Introduction
In the eastern flank of the Sierra de Maz, La Rioja Province, NW Argentina (29.23°S, 68.35°W) there is a deformed silicocarbonatite-(nepheline) syenite complex which, unlike other carbonatites known in western South America, is the only known body of Neoproterozoic age (ca. 570 Ma, SHRIMP U–Pb in zircon, Casquet et al., Reference Casquet, Pankhurst, Galindo, Rapela, Fanning, Baldo, Dahlquist, González-Casado and Colombo2008a). Silicocarbonatite (SiO2 wt.% ranging from 17.16 to 28.92, Biglia Reference Biglia2015) is the main lithological type, with syenite (with or without nepheline) occurring as metre-sized blocks and smaller rounded inclusions within the carbonatite. Together, the subparallel dykes define a system that extends in a north–south direction for ~4 km, with a maximum width of ~120 m, hosted by biotite–garnet ± hornblende gneisses, orthoamphibolites, metagabbros and local meta-peridotites. They are located within the Maz Central Domain, which records two main metamorphic episodes, one during the Neoproterozoic (1.2 Ga) and another one at 431 ± 40 Ma (Casquet et al., Reference Casquet, Pankhurst, Rapela, Galindo, Fanning, Chiaradia, Baldo, González-Casado and Dahlquist2008b).
Xenocrysts of aegirine–augite, amphibole and biotite are widespread as irregular masses scattered in the medium-grained biotite–plagioclase silicocarbonatite (Fig. 1a). With the exception of those found in a small dyke that runs parallel to the main system, xenocrysts display markedly Fe-dominant compositions, which is unusual for ferromagnesian silicates in carbonatites.
‘Katophorite amphibole’ (ferro-ferri-katophorite according to the electron microprobe chemical data, but not specifically identified as such) was found by C. Galindo and reported by Casquet et al. (Reference Casquet, Pankhurst, Galindo, Rapela, Fanning, Baldo, Dahlquist, González-Casado and Colombo2008a) as a replacement product of aegirine–augite that forms enclaves composed of albite, magnetite, ferroan calcite and amphibole. A later systematic survey, undertaken to document the chemical variability and textural relationships among ferromagnesian silicates, showed that the new amphibole is rather widespread, both as reaction rims and as xenocrysts. The new species and name (symbol Ffktp) has been approved by the Commission on New Minerals, Nomenclature and Classification of the International Mineralogical Association (IMA2016–008, Colombo et al., Reference Colombo, Rius, Molins, Biglia, Galliski, Márquez-Zavalía, Baldo and Kriscautzky2016) The name conforms to the current nomenclature scheme for amphiboles of Hawthorne et al. (Reference Hawthorne, Oberti, Harlow, Maresch, Martin, Shumacher and Welch2012). Type material (fragment of the holotype) has been deposited in the collection of the Museo de Mineralogía “Dr. Alfred Stelzner”, Universidad Nacional de Córdoba (Argentina), under catalogue number MS003341.
Occurrence
Ferro-ferri-katophorite occurs as granular masses (Fig. 1b) or as replacement rims around aegirine–augite crystals (Fig. 1c). In the first case individual crystals may reach 3 cm and show the {110} monoclinic prism elongated along [001], with ragged terminations. They are poikilitic, with inclusions of albite and calcite. When ferro-ferri-katophorite occurs as a replacement of aegirine–augite, it forms polycrystalline rims showing lobate contacts against the clinopyroxene. Crystals have an approximately parallel orientation, as they display similar colours under parallel polarised light.
All of the properties and the chemical data reported in this paper were measured on a single crystal of ferro-ferri-katophorite.
Physical and optical properties
Macroscopically, ferro-ferri-katophorite is black and opaque, with vitreous lustre and a pale greyish green streak. It does not fluoresce under ultraviolet radiation (either long- or shortwave). Grains display the typical perfect {110} cleavage of amphiboles. No twinning was observed. The mineral is brittle, with irregular or splintery fracture. Mohs hardness is 6.0 The measured density is 3.32(1) g/cm3 (determined by immersion in toluene, n = 2), while the calculated value is 3.358 g/cm3 based on the empirical formula using single-crystal unit-cell parameters.
