New data on S-bearing katoite from Tadano, Fukushima Prefecture, Japan: an implication of the presence of both (SO₃)²⁻ and (SO₄)²⁻ in the garnet structure

Abstract

Katoite, ideally denoted as {Ca₃}[Al₂](□_xSi_3−x)(OH)_4xO_12−4x, where 1.5 < x ≤ 3, was found in skarn xenoliths from Tadano, Fukushima Prefecture, Japan, and investigated by electron microprobe analyses, X-ray single crystal structure refinement, and infrared (IR) spectroscopy. Katoite characteristically contains up to 0.18 atoms per formula unit (apfu) of S and has a range of compositions, wherein Si = 0.96-1.25 apfu, Al = 1.47-1.74 apfu, Fe = 0.13-0.28 apfu, and Mg = 0.01-0.10 apfu. A difference Fourier map revealed the residual electron density near the octahedral Y site (Wyckoff position = 16a), and we interpreted that S with three-fold coordination occurs at the position of the residual electron density, which is represented as the Y’ site (Wyckoff position = 32e). The final refinement introduced the Y’ site to yield R1 [F² > 2σ(F²)] = 0.0353 with space group Ia3d and unit cell parameter a = 12.24095(8) Å. The IR spectrum in the range of 800-1200 cm⁻¹ shows a band at 1124 cm⁻¹ due to ν₃(SO₄) and doublet bands at 879 and 931 cm⁻¹ with different absorbances interpreted as ν₃(SiO₄) overlapping with ν₃(SO₃). The combined results of IR spectroscopy and structure refinement imply that in the Tadano katoite, S⁴⁺ and S⁶⁺, forming (SO₃)²⁻ and (SO₄)²⁻ coordinations, are placed at the Y’ and tetrahedral Z sites, respectively. Assuming that S is preferentially allotted to the Y’ site as S⁴⁺ to compensate for the deficiency of the octahedral cations Al, Fe, Mg, Mn, and Ti at the Y site, chemical formulae, including possible S⁴⁺ and S⁶⁺ contents, may be calculated. The average chemical formula for 14 different spots is as follows: {Ca₃}[(Al_1.614Fe³⁺_0.208Mg_0.063Ti_0.019)_Σ1.90S⁴⁺_0.096]_Σ2.00(□_1.867Si_1.120S⁶⁺_0.013)_Σ3.00[(OH)_7.192O_4.612F_0.147Cl_0.048]_Σ12.00.

INTRODUCTION

Katoite was first identified and described as an isomorphous series, Ca₃Al₂(SiO₄)₃-Ca₃Al₂(OH)₁₂ (the “hydrogrossular series”), and its ideal formula was initially defined as Ca₃Al₂(SiO₄)(OH)₈ (Passaglia and Rinaldi, 1984). Recently, the International Mineralogical Association Commission on New Minerals, Nomenclature and Classification approved a nomenclature for the garnet supergroup, and katoite was defined as a member of the “kenogarnet group” in the garnet supergroup (Grew et al., 2013). Cubic garnets have only four symmetrically unique atomic sites: dodecahedral {X}, octahedral [Y], and tetrahedral (Z) cation sites, as well as an anionic site designated as φ to indicate O, OH, and F, thereby providing a generalized chemical formula for the garnet supergroup {X₃}[Y₂](Z₃)φ₁₂ (Grew et al., 2013). Additionally, the three cation sites are at special positions with Wyckoff positions of 24c, 16a, and 24d, respectively (Grew et al., 2013). Grew et al. (2013) redefined the ideal formula for katoite as {Ca₃}[Al₂](□₃)(OH)₁₂, where □ is a vacancy. Therefore, the formula for the “hydrogrossular series” can be expressed as {Ca₃}[Al₂](Si_3−x□_x)O_12−4x(OH)_4x, where 0 (grossular endmember) ≤ x ≤ 3 (katoite endmember) relates to the substitution ^Z□ + 4OH⁻ = ^ZSi⁴⁺ + 4O²⁻ (Grew et al., 2013). The crystallographic and physical properties of katoite have been extensively investigated in recent years (Antao, 2015; Lacivita et al., 2015; Geiger and Rossman, 2018; Kato and Kyono, 2019; Kyono and Arora, 2019; Kyono et al., 2019; Geiger and Rossman, 2020).

Katoite and hydrous grossular rarely contain significant amounts of sulfur [0.11 S atoms per formula unit (apfu), Passaglia and Rinaldi, 1984; 0.17 S apfu, Rinaldi and Passaglia, 1989; 0.15 S apfu, Galuskin et al., 2009], and all the sulfur in the hydroxyl garnet was presumed to occur in the form of (SO₄)²⁻ and be located at the Z site as S⁶⁺. This assumption was made using only the electron microprobe data. Therefore, S-bearing hydroxyl garnets should be investigated using crystallographic and spectroscopic methods to confirm this assumption. We found S-bearing katoites in the skarn xenoliths in Tadano, Japan. As it contains a large amount of S (0.18 apfu), the katoite from Tadano is one of the most suitable samples for investigating the crystal chemistry of sulfur in the garnet structure. Here, we present infrared (IR) spectroscopic and X-ray structural data for Tadano katoite and discuss the valence state and site assignment for S in this mineral.

OCCURRENCE

Katoite was found in skarn xenolith specimens within the Tertiary andesite in a quarry in Tadano, Fukushima Prefecture, Japan (37°23′38′′N, 140°15′21′′E). The skarn xenoliths mainly comprised calcite, katoite, wadalite, andradite, and thaumasite. Katoite partially replaced wadalite crystals with sizes of up to 1 mm along its margins and fractures. Moreover, unidentified hydrous minerals containing Ca, Fe, Mg, Al, and Si often replace the wadalite. In many cases, katoite with or without hydrous minerals completely replaces wadalite crystals and occurs as pseudomorphs after wadalite (Banno et al., 2018). Katoite contains magnetite as the inclusion material.

