  Gatta,  G.D.,  Meven,  M. and  Bromeley,  G. (2010) Effects of temperature on the crystal structure of epidote: a neutron single-crystals diffraction study at 293 and 1,070K. Physics and Chemistry of Minerals, 37, 475-485.
• Gieré, R. and Sorensen, S.S. (2004) Allanite and other REE-rich epidote-group minerals. In Reviews in Mineralogy and Geochemistry, Vol. 56 Epidotes, Mineralogical Society of America, Washington, 431-493. https://doi.org/10.2138/gsrmg.56.1.431
•  Giuli,  G.,  Bonazzi,  P. and  Menchetti,  S. (1999) Al-Fe disorder in synthetic epidotes: An single-crystal X-ray diffraction study. American Mineralogist, 84, 933-936.
• Harada, S., Nishio-Hamane, D. and Matsubara, S. (2020) Uedaite-(Ce)·Hydroxylbastnäsite-(Ce) and Thalénite like mineral from Karasugawa mine, Fukushima Prefecture, Japan. Abstracts for Annual Meeting of Japan Association of Mineralogical Sciences, R1-08 (in Japanese).
•  Hasegawa,  S. (1957) Chemical studies of allanites and their associated minerals from the pegmatites in the northern part of the Abukuma Massif. Science Report, University of Tohoku, Third series, 5, 345-371.
•  Hasegawa,  S. (1960) Chemical composition of allanite. Science Report, University of Tohoku, Third series, 6, 331-387.
•  Hoshino,  M.,  Kimata,  M.,  Nishida,  N.,  Kyono,  A., et al (2005) The chemistry of allanite from the Daibosatsu Pass, Yamanashi, Japan. Mineralogical Magazine, 69, 403-423.
•  Hoshino,  M.,  Kimata,  M.,  Nishida,  N. and  Shimizu,  M. (2008) Crystal chemical significance of chemical zoning in dissakisite-(Ce). Physics and Chemistry of Minerals, 35, 59-70.
•  Hoshino,  M.,  Kimata,  M.,  Chesner,  C.A.,  Nishida,  N., et al (2010) Crystal chemistry of volcanic allanites indicative of naturally induced oxidation-dehydrogenation. Mineralogy and petrology, 99, 133-141.
•  Ishihara,  S. and  Chappell,  B.W. (2008) Chemical compositions of the late Cretaceous granitoids across the central part of the Abukuma Highland, Japan-Revisited. Bulletin of the Geological Society of Japan, 59, 151-170.
•  Ito,  T.,  Morimoto,  N. and  Sadanaga,  R. (1954) On the structure of epidote. Acta Crystallographica, 7, 53-59.
•  Kartashov,  P.M.,  Ferraris,  G.,  Ivaldi,  G.,  Sokolova,  E. and  McCammon,  C.A. (2002) Ferriallanite-(Ce), CaCeFe³⁺AlFe²⁺(SiO₄)(Si₂O₇)O(OH), a new member of the epidote group: description, X-ray and Mössbauer study. The Canadian Mineralogist, 40, 1641-1648.
• Kato, A. (1973) Sakurai Mineral Collection (Sakurai Kobutsu Hyohon). The Celebration Meeting of Dr. Sakurai’s 60^th Birthday, pp. 177 (in Japanese).
•  Kolitsch,  U.,  Mills,  S.J.,  Miyawaki,  R. and  Blass,  G. (2012) Ferriallanite-(La), a new member of the epidote supergroup from the Eifel, Germany. European Journal of Mineralogy, 24, 741-747.
• Kristiansen, R. (2009) A unique assemblage of scandium-bearing minerals from the Heftetjern-pegmatite, Tørdal, south Norway. Kongsberg Mineralsymposium 2009 Norsk Bergverksmuseum Skrift 2009, 41, 75-104.
•  Kristiansen,  R. (2017) Fire nye mineraler for Norge. Stein, 44, 24-27 (in Norwegian).
• Kristiansen, R. (2018) Nye mineralfunn fra Heftetjern-pegmatitten i Tørdal. Norsk Mineralsymposium 2018, 79-92 (in Norwegian with English abstract).
• Kubo, K., Yanagisawa, Y., Yamamoto, T., Komazawa, M., et al. (2003) Geological Map of Japan 1:200,000, Fukushima. Geological Survey of Japan, AIST.
•  Kvick,  A.,  Pluth,  J.J.,  Richardson,  J.W. and  Smith,  J.V. (1988) The ferric ion distribution and hydrogen bonding in epidote: a neutron diffraction study at 15 K. Acta Crystallographica, B44, 351-355.
