Beryllium Boride Carbide DOI: 10.1002/anie.200705023 The h 6 ,h 1 -Coordination of Beryllium Atoms in the Graphite Analogue BeB 2 C 2 ** KathrinHofmann,XavierRocquefelte,Jean-FrançoisHalet,CarstenBähtz,andBarbaraAlbert* DedicatedtoDr.JosephBauerontheoccasionofhis65thbirthday The coordination of beryllium ions in homoleptic beryllocene, [Be(C 5 H 5 ) 2 ], has for decades been the subject of debate and theoretical as well as experimental investigations. It was not until quite recently that Schurko and co-workers [1] were able to show beyond doubt that Be is present in [Be(C 5 H 5 ) 2 ] in the h 5 ,h 1 -coordination mode, which is consistent with the octet rule for Be. We have now found an analogous disposition for Be in a solid-state compound, namely BeB 2 C 2 , in which six- membered rings of boron/carbon (B/C) layers coordinate to beryllium atoms in a h 6 ,h 1 fashion. BeB 2 C 2 is the first boride carbide with slipped 6 3 B/C layers as in graphite. Initially, we were unable to determine its structure with diffraction methods; we thus solved the structure by means of electron energy loss spectroscopy (EELS)—whereby a combination of theoretical and exper- imental methods was indispensable for the analysis of the energy loss near-edge structure (ELNES)—and further refinement was achieved by X-ray powder diffractometry. This beryllium diboride dicarbide is one of two com- pounds that were described in the Be–B–C system about forty years ago. [2] But although this substance was accessible as a single crystal (and its diffraction diagram was indexed in the Laue class 6/mmm, a = 1082 pm, c = 618 pm), its structure had not been resolved before now. Indications from EELS [3] that BeB 2 C 2 is isostructural to LiBC could not be confirmed from X-ray powder diffractometry. LiBC and MgB 2 C 2 crystallize in layer structures in which the boron and carbon atoms form covalent, two-dimensional, planar (in analogy to the hexag- onal boron nitride) or slightly corrugated networks of condensed six-membered rings. [4,5] Similar structures are interesting in the context of the discussion of high-temper- ature superconductors, [6] since they are topologically closely related to MgB 2 . [7] We were recently able to show that it is possible to distinguish between several possible structural models for MB 2 C 2 compounds (M = Ca, La) by comparing the exper- imental fine-edge structures of the B K ionization edges with those obtained by DFT calculations. [8] This result was later confirmed by independent DFT calculations. [9] The fine-edge structure of the B K edge in borides and boride carbides is highly variable with respect to weak structural and electronic influences. [10] The work described herein derives an otherwise inacces- sible, coherent structural model for BeB 2 C 2 by calculating the energy loss near-edge structures (obtained with the WIEN2k software) [11] for a number of atomic distributions. The structure was then refined on the basis of X-ray powder diffractograms and confirmed by theoretical quantum calcu- lations. We were able to obtain BeB 2 C 2 in the form of a crystalline powder at a temperature of 1950 8C. With EELS, the Be/B/C ratios were established to be 1:2:2. The diffractograms, which were obtained by high-resolution Guinier diffractometry and Cu Ka1 irradiation (flat specimen, transmission) as well as on the synchrotron (Hasylab, DESY, l = 113.96101 pm, Ge(111) double monochromator, Ge(111) analyzer, capillaries, Debye–Scherrer geometry), did not permit us to find a solution for the structure. [12] Although it was possible to index the diffractograms for the first time in an orthorhombic crystal system similarly to those of magnesium diboride dicarbide (space group no. 64, Cmce, a = 1083.7, b = 939.6, c = 613.6 pm; compare MgB 2 C 2 : a = 1092.2, b = 946.1, c = 745.9 pm), the distortions of the network obtained by Rietveld refinement of the analogous structural model did not make sense, and the difference Fourier maps for the structure model without cations showed no atomic positions for the beryllium atoms. The measured B K ionization edges of the compounds LiBC, MgB 2 C 2 , and BeB 2 C 2 are very similar to one another (Figure 1). If one calculates the B K fine-edge structures of LiBC and MgB 2 C 2 on the basis of structures described in the literature and then compares these with the experimental ELNES, the agreement is very convincing (Figure 2a,b). On the other hand, the structural models of LiBC and MgB 2 C 2 do not allow a correct simulation of the experimental B K ELNES of BeB 2 C 2 (Figure 2c). As soon as the B/C layers are slipped with respect to each other, however, to make a B/C arrange- ment analogous to that of graphite, the agreement becomes striking between the ELNES calculated on the basis of this structural model and the experimental one. This is true for the B K as well as for the C K ionization edges (Figure 3a,b). One [*] Dr. K. Hofmann, Prof. Dr. B. Albert Eduard-Zintl-Institut für Anorganische und Physikalische Chemie, Technische Universität Darmstadt Petersenstrasse 18, 64287 Darmstadt (Germany) Fax: (+ 49)6151-166-029 E-mail: albert@ac.chemie.tu-darmstadt.de Dr. X. Rocquefelte, [+] Prof. Dr. J.-F. Halet Sciences Chimiques de Rennes, UMR 6226 CNRS-UniversitØ de Rennes 1 (France) Dr. C. Bähtz Hasylab/Desy (Germany) (now at ESRF, France) [ + ] Present address: Institut des MatØriaux Jean Rouxel, UMR 6502, CNRS-UniversitØ de Nantes (France) [**] We thank the Deutsche Forschungsgemeinschaft for financial support and Dr. K. Schmitt for preparatory work. Angewandte Chemie 2301 Angew. Chem. Int. Ed. 2008, 47, 2301 –2303 # 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim