Non-isovalent substitution in a Zintl phase with the TiNiSi type structure, CaMg1–xAgxGe [x = 0.13 (3)]

Single crystals of the title Ag-substituted calcium magnesium germanide, CaMg1–xAgxGe [x = 0.13 (3)] were obtained from the reaction of the corresponding elements at high temperature. The compound crystallizes with the TiNiSi structure type (Pearson code oP12) and represents an Ag-substituted derivative of the Zintl phase CaMgGe in which a small fraction of the divalent Mg atoms have been replaced by monovalent Ag atoms. All three atoms in the asymmetric unit (Ca, Mg/Ag, Ge) occupy special positions with the same site symmetry (.m.). Although the end member CaAgGe has been reported in an isomorphic superstructure of the same TiNiSi type, higher Ag content in solid solutions could not be achieved due to competitive formation of other, perhaps more stable, phases.

Single crystals of the title Ag-substituted calcium magnesium germanide, CaMg 1-x Ag x Ge [x = 0.13 (3)] were obtained from the reaction of the corresponding elements at high temperature. The compound crystallizes with the TiNiSi structure type (Pearson code oP12) and represents an Ag-substituted derivative of the Zintl phase CaMgGe in which a small fraction of the divalent Mg atoms have been replaced by monovalent Ag atoms. All three atoms in the asymmetric unit (Ca, Mg/Ag, Ge) occupy special positions with the same site symmetry (.m.). Although the end member CaAgGe has been reported in an isomorphic superstructure of the same TiNiSi type, higher Ag content in solid solutions could not be achieved due to competitive formation of other, perhaps more stable, phases.
Hence, the structure consist of a three-dimensional four-connected anionic [Mg 1-x Ag x Ge] networks with the Ca cations sitting in large channels (Fig. 1). The anionic network may be constructed from two-dimensional sheets, similar to those in black phosphorus and running perpendicular to the a-axis, which are linked along the a-direction to form one-dimensional ladders of edge-sharing four-rings and channels of eight-rings running along the b-direction. The TiNiSi type is known to be very versatile and shows remarkable structural and electronic flexibility (Landrum et al., 1998). Meanwhile, a large number of compounds with the TiNiSi type like CaMgGe can be rationalized as Zintl phases (Kauzlarich, 1996) according to the ionic formulation Ca 2+ (Mg 2+ Ge 4-). Zinlt phases are known to be very sensitive to the electron count (Ponou et al., 2007).
But, because of the above mentioned flexibility of the TiNiSi type, non-isovalent substitutions was expected without major structural distortion. Thus, since CaAgGe crystallizes in the isomorphic (i 3 ) superstructure of the TiNiSi type with a tripling of the a-axis (Ponou & Lidin, 2008), a wide stoichiometry breadth was expected in the system CaMg 1-x Ag x Ge. Eventually, the reaction of different starting mixtures with x = 1/4, 1/2, and 3/4, yielded almost the same composition (x = 0.10 -0.13) within 3σ standard deviation. This indicates a narrow homogeneity range, meaning that in this class of materials, the inherent electronic rigidity of the Zintl phase may be conflicting with the otherwise remarkable flexibility of TiNiSi type. CaMg 0.87 (1) Ag 0.13 (1) Ge is the Ag-richest phase that was structurally characterized. The unit-cell volume here (V = 280.99 (1) Å 3 ) is quite similar to that of the non-substituted phase CaMgGe (V = 280.89 Å 3 ), though the size of Mg is significantly larger than [Pauling's (1960) radii: Mg 1.600 Å, Ag 1.440 Å]. But, it should be noted that the later cell parameters were determined with much higher standard deviation (Eisenmann et al., 1972). In a reinvestigation of the CaMgGe structure (not reported), no indications of any superstructure were found.

Experimental
Three different mixtures of the elements (all from ABCR GmbH, Karlsruhe, Germany) Ca (granule, 99.5%), Mg (pieces, 99.9%), Ag (60m powder, 99.9%), and Ge (50m powder, 99.999%), with compositions along the CaMg 1-x Ag x Ge series with x = 1/4, 1/2, and 0.75 were loaded in a Niobium ampoules (approx. 9 mm diameter and 30 mm length) which were sealed on both ends by arc-melting and, in turn, enclosed in evacuated fused silica Schlenk tube to protect the former from air oxidation at high temperature. The ampoules were heated at 1273 K for 2 h, and cooled at a rate of 6 K/h to 923 K, where they are annealed for 24 h, then cooled down to room temperature by turning off the furnace. Semi quantitative EDX supplementary materials sup-2 analysis of the single crystals confirmed the presenced of the four elements, and no eventual contaminant at the detection limit could be observed.

Refinement
The refinement was straightforward, the full occupancies for all sites were verified by freeing the site occupation factor for an individual atom, while keeping that of the other atoms fixed. This proved that all positions but one, the Mg site, were fully occupied. The refined occupancy at Mg assigned position was higher than 100%, indicating a mixing with heavier element. Therefore, this position (labelled Mg/Ag) was modelled as a statistical mixture of Mg and Ag, and refined as 86.9 (1)% Mg and 13.1 (1)% Ag. Fig. 1. : A perspective view of (I) with displacement ellipsoids drawn at 95% probability level. Ca, Mg/Ag, and Ge atoms are drawn as grey crossed, orange and blue spheres, respectively.

Figures
calcium magnesium silver germanide Refinement. Refinement of F 2 against ALL reflections. The weighted R-factor wR and goodness of fit S are based on F 2 , conventional R-factors R are based on F, with F set to zero for negative F 2 . The threshold expression of F 2 > σ(F 2 ) is used only for calculating Rfactors(gt) etc. and is not relevant to the choice of reflections for refinement. R-factors based on F 2 are statistically about twice as large as those based on F, and R-factors based on ALL data will be even larger.