Table of contents

Following the tradition of the ACCMS-1 held in Bangalore, India, in November 2001, and the ACCMS-2 in Novosibirsk, Russia, during July 14–16, 2004, this conference, held at the Institute of Physics, Chinese Academy of Sciences (CAS), Beijing, 8–11 September 2005, has been set up to promote research and development activities in computational materials science in Asian countries. Computational materials science has emerged as a distinct multidisciplinary branch of science whose relevance and importance has come from (a) the desire to have a microscopic understanding of complex materials and phenomena, (b) the need to design novel materials with a desired combination of physical, chemical and metallurgical properties, and (c) the possibility to describe the basic interatomic interactions in materials via appropriate quantum mechanical and statistical mechanical tools. With the unprecedented growth of computer power and the developments of efficient and smart algorithms and codes, it is now possible to do large scale simulations of real materials with increasing complexity. A synergy amongst a wide variety of disciplines such as physics, chemistry, metallurgy, geology, biology, computer science and information technology is gradually coming to a reality due to advances in computer simulations. What follows here are the written ACCMS-3 proceedings based on the presentations, oral and poster, of the research pursued by the various participating groups, which cover topics, such as density functional theory-based methods, Monte Carlo, molecular and lattice dynamics simulations, tight-binding and effective medium approaches, coarse graining and mesoscopic modeling, continuum and quasi-continuum approaches, etc and their applications to different materials.

ACCMS-3 was chaired by Professor Enge Wang, and co-chaired by professors B.L. Gu, Y. Kawazoe and G.P. Das, and organized and supported by the Institute of Physics (Beijing), Chinese Academy of Sciences, and the Department of Physics, Tsinghua University. In addition, financial supports from Chinese Academy of Sciences and National Natural Science Foundation of China (NSFC) are gratefully acknowledged.

The conference was held at the Institute of Physics, CAS, in its newly built conference hall. The excellent facilities and the competent support of the professional staff, headed by Dr Jian-tao Wang, created a genial environment for scientific discussion and exchange. Their invaluable assistance is also gratefully acknowledged.

The major generalization of the existing theory of clathrate hydrates, so that it can account for phenomena such as multiple occupancy of individual cages and mutual guest-host couplings and guest-guest interaction, are suggested. The new model allows taking into account the influence of guest molecules on the host lattice. Atomistic modeling of structural, dynamical and thermodynamic properties of ices and different hydrates at high pressures and a range of temperature were performed. The influence of guest molecules (argon, methane and xenon) on the host lattice of hydrate of cubic structures I and II was investigated. Results of these calculations agree with known experimental data.

The ability of metal beta-diketonates to pass into gas phase without decomposition is widely used in the processes involving the deposition of various coatings by CVD. In order to optimize the CVD method (e.g the choice of the precursors etc.) the study of volatility of metal-beta-diketonates are needed. In this work the acetylacetonates of III-valent metals: Al(Acac)₃, Cr(Acac)₃, Fe(Acac)₃ and Ir(Acac)₃ were chosen for calculation of sublimation enthalpy in the atom-atomic molecular mechanics approximation. The different types of interaction potential is considered: the Buckingham potential (6-exp), the Lennard-Johns potential (6-12) and modified Buckingham potential (M-6-exp), the last two potentials are used in a UFF framework. The most reliable one is found to be the Buckingham potential (6-exp).

In an earlier paper, the second author of this letter proved an algorithm for computing the symmetry of molecules. He applied his algorithm to calculate the symmetry of the smallest fullerene C₂₀. In this letter, we improve the mentioned algorithm to compute the automorphism group of the fullerene C₆₀ with connectivity and geometry of I_hsymmetry point group.

DNA nanocompartment is a typical DNA-based machine whose function is dependent of molecular collective effect. Fundamental properties of the device have been addressed via electrochemical analysis, fluorescent microscopy, and atomic force microscopy. Interesting and novel phenomena emerged during the switching of the device. We have found that DNAs in this system exhibit a much steep melting transition compared to ones in bulk solution or conventional DNA array. To achieve an understanding to this discrepancy, we introduced DNA-DNA interaction potential to the conventional Ising-like Zimm-Bragg theory and Peyrard-Bishop model of DNA melting. To avoid unrealistic numerical calculation caused by modification of the Peyrard-Bishop nonlinear Hamiltonian with the DNA-DNA interaction, we established coarse-gained Monte Carlo recursion relations by elucidation of five components of energy change during melting transition. The result suggests that DNA-DNA interaction potential accounts for the observed steep transition.

