Magnetic properties of RE₁₀TCd₃ (RE = Ho, Er, Tm, Lu; T = Fe, Co, Ni, Ru) and ⁵⁷Fe Mössbauer spectroscopic data of Y₁₀FeCd₃

2.1 Synthesis

Starting materials (all with stated purities better than 99.5 %) for the synthesis of the RE₁₀TCd₃ samples were pieces of the rare earth elements (Smart Elements), ruthenium (Allgemeine Gold- und Silberscheideanstalt, Pforzheim), nickel powder (Merck), cobalt and iron pieces (Alfa Aesar), and a cadmium rod (Sigma-Aldrich). The elements were weighed in the ideal atomic ratios and sealed in niobium ampoules [13] under an argon pressure of ca. 700 mbar. The argon was purified with titanium sponge (870 K), silica gel and molecular sieves. The ampoules were placed in a water-cooled sample chamber of a high-frequency furnace (Hüttinger Elektronik, Freiburg, type TIG 1.5/300) [14]. They were rapidly heated to ca. 1500 K and kept at that temperature for 2 min. The temperature was then directly reduced to 1300 K and kept for another 45 min, followed by cooling to 900 K within 30 min and further annealing at that temperature for 2 h. The temperature was controlled through a Sensor Therm Methis MS09 pyrometer with an accuracy of ±30 K. The niobium ampoules were then sealed in evacuated silica tubes and annealed at 873 K for 5 d in a muffle furnace to achieve homogeneity. Finally, the samples were separated from the niobium tubes by mechanical fragmentation. The RE₁₀TCd₃ phases are stable at air over month.

2.2 X-ray diffraction

As a purity check, all polycrystalline RE₁₀TCd₃ samples were 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 experimental patterns matched with simulated ones [15], taking the positional parameters of the previously refined structures. The samples studied herein were pure on the level of powder X-ray diffraction.

2.3 Physical property measurements

Magnetic susceptibility and heat capacity investigations were performed with a Quantum Design Physical-Property-Measurement-System (PPMS) using the VSM (Vibrating Sample Magnetometer), ACMS and heat capacity options, respectively.

For the VSM measurements approximately 20 mg of the powdered samples was packed in polypropylene capsules and attached to the sample holder rod. The sample was investigated in the temperature range of 2.5–305 K and with magnetic flux densities up to 80 kOe (1 kOe = 7.96 × 10⁴ A m^–1).

For ac-MS measurements 67.754 mg of the powdered Ho₁₀RuCd₃ sample was enclosed in a thin-walled gelatine capsule. These investigations were performed in the temperature range of 5–100 K with internal frequencies from 41 to 9999 Hz and different magnetic field strengths up to 20 kOe.

The heat capacity of Ho₁₀RuCd₃ was measured with a 2.350 mg piece of the sintered tablet, which was fixed to a pre-calibrated heat capacity puck using Apiezon N grease and investigated between 2.5 and 300 K.

2.4 ⁵⁷Fe Mössbauer spectroscopy

A ⁵⁷Co/Rh source was available for the ⁵⁷Fe Mössbauer spectroscopic investigation. The Y₁₀FeCd₃ sample was placed in a PMMA container and the measurement was run in the usual transmission geometry at ambient temperature. Fitting of the spectrum was performed with the normos-90 program system [16].

3.1 Crystal chemistry

The intermetallic cadmium compounds RE₁₀TCd₃ [11] crystallize with a ternary ordered version of the Co₂Al₅ type [17]. The three crystallographically independent rare earth atoms occupy the aluminum sites, while the transition metal and cadmium atoms reside on the Co2 and Co1 sites, respectively. The crystal chemical details have been discussed in [11]. Herein we briefly focus on the structural details that are relevant with respect to the magnetic characterization.

The RE₁₀TCd₃ structures contain three different polyhedral building units (Fig. 1). The RE3 atoms have slightly distorted icosahedral coordination by six cadmium and six RE1 atoms. These RE3@Cd₆RE1₆ icosahedra are condensed via common triangular faces to infinite polyhedral chains along c, resulting in the motif of a rod packing. The space between these rods is filled by a second rod type which consists of condensed empty RE₆ octahedra and transition metal filled T@RE₆ trigonal prisms. For a detailed discussion of the interatomic distances we refer to [11].

Fig. 1:

The crystal structure of the RE₁₀TCd₃ phases in polyhedral presentation.

Figure 2 shows the two different layers of rare earth atoms as a projection onto the ab plane. As a consequence of the hexagonal space group symmetry, the rare earth substructure shows several equilateral, triangular units that might cause frustration in magnetic ordering. The layer around z ≈ 0 of the RE3 and RE1 atoms is slightly puckered, while the RE2 atoms are located on the mirror plane at z = 1/4. Due to the 6₃ screw axes, these motifs repeat around z ≈ 1/2 and z = 3/4.

