High-pressure synthesis of SmGe3

: The new samarium germanide SmGe 3 is obtained by high-pressure high-temperature synthesis of pre-reacted mixtures of samarium and germanium at a pressure of 9.5 GPa and temperatures between 1073 and 1273 K. SmGe 3 decomposes at 470(5) K into SmGe 2 , α -Sm 3 Ge 5 and a hitherto unknown phase. SmGe 3 exhibits a superstructure of the cubic Cu 3 Au-type. Transmission electron microscopy measurements of crystalline particles and prepared lamellae indicate a high density of defects on the nano-scale. Selected area electron diffraction and elaborate X-ray powder diffraction measurements consistently indicate a 2 a 0 × 2 a 0 × 2 a 0 superstructure adopting space group Fm 3 m with a = 8.6719(2) Å.


Introduction
The elemental semiconductors silicon and germanium form a rich variety of binary phases with electropositive partners of the alkaline, alkaline-earth or rare-earth metal groups, and their chemical bonding as well as their electron count is often within the scope of the Zintl-Klemm concept [1,2]. The tetrel atoms form one-, two-or three-dimensional partial structures with two-center two-electron bonds, frequently yielding electron-precise electron balances. The application of high-pressure techniques has proven to be a productive strategy to grant access to structural patterns that violate classical electron counting rules.
Systematic studies of tetrel connectivities in MT 3 compounds (M: alkaline earth or rare-earth metal; T: Si, Ge)  disclose a number of structural motifs that go beyond the scope of the 8-N rule. Moreover, the phases formed with diamagnetic metal ions repeatedly exhibit superconductivity with strong electron-phonon coupling. Within the set of manufactured phases, the absence of a corresponding samarium compound is striking. Earlier density functional theory calculations [24] predict an atomic arrangement with space group P6 3 /mmc for the compound SmGe 3 . However, by high-pressure high-temperature synthesis, we obtain a Cu 3 Au-like SmGe 3 phase. The finding of superstructure reflections motivated further investigations by energy dispersive X-ray spectroscopy, Transmission electron microscopy (TEM) and extensive X-ray powder diffraction experiments.
2 Experimental section 2.1 Synthesis SmGe 3 was synthesized under high-pressure high-temperature conditions. All sample handling, except for high-pressure synthesis itself, was performed in argon-filled glove boxes (MBraun, H 2 O and O 2 < 0.1 ppm). The precursor mixture was prepared by arc melting of samarium (Lamprecht, 99.9%) and germanium (Chempur, 99.9999+%) in the ratio 1:3 plus 6% excess of samarium for the compensation of evaporation losses during heating. The material was ground and filled into BN crucibles before being transferred into MgO octahedra with an edge length of 18 mm. High-pressure high-temperature syntheses were conducted in a Walker-type module [25] for 30 min to 5 h at pressures between 9(1) to 9.5(10) GPa and temperatures from 1073(107) and 1273(127) K before quenching under load. The calibration of pressure and temperature has been realized prior to the experiments by the observation of resistance changes of bismuth [26] and thermocouplecalibrated runs, respectively.