We present here a calculation of the configuration averaged optical conductivity of random binary alloys CuAu and AgAu. Our formulation is based on the augmented space formalism proposed by Mookerjee [J. Phys. C : Solid State Phys. 6 1340 (1973)] and the optical conductivity is obtained directly through a recursive procedure suggested by Viswanath and Müller [''The user friendly recursion method'', Troisieme Cycle de la Physique, en Suisse Romande (1993)].

A systematic theoretical study of the spectra of the electronic excited states of molecules is carried out by means of the first-principles Green's function methods based on the many-body perturbation theory. A merit of these methods is that all excitation spectra can be obtained accurately all at once together with the information on the corresponding quasiparticle wave functions. One-particle excitation spectra are obtained by the GW approximation, while the two-particle excitation spectra are obtained by starting from the GW approximation and solving the Bethe-Salpeter equation for the T-matrix. Here we present some of our recent results of CO and C₂H₂ molecules by using this state-of-the-art calculation. The results are in excellent agreement with experiments.

A doped electron or hole will exist in a self-trapping state even in a π-conjugated oligomer. It can be excited by absorbing photons. Two possible optical-excitation modes, which are respectively called low-energy excitation and high-energy excitation are demonstrated in this paper. It is found that the high-energy excitation will result in a reverse polarization of the molecule. The polarizability is calculated and it is obtained that the strength of the reverse polarization is dependent upon the conjugation-length of the molecule.

We have studied the electronic structure of the ferromagnetic double perovskites Sr₂CrReO₆, Sr₂CrWO₆ and Ba₂FeReO₆ by means of a full-potential linear muffin-tin orbital density-functional method. Our scalar-relativistic calculations predict these compounds to be half-metallic with a total magnetic moment of 1, 2, and 3 µ_B respectively. However, when the spin-orbit coupling is included, the 5d transition Re and W ions exhibit substantial unquenched orbital magnetic moments, resulting in a significant increase of the total magnetic moment. The half-metallic gap turns into a pseudo-gap in Sr₂CrReO₆ and Ba₂FeReO₆ when the spin-orbit coupling is included whereas Sr₂CrWO₆ remains half-metallic even with spin-orbit coupling. The calculated spin and orbital magnetic moments agrees well with the recent experimental XMCD measurements.

In order to gain a valuable insight into the essence of 4f electron in the δ phase cerium (bcc structure), full-potential linearized augmented plane wave method supplemented with GGA + U correction has been employed to study its bulk, electronic, and optical properties. The present equilibrium volume and bulk modulus are in accord with the experimental values. The 4f spectrum only consists two separate Hubbard bands, and at Fermi level the spectral weight of 4f characters can be neglect. It is argued that the f electron in δ phase is highly correlated. The surprising similarity between the dielectric functions of δ-Ce and γ-Ce is also addressed, which can be understood in the resemblance of degree of 4f electron correlation.

By using first principles molecular dynamics, we study postperovskite phase transition of MgSiO₃, which has important meanings in earth science, and its analogue materials, MgGeO₃ and CdGeO₃. We try to understand the mechanism and conditions of postperovskite phase transition.

In this paper, forecast of piezoelectric tensors are presented. Piezo crystals including quartz, quartz-like crystals, known and novel crystals of langasite-type structure are treated with density-functional perturb theory (DFPT) using plane-wave pseudopotentials method, within the local density approximation (LDA) to the exchange-correlation functional. Compared with experimental results, the ab initio calculation results have quantitative or semi-quantitative accuracy. It is shown that first principles calculation opens a door to the search and design of new piezoelectric material. Further application of first principles calculation to forecast the whole piezoelectric properties are also discussed.

By using statistical analysis and Monte Carlo simulation, we investigate the nature of step effect in the hysteresis loop with ±J Ising model and Heisenberg model respectively. The results indicate that it is the strong anisotropy, weak magnetic dipolar interaction, and low temperature that together induce the step effect of ferromagnetic (FM) and antiferromagnetic (AFM) mixed magnetic system, and the results are in good agreement with experiments phenomenally. That makes us view spin glass and magnetic sensor from a new aspect.

