The thermal motion of atoms and functional groups is a key characteristic of any molecular crystal, and such motion derived from scattering experiments is conveniently visualised by means of thermal ellipsoids (the famous "ORTEP" drawings). Unfortunately, it is often impossible to obtain the underlying anisotropic displacement parameters (ADPs) for hydrogen atoms, due to their low X-ray scattering power, and sometimes no ADPs can be refined at all even for heavier atoms. In these cases, it would seem advantageous to estimate ADPs by first-principles techniques, and indeed such ab initio ORTEP plots have become available very recently. Here, we test this young method for a representative set of hydrogen-bonded molecular crystals: first, we study urea (CON2H4)as a well-known benchmark, then, its all-nitrogen analogue guanidine (CN3H5); finally, we move on to rubidium guanidinate (RbCN3H4) as a specimen with pronounced ionic interactions. For all three systems, ADPs have been obtained from density-functional theory (DFT) based phonon computations using the PHONOPY software. The results are compared with neutron-diffraction data as the experimental "benchmark" in this regard, and a critical discussion of experimental aspects is given. We observe excellent agreement between experiment and theory for the hydrogen-bonded systems urea and guanidine at low temperature, whereas high-temperature data for guanidine deviate visibly, and the more salt-like RbCN3H4 may suffer from a less-than-ideal description even at 12 K. Both are discussed in depth as there are possible solutions and directions for further research. Generally, the present results shine a favourable light on a future, more routine application of combined experimental/theoretical approaches in chemical crystallography.

We report about a source-code modification of the density-functional program suite VASP which greatly benefits from the use of graphics-processing units (GPUs). The blocked Davidson iteration scheme (EDDAV) has been optimized for GPUs and gains speed-ups of up to 3.39 on S1070 devices and of 6.97 on a C2050 device. Using the Fermi card, the code reaches an impressive 61.7% efficiency but does not suffer from any accuracy losses. The algorithmic bottleneck lies in the multiplication of rectangular matrices. We also give some initial thoughts about introducing a different level of parallelism in order to harness the computational power of multi-GPU installations.

We report on a source-code modification of the density-functional program suite VASP which benefits from the use of graphics-processing units (GPUs). For the electronic minimization needed to achieve the ground state using an implementation of the blocked Davidson iteration scheme (EDDAV), speed-ups of up to 3.39 on S1070 devices or 6.97 on a C2050 device were observed when calculating an ion–conductor system of actual research interest. Concerning the GPU specialty – memory throughput – the low double-precision performance forms the bottleneck on the S1070, whereas on Fermi cards the code reaches 61.7% efficiency while not suffering from any accuracy losses compared to well-established calculations performed on a central processing unit (CPU). The algorithmic bottleneck was found to be the multiplication of rectangular matrices. An initial idea to solve this problem is given.

Ternary sodium borosilicide, Na(8)B(74.5)Si(17.5), was newly synthesized by heating a mixture of sodium, silicon and amorphous or crystalline boron at 1073-1273 K. The crystal structure of the black hexagonal prismatic single crystal obtained at 1273 K was analyzed. The X-ray diffraction reflections of the crystal were indexed with hexagonal cell parameters a = 10.2392(3) Å and c = 10.9215(4) Å (space group P6(3)/mmc, No. 194). The structural formula could be represented as Na(8)(B(12))(6)Si(16)[BSi](1.5)[B(2)](0.5). B(12) icosahedra form a three-dimensional framework having small triangular and large hexagonal channels along the c axis. Chains of [-Si-(Si-Si)(3)-Si-] surrounded by Na atoms are located in the large channels. Trigonal prism cages for Na atoms and for atom pairs of Si-B or B-B are alternately arrayed in the small channels. The ratio of Si-B and B-B pairs in the cage is around 3:1.

The crystal structures of two new ternary phases, La4Ag10Mg3 and La4Ag10.3Mg12, were refined from X-ray single crystal diffraction data. La4Ag10Mg3 crystallizes in the Ca4Au10In3 structure type, an ordered variant of the binary Zr7Ni10 compound: orthorhombic, Cmce, oS68, a=14.173(5), b=10.266(3), c=10.354(3)Å, Z=4, wR2=0.0826, 676 F2 values, 50 variables. La4Ag10.3Mg12 represents a new structure type: orthorhombic, Cmmm, oS116-10.32, a=9.6130(3), b=24.9663(8), c=9.6333(2)Å, Z=4, wR2=0.0403, 1185 F2 values, 101 variables. The structural analysis of both compounds, highlighting a significant contraction of the Ag–Mg distances, suggests the existence of three-dimensional [Ag–Mg] networks hosting La atoms. LMTO calculations applied to La4Ag10Mg3 indicate that the strongest bonds occur for Ag–Ag and Ag–Mg interactions, and confirm the presence of a 3D∞[Ag10Mg3]δ− polyanionic framework balanced by positively charged La atoms.

