Computations in Science Seminars

Previous Talks: 2005

January 12, 2005
Vakhtang Putkaradze, Dept. of Mathematics and Statistics, University of New Mexico
Faculty contact:Wendy Zhang,
Mathematical models of self-aggregation of particles at nano-scales
We derive a non-local evolution equation for self-assembly of particles at the nano-scale. The physical assumptions underpinning the problem are the presence of interaction potential between the particles and the dependence of mobility on averaged local density. We show that for almost any choice of sufficiently smooth potential and averaging functions the evolution is reduced to a finite-dimensional system of ODEs describing the dynamics of particle clumps, and the order of that system is generally very low. We also discuss how our equation can be reduced in some (singular) limits to several well-known models of different physical phenomena, including chemotaxis models, Cahn-Hilliard equation and some crystal growth phenomena.

January 19, 2005
Rho Shin Myong , Gyeongsang National University
Faculty contact: Timur Linde
Modeling Micro Flows: Surface Chemistry, Boltzmann Equation, and Irreversible Thermodynamics
The interfacial interaction between the gas (or liquid) molecules and the material surface plays a dominant role in the nonlinear transport phenomena associated with micro- and nano-devices. A slip model by Maxwell based on the concept of accommodation coefficient is usually employed to describe this effect. In this talk, an alternative model based on Langmuir°Øs theory of adsorption of gases on solids is presented. It turned out that the new model is capable of providing a physical meaning to the accommodation coefficient in the Maxwell model. In addition, a computational model for micro and rarefied gas flows beyond the Navier-Stokes-Fourier equations is presented. It is derived from the Boltzmann equation by employing the Eu's modified moment method in which a formal mathematical structure of the non-equilibrium entropy compatible with the 2nd law of thermodynamics is developed. Finally, various benchmark problems--internal flows in micro-channels and concentric rotating micro-cylinders, external flow over a micro-sphere, heat transfer in forced laminar flow through a circular micro-tube, and shock-dominated hypersonic flow around a blunt body--will be considered in order to validate the prediction of the new models.

February 2, 2005
Jointly sponsored by the Institute for Biophysical Dynamics
Amit Meller, Harvard University
Faculty contact: Tom Witten
Translocation and unzipping kinetics of DNA molecules using a nanopore
The discovery that a voltage gradient can be used to drive single-stranded DNA or RNA molecules through the ~1.5 nm a-Hemolysin nanopore, has opened up the possibility of detecting and characterizing unlabelled individual polynucleotide molecules. By measuring the ion current flowing through the pore when the biopolymer is threaded, we detect the passage time of each DNA molecule. We found that the translocation dynamics is dominated by the DNA-pore interactions that depend strongly on the DNA sequence as well as on its direction of entry inside the pore. Recently we have developed a method to dynamically set the voltage that drives the polynucleotides. This method makes it possible to study the unzipping kinetics of individual DNA molecules under fixed or time-varying forces. In particular, we have characterized the unzipping kinetics of DNA hairpin molecules under fixed voltage amplitudes (V) or steady voltage ramps (dV/dt). At high voltages (V>30mV) or at high voltage ramps (dV/dt>5 V/s) the unzipping process can be described by a single step kinetics model width negligible re-zipping probability. But at the low voltage (or voltage ramp) regime re-zipping probability must be included to account for our data. A model that includes re-zipping is introduced and is used to fit our data at low and at high voltages. From the fits we estimate the effective DNA charge inside the nanopore and the unzipping rate of the hairpins at the limit of zero force.

February 9, 2005
Douglas MacAyeal, University of Chicago
Faculty contact: Leo kadanoff,
Ice-shelf explosions, iceberg calving and other oddities of the glaciological world
Over the last 4 years, I have been observing the processes by which large parts of the Antarctic ice sheet have disintegrated (e.g., since Y2K, a surface area of glacial ice equivalent to all the main New England states has broken into small bits and drifted off to oblivion in the warm waters of the Southern Ocean). Current study involves little computation, so expect a descriptive travelogue-style presentation of the main ideas at play in glaciodyanmics of ice-sheet breakup.

TUESDAY, February 15, 2005: 12 p.m., Crerar Library, Lower Level Conference Room
Jointly sponsored by the Institute for Biophysical Dynamics
Mark Goulian, Department of Physics, University of Pennsylvania
Faculty contact: Aaron Dinner,
Perturbing, imaging, and modeling two-component signaling in bacteria

February 16, 2005
Mary Silber, Northwestern University
Faculty contact: Wendy Zhang,
Controlling Pattern Formation with Multi-Frequency Parametric Forcing
I will describe our work on manipulating the weakly nonlinear interactions of parametrically-excited standing waves on the surface of a fluid layer. Our mathematical analysis is motivated by recent experiments on standing wave patterns, first reported by Michael Faraday in 1831. The surface waves are generated by subjecting the fluid container to a periodic vibration in the vertical. It has been demonstrated in the lab that one can switch between different, exotic nonlinear wave patterns by adjusting the frequency content of the forcing function. We have focused on three-wave interactions, as building blocks of more complicated nonlinear wave interactions. In the weak forcing limit, we use methods of equivariant bifurcation theory to determine which additional frequency components to add to a sinusoidal forcing function in order to enhance or suppress particular resonant triad interaction. We have used our results to interpret experimental findings, suggest new experiments, and to explain the results of our companion analysis on impulsively-forced Faraday waves. In the latter problem, we consider surface waves excited by a periodic sequence of delta-function impulses, which is an idealization of smooth periodic forcing that proves analytically tractable in the linear and weakly nonlinear regimes.

