CSM seminar on Statistical Physics of Materials (Archive)

Former seminars:

Thursday, May 4, 14:15 at Y404 

Speaker: Maryam Hanifpour

 

Onset of crack propagation in an ordinary paper

 

In a venture to understand what happens as a small notch propels in a disordered medium, onset  of propagation of one initial crack in ordinary paper sheets is being investigated by three different tools:

i.  Digital camera: to observe strain field around a crack tip

ii. Infrared camera: to record temperature fluctuations around the crack tip

iii. Force measurement device: to propel the initial crack and record the force fluctuations

In this talk I will present some early results of the experiments.

 


Thursday, April 20, 14:00 at Y404 

Speaker: Tero Mäkinen

 

Simulating the Portevin-Le Chatelier effect in tensile experiments

During deformation some alloys exhibit collective dislocation dynamics called the Portevin-Le Chatelier effect which can be observed as serrations in the stress-strain curve and visible deformation bands moving on the sample. As directly simulating the dislocation dynamics is computationally infeasible due to the large number of dislocations some form of coarse-graining is used. Models based on this idea have shown promising results but for example the fluctuations in the band velocities have not been thoroughly studied.
In this talk I will present my work on some of these 1D coarse-grained models and compare the results to previously gathered experimental data obtained from tensile tests on aluminum alloy samples.


Thursday, March 30, 14:00 at Y312 

Speaker: Amandine Miksic

 

Intermittent crack growth in fatigue

Fatigue occurs under cyclic loading at stresses below a material's static strength limit. Most applications of structural materials involve cyclic loading and fatigue failures appear often catastrophic. We consider fatigue crack growth as a stochastic process and perform crack growth experiments in a metal (copper). We follow optically cracks propagating from initial edge notches. The main interest is in the dynamics of the crack growth - the Paris' law and the initiation phase prior to that - and especially the intermittency this displays. The sampling of the crack advancement, performed at regular intervals, is analyzed by the analogy of planar crack dynamics in slow, driven growth.

 


Tuesday, March 7, 13:00 at Y404

Speaker: Ezequiel Ferrero (Università degli Studi di Milano)

 

Statistics of collective dynamics on driven glassy systems

 

This seminar will be divided in two short presentations of related topics:


1) "Spatio-temporal patterns in ultra-slow creep dynamics of magnetic interfaces”

In presence of impurities or structural disorder, ferromagnetic domain walls advance only when a sufficiently large external field is applied. Close to this depinning threshold, the wall proceeds by abrupt jumps called depinning avalanches; while, at much smaller fields, it can only creep assisted by thermal activation. We have developed a novel numerical technique that captures this ultra-slow creep regime over arbitrarily large time scales for the paradigmatic problem of an elastic interface moving in a random media.

We point out the existence of activated events that involve a collective reorganization of substantial portions of the system, with a cutoff size "S_opt" that diverges at vanishing applied field. At variance with uncorrelated depinning avalanches, creep events display spatio-temporal patterns similar to the sequence of aftershocks observed after a large earthquake, they are highly correlated. We show that events assembly in independent "clusters". Remarkably, this clusters show a scale-free statistics identical to the depinning avalanches when clusters bigger than S_opt are considered and display a power-law distribution with an exponent proper of equilibrium for smaller clusters.
Our results are compatible with the celebrated creep law for the velocity of the interface below threshold, but on the other hand they challenge the commonly assumed scenario of uncorrelated creep events. We expect this spatio-temporal patterns to be experimentally accessible by magneto-optical imaging of ferromagnetic films and, for that, we provide some quantitative predictions.


2) "Avalanche statistics at the yielding transition of amorphous solids: driving rate dependence and inertial effects."
We study stress time series caused by plastic avalanches in athermally sheared disordered materials. Using extensive simulations of a mesoscopic elastoplastic model, we analyze system size and shear-rate dependence of the stress-drop duration and size distributions together with their average temporal shape.
We find critical exponents different from the usual mean-field depinning description, and a clear asymmetry for individual avalanches. The probability distribution for local yielding and its driving rate dependence are also discussed in the marginal stability picture. We probe scaling relations for the rate dependency of the dynamics and we report a crossover towards and effective depinning statistics at strong driving. 
By means of a finite-element extension of the model that allows for the control of damping, we additionally address the effects of inertia in the avalanche statistics. In contrast to avalanches in the overdamped case, dominated by the yielding point exponents, inertial avalanches are controlled by a nonuniversal damping-dependent feedback mechanism, eventually turning negligible the role of correlations.
-----------
References:
1) Spatio-temporal patterns in ultra-slow domain wall creep dynamics 
E.E. Ferrero, L. Foini, T. Giamarchi, A.B. Kolton, and A. Rosso
arXiv:1604.03726
https://arxiv.org/abs/1604.03726
2a) Driving Rate Dependence of Avalanche Statistics and Shapes at the Yielding Transition
C. Liu, E.E. Ferrero, F. Puosi, J.-L. Barrat, and K. Martens
Phys. Rev. Lett. 116, 065501 (2016)
https://doi.org/10.1103/PhysRevLett.116.065501
2b) Inertia and universality of avalanche statistics: The case of slowly deformed amorphous solids
K. Karimi, E.E. Ferrero, J.-L. Barrat
Phys. Rev. E 95, 013003 (2017)

