Boulder School 2017: Student Poster Abstracts

Student Poster Sessions

Please note that while there is not one required or mandatory size, all posters should fit in a space no longer than 36” (91 cm) wide and preferably be no more than about 48” (122 cm) in height. Arrangements can be made for other sizes as necessary.

Session #1 Session #2 Session #3

Session #1

Aghil Abed Zadeh (Duke University)

Title: Avalanches and local force evolution in a granular stick-slip experiment

Abstract: We perform a stick-slip experiment to characterize avalanches for granular materials. In our experiment, a constant speed stage pulls a slider which rests on a vertical bed of circular photoelastic particles in a 2D system. The stage is connected to the slider by a spring. We measure the force on the spring by a force sensor attached to the spring. We study the PDF of energy release and slip size, avalanche shape in time, and other seismicity laws during slip avalanches. We analyze the power spectrum of the force signal and probability distributions to understand the effect of the loading speed and of the spring stiffness on the statistical behavior of the system. From a more local point of view and by using a high speed camera and the photoelastic properties of our particles, we characterize the local stress change and flow of particles during avalanches. By image processing we detect the avalanches, as connected components in space and time, and the energy dissipation inside the granular medium and their PDFs. The PDFs of avalanches obey power laws both at global and local scales, but with different exponents. We try to understand the distribution and correlation of local avalanches in space and the way they coarse grain to the global avalanches.

Irem Altan (Duke University)

Title: Water in Protein Crystals

Abstract: Water occupies typically 50% of a protein crystal, and thus significantly contributes to diffraction signals obtained in crystallography experiments. Separating its contribution from that of the protein is, however, challenging. The intricate protein-water interface and the intrinsically probabilistic nature of solvation make describing the structure of water in those systems rather complex. We compare the solvent structure obtained from diffraction data for which experimental phasing is available to that obtained from constrained molecular dynamics (MD) simulations. The resulting spatial density maps show that commonly used MD water models are only partially successful in capturing biomolecular solvation. In general, their averaged radial distribution is captured with only slightly higher accuracy than their angular distribution, and only a fraction of the waters assigned with high reliability to the crystal structure are recovered. Despite these differences, which are likely due to shortcomings of both the water models and the protein force fields, we nevertheless attempt to infer protonation states of side chains utilizing MD-derived densities, and observe some cases for which reasonably confident assignment is possible.

Ada Altieri (Roma ‘Sapienza’ and Université Paris-Sud XI)

Title: Effective potential in high-dimensional systems near the jamming transition

Abstract: A recent challenging problem concerns the study of glassy systems at low temperature. We present a parallel derivation of the Thouless-Anderson-Palmer equations and of an effective thermodynamic potential in two high-dimensional models: the negative perceptron and its generalization to sphere systems. They both define continuous constrained satisfaction problems with a critical jamming transition characterized by the same exponents. The disclosed method enables us to study the vibrational spectrum and to deepen the diverging behavior of the stiffness coefficients. A pivotal feature emerging close to jamming is that the effective thermodynamic potential has a subleading logarithmic contribution, which turns out to be dominant in a suitable scaling limit. A more detailed analysis of higher order corrections to the potential, beyond the first two moments of the expansion, might also help in accounting for finite-dimensional systems and interpolating between different regimes.

Daria Atkinson (University of Massachusetts, Amherst)

Title: Geometric Constraints on Isodistance and the Intrinsic Geometry of Filament Bundles

Abstract: Assemblies of one-dimensional filaments appear in a wide range of physical systems, from biopolymer bundles, organogel fibers, columnar liquid crystals, vortex arrays, and nanotube yarns, to everyday macroscopic examples, like ropes, textiles or even hair in a ponytail. The strength of interactions between the different chains in these diverse examples is dominated by the shortest distance between them, leading to common geometric constraints.  When the bundle is twisted—which is often the case for self assembled chiral molecules—or has an otherwise nontrivial geometry, the packings of the filaments are frustrated, and the local ordering is decidedly nontrivial.  Frustration in filamentous assemblies is particularly insidious since there is in general neither uniform spacing between filaments at different points in the bundle or at different positions between the same fibers. Remarkably, for certain geometries of these filament bundles, the packing of filaments can be mapped onto another class of geometrically frustrated problems: the close packing of disks on curved surfaces.  For the bundle to map to a unique 2D geometry, the distances of closest approach between any two filaments must be constant.  Here we show that the only families of twisted curves that satisfy this constraint of isodistance are helical bundles, and describe several generalizations of the approach.

Christopher Baldwin (Boston University)

Title: Hamiltonian Dynamics of Quantum Spin Glasses

Abstract: I study the behavior of quantum spin glasses, when treated as isolated systems evolving under the Schrodinger Equation. An important question is whether the dynamics is ergodic, and how that ergodicity correlates with the thermodynamic phase diagram. I show that the canonical mean-field models exhibit dynamical phase transitions as the strength of an applied transverse field varies. These transitions have little to do with the equilibrium glass transition, but are much more related to the clustering transition that mean-field models exhibit. These results highlight the important role that configuration-space clustering plays for mean-field and long-range models, influencing not only their thermodynamic behavior but their isolated Hamiltonian dynamics as well.

Karsten Baumgarten (Delft University of Technology)

Title: Nonlocal Elasticity Near Jamming

Abstract: We use simulations of frictionless soft sphere packings to identify novel constitutive relations for linear elasticity near the jamming transition. By forcing packings at varying wavelengths, we directly access their transverse and longitudinal compliances. These are found to be wavelength dependent, in violation of conventional (local) linear elasticity. Crossovers in the compliances select characteristic length scales, which signify the appearance of nonlocal effects. Two of these length scales diverge as the pressure vanishes, indicating that critical effects near jamming control the breakdown of local elasticity. We expect these nonlocal constitutive relations to be applicable to a wide range of weakly jammed solids, including emulsions, foams, and granulates.

Farzan Beroz (Princeton University)

Title: Physical limits to biomechanical sensing

Abstract: Cells actively probe and respond to the stiffness of their surroundings. Since mechanosensory cells in connective tissue are surrounded by a disordered network of biopolymers, their mechanical environment can be extremely heterogeneous. Here, we investigate how this heterogeneity impacts mechanosensing by modeling the cell as an idealized local stiffness sensor inside a disordered fiber network. For all types of networks we study, including experimentally-imaged collagen and fibrin architectures, we find that measurements applied at different points yield a strikingly broad range of local stiffnesses, spanning roughly two decades. We verify via simulations and scaling arguments that this broad range of local stiffnesses is a generic property of disordered fiber networks. Finally, we show that to obtain optimal, reliable estimates of global tissue stiffness, a cell must adjust its size, shape, and position to integrate multiple stiffness measurements over extended regions of space.

Cong Cao (Emory University)

Title: Aging near the wall in dense colloidal samples

Abstract: We study dense (glassy) bidisperse colloidal samples near a wall with both smooth and rough boundary conditions. Using a confocal microscope we directly observe the slowing of particles’ motion (aging phenomena) in both samples. With a smooth boundary, due to the wall induced layer-like structures, we notice that particles’ dynamics slow down when near the wall, with aging process quench faster than the bulk area. Motions along the z direction are severely restricted by the wall, leading to fewer particle exchanges between layers. With the rough boundary, the layer-like structures greatly diminishes. At the same time, the particles’ dynamics shows no distinct differences between boundary and bulk. 

Shyr-Shea Chang (UCLA)

Title: Optimal occlusion uniformly partitions red blood cells fluxes within a microvascular network

Abstract: In animals, gas exchange between blood and tissues occurs in narrow vessels, whose diameter is typically less than that of a red blood cell. Red blood cells must deform to squeeze through these narrow vessels transiently blocking or occluding the vessels they pass through. Although the dynamics of vessel occlusion have been studied extensively, it remains an open question why microvessels need to be so narrow. We study occlusive dynamics within a model microvascular

network: the embryonic zebrafish trunk. We show that pressure feedbacks created when red blood cells enter the finest vessels of the trunk act together to uniformly partition red blood cells through the microvasculature. Using models as well as direct observation, we show that these occlusive feedbacks are tuned throughout the trunk network to prevent the vessels closest to the heart from short-circuiting the network. Thus occlusion is linked with another open question of microvascular function: how are red blood cells delivered at the same rate to each micro-vessel? Our analysis shows that tuning of occlusive feedbacks increase the total dissipation within the network by a factor of 11, showing that uniformity of flows rather than minimization of transport costs may be prioritized by the microvascular network.

Hanrong Chen (Harvard University)

Title: HIV population dynamics on a high-dimensional fitness landscape

Abstract: When HIV infects a host, its high mutability allows it to rapidly generate mutants that escape recognition by the host’s immune response. The HIV population within a host accumulates diversity over time, which depends on the host’s immune response, as well as intrinsic fitness constraints of the virus.

Here, we describe statistical physics inspired methods used to study aspects of intra-host HIV dynamics. First, we recap the HIV “fitness landscape” which captures information about the replicative capacity of the virus as a function of its protein sequence. We then introduce our method of simulating HIV population dynamics, which is based on extending the Wright-Fisher model to time-varying population size. Importantly, our model treats viral protein sequences, and immune targeting of specific regions of the viral sequence, explicitly, which enables us to study how HIV dynamics depends on the strength and locations of the host’s immune response. Simulating population dynamics on a high-dimensional fitness landscape is computationally intensive, but we have developed a mean-field–like approximation that considerably improves simulation time. Finally, we present an application of our methodology to remarkable experiments in monkeys involving a novel form of vaccination, which leads to protection in roughly half of them from SIV infection.

Xiaowen Chen (Princeton University)

Title: Inference of whole brain interaction with a minimal pairwise probability model

Abstract: Brain dynamics emerges as the collective behavior of many interacting neurons such that exhaustive sampling is impossible. In recent years, binary pairwise probability models, which are equivalent to Ising models, have found much success in reducing the dimension while describing network behavior quantitatively for studying the statistics of spike trains of neuronal populations. Using Caenorhabditis elegans as a model system, we show that this formulation can be generalized to describe 1) neural networks with graded (continuous) potential and 2) systems of neurons with size close to the whole brain by mapping neuronal signals to q-Potts models.

Simone Ciarella (Eindhoven University of Technology)

Title: Mechanic of reversible bonding in polymer networks

Abstract: Vitrimers are a recent new class of polymer networks which are characterized by a dynamic bond exchange mechanism. This feature allow them to interpolate between thermosets and thermoplasts and offers a unique combination of malleability and performance. In my project I am studying through computer simulation the effect of bond reversibility on the static and the dynamic of a polymer network, trying to model a framework to encompass the behaviour of vitrimers and exchange materials in general.

Peter Crowther (University of Bristol)

Title: Moving closer to the glass transition with real space imaging of colloidal suspensions

Abstract: Suspensions of colloidal particles can provide insight into the behaviour of condensed matter since their equilibrium behaviour is thermodynamically equivalent to that of molecules. Three dimensional measurements of colloidal suspensions can be made using confocal microscopy from which the structure and dynamics of the colloids can be determined. The limitation of this technique is that the smallest particles that can be reliably measured are around $2,\mathrm{\mu m}$ in size, which limits the dynamical regime that can be studied to around $10^5$ the single particle relaxation time ($\mathrm{tau_0}$). Since the molecular glass transition is around $10^{14},\mathrm{tau_0}$, the observed behaviour is a long way from the glass transition. We push further towards the glass transition using STED microscopy and differential dynamic microscopy methods to probe relaxation dynamics and dynamical heterogeneity, measuring particles of $100,\mathrm{nm}$ diameter and accessing relaxation times 4 orders of magnitude greater than that previously accessible in particle resolved studies.

Eduardo Domínguez (University of Havana)

Title: Cluster Variational Method for disordered quantum systems

Abstract: Based on the well known Cluster Variational Method [1] and with the aim at describing the local properties of disordered quantum models we develop an equivalent of a message passing algorithm that can be regarded as an extension of the Generalized Belief Propagation[2].  This technique allows for the determination of microscopic observables with better results than the standard mean field approach. To the level of the Bethe approximation for tree-like lattices we obtain numerical results for the random field Ising Model that are in good agreement with the quantum cavity method[3] and stochastic Monte Carlo simulations. We also report the results obtained for the plaquette approximation in two dimensional lattices.

[1] Pelizzola A 2005 J. Phys. A: Math. Gen. 38 R309
[2] Yedidia J, Freeman W T and Weiss Y 2005 IEEE Trans. Inf. Theory 51 2282
[3] F. Krzakala, A. Rosso, G. Semerjian and F. Zamponi, Phys. Rev. B 78, 134428 (2008).

Carlos M. Duque (University of Massachusetts, Amherst)

Title: Surface Design Based on Discrete Conformal Transformations

Abstract: Conformal transformations are angle-preserving maps from one domain to another. Although angles are preserved, the lengths between arbitrary points are not generally conserved. As a consequence there is always a given amount of distortion associated to any conformal map. Different uses of such transformations can be found in various fields, but have been used by our group to program non-uniformly swellable gel sheets to buckle into prescribed three dimensional shapes. I discuss how circle packings can be seen as a kind of discrete conformal map in order to find conformal maps from the sphere to the plane that can be used as nearly uniform swelling patterns to program non-Euclidean sheets to buckle into spheres. 

Davide Facoetti (King’s College London)

Title: From non-ergodic eigenvectors to local resolvent statistics and back: a random matrix perspective

Abstract: The phenomenon of many-body localisation (MBL) has attracted considerable interest in recent years in the study of non-equilibrium quantum sytems. A feature that is not well understood is the existence of an intermediate extended non-ergodic “bad metal”phase between the localised and extended ones. To help the debate about the existence and role of extended, non-ergodic phases in MBL and the related problem of Anderson localisation on the Bethe lattice, we study such a phase in the simpler setting of the generalised Rosenzweig-Porter random matrix model. We confirm previous results about the multifractality of the eigenstates, and show that the intermediate phase can be characterised studying the local resolvent in a non-standard scaling limit.

Reference: DF, P. Vivo, G. Biroli, EPL 115 (2016) 47003, arXiv:1607.05942

Session #2

Gabriel Canova (Universidade Federal do Rio Grande do Sul)

Title: Competing nematic interactions in a generalized XY model

Abstract: Long-range order at finite temperature is not stable in two-dimensional systems with continuous symmetry (the variables are angles) because a spin wave with long wavelength costs too little energy (along a given direction, a wave sets up with neighboring spins having a small angle difference between them). Nonetheless, it is well known that the 2D XY model, whose energy decreases when neighboring spins are parallel, exhibits an unusual phase transition, named after Kosterlitz, Thouless and Berezinskii (BKT). We study a generalization of this model which includes nematic-like interactions (this energy term, instead, decreases when the alignment is skewed by 180º, 120º, 90º, …, 2π/q, where q is an integer). The competition between these two terms leads to new ordered phases with new topological defects, like fractional vortices and domain walls, and, consequently, a more complex phase diagram. We study and characterize the new properties of  this model through Monte Carlo simulations on GPUs.

Joel Clemmer (Johns Hopkins University)

Title: Anisotropic Exponents for Avalanche Correlation Lengths in Self-Affine Growth of Magnetic Domains

Abstract: Driven interfaces in a wide variety of systems undergo a critical depinning transition as the driving force is increased to a critical value. Near this transition, growth consists of discrete avalanches with a power law distribution of sizes and a diverging length scale along the interface. We simulate depinning of a self-affine domain wall in the 3D random field Ising model to determine the how the ratio of avalanches sizes scale as avalanches grow larger. Analyzing individual avalanches show that the height and width along the interface scale sublinearly as a power law with an exponent of about 0.9 over 3 decades in systems of 1010 spins. This value of the scaling exponent is considerably larger than the roughness exponent. Scaling theories often assume the anisotropy is described by the roughness exponent. We conrm the value of the anisotropy exponent through finite size scaling of various system properties.

Varda Faghir Hagh (Arizona State University)

Title: Tunneling states in jamming and glasses

Abstract: We study the tunneling states in jammed and glassy networks, by employing a theorem that says generic isostatic frameworks are not globally rigid [1]. According to this theorem, any generic isostatic framework has an even number of equivalent configurations that are all locally rigid. To find such configurations, we remove a bond from the network, resulting in a mechanism that has a configuration space that is a closed, continuous curve. On this curve, there are an even number of points in which the removed bond reverts to its original length. These configurations provide a possible explanation for tunneling modes in glasses [2] but are a surprise in jamming. In a large system with N bonds, we can perform this procedure by removing every bond in turn. However only a few of these removals would lead to an independent configuration and in case of jamming, the number of re-packable configurations is even smaller.

The results are presented and compared for two-dimensional systems of both corner sharing triangles and jammed systems of disks [3]. We conjecture that these tunneling states collapse onto a single state at the densest random packing, making that state non-generic.
[1] Gortler SJ, Healy AD, Thurston DP. Characterizing generic global rigidity. American Journal of Mathematics. 2010;132(4):897-939.; Hendrickson B. Conditions for unique graph realizations. SIAM journal on computing. 1992 Feb;21(1):65-84.
[2] Phillips WA. Tunneling states in amorphous solids. Journal of Low Temperature Physics. 1972 May 1;7(3):351-60.; Anderson PW, Halperin BI, Varma CM. Anomalous low-temperature thermal properties of glasses and spin glasses. Philosophical Magazine. 1972 Jan 1;25(1):1-9.
[3] Work in collaboration with M.F. Thorpe, M. Sadjadi, E. Corwin, S.J. Gortler, R. Connelly, L. Theran, M. Sitharam and M. Holmes-Norton.