Under the polarising microscope, ferro-ferri-katophorite shows a very strong absorption and is transparent only in very thin fragments. Optical properties were determined using white light and calibrated immersion oils. Ferro-ferri-katophorite is biaxial (–), with α = 1.688(3), β = 1.697(3), γ = 1.698(3) and 2Vcalc = 36.7°. The observed 2V was small, but could not be measured accurately because of indistinct interference figures and very strong absorption. It displays very strong parallel dispersion, with r>v. The orientation is: Z ∥ b, and X forms a small angle with [001]. Pleochroism is very strong, with X = light greenish brown, Y = dark greyish brown and Z = dark greyish olive green. Absorption is Z > Y > X. It should be noted that amphibole (identified as ferro-ferri-katophorite by electron microprobe analyses) rimming clinopyroxene may display pleochroism with bluish colours.
The compatibility index 1 – (K p/K c) is 0.030 (excellent) for the empirical formula using the density derived from the single-crystal unit cell, and 0.019 (superior) for the empirical formula using the density derived from powder X-ray diffraction data.
Chemical composition
Electron microprobe analysis
Ferro-ferri-katophorite was analysed using wavelength dispersive spectroscopy on a JEOL JXA-28230 electron microprobe (LAMARX-Universidad Nacional de Córdoba, Argentina) operated at an accelerating voltage of 15 kV and a beam current of 20 nA. The beam was defocussed to 10 μm. We measured Kα lines, except for Zr (Lα). An empirical Ti–V overlap correction was applied. Data reduction was made with ZAF, as implemented by the JEOL software. The chemical composition of the type specimen appears in Table 1, along with standards and monochromator crystals. No zoning was detected using back-scattered electron images. An average of the composition of replacement rims around aegirine–augite (taken from Casquet et al., Reference Casquet, Pankhurst, Galindo, Rapela, Fanning, Baldo, Dahlquist, González-Casado and Colombo2008a) is also listed for comparison. The calculated structural formulae appear in Table 2, column 1 (for the type specimen) and column 5 (for the rims).
S.D. – standard deviation
* Total Fe expressed as Fe2O3 and FeO based on the Mössbauer spectroscopy results. O≡Cl = 0.002 wt.%, eliminated by rounding off.
*Blank: not measured. Normalisation procedures used for formula, as described in Hawthorne et al. (Reference Hawthorne, Oberti, Harlow, Maresch, Martin, Shumacher and Welch2012): [1] average between Si+Ti+Zr+Al+Fe+Mn+Mg+V+Zn = 13 apfu and Σcations = 16 apfu; [2] Si+Ti+Zr+Al+Fe+Mn+Mg+V+Zn = 13 apfu; [3] per 24 (O,OH,Cl).
H2O and O2– contents were calculated based on a combination of results from the electron microprobe (EMP) and Mössbauer analyses, and the crystal structure refinement (as there is a relationship between Ti at the M(1) site and O2– content). For the rims around aegirine–augite, the information is more limited, as neither Mössbauer spectroscopy nor a structural study could be performed. In that case the formula was calculated under the assumption that per each Ti atom that enters at the M(1) position replacing a divalent cation, two protons are lost.
Chromium (detection limit 120 ppm, LIFH, chromite standard) was also sought but not found. A single analysis shows an F content of 0.13 wt.%; however, the average value is below the detection limit for this element (670 ppm, TAP, fluorapatite standard). For the calculation of average values, analyses below the detection limit were considered as 0.00.
