Date :

 05 July 2010

Place :

The meeting room on 7th floor, Institute of Physics, Academia Sinica, Taipei

 National Center for Theoretical Sciences (Critical Phenomena and Complex Systems focus group)

 Institute of Physics, Academia Sinica (Taipei)

Miss Chia-Chi Liu (Secretary, Physics Division, NCTS)
Tel:(886)-2-33665566; Fax:(886)-2-33665565; E-mail: ccliu@phys.ntu.edu.tw

Speakers :

Asymptotic Behavior for a Version of Directed Percolation on Square Lattices
    We consider  a version of directed bond percolation on square lattices  whose vertical edges are directed upward with probabilities p_v and horizontal edges are directed rightward with probabilities p_h and one in alternate rows.   Let τ(M,N) be the probability that there is a  connected-directed path of occupied edges from (0,0) to (M,N). For each p_h∈[0,1], p_v=(0,1) and aspect ratio α=M/N fixed,  it was established that there is an α_c=[1-p_v^2-p_h(1-p_v)^2]/2p_v^2 such that as N→∞,  τ(M,N) is 1, 0 and 1/2  for α>α_c,α<α_c and α=α_c, respectively. Particularly, for p_h=0 or 1 the model reduces to the Domany-Kinzel model.  In this talk,  we investigate the asymptotic behavior of τ(M,N) and τ(M^-_n,N), τ(M⁺_n,N), where τ(M^-_n,N)/N ↑ α_c and τ(M⁺_n,N) ↓ α_c as N→∞.
　

Soft particle and polymer conformation and dynamics in micro and nanochannels

    We investigate the dynamics of deformable, soft particle suspensions in microfluidic flow. A lattice Boltzmann-Brownian Dynamics(LB-BD) approach is developed to capture coupling between particle deformation and fluid field perturbations. The influence of hydrodynamic interactions on particle dynamics is characterized and found in agreement with single and two particle studies. For single component suspensions, it is found that the more elastic particles achieve higher average velocity due to wall-induced hydrodynamic migration. In binary suspensions, a novel mechanism for particle fractionation based on particle elasticity is found.
　

Rate determinants for protein fibrillization: an amyloid beta story

A Monte Carlo Renormalization-Group Study for the q-state Manna Sandpile

    A renormalization group (RG) approach to the q-state Manna model where q=2, 3, 4, 6, 8, 12, and 20 is extensively investigated by sampling RG relaxation events from a Monte Carlo procedure. For q≤4, through comparing the RG results of the exact enumeration with those of our Monte Carlo procedure, we show that our Monte Carlo procedure is efficient and reliable. For q >4, according to our Monte Carlo procedure, we find that the resulting height probabilities, the toppling number exponent and the dynamical exponent are consistent with the results obtained from numerical simulations. Finally we conclude that the grain-redistribution effect dominates the accuracy of the RG results.

q-statistics in systems of polymer chains and Lennard-Jones fluid

Exact probability distribution function for multi-fractal walk models

Prof. Fa-Yueh Wu
Department of Physics, Northeastern University, USA
E-mail: fywu@lepton.neu.edu

Potts and percolation models on lattices with hexagonal symmetry: Closed-form expressions of critical frontiers and numerical analysis

    We consider the Potts and percolation models on lattices with triangular and kagome symmetries. For triangular-type lattices we use a known exact result to obtain new closed-form expressions of critical frontiers of the Potts and percolation models for lattices with subnet structures. This extensively extends results on percolation models obtained recently by Ziff and co-workers. For kagome-type lattices we deduce a new closed-form expression of the critical frontier under a homogeneity assumption, by using a duality relation formulated by Essam more-than-30 years ago, and apply it to determine the critical frontier of various models. To assess the accuracy of the determinations, we carry out finite-size analyses on critical frontiers for a host of Potts, bond, and site percolation models on various lattices. Comparisons with known exact results, where they are known, the finite-size analyses are shown to be accurate to within 7 and 8 digits. This in turn determines the accuracy of the homogeneity assumption to be within 5 digital places.