Ternary aurides RE₄Mg₃Au₁₀ (RE = La, Ce, Pr) and RE₄Cd₃Au₁₀ (RE = Y, La–Nd, Sm, Gd–Dy) – ordering variants of the Zr₇Ni₁₀ type

2.1 Synthesis

Starting materials for the syntheses of the RE₄Mg₃Au₁₀ and RE₄Cd₃Au₁₀ phases were pieces of sublimed rare earth elements (Smart Elements, >99.9 %), gold granules (Allgussa AG, >99.9 %), a magnesium rod (Johnson Matthey, >99.5 %), and a cadmium bar (Johnson Matthey, >99.9 %). Smaller pieces of the moisture sensitive rare earth ingots were cut under paraffin oil and washed with cyclohexane. The paraffin oil and cyclohexane were dried over sodium wire. The pieces were then kept under dry argon in Schlenk tubes. The argon was purified over titanium sponge (900 K), silica gel, and molecular sieves.

The low boiling temperatures of magnesium and cadmium requested sealed reaction containers. The ideal elemental mixtures 4RE: 3Cd/Mg: 10Au were thus arc-welded [22] in small niobium ampoules under an argon pressure of ca. 700 mbar. The ampoules were subsequently placed in a water cooled sample chamber of a high-frequency furnace [23] (Typ TIG 2.5/300, Hüttinger Elektronik, Freiburg, Germany). The annealing sequence was similar to that used for the previously synthesized RE₄Mg₃Au₁₀ phases with the heavier rare earth elements [18]: (i) rapid heating to ca. 1450 K, (ii) keeping that temperature for 1 min, (iii) reducing slowly to 1300 K, (iv) 30 min at 1300 K, (v) cooling to 900 K within 30 min. (vi) 900 K for two hours, and (vii) quenching the annealing sequence by turning off the high frequency generator. The magnesium and cadmium samples were separated from the niobium ampoules by careful mechanical fragmentation. There was no visible attack of the crucible material. The polycrystalline samples are dark gray. Small single crystals appear light gray with metallic luster. The samples are stable in air over several weeks.

The series of RE₄Mg₃Au₁₀ phases with the heavier rare earth elements [18] formed peritectically. Consequently, also the polycrystalline samples prepared in the induction furnace were sealed in evacuated silica ampoules and annealed at 823 K for 10 days.

2.2 X-ray diffraction

Parts of the annealed RE₄Mg₃Au₁₀ and RE₄Cd₃Au₁₀ samples were ground to fine powders and characterized through Guinier powder patterns (Enraf-Nonius camera, type FR 552): imaging plate detector, Fujifilm BAS-1800, CuK_α1 radiation and α-quartz (a = 491.30, c = 540.46 pm) as an internal standard. The lattice parameters (Table 1) were deduced from least-squares fits. The experimental patterns were compared to simulated ones [24] to ensure correct indexing. Most samples were obtained in X-ray-pure form. As an example we present the experimental and simulated powder pattern of Tb₄Cd₃Au₁₀ in Fig. 1.

Fig. 1:

Experimental and simulated Guinier powder pattern of Tb₄Cd₃Au₁₀.

Irregularly shaped single crystals were selected from the crushed Pr₄Cd₃Au₁₀ and Tb₄Cd₃Au₁₀ samples and fixed to quartz fibres using bees wax. The crystals were first characterized on a Buerger camera (using white Mo radiation) to check their quality. Intensity data of the Pr_4.46Cd_2.54Au₁₀ crystal were collected at room temperature on a Stoe IPDS-II image plate system (graphite monochromatized MoK_α radiation; λ = 71.073 pm) in oscillation mode. The Tb_4.38Cd_2.62Au₁₀ data set was measured on a Stoe Stadi Vari diffractometer equipped with a Mo micro focus source and a Pilatus detection system and scaled subsequently due to the Gaussian-shaped profile of the X-ray source. Numerical absorption corrections (along with scaling for the Stadi Vari data set) were applied to the data sets. Details about the data collections and the crystallographic parameters are summarized in Table 2.