https://doi.org/10.1103/PhysRevE.95.013003


Friday, March 03.03.2017, 11:00 - 12:00

Speaker: Ilari Rissanen

Discrete convolution in Fourier space with GPUs: a fast method for calculating long range interactions in a finite difference framework​

In physics simulations, long range interactions are a source of a significant

computational load that scales strongly (typically as a square) with system size.

In order to simulate larger systems, various techniques have been developed to

deal with these interactions. One of the techniques, applied  in finite difference

calculations for interactions of certain form, is to interpret the interaction as a

discrete convolution of a source term and a convolution kernel. Transforming and

calculating the convolution in Fourier space reduces the scaling from square to

linearithmic scaling of Fast Fourier transforms. The  calculation can be further sped

up by using graphics processing units due to high-performance FFT algorithms and

the easy parallelization of the calculation in Fourier space. In my talk I present this

technique and an example of its use in micromagnetic simulations.


Thursday, February 02.02.2017, 11:00 - 12:00

Speaker: Oleksandr Chepizhko

Physics of collective motion of biological cells

Collective cell motion is an important biological phenomenon that occurs during organism development, wound healing or in cancer metastasis. 
We showed experimentally that such motion can occur in bursts reminiscent of ones recorded in the propagation of cracks, fluid fronts in porous media, and ferromagnetic domain walls.
The computer modeling of this process is an important approach to understand the underlying physics. An agent based model for numerical simulations of collective cell motion will be presented. Fitting of experimental results, such as distributions of the velocity components, the progression of the average front position, etc. will be discussed.


Thursday, January 19, 2017, 11:00-12:00

Speakers: Maryam Hanifpour and Charlotte Petersen

3D Printing of Disordered Auxetic Structures

Auxetic structures, unlike conventional structures, exhibit negative Poisson’s ratios in one or more directions. Counter intuitively,  they become wider in the direction perpendicular to an applied tension. Their unique mechanical properties have been the subject of many recent studies. We have investigated the effects of random topological changes in auxetic structures subjected to tensile and compression loads. The samples have been prepared by means of 3D printing technique, using the FDM method. These results are compared with simple simulations. Also, the mechanism of failure in auxetic bowtie structures made of somewhat brittle material reveals some interesting features. 


Friday, December 16, 10:15-11:00

Speaker: Anniina Salonen

Arresting foams, and arresting in foams

Foams are made by dispersing gas in a fluid. Such mixtures are inherently unstable and separate in time. A major challenge in current research is understanding the mechanisms leading to the destabilisation and finding ways in which foams can be made to last. We know that either the bubble surfaces need to be fortified or the continuous fluid phase made viscoelastic to increase the stability of the foams. The continuous phase can be made viscoelastic by making it gel (solid particles that form a network) or by jamming soft particles (oil droplets).

However, the exact criteria to determine whether the foam will last or disappear are still unclear. I will review our current work on understanding of the evolution of foams made with soft or solid particles (and the evolution of the particles inside the foam channels). 


Thursday, December 1, 11:00-12:00

Speaker: Touko Herranen

Influence of sample thickness on domain wall dynamics in uniaxial magnetic garnets

We study field-driven magnetic domain wall dynamics in garnet strips by  large-scale three-dimensional micromagnetic simulations. By considering a wide range of strip thicknesses from 30 nm to 1.89 µm, we find a non-monotonic  thickness dependence of the threshold field for excitations of the domain wall internal degrees of freedom, i.e., nucleation of Bloch lines, and the concurrent drop in the domain wall propagation velocity. We identify a critical strip thickness above which nucleation of horizontal Bloch lines takes place, while for thinner strips the velocity instability is associated with generation of vertical Bloch lines within the domain wall and uniform precession of the domain wall magnetization.


Thursday, November 10, 14:00-15:00

Speakers: Lasse Laurson and Mikko Alava

Nobels for what?

In this seminar we will briefly discuss the science behind the Nobel Prize in Physics 2016, awarded to David J. Thouless, F. Duncan M. Haldane, and J. Michael Kosterlitz "for theoretical discoveries of topological phase transitions and topological phases of matter".


Thursday, October 27, 11:00 at Y404

Speaker: Alberto Petri

Rheology of slowly sheared grains

Granular matter is ubiquitous in both natural environment and human activities. One main issue is the way it reacts to shear, since in practical situations this can be source of undesired mechanical instabilities. It deserves interest also from a fundamental point of view, as an instance of non-equilibrium system that can be possibly driven close to a critical point.
The seminar focuses mainly on this last situation, characterized by so called stick-slip motion, and accounts for some results obtained  in our lab by means of experimental and numerical investigation.
Specifically, our mechanical measurements and numerical simulations show that, under low shear rate, several physical quantities exhibit the intermittent and self-similar fluctuations expected from criticality.
Detection of the related acoustic emission signals shows consistent statistical features. Moreover, unveiling a direct relation between signal intensity and shear rate, it supplies a new way for probing the system dynamics.
We also briefly discuss some possible theoretical modeling and show, from experimental and numerical data, that some expected scaling are sometimes broken by the presence of non-Markovian friction. We conclude with some perspectives from diffusion wave spectroscopy.


Thursday, October 6, 11:00 at Y404

Speaker: Oleksandr Chepizhko

Bursts of activity in collective cell migration

Dense monolayers of living cells migrate collectively, for example during wound healing and in cancer invasion. They are driven by active forces and invade when free space is available. Here we show that this motion occurs in bursts similar to the ones observed in other driven systems, such as the propagation of cracks, fluid fronts in porous media, and ferromagnetic domain walls [1]. In analogy with these systems, the distribution of activity bursts displays scaling laws that are universal in different cell types and for cells moving on different substrates. This main feature of the dynamics is captured by a model of interacting active particles moving through disordered landscape. Our results demonstrate that living systems display universal nonequilibrium critical fluctuations, that are usually associated with externally driven inanimate media.

1. O. Chepizhko, C. Giampietro, E. Mastrapasqua, M. Nourazar, M. Ascagni, M. Sugni, U. Fascio, L. Leggio, C. Malinverno, G. Scita, S. Santucci, M. J. Alava, S. Zapperi, and C. A. M. La Porta, PNAS 2016 ; published ahead of print September 28, 2016, doi:10.1073/pnas.1600503113


Thursday, September 29, 11:00 at Y404

Speaker: Markus Ovaska

Deformation and fracture of collagen networks

Collagen networks are largely responsible for providing strength to animal tissues, and they are relevant for many biomedical applications. Here we characterize the mechanical properties of stiff collagen networks derived from three different marine animals. Tensile tests show that the networks exhibit nonlinear stiffening followed by brittle fracture, with large sample-to-sample fluctuations in the Young modulus and fracture strength. We also construct a three-dimensional numerical model for the deformation of a cross-linked network of elastic fibers and study how e.g. properties of the cross-links affect the mechanical response.

 


Thursday, September 22, 14:00 at Y404

Speaker: Charlotte Petersen

Magnetic charge screening in artificial spin ice

Electric charges are screened in a variety of natural systems through reconfiguration of the local environment, effectively decreasing the Coulomb interaction between charges. This phenomena has not been directly observed for magnetic charges. Artificial spin ice is a recently developed metamaterial, whose microstate can be directly imaged. It consists of two dimensional nanomagnets placed in periodic arrays. There has been particular interest in observing emergent magnetic charges in these systems, and they are an ideal testbed in which to explore the possibility of screened charges. We introduce a new lattice geometry we call the Dice lattice, which contains vertices of mixed coordination numbers, and allows for the direct observation of magnetic charge screening. With recent advances in experimental methods, it is now possible to take real time measurements of the dynamic configurations of thermal artificial spin ice. This allows us to measure the thermal stability of screened magnetic charges. We also model the dynamics of this new lattice using kinetic Monte Carlo.