Sample characterization
For metallographic analysis, samples were prepared by polishing with diamond powder disks (grain size 6, 3 and 0.25 μm) after embedding in paraffin. The investigation was realized with a Philips XL 30 scanning electron microscope (SEM) (LaB 6 cathode), comprising an EDAX Si(Li) detector for energy-dispersive X-ray spectroscopy (EDXS).
Dedicated to Professor Dr. Ulrich Müller on the occasion of his 80th birthday.
Differential scanning calorimetry (DSC) experiments were performed in a Netzsch DSC 404 C device (Netzsch-Gerätebau GmbH, Selb, Germany) operated with heating and cooling rates of 10 K/min under argon atmosphere using corundum crucibles.
Phase designation was conducted by X-ray powder diffraction experiments with a Huber Image Plate Guinier Camera G670 (Huber Diffraktionstechnik GmbH & Co. KG, Rimsting, Germany), using CuKα 1 radiation (λ = 1.54056 Å). X-ray diffraction experiments for structure refinement were realized with synchrotron radiation (λ = 0.20709 Å) at DESY Group PETRA III using beamline P02.1, and with a Stoe Stadi MP in Bragg-Brentano geometry. The diffractometer was equipped with a DECTRIS MYTHEN2 1K silicon strip detector and operated with Cu-Kα 1 radiation (λ = 1.54056 Å, curved germanium (111) Johann-type Monochromator). The sample with cylindrical shape was mounted on a zero background sample holder with one of the cross sectional area at its end oriented to the incident X-ray beam. For better data point statistics, the final powder pattern results from the sum of three individual intensity data sets, collected in the angular range 5.00°≤ 2θ ≤ 120°( scan step = 0.06°, time pro step 40 s, t total ≈ 24 h).
All crystallographic calculations including determination of diffraction peak positions as well as lattice parameter and structure refinements on basis of full diffraction profiles (Rietveld technique) were performed with the WinCSD program package [27].
Thin samples for the TEM study were prepared by the focused-ionbeam (FIB) lift-out technique. Thin cross-sections of micro-crystalline grains were extracted from a broken bulk fragment. Defined crystallographic oriented cross-sections were prepared perpendicular to the staking-faults visible at the fracture surface. The FIB lift-out technique was performed on a FEI Quanta 200 3D ion/electron dual beam device (FEI Company, Eindhoven, the Netherlands) equipped with an Omniprobe micro-manipulator (tungsten needle), and can be used both as a SEM and a scanning ion microscope (SIM). First, protecting Pt layers (24 × 2 µm thickness, 2 µm high) were deposited on selected parts (parallel and perpendicular to the c axis of microcrystalline arrays) using an acceleration voltage of 30 kV and a current of 0.1 nA. Each cross section (2 μm thickness) was prepared by applying a Ga-ion beam using an acceleration voltage of 30 kV and a current of 1-0.5 nA. The manufactured cut was transferred onto a copper Omniprobe TEM holder using the in-situ lift-out technique [3]. Finally, the cross section was thinned in several discrete steps down to a thickness of about 40 nm by applying an acceleration voltage of 30 kV with currents of 0.5-0.01 nA of the Ga-ion beam.
Magnetic susceptibility was measured using a polycrystalline sample of cylindrical shape (3.0 × 2.5 × 3.4 mm) on an SQUID magnetometer (MPMS XL-7, Quantum Design) between 1.8 and 300 K in external fields of 0.01 to 3.5 T. Electrical resistivity measurements were conducted in a temperature range from 1.8 to 300 K by using the four-point probes method with a Keithley DC current source and a Hewlett Packard nanovoltmeter inside a helium flow cryostat at zero magnetic field. The contacting of the sample with Cu wire was done inside the glove box with Ag paint.

Composition and properties
The new samarium germanide is synthesized by highpressure high-temperature treatment of pre-reacted Sm 25.4 Ge 74.6 mixtures. The average chemical composition of the product as determined by energy dispersive X-ray spectroscopy amounts to Sm 25.5(5) Ge 74.5 (5) or SmGe 2.92 (8) .
Differential scanning calorimetry measurements ( Figure 1) of SmGe 3 reveal a decomposition into SmGe 2 [28], Sm 2 Ge 5 [29] and a hitherto unknown phase at 470(5) K. The feature at 737(5) K is attributed to a reaction of the decomposition products into SmGe 2 [28] and Ge [30]. In full agreement with the ambient-pressure phase diagram [31], effects at 1030(5) K and 1103(5) K correspond to the peritectoid decomposition of SmGe 2 into Sm 2 Ge 3 and Ge as well as to the melting of the resulting eutectic, respectively.
The magnetic susceptibility χ measured in an external field of μ 0 H = 3.5 T between 1.8 and 300 K ( Figure 2) indicates van Vleck-type paramagnetic behavior, denoting the influence of the external field on the wave function and the transition towards excited states. Under consideration of the slope of the experimental data, the typical minimum of χ(T) for Sm 3+ compounds is expected to occur around 400 K. The electrical resistivity ρ at zero-field between 1.8 and 300 K (Figure 2, inset) denotes metallic behavior with a room temperature value ρ(300 K) of 122 µΩcm and an inflection, which is attributed to the reduced scattering of charge carriers in the magnetically ordered phase [32]. The ordered state is antiferromagnetic as indicated by the cusp at 23 K in χ(T).

Substructure
X-ray powder diffraction patterns of SmGe 3 evidence that the strongest reflections indicate a Cu 3 Au-type structure motif [33]. A similar atomic arrangement is also reported for analog phases MGe 3−x (x ≈ 0.15: M = Tb, Dy, Yb [34,35]; x = 0: M = Ce [14], Np [36], Pu [37] and U [38], as well as for SmSn 3 [39]), the corresponding tin compound of samarium. The crystal structure represents an ordered variety of an fcc lattice, with germanium atoms occupying the centers of cube faces and samarium being positioned on the vertices.
The refinement of a Cu 3 Au-type model using full diffraction profiles (Figure 3) results in a 0 = 4.337(3) Å with R(P) = 0.043 and wR(P) = 0.063. As some other isostructural phases exhibit germanium deficiencies, partial occupation of the Ge site has been tested. The refinement proceeds without improvement of the residuals so that the models for phases SmGe 3 and SmGe 2.84(2) yield R(P) = 0.0429 and wR(P) = 0.0629 in the substructure.