We exploit theoretically a new class of magneto-controlled nonlinear optical material based on ferrofluids in which ferromagnetic nanoparticles are coated with a nonmagnetic metallic nonlinear shell. Such an optical material can have anisotropic nonlinear optical properties and a giant enhancement of nonlinearity, as well as an attractive figure of merit.

The electronic structure is calculated for CaAl₂Si₂ by using the full-potential, densityfunctional-based method. In CaAl₂Si₂, there are two AlSi layers per unit cell, each one having in-plane sp²_xy bonding and out-of-plane spz bonding. These orbitals interact between the layers to form a bound AlSi double layer. It is found that CaAl₂Si₂ is a semimetal with a small overlap between conduction and valence bands. Its Fermi surface consists of two small hole pockets centered at the Γ point and one small electron pocket centered at the M point. The behavior of the Hall coefficient R_H and the resistivity ρ with the temperature is discussed in combination with its special electronic structure.

The Raman coupling coefficient (CC), which was defined by Shuker and Gammon, of silicate melt in high-frequency region, is still an untouched subject but the key parameter for quantitatively processing the Raman spectra (RS) to achieve the abundance of microstructural units. In this paper, we present the results of a newly constructed model - SiOT model (Wu et al) about the Raman coupling coefficients of calcium and sodium silicate melts, especially in the high-frequency region. This new model combines classical MD simulation with decomposition of simulated configurations and vibrational analysis including Wilson's GF matrix method, electro-optical parameter method (EOPM) and bond polarizability model (BPM). The displacement dependence of the cluster polarizability through the combination of EOPM and BPM connotes the description of the frequency dependence of CC. This allows us to make a direct comparison for the first time between the calculated VDOS (vibrational density of states) and RS within the entire frequency range. A strong conclusion was given that the partial VDOS, RS and CC are all the intrinsic properties of respective Q_i species, and the only variable inducing the change in total VDOS, RS and CC is the change of microstructure, i.e. the distribution of the Q_i.

Transport process of a charged polaron in impurity-doped conjugated polymers is investigated by using a nonadiabatic evolution method. Effect of the impurity ion on the polaron motion in a system of two-chains is focused. It is found that an impurity ion acts as a barrier or well rather than a bridge, which is not favourable for the polaron transport between chains.

In order to investigate the role of the electrodes in a molecular conductor, we have studied the effects of diatomic molecules H₂ and O₂ on the eigenchannels of a nanocontact formed by two Au(100) electrodes with finite cross sections by first principles calculations combined with non-equilibrium Green's function technique. Two cases where the axes of the molecules are placed along the electrode axis and vertically to the electrode axis are studied. We find that the maximum number of the eigenchannels and their energy ranges are always determined by the band structure of the electrodes. The molecules only modify the electron transmission probability of each channel.

Combining the Su-Schrieffer-Heeger (SSH) model and the non-equilibrium Green's function (NEGF) formalism, we have theoretically studied controllable conductance switching of the organic spintronics and interpreted it with the voltage-induced polaron picture. The exponential decay of magnetoresistance with bias voltage is qualitatively consistent with the experimental result.

An extended Blonder-Tinkham-Klapwijk (BTK) approach is applied to study the coherent spin polarized transport in an ferromagnetic semiconductor (FS)/superconductor (SC)/FS double tunelling junction, in which the Andreev reflection, mismatches in the effective mass and Fermi velocity between the FS and SC, as well as strengths of potential scattering at the interfaces are included. It is demonstrated that they have different effects on the oscillations of differential conductances with energy, which is quite different from that in an ferromanet (FM)/SC/FM double tunneling junction. A direct robust measure of the spin polarization of the FS using conductance spectroscopy, being very weakly dependent on the strengths of potential scattering and energy, is also revealed.

We have performed the first principles calculations on doped δ-Bi₂O₃ to investigate ionic conductivity. The crystal structure and the ionic conductivity are discussed in total energy and density of states (DOS) from the calculations. The stabilized δ-phase Bi₂O₃ doped rare-earth metal was explained from DOS data. By doping Ca, Sr, La, Gd or Sm, the ion conductivity monotonically decreases, while doping with impurity Y, Tb, Dy, Er or Tm, the ion conductivity firstly increases and then decrease. Our results support the effective oxygen vacancies mechanism.

Influences of electrode distances on geometric structure of molecule and on electronic transport properties of molecular junctions have been investigated by means of a generalized quantum chemical approach based on the elastic scattering Green's function method. Numerical results show that, for organic molecule 4,4'-bipyridine, the geometric structure of the molecule especially the dihedral angle between the two pyridine rings is sensitive to the distances between the two electrodes. The currents of the molecular junction are taken nonlinearly increase with the increase of the bias. Shortening the distance of the metallic electrodes will result in stronger coupling and larger conductance.

Electric transport of DNA is closely related to the itinerant π-electrons due to the soft characteristic of DNA molecules. We suggest a model with adjustable number itinerant electrons (ANIE). The density of states (DOS), the transmission coefficient and the electric conductance were studied. It was found that the conductance depends sensitively upon the number of the itinerant electrons. Resistivity of Poly(dG)-Poly(dC) was obtained, which is well consistent with the experimental data.

To check the relevance of charged-slab calculation, work functions of Al(111), Si(111) and TiO₂(110) surfaces are calculated using repeated charged slabs in density functional theory. It is shown that the charged-slab calculation can provide work functions which are in agreement with those by the conventional density-functional method using neutral surfaces. Moreover, it is shown that the charged-slab calculation is sensitive to the boundary condition; the charged surfaces often induces a non-uniform electricfield in a vacuum region, which produces unphysical interactions between repeated charged slabs and errors in calculated results unless the vacuum thickness is enough large.

A density matrix based fictitious electron dynamics method for calculating electronic structure has been implemented within a semi-empirical quantum chemistry environment. This method uses an equation of motion that implicitly ensures the idempotency constraint on the density matrix. Test calculations showed that this method has potential of being combined with simultaneous atomic dynamics, in analogy to the popular Car-Parrinello method. In addition, the sparsity of the density matrix and the sophisticated though flexible way of ensuring idempotency conservation while integrating the equation of motion creates the potential of developing a fast linear scaling method.

The spin-polarized, all-electron, ab initio calculations have been performed for the electronic structure of the surface of crystal LaNi₅ using the self-consistent cluster-embedding (SCCE) calculation method. The geometrical surface structure of crystal LaNi₅ and its electronic structure having the lowest ground-state energy are obtained, with the full relaxation of atomic positions along the direction perpendicular to the surface. On the surface of crystal LaNi₅, the La atom protruded, and the Ni atoms caved in, so the surface becomes uneven which increases the contacting area. The effective volume of the surface layer rises by 9%. The Fermi level of the LaNi₅ surface, which is contributed by mainly Ni 3d electrons, is much higher than that of body LaNi₅. The valence band is not fully filled. For the first two layers of LaNi₅ surface, there are 1.15 electrons transferred from La to Ni, and the two layers have small opposite spin magnetic moments which shows the paramagnet. All calculated results show that the properties of LaNi₅ surface are significantly different from that of body LaNi₅, but very similar to that of hydride LaNi₅H₇. So the geometrical surface structure of crystal LaNi₅ is in favor of the absorption of hydrogen.

Finite Element Analysis (FEA) and atomistic simulations are used to model Ge/Si quantum dots. The three dimensional non-uniform composition profile in Ge(Si)/Si(001) quantum dots is calculated using atomistic modelling. The results are compared to experimental data from the literature. FEA is used to model the contact angle dependence of the strain energy of the QD. An equation is fitted to the modelled dataset describing the strain energy of a uniformly alloyed, pyramid shaped Ge/Si quantum dot as a function of the contact angle.

The interaction of molecules (HCl, NH₃ and H₂O) and the hydrogenterminated diamond (001) surface is studied with ab inito simulation. Both the energy and the geometry indicate that the surface H atoms are more favourable to pull the H atoms of these polar molecules. The calculated charge redistribution reveals that a weak covalent interaction formed between H atom of adsorbate molecules and the H atom of the diamond surface.