The backscattered electron image of a katoite crystal shows that it is chemically heterogeneous, with bright and dark areas, and the bright area is richer in S than the dark area (Fig. 1). The bright core was surrounded by a dark rim. An additional dark part cuts the bright host to form veins. This texture indicates that the bright parts were formed earlier than the dark parts.

Figure 1. Backscattered electron image of katoite with relic wadalite. Numbers indicate the content of S (apfu) at the point. Ka, katoite; Wa, wadalite; Mag, magnetite; Adr, andradite; and Cal, calcite.

INFRARED (IR) SPECTROSCOPY

The samples for IR spectroscopy were prepared as follows: first, a single katoite fragment (100 µm × 50 µm × 30 µm) extracted from a thin section was crushed between two diamond plates. Subsequently, a fine-grained portion of the crushed fragments was pressed using a hand press machine. They were then placed on a diamond plate and analyzed. The micro-IR spectrum of katoite was recorded using a JASCO FT/IR-4700 with IRT-5200, employing a transmission method that collected 100 scans over a range of 600-4000 cm⁻¹ with a resolution of 4 cm⁻¹, as shown in Figure 2a. The aperture size was 34 × 109 µm. The IR spectrum was assigned by referring to the data reported by Kobayashi and Shoji (1983, 1984), Nyquist and Kagel (1997), Nyquist et al. (1997a, b), and Makreski et al. (2011). The spectrum of katoite showed absorption bands at 3656 cm⁻¹ attributed to O-H stretching vibrations, and at 1124 cm⁻¹ attributed to the vibration of the (SO₄)²⁻ group [the ν₃(SO₄) mode]. The absorption bands at 931 and 879 cm⁻¹ were tentatively assigned to the ν₃(SiO₄) mode, but another possible assignment for these bands is discussed later.

Figure 2. Infrared spectrum of (a) S-bearing katoite (this study) compared to those of (b)-(d) synthetic S-free hydroxyl garnets.

MINERAL CHEMISTRY

The chemical composition of katoite was determined using a JEOL JXA-8900R electron microprobe analyzer installed at the Geological Survey of Japan (GSJ) laboratory, Advanced Industrial Science and Technology (AIST), Tsukuba, Japan. Accelerating voltage, beam current, and beam diameter were 15 kV, 12 nA, and 3 µm, respectively. Synthetic quartz (for Si), rutile (Ti), corundum (Al), MnO (Mn), hematite (Fe), periclase (Mg), wollastonite (Ca), celestine (S), F-phlogopite (F), and natural Cl-rich hastingsite (Cl) were used as the standard materials. Matrix correction was conducted using the ZAF method. The detection limits for Cl and F were ∼0.02 and ∼0.05 wt% (2σ level), respectively. Na was not detected. Tables 1-3 list the chemical composition of katoite. All the analytical positions were carefully selected from the backscattered electron images to avoid contamination by adjacent minerals.

CHEMICAL COMPOSITION

First, the chemical formulae of katoite were normalized by assuming that the sum of the anions, O + OH + F + Cl, was equal to 12.00 apfu, wherein the H₂O content was calculated based on the stoichiometric assumption of Si + S + (OH + F + Cl)/4 = 3.00 apfu. This equation implies that (1) the Z cations are Si and S; (2) values of (OH + F + Cl) are controlled by the coupled substitution of ^Z□ + 4(OH⁻, F⁻, Cl⁻) = ^ZSi⁴⁺ + 4O²⁻; and (3) the sum of the Z cations and ^Z□ is equal to the ideal value of 3.00 apfu at the Z site. Vacancies were consequently introduced at the Z site. Table 1 presents these results. In many cases, Ca content was higher than the ideal value of 3.00 apfu at the X site (Table 1). The highest Ca content and analytical uncertainty (2σ) were 3.077 apfu (Table 1, No. 28) and ±0.025 apfu, respectively. As Ca resides exclusively at the X site (Grew et al., 2013), the chemical formula normalized by this method was not considered to be appropriate for katoite. Second, the chemical formula of katoite was normalized based on the assumption that the sum of the X and Y cations, Ca + Al + Fe + Mg + Mn + Ti, was equal to 5.00 apfu. H₂O was calculated based on both the charge balance and the assumption that the sum of the anions was equal to 12.00 apfu. Table 2 presents these results. Subsequently, the Ca content was higher than the ideal value of 3.00 apfu at the X site (Table 2). In our study, katoite with higher S content (SO₃ = 1.48-3.47 wt%) tends to have higher Ca content than that with lower S content (SO₃ = 0.09-0.13 wt%) (Table 2). The highest Ca content and analytical uncertainty (2σ) were 3.119 apfu (Table 2, No. 28) and ±0.025 apfu, respectively. The chemical formula normalized using this method is also unlikely to be appropriate for katoite. Third, the chemical formulae of katoite were normalized based on the assumption that Ca content was equal to 3.00 apfu, implying that the X site was fully occupied by Ca. H₂O was calculated based on both the charge balance and the assumption that the sum of the anions was equal to 12.00 apfu with all iron and sulfur ions being treated as Fe³⁺ and S⁶⁺, respectively. Consequently, vacancies were introduced at the Y site. Table 3 presents these results. Another possible normalization method is based on the assumption of (Ca + Fe²⁺ + Mg + Mn) = 3.00 apfu, implying that the X site is fully occupied by Ca, Fe²⁺, Mg, and Mn. However, this normalization method leads to higher amounts of vacancies introduced at the Y site compared to the assumption of Ca = 3.00 apfu. Significant number of vacancies at the Y site or significant garnet coupled heterovalent substitutions involving vacancies at the Y site have not been reported from natural garnets (Grew et al., 2013; Deer et al., 1997). From the perspective of both the stoichiometric limit of Ca at the X site (Ca ≤ 3.00 apfu) and the minimum amounts of vacancy at the Y site in the garnet structure, we considered that the normalization method based on the assumption of Ca = 3.00 apfu leads to the most appropriate chemical formula for the katoite in this study. Therefore, we adopted the data presented in Table 3 for the description and discussion of Tadano katoite. All sulfur was tentatively treated as S⁶⁺, but another possible valence state is discussed later. The normalization method based on the assumption that the X site is fully occupied by Ca was applied to the S-bearing katoite from the Northern Caucasus, Russia, as well (Galuskin et al., 2009).

