
Computational Physics Seminar 20th May 2015, Professor Robin Seligman, 
Speaker: Noa Marom, Department of Physics and Engineering Physics, Tulane University, New Orleans, LA 70118
In this talk I will provide an introduction to GW methods and demonstrate their applications to the components of organic and dyesensitized solar cells: TiO_{2} clusters [1], organic semiconductors [2,3], dyes [4,5], and dyesensitized TiO2 clusters [6,7].
Speaker: Irina Paykin, Physics Department, Technion
MSc seminar (In Hebrew): Supervisors: Drs Erez Ribak and Joan Adler.
Abstract
This study concentrates on an approach for aligning multiaperture optical systems by using information available only in the image plane itself, thereby correcting the image without any information on the wavefront. Previous theoretical work and simulations have shown that the optical problem can be mapped onto a model for crystal roughening that has served as a motivation to implement the simulated annealing algorithm in adaptive optical systems.
We present the first simulations which are carried out in tandem with a hardware realization of a simulated annealing algorithm by Dr. E. Ribak and Lee Yacobi in a specially designed multiaperture active optical system. The results of both simulations and laboratory experiments demonstrate the ability of the simulated annealing algorithm to correct a piston and tip/tilt errors. In addition we explored image restoration techniques, required for multiaperture systems. We implemented and investigated several classic deconvolution algorithms, such as WienerHelstrom, LucyRichardson and blind deconvolution. Finally, we analyzed diffraction and aberration effects related to specific multiaperture pupil configurations.
Speaker: Joan Adler
Affiliation: Physics, Technion
Completely different architecture to other top 500 machines.
Speaker: Dr Adham Hashibon
Affiliation: Fraunhofer Institute for Mechanics of Materials IWM, Wöhlerstr. 11, 79108 Freiburg
Date  Speaker  Topic 
16th May, 2012  Grzegorz Kamieniarz  Nanomagnetism en route from molecules to materials 
Title: Electronic Structure and Electron Transport in Nanoscale Systems Speaker: Professor Jerry Bernholc, Drexel Professor of Physics and Director Center for High Performance Simulation North Carolina State University, Raleigh, NC 276957518 Date: Thursday 28th July, 2011 Abstract: Nanoscale and molecular electronics promise to revolutionize computing, sensing, and harvesting of solar energy. However, molecularscale control and manufacturing are difficult tasks, which require major advances to become practical in largescale applications. The development of molecular scale devices and circuits can be greatly enhanced by predictive simulation of their components and by formulating design principles that will make molecular circuitry smaller, more efficient and more reliable. This talk will discuss two recent applications: (i) We show that the celebrated Negative Differential Resistance (NDR) effects in molecular electronics can be expected for a wide range of organic molecules attached to semiconductor and metallic leads. In particular, the NDR position and strength can be made tunable both by simple atomic substitutions and by doping of the leads. In multiterminal structures, NDR can also emerge through quantum interference effects, which open new avenues for device design and logic. (ii) In the second part of the talk we will discuss the electronic structure and spin polarization of nitrogendoped carbon nanoribbons, which are candidate materials for ultrahigh speed nanodevices. While the ground state of zigzag ribbons (ZR) is spin polarized, defects at the edges destroy the polarization and lead to a nonmagnetic ground state. We also find enhanced N segregation in ZR, due to interplay between impurity states in the valence bands and the edge states. Spin distribution is significantly affected, even at edges that are quite far from the dopant. Turning to armchair nanoribbons (AR), the three AR families, defined by mod (n, 3), behave differently in doping, with family 1 AR being the most attractive for ntype semiconductor applications.