 

 


Thursday, September 8, 13:00 at Y404

Speaker: Juha Koivisto

Flow and Clogging of Submerged Hoppers

In accord with the Beverloo equation the flow rate of granular material from a hopper is constant, irrespective of the filling height. For pure fluids, by contrast, the flow rate is set by the hydrostatic pressure and decreases as the filling height decreases. In this talk two situations are compared: a case of dry non-cohesive grains and a case where the entire hopper is submerged under water. The experimental setup consists of a cylindrical flat bottomed hopper with a centered orifice at the bottom. The diameter of the hopper varies from 3 to 200 mm while the orifice has values from 1 to 10 mm. The grains are spherical polydisperse glass beads with 0.9-1.1 mm diameter and granular density of 1.6 g/ml. The density of glass is 2.5 g/ml. The most noticeable difference between the two cases is that in the submerged case, the flow rate clearly depends on the filling height. However, the behavior is rather counterintuitive: The flow rate increases as the filling height decreases. Similar behavior is now discovered in the dry case, but the rate increase is much smaller and is merely just before the hopper runs out of grains. Previous studies have only reported the final decrease of flow rate, and disregarded it as a boundary effect. For small orifice sizes the hopper clogs.It is found that the qualitative behavior is the same in both submerged and dry cases while only quantitative behavior changes. This indicates that the interstitial medium does not affect the geometry of the clogging process, just the dynamics.

 


Thursday, September 1, 11:00 at Y404

Speaker: Pritam Kumar Jana

Effect of additives and external fields on lubricating properties of nematic liquid crystals

Several studies have been performed to understand the structural and dynamical properties of liquid crystal lubricants under external pressure and shear stress, mainly due to the observations of ultra-low friction in such systems [1]. However, to balance between efficiency and effective cost, more extensive investigations are required. In the present study, we focus on the influences of short alkane chains as additives on the lubricating properties of liquid crystals by performing molecular dynamics simulations. To be specific, we construct a full atomistic model where a nematic liquid crystal, 4-cyano-4-hexylbiphenyl (6CB), and a short alkane chain, hexane (C6H14), are used as lubricant and additives, respectively, and mica serves as the confining surfaces. When the sliding velocity of the upper mica plate is low enough, thin films of both pure 6CB and pure C6H14 show stick-slip dynamics. However, for the liquid crystal lubricant, stick-slip appears at lower driving forces as compared to the hexanelubricant, which demonstrates liquid crystal as a better lubricant in comparison to hexane in nano-scale friction. We also consider mixtures of 6CB and C6H14, by varying their proportions, as well as the effect of external electric fields, to optimize the resulting lubrication properties.

[1] C. Manzato, A. S. Foster, M. J. Alava, and L. Laurson. Physical Review E 91, 012504 (2015)

 


Thursday, May 12, 11:00 at Y404

Speaker: Marcelo Dias

Thin elastic structures: a roadmap from forms to functions

In recent years, there has been a growing research trend in the mechanics of highly deformable and soft complex structures. A theoretical mechanics treatment of such systems, so as to include fundamental geometric non-linearities from large deflections and/or micro-mechanical instabilities, is timely in order to unleash their full potential for functionalization. In this talk, we will discuss some recent ideas in thin elastic structures by focusing on controlled-engineering designs. First, we will begin with the study of how micro-architectures in elastic sheets affect their macroscopic effective elastic responses. Secondly, we will discuss a deeper understanding of adaptability and stiffness modulation of thin shells and combine it with questions concerning biomechanics function and the evolution of human bipedal locomotion. All these questions are challenging matters of mechanical optimization appearing in a variety of length-scales for both man-made and natural systems.

 


Thursday, March 31, 11:00 at Y404
Speaker: Maryam Hanifpour

Characterization of dense granular material

Hard sphere packings have always been the subject of interest for physicists and mathematicians. Apart from the strong capabilities of hard sphere model in describing different physical systems, the model itself presents its own challenges. The history of experimental studies on monosized hard spheres dates back to 1950s when J. D. Bernal used 8000 glass balls gently shaken in a box, fixing their positions with pouring wax and recording the positions and number of contacts for each ball by tediously detaching all the balls one by one. Bernal’s experiment reported a packing fraction (the fraction of volume occupied by the balls) of 64 % and an average number of 6 contacts for each ball, both values being far from the optimum ordered structure of FCC or HCP with 12 number of contacts per each sphere and a packing fraction of 74%. Since Bernal, numerous experiments and simulations have confirmed that a stable configuration of frictionless monodisperse spheres can not exist with a packing fraction below Bernal limit known as the the Random Close Packing (RCP) limit. However by introducing friction it is possible to generate randomly packed mono-sized hard spheres over a range of densities between 0.55 (know as random loose pacing) to 0.64 and still retain their resistance to shear. Until recently the limit beyond RCP has been accessible only through numerical simulations and a handful of 2D experiments. These studies show that by passing beyond the RCP limit, locally ordered structures seed and develop in the system and by further increasing the packing density, these local orders expand and eventually dominate the whole system.

