Flux-growth synthesis and structural characterization of R ₆Nb₄Al₄₃ (R = Sm, Gd, Tb, Dy, Ho, Er, Tm, Yb, and Lu)

2.1 Synthesis

All starting materials were pure elements sourced from Thermo Scientific with a stated purity 99.9 %wt. or higher. They were stored and handled inside an argon-filled glovebox (p(O₂) ≤ 1 ppm). Each element was weighed to the stoichiometric ratio 1:1:20. After weighing, the elements were put in Coorstek 2 cm³ alumina crucibles, which were subsequently transferred to fused silica tubes and capped with quartz wool. The tubes were evacuated and flamed-sealed in order to maintain an oxygen-free and airtight environment during the heat treatment. The sealed tubes were placed upright in muffle furnaces equipped with programmable temperature controllers. The heating profile included a gradual increase in temperature to 1,373 K at a rate of 100 K/h, where the temperature was maintained for 24 h. Following this, the temperature was decreased to 1,023 K at a rate of 50 K/h and held for 3 h before the remaining molten Al was centrifuged while still orange-hot.

This process effectively separated the R ₆Nb₄Al₄₃ crystals from the aluminum growth medium, although not to the fullest extent. Residual Al metal was partially removed by soaking the samples in concentrated NaOH solutions for several hours and washing them thoroughly with distilled water. This harsh treatment did not appear to affect the R ₆Nb₄Al₄₃ crystals, although we note that their surfaces were left tarnished post-treatment. The crystallographic quality of such pieces was not severely diminished, although we preferred to use untreated crystals for our crystallography work.

The as-grown crystals of all title phases exhibit metallic luster. They are stable in air over periods greater than 6 months.

2.2 Structural characterization

The synthesized samples were initially characterized by powder X-ray diffraction (PXRD) to confirm phase purity. Small portions of each sample were ground into a fine powder before undergoing PXRD analysis. The collected powder patterns (a Rigaku MiniFlex diffractometer operating in Bragg-Brentano configuration and equipped with a sealed Cu-source energized at 450 W of power) were cross-referenced with the custom structure database to confirm the outcome of each synthesis. Once phase identifications were verified by PXRD, crystals from each sample were subjected to single-crystal X-ray diffraction (SCXRD) in order to gather precise structural data. SCXRD analysis provided detailed insights into the spatial arrangement of atoms within the unit cell, and helped establish all structures as disorder-free, which allows for a comprehensive understanding of the crystal structure and ensured robust structural characterization.

All single-crystal X-ray diffraction experiments were done on a Bruker APEX III diffractometer equipped with monochromatized Mo Kα radiation (λ = 0.71073 Å). Data were collected at 200 K, a temperature that was maintained by a cryostat using cold flow of gas evaporated from liquid nitrogen. Suitable-sized single crystals were selected under the microscope and mounted on plastic loops. Raw data processing, including absorption corrections, was performed with the Bruker-supplied software. Standard coordinates from the known structure of Ho₆Mo₄Al₄₃ were used as a starting model for refinements, which were performed with the full-matrix least-squares methods on F ² implemented in the program SHELXL. ²² The refinements converged quickly to low conventional residuals and featureless difference Fourier maps, confirming the validity of the model. Refinements of site occupation factors showed no statistically significant deviations from 100 %, which is indicative of all structures exhibiting long-range ordering without measurable defects and/or admixing of atoms on the same sites, as suggested previously for some “6–4–43” aluminides, particularly those with Cr. ^{1 , 2 , 3 , 4}

Selected details of the data collection and crystallographic parameters are provided in Tables 1–3.

3.1 Synthesis

The initial observation for the formation of a ternary R–Nb–Al compound was made during the attempt to synthesize another rare-earth metal-containing compound from Al flux in a crucible made from Nb tubing. The experiment was conducted because the standard-design Al₂O₃ crucibles would crack at high temperature and the contents would leak, causing failures. Using a metallic crucible was considered, and Nb was chosen because the laboratory uses 3/8˝ tubing for other types of experiments. Under the stated conditions (see Experimental), crystals of R ₆Nb₄Al₄₃ could be indentified in several batches, which prompted our attention to systematically work out the flux-synthesis of all possible R ₆Nb₄Al₄₃ phases. So far, we have succeeded in obtaining 9 of the possible 14 (all metals from La through Lu, excluding the radioactive Pm).