February 23, 2005
Sung Joon Moon, Princeton University
Faculty contact: Wendy Zhang,
Modeling grains and powders: Multiscale methods for multiphase flow problems
Discrete particles are pervasive in nature and industry, but many industrial processes are far from cost-effectiveness because of a lack of understanding of particle flow behavior. Numerical methods for particulate flows have significant potential for uncovering the underlying physics, cost savings, and productivity enhancements.
In this talk, two examples that have been investigated with experiments or that have significant practical importance will be considered: vertically vibrated shallow granular layers and vibrated gas fluidized beds of cohesive fine powders. Oscillating shallow granular layers, where the interparticle interaction can be modeled as an instantaneous binary, frictional dissipative collision, exhibit various phenomena including the spatial pattern formation and the so called "horizontal Brazil-nut effect".
They will be discussed, together with the issues in modeling. The second problem examined is the gas fluidized bed of fine particles, often referred to as the powders, where the interstitial gas effects and interparticle forces such as cohesion play an important role. A particle dynamics-based hybrid model is used to study this system, and a multiscale computational method is used to facilitate the efficient computation. It will be shown how the vibration enables fluidizing highly cohesive powders, which are unable to get fluidized otherwise.

March 2, 2005
Mohammad Islam, Physics Department, University of Pennsylvania
email:, Faculty contact: Sidney Nagel,
Single-Wall Carbon Nanotube Dispersions: Gels, Liquid Crystalline Phases, Optical and Magnetic Anisotropies
I will describe our explorations of carbon nanotube science and technology from a soft materials perspective. We first created stable dispersions of purified single-wall carbon nanotubes (SWNTs) using an anionic surfactant, sodium dodecyl benzene sulfonate (NaDDBS), and then studied their structure and rheology in suspension, demonstrating interconnected networks of stiff filaments. Our attempts to induce nematic liquid crystalline alignment of SWNTs in suspension did not succeed, but eventually led us to create a new class of nanocomposite: nematic nanotube gels. These gels exhibit rich physical properties due to a coupling between the nematic alignment and the polymer network elasticity. Finally, these gels and dispersions have been enabling technologies for fundamental measurements of single tube optical and magnetic anisotropy.

March 9, 2005
Norbert Schorghofer, Institute for Astronomy, University of Hawaii
e-mail:, Faculty contact: Wendy Zhang,
The Stability of Ice on Mars
An extensive reservoir of ground ice has been discovered on Mars using neutron and gamma spectroscopy from orbit. Atmospheric water vapor diffuses into the ground and deposits frost and, vice versa, ground ice sublimes and diffuses upward, while exposed to temperature cycles and adsorption. The ground ice should rapidly evolve toward vapor equilibrium, but it is unknown if this equilibrium solution depends on history. Vapor diffusion on a small scale determines the distribution of ground ice on a planetary scale. The theory is used to predict the occurrence of ice elsewhere on the planet and to explain how the oldest ice on Earth can survive 20 cm beneath the surface.

March 30, 2005
Snezhana Abarzhi, Center for Turbulence Research, Stanford University
e-mail:, Faculty contact: Leo Kadanoff,
Turbulent mixing and beyond
Whenever fluids of different densities are accelerated against the density gradient we observe the development of the Rayleigh-Taylor instability (RTI), which causes extensive interfacial mixing of the fluids. The turbulent mixing plays a key role in preventing the formation of "hot spot" in inertial confinement fusion, providing proper conditions for the synthesis of heavy mass elements in thermonuclear flashes on the surface of compact star, in core-collapse supernovae, and many other applications. The properties of the Rayleigh-Taylor turbulent mixing flow differs significantly from those of the classical Kolmogorov turbulence. We study theoretically the dynamics of the large-scale coherent structures in RTI, identify the invariants of the flow and show that the instability evolution has a non-local and multi-scale character. A phenomenological model is suggested to describe the turbulent mixing of immiscible, miscible and stratified fluids, and to account for the stochastic properties of the process. We discuss the applications of the results obtained in stellar non-Boussinesq convection, supernovae, and reactive flows.

April 6, 2005
Matthew Hastings, Los Alamos National Laboratory
e-mail:, Faculty contact: Wendy Zhang,
Lieb-Schultz-Mattis in Higher Dimensions
In 1961, Lieb, Schultz, and Mattis showed the absence of a gap in a class of one-dimensional spin chains: chains with half-integer spin per unit cell and SU(2)- invariant short-range interactions. This basic result has guided research on spins chains ever since. For example, the discovery of the Haldane gap in chains with integer spin was surprising as it indicated a fundamental difference between integer and half- integer spins.

Since then, there has been much work searching for higher dimensional extensions of this result, in particular due to possible connections to high- temperature superconductivity. The clearest statement of the basic physical reasons to expect such an extension are due to Misguich et. al, who argued that any such system would either have long- range spin order, and hence have gapless spin wave excitations, or else would have a class of topological excitations with vanishing gap. Thus, showing this result in higher dimensions would connect directly to recent ideas on topological order in quantum systems. I will sketch my recent proof of this result, emphasizing connections to these basic physical ideas. In the process, I will derive various results about locality of correlation functions in these systems.

April 13, 2005
Jointly sponsored by the Institute for Biophysical Dynamics
Alexander Van Oudenaarden, MIT
e-mail:, Faculty contact: Philippe Cluzel
Information storage and propagation in genetic networks
The ability of a living cell to grow and divide, and to sense and respond to its environment is determined by a complex web of intracellular, and sometimes intercellular, protein and gene networks. During the last decade new technologies such as high-throughput genome sequencing and gene arrays have enabled a large-scale identification of these interaction networks. Although many of these networks have already been mapped, surprisingly little is known about the function of specific network architectures. Rather than taking a genome-wide approach, our lab focuses on the smaller, recurring network motifs buried in the larger networks. These motifs are built from a handful of genes and proteins and display a network structure that appears over and over again in different networks and different organisms. The underlying idea is that these motifs define autonomous functional modules that are the building blocks for the entire cellular network. In this talk I will focus on two elementary motifs: positive feedback loops that can be used to store information and generate steep switches; and feed-forward loops that are used to propagate signals and the concomitant noise through the network. I will present both theoretical models and experiments on natural and synthetic genetic networks in the bacterium Escherichia coli and the budding yeast Saccharomyces cerevisiae.

April 20, 2005
Pavel M. Lushnikov, University of Notre Dame and Landau Institute
email:, Faculty contact: Paul Wiegmann,
Singularity formation in hydrodynamics, nonlinear optics and plasma
Many nonlinear systems have a striking phenomenon, an explosive instability, which occurs if the system is linearly unstable and nonlinearity does not saturate an exponential growth of small perturbations, but, on the contrary, results in singularity formation in a finite time. Near singularity point there is usually a qualitative change in underlying physical phenomena, reduced models loose their applicability and other physical mechanisms become important such as inelastic collisions in the Bose-Einstein condensate; optical breakdown and dissipation in nonlinear optical media and plasma, wave breaking in hydrodynamics. Special focus will be on two examples of explosive instability. First is the spontaneous breaking of rotational symmetry and formation of coherent hexagonal pattern in photorefractive crystals. Second is the wave breaking and foam formation on crest of sea waves.

TUESDAY, April 26, 2005: 12 p.m., Crerar Library, Lower Level Conference Room
Jointly sponsored by the Institute for Biophysical Dynamics
William Gelbart, Department of Chemistry, University of California, Los Angeles
e-mail:, Faculty contact: Aaron Dinner,
Physical aspects of viral infection
The sole job of a virus is to get its genome into a host cell. It is then replicated a large number of times and each new copy in turn infects neighboring cells. In this talk I outline the essentials of viral life cycles for the cases of infection by viruses of bacteria, plants, and animals, highlighting the basic differences between the modes of entry in these three general situations. In virtually all plant and animal cases, the entire viral particle enters the host cell, i.e., the (nucleic acid) genome and the (protein) capsid. In the case of viruses that infect bacteria, on the other hand, the viral particle remains outside the host cell and injects its genome by virtue of the high pressure in its capsid. Experiments are described in which these pressures are measured and in which it is shown that genome delivery is necessarily incomplete due to osmotic pressure in the host cell (bacterial) cytoplasm. It is then conjectured that the remainder of the genome enters the cell via a ratcheted diffusion mechanism due to nonspecific binding of proteins. We argue that the translocation process is significantly faster than predicted by ideal ratcheting, because of an entropic force associated with particle binding. This speed-up is estimated by both molecular dynamics simulation and by analytic and numerical solutions to the equations describing the coupling of chain diffusion and particle binding. Explicit account is taken of both the driving forces due to capsid pressure and the resisting forces arising from osmotic pressure in the host cell cytoplasm. I finish by outlining new experiments, both in vitro and in vivo, for testing this general theory of the kinetics of genome delivery by bacterial viruses.

April 27, 2005
Shenda Baker, Harvey Mudd College
e-mail:, Faculty contact: Leo Kadanoff,
Pasta science: self-assembling diblock copolymers in two dimensions
We have developed and modeled a means to deposit nanoscopic structures such as dots or lines of a monolayer of diblock copolymer at a liquid interface (that can be subsequently deposited onto a solid) by a very simple preparation. A droplet of known concentration of polystyrene-polyethylene oxide PS-PEO with a particular ratio of block sizes is spread on an air-water interface. A competition between spreading, caused by Marangoni effects, and solvent evaporation leading to aggregation produces 2-dimensional features that can be controlled by judicious choice of solution and polymer variables. We have also developed a fluid hydrodynamic model that captures the details of the process quantitatively with no free fitting parameters.

If there is time, I shall morph to a brief discussion of "Strange Matter" --- a traveling museum exhibition that highlights Materials Science. The exhibit was a partnership of he Materials Research Society and Industry with funding through NSF and a contracted museum design team.

May 4, 2005
Constantino Tsallis, Centro Brasileiro de Pesquisas Fisicas
e-mail:, Faculty contact: Leo Kadanoff,
Finally, the entropy Sq is extensive or nonextensive?
Boltzmann-Gibbs entropy and associated statistical mechanics constitute what is well known to be the correct thermostatistical theory to be used for systems whose microscopic dynamics is strongly chaotic (e.g., positive Lyapunov exponents for classical systems). For a vast class of the so called complex systems -- typically nonlinear dynamical systems at some kind of edge of chaos (e.g., with zero Lyapunov exponents) -- this theory needs to be generalized. Nonextensive statistical mechanics constitutes a possible such generalization, introduced in 1988, and addressing non equilibrium systems. A brief introduction to the theory, as well as to its dynamical foundations (calculation from first principles of the entropic index q) will be provided. Some selected applications will be mentioned. Emphasis will be given to the analysis of the extensivity of the entropy Sq which generalizes the Boltzmann-Gibbs form. Bibliography: (i) ; (ii) M. Gell-Mann and C. Tsallis, eds., Nonextensive Entropy - Interdisciplinary Applications (Oxford University Press, New York, 2004); (iii) C. Tsallis, M. Gell-Mann and Y. Sato, Special scale-invariant occupancy of phase space makes the entropy Sq additive, cond-mat/0502274.

May 11, 2005
Jim Sethna, Cornell Center for Materials Research, Cornell University
e-mail:, Faculty contact: Wendy Zhang,
Sophisticated Statistical Mechanics of Sloppy Models
Science is filled with multiparameter models that must be fit to observations. An ecosystem has many interacting species, a cell has interacting proteins and genes, and a material has many atoms whose forces are governed by quantum-mechanical electronic calculations. A key question for these models is when we can trust their predictions: usually only wisdom and experience can judge for which problems a given model will likely be reliable. One source of unreliability in these models is that they are sloppy: the parameters are ill-determined by the data, with enormous ranges giving roughly equivalent fits. These parameters giving roughly equivalent fits, however, do not yield the same predictions! By using an ensemble of good parameter sets, we have been able to produce falsifiable predictions for regulatory networks in cells with fifty unknown parameters. We have also used them to generate estimates for the `sloppy model' component of the systematic `transferability' errors in interatomic potentials.

June 1, 2005
Thomas Witelski, Dept. of Mathematics, Duke University
e-mail:, Faculty contact: Wendy Zhang,
The coarsening dynamics of dewetting fluid films
The study of instabilities of thin fluid films on solid surfaces is of great importance in understanding coating flows. These instabilities lead to rupture, the formation of dry spots, and further morphological changes that promote non-uniformity of coatings; these behaviors in unstable thin films are generally called dewetting. Following rupture and subsequent transient behavior, the long-time structure of films takes the form of an array of droplets. The evolution of this system can be represented in terms of coupled ODEs for the masses and positions of the droplets. Regimes where droplet coarsening by each of two mechanisms (collision and collapse) are identified, and power laws for the statistics of the coarsening processes are explained. This is joint work with Karl Glasner, University of Arizona.

June 8, 2005
Jointly sponsored by the Institute for Biophysical Dynamics
Michael Brenner, Harvard University
e-mail:, Faculty contact: Leo Kadanoff,
Towards models for the structure and evolution of ion channels
This talk summarizes some very preliminary theoretical ideas and calculations regarding the structure and evolution of ion channels. The great advantage of ion channels is that they are individual proteins whose function has long been known and is readily inferred through voltage measurements. Their evolution can be readily tracked through the growing data base of sequences. The kinetic schemes of ion channels (regulating membrane permeability to ions) have been studied for more than 50 years, and in general can be quite complex. We will describe our endeavors to ask a somewhat dangerous question: Why are they this way? Examples of questions we would like to answer include: to what extent do design principles dictate the details of the kinetic schemes of ion channels, such as (a) the symmetry of the sodium and potassium channels (or lack thereof), as reflected in their kinetic schemes ; (b) the coupling of sodium channel kinetics to potassium channel kinetics; or (c) activation/inactivation of the channels themselves.
The talk will combine a description of what is known and not know about ion channels, and their evolution; a description and some analysis of models; and some preliminary analysis of using data for the evolution of the sequences of voltage gated sodium and potassium channels in conjunction with the models. The great hope is to be able to draw conclusions that are not 'just so' stories.