2.3 Structure refinements

Both diffractometer data sets showed C-centred orthorhombic lattices and the systematic extinctions were compatible with space group Cmce. The atomic parameters of Gd_4.43Mg_2.57Au₁₀ were taken as starting values [18] and the structures were refined using jana2006 [25] with anisotropic atomic displacement parameters for all atoms. Since Gd_4.43Mg_2.57Au₁₀ [18] and the isotypic cadmium compound Ca₅Cd₂Au₁₀ [16] revealed Gd/Mg, respectively, Ca/Cd mixing on two of the cation sites, we carefully checked the occupancy parameters of the rare earth sites. Crystals of both cadmium compounds revealed Cd/RE mixing only for the 8e site, while all other sites were fully occupied within two standard deviations. The Cd/RE mixed occupancies were refined as least-squares variables in the final cycles, leading to the compositions listed in Table 2. The final difference Fourier syntheses revealed no residual peaks. The refined atomic positions, displacement parameters, and interatomic distances (exemplarily for Pr_4.46Cd_2.54Au₁₀) are given in Tables 3 and 4.

Further details of the crystal structure investigations may be obtained from Fachinformationszentrum Karlsruhe, 76344 Eggenstein-Leopoldshafen, Germany (fax: +49-7247-808-666; e-mail: crysdata@fiz-karlsruhe.de, http://www.fiz-karlsruhe.de/request_for_deposited_data.html) on quoting the deposition numbers CSD-429863 (Pr_4.46Cd_2.54Au₁₀) and CSD-429864 (Tb_4.38Cd_2.62Au₁₀).

2.4 EDX data

The Pr_4.46Cd_2.54Au₁₀ and Tb_4.38Cd_2.62Au₁₀ crystals studied on the diffractometers were semiquantitatively analyzed by EDX using a Zeiss EVO^® MA10 scanning electron microscope in variable pressure mode. The rare earth trifluorides, gold, and cadmium were used as standards. The experimentally observed compositions were close to the ideal ones. No impurity elements (especially from the container material) were detected.

2.5 Physical property measurements

The magnetic and heat capacity measurements were carried out on a Quantum Design Physical Property Measurement System (PPMS) using dc-MS and heat capacity options, respectively. Dc magnetic measurements were performed using the VSM (Vibrating Sample Magnetometer) option. For this measurement 56.323 mg of the powdered Tb_4.38Cd_2.62Au₁₀ sample were packed in a polypropylene capsule and attached to the sample holder rod. For the heat capacity measurement a piece of a sintered tablet (12.929 mg) was fixed to a pre-calibrated heat capacity puck using Apiezon N grease. Magnetic investigations were performed in the temperature range of 2.5–305 K with magnetic flux densities up to 80 kOe (1 kOe = 7.96 × 10⁴ A m^–1). Heat capacity measurements were done in the temperature range of 2.5–310 K.

3.1 Crystal Chemistry

The magnesium and cadmium compounds RE₄Mg₃Au₁₀ (RE = La, Ce, Pr) and RE₄Cd₃Au₁₀ (RE = Y, La–Nd, Sm, Gd–Dy) extend the series of intermetallic compounds with ordered Zr₇Ni₁₀ type structure, space group Cmce, Pearson symbol oS68 and Wyckoff sequence g²f²eda. The course of the cell volumes within both series (including the data from [18]) is presented in Fig. 2. As expected from the lanthanide contraction, the cell volumes decrease almost monotonically from the lanthanum to the dysprosium compound with larger cell volumes for the cadmium series. Ce₄Mg₃Au₁₀ and Ce₄Cd₃Au₁₀ fit into the smooth function, indicating trivalent cerium. Both yttrium phases fit in between the cell volumes of the corresponding terbium and dysprosium phases as frequently observed for rare earth based magnesium and cadmium intermetallics [20, 21].

Fig. 2:

Plot of the cell volumes of the RE₄Mg₃Au₁₀ and RE₄Cd₃Au₁₀ aurides.

Exemplarily we discuss the structure of Pr_4.46Cd_2.54Au₁₀. In the first step of the discussion we refer to the ideal composition, neglecting the 23 % praseodymium occupancy on the 8e cadmium site. A projection of the Pr₄Cd₃Au₁₀ structure along the c axis is presented in Fig. 3. Although the unit cell seems quite complex at first sight, it can easily be subdivided into two substructures.