The phase diagram of (Fe(1-x) Mn(x))(3)C has been investigated by means of density-functional theory (DFT) calculations at absolute zero temperature. The atomic distributions of the metal atoms are not random-like as previously proposed but we find three different, ordered regions within the phase range. The key role is played by the 8d metal site which forms, as a function of the composition, differing magnetic layers, and these dominate the physical properties. We calculated the magnetic moments, the volumes, the enthalpies of mixing and formation of 13 different compositions and explain the changes of the macroscopic properties with changes in the electronic and magnetic structures by means of bonding analyses using the Crystal Orbital Hamilton Population (COHP) technique.

The electronic structures and magnetic properties of MNCN (M = Fe, Co, and Ni) have been investigated by density-functional theory including explicit electronic correlation through an ad hoc Coulomb potential (GGA+U). The results evidence CoNCN and NiNCN as type-II anti-ferromagnetic semiconductors (that is, intralayer ferromagnetic and interlayer anti-ferromagnetic), in accordance with experimental observations. Just like the prototype MnNCN, the MNCN phases, with M = Ni and Co, thus resemble the corresponding MO monoxides with respect to their magnetic and transport properties. By contrast, FeNCN remains (semi)metallic even upon applying a strong Coulomb correlation potential. This, most probably, is in contradiction with its observed optical transparency and expected insulating behavior and points toward a serious density-functional theory problem.

In this contribution we present results of our theoretical studies of nitrogen mobility in solid M–Ta–O–N systems. Periodic supercell calculations at density-functional level have been performed to investigate the local structure of N-doped oxynitrides and the anion-diffusion mechanisms. The migration pathways and activation barriers were calculated using the nudged elastic band method with the climbing-image enhancement. We show that the defect migration is mainly caused by the diffusion of oxygen anions. The activation energy can be lowered by increasing the defect concentration, and it is, to a large extent, depending on the dopant size.

Single crystals of Ti2Rh6B were synthesized by arc-melting the elements in a water-cooled copper crucible under an argon atmosphere. The new silver-like compound with metallic luster crystallizes in space group Fm (3) over barm (no. 225) with a = 7.8191(5) angstrom, V = 478.05(5) angstrom(3), and Z = 4. The refinement converged to R-1 = 0.0169 and wR(2) = 0.0486 for all 96 unique reflections and 7 parameters. The structure can be described as a defect double perovskite, A(2)BB'O-6, where the A site is occupied by titanium, the B site by boron, the 0 site by rhodium but the B' site is vacant leading to the formation of perfectly octahedral Rh-6 clusters and BRh6 units. Rh-B, Rh-Ti, and Rh-Rh interactions are observed. According to density-functional (LMTO) electronic structure calculations, the strongest bonding a occurs for the Rh-B contacts, and the Rh-Rh bonding Within the clusters is more than two times stronger than in the BRh6 units.

A new rare-earth rich Zintl phase Yb11GaSb9 was synthesized by direct fusion of the corresponding elements, and large single crystals of the compound were obtained from high temperature flux synthesis. Its crystal structure was determined by single-crystal X-ray diffraction to be orthorhombic in the non-centrosymmetric space group Iba2 (No. 45), Z=4 (R1=3.24%, wR2=6.40%) with a=11.7257(12)Å, b=12.3204(13)Å, c=16.633(2)Å measured at 90(3)K. The structure belongs to the Ca11InSb9-type and can be viewed as built of isolated Sb4-tetrahedra centered by Ga, Sb-dimers and isolated Sb anions, which are separated by Yb2+ cations. Electron count according to the Zintl formalism suggests that the phase is electron-precise and charge-balanced, which is supported by the virtually temperature-independent magnetization for Yb11GaSb9. Electrical resistivity data from 2 to 400K confirm that Yb11GaSb9 is a small band-gap semiconductor with room temperature resistivity ρ298=45.1mΩcm, and low-temperature resistivity at 2K ρ2=1.9Ωcm. As such, Yb11GaSb9 and related compounds might be promising materials for thermoelectric applications, and currently, efforts to synthesize new members of this family and test their thermoelectric performance are under way.