Table 1. Chemical compositions of katoite with formulae normalized based on 12(O, OH, F, Cl) and H₂O contents calculated based on the assumption that the sum of Z cations and ^Z□ was equal to 3.00 apfu

Sample	1811	1901	1711	1711	1811	1711	1901	1901	1901	1901	1711	1811	1901a	1811
No.	49	29	41	40	18	32	4	41	47	62	53	64	76#	28
Weight percentage
SiO2	14.18	14.48	16.65	16.65	15.70	16.51	16.90	16.54	15.81	17.12	17.80	16.96	17.08	15.36
TiO2	0.55	0.36	0.29	0.22	0.44	0.30	0.52	0.40	0.45	0.39	0.23	0.39	0.30	0.29
Al2O3	21.62	21.57	21.34	21.62	21.14	19.93	19.79	19.61	20.51	18.70	18.41	18.20	17.83	18.21
Fe2O3	4.56	4.41	2.59	2.65	2.88	3.35	4.71	3.76	3.96	3.95	4.69	5.23	4.96	4.42
MnO	0.00	0.04	n.d.	0.00	n.d.	n.d.	n.d.	n.d.	n.d.	0.00	n.d.	n.d.	n.d.	n.d.
MgO	0.14	0.09	0.58	0.59	0.48	0.69	0.29	0.73	0.46	0.83	0.95	0.81	0.99	0.96
CaO	41.55	40.90	41.21	40.93	40.68	40.43	40.55	40.96	40.77	40.45	39.75	40.00	40.14	40.92
SO3	0.09	0.13	1.48	1.66	1.80	2.04	2.23	2.32	2.65	2.87	2.92	2.94	2.95	3.47
F	0.93	0.80	0.78	0.71	0.86	0.72	0.58	0.80	0.53	0.49	0.54	0.51	0.54	0.66
Cl	0.11	0.09	0.26	0.23	0.58	0.28	0.20	0.63	0.90	0.85	0.47	0.41	0.44	0.33
-O=F+Cl	0.42	0.36	0.39	0.35	0.49	0.37	0.29	0.48	0.43	0.40	0.33	0.31	0.33	0.35
H2O*	17.67	17.17	15.06	15.12	15.28	14.35	14.52	14.35	15.16	13.43	12.94	13.62	13.33	14.46
Total	100.98	99.68	99.85	100.03	99.35	98.23	100.00	99.62	100.77	98.68	98.37	98.76	98.23	98.73
Formulae based on the sum of anions: O + OH + F + Cl = 12.00 apfu
Ca	3.002	2.996	3.038	3.006	3.009	3.046	3.008	3.054	2.986	3.065	3.028	3.028	3.062	3.077
Al	1.718	1.738	1.730	1.747	1.720	1.652	1.615	1.608	1.653	1.559	1.543	1.515	1.496	1.506
Fe3+	0.231	0.227	0.134	0.137	0.150	0.177	0.245	0.197	0.204	0.210	0.251	0.278	0.266	0.233
Mg	0.014	0.009	0.059	0.060	0.049	0.072	0.030	0.076	0.047	0.088	0.101	0.085	0.105	0.100
Mn	0.000	0.002	n.d.	0.000	n.d.	n.d.	n.d.	n.d.	n.d.	0.000	n.d.	n.d.	n.d.	n.d.
Ti	0.028	0.019	0.015	0.011	0.023	0.016	0.027	0.021	0.023	0.021	0.012	0.021	0.016	0.015
Sum Y	1.991	1.995	1.939	1.955	1.942	1.917	1.918	1.902	1.926	1.877	1.907	1.900	1.883	1.856
S	0.005	0.007	0.076	0.085	0.093	0.108	0.116	0.121	0.136	0.152	0.156	0.156	0.158	0.183
Si	0.956	0.990	1.146	1.141	1.084	1.161	1.170	1.151	1.081	1.211	1.266	1.198	1.216	1.078
F	0.198	0.173	0.170	0.154	0.188	0.160	0.127	0.176	0.115	0.110	0.121	0.114	0.122	0.146
Cl	0.013	0.010	0.030	0.027	0.068	0.033	0.023	0.074	0.104	0.102	0.057	0.049	0.053	0.039
O	3.842	3.986	4.889	4.906	4.709	5.077	5.144	5.089	4.868	5.453	5.685	5.418	5.494	5.044
OH	7.947	7.831	6.911	6.913	7.035	6.730	6.706	6.661	6.913	6.336	6.137	6.419	6.331	6.770

^*H₂O was calculated based on both Si + S + (OH + F + Cl)/4 = 3.00 apfu and the assumption that the sum of the anions was equal to 12.00 apfu.

^#Used for single-crystal X-ray diffraction study.

n.d., not detected.