Fig. 3:

View of the Pr₄Cd₃Au₁₀ unit cell along the crystallographic c axis. Praseodymium, cadmium, and gold atoms are drawn as light grey, magenta, and blue circles, respectively. The two-dimensional gold substructure and the location of the Cd1/Pr2 and Pr1/Cd2 atoms in the different layers are emphasized. Note that the Cd2 site shows mixing with praseodymium (Table 3).

The gold atoms build up a two-dimensional substructure around the mirror planes at x = 0, 1/2 and 1. These blocks show a quite large range of Au–Au distances from 275 to 327 pm, shorter and longer than in fcc gold (288 pm) [26]. Each gold atom has four or five nearest gold neighbors. The blocks leave cages of eight, respectively, 11 gold atoms (Fig. 4) which are filled by the Cd1 and Pr2 atoms.

Fig. 4:

Cutout of the Pr₄Cd₃Au₁₀ structure. Praseodymium, cadmium, and gold atoms are drawn as light grey, magenta, and blue circles, respectively. One layer of the gold substructure along with the Pr2 and Cd1 atoms is drawn. Atom designations and relevant interatomic distances are given.

The Au–Cd distances within the slightly distorted Cd1@Au₈ and the strongly distorted Cd2@Au₈ cubes (Fig. 5) range from 283 to 293 pm, slightly longer than the sum of the covalent radii for gold and cadmium of 275 pm [27]. Similar ranges of Au–Cd distances were observed in the [AuCd] network of EuAuCd (288–295 pm) [28] and the [Au₇Cd₁₆] network of Na₆Au₇Cd₁₆ (278–300 pm) [29]. Similar to EuAuCd and Na₆Au₇Cd₁₆, the Au–Cd interactions certainly play an important role for the stability of the RE₄Cd₃Au₁₀ phases.

Fig. 5:

Cutout of the Pr₄Cd₃Au₁₀ structure. Praseodymium, cadmium, and gold atoms are drawn as light grey, magenta, and blue circles, respectively. One layer with the ondensed Pr1@Au₁₀ and Cd2@Au₈ polyhedra is emphasized.

The gold containing blocks are separated by the Pr1 and Cd2 atoms. One of these layers is presented in Fig. 5. It can be explained as a 1:1 intergrowth of Pr1@Au₁₀ and Cd2@Au₈ polyhedra in a checkerboard pattern. The distortion of the Cd2@Au₈ unit is much stronger than for the almost regular Cd1@Au₈ cubes (Fig. 4). For an alternative description of these structures through a condensation of (Mg/Cd)@Au₈RE2₄ and RE1@Au₁₀(Mg/Cd)₄ polyhedra we refer to [18].

Finally we draw back to the homogeneity ranges. Already for the series of RE₄Mg₃Au₁₀ aurides we observed significant RE/Mg mixing on 4a and 8e sites, leading to the solid solutions RE_4+xMg_3–xAu₁₀. Single crystal data for Gd_4+xMg_3–xAu₁₀ revealed the highest x value of 1.61 [18] and a value of x = 1 for the alkaline earth compound Ca₅Cd₂Au₁₀ [16]. The four available structure refinements showed a strong variation of the occupancy parameters without any systematics. Interestingly Ca₄In₃Au₁₀ [15] shows complete calcium/indium ordering and the RE₄Cd₃Au₁₀ aurides show only RE/Cd mixing on the 8e sites, at least for the investigated crystals. These mixed occupancies play an important role for the magnetic behavior and are re-addressed below.

In analogy to the RE₄Mg₃Ag₁₀ series we also tried the synthesis of the related silver-cadmium phases. The complex powder patterns gave hints for the existence of such phases, but no phase-pure materials could be obtained under the present synthesis conditions. The very small difference in scattering power between silver and cadmium further hampers the structure elucidation of such phases. Synthesis attempts with copper led to the Laves phases RECu₄Mg and RECu₄Cd which will be reported separately.

3.2 Magnetic behavior of Tb₄Cd₃Au₁₀

The characterization of the magnetic properties of the RE₄Cd₃Au₁₀ aurides is a challenging task. The severe problem concerns inhomogeneities within the bulk samples as a consequence of the RE/Cd mixing on the 8e sites (vide supra). We discuss this problem exemplarily for the Tb₄Cd₃Au₁₀ sample, which has been measured after an annealing period of two and four weeks. All figures shown herein were obtained from the measurement of the four weeks sample.