Density functional theory calculations with local spin density approximation (LSDA) have been performed to study the properties of VAs/GaAs interface. It was found that the half-metallicity of Cr layers close to the interface remains unchanged for an abrupt interface. The valence band minimum (VBM) of the minority spin lies well below the Fermi level of the majority spin (1.0 eV). The Schottky barrier height (SBH) to GaAs (n-type) is calculated to be 1.09 eV, which is acceptable regarding the band gap of GaAs. The conservation of halfmetallicity at the interface indicates that such a heterostructure is a promising candidate for high efficiency spin current injection at high temperature.

The stress reduction behavior in metal-incorporated amorphous carbon films was investigated by the first-principle calculation. We calculated the total energy of the system with changes in bond angles between the incorporated metal (Ti, Mo, Cr, W, Ag, Au, Al, Si, etc) and the carbon atoms by using DMOL³computational software package. The four carbon atoms are arranged as a tetrahedron, with a carbon or metal atom at the center. The total energy increased substantially as the bond angle deviated from the equilibrium value when a carbon atom is located at the tetrahedron center. However, with a replacement by a metal atom at the center of the tetrahedron, the increase in the total energy due to the distortion in bond angle was significantly reduced. The pivotal action of the metal atoms dissolved in the carbon matrix would be more significant when noble metals having filled d-shells are incorporated compared to the transition metals having unfilled d-shells. These atoms have a weak and more isotropic bond with carbon atoms as confirmed by the electron density distribution.

The PI index is a graph invariant defined as the summation of the sums of edges of n_eu and n_ev over all the edges of connected graph G, where n_eu is the number of edges of G lying closer to u than to v and n_ev is the number of edges of G lying closer to v than to u. The index is very simple to calculate and has disseminating power similar to that of the Wiener and the Szeged indices. The comprehensive studies show that the PI index correlates highly with W and Sz as well as with physicochemical properties and biological activities of a large number of diversified and complex compounds. In this paper we prove an algorithm which is very simple for computing PI index of nanotubes. Using this algorithm the PI index of a polyhex zig-zag nanotube is computed.

The lowest-energy geometric and isomers of freestanding Co_n clusters (n = 2-13) and their corresponding magnetic moments have been systematically studied using the first principles (DMol method) based on the density-functional theory. The calculated results show that the Jahn-Teller effect plays an important role in the process because there are many isomers near the ground state. The difference of magnetic moment is in general very similar between the global minima and the second isomers. The second derivative of binding energy shows strong odd-even alternation and that the 6-, 10- and 12- atom cluster are all magic.

Ab initio Density Functional Theory (DFT) calculations are performed to study the diffusion of atomic hydrogen on a Mg(0001) surface and their migration into the subsurface layers. A carbon atom located initially on a Mg(0001) surface can migrate into the sub-surface layer and occupy a fcc site, with charge transfer to the C atom from neighboring Mg atoms. The cluster of postively charged Mg atoms surrounding a sub-surface C is then shown to facilitate the dissociative chemisorption of molecular hydrogen on the Mg(0001) surface, and the surface migration and subsequent diffusion into the subsurface of atomic hydrogen. This helps rationalize the experimentally-observed improvement in absorption kinetics of H₂ when graphite or single walled carbon nanotubes (SWCNT) are introduced into the Mg powder during ball milling.

Stability of rhombic (tetragonal-like) and cubic-like zirconia nanoparticles is investigated using the density functional pseudopotential method in a general gradient approximation. The binding energy (E_b) increases with increasing of rhombic particles, but the cubic-like particle is more favorable than rhombic ones. All studied nanoparticles have nonzero energy distances (ΔE) between occupied and non-occupied states. A simple model of nanostructured zirconia has been constructed using a Zr₁₆O₃₂ cluster as a unit element for a three dimensional superlattice with cubic symmetry. This work was supported by the Russian Foundation of Basic Researches, the grant # 04-02-9700.

The conventional tight-binding method, that based on the total energy functional proposed by Chadi [D. J. Chadi, Phys. Rev. B 19, 2074 (1979)], and the non-conventional tight-binding method, that based on total energy functional proposed by the author [Z.M. Khakimov, Comput. Mater. Sci. 3, 95 (1994)], are comparatively discussed. Principal possibility of the latter to be competitive with the first-principle methods for reliable prediction of several properties (geometry, cohesive energy, ionization potentials, electronic affinity) of atomic clusters is demonstrated for silicon case.