Table 2. Chemical compositions of katoite with formulae normalized based on the assumption that the sum of the X and Y cations, Ca + Al + Fe + Mg + Mn + Ti, was equal to 5.00 apfu

Sample	1811	1901	1711	1711	1811	1711	1901	1901	1901	1901	1711	1811	1901a	1811
No.	49	29	41	40	18	32	4	41	47	62	53	64	76#	28
Weight percentage
SiO2	14.18	14.48	16.65	16.65	15.70	16.51	16.90	16.54	15.81	17.12	17.80	16.96	17.08	15.36
TiO2	0.55	0.36	0.29	0.22	0.44	0.30	0.52	0.40	0.45	0.39	0.23	0.39	0.30	0.29
Al2O3	21.62	21.57	21.34	21.62	21.14	19.93	19.79	19.61	20.51	18.70	18.41	18.20	17.83	18.21
Fe2O3	4.56	4.41	2.59	2.65	2.88	3.35	4.71	3.76	3.96	3.95	4.69	5.23	4.96	4.42
MnO	0.00	0.04	n.d.	0.00	n.d.	n.d.	n.d.	n.d.	n.d.	0.00	n.d.	n.d.	n.d.	n.d.
MgO	0.14	0.09	0.58	0.59	0.48	0.69	0.29	0.73	0.46	0.83	0.95	0.81	0.99	0.96
CaO	41.55	40.90	41.21	40.93	40.68	40.43	40.55	40.96	40.77	40.45	39.75	40.00	40.14	40.92
SO3	0.09	0.13	1.48	1.66	1.80	2.04	2.23	2.32	2.65	2.87	2.92	2.94	2.95	3.47
F	0.93	0.80	0.78	0.71	0.86	0.72	0.58	0.80	0.53	0.49	0.54	0.51	0.54	0.66
Cl	0.11	0.09	0.26	0.23	0.58	0.28	0.20	0.63	0.90	0.85	0.47	0.41	0.44	0.33
-O=F+Cl	0.42	0.36	0.39	0.35	0.49	0.37	0.29	0.48	0.43	0.40	0.33	0.31	0.33	0.35
H2O*	17.59	17.08	14.82	14.71	14.76	13.97	13.75	13.89	14.24	12.85	12.28	12.88	12.78	13.77
Total	100.91	99.59	99.61	99.62	98.83	97.86	99.23	99.16	99.86	98.10	97.71	98.02	97.68	98.04
Formulae based on Ca + Al + Fe + Mg + Mn+ Ti = 5.00 apfu
Ca	3.006	3.001	3.052	3.030	3.039	3.069	3.054	3.081	3.039	3.101	3.068	3.072	3.096	3.119
Al	1.720	1.741	1.738	1.760	1.737	1.664	1.639	1.623	1.682	1.577	1.563	1.538	1.513	1.527
Fe3+	0.232	0.227	0.135	0.138	0.151	0.179	0.249	0.199	0.207	0.213	0.254	0.282	0.269	0.237
Mg	0.014	0.009	0.060	0.061	0.050	0.073	0.030	0.076	0.048	0.089	0.102	0.087	0.106	0.102
Mn	0.000	0.002	n.d.	0.000	n.d.	n.d.	n.d.	n.d.	n.d.	0.000	n.d.	n.d.	n.d.	n.d.
Ti	0.028	0.019	0.015	0.011	0.023	0.016	0.027	0.021	0.024	0.021	0.012	0.021	0.016	0.016
Sum Y	1.994	1.999	1.948	1.970	1.961	1.931	1.946	1.919	1.961	1.899	1.932	1.928	1.904	1.881
S	0.005	0.007	0.077	0.086	0.094	0.108	0.118	0.122	0.138	0.154	0.158	0.158	0.159	0.185
Si	0.957	0.992	1.151	1.150	1.095	1.170	1.188	1.161	1.100	1.225	1.282	1.216	1.230	1.093
F	0.199	0.173	0.170	0.155	0.190	0.161	0.129	0.178	0.117	0.111	0.123	0.116	0.123	0.148
Cl	0.013	0.010	0.030	0.027	0.069	0.034	0.024	0.075	0.106	0.103	0.057	0.050	0.054	0.040
O	3.865	4.013	4.967	5.039	4.878	5.203	5.401	5.242	5.167	5.656	5.919	5.675	5.689	5.277
OH	7.924	7.804	6.832	6.779	6.864	6.602	6.447	6.505	6.610	6.130	5.901	6.160	6.135	6.534

^*H₂O was calculated based on both the charge balance and the assumption that the sum of the anions was equal to 12.00 apfu.

^#Used for single-crystal X-ray diffraction study.

n.d., not detected.

Table 3. Chemical compositions of katoite with formulae normalized based on the assumption that the X site was fully occupied by Ca