Marylou Gabrié (École Normale Supérieure de Paris)

Title: Mean-Field Framework for Unsupervised Learning

Abstract: Boltzmann machines are undirected neural networks useful for unsupervised machine learning. In particular, a simple bipartite version - called Restricted Boltzmann machines (RBMs) - has been widely popularized by the discovery of fast training algorithms, relying on approximate Monte Carlo Markov Chains. Realizing that training RBMs is closely related to the inverse Ising problem, a notoriously hard statistical physics problem, we designed an alternative deterministic procedure based on the Thouless-Anderson-Palmer approach. Our algorithm, improving on the naive mean-field approximation, provides performance equal to the commonly used MCMC algorithms while also providing a clear and easy to evaluate objective function to follow progress along training. Moreover, this strategy can be generalized in many ways, including for new network architectures or for new types of data. Finally, a particularly exciting application of this new framework is the integration of learned priors in Bayesian inference problems. 

Marco Galvani (Johns Hopkins University)

Title: Effect of Flow-Induced Molecular Alignment on Welding of Polymer Interfaces 

Abstract: Additive manufacturing, a process of successive deposition of layers of polymer used to synthesize objects, is quickly developing into an effective method of creating polymer-based materials. The physical properties of materials produced by additive manufacturing depend strongly on the mechanics of welded interfaces where polymer chains diffuse between contacting layers. The degree of interdiffusion and the resulting entanglement structure are the dominant factors that give rise to the interfacial strength. In this project, we use large molecular dynamics simulations to examine how flow-induced molecular alignment from the deposition process affects weld formation and strength. First the relation between alignment and entanglement loss is studied for model polymers of different entanglement lengths. Then the effect on the rate of interdiffusion is determined and the formation of interfacial entanglements is correlated to the rate of increase in interfacial strength towards the bulk value.

Guram Gogia (Emory University)

Title: Stochastic Resonance-induced Periodic Melting in 2D Dusty Plasma Crystals

Abstract: Stochastic Resonance (SR) is a common cooperative phenomenon identi ed by enhancement of weak signal by certain amount of noise. It has been observed in wide variety of electrical, biological, chemical and physiological systems. The effects of stochastic resonance in excitable spatially extended media has been mostly studied numerically and analytically. There has been lack of experimental model systems that would allow study of SR in a collective system on individual particle length-scale. In this work, we present a direct observation of SR in two dimensional dusty plasma crystals. At low plasma pressures, density fluctuations force particles in the crystal mono-layer to oscillate vertically in and out of the sample plane. Gradual increase of the oscillation amplitude leads to melting of the crystal and the system enters a gas-like phase, characterized with high kinetic energy. Without changing any external parameters, collisions with neutral atoms slow down the particles and the system recrystallizes and remains stable until next “punctuating” instability. By treating the polydispersity of the particles and plasma density fluctuations  as two separate noise sources, we captured the salient features of SR through simple Langevian simulations - for certain polydispersity and the magnitude of driving, we observe a characteristic stochastic resonance peak in intensity of hopping frequency between melted and crystalline states.

Yi Hu (Duke University)

Title: Clustering of Microphase Former in One-dimensional Continuous Space: Exact Results and Event-chain Simulations

Abstract: Microphases are ubiquitously formed by particles interacting through short-range attraction and long-range repulsion (SALR), and microscopic details of interactions significantly affect the resulting rich microphase transition behavior. The strong frustration generically present in these systems, however, impedes the determination of the equilibrium phase behavior. In one-dimensional case the partition function nonetheless can be calculated exactly. Here, we calculate the partition function of SALR microphase formers in one dimension by the transfer-matrix method, considering the interaction range up to next-next-nearest-neighbor particles. By analyzing the equation of state, the critical micelle concentration (CMC) can be found at low temperatures while such a micellization behavior is absent above certain critical temperature. This qualitative picture is also verified through the direct calculation of virial coefficients. Event-chain based Monte Carlo simulations, introduced to overcome the sampling problem of naive Metropolis algorithm at low temperatures, further produce the results consistent with those obtained by the transfer-matrix method.

Andrea Kadović (Ruđer Bošković Institute, Zagreb, Croatia)

Title: Percolation Critical Behaviour of Colored Networks

Abstract: Network has typically been abstracted to be a set of identical objects with identical kind of bonds in the studies of percolation on complex networks. Recently, it was shown that the colored network in which the color is an additional degree of freedom allocated to every either vertex or edge, is a more sensible representation of complex systems subject to different classes of vulnerabilities. On this poster we provide analytical and numerical evidence of a color affected mean-field critical behaviour for a class of networks. To the contrary, we also find that the critical behaviour stays topological (color independent) for networks defined with the power law degree distribution of exponent α ∈ <2, 3> ∪ <3, 4].

Shachi Katira (University of California, Berkeley)

Title: Pre-transition effects in trajectory space

Abstract: Thermodynamic first-order phase transitions exhibit pre-transition effects such as pre-melting, which can occur in regions of the phase diagram far from co-existence. Here we illustrate pre-transition effects for dynamical first-order phase transitions using the East model as an example.

Dion Koeze (TU Delft)

Title: Jamming in constrained geometries

Abstract: Does jamming play a role when forming a Pickering emulsion? What is the influence of the narrow domain in an inclined plane flow? The role of constraining the geometry of a jammed packing is largely unexplored. By studying systems with extreme aspect ratios or by pinning particles to a curved surface we explore the role of geometry in jamming.