The top panel of Fig. 6 displays the temperature dependence of the magnetic and inverse magnetic susceptibility (χ and χ^–1 data) of Tb₄Cd₃Au₁₀ measured at 10 kOe. A fit of the χ^–1 data above 50 K using the Curie–Weiss law revealed an effective magnetic moment of μ_eff = 9.74(1) μ_B per Tb atom and a Weiss constant of θ_p = 7.9(5) K. The effective magnetic moment fits very well to the theoretical value of 9.72 μ_B for a free Tb³⁺ ion. Despite of the Tb/Cd mixing no reduced or enlarged magnetic moment was identified. The observed Weiss constant points towards weak dominating ferromagnetic interactions in the paramagnetic range. Furthermore a significant, but broad increase around 20 K temperature indicates a ferromagnetic ordering.

Fig. 6:

Magnetic properties of Tb₄Cd₃Au₁₀: (top) temperature dependence of the magnetic susceptibility (χ and χ^–1 data) measured at 10 kOe; the inset presents the magnetic susceptibility in zero-field-cooled/field-cooled (ZFC/FC) mode at 100 Oe in the low-temperature range; (bottom) magnetization isotherms at 3, 16, 30, 50 and 100 K.

To obtain more precise information about this magnetic ordering, a low-field measurement with external field strength of 100 Oe was performed in a zero-field- and field-cooled (ZFC/FC) mode. This measurement is depicted in the inset of Fig. 6 (top panel). In contrast to the ZFC measurement with 10 kOe, a Néel temperature of T_N = 12(1) K is determined from the FC data. Furthermore it is obvious that a broad shoulder is present around 20 K, which is most likely related to a second magnetic ordering. This might be related to a significant amount of Tb on the different cadmium sites. The related phase width finds expression in significant broadening of both anomalies. The sample with an annealing period of only 2 weeks exhibits even broader peaks and a third anomaly can be identified below 10 K. Due to this complex magnetic behavior we cannot definitely exclude slight amounts of impurity phases, but the overall consistent behavior points towards intrinsic magnetic “impurities”, which are caused by the mixing of Tb and Cd. These would also explain the broad bifurcation of the ZFC and FC curve, which is only observed in the case of non-zero remanent magnetizations and cannot be caused by a simple antiferromagnetic ordering.

The bottom panel in Fig. 6 shows the magnetization isotherms measured at 3, 16, 30, 50, and 100 K. The three latter isotherms (above the magnetic ordering temperature) confirm the paramagnetic character in this temperature range. Only very weak saturation effects can be observed for the 30 K isotherm, while the two others exhibit a totally linear field dependency. The 16 K isotherm exhibits a much steeper increase and more pronounced saturation effects due to the onset of the antiferromagnetic ordering. However, no hint for a wide-range magnetic ordering above this temperature is identified. This underlines the assumption of short range-magnetic phenomena as the reason for the broad anomaly around 20 K. At 3 K the magnetization increases even steeper and a very weak s-shaped character can be identified around 5 kOe, which is attributed to a reorientation of the spins confirming the antiferromagnetic ground state. Furthermore it explains why a ferromagnetic character is observed for the ZFC curve measured at 10 kOe. Around 10 kOe another kink is observed pointing towards a somehow canted magnetic behavior, followed by saturation effects at higher fields. At 3 K and 80 kOe a magnetization of 6.0(1) μ_B per Tb atom is measured, which corresponds to 67 % of the expected saturation magnetization of 9 μ_B per Tb atom according to g_J × J. This reduction may be ascribed to the polycrystalline character of the sample. A weak hysteresis observed of the 3 K isotherm indicates again the existence of a non-zero remnant magnetization caused by very weak ferromagnetic interactions.

In Fig. 7 the investigated heat capacity of Tb₄Cd₃Au₁₀ is displayed for the temperature range of 2.5–300 K. According to the magnified low-temperature area in the inset a weak, broad λ anomaly can clearly be identified at 11.0(5) K. This underlines the observed antiferromagnetic ordering around 12 K as the main magnetic ordering of Tb₄Cd₃Au₁₀. The absence of a second anomaly definitely excludes the existence of a second, wide range magnetic ordering around 20 K.

Fig. 7:

Heat capacity of Tb₄Cd₃Au₁₀ measured in the temperature range of 2.5–300 K without an applied field. The inset shows the magnified low-temperature area to highlight the observed λ anomaly.