Sample	1811	1901	1711	1711	1811	1711	1901	1901	1901	1901	1711	1811	1901a	1811
No.	49	29	41	40	18	32	4	41	47	62	53	64	76#	28
Weight percentage
SiO2	14.18	14.48	16.65	16.65	15.70	16.51	16.90	16.54	15.81	17.12	17.80	16.96	17.08	15.36
TiO2	0.55	0.36	0.29	0.22	0.44	0.30	0.52	0.40	0.45	0.39	0.23	0.39	0.30	0.29
Al2O3	21.62	21.57	21.34	21.62	21.14	19.93	19.79	19.61	20.51	18.70	18.41	18.20	17.83	18.21
Fe2O3	4.56	4.41	2.59	2.65	2.88	3.35	4.71	3.76	3.96	3.95	4.69	5.23	4.96	4.42
MnO	0.00	0.04	n.d.	0.00	n.d.	n.d.	n.d.	n.d.	n.d.	0.00	n.d.	n.d.	n.d.	n.d.
MgO	0.14	0.09	0.58	0.59	0.48	0.69	0.29	0.73	0.46	0.83	0.95	0.81	0.99	0.96
CaO	41.55	40.90	41.21	40.93	40.68	40.43	40.55	40.96	40.77	40.45	39.75	40.00	40.14	40.92
SO3*	0.09	0.13	1.48	1.66	1.80	2.04	2.23	2.32	2.65	2.87	2.92	2.94	2.95	3.47
F	0.93	0.80	0.78	0.71	0.86	0.72	0.58	0.80	0.53	0.49	0.54	0.51	0.54	0.66
Cl	0.11	0.09	0.26	0.23	0.58	0.28	0.20	0.63	0.90	0.85	0.47	0.41	0.44	0.33
-O=F+Cl	0.42	0.36	0.39	0.35	0.49	0.37	0.29	0.48	0.43	0.40	0.33	0.31	0.33	0.35
H2O**	17.70	17.11	15.72	15.23	15.43	15.13	14.67	15.28	14.92	14.54	13.42	14.09	14.38	15.78
Total	101.01	99.62	100.51	100.14	99.49	99.02	100.15	100.55	100.54	99.79	98.84	99.24	99.28	100.05
Formulae based on Ca = 3.00 apfu
Ca	3	3	3	3	3	3	3	3	3	3	3	3	3	3
Al	1.717	1.740	1.709	1.743	1.715	1.627	1.611	1.580	1.660	1.526	1.528	1.501	1.466	1.469
Fe3+	0.231	0.227	0.132	0.136	0.149	0.175	0.245	0.193	0.205	0.206	0.249	0.275	0.260	0.228
Mg	0.014	0.009	0.059	0.060	0.049	0.071	0.030	0.074	0.047	0.086	0.100	0.085	0.103	0.098
Mn	0.000	0.002	n.d.	0.000	n.d.	n.d.	n.d.	n.d.	n.d.	0.000	n.d.	n.d.	n.d.	n.d.
Ti	0.028	0.019	0.015	0.011	0.023	0.016	0.027	0.021	0.023	0.020	0.012	0.021	0.016	0.015
Sum Y	1.990	1.998	1.915	1.951	1.936	1.888	1.912	1.868	1.935	1.837	1.889	1.882	1.845	1.809
S*	0.005	0.007	0.075	0.085	0.093	0.106	0.116	0.119	0.137	0.149	0.154	0.154	0.154	0.178
Si	0.956	0.991	1.131	1.139	1.081	1.143	1.167	1.131	1.086	1.185	1.254	1.187	1.191	1.051
F	0.198	0.173	0.168	0.154	0.187	0.158	0.127	0.173	0.115	0.107	0.120	0.113	0.119	0.143
Cl	0.013	0.010	0.030	0.027	0.068	0.033	0.023	0.073	0.105	0.100	0.056	0.049	0.052	0.038
O	3.834	4.005	4.678	4.872	4.662	4.819	5.095	4.788	4.944	5.081	5.521	5.258	5.140	4.617
OH	7.955	7.811	7.124	6.948	7.083	6.991	6.755	6.966	6.836	6.712	6.303	6.581	6.689	7.202
S4+***	0.005	0.002	0.075	0.049	0.064	0.106	0.088	0.119	0.065	0.149	0.111	0.118	0.154	0.178
S6+***	0.000	0.004	0.000	0.036	0.029	0.000	0.028	0.000	0.072	0.000	0.043	0.036	0.000	0.000
O***	3.825	4.000	4.528	4.774	4.534	4.607	4.919	4.550	4.814	4.783	5.299	5.022	4.831	4.261
OH***	7.964	7.816	7.275	7.046	7.211	7.203	6.931	7.204	6.966	7.010	6.525	6.817	6.998	7.558
Weight percentage
SO2***	0.07	0.04	1.18	0.76	0.99	1.63	1.36	1.86	1.01	2.30	1.68	1.80	2.36	2.78
SO3***	0.00	0.08	0.00	0.71	0.56	0.00	0.53	0.00	1.39	0.00	0.82	0.70	0.00	0.00
H2O***	17.72	17.12	16.05	15.44	15.71	15.59	15.05	15.80	15.21	15.18	13.89	14.60	15.04	16.56
Total***	101.01	99.62	100.55	100.16	99.53	99.07	100.19	100.61	100.57	99.86	98.89	99.29	99.35	100.13

^*Total S as SO₃.

^**H₂O was calculated based on both the charge balance and the assumption that the sum of the anions was equal to 12.00 apfu.

^***Recalculated (see the text for the recalculation procedures).

^#Used for single-crystal X-ray diffraction study.

n.d., not detected.

The compositional range of Si is 0.96-1.25 apfu. Katoite has a range of compositions, wherein Al = 1.47-1.74 apfu, Fe = 0.13-0.28 apfu, Mg = 0.01-0.10 apfu, and S = 0.01-0.18 apfu, and contains appreciable amounts of F (0.11-0.20 apfu) and Cl (0.01-0.11 apfu). The compositional variations in S versus (Al + Fe + Mg + Mn + Ti) for the data listed in Table 3 are shown in Figure 3. A negative correlation was observed between these two parameters. The value of (Al + Fe + Mg + Mn + Ti) represents the sum of the cations that can occupy the Y site. This value tends to decrease from 2.00 to 1.81 apfu along the straight line of S + Al + Fe + Mg + Mn + Ti = 2.00 apfu with increasing S content (Fig. 3), implying that sulfur is incorporated into the Y site.

Figure 3. S versus (Al + Fe + Mg + Mn + Ti) diagram for katoite. Data are obtained from Table 3 in this study. Error bars in the Figure represent the 2σ analytical uncertainty.