June 15, 2005
Adriana Pesci, University of Arizona
e-mail:, Faculty contact: Leo Kadanoff, .
Connections between classical statistics and the Schroedinger and Pauli equations.
In the year 1926 Madelung found a transformation that connected the Schroedinger operator of quantum mechanics with the continuity and Euler equations of fluid mechanics in which the ``pressure" is proportional to the Laplacian of the density of the fluid. Later on, Bohm and Takabayasi found a similar transformation connecting the Pauli and the ideal fluid equations where the new ``pressure" term involved the same Madelung term plus a vortical component.
We know that the Euler and continuity equations can be derived from statistical descriptions of fluids. Correspondingly we might ask: What statistical description, if any, stands behind the Madelung and Bohm-Takabayasi equations?
Here we suggest a possible answer to such question in the form of an irreversible mapping in Fourier space with a single free parameter. We find a particular class of probability functions that give rise to the Schroedinger and Pauli equations for which the averages of physical quantities read like the postulates of quantum mechanics. This procedure seems to provide a statistical representation of quantum mechanics.

June 22, 2005
Jointly sponsored by the Institute for Biophysical Dynamics
John Bechhoefer, Simon Fraser University
e-mail:, Faculty contact: Leo Kadanoff, .
Kinetic model of DNA replication in higher organisms
Higher organisms contain about three billion DNA base pairs. Although the genome is replicated in times as short as fifteen minutes, individual base pairs are copied at only ten bases/second. The apparent paradox is resolved by having many "origins" of replication distributed along the genome and initiating stochastically. New experimental and theoretical techniques are beginning to shed light on the organization of replication -- giving, for example, the rate of initiation of origins at different moments during the replication cycle. We discuss evidence that the looping of chromatin -- the DNA-protein complex that makes up chromosomes -- plays an important biological role in organizing DNA replication.

June 29, 2005
Anette Hosoi, Department of Mechanical Engineering, MIT
e-mail:, Faculty contact: Leo Kadanoff,
Building a better snail: lubrication and adhesive locomotion
Many gastropods, such as slugs and snails, crawl via an unusual mechanism known as adhesive locomotion. We investigate this method of propulsion using two mathematical models, one for direct waves and one for retrograde waves. We then test the effectiveness of both proposed mechanisms by constructing two mechanical crawlers. Each crawler uses a different mechanical strategy to move on a thin layer of viscous fluid. The first uses a flexible flapping sheet to generate lubrication pressures in a Newtonian fluid which in turn propel the mechanical snail. The second generates a wave of compression on a layer of Laponite, a non-Newtonian, finite-yield stress fluid with characteristics similar to those of snail mucus. This second design can climb smooth vertical walls and perform an inverted traverse.

July 6, 2005 (&)
Uri Hershberg, Yale University School of Medicine
Finding rules of immune selection dynamics by analyzing their genetic starting point.
Selection, whether in evolution or in the process of immune reaction to pathogen takes place on two landscapes. Mutation occurs at the level of the genotype while selection occurs at the level of the phenotype. In the immune reaction to invading pathogens B cells undergo high levels of mutation of the DNA encoding their antigen receptors. In parallel they undergo rapid proliferation and death in a pattern dependent on the affinity of their receptors to antigens exhibited by the patogen. The relationship between the genotypic and phenotypic landscapes is complex. Movement in one does not imply movement in the other. For a given phenotype there may be multiple states in the genotype landscape. This implies that two individuals with the same phenotype can nonetheless differ in the potential for change that the genotype encodes i.e. what they can mutate to. Ultimately this potential is determined by the codons and the amino acids they specify.

We have developed a network view of the amino acid table in which every codon is a node and every edge is a mutation. We have used the measures this view generates to analyze the DNA sequences that start the process of affinity maturation in the immune reaction. The results of our analysis suggest three new ideas about selection: First, the traits of amino acids and the potential to change them are a meaningful signal for selection. Second, we found that while the DNA encoding B cell receptors has evolved to generate variable progeny under high rates of mutation, the different gene families differ in the extent to which they will risk their potential viability. Finally, the existence of a transition bias in mutations means that not all movements on the amino acid network are equal. Codons tend to mutate to codons that are a Transiton mutation away from them, dividing the amino acid network into Transition Neighborhoods.

July 13, 2005
Eric Isaac Corwin, University of Chicago
e-mail:, Faculty contact:Wendy Zhang,
Structural signature of jamming in granular media
Glasses are rigid but flow when the temperature is increased. Likewise, granular materials are rigid but become unjammed and flow if sufficient shear stress is applied. The rigid and flowing phases are strikingly different, yet measurements of the structure in glasses and liquids are virtually indistinguishable. Is there a structural signature of the jammed state that distinguishes it from its unjammed counterpart? We address this question with a novel experiment that accesses the contact-force distribution measured during shearing. Because forces are sensitive to minute variations in particle position, the distribution of forces can serve as a microscope with which to observe nearest-neighbor position correlations. We find a qualitative change in the force distribution at the onset of jamming. If, as has been proposed, the jamming and glass transitions are related, this observation of a structural signature at jamming hints at a similar structural difference, too subtle for conventional scattering techniques to uncover, at the glass transition. Our measurements also provide a determination of a new granular temperature that would be the counterpart in granular systems to the glass-transition temperature in liquids.