•  Lavina,  B.,  Carbonin,  S.,  Russo,  U. and  Tumiati,  S. (2006) The crystal structure of dissakisite-(La) and structural variations after annealing of radiation damage. American Mineralogist, 91, 104-110.
•  Miyajima,  H.,  Matsubara,  S.,  Miyawaki,  R. and  Hirokawa,  K. (2003) Niigataite, CaSrAl₃(Si₂O₇)(SiO₄)O(OH): Sr-analogue of clinozoisite, a new member of the epidote from the Itoigawa-Ohmi district, Niigata Prefecture, central Japan. Journal of Mineralogical and Petrological Sciences, 98, 118-129.
•  Miyawaki,  R.,  Yokoyama,  K.,  Matsubara,  S.,  Tsutsumi,  Y. and  Goto,  A. (2008) Uedaite-(Ce), a new member of the epidote group with Mn at the A site, from Shodoshima, Kagawa Prefecture, Japan. European Journal of Mineralogy, 20, 261-269.
•  Momma,  K. and  Izumi,  F. (2011) VESTA3 for three-dimensional visualization of crystal, volumetric and morphology data. Journal of Applied Crystallography, 44, 1272-1276.
•  Nagashima,  M.,  Akasaka,  M.,  Kyono,  A.,  Makino,  K. and  Ikeda,  K. (2007) Distribution of chromium among the octahedral sites in chromian epidote from Iratsu, central Shikoku, Japan. Journal of Mineralogical and Petrological Sciences, 102, 240-254.
•  Nagashima,  M.,  Armbruster,  T.,  Akasaka,  M. and  Minakawa,  T. (2010) Crystal chemistry of Mn²⁺-, Sr-rich and REE-bearing piemontite from the Kamisugai mine in the Sambagawa metamorphic belt, Shikoku, Japan. Journal of Mineralogical and Petrological Sciences, 105, 142-150.
•  Nagashima,  M.,  Imaoka,  T. and  Nakashima,  K. (2011) Crystal chemistry of Ti-rich ferriallanite-(Ce) from Cape Ashizuri, Shikoku Island, Japan. American Mineralogist, 96, 1870-1877.
•  Nagashima,  M.,  Nishio-Hamane,  D.,  Tomita,  N.,  Minakawa,  T. and  Inaba,  S. (2013) Vanadoallanite-(La): a new epidote-supergroup mineral from Ise, Mie Prefecture, Japan. Mineralogical Magazine, 77, 2739-2752.
•  Nagashima,  M.,  Nishio-Hamane,  D.,  Tomita,  N.,  Minakawa,  T. and  Inaba,  S. (2015) Ferriakasakaite-(La) and ferriandrosite-(La): new epidote-supergroup minerals from Ise, Mie Prefecture, Japan. Mineralogical Magazine, 79, 735-753.
•  Nagashima,  M.,  Nishio-Hamane,  D.,  Nakano,  N. and  Kawasaki,  T. (2019) Synthesis and crystal-chemistry of mukhinite, V-analogue of clinozoisite on the join Ca₂Al₃Si₃O₁₂(OH)-Ca₂Al₂VSi₃O₁₂(OH). Physics and Chemistry of Minerals, 46, 63-76.
•  Nagashima,  M.,  Morifuku,  Y. and  Mihailova,  B. (2024) Distribution of Sc³⁺ at the octahedral sites and its effect on the crystal structure of synthetic Sc-bearing clinozoisite on the Ca₂Al₃Si₃O₁₂(OH)-Ca₂Al₂ScSi₃O₁₂(OH) join. Physics and Chemistry of Minerals, 51, 17.
•  Nagashima,  M.,  Nishio-Hamane,  D.,  Ohnishi,  M.,  Miyajima,  H. and  Harada,  A. (2026) New epidote-supergroup minerals from the Mogurazawa mine, Gunma, Japan: akasakaite-(Ce), akasakaite-(La), vanadoakasakaite-(Ce), and vanadoakasakaite-(La). Journal of Mineralogical and Petrological Sciences, 121, 002.
•  Orlandi,  P. and  Pasero,  M. (2006) Allanite-(La) from Buca della Vena mine, Apuan Alps, Italy, an epidote-group mineral. The Canadian Mineralogist, 44, 63-68.
