Intermetallics of the types REPd₃X₂ and REPt₃X₂ (RE=La–Nd, Sm, Gd, Tb; X=In, Sn) with substructures featuring tin and In atoms in distorted square-planar coordination

2.1 Synthesis

The starting materials for the synthesis of the ternary indium and tin compounds were ingots of the rare earth elements (Chempur or smart elements, 99.9%), sheets of palladium and platinum powder (Agosi, >99.9%), an indium bar (smart elements 99.99%) and tin droplets (Merck 99.9%). The moisture sensitive rare earth elements were cut under dried cyclohexane and kept under argon in Schlenk tubes prior to the reactions. For the platinum containing compounds, the Pt powder was cold-pressed into pellets, arc-melted [24] under an atmosphere of 800 mbar dry argon (Westfalen, 99.998%; purified over titanium sponge (T=870 K), silica gel, and molecular sieves) and subsequently cut into suitable pieces. The elements were weighed in the ideal ratios 1:3:2 and arc-melted under an argon atmosphere of 800 mbar. The product buttons were re-melted three times to ensure homogeneity. Afterwards, the samples were sealed in carbon coated evacuated silica ampoules and annealed with different temperature sequences using resistance furnaces. The purity of the samples and the crystal growth strongly depended on the annealing sequences. These are summarized in detail in the paragraphs below.

The REPd₃In₂ samples and LaPd₃Sn₂ were annealed at T=773 K for 14 days and subsequently cooled at a rate of 5 K h⁻¹ to room temperature. The single crystals of PrPd₃In₂ and NdPd₃In₂ were selected from those samples. The PrPd₃In₂ sample for the magnetic measurement had a different heat treatment history. It was first annealed at 873 K for 7 days and then cooled to room temperature at a rate of 20 K h⁻¹.

The REPt₃In₂ samples were heated at 1023 K for 35 days and cooled to room temperature at a rate of 5 K h⁻¹ (11 days and 10 K h⁻¹ for LaPt₃In₂). For the growth of PrPt₃In₂ crystals, a second sample was inductively heated (high-frequency furnace type 1.5/300, Hüttinger Elektronik, Freiburg, Germany) in a water-cooled evacuated silica ampoule (the experimental setup is documented elsewhere [25]). The sample was annealed for 10 min barely below the melting point and gradually cooled to room temperature within 7 h. Afterwards the ampoule was placed in a resistance furnace, heated to 973 K and cooled at a rate of 10 K h⁻¹ to room temperature.

The remaining REPd₃Sn₂ samples were annealed at 1173 K for 7 days and cooled to room temperature at a rate of 5 K h⁻¹ (5 days and 20 K h⁻¹ for CePd₃Sn₂). The single crystals of PrPd₃Sn₂ and NdPd₃Sn₂ were isolated from these samples. The magnetic measurement of PrPd₃Sn₂ was performed on a sample heated at 1173 K for 7 days and cooled to room temperature at a rate of 20 K h⁻¹.

The CePt₃Sn₂ and PrPt₃Sn₂ samples for the magnetic measurements were annealed at 873 K for 14 days, then at 1073 K for 7 days and finally cooled to room temperature at a rate of 10 K h⁻¹. The NdPt₃Sn₂ sample was annealed at 973 K for a period of 14 days and afterwards at 1073 K for 16 days, followed by quenching in water. The single crystals were obtained from different samples. The PrPt₃Sn₂ single crystal was isolated from a sample which was annealed at 1223 K for 21 days and cooled to room temperature at a rate of 5 K h⁻¹. The NdPt₃Sn₂ single crystal was isolated from an inductively annealed sample, treated similarly as PrPt₃In₂. The sample was annealed barely below the melting point for 10 min and gradually cooled down to room temperature within 6 h. Then, the sample was placed in a resistance furnace, heated to 1073 K for 2 days followed by cooling to room temperature at a rate of 2 K h⁻¹.

The different annealing sequences tested for the SmPd₃Sn₂ and SmPt₃Sn₂ samples always resulted in increasing amounts of impurity phases. The unit cell parameters were thus refined from the arc-melted samples. The SmPd₃In₂, GdPt₃In₂ and TbPt₃In₂ samples were not single-phase. They always contained by-products which were not diminished after the heat treatments.