Superstructure
Closer inspection of the substructure refinement reveals significant extra reflections not being accounted for by the Cu 3 Au-aristotype, e.g., that at 2.36°. The finding points at a larger unit cell for SmGe 3 and motivated additional investigations of the crystal structure. For this purpose, annealed samples (55 h at 823 K and 9.5 GPa) are selected and cleaved in order to avoid peak broadening because of grinding. Nevertheless, transmission electron microscopy investigations of a selected lamella prepared by the focused ion-beam method still reveal the presence of extended structural defects in the microstructure (Figure 4a). SAED (Figure 4b-d) and X-ray powder diffraction patterns consistently evidence a 2a 0 × 2a 0 × 2a 0 superstructure (a 0 unit cell of the primitive cubic phase). Systematic absences are compatible with a face-centered   lattice. As the synchrotron X-ray diffraction experiment evidences significant broadening of the diffraction profiles upon grinding of the material, subsequent X-ray measurements are performed with as-grown ingots at the cost of reduced powder average. Thus, diffraction data of several samples had to be collected in order to check for reproducibility of the intensity information. The best result has been achieved for a sample having been manufactured at 9.5 GPa by heating the starting mixture to 1273 K for 30 min before annealing at 823 K for 5 h to increase the size of the crystal domains. The corresponding diffraction patterns evidence a number of weak reflections, which are absent in the non-annealed samples. In order to test if the extra lines have to be attributed to (potentially incommensurate) modulations of the superstructure, metallographic investigations have been performed. The measurements reveal the presence of a small amount of a side product with composition Sm:Ge ≈ 1:2. As the diffraction peak positions of the previously reported compounds with similar composition do not match the extra lines, a reference sample was synthesized at the same conditions as the 1:3 phase. The X-ray powder diffraction diagram of this component is compatible with a Pu 3 Pd 5 -type [40] crystal structure and fits the extra lines in the pattern of SmGe 3 . With these pieces of information at hand, a model for the superstructure of SmGe 3 was developed in space group Fm3m and refined (Tables 1 and 2 and Figure 5). Decrease to tetragonal symmetry (sg I4/mmm, a t = √2 a 0 , c t = 2 a 0 ) does not result in improvements of the residuals of the least squares refinements. Further details on an ordered variety with presumably monoclinic symmetry and on the Pu 3 Pd 5 -type phase will be reported elsewhere.
The atomic arrangement of the 2a 0 × 2a 0 × 2a 0 superstructure represents a new variety of the Cu 3 Au-type. The decrease of symmetry with respect to the lt-Cu 3 Au type [33] adds a degree of freedom to the positions of the germanium atoms. While germanium is coordinated by eight germanium and four metal atoms at the same  Atom  distance in the undistorted lt-Cu 3 Au arrangement (corresponding to 12 distances of a 0 /√2 = 3.067(2) Å for the substructure), the distances amount to d 1 (Ge-Ge) = 2.903(2) Å and d 2 (Ge-Ge) = 3.229(2) Å, while the contacts d(Sm-Ge) = 3.0682(1) Å remain essentially unchanged in the superstructure of SmGe 3 ( Figure 6). The symmetry relation between the sub-and the superstructure is concisely described by a group subgroup relationship [41,42] evidencing a klassengleiche transition of order 2. The relationship considering also some other fcc varieties is summarized in form of a small family tree (Figure 7) and the atomic arrangements are shown (Figure 8).
The interatomic distances d(Ge-Ge) of all structure types except Cu 3 Au fall into the range from 2.4 to roughly 3 Å (Figure 9). Within the series of isotypic DyGe 3 -type compounds, the distances d(Ge-Ge) remain more or less constant (distances d(M-Ge) decrease with increasing atomic number in accordance with the lanthanide contraction, not shown). Compounds, which are reported to adopt the lt-Cu 3 Au-type structure, exhibit unusually long interatomic distances d(Ge-Ge). This abnormality is attenuated by the symmetry breaking in the new SmGe 3type arrangement. The distortion induces a subdivision of the Ge-Ge distances into two groups. The shorter contacts, representing half of the distances d(Ge-Ge), fall now into the upper part of the range observed for DyGe 3 -type compounds and phases REGe 3+x (RE = Pr, Nd). Whether the underlying reason for the distortion is the formation of Ge-Ge bonds or the stereochemical activity of lone pairs located at the Ge − species (assuming the electron balance Sm 3+ [Ge − ] 3 ) will be the topic of future investigations by quantum chemical methods.