Fig. 2:

Projection of the rare earth layers of the RE₁₀TCd₃ structures onto the ab plane. The heights of the atoms are indicated. Both layer types repeat at z ≈ 1/2 and z = 3/4 as a consequence of the 6₃ screw axis.

3.2 Magnetic properties

At first we discuss the magnetic properties of the three lutetium compounds Lu₁₀TCd₃ (T = Fe, Co, Ru). The temperature dependence of the magnetic susceptibility measured at 10 kOe is shown in Fig. 3. The overall susceptibility values are small. The susceptibilities of Lu₁₀CoCd₃ and Lu₁₀RuCd₃ are almost independent of temperature down to about 100 K. This is compatible with Pauli paramagnetism originating from the conduction electrons. The room temperature values are 5.6(2) × 10^–4 and 4.6(2) × 10^–4 emu mol^–1 for Lu₁₀CoCd₃ and Lu₁₀RuCd₃, respectively. The upturns at low temperatures are due to tiny amounts of paramagnetic impurities. Also for Lu₁₀FeCd₃ we can conclude Pauli paramagnetic behavior, however, with higher susceptibility values and stronger temperature dependence (Fig. 3). This can be ascribed to iron containing impurity phases at the grain boundaries. Nevertheless, the absolute susceptibility values are two orders of magnitude lower than those of the paramagnetic holmium, erbium and thulium compounds discussed below. We can thus conclude that the transition metal atoms in the RE₁₀TCd₃ phases carry no magnetic moment. Similar behavior has been reported for the aristotype Co₂Al₅ [18].

Fig. 3:

Temperature dependence of the magnetic susceptibility of Lu₁₀TCd₃ (T = Fe, Co, Ru) measured at 10 kOe.

The high-temperature parts of the susceptibility data of all remaining compounds could be fitted with the Curie-Weiss law. The resulting experimental magnetic moments and Weiss constants are listed in Table 1. Most experimental moments are close to the theoretical values of the RE³⁺ free ion values, indicating stable trivalent rare earth elements. The enhanced value observed for Ho₁₀FeCd₃ is again a consequence of iron-containing impurity phases and/or elemental iron at the grain boundaries. Such a behavior is frequently observed for iron containing intermetallics.

Except for Ho₁₀FeCd₃ and Ho₁₀NiCd₃, all Weiss constants have low values close to zero. This is somehow in contrast to the magnetic ordering temperatures and is indicative of compensating magnetic interactions in the paramagnetic regimes in the case of the Ho and Er compounds. For instance, compensating antiferromagnetic and ferromagnetic interactions can result in small values for the Weiss constant. Nevertheless we can exclude magnetic frustration, since this would cause higher Weiss constants. According to Schiffer and Ramirez, geometrical frustration is evident if the frustration index (the quotient of the amount of the Weiss constant and the Néel temperature) is higher than 10 [19]. Our compounds have much smaller indices.

Exemplarily we describe the magnetic properties of Tm₁₀RuCd₃ and Ho₁₀RuCd₃ herein in detail, whereas the important parameters of the remaining compounds are listed in Table 1. The temperature dependence of the magnetic and inverse magnetic susceptibility (χ and χ^–1 data) of Tm₁₀RuCd₃ are displayed in the top panel of Fig. 4 (10 kOe data). The Néel temperature of 9.8(5) K was deduced from a zero-field-cooled (ZFC)/field-cooled (FC) measurement at 100 Oe (inset of Fig. 4). This is in line with the Weiss constant of –4.8(5) K. Similar behavior has been observed for Tm₁₀TCd₃ with T = Fe, Co and Ni (Table 1).

Fig. 4:

Magnetic properties of Tm₁₀RuCd₃: (top) Temperature dependence of the magnetic susceptibility and its reciprocal (χ and χ^–1 data) measured at 10 kOe. The inset displays the magnetic susceptibility in ZFC/FC mode at 100 Oe; (bottom) Magnetization isotherms of Tm₁₀RuCd₃ at 3, 8, 20 and 50 K.

Magnetization isotherms of Tm₁₀RuCd₃ taken at 3, 8, 20 and 50 K are presented in the bottom panel of Fig. 4. At 20 and 50 K, well above the magnetic ordering temperature we observe linear field dependence as expected for a paramagnet. An s-shaped field dependence is observed for the 3 K isotherm. The spin reorientation (antiparallel to parallel realignment) takes place at a critical field strength of 23(3) kOe, underlining the antiferromagnetic ground state. The saturation moment at 3 K at the highest achievable field of 80 kOe is 3.9(1) μ_B per thulium atom, significantly lower than the theoretical value of 7 μ_B. The other three thulium compounds (Table 1) show comparable values for the saturation magnetization.

Data of the magnetic measurements of Ho₁₀RuCd₃ are shown in Fig. 5. The ZFC measurement (top panel of Fig. 5), carried out at an external field strength of 10 kOe, manifests antiferromagnetic ordering below approximately 50 K. The Weiss constant of 2.2(5) K was deduced from this measurement. ZFC/FC data at 50, 100 and 500 Oe are shown in the middle panel of that figure. We observe a first broad transition at T_N = 46.5(5) K and a second anomaly below 20 K. The curves show a steep increase in this temperature range and the splitting between the ZFC/FC measurements is smaller with increasing field strength. This is indicative of weak intrinsic magnetic interactions or due to tiny amounts of impurities.