All RET₃X₂ samples are silvery with metallic lustre and stable in air over several months.

2.2 X-ray diffraction

The polycrystalline RET₃X₂ samples were characterized through their Guinier powder diffraction patterns using an Enraf-Nonius FR552 camera, equipped with an imaging plate detector (Fuji film BAS-1800) and CuKα₁ radiation. α-Quartz (a=491.30, c=540.46 pm) was taken as an internal standard. The orthorhombic lattice parameters (Table 1) were obtained from least-squares refinements. All experimental patterns were compared to calculate ones (Lazy Pulverix routine [26]) in order to ensure correct indexing. The powder data (Table 1) is in good agreement with the single crystal lattice parameters (Tables 2 and 3).

The shape of the small single crystals depended on the element combination. After careful mechanical fragmentation either lath-shaped crystals or small blocks were obtained. The crystals were glued to quartz fibres using beeswax and their quality was first tested on a Buerger precession camera (white Mo radiation). Most intensity data sets were collected on a STOE IPDS-II diffractometer (graphite-monochromatized MoKα radiation) in oscillation mode or on a STOE StadiVari diffractometer equipped with a Mo micro focus source and a Pilatus detection system. Numerical absorption corrections were applied to all data sets. Due to a Gaussian-shaped profile of the micro focus X-ray source scaling was applied along with numerical absorption corrections to the StadiVari data sets. Details of the data collections and the refinement data are summarized in Tables 2 and 3.

2.3 Structure refinements

The seven data sets showed primitive orthorhombic lattices and the systematic extinctions were compatible with space group Pnma, in agreement with the previous work on the prototype CePd₃In₂ [20]. The atomic parameters of the cerium compound were taken as starting values and the structures were refined with full-matrix least-squares on Fo2 using the program Jana2006 [27] with anisotropic displacement parameters for all atoms. Separate refinements of the occupancy parameters gave no hint for deviations from the ideal composition. All sites were fully occupied within two standard deviations. The final difference Fourier syntheses revealed no significant residual peaks. The atomic positions, displacement parameters, and interatomic distances (exemplarily listed for the praseodymium compounds) are given in Tables 4 and 5.

The NdPt₃Sn₂ crystal showed slightly enhanced displacement parameters for all atoms. We collected an additional data set at 90 K. The lattice and displacement parameters showed an almost isotropic shrinking, giving no hint for a lowering of the space group symmetry.

CCDC 1959118–1959122, 1959124 and 1959211 contain the supplementary crystallographic data for this paper. These data can be obtained free of charge from The Cambridge Crystallographic Data Centre via www.ccdc.cam.ac.uk/data_request/cif.

2.4 EDX data

The single crystals studied on the diffractometers were additionally semi-quantitatively analysed by EDX, using a Zeiss EVO^® MA10 scanning electron microscope in variable pressure mode (60 Pa) with PrF₃, NdF₃, Pd, Pt, InAs and Sn as standards. Four point analyses were carried out for each crystal. The averaged values were within 1–2 at% close to the ideal composition. No impurity elements were observed.

2.5 Physical property measurements

Magnetic property measurements were conducted on X-ray phase pure samples of the general composition RET₃X₂ (RE=La–Nd, Sm; T=Pd, Pt; X=In, Sn) by means of magnetic susceptibility measurements. For this purpose, approximately 20 mg of each sample were attached to a brass sample holder rod of a vibrating sample magnetometer (VSM) of a Quantum Design Physical-Property-Measurement-System (PPMS) using Kapton foil. Magnetization experiments were carried out in the temperature range of 2.5–300 K with external magnetic field strengths of up to 80 kOe (1 kOe=7.96×10⁴ A m⁻¹). For heat capacity measurements performed on SmPt₃In₂, a small piece of the previously investigated sample was fixed to a pre-calibrated heat capacity puck using Apiezon-N^® grease and investigated in the temperature range of 2–25 K without an applied external field.