In this talk I try to give a brief description of the research I have done during my PhD. I have studied partially crystallized packings of monozised hard sphere produced experimentally by applying effective vibrational protocols. Then by means of X-Ray Computed Tomography and image analysis technique, the position and radii of each sphere from the experimental configurations were determined. I then carried out some molecular dynamics simulations on the experimental data to further understand the experimental results. The post processing analysis revealed very interesting features of crystallization both in the mechanical and geometrical characteristics of the system.

 


Wednesday, March 16, 14:00 at Y404
Speaker: Tero Mäkinen

Portevin-Le Chatelier effect in an aluminum alloy

The deformation of some alloys under certain strain rate and temperature conditions exhibits serrated flow and strain localization into deformation bands, which is known as the Portevin-Le Chatelier (PLC) effect. Until fairly recently the study of the kinetics of the PLC bands using direct measurements (for example digital image correlation) has been limited to a fairly low acquisition rate.
In this talk I will discuss my master's thesis project where I have used a laser speckle technique with a high speed camera to study the PLC bands with a significantly higher framerate. This helps to shed some light on the phenomena on smaller timescales.

 


Thursday, March 3, 11:00 at Y404

Speaker: Kseniia Khakalo

Rheology of foams: from bubble dynamics to the continuum

Foam is a typical example of soft jammed material, i.e. material, that behaves like a liquid until volume fraction of gas reaches some critical value (jamming transition) and after that it acts like an amorphous solid. However, the gaseous nature of foams distinguishes it from other jammed materials. Gas can diffuse through bubble walls and this leads to coarsening. If a foam is watched over time, the small bubbles shrink and disappear, and the larger bubbles grow. This leads to a polydisperse distribution of bubble sizes, which turns out to have a universal form [1]. The average bubble radius changes as tα, so the coarsening exhibits scaling behaviour.
In our research we are modelling foams using particle-based dynamics approach (Durian bubble model [2]). At this stage we are modelling coarsening for systems with different volume fractions and our results show good agreement with the experiments. The next step would be to investigate the influence of coarsening on the foam rheology.

[1] D. J. Durian, D. A. Weitz, D. J. Pine, Science 252, 686 (1991)
[2] D. J. Durian Phys. Rev. Lett. 75, 4780 (1995)


Thursday, February 18, 11:00 at Y404

Speaker: Pritam Kumar Jana

Plastic events of polycrystals under cyclic deformation

Recently, Tamborini et al. have investigated plasticity of colloidal polycrystals, which are prepared by crystallizing thermo sensitive block copolymer Pluronics F108 with a small amount of nanoparticles as impurities, under cyclic shear [1]. Guided by their experiment, we perform plastic deformation study on 2D polycrystalline samples using molecular dynamics simulations. Samples are prepared by reducing the temperature of binary Lennard-Jones liquids with a fixed cooling rate. A small amount of bigger particles, which basically acts as an impurity, creates dislocations in the system because of size mismatch. A string of dislocations are defined as grain boundaries and the amount of impurities can control the number of grains. In the next stage, samples are undergone a cyclic deformation where the maximum strain amplitude is varied by changing the strain rate.  For a small value of maximum strain  amplitude, the system becomes quiescent within a few cycles whereas for larger maximum strain amplitude, particles move irreversibly which leads to the grain boundary motion and the annihilation of dislocations. 

 

[1] Plasticity of a colloidal polycystal under cyclic shear. E. Tamborini, L. Cipelletti, and L. Ramos. PRL 113, 078301 (2014)