July 20, 2005
Marko Kleine Berkenbusch, University of Chicago
e-mail:, Faculty contact: Wendy Zhang,
Numerical studies of the selective withdrawal transition
Emulsification studies show that droplets immersed in an external straining flow are easily pulled apart. In situations of comparable viscosity of the inner and outer liquid, the only stable steady state shapes are slightly deformed drops with rounded edges.

In selective withdrawal experiments, two immiscible fluids form a layered system with a horizontal interface. Fluid is withdrawn from the top layer through a straw suspended closely above the interface. Cohen et al. found in their setup that steady state interfaces with sharp tips could be produced even when the two fluid viscosities were almost matched. This is surprising since the stress balancing mechanisms in the breakup and withdrawal situation are closely related. Furthemore, at a certain flow rate the tips undergo a topological transition into a steady spout state in which both liquids are entrained simultaneously. We investigate the mechanisms of this transition numerically in a simplified boundary integral model. In the tip state, two lengthscales emerge naturally, the deflection of the interface and the radius of the tip. We study the dependencies of these two lengthscales on each other and on external flow parameters and on boundary conditions.

August 10, 2005
Francois Blanchette, University of Chicago
Simulations of interfacial flows: from multiple coalescence to air bubble pinch off
In this talk, I will present highly accurate simulations of fluid flows in which surface tension is dominant. The position of the interface is modelled using markers, which allows to determine its location very precisely and thus greatly reduces numerical errors near the interface. Simple topological changes such as drop or bubble pinch off and coalescence may also be handled by our method. Applications to the problem of multiple coalescence of a drop coming into contact with horizontal surface will be presented. The mechanism allowing multiple coalescence is described in details for the first time and the conditions in which multiple coalescence may occur are determined. I will also present a second application of my simulations to air bubble pinching off from a nozzle and discuss how such computations may help understand ongoing experiments.

September 7, 2005 (&)
Thierry Emonet, University of Chicago
e-mail:, Faculty contact: Leo Kadanoff, .
Towards a digital bacterium
In recent years, single-cell biology has focused on the relationship between the stochastic nature of molecular interactions and variability of cellular processes. In addressing this problem, most efforts in computational biology have isolated one particular scale of interest, concentrating on either intracellular, cellular or population dynamics. Our long-term goal is to develop a modular computational framework able to cross scales and relate stochastic events at the intracellular level to the behavior of a single cell and ultimately to the dynamics of a population of cells. In this talk I will relate our first steps towards that goal.

As a test-bed for our approach we are using bacterial chemotaxis. /E. coli /bacteria can sense their environment and use that information to control their flagellar motors and move closer to sources of nutrients. To understand how a single /E. coli/ processes information we decomposed the chemotaxis pathway into simpler information processing units. We then modeled each unit analytically and/or numerically, and when possible tested model predictions with data obtained from measurements in single cells. Finally, we constructed a digital bacterium equipped with the necessary modules to perform chemotaxis: receptors, adaptation module, intracellular signal carriers (response regulator), motors and flagella. Digital chemotaxis assays consisting of more than 1000 independent digital cells swimming in a 3D environment reproduced experimental data from both single cells and bacterial populations.

September 21, 2005 (^)
Predrag Cvitanovic, Georgia Institute of Technology
e-mail:, Faculty contact: Wendy Zhang,
Recurrent Coherent States in Turbulent Flows
Elucidating the onset and nature of turbulence in flows such as channel and pipe flows is arguably one of the longest-standing and most fundamental questions in fluid dynamics.

Experimental and computational studies point to the existence and importance of coherent structures. Waleffe's `Self-Sustaining Process' theory together with recent full Navier-Stokes computations of unstable traveling waves in plane Couette, Poiseuille, and pipe flows captures remarkably well qualitatively and quantitatively the turbulent structures recently observed in great detail in several 3-d PVI experiments.

However, turbulence itself does not occur on the steady solutions, but on nearby ergodic attractors. We test the ``recurrent coherrent states'' description of turbulence on a Kuramoto-Sivashinsky model, deploying a new variational method that yields a large number of numerical unstable spatiotemporally periodic solutions. For a small but turbulent system, the attracting set appears surprisingly thin. Its backbone are several Smale horseshoe repellers, well approximated by local return maps, each with good symbolic dynamics.

September 28, 2005
Tobin Sosnick, University of Chicago
e-mail:, Faculty contact: Leo Kadanoff,
Integrating multiple length scales in protein folding
One of the central problems in structural biology is the prediction of a protein's 3D structure from its 1D amino acid sequence. The refolding polypeptide chain and its surrounding solvent generally are too computationally expensive to simulate at the atomic level. However, folding is sensitive to properties of individual atoms. Hence, a major hurdle is the development of a reduced representation of the system which retains the information necessary to accurately specify a proteinís fold. We propose a representation that includes realistic chain conformations along with a potential that depends on detailed chemical properties throughout the chain. In our computational approach, backbone dihedral rotations are the only allowed motions and each side-chain is represented by a single pseudo-atom. Rotation probabilities are obtained from a rotamer library that is built from 1000's of known crystal structures by combining joint information about amino acid sequence and Ramachandran basin occupancies. A statistical potential of mean force for backbone heavy atoms and the side-chain pseudo-atom is obtained from the crystal structures. These two facets are combined to predict native structures of small proteins. Successes, failures, and needed improvements to the model will be discussed.