2.6 Mössbauer spectroscopy

A Ca^119mSnO₃ source was used for the Mössbauer spectroscopic characterization of the stannides REPd₃Sn₂ (RE=La, Ce, Pr, Nd) and REPt₃Sn₂ (RE=Ce, Pr, Nd). The measurements were conducted in the usual transmission geometry using a commercial nitrogen bath cryostat. The source was kept at room temperature. The REPt₃Sn₂ samples were cooled to 78 K while the REPd₃Sn₂ samples were measured at room temperature. All samples were placed in thin-walled PMMA containers and their ideal absorber thickness was calculated according to the work of Long et al. [28]. The total counting time per spectrum was between 6 and 24 h. Only in the case of CePt₃Sn₂ the counting time was extended to 3 days to obtain better resolution. The spectra were fitted with the WinNormosforIgor6 software package [29].

3.1 Crystal chemistry

The CePd₃In₂-type structure was originally described in 2004 by Nesterenko et al. [20], space group Pnma, Pearson symbol oP24 and Wyckoff sequence c⁶. Besides the prototype, isotypic LaPd₃In₂ [17] was characterized as the Pauli-paramagnetic counterpart of the cerium compound. Recently, also the isotypic stannides LaPd₃Sn₂ and CePd₃Sn₂ [23] have been reported; however, only the cerium compound has been structurally characterized. Herein we report on a complete structural characterization of the so far existing series of REPd₃In₂ and REPd₃Sn₂ phases and their platinum counterparts.

The refined lattice parameters of all CePd₃In₂-type phases along with some literature data are summarized in Table 1. This structure type shows some geometrical restrictions. The existence range is not similar for all Pd(Pt)/In(Sn) combinations. Within all four series we observe a decrease of the cell volume with increasing atomic number, a consequence of the lanthanide contraction. All cerium compounds nicely fit into the volume plots, indicating trivalent cerium. From Table 1 it is readily evident that the a and c lattice parameters of the indium and tin series show drastic differences. We address this feature in more detail below when discussing the p-element substructures.

As an example we discuss the PrPd₃In₂ structure in more detail. A projection of its unit cell is presented in Fig. 1. All atoms lie on mirror planes at y=1/4 and y=3/4, indicated by thin and thick lines in the drawing. A striking geometrical motif is the slightly distorted pentagonal prismatic coordination of the praseodymium atoms by two Pd₃In₂ pentagons. These prisms are capped on the rectangular faces by further palladium and tin atoms. Such pentagonal prisms are a general structural motif in many of the rare earth-transition metal-indides and stannides [14], [15]. Two prominent examples are the indium-based structures of Ce₄Pd₁₀In₂₁ [30] and Yb₂Pd₆In₁₃ [31]. A complex tessellation of different pentagons has been observed in the stannides Sm₃Pd_4.95Sn₅ and Sm₆Pd_9.82Sn₁₁ [32].

Fig. 1:

The crystal structure of PrPd₃In₂. Praseodymium, palladium and indium atoms are drawn as medium grey, blue and magenta circles, respectively. (bottom) Projection of the structure along the b axis. Atoms connected by thin (y=1/4) and thick (y=3/4) lines are shifted by half a translation period. (top) Cutout of the indium substructure. The crystallographically independent sites and relevant In–In distances in the slightly corrugated square net are given.

The shortest interatomic distances in the PrPd₃In₂ structure occur between the palladium and indium atoms. The Pd–In distances range from 271 to 284 pm, close to the sum of the covalent radii [2] for Pd+In of 278 pm. Similar ranges also occur in other indides, e.g. Ce₄Pd₁₀In₂₁ (269–288 pm Pd–In) [30] or EuPdIn₄ (265–280 pm Pd–In) [33]. All these short Pd–In distances are indicative of substantial covalent Pd–In bonding, in agreement with electronic structure calculations of structurally closely related compounds [34], [35].

Each of the three crystallographically independent palladium atoms in PrPd₃In₂ has at least two short Pd–Pd contacts. The Pd–Pd distances range from 283 to 316 pm. The shorter ones are close to the Pd–Pd distances of 275 pm in fcc palladium [36]. Within the PrPd₃In₂ structure we observe a clustering of the indium atoms into a slightly corrugated and distorted square network in form of a strand (Fig. 1, top). The In1 atoms have distorted square planar indium coordination with 2×325 pm In1–In1 and 2×332 pm In1–In2, while the In2 atoms have only two indium neighbors. Again, also the In–In distances compare well with those in elemental indium (tetragonal body-centered indium: 4×325 and 8×338 pm In–In) [36].