Thursday, February 4, 14:00 at Y404

Speaker: Arttu Lehtinen

Crystals as complex systems: Avalanche dynamics in 3D dislocation configuration

Crystal plasticity occurs by deformation bursts due to the avalanche-like motion of line defects called dislocations. In order to study the statistical properties of these avalanches, we have performed extensive numerical simulations of Al with three-dimensional dislocation dynamics code ParaDis [1]. We have modified the default version of the code  by adding a quasistatic stress loading feature where the stress-rate is controlled by the average velocity of the dislocations. Our results show that dislocation avalanches are power-law distributed and display peculiar stress and sample size dependence: The average avalanche size grows exponentially with the applied stress, and the amount of slip increases logarithmically with the system size.
These results suggest that intermittent deformation processes in materials with  FCC crystal structure exhibit an extended critical-like phase in analogy to glassy systems, instead of originating from a non-equilibrium phase transition critical point. Real metals contain also point like defects like precipitates, solute atoms and voids. These objects serve as pinning obstacles for the dislocation motion. The relevant question is then that does these obstacles and the frozen disorder that they generate,  change the avalanche dynamics of the deformation bursts in 3D systems as it does in simple 2D simulations? We have implemented a new obstacle feature into the ParaDis which can  help in shedding light to this problem [2].
In this talk i will introduce the basic properties of the dislocations and our simulation results concerning the avalanche dynamics of pure dislocation system. I then describe our new obstacle implementation and some future research goals .  

[1]A. Arsenlis, W. Cai, M. Tang, M. Rhee,T.Oppelstrup, G.Hommes, T.G. Pierce, and V.Bulatov. Enabling strain hardening simulations with dislocation dynamics. Modelling and Simulation in Materials Science and Engineering, 15(6):553, 2007
[2]A.Lehtinen, F.Granberg, L.Laurson, K. Nordlund, and M. J. Alava.
"Multiscale modeling of dislocation-precipitate interactions in Fe: From molecular dynamics to discrete dislocations"  Phys. Rev. E 93, 013309

 


 

Thursday, January 21, 11:00 at Y404

Speaker: Charlotte Petersen

Artificial Spin Ice

Frustrated materials exhibit an inability to simultaneously minimize all of their interactions, resulting in a disordered equilibrium state. Artificial spin ice is an analog to other forms of frustrated condensed matter, such as Pyrochlore spin ice or water ice, but it is more experimentally accessible. It consists of lattice arrays of nanomagnets, whose magnetic spins can become geometrically frustrated. The direction of the spin of each nanomagnet can be measured directly experimentally using synchrontron-based photoemission electron microscopy. In this talk I discuss two new lattice arrangements, and how we modeled the experiments with Monte Carlo simulations. In addition to observing geometric frustration in these systems, we also see defects analogous to the lattice polaron. A polaron is an electron moving in a crystal lattice, causing displacements to the surrounding ions. In our system, we observe magnetic charge defects influencing their immediate environment, resulting in a screening effect, and what we will call a magnetic polaron. These polarons are observed experimentally and in simulation as the system relaxes from an excited state towards equilibrium. 

 


 

Thursday, September 24, 11:00 at Y404

Speaker: Filippo Federici Canova

Machine-learning approach to lubricant optimization

Friction and wear are the main source of failure for any kind of machinery. As technology scales down towards micro- and nano-mechanical systems, the detrimental effects of friction can seriously hamper the functionality of these devices. Searching for appropriate lubricant is often a daunting task, due to the wide range of possible chemical compounds, and the cost of evaluating each candidate experimentally.
Being able to predict the viscous properties of a lubricant based on its composition could make the screening process much faster and efficient. However, the relationship between the atomic details of a fluid, and its physical properties, which intuitively exist, are not known.
Machine-learning has been widely used in computer science to mimic human intuition and categorisation skills, and recently found its way into chemical and physical problems [1].
In principle then, the correlation between the frictional properties and the molecular structure can be machine-learned from experimental or simulation data, however, it remains challenging to formulate a descriptor for the lubricants, that is flexible enough to span the chemical space and computer-readable.
We developed a graph based neural network [2] that satisfies both requirements. This neural network abstracts the system hierarchically, from atoms to molecules, and to the whole fluid mixture, in a deep learning fashion, where each stacked layer is trained to recognise features in the structure of the fluid, before calculating its physical properties.
Training such large and complex model is possible, provided a large enough database: none could be found in literature. We then focussed on simple toy models for liquids and attempt to construct such database from molecular dynamic simulations.
[1] K. Hansen, et al., J. Chem. Theor. Comput., 9, 3404, 2013
[2] F. Scarselli, et al., Neural Networks, 20, 61, 2008


Thursday, September 10, 11:00 at Y404

Speaker: Stephane Santucci

Multiscale Stick-Slip Dynamics of Adhesive Tape Peeling

Using a high-speed camera, we follow the propagation of the detachment front during the peeling of an adhesive tape from a flat surface. In a given range of peeling velocity, this front displays a multi-scale unstable dynamics, entangling two well-separated spatiotemporal scales, which correspond to microscopic and macroscopic dynamical stick-slip instabilities. While the periodic release of the stretch energy of the whole peeled ribbon drives the classical macro-stick-slip, we show that the micro-stick-slip – due to the regular propagation of transverse dynamic fractures – is related to a high-frequency periodic release of the elastic bending energy of the adhesive ribbon concentrated in the vicinity of the peeling front.