The Padmakar-Ivan (PI) index of a graph G is defined as PI(G) = ∑[n_eu(e|G) + n_ev(e|G)], where n_eu(e|G) is the number of edges of G lying closer to u than to v, n_ev(e|G) is the number of edges of G lying closer to v than to u and summation goes over all edges of G. The PI Index is a Szeged-like topological index developed very recently. In this paper we report on new results about computing PI index of nanotubes.

A Kinetic Monte Carlo (KMC) method is presented to describe the growth of metallic nanostructures through atomic and cluster deposition in the mono -and multilayer regime. The model makes provision for homo- and heteroepitaxial systems with small lattice mismatch. The accuracy of the model is tested with simulations of the growth of gold nanostructures on HOPG and comparisons are made with existing experimental data.

Molecular dynamics calculations are performed to calculate the formation energy for helium in Ni-vacancy and Pd-vacancy clusters. The binding energies of helium and metal self-interstitial atoms (SIA) to the helium-vacancy cluster are also determined. The comparison of these energies indicates that helium to vacancy ratio (He/V) for helium in Pd is much higher.

The existence of a large induced magnetic moment in defect single-walled carbon nanotube(SWNT) is predicted using the Green's function method. Specific to this magnetic moment of defect SWNT is its magnitude which is several orders of magnitude larger than that of perfect SWNT. The induced magnetic moment also shows certain remarkable features. Therefore, we suggest that two pair-defect orientations in SWNT can be distinguished in experiment through the direction of the induced magnetic moment at some Specific energy points.

Thermodynamic properties and the pressure of hydrate phases immersed in the ice phase with the aim to understand the nature of self-preservation effect of methane hydrate in the framework of macroscopic and microscopic molecular models was studied. It was show that increasing of pressure is happen inside methane hydrate phases immersed in the ice phase under increasing temperature and if the ice structure does not destroy, the methane hydrate will have larger pressure than ice phase. This is because of the thermal expansion of methane hydrate in a few times larger than ice one. The thermal expansion of the hydrate is constrained by the thermal expansion of ice because it can remain in a region of stability within the methane hydrate phase diagram. The utter lack of preservation behavior in CS-II methane- ethane hydrate can be explain that the thermal expansion of ethane-methane hydrate coincide with than ice one it do not pent up by thermal expansion of ice. The pressure and density during the crossing of interface between ice and hydrate was found and dynamical and thermodynamic stability of this system are studied in accordance with relation between ice phase and hydrate phase.

Effect of self-preservation of gas hydrates was explored over many years but there is no complete understanding how can hydrates exist in their thermodynamic instability region. We are suggesting the microscopic-level model of methane hydrate clusters immersed in ice matrix. Due to differences in thermal expansion of methane hydrate and Ice Ih the additional pressure appears in the hydrate phase and this moves it into its stability field. MD simulations were performed to find local pressure and density profiles. Results are well confirming our assumption.

In this paper, two phase field models for free interface of dendrite growth as sharp interface model and diffusion interface have been introduced; and microstructure evolution in solidification process is studied by phase field model, model of the PFM and its numerical solution are introduced, the dendrite growth process of Al-Si alloy in solidification process is simulated, and the simulation results has been analyzed, The comparisons of dendrite morphology between simulated results and investigations by others are also presented, which proved that the dendrite morphology are similar in trunks and arms growth, so proved the developed phase field program is accurate. But this study is still at the stage of theoretical exploration, and it is necessary to be linked with engineering application further.

First principles calculations were carried out to investigate the energetics of point defects, including solute atoms, vacancies, and antisite defects, in titanium solid solution and intermetallics. Their influence on the mechanical behavior of titanium alloys and intermetallic compounds were discussed. The self consistent procedure proposed by the authors was applied to TiAl intermetallic compounds. The ordering parameter was redefined to include the contribution of vacancies to the disordering of intermetallic compounds, so that the new approach can be applied to both strongly and weakly ordered compounds. Concentrations of point defects of various kinds can be estimated for intermetallics of different composition at different temperature. The solid solution strengthening of alloying elements through short range ordering was studied for titanium alloys and the solid solution hardening rate was estimated for various alloying elements. The solute-vacancy interaction in titanium alloys was calculated and its influence on the diffusion and creep behavior was discussed. These calculations provided some useful information for the selection of alloying elements in designing new titanium alloys.