Keeping the courses of the interatomic distances in mind we can draw the following picture on chemical bonding in PrPd₃In₂. The praseodymium atoms as the least electronegative component transfer valence electrons to the palladium and indium atoms, enabling formation of the polyanionic [Pd₃In₂]^δ− network which is primarily stabilized through strong Pd–In bonds but also through In–In and Pd–Pd bonding. This rigid network (Fig. 2) leaves the pentagonal channels that are filled by the praseodymium atoms.

Fig. 2:

The complex, three-dimensional polyanionic [Pd₃In₂]^δ− network in the structure of PrPd₃In₂. Praseodymium, palladium and indium atoms are drawn as medium grey, blue and magenta circles, respectively.

Now we compare the REPd₃In₂ compounds with the platinum-based series REPt₃In₂. Although platinum (129 pm) has an only slightly larger covalent radius than palladium (128 pm) [2], we observe small shifts in the lattice parameters when directly comparing the individual unit cells of REPd₃In₂ and REPt₃In₂ (Table 1). This accounts for the individual distances in Pr–Pd, Pd–Pd and Pd–In vs. Pr–Pt, Pt–Pt and Pt–In as evident from the direct comparison of the data shown in Table 5.

Finally we turn to the stannides REPd₃Sn₂ and REPt₃Sn₂. Already the course of the lattice parameters (Table 1) indicates severe differences with respect to the corresponding indium intermetallics. The a parameters of the stannides are drastically shorter than those of the indium compounds while the b and c parameters show the inverse behavior. This is directly related to the different extension of the tin and indium substructures. As emphasized in Fig. 1, the p-element substructure extends in bc direction. While the indium network (Fig. 1) shows closer contacts, similar to those of elemental indium, the Sn–Sn distances within the unique tin substructure (distorted square planar!; exemplarily shown for PrPt₃Sn₂ in Fig. 3) of 344 and 356 pm are distinctly longer than e.g. in β-Sn (4×302 and 2×318 pm Sn–Sn) [36], indicating much weaker Sn–Sn bonding. Thus, the stannide unit cells are expanded in bc direction and in order to keep the overall bonding, they are contracted in a direction. The direct comparison of the PrPd₃In₂ and PrPt₃Sn₂ interatomic distances (Table 5) shows that the weaker Sn–Sn bonding is compensated by stronger Pt–Sn bonding. The Pt–Sn distances in PrPt₃Sn₂ range from 264 to 284 pm, close to the sum of the covalent radii for Pt+Sn of 269 pm [2]. Keeping the differences in the lattice parameters and the chemical bonding in mind, the relationship between the indium and tin series is rather isopointal [37], [38] than isotypic.

Fig. 3:

The tin substructure of PrPt₃Sn₂. The crystallographically independent tin sites are shown and relevant Sn–Sn distances are indicated.

Another crystal chemical issue concerns the individual coordination of the two crystallographically independent tin sites in the REPd₃Sn₂ and REPt₃Sn₂ series, exemplarily drawn for PrPt₃Sn₂ in Fig. 4. Both tin sites have site symmetry 4c and coordination number 13; however, with different atoms, i.e. 6 Pt+4 Sn+3 Pr for Sn1 and 7 Pt+2 Sn+4 Pr for Sn2. This leads to small changes in the tin local electron density and is thus reflected in the ¹¹⁹Sn Mössbauer spectra discussed below.

Fig. 4:

Coordination of the two crystallographically independent tin sites (both with site symmetry .m.) in PrPt₃Sn₂. Praseodymium, platinum and tin atoms are drawn as medium grey, blue and magenta circles, respectively. The crystallographically independent platinum and tin sites are indicated.