Title: GPUs as a computational engine  basic ideas Speaker: Mark Silberstein, EE, Technion Date: Sunday January 9th, 2011  This is the last seminar for this semester. Seminar restarts next semester at 10:30 on Tuesdays.  Title : Effects of disorder in microscopic models of heat conduction in one dimensional dielectrics Speaker: Tal Kachman Aviv Date: November 28, 2010 Abstract: Understanding transport phenomena in the micro scale is of interest from both the fundamental theoretical point of view as well as the practical one. The problem of classical heat conduction in 1D system is a well known one. The problem regards the validity of Fourier law in this setting. In our research we investigate the heat transfer in broader context, taking into account the nonstationary processes. In this talk we will report the investigation of the phenomena of non stationary heat conduction in onedimensional linear and nonlinear chains (classical models of dielectrics), by combining molecular dynamics simu lations and analytic investigation. Two main aspects of our research were the transport of kinetic energy through periodic chains (both integrable and nonintegrable) and through disordered chains. We also investigate how the effects of disorder affect the transport mechanism  Speaker: Dr Dan Mordehai Affiliation: Materials, Technion Topic: Size Effect in trength of Microparticles: A combined Experimental/FEM/Molecular Dynamics Study Date: 20101114  Speaker: Joan Adler Date: 7th November Title: Visualization of electron density in general and in carbon allotropes ABSTRACT: Quantum mechanics tells us that electrons don't run in circles around nuclei, they have a probablity of being somewhere at sometime. This is a concept that we try to explain in Modern Physics courses, but its hard to ``see''. The distribution of electron density around the carbon atoms of allotropes, including diamond, amorphous carbon and nanotubes, turns out to be very important for understanding their properties. I will describe the AViz approach to visualizing electron density with examples from both Modern Physics courses and research projects in the Computational Physics group. The talk will be based on slides from my recent conference presentations: Hydrogen atom visualization  http://phycomp.technion.ac.il/~phr76ja/TW/talkindex.html and carbon allotrope visualization  http://phycomp.technion.ac.il/~phr76ja/CCP2010/talktitle.html and will present results obtained in collaboration with Joey Fox, Or Cohen David Saada, Anastassia Sorkin, Rafi Kalish, Jeremie Zaffran, Amihai Silverman and Polina Pine.Speaker: Oleg Gendelman Affiliation: Mechanical Engineering, Technion Title: Peculiarities of heat conduction in nanosystems Date: 24Oct2010, Sunday Time: 11:00 (note change from earliest announcement) Place: Lidow Complex, room 620 Remarks: Coffee/tea at 10:45The presentation is devoted to applicability of Fourier law of heat conduction in at nanoscale and in low – dimensional systems. Two main aspects of this problem will be discussed, with brief review of existing analytic, numeric and experimental results.
The first aspect is existence (size independence) of the heat conduction coefficient for the case of stationary conduction. Numeric simulations, supported by some analytic evidence, suggest crucial effect of dimensionality – only for 3D systems the heat conduction coefficient reveals size independence.
The other aspect is related to the nonstationary heat conduction. In this case, one can demonstrate that for extremely small times or extremely small space scales, the parabolic equation of the heat conduction is no more applicable and hyperbolic models should be used instead. Such models use even more empiric coefficients than the common Fourier law. Applicability of popular hyperbolic extensions of Fourier law (Cattaneo – Vernotte and some others) will be discussed quantitatively and qualitatively.
Last semester's talks
Monday 4th October at 11:00 in rm 620
Jonathan Gross, Juelich Massively parallelized computer simulations on GPUs with CUDA. We discuss the advantages of parallelization on graphics processing units (GPUs) for parallel tempering Monte Carlo computer simulations of an exemplified beadspring model for homopolymers. Since the sampling of a large ensemble of conformations is a prerequisite for the precise estimation of statistical quantities such as typical indicators for conformational transitions like the peak structure of the specific heat, the advantage of a strong increase in performance in Monte Carlo simulations cannot be overestimated. Employing multithreading and utilizing the massive power of the large number of cores on GPUs, available in modern but standard graphics cards, we find a noticeable increase of efficiency when porting parts of the code to the GPU. Not only is it possible to simulate multiple replica simultaneously, but also to parallelize important parts of a single simulation itself. With a deeper understanding of the GPU architecture and different memory layers further optimizations can be done.