arXiv subjects

G. T. Barkema

Publications and source records attributed to G. T. Barkema.

At least 37 records · Page 2Linked to original sources

arXiv2005

Activated sampling in complex materials at finite temperature: the properly-obeying-probability activation-relaxation technique

While the dynamics of many complex systems is dominated by activated events, there are very few simulation methods that take advantage of this fact. Most of these procedures are restricted to relatively simple systems or, as with the activation-relaxation technique (ART), sample the conformation space efficiently at the cost of a correct thermodynamical description. We present here an extension of ART, the properly-obeying-probability ART (POP-ART), that obeys detailed balance and samples correctly the thermodynamic ensemble. Testing POP-ART on two model systems, a vacancy and an interstitial in crystalline silicon, we show that this method recovers the proper thermodynamical weights associated with the various accessible states and is significantly faster than MD in the diffusion of a vacancy below 700 K.

cond-mat.mtrl-sci↗

arXiv2005

Phase separation driven by surface diffusion: a Monte Carlo study

We propose a kinetic Ising model to study phase separation driven by surface diffusion. This model is referred to as "Model S", and consists of the usual Kawasaki spin-exchange kinetics ("Model B") in conjunction with a kinetic constraint. We use novel multi-spin coding techniques to develop fast algorithms for Monte Carlo simulations of Models B and S. We use these algorithms to study the late stages of pattern dynamics in these systems.

cond-mat.stat-mech↗

arXiv2004

Binary continuous random networks

Many properties of disordered materials can be understood by looking at idealized structural models, in which the strain is as small as is possible in the absence of long-range order. For covalent amorphous semiconductors and glasses, such an idealized structural model, the continuous-random network, was introduced 70 years ago by Zachariasen. In this model, each atom is placed in a crystal-like local environment, with perfect coordination and chemical ordering, yet longer-range order is nonexistent. Defects, such as missing or added bonds, or chemical mismatches, however, are not accounted for. In this paper we explore under which conditions the idealized CRN model without defects captures the properties of the material, and under which conditions defects are an inherent part of the idealized model. We find that the density of defects in tetrahedral networks does not vary smoothly with variations in the interaction strengths, but jumps from close-to-zero to a finite density. Consequently, in certain materials, defects do not play a role except for being thermodynamical excitations, whereas in others they are a fundamental ingredient of the ideal structure.

cond-mat.dis-nn↗

arXiv2004

Accelerated dynamics with the dynamical activation-relaxation technique

The dynamics of many atomic systems is controlled by activated events taking place on a time scale which is long compared to that associated with thermal vibrations. This often places problems of interest outside the range of standard simulation methods such as molecular dynamics. We present here an algorithm, the dynamical activation-relaxation technique (DART), which slows down thermal vibrations, while leaving untouched the activated processes which constitute the long-time dynamics. As an example, we show that it is possible to accelerate considerably the dynamics of self-defects in a 1000-atom cell of c-Si over a wide range of temperatures.

cond-mat.mtrl-sci↗

arXiv2003

Universality in the pair contact process with diffusion

The pair contact process with diffusion is studied by means of multispin Monte Carlo simulations and density matrix renormalization group calculations. Effective critical exponents are found to behave nonmonotonically as functions of time or of system length and extrapolate asymptotically towards values consistent with the directed percolation universality class. We argue that an intermediate regime exists where the effective critical dynamics resembles that of a parity conserving process.

cond-mat.stat-mech↗

arXiv2003

Efficient sampling in complex materials at finite temperature: the thermodynamically-weighted activation-relaxation technique

We present an accelerated algorithm that samples correctly the thermodynamic ensemble in complex systems where the dynamics is controlled by activation barriers. The efficiency of the thermodynamically-weighted activation-relaxation technique (THWART) is many orders of magnitude greater than standard molecular dynamics, even at room temperature and above, in systems as complex as proteins and amorphous silicon.