3.2 Physical properties

The basic magnetic properties of the compounds with nominal composition RET₃X₂ (RE=La–Nd, Sm; T=Pd, Pt; X=In, Sn) are summarized in Table 6. The lanthanum palladides all exhibit nearly temperature-independent Pauli-paramagnetic behavior (data not shown hear), while in the case of the corresponding platinide LaPt₃In₂ the intrinsic-core diamagnetism overcompensates the Pauli-paramagnetic contribution of the conduction electrons. All other investigated compounds except for CePt₃In₂ and SmPt₃In₂ exhibit paramagnetic behavior over the entire investigated temperature range. Long-range magnetic ordering for these compounds was not detected using zero-field-cooled and field-cooled (ZFC/FC) experiments above 2.5 K. The magnetic properties of CePt₃In₂ and SmPt₃In₂ will be discussed in further detail below. The basic magnetic properties of all compounds, except SmPt₃In₂, were obtained by fitting the experimental magnetic susceptibility data, recorded at 10 kOe, between T=50 and 300 K using the modified Curie-Weiss law. The deduced effective magnetic moments for compounds containing paramagnetic rare earth elements are in close agreement with the theoretically expected value of the free RE³⁺ ions and therefore pointing towards stable trivalent oxidation states. PrPd₃In₂ and PrPd₃Sn₂ show slightly enhanced effective magnetic moments (Table 6), most likely due to minor amounts of impurity phases (not visible in the Guinier patterns). Both compounds were prepared several times. The measurements with the lowest moments are reported in Table 6.

The results of the magnetic and inverse magnetic susceptibility measurements of CePt₃In₂, obtained at an applied external field strength of 10 kOe, are shown in Fig. 5. The effective magnetic moment per cerium atom was deduced as μ_eff=2.48(1) μ_B, close to the value expected for the free Ce³⁺ ion (μ_eff,theo=2.54 μ_B) indicating a stable trivalent oxidation state of the cerium atoms in CePt₃In₂. The paramagnetic Curie temperature of θ_P=0.8(5) K indicates the absence of dominant magnetic interactions in the paramagnetic range. In order to check for long-range magnetic ordering at low temperature, additional ZFC/FC experiments at an external magnetic field of 100 Oe were performed and the results are depicted in the middle of Fig. 5. CePt₃In₂ is ordered antiferromagnetically below T_N=4.0(1) K. The magnetization isotherms recorded at T=3, 10 and 50 K, are depicted in the bottom panel of Fig. 5. While those isotherms recorded above the magnetic phase transition show a linear field dependency, the one recorded with CePt₃In₂ being in its antiferromagnetic ground state clearly shows a metamagnetic step (antiparallel to parallel spin alignment; emphasized by the straight red line). The high critical field strength of 47(3) kOe, determined at the inflection point, suggests a stable antiferromagnetic ground state. The saturation magnetization at 80 kOe and 3 K reaches only μ_sat=1.0(1) μ_B, significantly below the theoretical value of μ_sat,theo=2.14 μ_B, frequently observed in intermetallic cerium compounds such as CeAuSn [39] and CePtAl [40], [41].

Fig. 5:

Magnetic properties of CePt₃In₂: (top) temperature dependence of the magnetic and inverse magnetic susceptibility measured at 10 kOe; (middle) temperature dependence of the magnetic susceptibility in ZFC/FC mode (100 Oe); (bottom) isothermal magnetization recorded at T=3, 10 and 50 K. The red line acts as a guide to the eye.

The temperature dependence of the magnetic susceptibility and its inverse of SmPt₃In₂ is shown in Fig. 6. The compound exhibits the typical van Vleck paramagnetism arising from the population of excited J=7/2 states in the proximity of the J=5/2 ground state of the Sm³⁺ ions [42]. Hamaker et al. developed a model to fit experimentally the temperature dependent molar susceptibilities of polycrystalline, metallic samarium compounds [43]:

Fig. 6:

Magnetic properties of SmPt₃In₂: (top) temperature dependence of the magnetic and inverse magnetic susceptibility measured at 10 kOe. The red curve represents the fit of the experimental data using the model introduced by Hamaker et al. [43]; (middle) temperature dependence of the magnetic susceptibility in ZFC/FC mode at various magnetic fields; (bottom) isothermal magnetization recorded at T=3, 10 and 50 K. The inset represents an enlarged section of the low-field region at T=3 K.

χ(T)=NAkB(μeff23(T−θP)+μB2δ)

In this equation, N_A represents the Avogadro constant, k_B the Boltzmann constant, μ_eff the effective magnetic moment, θ_P the Curie temperature, and μ_B the Bohr magneton. The term δ corresponds to an energy scale and is defined as δ=20ΔE/7, where ΔE describes the energy difference between the ground and first excited state.