Oxygen (as O2–) at the O3 site was calculated by charge balance, with (OH)– obtained as (2–O2–F––Cl–) atoms per formula unit (apfu). The empirical formula of the single crystal is: (Na0.74K0.23)Σ0.97(Ca1.08Na0.91Mn0.01)Σ2.00(Fe2+1.78Mg1.57Fe3+1.07Ti4+0.32Al0.19Mn2+0.04Zr0.01V3+0.01Zn0.01)Σ5.00(Si6.61Al1.39)Σ8.00 O22 (OH1.59O0.61)Σ2.00. The amount of Cl is above the electron microprobe detection limit, however it is only equivalent to 0.002 Cl apfu and is thus eliminated by rounding off. The ideal formula is Na(NaCa)(Fe2+4Fe3+)(Si7Al)O22(OH)2, which requires (in wt.%) SiO2 43.14, Al2O3 5.23, Fe2O3 8.19, FeO 29.48, CaO 5.75, Na2O 6.36, H2O 1.85, total 100.00%.
Mössbauer spectroscopy
The Mössbauer spectrum of a powdered sample was acquired in transmission geometry at 298 K using a conventional transmission Mössbauer spectrometer with a 57Co/Rh source. The spectrometer was calibrated with the room-temperature spectrum of α-Fe. The program package Fit;o) (Hjøllum and Madsen Reference Hjøllum and Madsen2009) was used to fit the spectrum using doublets (Fig. 2). Both peaks of each doublet were constrained to have the same area and same width. The results are shown in Table 3, where the distribution of Fe2+ and Fe3+ obtained by single-crystal structural refinement has been reported for comparison.
SREF: single-crystal structural refinement.
Crystallography
Powder X-ray diffraction
Powder X-ray diffraction data were recorded using a Siemens D-5000 diffractometer in Bragg-Brentano geometry using CuKα1+2 radiation. Observed d values and intensities were derived by full-profile fitting using the FullProf software (Rodríguez-Carvajal, Reference Rodríguez-Carvajal2001). Observed intensities are affected by preferred orientation due to the perfect {110} cleavage. The strongest five lines in the powder X-ray diffraction pattern [d in Å (I)(hkl)] are: 8.416(100)(110), 3.135(50)(310), 2.815(26)(330), 2.720(18)(151) and 1.4422(15)($\bar{6}$61). The complete dataset (in Å for CuKα1, Table S1), along with the calculated pattern, has been deposited with the Principal Editors of Mineralogical Magazine and is available as Supplementary material. Unit cell parameters measured from powder X-ray diffraction data are a = 9.8697(5), b = 18.094(1), c = 5.3320(4) Å, β = 104.603(4)° and V = 921.4(1) Å3.
Single-crystal structure refinement
Data collection and refinement details for the single-crystal structural refinement (SREF) are given in Table 4. Intensities were corrected for Lorentz and polarisation effects. The structure was refined to R 1 = 2.77% for 1064 reflections with I>2σ(I) using SHELXL (Sheldrick Reference Sheldrick2008), starting from the atomic coordinates of katophorite (sample A7) published by Hawthorne et al. (Reference Hawthorne, Oberti and Sardone1996). We used ionised atom scattering curves for all atoms, except for H (neutral) and Si, where neutral vs. ionised curves were applied (cf. Hawthorne et al., Reference Hawthorne, Ungaretti and Oberti1995). The H atom was located by Fourier synthesis; its U eq was constrained to be 1.2 times the U eq of O(3).
Atomic coordinates and atomic displacement parameters (ADP) are listed in Table 5, whereas selected bond lengths appear in Table 6.
1 Set as 1.2 times the U eq of O3.
Symmetry transformations used to generate equivalent atoms: (i) x–½, y+½, z; (ii) –x+½, –y+½, –z; (iii) –x+½, y+½, –z.
Crystal structure and site populations
Site populations calculated on the basis of EMP, Mössbauer spectroscopy and SREF analyses are reported in Table 7.
N: number of equivalent sites in the structural formula. apfu: atoms per formula unit; epfu: electrons per formula unit; SREF: single crystal structural refinement; EMP: electron microprobe.
The <T(2)–O> distance, 1.627 Å, is very short and suggests that no Al occurs at this site; T(2) seems to be fully occupied by Si.
