Occurrence and mineral assemblages

Nickolayite was found in a small quarry operated for phosphorites, located in the Jizah District, Amman Governorate, Jordan (31°21′52″N, 36°10′55″E) (Britvin et al., Reference Britvin, Murashko, Vapnik, Polekhovsky and Krivovichev2015). The quarry exposes a suite of pyrometamorphic rocks belonging to the Hatrurim Formation, in general comprising severely metamorphosed chalks and marls. Phosphide assemblages, described previously, are confined to the fused, originally sedimentary lithologies, known in the literature under the term ‘paralava’ (Vapnik et al., Reference Vapnik, Sharygin, Sokol and Shagam2007) (Fig. 1). The paralavas containing nickolayite are dense rocks of dark-brown colour, consisting of submillimetre-sized elongated crystals of clinopyroxene and anorthite (Fig. 2). The mineral composition and texture of these paralavas resembles that of gabbro–dolerites, but they exhibit some specific features, such as a spatially high content of tridymite (confirmed by electron back-scatter diffraction - EBSD). Plagioclase crystals are chemically homogeneous and show compositions within a narrow range of Na/Ca ratios, approaching an anorthite end-member (Table 1). In contrast, the clinopyroxenes exhibit an extraordinary compositional zonation, varying from diopside (core) to hedenbergite (rim) within the same crystal (Table 2, Fig. 1) and can contain considerable amounts of Cr, V and Ti (Table 2). In addition, clinopyroxene can contain considerable amounts of Cr, V and Ti (Table 2). The accessory minerals are represented by baryte, chromite, hematite, trevorite, pyrrhotite, fluorapatite, titanite and powellite. The paralavas bear numerous submillimetre- to millimetre-sized cavities typically filled with secondary calcite (Fig. 1).

Fig. 1. The dark green–grey paralava schlieren, along with white calcite veinlets dissecting metamorphosed chalk. Phosphorite quarry, Daba Siwaqa complex, West Jordan.

Fig. 2. The microstructure of nickolayite-bearing paralava. The cross-sections of euhedral clinopyroxene crystals (Cpx) exhibit sharp compositional zoning in back-scattered electrons (BSE), due to differences in Fe/Mg contents (cf. Table 2). Plagioclase (Pl) (Table 1) and tridymite (Trd), along with pyrrhotite (white inclusions) fill up the interstices between clinopyroxene crystals. Polished section, SEM BSE image.

Table 1. Representative chemical compositions of anorthite associated with nickolayite (wt.%).*

* b.d.l. – below detection limit

Table 2. Representative chemical compositions of clinopyroxene associated with nickolayite (wt.%).*

* b.d.l. – below detection limit

Appearance and physical properties of nickolayite

Nickolayite forms irregularly shaped rounded single-crystal grains up to 80 μm, disseminated in the host silicate matrix (Fig. 3). The grains show no signs of alteration (reaction) at the contact with the rock-forming minerals. In reflected light, the mineral has a purely white colour with no observable birefringence or pleochroism. It exhibits very weak anisotropy (ΔR ₅₈₉ = 1.95%). Reflectance values are presented in Table 3. Loose nickolayite grains extracted from the rock matrix have metallic light-grey to greyish-white colour. The mineral is ductile and has no cleavage. The mean micro-indentation hardness (VHN load of 50 g) is 538 mm^–2, ranging from 490 to 588 kg mm^–2, which corresponds to a Mohs hardness of 5 to 6. The density of the holotype crystal calculated from the empirical formula and the single-crystal unit-cell parameters is 7.840 g cm^–3.

Fig. 3. The holotype nickolayite grain (specimen number 5290/1) in paralava. (a) General view. (b, c) The enlarged area, photomicrograph in reflected light (b) and BSE image (c). Abbreviations: Cpx – clinopyroxene; Trd – tridymite; and Pl – plagioclase.

Table 3. Reflectance values for nickolayite (%).*

* Data for wavelengths recommended by the IMA Commission on ore microscopy (COM) are marked in boldtype.

Chemical composition

The composition of three nickolayite grains was investigated (Table 4). The individual grains are homogeneous, with no chemical zoning. However, each one differs significantly in composition from the other two. The holotype crystal (#1 in Table 4) has the highest Mo content (0.87 atoms per formula unit), whereas grain #3 is depleted in Mo (0.44 Mo apfu) and, from the formal point of view, represents a composition intermediate between nickolayite, FeMoP, and Mo-rich allabogdanite, (Fe,Mo)₂P (Fe>Mo). The crystal structure determinations performed for grains #1 and #3 (see below) confirm the results of electron microprobe analyses. The plot of the chemical compositions presented in Fig. 4 reveals that the change in the Mo contents linearly and inversely correlates with the change in the Fe contents, implying the substitution of 1 Fe atom for 1 Mo atom. At the same time, the Ni concentration remains almost constant: a feature which cannot be explained on the basis of the available data.

Fig. 4. Part of the Fe–Mo–Ni ternary diagram illustrating the chemical composition of the nickolayite crystals studied.

Table 4. Chemical composition of nickolayite (wt.%).