Ref: M.-J. Dalbe, P.-P. Cortet, M. Ciccotti, L. Vanel, S. Santucci, to appear in PRL (2015)


Thursday, August 13, 14:00 at Y404

Speaker: Charlotte Petersen

The dissipation function in non-equilibrium statistical mechanics

Fluctuation relations are one of the new exact results in non-equilibrium statistical mechanics. They extend our understanding of small systems, and systems far from equilibrium through making quantitative predictions about fluctuations that are monitored over short periods. This talk will discuss two consequences of the Evans-Searles fluctuation theorem, given in terms of the dissipation function, a key quantity in non-equilibrium statistical mechanics.

I’ll introduce a recently derived fluctuation theorem that gives the relative probability that the instantaneous values of a phase function take on opposite values, ±A in terms of the path integral of the dissipation function. The form of this new relation is not obvious, and involves observing the system along its transient phase space trajectory both before and after the point in time at which the fluctuations are being compared.

The non-equilibrium partition identity is a direct result of the Evans-Searles fluctuation theory. While its derivation is straightforward, calculation of this quantity can be very difficult and it is often observed to converge to a value lower than expected. I will discuss the mechanism for this asymmetric bias.


Thursday, August 6, 14:00 at Y404

Speaker: Zoe Budrikis

Avalanches in plasticity of amorphous material

Amorphous materials are interesting from the point of view of statistical mechanics of complex systems, because when subject to loading they exhibit a rich variety of collective phenomena, such as strain localization, intermittent dynamics and power-law distributed avalanches. The statistics of avalanches can be used to determine the universality class of the transition to flow in the material, but the nature of the universality class has been controversial, with disagreement among workers on whether the dynamics can be captured by mean field models. In this seminar, I will discuss avalanche statistics and strain patterning in amorphous materials in 2 and 3 dimensions and present evidence the universality class is not mean field.


Friday, June 12, 14:00 at Y404

Speaker: Daniel Rayneau-Kirkhope

Novel mechanical properties through elastic instability

Mechanical metamaterials behave in a manner defined by their structure/geometry rather than their material composition. Through the manipulation of structural parameters, one can create a material that will behave in a predetermined, novel manner. In this talk, I will focus on the creation of a 2-dimensional auxetic metamaterial (a material with a negative Poisson ratio). In this particular example, the auxetic nature of the material is activated through elastic instability, thus we have a system that for small loading (before elastic instability) the material behaves with a positive Poisson ratio, and for larger, compressive loading the material will exhibit auxetic bahaviour. We describe this system analytically through Euler-Bernoulli beam theory. We derive the parameter space for which the auxetic material properties exhibited, and the critical loading at which the auxetic behaviour is “switched on”. The work presented is part of an ongoing collaboration with Marcelo Dias (Aalto Science Institute).


Thursday, June 4  at Y404

Speaker: Touko Herranen

Extended domain walls with non-uniform internal degrees of freedom

We study field-driven dynamics of magnetic domain walls in ferromagnetic strips with a high perpendicular magnetic anisotropy, using CoPtCr as an example system. For sufficiently wide strips and above the Walker breakdown, our micromagnetic simulations reveal that the internal degrees of freedom of the domain wall are excited in a spatially non-uniform fashion, in contrast to typical nanostrip geometries where the internal magnetization of the domain wall precesses uniformly. We show that such inhomogeneities, nucleating at the strip edges or due to disorder within the strip, can be well-described as a set of internal in-plane domains along the domain wall, separated by internal domain walls. We study the field-driven dynamics of the internal domain walls along the main domain wall, and show that due to topological constraints, they do not experience a Walker breakdown.


Thursday, May 21  at Y404

Speaker: Ilari Rissanen

Contact aging and viscoelasticity in a minimalistic friction model

The stick-slip motion of a driven elastic chain on a disordered substrate is studied. The primary goal is to see  how said motion changes when viscoelasticity and contact aging are incorporated in the system.  A master’s thesis presentation.