October 5 , 2006
Wolfgang Losert, University of Maryland
e-mail:, Faculty contact: Wendy Zhang,
Cell motility: Dynamic networks and flexible membranes
Motion of cells in response to external signals is crucial for many biological processes, from wound healing to the spread of cancer. Experimental studies of two physical processes involved in cell motion will be described:
(1) the continusously breaking, rearranging and growing actin scaffolding that gives a cell mechanical strength while it moves. Our work focuses on gradients in actin network properties that play an important role at the leading edge of a moving cell.
(2) the deformations of a model cell membrane to local forcing. Shape instabilities and phase separation characterize the membrane response.

October 12, 2005
Maximino Aldana Gonzalez, Universidad Nacional Autónoma de México
e-mail:, Faculty contact: Leo Kadanoff,
Robustness and Evolvability in Genetic Regulatory Networks.
Living organisms are robust to a myriad of random perturbations. At the same time, they are evolvable, which means that internal perturbations can eventually make the organism acquire new functions and adapt to new environments. It is still an open problem to determine how robustness and evolvability blend together to produce stable organisms that yet can change and evolve. In this talk I will address this problem by studying the dynamical stability of genetic regulatory network models under the process of gene duplication and functional divergence. I will show that an intrinsic property of this kind of networks is that, after the divergence of the parent and duplicate genes, with a high probability the previous functions of the network are preserved and new ones might appear. Furthermore, the robustness observed in the network dynamics is not associated with any kind of gene redundancy. Rather, it seems to be a distributed robustness produced by the collective behavior of the entire network.

October 19, 2005(^)
Sanjay Sampath ,State University of New York Stony Brook
e-mail:, Faculty contact: Wendy Zhang,
The Science and Technology of Thermal Sprays and Droplet based Deposition
Thermal spray, a generic term used to describe a family of deposition processes involving heating/melting of powdered materials in either combustion flame or thermal plasmas, providing momentum transfer and acceleration, and finally impinging the droplets onto prepared surfaces. The resulting impacted and solidified droplets (splats) are the building block of a consolidated thick film coating offering wide ranging opportunities for tailored surface engineering.

An important advantage of thermal spray is the flexibility with respect to feed materials (most metals and ceramics), a single step material consolidation, limited substrate heating, and ability to process under ambient conditions. These benefits have resulted in a highly versatile and flexible process which has translated to a rapidly growing industry (estimated $4B worldwide). However, complexities associated with far-from- equilibrium treatment of materials involving two rapid phase change operations (melting and solidification) have challenged the fundamental understanding of the process. In the same vein, the extremes that the materials are subjected also offer exciting opportunities for exploratory materials research.

This presentation will provide a brief overview of the process and experimental measurements for splat formation on cold and "warm" surfaces, under low pressure conditions, and at various impact velocities.

FIG. CAPTION: Top view micrograph of Zirconia splats formed on cold (A) and warm (B) stainless steel substrates displaying elimination of fragmentation in the latter case. Both were processed under identical process conditions under ambient environments. Similar phenonmena has been observed for wide ranging combinations of droplets and substrates.

October 26, 2005 (^)
Stephen Berry, University of Chicago
e-mail:, Faculty contact: Leo Kadanoff,
The strange ways small systems differ from bulk but tell us about it anyway
While thermodynamics is perfectly valid for very small systems, the behavior of such systems, e.g. atomic and molecular clusters, sometimes seems to be in sharp disagreement with our traditional concepts. The phases and phase changes of small systems, and properties such as their heat capacities (under appropriate conditions) seem very unlike what we have come to expect. However by analyzing such behavior suitably, we not only learn why the small systems seem strange and "badly behaved"; we also get new insights into the behavior of bulk matter. This discussion will concentrate on the ways phase behavior of small systems violates the Gibbs phase rule and our entire notion of sharp phase transitions, yet is really in harmony with fundamental thermodynamics.

November 2, 2005
Ian Foster, Argonne National Laboratory
e-mail:, Faculty contact: Leo Kadanoff,
Service-Oriented Science
My work is frequently motivated by the information technology concerns of "big science", a frequently fascinating source of problems for the computer scientist due to the broad scope and ambitious goals of many scientific communities. I speak here about work that seeks to rethink science's information technology foundations in terms of service-oriented architecture. In principle, service-oriented approaches can have a transformative effect on scientific communities, allowing tools formerly accessible only to the specialist to be made available to all, and permitting previously manual data-processing and analysis tasks to be automated. However, while the potential of such "service-oriented science" has been demonstrated, its routine application across many disciplines raises challenging technical problems. One important requirement is to achieve a separation of concerns between discipline-specific content and domain-independent infrastructure; another is to streamline the formation and evolution of the "virtual organizations" that create and access content. I describe the architectural principles, software, and deployments that I am and my colleagues have produced as we tackle these problems, and point to future technical challenges and scientific opportunities.