The application of the relationship T (1)Al = 32.17 <T(1)–O> – 52.12, proposed by Hawthorne and Oberti (Reference Hawthorne, Oberti, Hawthorne, Oberti, Della Ventura and Mottana2007), gives T (1)Al = 1.14 apfu, lower than the 1.40 Al apfu necessary to completely fill the site; nevertheless, the 1.40 value is well within the general trend shown by Hawthorne and Oberti (2007, their figure 17). Other cations that could occupy tetrahedral sites in amphiboles are Fe3+ and Ti4+. However, no fourfold-coordinated Fe3+ was detected by Mössbauer spectroscopy. Titanium has been shown to order at the T(2) site, where it causes an increase of <T(2)–O> distance and of the mean atomic number (Oberti et al., Reference Oberti, Ungaretti, Cannillo and Hawthorne1992); none of these features are present in the sample of ferro-ferri-katophorite studied. Therefore, we conclude that the T(1) site is only occupied by Si and Al, and that this last element is strongly ordered at this site.
Atomic populations in the octahedral M(1), M(2) and M(3) sites were assigned taking into account the refined site-scattering values and <M–O> distances. Calculated distances were based on the values typical for each element as given by Hawthorne (Reference Hawthorne1983, table 28). When no values were available (for Zr, Zn and V3+, all of them present in trace amounts), they were calculated from Shannon (Reference Shannon1976). They were assigned to the M(2) site based on the results obtained by Oberti et al. (Reference Oberti, Vannucci, Zanetti, Tiepolo and Brumm2000) (for Zr) and Hawthorne et al. (Reference Hawthorne, Ungaretti, Oberti, Bottazzi and Czamanske1993) (Zn). Nevertheless, their assignments remain tentative because of their very low concentration in ferro-ferri-katophorite.
The M(4) site was modelled as a split site [M(4) + M(4´)], with full occupancy; a small amount of Mn2+ was assigned to the M(4´) site. Oberti and Ghose (Reference Oberti and Ghose1993) mentioned that the M(4´) site has a distorted [6+2] coordination, and can host divalent cations more suitably than the M(4) site.
As in other amphiboles belonging to the sodium–calcium group, alkalis are distributed over two subsites [A(m) and A(2)] within the A cavity (Hawthorne et al., Reference Hawthorne, Oberti and Sardone1996). The large ADP ellipsoids of atoms located at this cavity is related to the difficulties in accurately modelling the electron density of this site.
Relationship with other species
Ferro-ferri-katophorite belongs to the sodium–calcium amphibole group. In particular, species bearing the katophorite root name fulfil the conditions: A(Na+K+2Ca)> 0.5 apfu and 0.5 apfu < C(Al+Fe3++2Ti) < 1.5 apfu.
The composition Na(NaCa)(Fe2+4Fe3+)(Si7Al)O22(OH)2 (currently ferro-ferri-katophorite) was named ferri-katophorite in the classification schemes of Leake (Reference Leake1978) and Leake et al. (Reference Leake, Woolley, Arps, Birch, Gilbert, Grice, Hawthorne, Kato, Kisch, Krivovichev, Linthout, Laird, Mandarino, Maresch, Nickel, Rock, Schumacher, Smith, Stephenson, Ungaretti, Whittaker and Youzhi1997). However, no phase with this composition was ever submitted for formal approval to the IMA–CNMNC.
There are currently 9 species (including the one described in this proposal) that justify the katophorite root name (Table 8). To date only katophorite and ferri-fluoro-katophorite have been formally described as such; the others have the ‘Redefined’ status, or have been identified by chemical analyses (sometimes complemented by single-crystal structural refinement) in publications, but not further characterised.
The samples described by Brøgger (Reference Brøgger1894) as katophorite from near Grorud, Oslo (Norway) are strongly zoned, and thus the chemical analysis does not correspond to a single phase. If the formula is calculated on a 23O eq basis, the mineral classifies as ferro-ferri-katophorite, however some significant violations to amphibole stoichiometry (such as ΣA = 1.44 apfu) and the presence of Fe3+ in tetrahedral coordination indicate that there are analytical problems.