Charuhansini Kulkarni (National Centre for Biological Sciences, Tata Institute of Fundamental Research)


Abstract: How do the intrinsic physical properties of the cytoplasm facilitate survival of cells during freezing and thawing? Not all organisms can survive (i.e. restore normal structure and function) complete freezing and thawing of their bodies. Many freeze-thaw tolerant organisms produce molecules such as trehalose, facilitating the liquid cytoplasm to transition to a glassy state, which is speculated to aid in their survival upon thawing. Indeed, glassiness of the cytoplasm is a recurrent feature in many organisms which aids in their survival in stressful conditions. Under normal conditions, the cytoplasm is known to be an active liquid fluidised by the various forces generated inside the cell. [1,2] We study the factors that govern the active liquid–glass transition of the cytoplasm during freezing and the glass–active liquid re-entry upon thawing. In particular, we aim to understand how these transitions are set by physical parameters (temperature, cooling rate, thawing rate)  and to what extent they are affected by the active regulation of the biochemical state of the cell.

We study these questions using yeast cells which are known to utilise enhanced trehalose production in response to freezing. Currently, we are setting up experiments using microrheology techniques to measure the physical properties of the cytoplasm during the freeze thaw process. These measurements are complemented by biochemical assays to quantify the amount and rates of trehalose production under various physical parameters and how they relate to the rheological state of the cellular cytoplasm. Further, we are evolving increased freeze-thaw tolerance  in yeast in order to elucidate the different adaptations that might aid in surviving freezing and thawing, namely physical adaptations to the cytoplasm and the biochemical adaptations to the regulatory pathways triggered upon freeze induction. These evolutionary outcomes are identified through genome sequencing and in combination with the physico-chemical measurements mentioned earlier, they may then be used to inform us about the fitness landscape across which the populations traverse.
1. Fakhri, N., Wessel, A. D., Willms, C., Pasquali, M., Klopfenstein, D. R., MacKintosh, F. C., & Schmidt, C. F. (2014). High-resolution mapping of intracellular fluctuations using carbon nanotubes. Science (New York, N.Y.), 344(6187), 1031–5.
2. Parry, B. R., Surovtsev, I. V., Cabeen, M. T., O’Hern, C. S., Dufresne, E. R., & Jacobs-Wagner, C. (2014). The bacterial cytoplasm has glass-like properties and is fluidized by metabolic activity. Cell, 156(1–2), 183–194.

Cathy Li (Duke University)

Title: Free energy of frustrated spin chain model

Abstract: How should the free energy of a frustrated spin system be calculated? We here apply the Monte Carlo algorithm, the replica trick, and the cycle-expansion method to study disordered Ising spin chains with various interaction ranges. Besides the standard difficulties associated with the replica trick, a sharp turn of the extrapolating function near the zero replica limit at low temperature leads to the effective failure of directly applying this scheme. The cycle expansion, which has a subtle connection to the replica trick, overcomes this obstacle and successfully obtains higher-order derivatives of the extrapolating function, thereby attaining detailed information about the distribution of the free energy. In particular the cycle-expansion method, when computationally feasible, yields a far more effective way of accessing the tail of the sharply-peaked distribution that is hard to obtain through direct Monte Carlo simulations.

Maxwell Marple (University of California, Davis)

Title: Stoichiometric Glasses in the System Li2S-Ga2Se3-GeSe2 : Structural Characterization and Fast Li ion Conduction

Abstract: Homogeneous glasses in the mixed-chalcogen pseudo-ternary system Li2S-Ga2Se3-GeSe2 are synthesized and their structure is characterized using Raman and one– and two-dimensional 6Li, 77Se, and 71Ga nuclear magnetic resonance (NMR) spectroscopy. The structure of these glasses can be described as a charge-compensated network predominantly consisting of corner sharing (Ga/Ge) (Se,S)4/2 tetrahedra. The compositional evolution of the atomic structure is heavily influenced by the Li2S:Ga2Se3 ratio R where charge compensation is accommodated by the formation of different structural units and preferential chemical ordering for S atoms.  Glasses with R<1 are deficient in chalcogens required to satisfy the tetrahedral coordination of Ga and consequently form ethane-like X3Ge-GeX3 (X=S, Se) units with S atoms preferentially participating in these structural units.  On the other hand, the structure of chalcogen-excess glasses with R>1 are characterized by the formation of non-bridging Se (NBSe) and S (NBS) sites.  The Se atoms show a preference over S for these non-bridging sites and form Ge-NBSe linkages, while the S atoms preferentially bond to Ga, resulting in the formation of GaS4/2 tetrahedra.  Ionic conductivity of these glasses is measured using electrical impedance spectroscopy and is found to monotonically increase with increasing Li concentration, displaying the highest dc conductivity of ~ 10-4 S/cm at ambient temperature. Motional narrowing of the 7Li NMR linewidth and the frequency dependence of conductivity reveal similar hopping frequencies suggesting the concentration of mobile Li ions rather than their mobility is the limiting factor for achieving high ionic conductivity in this system. 

Zhe Mei (Yale University)

Title: Void Study of Protein Cores

Abstract: There has been continued interest in the arrangements of hydrophobic amino acid residues in protein cores due to their large contribution to the stability of proteins.  Prior study has reported that protein cores possess a packing fraction of φ ≈ 0.56, which is similar to values for random close packing of nonspherical particles. This packing fraction is recovered by packing simulations of mixtures of residues that are isotropically compressed to jamming onset. Here we apply the analysis of void spaces and compare the void distributions in protein crystal structures, hydrophobic residue packing and other random close packing systems. The similarity of void spaces between protein cores and amino acid residue packing shows the potential of simplifying the complicated protein stability and mutation problem into a simple mechanical system.

Peter Morse (Syracuse University)

Title: Rearrangements and Reversibility in Sheared 2d Systems

Abstract: Under shear, jammed packings of particles transition between mechanically stable states through saddle points in the energy landscape. These saddle points correspond with a change in the contact network of the packing. It has long been assumed that this process is one-to-one and that all contact changes correspond with a saddle point. However, I will demonstrate that a vast majority of contact changes do not have an associated saddle point, and thus remain in the same well. We call these “network events” and we call contact changes that pass through a saddle point “rearrangements.” I will show that network events are reversible under cyclic strain and that rearrangements can be either reversible or irreversible, and I will explore the conditions that separate those two scenarios.

Aya Nawano (Yale University)

Title: Crystallization and glass formation in binary metallic glasses

Abstract: Bulk metallic glasses (BMGs) are amorphous alloys that have favorable mechanical properties such as enhanced yielding strength and elasticity compared to conventional alloys. We perform molecular dynamics simulations of BMGs using the embedded atom model, which includes both pairwise and many-body interactions, to study the glass forming ability as a function of different atomic species and relative concentrations for metal-metalloid binary alloys. We seek to determine the parameter regime that yields the lowest critical cooling rates and best glass-formers. We also present measurements of the nucleation and growth of crystal clusters as a function of cooling rate, system size, and the nature of the boundary conditions. In particular, we will clarify why the critical cooling rate decreases monotonically with increasing system size for numerical simulations with periodic boundary conditions, whereas experimental studies show non-monotonic behavior for the critical cooling rate versus system size.

Fernanda Pereira da Cruz Benetti (Sapienza Università di Roma)

Title: Effective Mean-Field Model for Jammed Spheres

Abstract: Systems of athermal soft spheres interacting through repulsive harmonic potentials are mechanically stable if the total number of contacts between spheres is at least dN, where d is the dimensionality of space and N is the number of spheres. For exactly dN contacts, it is marginally stable: if any contact is broken, the system gains a degree of freedom. Near this point, corresponding to the jamming transition, the frequency spectra presents anomalies such as an excess of soft modes in relation to the prediction of the Debye model.

In order to gain insight into these modes and their localization, we study a model based on random regular graphs (RRG): a random spring model. The equilibrium position of each soft sphere corresponds to a vertex on the graph, and the edges connecting the vertices are related to their interaction energy. For a pair of overlapping spheres, this energy is proportional to the square of their overlap; otherwise, it is zero. By taking a Hamiltonian in which contacts occur between random spheres instead of neighboring ones, we may apply the RRG theoretical framework, allowing us to compare analytical results with numerics.

Abby Plummer (Harvard University)

Title: Population genetics in compressible turbulence

Abstract: Many classic population genetics results, such as the probability and time of fixation, change dramatically in the presence of turbulent flows. Using a shell model for turbulence and agent-based simulations in one dimension, we investigate these departures from the well-mixed theory. The deviations are found to be linked to the spatial structure of the flow, rather than its time dependence. By both characterizing the flow profile in terms of sources and sinks and examining the behavior of the population under various deterministic flows, we gain insight into how turbulence alters fixation events.

Session #3

Benjamin Aubin (Ecole Normale Supérieure de Paris)

Title: Capacity storage of the perceptron

Abstract: Multi-layer neural networks are nowadays a top-research priority in reason of their increasing use in various areas. However the theoretical foundations remain essentially empirical and we suggest deepening our understanding of the neural networks with the study of a simple machine learning structure: the perceptron. The perceptron capacity problem was studied by the replica method in a series of physics works in the 80’s. As a continuation of these former works and within the statistical physics framework, we aim to retrieve well known results in this simple case, in order to establish assuredly our methodology, before tackling more complex multi-layer neural networks. We present a fairly general method to study the information storage capacity of the perceptron in various settings and in particular we emphasize the impact of the activation function choice on the storage capacity.

Noah Rahman (Tulane University)

Title: Many-body localization in Ising-like models with long range interactions

Abstract: The phenomenon of many-body localization in random (disordered) long-range spin-spin interacting Ising-like models (in the XY universality class) is studied via methods including spectral statistics and numerical renormalization group, with an eye towards characterizing the many-body delocalization transition for certain ranges of interaction (with the exponent between 1 and 2). Predictions are made with an eye to experimental testability in ion trap emulators with programmable long-range interactions, including glassy dynamics in the zero disorder limit.

Riccardo Ravasio (Ecole Polytechnique Federale de Lausanne)

Title: Architecture and coevolution of allosteric materials

Abstract: We consider the phenomenon of allosteric regulation, which requires the protein to respond specifically to a stimulus at a distant active site. Several proteins are reported to use two specific mechanical architectures to reorganise themselves after a perturbation: hinge- and shear-like movements. At the same time, not all structural rearrangements follow this classification. We build a disordered elastic model which is able to reproduce such architectures. It  also unravels a new mechanism for allosteric regulation, extending the aforementioned classification. Moreover, given a family of structures accomplishing the same task, we ask whether we can identify constraints in evolution associated to this specificity. Our aim is to extract relevant information on function, structure and mechanics just from a statistical analysis of the different solutions we obtain from the model. Finally, we compare the results of the model with the information one can extract from real data. 

Dominic “Nick” Robe (Emory University)

Title: Dynamic Heterogeneity and Aging in a 2D Colloidal Simulation

Abstract: Physical aging is a non-equilibrium process with a time scale which grows after the quench. Glassy systems also exhibit dynamic heterogeneity, or localized rearrangements, which cover a growing length scale and become rarer as the system ages. The theoretical framework of record dynamics suggest that the growing time scale during aging is related to these rearrangements. Here we study the aging of dynamic heterogeneity in a system of strongly repulsive disks. We find that the dynamics collapse to the forms predicted by record dynamics, and we are working on a way to characterize the size of rearrangement events.

Joshua Robinson (University of Bristol)

Title: Free energy of local structure in hard spheres

Abstract: Local structure is key to understanding disorder and frustration. However, a priori predictions of local structure is challenging in systems with annealed disorder. We formulate the free energy of a local arrangement of particles within a bulk system in terms of solvation forces. For hard spheres we can then employ a morphological formulation of solvation to compute the local library of states. From this we obtain all of the thermodynamics for local structures in this system. We find the locally favoured structures to be rich in 5-fold symmetry, in line with geometric frustration arguments for dynamical arrest.

Stefano Sarao (Politecnico di Torino)

Title: Phase Transitions in Hierachical Community Detection

Abstract: In this work we use an approximate version of Belief Propagation, called Approximate Message Passing (AMP), to study the possibility of doing inference in a generalized Community Detection problem. The generalization consists in having different levels of communities nested hierarchically. Our aim is to understand under what conditions an algorithm is able to reconstruct the communities at the various levels. AMP provides us both a powerful algorithm and a tool to analyze the evolution of the system in the thermodynamic limit.

Camille Scalliet (Universite de Montpellier)

Title: Is there a Gardner transition in soft glasses?

Abstract: Recent theoretical advances in the mean-field theory of hard-sphere glasses predict the existence of the so-called Gardner transition, an ergodicity breaking transition that takes place deep in the glass phase. In hard-sphere glasses, this transition is crucial to make theoretical predictions for the jamming transition occurring at higher densities.  Our goal is to determine if the Gardner transition is also a crucial element to understand the low temperature behavior of soft glasses, in which particles interact via a continuous potential. We use two complementary theoretical approaches: a mean-field study of a soft glassy model in the limit of infinite dimensions, combined to a numerical investigation of the transition in a simple three-dimensional soft glass-former. Analytical results confirm the existence of the Gardner transition even in soft glasses. Numerical investigations of three-dimensional soft glasses, however, do not detect the presence of a sharp Gardner crossover at low temperatures.

Alexandra Signoriello (Yale University)

Title: Computational Modeling of Two-Dimensional Epithelial Sheets

Abstract: Structural and mechanical properties regulate cell migration, interaction forces, and packing geometry during tissue development. We have developed a new model for tissue development in two spatial dimensions (2D) that includes different rates for cell growth, cell-cell interactions, and extracellular matrix. Cells are represented as polygons and the total energy of the system includes contributions from cell elasticity, contraction, and excluded volume. We study the formation of confluent tissues by slowly increasing cell sizes followed by energy minimization. We then measure the structural and mechanical properties of the tissue as a function of the cell density. The results from our simulations will be compared to experiments that are able visualize the spatiotemporal evolution of monolayers of epithelial sheets.

Stefano Spigler (LPTMS, CNRS, Univ. Paris-Sud, Université Paris-Saclay)

Title: ”Distribution of avalanches in disordered systems

Abstract: Disordered systems are characterized by the existence of many sample dependent local energy minima, that cause a step-wise response when the system is perturbed. We used an approach based on elementary probabilistic methods to compute the complete probability distribution of the jumps (static avalanches) in the response of mean-field systems described by replica symmetry breaking. In particular we found a power-law distribution in infinite-dimensional systems of soft-spheres, with different exponents at jamming and in the marginal glass phase (at higher packing fractions); a consequence of such a distribution is a diverging sample-to-sample variance of the shear modulus. Remarkably, numerical simulations of 3-dimensional systems confirm these results.

Watee Srinin (New York University)

Title: Scale-Dependent Viscosity in Ring and Linear Polymer Fluids

Abstract: Traditional polymer rheology considers the response of the polymer fluids caused by external forces applied on a macroscopic scale. By contrast, in live cells, and especially in the nucleus, one encounters the situation in which forces are applied on a molecular scale, for instance, by molecular motors or other active ATP-consuming molecules. The influence of these locally applied forces can cascade upwards and lead to large-scale motion of the surrounding fluids which can not be captured by the traditional notion of viscosity. In this work, we examine the response of polymer fluids to forces applied at finite wave vector and frequency. Using simple physical arguments we construct a “phase diagram” of various frequency and wave vector-dependent regimes of effective viscosity for polymer fluids, including non-entangled and entangled melts, semidilute solutions with and without hydrodynamic interactions, as well as the more exotic case of a melt of unconcatenated ring polymers.

Ethan Stanifer (Syracuse University)

Title: Investigating the Vibrational Properties of Disordered Solids Using Random Matrix Models

Abstract: Glasses and jammed packings exhibit interesting low-frequency vibrational behavior, which is important for understanding flow and failure in these materials. Recent work indicates simple random matrix models can provide a simple explanation for universal vibrational properties in glasses [Manning and Liu EPL 2015]. Here we analyze 2-regular random graphs and study two sources of disorder: randomized bond strengths and bonds added to distort the network. In the model with both types of disorder, we find a low frequency scaling consistent with simulation results [Lerner During Bouchbinder PRE 2016]. We consider pure crystals, random matrices with no disorder, and verify they cannot reproduce this effect. We find bond disorder is sufficient to produce the low-frequency scaling while network disorder controls the scaling transition.

Menachem Stern (University of Chicago)

Title: Controlling large deformations of marginal disordered structures

Abstract: Metamaterials are typically sought to demonstrate specific responses in the non-linear deformation regime. However, analytic methods for these systems are often based on linear approximations. We find that practical questions of actuation in origami have counter-intuitive answers due to a strong mismatch between linear and non-linear theory near the special flat state, where all modes meet. Linear-non-linear mismatch generically leads to an exponential number of `dead end’ folding modes, resulting in an emergent glassy energy landscape around the flat state. This landscape makes refolding of a pre-folded creased sheet much more difficult than one would expect. Conversely, borrowing results from associative memory in neuroscience, we show that structures with multiple programmed folding motions can be much easier to control than expected, as long as the flat state is avoided.