•  Pautov,  L.A.,  Khorov,  P.V.,  Ignatenko,  K.I.,  Sokolova,  E.V. and  Nadezhina,  T.N. (1993) Khristovite-(Ce)-(Ca, REE)REE(Mg, Fe)AlMnSi₃O₁₁(OH)(F, O): A new mineral in the epidote group. Proceedings of the Russian Mineralogical Society, 122, 103-111 (in Russian) (American Mineralogist, 1995, 80, 405).
•  Peacor,  D.R. and  Dunn,  P.J. (1988) Dollaseite-(Ce) (magnesium orthite redefined): Structure refinement and implications from F + M²⁺ substitutions in epidote-group minerals. American Mineralogist, 73, 838-842.
•  Pieczka,  A.,  Kristiansen,  R.,  Stachowicz,  M.,  Dumańska-Słowik,  M., et al (2024) Heflikite, ideally Ca₂(Al₂Sc)(Si₂O₇)(SiO₄)O(OH), the first scandium epidote-supergroup mineral from Jordanów Śląski, Lower Silesia (Poland) and from Heftetjern, Tørdal (Norway). Mineralogical Magazine, 88, 228-243.
•  Pieczka,  A.,  Gołębiowska,  B.,  Włodek,  A.,  Stachowicz,  M., et al (2025) Allanite-(Sm), CaSm(Al₂Fe²⁺)(Si₂O₇)(SiO₄)O(OH), the third samarium mineral from Jordanów Śląski, Lower Silesia, Poland. American Mineralogist, 110, 808-819.
•  Raade,  G. and  Kristiansen,  R. (2000) Mineralogy and geochemistry of the Heftetjern granite pegmatite, Tørdal: a progress report. Norsk Bergverksmuseum Skrift, 17, 19-25.
•  Ragu,  A.,  Bindi,  L.,  Bonazzi,  P.,  Remusat,  L. and  Chopin,  C. (2025) Vielleaureite-(Ce), a new epidote of the dollaseite group, coexisting with Mn-dominant hellandite in Mn deposits of the central Pyrenees. European Journal of Mineralogy, 37, 627-638.
•  Robinson,  K.,  Gibbs,  G.V. and  Ribbe,  P.H. (1971) Quadratic elongation: a quantitative measure of distortion in coordination polyhedra. Science, 172, 567-570.
•  Rouse,  R. and  Peacor,  D.R. (1993) The crystal structure of dissakisite-(Ce), the Mg analogue of allanite-(Ce). The Canadian Mineralogist, 31, 153-157.
•  Sakurai,  K.,  Wakita,  H.,  Kato,  A. and  Nagashima,  K. (1969) Chemical studies of minerals containing rarer elements from the Far East. LXIII. Bastnaesite from Karasugawa, Fukushima Prefecture, Japan. Bulletin of the Chemical Society of Japan, 42, 2725-2728.
•  Segalstad,  T.V. and  Eggleston,  T.L. (1993) Pegmatittene i Tørdal, Telemark. Stein, 20, 190-195 (in Norwegian).
•  Sheldrick,  G.M. (2015) Crystal structure refinement with SHELX. Acta Crystallographica, C71, 3-8.
•  Škoda,  R.,  Cempírek,  J.,  Filip,  J.,  Novák,  M., et al (2012) Allanite-(Nd), CaNdAl₂Fe²⁺(SiO₄)(Si₂O₇)O(OH), a new mineral from Åskagen, Sweden. American Mineralogist, 97, 983-988.
•  Sokolova,  E.V.,  Nadezhina,  T.N. and  Pautov,  L.A. (1991) Crystal structure of a new natural silicate of manganese from the epidote group. Soviet Physics Crystallography, 36, 172-174.
•  Steffenssen,  G.,  Müller,  A.,  Munnik,  F.,  Friis,  H., et al (2020) Unusual scandium enrichments of the Tørdal pegmatites, south Norway. Part I: Garnet as Sc exploration pathfinder. Ore Geology Reviews, 126, 103729.
•  Stergiou,  A.C.,  Rentzeperis,  P.J. and  Sklavounos,  S. (1987) Refinement of the crystal structure of a medium iron epidote. Zeitschrift Kristallographie, 178, 297-305.
•  Števko,  M.,  Myšľan,  P.,  Biagioni,  C.,  Mauro,  D. and  Mikuš,  T. (2023) Ferriandrosite-(Ce), a new member of the epidote supergroup from Betliar, Slovakia. Mineralogical Magazine, 87, 887-895.
•  Wakita,  H. and  Nagashima,  K. (1968) Rare Earth elements in xenotimes from Kawabe and Karasugawa Mine, Fukushima Pref., Japan. Journal of the Mineralogical Society of Japan, 9, 92-98 (in Japanese).