X-RAY CRYSTALLOGRAPHY

A single-crystal fragment of katoite extracted from a thin section used for electron microprobe analysis was mounted on a glass fiber and analyzed by X-ray diffraction. Single-crystal X-ray studies were conducted at room temperature with monochromated MoKα radiation on a Rigaku Synergy-Custom diffractometer equipped with a HyPix-6000He detector housed at the National Museum of Nature and Science, Tsukuba, Japan. Data collection and reduction, including empirical absorption correction, were conducted using the CrysAlis Pro software. The crystal structure was solved by the charge flipping method using the Superflip software (Palatinus and Chapuis, 2007). After confirming the structure type, structure refinement was conducted using SHELXL (Sheldrick, 2015) using the atomic scattering factors of the neutral atoms. The first refinement was conducted based on a normal model containing X, Y, and Z sites in the garnet structure, yielding an R1 value [F² > 2σ(F²)] of 0.0380. The occupancies of the Y and Z sites were constrained to 1 and refined as Al versus Fe and Si versus □, respectively. Refinement of the site occupancies yielded [0.861(8)Al + 0.139(8)Fe] for the Y site and 0.403(7)Si for the Z site. The occupancy of Ca at the X site was fixed at 1. However, the position of the H atom could not be determined. The chemical formula of the fragment used for X-ray diffraction was {Ca₃}[Al_1.466Fe_0.260Mg_0.103Ti_0.016]_Σ1.85(□_1.655Si_1.191S_0.154)_Σ3.00[(OH)_6.689O_5.140F_0.119Cl_0.052]_Σ12.00, assuming that all S occurred at the Z site (Table 3, No. 76). Based on the aforementioned formula (model 1), the calculated Y-site-scattering value was found to be smaller than the refined value, whereas the calculated Z-site-scattering value is larger than the refined value (Table 4). Subsequently, we used the formula assuming that all S occurred at the Y site (model 2), and the resultant calculated site-scattering values were close to the refined values (Table 4). The residual electron density near the Y site (Wyckoff position = 16a) was detected on a difference Fourier map after the convergence of the first refinement (Fig. 4). Therefore, we interpreted that S occurred at the position of the residual electron density, which is represented as the Y’ site (Wyckoff position = 32e). We then conducted refinement to introduce the Y’ site, and this yielded a smaller R1 value [F² > 2σ(F²)] of 0.0353 than the first refinement (0.0380). Based on the chemical formula, the sum of the occupancies of Al and Fe at the Y site was constrained to 0.9225, and the occupancy of S at the Y’ site was fixed at 0.0385. The details of the sample, data collection, and structural refinement are presented in Table 5. The final atomic coordinates, cation occupancies, and displacement parameters are presented in Table 6. The final site populations and a comparison between the refined and calculated site-scattering values are presented in Table 7. The projection of the structure along the [001] direction is shown in Figure 5. The position of the Y’ site and the interatomic distance of Y’-O [1.60(2) Å] suggest that S at the Y’ site has three-fold coordination. We interpreted that S occurs as either (S₂O₆)²⁻ (two-sulfur model) or (SO₃)²⁻ (one-sulfur model) in an octahedron. In the latter case, only one of the two symmetrically equivalent positions is occupied by S because the multiplicity of the Y’ site is double that of the Y site.

Table 4. Site populations and a comparison between the refined and calculated site-scattering values for the first refinement

Site	Site population (apfu)	Site-scattering value (epfu*)
	Model 1	Refined	Calculated
Y	1.466Al + 0.260Fe3+ + 0.103Mg + 0.016Ti	29.6	27.4
Z	1.191Si + 0.154S	16.9	19.1
	Model 2	Refined	Calculated
Y	1.466Al + 0.260Fe3+ + 0.103Mg + 0.016Ti + 0.154S	29.6	29.9
Z	1.191Si	16.9	16.7

^*Electrons per formula unit.

Figure 4. Residual electron density near the Y site, suggesting that the S atom occurs at the positions (Wyckoff position = 32e) represented as Y’ site. Blue and red spheres indicate Y site and oxygen atoms, respectively. Isosurface level is 0.5 e/ų. Residual electron densities except the Y site were below this level.

Table 5. Experimental details and crystallographic data for the final refinement

a (Å)	12.24095(8)
V (Å3)	1834.20(4)
Space group	Ia3d
Z	8
Dcalc. (g/cm3)	3.013(5)
μ (mm−1)	2.635
Radiation	MoKα
Reflection range	−21 ≤ h ≤ 21
	−20 ≤ k ≤ 20
	−20 ≤ l ≤ 21
No. of measured reflections	51942
No. of unique reflections	425
No. of observed reflections [F2 > 2σ(F2)]	394
Rint	0.0248
Variable parameters	23
R1 [F2 > 2σ(F2)], R1 (all reflections)	0.0353, 0.0388
wR2 (all reflections)	0.0909
Weighting parameters, a, b	0.0261, 6.5262
Goodness of Fit	1.297
Final Δρmax (e/Å3)	0.368
Final Δρmin (e/Å3)	−0.370

R1 = Σ| |F_o| − |F_c| |/Σ|F_o|.

wR2 = {Σ[w(F_o² − F_c²)²]/Σ[w(F_o²)²]}^0.5.

w = 1/[σ²(F_o²) + (aP)² + bP].

P = (2F_c² + F_o²)/3.

Table 6. Final atomic coordinates, cation occupancies, and displacement parameters (Ų)

Site	W*	Occupancy	x	y	z	Ueq
X	24c	Ca	0.25000	0.12500	0	0.01666(15)
Y	16a	0.743(3)Al + 0.180(3)Fe	0	0	0	0.0181(2)
Y’	32e	0.03850S	0.036(5)	0.036(5)	0.036(5)	0.12(3)**
Z	24d	0.406(6)Si	0.37500	0	0.25000	0.0102(5)
O	96h	O	0.20036(13)	0.28347(13)	0.10372(12)	0.0278(3)
Site	U11	U22	U33	U23	U13	U12
X	0.01723(19)	0.0155(2)	0.01723(19)	0	0.00415(19)	0
Y	0.0181(2)	0.0181(2)	0.0181(2)	−0.0009(2)	−0.0009(2)	−0.0009(2)
Z	0.0102(8)	0.0101(6)	0.0101(6)	0	0	0
O	0.0240(6)	0.0333(8)	0.0260(7)	−0.0114(5)	−0.0062(5)	0.0072(5)

^*Wyckoff position.

^**Isotropic displacement parameter.

Table 7. Site populations and a comparison between the refined and calculated site-scattering values for the final refinement

Site	Site population (apfu)	Site-scattering value (epfu)
		Refined	Calculated
Y	1.466Al + 0.260Fe3+ + 0.103Mg + 0.016Ti	28.7	27.4
Y’	0.154S	2.5*	2.5
Sum of Y and Y’	1.466Al + 0.260Fe3+ + 0.103Mg + 0.016Ti + 0.154S	31.1	29.9
Z	1.191Si	17.1	16.7

^*Fixed value based on the electron microprobe analysis.

Figure 5. Crystal structure of katoite projected along [001] drawn with VESTA 3 (Momma and Izumi, 2011).

DISCUSSION

Based on the results of the aforementioned structural refinements, we examined whether the characteristic bands of (S₂O₆)²⁻ and (SO₃)²⁻ were present or absent in the IR spectrum by referring to the IR data of inorganic materials containing these anions (Ross, 1972; Nyquist and Kagel 1997; Nyquist et al., 1997a, b). As shown in Figure 2a, katoite showed no distinct absorption band at approximately 1200 cm⁻¹, which is one of the strong bands of (S₂O₆)²⁻, thereby indicating the absence of (S₂O₆)²⁻. Inorganic materials containing (SO₃)²⁻, such as Na₂SO₃, MgSO₃·6H₂O, and SrSO₃ (Nyquist and Kagel 1997; Nyquist et al., 1997a), have a very strong IR band of the ν₃(SO₃) mode within 911-980 cm⁻¹; however, this band appears within a range of 800-1000 cm⁻¹ for strong absorbance due to the antisymmetric stretching vibration of the SiO₄ tetrahedra [the ν₃(SiO₄) mode] in synthetic “hydrogrossular” (cf. Kobayashi and Shoji, 1983). Based on the IR spectra of synthetic “hydrogrossular” and “hydrograndite” observed by Kobayashi and Shoji (1983, 1984, 1987), doublet ν₃(SiO₄) bands near 850 and 910 cm⁻¹ become weaker and more diffuse with decreasing Si content, as indicated by Żabiński (1966) and Passaglia and Rinaldi (1984). The absorbance of the two ν₃(SiO₄) bands became approximately equivalent when the Si content was ∼2 apfu (Kobayashi and Shoji, 1984) (cf. Figs. 2b and 2c). Moreover, synthetic “hydroandradite” with Si = ∼1 apfu showed only one band near 900 cm⁻¹ (Cohen-Addad, 1970) (cf. Fig. 2d). As the Si content of katoite (Si = 0.96-1.25 apfu) is close to that of the synthetic “hydroandradite” (Si = 1.15 apfu), only one broad band was expected to be observed in the range of 800-1000 cm⁻¹. However, in this study, the two bands were characterized by a much stronger absorbance band at 931 cm⁻¹ than that at 879 cm⁻¹ (Fig. 2a). Hence, the spectral pattern for the doublet ν₃(SiO₄) bands of the Tadano kaoite seemingly differs from that of the synthetic S-free hydroxyl garnets (Figs. 2b-2d). Thus, we interpret that this is due to the overlap of the ν₃(SO₃) band with the single ν₃(SiO₄) band. The absorption band at 1124 cm⁻¹ was attributed to the ν₃(SO₄) mode. Therefore, the combined results of IR spectroscopy and structure refinement imply that in Tadano katoite, S⁴⁺ and S⁶⁺, forming (SO₃)²⁻ and (SO₄)²⁻ coordinations, are placed at the Y’ and Z sites, respectively. The structural refinement results suggest that S is preferentially allotted to the Y’ site as S⁴⁺ to compensate for the deficiency of the octahedral cations Al, Fe, Mg, Mn, and Ti at the Y site. The chemical formulae of katoite in Table 3 indicate an appreciable variation in the value of (S + Al + Fe + Mg + Mn +Ti), ranging between 1.99 and 2.07 apfu (cf. Fig. 3). The values of (S + Al + Fe + Mg + Mn +Ti) that are appreciably higher than 2.00 apfu indicate the presence of S⁶⁺ at the Z site because the ideal sum of the cations at the Y’ and Y sites is 2.00 apfu. The S⁴⁺ and S⁶⁺ contents estimated in Table 3 provided the maximum and minimum values, respectively. The numbers of OH groups and O atoms were recalculated based on the S⁴⁺ and S⁶⁺ contents. The SO₂, SO₃, and H₂O wt% corresponding to the S⁴⁺, S⁶⁺, and OH contents are presented at the bottom of Table 3. The chemical formula of the averaged compositions listed in Table 3 is as follows {Ca₃}[(Al_1.614Fe³⁺_0.208Mg_0.063Ti_0.019)_Σ1.90S⁴⁺_0.096]_Σ2.00(□_1.867Si_1.120S⁶⁺_0.013)_Σ3.00[(OH)_7.192O_4.612F_0.147Cl_0.048]_Σ12.00.

Passaglia and Rinaldi (1984) reported the IR spectrum of the holotype katoite with S = 0.11 apfu, wherein all sulfur atoms bearing a charge of +6 are located at the Z site. Within the range of 800-1200 cm⁻¹, we observed an absorption band at ∼1100 cm⁻¹ attributed to the ν₃(SO₄) mode, and two broad bands at ∼860 and ∼940 cm⁻¹ ascribed to the ν₃(SiO₄) and ν₃(SO₃) modes, respectively. The band at ∼940 cm⁻¹ shows a higher absorbance than that at ∼860 cm⁻¹. This spectral pattern was very similar to that observed in our data. Moreover, the sum of the octahedral cations in the chemical formula [(Al + Mg) = 1.86 apfu] was less than the ideal value of 2.00 apfu. Therefore, it seems highly likely that in holotype katoite, sulfur atoms not only form (SO₄)²⁻ coordination at the Z site but also form (SO₃)²⁻ coordination at the Y’ site.

ACKNOWLEDGMENTS

We thank T. Sato, E. Hirabayashi, and H. Fukusako for preparing the polished thin sections used in this study. We are also grateful for the constructive reviews by Prof. S. Kobayashi and Dr. A. Kyono, and to the anonymous reviewer for improving the manuscript.