* The holotype material; S.D. – standard deviation.

Crystal structure and powder X-ray diffraction

X-ray structural studies were carried out on crystal #1 (the holotype one) and #3 (for the compositions, see Table 4). Their crystal parameters, data collection and structure refinement details are given in Table 5; atomic coordinates and thermal displacement parameters are listed in Table 6. Nickolayite, being an analogue of synthetic FeMoP, belongs to the structural type TiNiSi – a metal-ordered variant of the Co₂Si framework (Rundqvist and Nawapong, Reference Rundqvist and Nawapong1966; Guérin and Sergent, Reference Guérin and Sergent1977). As a consequence, the nickolayite structure (Fig. 5) has the same site geometry and lattice metrics as the structure of the only known natural Co₂Si-type phosphide, allabogdanite (Britvin et al., Reference Britvin, Rudashevsky, Krivovichev, Burns and Polekhovsky2002). There are two symmetrically independent metal sites: M(1), tetrahedrally coordinated by 4 phosphorus atoms, and M(2), which is five-fold coordinated by P, resulting in square–pyramidal coordination (Fig. 5). Phosphorus, is nine-fold coordinated by both metal and P atoms. The corner- and edge-sharing [M(1)P₄] and [M(2)P₅] polyhedra are arranged into infinite chains propagated along the b axis (Fig. 5). Both metal sites allow wide homogeneity ranges, where, in particular, Mo can be substituted by either Fe or Ni (Rundqvist and Nawapong, Reference Rundqvist and Nawapong1966; Guérin and Sergent, Reference Guérin and Sergent1977; Oliynyk et al., Reference Oliynyk, Lomnytska, Dzevenko, Stoyko and Mar2013).

Fig. 5. Crystal structure of nickolayite. (a) General view along the b axis. Chains of edge- and corner-sharing tetrahedra [FeP₄] (yellow) and square pyramids [MoP₅] (red). (b) Projection onto (010) plane. (c) A fragment showing interconnections between tetrahedra [FeP₄] (yellow) and square pyramids [MoP₅] (red).

Table 5. Crystal parameters, data collection and structure refinement details for nickolayite.

* SHELX weighting scheme applied: w ^–1 = [σ²(F _o²) + (0.0255P)² + 0.5924P] (crystal #1); w ^–1 = [σ²(F _o²) + (0.0203P)²] (crystal #3), where P = [(F _o² + 2F _c²)/3].

Table 6. Fractional atomic coordinates and displacement parameters (Ų) for nickolayite.*

* U ¹² and U ²³ displacement parameters are equal to zero owing to site symmetry.

** Occupancy of the M(1) site is Fe_1.00 for both crystals. Occupancy of the M(2) site is Mo_0.88Ni_0.12 (crystal #1) and Mo_0.45Fe_0.45Ni_0.10 (crystal #3).

The nickolayite crystals studied are distinguished by significantly different Mo contents; therefore they are of interest in view of Mo distribution across the structure. The free refinements of site populations reveal that Mo only incorporates into the square–pyramidal M(2) site. Taking into account that the atomic metal radius of Mo (1.39 Å) is greater than the corresponding radii of Fe and Ni (1.26 and 1.24 Å, respectively) (Pauling, Reference Pauling1960), the observed behaviour agrees well with the longer M–P distances at the M(2) site, in contrast to the considerably shorter M(1)–P bonds (Table 7). In addition, the M(2)–P bonds in the Mo-rich crystal #1 are longer than the corresponding bonds in the Mo-poor crystal #3 (Table 7). However, our data disagrees with the random distribution of Mo and Fe over the M(1) and M(2) sites as was determined in the synthetic (Mo_1–xFe_x)₂P solid solutions (Oliynyk et al., Reference Oliynyk, Lomnytska, Dzevenko, Stoyko and Mar2013). This discrepancy between natural nickolayite and synthetic (Mo_1–xFe_x)₂P may indicate that in the latter case, the synthesis time was insufficient for completion of the diffusion processes, so the metal site ordering seen in nickolayite was not achieved (cf. Britvin et al., Reference Britvin, Krzhizhanovskaya, Zolotarev, Gorelova, Obolonskaya, Vlasenko, Shilovskikh and Murashko2021c). The free occupancy refinements of the M(1) site show that the latter is populated solely by Fe. Nickel is thus incorporated into the M(2) position, leading to the structural formulae given in Table 5, which are consistent with the microprobe data (Table 4).

Table 7. Selected bond lengths (Å) in nickolayite.

Powder X-ray diffraction data for crystal #1 are given in Table 8. The unit-cell parameters refined from the XRD pattern are: a = 5.942(1), b = 3.703(1), c = 6.841(2) Å and V = 150.52(4) Å³.

Table 8. Powder X-ray diffraction data for holotype nickolayite (d in Å).*

* The theoretical pattern was calculated on the basis of atomic coordinates obtained from the structure refinement and unit-cell parameters refined from powder diffraction data. The intensities of seven strongest lines are highlighted in bold. Calculated lines with intensity <2 have been omitted.