Thursday, May 7, 14:00 at Y404

Speaker: Henri Salmenjoki

Modeling oscillatory rheology of viscoelastic and thixotropic complex fluids

Complex fluids are present in several processes from blood circulation in human bodies to numerous areas of industry. Understanding rheology of these fluids gives new opportunities in many fields such as medicine and quality control of various products including foods, paints and cements [1]. In this presentation, I will discuss recent modeling results of viscoelastic and thixotropic complex fluids in a Couette cell rheometer under oscillating shear. The obtained stress response of the fluid is more elastic when the structure is intact during oscillations whereas disintegration of structure leads to more viscous response. Also, the gap size of the rheometer is varied to study effects of the used geometry. Comparing behaviour of dynamic moduli in various gaps suggests that using notably larger measuring systems provides distorted rheological results while typical rheometer dimensions are capable to produce reliable results.

[1] Malkin, A. Y. and Isayev, A. I. Rheology – Concepts, Methods, and Applications. (2nd Edition). ChemTec Publishing, 2012.


Friday, April 24, 11:15 at Y404

Speaker: Dr. Virginia Estévez

Domain wall structures in wide permalloy strips

We analyze the equilibrium micromagnetic domain wall structures encountered in Permalloy strips of a wide range of thicknesses and widths, with strip widths up to several micrometers. By performing an extensive set of micromagnetic simulations, we show that the equilibrium phase diagram of the domain wall structures exhibits in addition to the previously found structures (symmetric and asymmetric transverse walls, vortex wall) also double vortex and triple vortex domain walls for large enough strip widths and thicknesses. Also several metastable domain wall structures are found for wide and/or thick strips. We discuss the details of the relaxation process from random magnetization initial states towards the stable domain wall structure and show that our results are robust with respect to changes of, e.g., the magnitude of the Gilbert damping constant and details of the initial conditions.  We also consider the field-driven dynamics of the domain wall structures.

[1] V. Estévez, L. Laurson, Phys. Rev. B. 91, 054407 (2015)


Thursday, April 9, 14:00 at Y404

Speaker: Dr. Alessandro Luigi Sellerio

Molecular Dynamics: a quick review of theory, applications and examples

Molecular Dynamics (or MD) is a computer simulations technique applied to N-body systems. During the simulation, the bodies are allowed to interact, under prescribed conditions, for a period of time, so that general properties of the system can be inferred. Molecular dynamics has proved a valuable tool to study a large variety of systems, ranging from material science to chemistry, biology and engineering, which are otherwise impossible to study analytically. In the seminar we will quickly review the basic principles of MD: thermodynamics, time integration schemes, energy minimization criteria, etc. We will focus on a few specific examples, simulated with the MD software toolbox “Lammps”. The final part of the seminar will provide a more detailed introduction to Lammps: installation, testing and deploying simple simulations, with the aim to provide a beginner’s quick-start guide.


Thursday, March 19, 14:00 at Y404

Speaker: Dr. Pritam Kumar Jana

Structure formation of anisotropic molecules on surfaces in a non-equilibrium setting

Understanding of STM images of different adsorbed molecules and the mechanism of structure formation is an important aspect in the field of surface science due to its technological relevance. Here, we are interested in the structure formation of lipophilic molecules, containing a nucleobase as the head group. Recently, the group of Prof. Chi (Muenster) has observed that N9-substituted adenine derivative in solution form two different type of structures (intercalation vs. stripes patterns) on the surface[1]. Extending that study to deposition experiments, information about the impact of deposition rate and substrate temperature on structure formation has been obtained [2]. It has been observed that higher deposition rates and lower substrate temperatures prefer to stabilize the intercalated structure. We present Monte Carlo simulations where the molecules are represented as model chains with one head-type and a few tail-type monomers. The key interaction properties, i.e., the formation of hydrogen bonds (intercalated structure) and pi-pi interactions (stripe interaction) among the adenine molecules, respectively, and the van der Waals chain-chain interactions, are reflected in this model, based on DFT-based parameters. The quality of structure formation in dependence on deposition rate and substrate temperature as well as the relative fraction of both phases is analyzed and compared with experimental data. Good qualitative agreement is achieved. The final structure emerges from a subtle interplay of the initial growth of the striped pattern and the subsequent thermodynamically preferred formation of the intercalated structure. More generally, this can be interpreted as a kinetically trapped non-equlibrium process.

[1] Mu et al. Langmuir, 29, 10737, 2013
[2] Wang et al. Chem. Comm., 50, 9192, 2014

Page content by: communications-phys [at] aalto [dot] fi (Department of Physics) | Last updated: 18.05.2017.