Fig. 5:

Magnetic properties of Ho₁₀RuCd₃: (top) Temperature dependence of the magnetic susceptibility and its reciprocal (χ and χ^–1 data) measured at 10 kOe; (middle) magnetic susceptibility in zero-field-cooled/field-cooled (ZFC/FC) mode at 50, 100 and 500 Oe; (bottom) magnetization isotherms of Ho₁₀RuCd₃ at 3, 40, 60 and 100 K.

A pronounced splitting of ZFC/FC curves is characteristic for ferromagnetic compounds or for short-range interactions (spin-glass phenomenon), which might be accompanied by magnetic frustration. This is in contrast to the antiferromagnetic transition at 46.5 K, which becomes more pronounced at higher field strengths, underlining the intrinsic nature. The magnetization isotherms of Ho₁₀RuCd₃ at 3, 40, 60 and 100 K up to maximum field strength of 80 kOe are presented in the bottom panel of Fig. 5. The isotherms in the paramagnetic range (60 and 100 K data) show a linear increase. The 3 K isotherm shows a distinct upturn above approximately 45 kOe, which can be attributed to a field induced spin-reorientation confirming the antiferromagnetic ground state, similar to Pr₂Pd₂Mg [20]. However, most metamagnetic materials show weak ferromagnetism in the high-field range and almost no hysteresis. The maximum magnetization of Ho₁₀RuCd₃ at 3 K and 80 kOe is 3.7(1) μ_B per holmium atom, significantly smaller than the theoretical value of 10 μ_B. All other holmium compounds of this series show higher saturation magnetization. A further metamagnetic step for Ho₁₀RuCd₃ at higher field strengths cannot be excluded.

In view of the two magnetic transitions observed for Ho₁₀RuCd₃ in the low-temperature regime, we also measured the temperature dependence of the specific heat from 2 to 300 K without an applied magnetic field (Fig. 6). The measurement shows a pronounced transition at 45.4(5) K, in good agreement with T_N = 46.5(5) K determined from the susceptibility data. No further anomalies are evident towards lower temperatures, excluding a second intrinsic long-range ordering at lower temperatures. This is in line with the suppression of the 20 K anomaly by higher field strengths.

Fig. 6:

Heat capacity of Ho₁₀RuCd₃ in the temperature range of 2.0–300 K without an applied field. The inset shows the magnified low-temperature area to highlight the magnetic ordering at 45.5 K.

The complex magnetic behavior of Ho₁₀RuCd₃ was additionally studied by ac-susceptibility data. The ZFC χ′ (ω, T) susceptibilities at frequencies ranging from 41 to 9999 Hz and an amplitude of 8 Oe are plotted in Fig. 7. In agreement with the dc-susceptibility data we observe two ordering phenomena at 45 and 15 K. The absence of a frequency dependence excludes dynamic processes. This is in line with the χ″ (ω, T) susceptibilities (Fig. 7, middle). The maximum value of χ″ (ω, T) corresponds to approximately one sixtieth of χ′ (ω, T). For common spin glasses a ratio of one tenth is usually observed [21]. We can thus conclude simple antiferromagnetic behavior for the first magnetic transition, whereby one possible reason for the broadening of these peaks is a weak homogeneity range of the sample. The anomaly below 20 K can most likely be attributed to a tiny impurity contribution. A wide-range magnetic phase transition can be excluded due to the absence of an anomaly in the heat capacity. Short-range magnetic interactions usually accompany dynamic processes, which are almost not evident in the ac-susceptibility measurements.

Fig. 7:

ac-Susceptibility of Ho₁₀RuCd₃: (top) Temperature dependence of the magnetic susceptibility χ′ (ω) recorded after ZFC with internal frequencies ranging from 41 to 9999 Hz and an amplitude of 8 Oe; (middle) Magnified temperature area around the magnetic ordering temperatures; (bottom) Temperature dependence of the out-of-phase susceptibility χ″ (ω) around the second anomaly.

3.3 ⁵⁷Fe Mössbauer spectroscopy

The room temperature ⁵⁷Fe Mössbauer spectrum of Y₁₀FeCd₃ along with a transmission integral fit is presented in Fig. 8. In agreement with the crystallographic data the spectrum could be well reproduced with a single iron site at an isomer shift of δ = –0.38(1) mm s^–1, subjected to quadrupole splitting of ΔE_Q von 0.82(1) mm s^–1 (a consequence of the trigonal prismatic yttrium coordination). The isomer shift value is within the expected range for iron in intermetallic compounds [22]. The experimental line width is Γ = 0.29(1) mm s^–1.

Fig. 8:

⁵⁷Fe Mössbauer spectrum of Y₁₀FeCd₃ at room temperature.