cond-mat.mtrl-sci↗

arXiv2003

Nucleation times in the 2D Ising model

A theoretical framework is presented for the estimation of nucleation times in systems with Brownian type dynamics. This framework is applied to a prototype system: the two-dimensional Ising model with spin-flip dynamics in an external magnetic field. Direct simulation results for the nucleation times, spanning more than four orders of magnitude, are compared with theoretical predictions. There are no free parameters; all system properties required by the theory are determined by simulations in equilibrium systems. Theoretical and simulation values are found to agree in most cases within 20%.

cond-mat.stat-mech↗

arXiv2003

Reaching large lengths and long times in polymer dynamics simulations

A lattice model is presented for the simulation of dynamics in polymeric systems. Each polymer is represented as a chain of monomers, residing on a sequence of nearest-neighbor sites of a face-centered-cubic lattice. The polymers are self- and mutually avoiding walks: no lattice site is visited by more than one polymer, nor revisited by the same polymer after leaving it. The dynamics occurs through single-monomer displacements over one lattice spacing. To demonstrate the high computational efficiency of the model, we simulate a dense binary polymer mixture with repelling nearest-neighbor interactions between the two types of polymers, and observe the phase separation over a long period of time. The simulations consist of a total of 46,080 polymers, 100 monomers each, on a lattice with 13,824,000 sites, and an interaction strength of 0.1 kT. In the final two decades of time, the domain-growth is found to be d(t) ~ t^1/3, as expected for a so-called "Model B" system.

cond-mat.soft↗

arXiv2002

Large well-relaxed models of vitreous silica, coordination numbers and entropy

A Monte Carlo method is presented for the simulation of vitreous silica. Well-relaxed networks of vitreous silica are generated containing up to 300,000 atoms. The resulting networks, quenched under the BKS potential, display smaller bond-angle variations and lower defect concentrations, as compared to networks generated with molecular dynamics. The total correlation functions T(r) of our networks are in excellent agreement with neutron scattering data, provided that thermal effects and the maximum inverse wavelength used in the experiment are included in the comparison. A procedure commonly used in experiments to obtain coordination numbers from scattering data is to fit peaks in rT(r) with a gaussian. We show that this procedure can easily produce incorrect results. Finally, we estimate the configurational entropy of vitreous silica.

cond-mat.mtrl-sci↗

arXiv2002

Magnetization reversal times in the 2D Ising model

We present a theoretical framework which is generally applicable to the study of time scales of activated processes in systems with Brownian type dynamics. This framework is applied to a prototype system: magnetization reversal times in the 2D Ising model. Direct simulation results for the magnetization reversal times, spanning more than five orders of magnitude, are compared with theoretical predictions; the two agree in most cases within 20%.

cond-mat.stat-mech↗

arXiv2002

DNA electrophoresis studied with the cage model

The cage model for polymer reptation, proposed by Evans and Edwards, and its recent extension to model DNA electrophoresis, are studied by numerically exact computation of the drift velocities for polymers with a length L of up to 15 monomers. The computations show the Nernst-Einstein regime (v ~ E) followed by a regime where the velocity decreases exponentially with the applied electric field strength. In agreement with de Gennes' reptation arguments, we find that asymptotically for large polymers the diffusion coefficient D decreases quadratically with polymer length; for the cage model, the proportionality coefficient is DL^2=0.175(2). Additionally we find that the leading correction term for finite polymer lengths scales as N^{-1/2}, where N=L-1 is the number of bonds.

cond-mat.soft↗

arXiv2002

Configurational entropy of network-forming materials

We present a computationally efficient method to calculate the configurational entropy of network-forming materials. The method requires only the atomic coordinates and bonds of a single well-relaxed configuration. This is in contrast to the multiple simulations that are required for other methods to determine entropy, such as thermodynamic integration. We use our method to obtain the configurational entropy of well-relaxed networks of amorphous silicon and vitreous silica. For these materials we find configurational entropies of 1.02 kb and 0.97 kb per silicon atom, respectively, with kb the Boltzmann constant.