The Zintl compound EuGe2 crystallizes in the trigonal space group P3̄m1 (No. 164) with the CeCd2-structure type. Its structure can be formally derived from the hexagonal AlB2-structure type by a strong puckering of the hexagonal layers. The chemical bonding in EuGe2 can be rationalized according to the Zintl concept as (Eu2+)(Ge1−)2, since the europium atoms are divalent and each germanium atom receives one additional valence electron. In that sense, EuGe2 is expected to be a closed-shell compound with semiconducting behavior. However, temperature dependent resistivity measurements show EuGe2 to be metallic. Subsequently, detailed crystallographic studies revealed the structure and the composition of EuGe2 to be free of defects and impurities, which, along with the confirmed divalent oxidation state of the europium atoms by means of magnetic measurements, make EuGe2 another example of a metallic Zintl phase. These results are in good agreement with the results of electronic structure calculations such as TB-LMTO-ASA (LDA) and FLAPW (GGA), which reveal non-zero DOS at the Fermi level.

Using crystal–chemical knowledge we show how to conceptionally partition the chemical bonding properties of any given inorganic crystal structure into a few fundamental types; their potentials are presented in analytical form. The parameterizations are based on the bond valence concept (BVC), the universal bonding energy-distance relationship (UBER), and the concept of absolute electronegativity and hardness (AEH). The approach has been implemented into the recently developed aixCCAD computer program, intended to establish solid-state atomistic simulations (molecular dynamics-type) in which the atomic charges are dynamical variables of freedom.

Based upon a partitioning and potential concept for the chemical bonding in solids, we illustrate a number of crystal–chemical simulations for various kinds of structures and bonding types on the picosecond time scale using the aixCCAD computer program. These include ionic/covalent materials (NaCl, ZnO, AlN), ternary oxides (LiAlO2 and its crystallographic phases), main-group (Ga, Al) as well as transition (3d, 4d, 5d) metals, various intermetallics (b.c.c.- and f.c.c.-like), as well as complex Fe/AlN nano composites. The simulations give access to detailed energetics, ionic mobilities, crystallographic structures, bulk moduli, and questions of chemical reactivity.

Nitrogen is well known to be crystallochemically different from neighboring atoms. For example, carbides and oxides are very common and quite easy to produce, but nitrides are more difficult. Using some examples from recent theoretical investigations in our group, we explore the differences between some model groups of compounds.Here we focus on two series of compounds where stoichiometrically similar carbides, nitrides, and phosphides are known: {CeC, CeN, CeP} and {FeC, FeN, FeP}. We will attempt to understand the differences in structures and magnetic properties within each of these series by bringing to bear our full arsenal of solid-state interpretive tools: band structures, densities-of-states (DOSs) and projected DOSs, and Crystal Orbital Hamilton Populations (COHPs).

Today clusters of PCs are powerful computing platforms offering an inexpensive alternative to traditional supercomputers. Besides the low price, the major reason for the emergence of PC clusters is the availability of a wealth of public domain as well as important commercial software on top of the Linux operating system. Furthermore, a PC cluster can be built in a modular fashion enabling a potential customer to perfectly meet his requirements. For instance, the number of nodes, the node types, and the network can be chosen. In this note, experiences with an SCI-based Linux cluster installed at the Computing Center of Aachen University of Technology are reported. Timing results for two different applications, a two-dimensional fast Fourier transform and a molecular dynamics simulation, are presented. Our experiences demonstrate the importance of a fast network when dealing with large-scale scientific applications on a Linux cluster.

Modern electronic structure theory is a valuable tool for the chemistry and physics of extended materials. This contribution illustrates some recent examples on how structures, bonding, and physical properties of solid nitrides containing transition metals, lanthanides, and main group elements may be theoretically accessed and, in selected cases, how their syntheses may be more rationally planned.

Layered nanocomposites made of metallic iron and aluminum nitride are subject to unexpected chemical reactions, resulting in a spontaneous formation of iron nitrides and a partial reduction to metallic aluminum. Since bulk thermochemical data are unable to rationalize the above finding, atomistic computer simulations based on the crystal-chemical atomic dynamics (CCAD) approach have been performed in the search for an explanation. The computational setup mimics a total number of about 1000 atoms moving over a time frame of 74 ps. When AlN molecules are sputtered on the iron surface under the experimental radio frequency (rf) conditions, the molecules are found to be chemically unstable upon hitting the surface, immediately breaking apart into individual atoms. Atomic nitrogen enters the Fe crystal to acquire quasi-octahedral coordination, leaving Al atoms behind on the surface. The reaction results in a stronger bonding of the nitride ion in the crystal compared to the covalently bonded nitrogen atom in the molecule. As a consequence, a small amount of Fe lattice expansion (2.5%) as well as a partial buildup of an iron/aluminum alloy is observed in the reaction zone near the surface of the bulk material.