In search of new candidates for ultra-hard materials: the ternary BC 3 N 3 stoichiometry R. Weihrich, S.F. Matar * , E. Betranhandy Institut de Chimie de la Matie`re Condense´e de Bordeaux, ICMCB-CNRS 87, Avenue du Docteur Albert Schweitzer, F-33608 Pessac Cedex, France Abstract Starting from formerly investigated graphitic like C 3 N 4 , selective substitution of nitrogen with boron led to model structures for the experimentally observed BC 3 N 3 stoichiometry. Investigations of the geometry optimisation and of the electronic properties were carried out using pseudo potential and full potential computations in the framework of the local density functional theory for the two and three dimensional structures (2D and 3D). They lead to propose a precursor (2D), a b-structure and a new ultra hard rhombohedral compound with a hardness (B 0 , 358 GPa) that reaches the range of formerly studied BC 2 N structures built from hexagonal and cubic diamond. q 2003 Elsevier Ltd. All rights reserved. 1. Introduction The on going interest in boron carbonitrides with general composition B x C y N z arises from the difficulty to obtain new materials for abrasives, heat sinks and protective coating applications. Ternary systems with a diamond-like structure in which some of the carbon atoms are replaced with nitrogen and boron are expected to show the same interesting properties found in diamond and cubic boron nitride, such as large hardness, wide band gap and high melting points. As a consequence their possible applications could be found in several mechanical and electronic devices [1–3]. Moreover the low oxidation resistance of diamond, which is one of the most important drawbacks for its applications, might be improved in the boron-based hard materials. As a matter of fact diamond can only be used at around 870 K in air, while cubic boron nitride avoids the oxidation up to 1370 K [4,5]. To improve these properties mixed phases like BC 2 N were investigated. As a first ternary ‘B–C–N’ system we have investigated the bulk modulus and elastic constants of 3D model structures of BC 2 N within a density functional theory (DFT) framework. While graphitic BC 2 N had been prepared experimentally in the late eighties [6,7] by chemical vapour deposition, no hard 3D phase could be prepared from the graphitic precursor [1–3,17–20] due to limited solubility [21,22] or segregation in a mixture of diamond and c-BN [2,19,20,23]. The recently successful high pressure preparation of a heterodiamond single BC 2 N phase [10] now seems to prove our prediction that BC 2 N is one of the best candidates for a new ultra hard material. Our former theoretical investigations in the framework of the DFT have led to characterise the stability and the hardness of six novel BC 2 N structures obtained from a full geometry relaxation of the substituted diamond [8,11]. B 0 of the order of 400 GPa were found for to be in between the calculated values for diamond , 464 GPa and cubic BN (397 GPa). This encouraged us to investigate in the same way structures of BC 3 N 3 stoichiometry, that has recently been found experimentally [9] in systems prepared from organic reactants proposed as a graphite like boron cyanide ‘B(CN) 3 ’. Again model structures were proposed from IR and EELS spectroscopies, but complete structure determi- nation is done yet. The relevant feature is that upon heating the linearly connected B–(CN)–B compounds, a 2D network similar to that of graphitic C 3 N 4 [12] was found. The authors claim that the C 3 N 3 hetero cycles were connected by trigonal planar boron with B – C connections. With the study of the BC 3 N 3 stoichiometry the aim of this work in the same framework as formerly [8,11] is twofold: (I) to provide a description of the structure and electronic properties of the model boron tricyanide and (II) 0022-3697/03/$ - see front matter q 2003 Elsevier Ltd. All rights reserved. doi:10.1016/S0022-3697(03)00171-9 Journal of Physics and Chemistry of Solids 64 (2003) 1539–1545 www.elsevier.com/locate/jpcs * Corresponding author. E-mail address: matar@icmcb.u-bordeaux.fr (S.F. Matar).