cond-mat.mtrl-sci↗

arXiv2001

Towards device-size atomistic models of amorphous silicon

The atomic structure of amorphous materials is believed to be well described by the continuous random network model. We present an algorithm for the generation of large, high-quality continuous random networks. The algorithm is a variation of the "sillium" approach introduced by Wooten, Winer, and Weaire. By employing local relaxation techniques, local atomic rearrangements can be tried that scale almost independently of system size. This scaling property of the algorithm paves the way for the generation of realistic device-size atomic networks.

cond-mat.mtrl-sci↗

arXiv2001

Partitioning 3D space for parallel many-particle simulations

In a common approach for parallel processing applied to simulations of many-particle systems with short-ranged interactions and uniform density, the simulation cell is partitioned into domains of equal shape and size, each of which is assigned to one processor. We compare the commonly used simple-cubic (SC) domain shape to domain shapes ch osen as the Voronoi cells of BCC and FCC lattices. The latter two are found to result in superior partitionings with respect to c ommunication overhead. Other domain shapes, relevant for a small number of processors, are also discussed. The higher efficiency with BCC and FCC partitionings is demonstrated in simulations of the sillium model for amorphous silicon.

cond-mat.mtrl-sci↗

arXiv2001

Diffusion with random distribution of static traps

The random walk problem is studied in two and three dimensions in the presence of a random distribution of static traps. An efficient Monte Carlo method, based on a mapping onto a polymer model, is used to measure the survival probability P(c,t) as a function of the trap concentration c and the time t. Theoretical arguments are presented, based on earlier work of Donsker and Varadhan and of Rosenstock, why in two dimensions one expects a data collapse if -ln[P(c,t)]/ln(t) is plotted as a function of (lambda t)^{1/2}/ln(t) (with lambda=-ln(1-c)), whereas in three dimensions one expects a data collapse if -t^{-1/3}ln[P(c,t)] is plotted as a function of t^{2/3}lambda. These arguments are supported by the Monte Carlo results. Both data collapses show a clear crossover from the early-time Rosenstock behavior to Donsker-Varadhan behavior at long times.

cond-mat.stat-mech↗

arXiv2001

Simulation of polysilane and polysilyne formation and structure

We present Monte Carlo simulations of the formation and structure of polysilane, hybrid polysilane/polysilyne and polysilyne networks. The simulation technique allows for the investigation of large networks, containing up to 1000 silicon atoms. Our results show that ring formation is an important factor for all three types of materials. For polysilyne networks, a random structure is found incorporating cyclic substructures, linear chains and branching points.

cond-mat↗

arXiv2000

Efficient tight-binding Monte Carlo structural sampling of complex materials

While recent work towards the development of tight-binding and ab-initio algorithms has focused on molecular dynamics, Monte Carlo methods can often lead to better results with relatively little effort. We present here a multi-step Monte Carlo algorithm that makes use of the possibility of quickly evaluating local energies. For the thermalization of a 1000-atom configuration of {\it a}-Si, this algorithm gains about an order of magnitude in speed over standard molecular dynamics. The algorithm can easily be ported to a wide range of materials and can be dynamically optimized for a maximum efficiency.

cond-mat.stat-mech↗

arXiv2000

Equilibrium crystal shapes in the Potts model

The three-dimensional $q$-state Potts model, forced into coexistence by fixing the density of one state, is studied for $q=2$, 3, 4, and 6. As a function of temperature and number of states, we studied the resulting equilibrium droplet shapes. A theoretical discussion is given of the interface properties at large values of $q$. We found a roughening transition for each of the numbers of states we studied, at temperatures that decrease with increasing $q$, but increase when measured as a fraction of the melting temperature. We also found equilibrium shapes closely approaching a sphere near the melting point, even though the three-dimensional Potts model with three or more states does not have a phase transition with a diverging length scale at the melting point.

cond-mat↗