Extensive attempts to make R ₆Nb₄Al₄₃ from stoichiometric melts were not made during this study. Only several tries employing an arc melter can be noted here. This method (without subsequent annealing) only led to the formation of binary phases, such as NbAl₂ and RAl₃, rather than the target ternary aluminides. It is possible that the very high temperatures of the arc caused partial volatilization, preventing the formation of the desired phase. An alternative synthesis in welded Nb-tubes involving an induction furnace and rection temperature of 1,373 K also failed, which points at the high melting point of niobium as the possible culprit. We speculate that much longer reaction times may be required for such experiments to succeed. Of course, it is also possible that R ₆Nb₄Al₄₃ are metastable phases and that excess Al creates a favorable reaction medium and allows for the crystallization of R ₆Nb₄Al₄₃ at lower temperatures, where it does not “compete” with thermodynamically stable binaries.

Attempts to make R ₆Nb₄Al₄₃ for R = La, Ce, Pr, Nd were also made, both by flux and direct reaction methods, but were not successful. Powder X-ray diffraction patterns from these samples were indicative of multi-phase products. The diffraction data also point at the possible formation of ternary phases, whose structures remain unresolved to date.

We will also note here that the synthesis of Lu₆Nb₄Al₄₃ was particularly challenging. The observation that during synthesis, multi-phase mixtures will always result, and that their relative amounts would vary from one experiment to another, points at the need for more careful control of the reaction condition in this case. Despite these challenges, crystals of Lu₆Nb₄Al₄₃ were successfully grown and analyzed, showcasing the expected lattice contraction and density increase. We posit that due to Lu’s small atomic radius, other interatomic contacts within the structure become “strained”. The reverse may also be true for the lighter R-elements, whose atoms are larger, which can cause less tighter packing, potentially leading to structural instabilities. Another possible explanation for the synthetic challenges encountered during the synthesis of Lu₆Nb₄Al₄₃ is the very high melting point of Lu, and/or its solubility in molten Al. This also underscores the impact of R on the crystal formation and stability, highlighting the need for tailored synthesis approaches to maintain structural fidelity when working with the elements from the lanthanide series.

3.2 Crystal structure and bonding

The archetype structure for the reported compounds is well known and does not require an extensive description in this article. In the following paragraphs, we will provide only a brief account, focusing on several important structural details.

Selected crystallographic data for R ₆Nb₄Al₄₃ (R = Sm, Gd, Tb, Dy, Ho, Er, Tm, Yb, and Lu) can be found in Tables 1–3. Additionally, Table 4 displays the atomic coordinates and the equivalent displacement for the atoms in the asymmetric unit of Sm₆Nb₄Al₄₃, chosen as a representative example. As seen from Table 4, there are 10 crystallographically independent atoms, all in special positions except Al1, which takes the 24l position with no symmetry constraints. A schematic view of this rather complicated structure is shown in Figure 1.

Figure 1:

Ball-and-stick representation of the hexagonal crystal structure of Sm₆Nb₄Al₄₃, viewed down the [0001] direction. Cylinders represent Al–Al contacts 3.1 Å and shorter (for clarity, only Al–Al bonding is taken into consideration here). Al atoms are drawn in turquoise, Nb atoms are in light green, and the atoms of the rare-earth metal are magenta colored. Unit cell is outlined.

As one might infer from the chemical formula, and the abundance of Al sites (seven), there is extensive Al–Al bonding, which is what we emphasize in Figure 1. Overall, the structure can be viewed as a framework of fused Al polyhedra, the largest ones hosting the rare-earth metal and Nb atoms. From the figure, one ought to be able to distinguish various Al₆-octahedra and trigonal prisms, Al₅-square prisms, as well as larger Al₁₂-icosahedral fragments. The most unusual polyherdron is perhaps that around Al6, which is showing 9 nearest Al neighbors (Figure 2). When three close by rare-earth metals are considered, the environment of Al6 becomes an icosahedron. It is noteworthy that the Al6 site in many analogs of Ho₆Mo₄Al₄₃ is known to host admixtures of Al and transition metal atoms. ^{1 , 2 , 3 , 4} For instance, in Yb₆Cr_4+x Al_43–x , Al6 is reported as 56 % Al and 44 % Cr. ⁴ In all of our refined structures, however, Al6 does not show contributions from Nb mixing in.

Figure 2:

Schematic view of the pseudo-icosahedral environment around Al6 in the structure of Sm₆Nb₄Al₄₃. Cylinders represent Al–Al contacts 3.1 Å and shorter; Sm−Al contacts 3.5 Å and shorter are depicted with thin solid lines. Al atoms are drawn in turquoise, and the atoms of the rare-earth metal are colored in magenta.

The polyhedron around the largest atom in the structure, Sm, on the representative example Sm₆Nb₄Al₄₃, is understandably the largest. According to the structural representation shown in Figure 3, Sm has 15 nearby Al atoms. Including into a consideration one Nb atom and one additional Sm atom, both located within 3.5 Å, makes for a polyhedron that has 17 vertices. Notably, two such Sm-centered polyhedra are integrown, sharing a “hexagonal” face made by four Al5 and two Al3 atoms.

Figure 3:

Schematic view of the pseudo-icosahedral environment around Sm in the structure of Sm₆Nb₄Al₄₃. Cylinders represent Al–Al contacts 3.1 Å and shorter; Sm−Al contacts 3.5 Å and shorter are depicted with thin solid lines. Al atoms are drawn in turquoise, and the atoms of the rare-earth metal are colored in magenta.

The polyhedra around the two independent Nb atoms in the structure are both with coordination number (CN) 12. Nb2, which occupies the more symmetric position with Wyckoff symbol 2b at the origin, has a nearly perfect icosahedron around itself (Figure 4a). The 12 vertices are Al2 and Al3 atoms only. The packing of the Nb2-centered, Al6-centered, and the Sm-centered polyhedra on the ab-plane is depicted in Figure 4b.

Figure 4:

Schematic view of the icosahedral environment around Nb2 in the structure of Sm₆Nb₄Al₄₃. Cylinders represent Al–Al contacts 3.1 Å and shorter; Nb–Al contacts 3.1 Å and shorter are depicted with thin solid lines. Al atoms are drawn in turquoise, and the Nb are colored in light green. (a) Nb2 atoms at the centers of icosahedra at the unit corners. (b) A slab of approximate thickness 8 Å, cut parallel to (0001) at z ∼ 0.05 and showing how the polyhedra around Sm, Nb2 and Al6 at packed.

Finally, we will briefly comment on the polyhedron around the Nb1 atoms in the structure, which are also with CN 12, but this count includes two neighboring rare-earth atoms. From the chosen representation on Figure 5, one can see how these less-regularly shaped icosahedra share corners and are intergrown together with the Al6-centered icosahedra, which also include three rare-earth metal atoms. The space around is packed with “empty” Al-polyhedra of smaller size (prisms and octahedra).

Figure 5:

Schematic view of the icosahedral environment around Nb1 in the structure of Sm₆Nb₄Al₄₃. Al and Nb atoms are drawn in turquoise and light green, respectively. The Sm atoms are colored in magenta. A slab of approximate thickness 8 Å, cut parallel to (0001) at z ∼ 0.25 shows showing how the polyhedra around Nb1 and Al6 at packed together.

3.3 Periodic trends

The lanthanide contraction is a prominent trend in this study, as evidenced by a systematic reduction in lattice parameters and unit cell volume as atomic number increases from Sm to Lu (Tables 1–3). This is the consequence from the limited shielding provided by f-electrons, which allows the nucleus to exert a stronger pull on outer electrons, effectively reducing the atomic radius. Moving across the lanthanide series, lattice parameters decrease steadily, which makes all Al–Al, Nb–Al, R–Al, and R–R contacts contract. The inability to synthesize R ₆Nb₄Al₄₃ for R = La, Ce, Pr, Nd showcases how the atomic size of the rare-earth metal influences overall crystal packing and underpins the stability of the crystal structure across the series. This trend also implies that structural and potentially electronic properties may vary predictably with the choice of R-element.

A clear deviation from the lanthanide contraction trend is the compound Yb₆Nb₄Al₄₃, presenting a slightly larger lattice parameter and unit cell volume than expected for its position in the series (Figure 6). The lattice parameters for Yb₆Nb₄Al₄₃ are larger relative to its neighbors in the series, a behavior likely stemming from Yb’s unique ability to exist in both 4f¹³ (Yb³⁺) and 4f¹⁴ (Yb²⁺) electronic configurations. This is suggestive that all, or a part of the Yb ions in Yb₆Nb₄Al₄₃ are Yb²⁺, having closed-shell electron arrangement that affects its atomic radius, and effectively causing this phase to behave differently than the other isotypic materials where the rare-earth elements have partially filled f-orbitals (except the case of Lu of course). Similar mixed-valent Yb²⁺/Yb³⁺ behavior has been reported for the isotypic Yb₆V₄Al₄₃ and Yb₆Ta₄Al₄₃ phases. ⁷

Figure 6:

Relationship between unit cell volume (V) and atomic number (Z) for the R ₆Nb₄Al₄₃ series. The values are those from Tables 1–3, obtained by single-crystal X-ray diffraction methods. A general decrease in volume with increasing atomic number is observed, consistent with lanthanide contraction. An anomaly at Yb (Z = 70) indicates that all, or part of, the Yb ions in Yb₆Nb₄Al₄₃ are Yb²⁺.