The corresponding fit of the experimental data of SmPt₃In₂ is shown as a solid red line in Fig. 6 (top) and results in an effective magnetic moment of μ_eff=0.76(1) μ_B which is below the expected value of μ_eff,calc=0.845(1) μ_B for the free Sm³⁺ ion. Nonetheless, as for the other members of the RET₃X₂ series, a stable trivalent oxidation state can be assumed. The negative paramagnetic Curie temperature of θ_P=–45(1) K points towards predominant antiferromagnetic interactions in the paramagnetic temperature range. The refined energy scale δ=515(1) K corresponds to an energy difference of ΔE=1471 K. Stewart predicted the energy difference to be at 1550 K [44]. However, in intermetallic Sm compounds like SmRh₄B₄ (1080 K) [43] or SmPdGa₃ (1240 K) [45] lower values are often observed (in magnetochemistry the energy difference between the quantum states is usually given in units of K while spectroscopic investigations usually list them in units of cm⁻¹).

The middle panel of Fig. 6 shows the results of ZFC/FC experiments performed at external magnetic fields of 250, 1000 and 2500 Oe, respectively. The huge bifurcation between both curves at 250 Oe suggests a ferromagnetic ground state below approximately T=8 K. However, increasing the magnetic field strength strongly reduces this effect, pointing towards an extrinsic phenomenon. The presence of a tiny ferromagnetic impurity leads to a hysteresis effect observed in a magnetization isotherm recorded at 3 K (see Fig. 6, bottom panel). The saturation magnetization of only μ_sat=0.05(1) μ_B at 3 K and an external field of 80 kOe is significantly below the expected value of μ_sat,calc=0.714 μ_B according to g_J×J, often observed in intermetallic samarium compounds. Those isotherms observed at 10 and 50 K exhibit a linear field dependence as expected for SmPt₃In₂ being in a paramagnetic state.

In order to prove that the observed anomaly is caused by a ferromagnetic impurity, additional heat capacity measurements were performed in zero-field and the results are depicted in Fig. 7. As expected, no heat tone is present in the corresponding temperature range excluding an intrinsic magnetic ordering. Instead, a λ-shaped anomaly appears at T_HC=3.4(1) K, correlating with an antiferromagnetic ordering below T_N=3.5(1) K detectable in the ZFC measurement at an applied field of 10 kOe. This magnetic transition can only hardly be observed in the low-field ZFC/FC experiments due to the presence of the aforementioned magnetic impurities.

Fig. 7:

Heat capacity (red line) of SmPt₃In₂ measured between 2 and 25 K without an applied field and ZFC measurements at 10 kOe (black line).

3.3 Mössbauer spectroscopy

The experimental and simulated ¹¹⁹Sn Mössbauer spectra of the stannides REPd₃Sn₂ (RE=La, Ce, Pr, Nd) and REPt₃Sn₂ (RE=Ce, Pr, Nd) are presented in Figs. 8 and 9. The corresponding fitting parameters are listed in Table 7. At first sight one may assume that all samples show simple quadrupole-split signals; however, fits with single signals resulted in too large line widths. This is a clear sign for overlapping individual sub-signals. The tin coordination has already been discussed in the crystal chemical section. Based on the two different coordination polyhedra we fitted the spectra with a 1:1 overlap of two quadrupole-split sub-spectra. It was possible to refine the data without any constraints with reasonable line width parameters. The quadrupol splitting for both sites accounts for the non-cubic site symmetry of the tin atoms. The isomer shifts are in the typical range observed for intermetallic tin compounds [16].

Fig. 8:

Experimental (data points) and simulated (colored lines) ¹¹⁹Sn Mössbauer spectra of the stannides REPd₃Sn₂ (RE=La–Nd) at room temperature.

Fig. 9:

Experimental (data points) and simulated (colored lines) ¹¹⁹Sn Mössbauer spectra of the stannides REPt₃Sn₂ (RE=Ce–Nd) at T=78 K.

Although we have observed well-resolved ¹¹⁹Sn spectra, it is difficult to assign the two sub-signals to the two individual tin sites. In ¹¹⁹Sn Mössbauer spectroscopy, a higher isomer shift is compatible with a higher s electron density at the tin nucleus. This has nicely been demonstrated for the stannides CaTSn₂ (T=Rh, Pd, Ir) [46], where the palladium compound shows the highest isomer shift, similar to the results for the series reported herein. The correct assignment of the tin sites for the series REPd₃Sn₂ and REPt₃Sn₂ deserves precise electronic structure calculations.