Ferro-ferri-katophorite has been identified by electron microprobe analysis from four other localities: the Xiangshan pluton, Funing Co., Qinhuangdao Prefecture, Hebi Province, China (Zhou et al., Reference Zhou, Li, Zhao and Wang1989); Olenii (or Olenyi) Creek, Mt. Kukisvumchorr, Khibiny Massif, Kola Peninsula, Murmanskaja Oblast', Northern Region, Russia (Pekov and Podlesnyi, Reference Pekov and Podlesnyi2004); the In den Dellen quarries, Mendig, Mayen-Koblenz district, Rhineland-Palatinate, Germany (Schäfer and Schäfer, Reference Schäfer and Schäfer2018); and the Motzfeldt center, Igaliku complex, Kujalleq, Greenland (Schönenberger and Markl, Reference Schönenberger and Markl2008). The Mindat database (www.mindat.org, accessed December 29th 2022) lists two other occurrences of ferro-ferri-katophorite, based on personal communications but for which there are no formal publications: the Abdong (or Aptong) Zr–Nb deposit, Pyonggang-gun, Kangwon Province, North Korea, and the Água de Pau volcano, San Miguel, Azores, Portugal.
Calculation of the Fe3+/Fe2+ ratio and the oxo component in ferro-ferri-katophorite
Note that when only microprobe chemical data are available, the Fe3+ content may be severely underestimated, depending on the normalisation scheme used. When the procedure recommended by Hawthorne et al. (Reference Hawthorne, Oberti, Harlow, Maresch, Martin, Shumacher and Welch2012) is followed, starting with all Fe as Fe2+, the Fe3+/FeTotal ratio is much lower than that obtained by Mössbauer spectroscopy [0.45 vs. 1.07 Fe3+ apfu, or Fe3+/(Fe2+ + Fe3+) = 0.16 vs. 0.38, Table 2, column 2]. Thus, it is likely that some ferro-ferri-katophorite will be misidentified as another related species when the analysis is based only on microprobe data. The situation does not improve much if the empirical formula is calculated based on Σcations = 13 apfu (excluding Ca, Na and K), and then adjusting the Fe2+/Fe3+ ratio to obtain 46 positive charges (Table 8, column 4). In addition, this neglects the presence of any Fe2+, Mn and Mg in the B site.
A better procedure is to assign two O2– atoms (located at the O3 position) per each atom of Ti (Table 2, column 3), following the Ti4+ + 2O2– → (Mg,Fe,Mn)2+ + 2(OH)– substitution (Oberti et al., Reference Oberti, Ungaretti, Cannillo and Hawthorne1992). Although this is an oversimplification, as only Ti at the M(1) site is associated with deprotonation, this nevertheless gives a better approximation. When this procedure is applied to the microprobe analysis of this paper (starting with all Fe as FeO), it gives a Fe3+/(Fe2+ + Fe3+) ratio which is equal to the value obtained by Mössbauer spectroscopy. In addition, a small amount of Mn is then assigned to the B site, as also indicated by the single-crystal structural refinement.
As a further check, we applied the equation proposed by Oberti et al. (2015b, p. 289), using an M(1)–M(2) distance of 3.133 Å for ferro-ferri-katophorite (as calculated from the structure refinement). We obtained a maximum amount of O at the O3 position equal to 0.59 O2– apfu, in good agreement with the 0.61 O2– apfu value obtained by charge balance (with Mössbauer data) or 0.64 O2– apfu calculated assuming the substitution mentioned above.
Acknowledgements
This research was funded by the grants PICT 2698 (2010), PICT-2017-0619, PIP 112-20200101489-CO and SECyT-UNC proyecto Consolidar 2019-2021. CONICET is gratefully acknowledged for a postdoctoral stay of FC at ICMAB (Spain). Members of the IMA-CNMNC are thanked for their comments. We are grateful to two anonymous reviewers for their constructive comments that improved the manuscript, and to Stuart Mills for the editorial handling. This paper is dedicated to the memory of Carmen Galindo, who first noticed the amphibole with peculiar composition at this locality.
Supplementary material
To view supplementary material for this article, please visit https://doi.org/10.1180/mgm.2023.2
Competing interests
The authors declare none.