November 9, 2005 (^)
Michael Tabor, University of Arizona
email:, Faculty contact: Wendy Zhang,
Modeling the growth, form and function of micro-organisms
Bacterial and fungal microorganisms are everywhere and play many important roles in our environment and health. For example, they can produce antibiotics, bore through concrete, and destroy crops. Many of their functions involve very interesting mechanical processes, often involving the equivalent of enormous forces. This talk will describe the modeling of: (i) actinomycetes, which are a type of filamentary bacteria that produces antibiotics; and (ii) the rice blast fungus, which destroys rice crops. Despite the many biological and functional differences of these two microorganisms, there are certain common modeling approaches at the biomechanical level, and we show how the use of exact, nonlinear, elasticity theory can be used successfully to explain the growth and form of both of them. Further aspects of the function of the rice blast fungus, such as its strong adhesive properties and powerful mechanism to penetrate biological and inert media will be discussed.

November 16, 2005 (^)
Detlef Lohse, Dept. of Applied Physics, University of Twente, Netherlands
e-mail:, Faculty contact: Wendy Zhang,
Bubbles in micro- and nanofluidics
I will flash several of our activities in micro- and nanofluidics.
(i) Ink jet printing
Ink-jet printing is considered as the hitherto most successful application of microfluidics. A notorious problem in piezo-acoustic ink-jet systems is the formation of air bubbles during operation. They seriously disturb the acoustics and can cause the droplet formation to stop. We could show that the air-bubbles are entrained at the nozzle and then grow by rectified diffusion. Both high-speed bubble visualizations and numerical results are presented.
(ii) Bubble nucleation
Bubble nucleation at surfaces is a poorly understood phenomenon. We did visualization experiments at structured hydrophobic surfaces and compared the results with model calculations, in particular focusing on bubble-bubble interactions. It is demonstrated that in the many bubble case the bubble collapse is delayed due to shielding effects.
(iii) Surface nanobubbles
It is a more than 200 year old dogma that fluid flowing along a solid wall does not ``slip'', i.e., it sticks to the boundary. On a macroscale this assumption seems to work extremely well. However, recently, a number of micro- and nano-fluidics experiments and simulations have shed increasing doubts on the dogma, because they show strong indications of partial slip. It has been suggested by de Gennes and others that the observed slip could be explained by the presence of surface nanobubbles. We performed controlled surface force atomic force microscope (AFM) studies for different surface materials and liquid conditions, in order to clarify whether the structures seen in the AFM experiments are indeed consistent with an interpretation as ``nanobubbles''. We have also performed molecular dynamics simulations in order to find out why nanobubbles do not dissolve.

The work is done in collaboration with Jos de Jong, Michel Versluis, Hans Reinten and colleagues from Oce (ink-jet printing), Nicolas Bremond, Manish Arora, and C.D. Ohl (cavitation), Stephan Dammer, S. Yang, and Harald Zandvliet and colleagues (surface nanobubbles).

November 30, 2005 (^)
Alex Barnett, Courant Institute of Mathematical Sciences, New York University
email:, Faculty contact: Wendy Zhang,
High-frequency cavity modes: efficient algorithms, and quantum chaos
The `drum problem' - finding the modes (eigenfunctions) of the Laplacian in a cavity - has a 150-year history including acoustics, quantum mechanics and electromagnetics. Modern applications, such as modeling dielectric micro-cavity lasers, can involve complex geometries and high frequencies: a challenging multiscale problem. I present efficient numerical methods which rely on global basis representations and boundary matching. I will overview the exciting recent convergence of methods developed independently by physicists and numerical analysts. A long-standing question in the field of `quantum chaos' is: To what, if anything, do eigenmodes tend in the high-frequency limit? Do they become spatially uniform, if so, how fast? Can `scars' persist? Using the above tools, I will present numerical evidence towards some answers in planar chaotic cavities.

December 7, 2005
Bernard Derrida, Laboratoire de Physique Statistique, Département de Physique,Ecole Normale Supérieure
e-mail:, Faculty contact: Leo Kadanoff,
Effect of noise on travelling waves (such as in the Fisher Equation)
Travelling wave equations such as the Fisher equation appear as the mean-field limit in a number of problems: reaction-diffusion, evolution in presence of selection, growth,etc. . The main effect of noise at the microscopic scale is to shift the velocity of the travelling waves by an amount which can be predicted by a simple cut-off theory. Other characteristics of the position of the front, such as the diffusion constant, can be predicted by a simple phenomenological theory. In the case of evolution models, the structure of the genealogical trees in presence of selection looks numerically universal and very close to those found for an exactly soluble case.

Brunet E, Derrida B Shift in the velocity of a front due to a cutoff Phys. Rev. E 56, 2597 (1997) Effect of microscopic noise on front propagation J. Stat. Phys. 103, 269 (2001)

Brunet E., Derrida B., A.H. Mueller Munier S. A phenomenological theory giving the full statistics of the position of fluctuating pulled fronts preprint 2005