REFERENCES

•  Antao,  S.M. (2015) Crystal chemistry of birefringent hydrogrossular. Physics and Chemistry of Minerals, 42, 455-474.
•  Banno,  Y.,  Bunno,  M. and  Tsukimura,  K. (2018) A reinvestigation of holotype wadalite from Tadano, Fukushima Prefecture, Japan. Mineralogical Magazine, 82, 1023-1031.
•  Cohen-Addad,  C. (1970) Etude du composé Ca₃Fe₂(SiO₄)_1.15(OH)_7.4 par absorption infrarouge et diffraction des rayons X et des neutrons. Acta Crystallographica, A27, 68-70.
• Deer, W.A., Howie, R.A. and Zussman, J. (1997) Rock-forming minerals. 1A (second edition), Orthosilicates. pp. 919, The Geological Society, London.
•  Galuskin,  E.V.,  Gazeev,  V.M.,  Lazic,  B.,  Armbruster,  T., et al (2009) Chegemite Ca₇(SiO₄)₃(OH)₂ - a new humite-group calcium mineral from the Northern Caucasus, Kabardino-Balkaria, Russia. European Journal of Mineralogy, 21, 1045-1059.
•  Geiger,  C.A. and  Rossman,  G.R. (2018) IR spectroscopy and OH⁻ in silicate garnet: The long quest to document the hydrogarnet substitution. American Mineralogist, 103, 384-393.
•  Geiger,  C.A. and  Rossman,  G.R. (2020) Micro- and nano-size hydrogarnet clusters and proton ordering in calcium silicate garnet: Part I. The quest to understand the nature of “water” in garnet continues. American Mineralogist, 105, 455-467.
•  Grew,  E.S.,  Locock,  A.J.,  Mills,  S.J.,  Galuskina,  I.O., et al (2013) Nomenclature of the garnet supergroup. American Mineralogist, 98, 785-811.
•  Kato,  M. and  Kyono,  A. (2019) An in situ Raman study on katoite Ca₃Al₂(O₄H₄)₃ at high pressure. Journal of Mineralogical and Petrological Sciences, 114, 18-25.
•  Kobayashi,  S. and  Shoji,  T. (1983) Infrared analysis of the grossular-hydrogrossular series. Mineralogical Journal, 11, 331-343.
•  Kobayashi,  S. and  Shoji,  T. (1984) Infrared analysis of the grossular-hydrogrossular series with a small amount of andradite molecule. Mineralogical Journal, 12, 122-136.
•  Kobayashi,  S. and  Shoji,  T. (1987) Infrared spectra and cell dimensions of hydrothermally synthesized grandite-hydrograndite series. Mineralogical Journal, 13, 490-499.
•  Kyono,  A. and  Arora,  S. (2019) Crystal structure change in grossular-Si-free katoite solid solution: Oxygen position splitting in katoite. Journal of Mineralogical and Petrological Sciences, 114, 189-200.
•  Kyono,  A.,  Kato,  M.,  Sano-Furukawa,  A.,  Machida,  S.I. and  Hattori,  T. (2019) Crystal structure change of katoite, Ca₃Al₂(O₄D₄)₃, with temperature at high pressure. Physics and Chemistry of Minerals, 46, 459-469.
•  Lacivita,  V.,  Mahmoud,  A.,  Erba,  A.,  D’Arco,  P. and  Mustapha,  S. (2015) Hydrogrossular, Ca₃Al₂(SiO₄)_3−x(H₄O₄)_x: An ab initio investigation of its structural and energetic properties. American Mineralogist, 100, 2637-2649.
•  Makreski,  P.,  Runčevski,  T. and  Jovanovski,  G. (2011) Minerals from Macedonia. XXVI. Characterization and spectra-structure correlations for grossular and uvarovite. Raman study supported by IR spectroscopy. Journal of Raman Spectroscopy, 42, 72-77.
•  Momma,  K. and  Izumi,  F. (2011) VESTA 3 for three-dimensional visualization of crystal, volumetric and morphology data. Journal of Applied Crystallography, 44, 1272-1276.
• Nyquist, R.A. and Kagel, R.O. (1997) The handbook of infrared and Raman spectra of inorganic compounds and organic salts. Volume 4, Infrared spectra of inorganic compounds (3800-45 cm⁻¹). pp. 495, Academic Press, San Diego.
• Nyquist, R.A., Putzig, C.L. and Leugers, M.A. (1997a) The handbook of infrared and Raman spectra of inorganic compounds and organic salts. Volume 1, Infrared and Raman spectral atlas of inorganic compounds and organic salts: text and explanations. pp. 119, Academic Press, San Diego.
• Nyquist, R.A., Putzig, C.L. and Leugers, M.A. (1997b) The handbook of infrared and Raman spectra of inorganic compounds and organic salts. Volume 3, Infrared and Raman spectral atlas of inorganic compounds and organic salts: infrared spectra. pp. 281, Academic Press, San Diego.
•  Palatinus,  L. and  Chapuis,  G. (2007) SUPERFLIP - a computer program for the solution of crystal structures by charge flipping in arbitrary dimensions. Journal of Applied Crystallography, 40, 786-790.
•  Passaglia,  E. and  Rinaldi,  R. (1984) Katoite, a new member of the Ca₃Al₂(SiO₄)₃-Ca₃Al₂(OH)₁₂ series and a new nomenclature for the hydrogrossular group of minerals. Bulletin de Minéralogie, 107, 605-618.
•  Rinaldi,  R. and  Passaglia,  E. (1989) Hibschite topotype: crystal chemical characterization. European Journal of Mineralogy, 1, 639-644.
• Ross, S.D. (1972) Inorganic infrared and Raman spectra. pp. 413, McGraw-Hill, London.
•  Sheldrick,  G.M. (2015) Crystal structure refinement with SHELXL. Acta Crystallographica, C71, 3-8.
•  Żabiński,  W. (1966) Hydrogarnets. Prace Mineralogiczne: Polska Akademia Nauk Oddział w Krakowie Komisja Nauk Mineralogicznych, 3, 1-61.