Michio Tateno (The University of Tokyo, Japan)

Title: Accessing origin of dynamic arrest in a colloidal gel at a single-particle level

Abstract: Gels are ubiquitous in our daily life and have unique properties as a state having characteristics of solid and fluid. The solid nature, or elasticity, comes from a permanent percolated network. For chemical gels, the origin of this permanency is obvious: covalent bonding. For colloidal gels formed by phase demixing, it is often ascribed to dynamic arrest of the network structure either by glass transition or by stabilization of local structures. Here we study the process of dynamic arrest at a single-particle level by confocal microscopy. This is achieved by following the trajectory of each particle from the beginning to the end of phase demixing and accessing structural evolution microscopically. We find that upon phase demixing, the volume of the colloid-rich liquid phase keeps decreasing with time, resulting in the slowing down of dynamics and eventually leading to dynamical arrest. Furthermore, during this slow densification process, the system seeks a low energy particle configuration by forming a network composed of locally favored structures.

Georgios Tsekenis (University of Oregon)

Title: The Effect of Interaction disorder on ordered packings

Abstract: We introduce interaction disorder in a 2D collection of athermal soft spheres and observe how the system changes from a global ground state of the hexagonal crystal to several marginally stable jammed configurations. Even for minuscule interaction disorder the jammed packings achieve isostaticity with characteristic power-law force and gap distributions and density of vibrational states akin to amorphous rigidity while maintaining their near crystalline structure. We discover that the crystalline symmetries (structural and bond-orientational) diminish into the amorphous system with characteristic new power-laws as the amount of frustration is increased.

Mike van der Naald (University of Oregon)

Title: Near Crystalline Sphere Packings at Isostaticity

Abstract: Using simulations of friction-less soft-spheres we measure the distribution of forces, density of states, and distribution of gaps of near-crystalline sphere packings in 3D.  We generate these near-crystalline packings by introducing small amounts of disorder in the form of polydispersity into crystalline packings and minimizing them to isostaticity from above jamming.  Remarkably, we find that these near-crystalline packings have novel power-law exponents that are unlike a perfect crystalline packing or that of an amorphous isostatic packing.  Specifically, we find different scaling for both the small gap distribution and buckler forces.  These power-laws persist down to the smallest polydispersity we can simulate and as we increase the polydispersity we begin to recover power-law scaling characteristic of amorphous packings.

Jun Wang (University of Colorado, Denver)

Title: Abonded-particle model for rock

Abstract: A numerical model for rock is proposed in which the rock is represented by a dense packing of non-uniform-sized circular or spherical particles that are bonded together at their contact points and whose mechanical behavior is simulated by the distinct element method using the two- and three-dimensional discontinuum programs PFC2D and PFC3D. The microproperties consist of stiffness and strength parameters for the particles and the bonds. Damage is represented explicitly as broken bonds, which form and coalesce into macroscopic fractures when load is applied. The model reproduces many features of rock behavior, including elasticity, fracturing, acoustic emission, damage accumulation producing material anisotropy, hysteresis, dilation, post-peak softening and strength increase with confinement. These behaviors are emergent properties of the model that arise from a relatively simple set of microproperties. Amaterial-genesis procedure and microproperties to represent Lac du Bonnet granite are presented. The behavior of this model is described for two- and three-dimensional biaxial, triaxial and Brazilian tests and for two-dimensional tunnel simulations in which breakout notches form in the region of maximum compressive stress. The sensitivity of the results to microproperties, including particle size, is investigated. Particle size is not a free parameter that only controls resolution; instead, it affects the fracture toughness and thereby influences damage processes (such as notch formation) in which damage localizes at macrofracture tips experiencing extensile loading. 

Keywords: Rock fracture; Distinct-element method; Numerical model; Micromechanics Reference: doi:10.1016/j.ijrmms.2004.09.01 

Shang Zhang (University of Michigan, Ann Arbor)

Title: Correlated rigidity percolation and gelation of colloidal particles

Abstract: Rigidity percolation on a lattice with sites or bonds randomly diluted is controlled by the isostatic point, where the degrees of freedom and constraints balance and the system is at the verge of mechanical instability.  In the case of triangular lattice rigidity percolation occurs very close to p=2/3 as predicted from isostaticity.  Interestingly, we found that when the site dilution is correlated, this transition occurs at a lower p, meaning that less “material” is needed for rigidity in the disordered structure.  This correlation may be seen as a consequence of short range attraction between the particles which make them cluster.  We characterized critical scaling associated with the site correlation parameter, and will discuss implication to understand experimental systems such as gelation of colloidal particles.

Yuchen Zhao (Duke University)

Title: Hyper-static Packings Made of Granular Hexapods

Abstract: Columns are made of convex non-cohesive grains like sand collapse after being released from initial positions. On the other hand, various architectures built by concave grains can maintain stability. We explore why these structures are stable, and how stable they can be. We performed experiments by randomly pouring identical star-shape particles into hollow cylinders left on glass and a rough base, and observed stable granular columns after lifting the cylinders. Particles have six 9mm arms, which extend symmetrically in the xyz directions. Both the probability of creating a stable column and mechanical stability aspects have been investigated. In order to understand structure leading to stability, 3D CT scan reconstructions of columns have been done and contact number, contact position and packing density will be discussed.

Di Zhou (University of Michigan, Ann Arbor)

Title: Topological Phonon Edge Modes in Active Mikado Networks

Abstract: Mechanical properties of disordered fiber networks are not only important in understanding a broad range of natural (such as the cytoskeleton and the extracellular matrix) and manmade materials (such as aerogels and porous media), but also exhibit interesting and rich physics.  In this talk, we discuss how topological floppy edge modes can emerge from these fiber networks as a result of active driving.  It is known that straight fibers in a network carries a state of self-stress and bears a bulk floppy mode.  We find that, interestingly, by driving the network with a tiny perturbation, the bulk modes evolve into edge modes.  We introduce a new transfer-matrix formulation that can be applied to this strongly disordered system, to characterize the topological edge modes.  We also discuss possible implications of these edge modes in biological processes.