Inherent Structures De<strong>for</strong>mations Detect Long-Range Correlations inSupercooled LiquidsM.Mosayebi 1 , E.Del. Gado 1 , P. Ilg 1 , H.C.Ottinger 11 <strong>ETH</strong> Zürich, Zürich, SwitzerlandINTRODUCTION: One of the most puzzlingfeatures of the supercooled regime and the glasstransition is the apparent lack of structural changesunderlying the dramatic slowing down of thedynamics. Recently, some progress has beenachieved in relating dynamical properties toinherent structures, which are the local minima ofthe collective potential energy [1].We have found striking similarities in the onset ofcooperative behaviour in dynamics and in the nonaffinepart of the inherent structure response toexternal de<strong>for</strong>mations [2]. That approach wasmotivated by a recent theory [3] based on a generalframework of non-equilibrium thermodynamics.This thermodynamic treatment suggests that thereversible part of glassy dynamics changesconsiderably when approaching the glass transition.Above the glass transition, the particles can followan imposed de<strong>for</strong>mation more or less freely,whereas closer to the glass transition the particlemovement becomes a hopping-like transitionbetween different basins of attraction of theunderlying inherent structures [1].M<strong>ETH</strong>ODS: We study numerically wellestablishedmodel glass <strong>for</strong>mers, in particularbinary Lennard-Jones and binary soft spheresystems consisting of N particles (N varies between2000 and 64000) from the high temperature liquiddown to the supercooled regime. From thermal,equilibrated particle configurations, we preparecorresponding inherent structures by minimizingthe potential energy with conjugate gradientmethods. We also apply instantaneous, affine shearde<strong>for</strong>mations to each thermal configuration and findthe closest inherent structure corresponding to thede<strong>for</strong>med state. The difference between bothinherent structures – subtracting the affine part - isthe nonaffine displacement field.RESULTS: The distance between two inherentstructures - that are related via shear de<strong>for</strong>mation ofrather small amplitude - sharply decreases withdecreasing temperature below the onset temperatureof the landscape-dominated regime. Furthermore,we observe a crossover between two regimes that isalso present in a qualitative change of the nonaffinedisplacement distribution from an exponential to apower-law shape. The exponent of the latter can beexplained by elasticity arguments. Qualitatively, wefind similar correlations of the nonaffinedisplacement field as experiments on correlatedmotions in colloids. Quantitatively, the staticcorrelation length extracted from nonaffinedisplacements of inherent structures is growingstronger than the dynamical one [4].Fig. 1: Nonaffine displacements <strong>for</strong> one sample ofthe binary Lennard-Jones system at high (left) andlow (right panel) temperature. Homogeneous shearde<strong>for</strong>mation of amplitude 10 -4 was applied. Forclarity, only particles near one face of the threedimensionalsimulation cell are shown.DISCUSSION & CONCLUSIONS: <strong>Here</strong>, wehave shown by extensive molecular simulations thatcorrelations in neighbouring inherent structures arequite reminiscent of cooperatively rearrangingregions that are observed in the system's dynamics[4]. Due to the absence of thermal fluctuations, ourmethod provides a very sensitive and efficient toolto investigate such correlated regions in great detail.REFERENCES:1 A. Heuer; 2008; J. Phys.:Condens. Matter; 20:373101. 2 E. Del Gado, P. Ilg,M. Kröger, H.C. Öttinger; 2008; Phys. Rev. Lett;101:095501. 3 H.C. Öttinger; 2006; Phys. Rev. E74:095501. 4 M. Mosayebi, E. Del Gado, P. Ilg,H.C. Öttinger; 2010; Phys. Rev. Lett; 104:205704.ACKNOWLEDGEMENTS: We gratefullyacknowledge stimulating discussions with LudovicBerthier, Andrea Cavagna, Walter Kob, SrikanthSastry, Anne Tanguy and Asaph Widmer-Cooper.38
Controlling building materials microstructure through the mixing stepHélène Lombois-BurgerLafarge – LCR (France)Mixing is a key step <strong>for</strong> the production of cement pastes or concretes. Indeed, eventhough very short this stage has a major influence on the quality of the microstructureof the mix and thus on the materials per<strong>for</strong>mances. Its impact extends on the wholematerial’s service life, from the rheology to the mechanical strength and even thedurability. These microstructural characteristics due to the mixing are the result of acomplex interplay between process and mix design parameters.Using several techniques to monitor the mixing extent in terms of homogenization,de-agglomeration of the fine particles, prevailing inter-particle <strong>for</strong>ces, allows:(i) To unveil mixing mechanisms, and demonstrate the role of both processparameters (sequence <strong>for</strong> the introduction of the components, mixingtime/speed/energy, volume effects, etc) and mix design ones such as coarse aggregateproportion <strong>for</strong> instance. Mixing optimization consists then in obtaining the bestmicrostructure while shortening the process duration and required energy.(ii) To monitor mix design levers such as packing effects, water content, competitionbetween admixtures, etc39