(53i) An Information-Driven Approach to Quantifying and Controlling Emergent Order

Controlling the emergence of order out of disorder is an overarching goal in the engineering of biological and soft materials. Examples in which such control has been targeted include the design of self-assembling functional materials such as metal organic frameworks [1], as well as the identification of strategies to halt aggregation of disease-implicated unstructured proteins [2]. Computational approaches traditionally rely on simulations at varying time and length scales, aided by enhanced sampling techniques [3], including the recent proliferation of machine learning based methods [4-6]. An exciting direction in the computational design of soft materials incorporates ideas from information theory into these approaches, drawing upon the connection between thermodynamic entropy and information entropy. In this talk, I consider a simple case with an information-driven analysis, which may inform future studies of real systems with greater complexity.

Here, I will discuss the emergence of order in a continuous dynamical absorbing state model known as Biased Random Organization (BRO), in which overlapping particles are considered active and are displaced away from one another by a random distance. This results in a phase transition between absorbing states containing no active particles and active states where particles continue to be displaced. I will focus on the behavior of BRO in 2D, which results in two distinct active phases unlike in other dimensions [7]. One such active phase is crystalline, which emerges for small displacement magnitudes. The second active phase is disordered, which occurs for large displacement magnitudes. Although this phase is disordered, I will show that unlike in the crystalline phase, it exhibits hyperuniformity at criticality, a form of “hidden order” characterized by vanishing density fluctuations that holds potential for unique optical properties. As part of characterizing the three phase transitions found in 2D BRO, I apply a computable information density approach, in which a data compression algorithm is used to quantify changes in the system entropy. I will conclude with my plan for extending this approach to the computational study and design of self-assembling materials.

[1] Y.J. Colón, A.Z. Guo, L.W. Antony, K.Q. Hoffmann, J.J. de Pablo, "Free energy of metal-organic framework self-assembly," (2019). J Chem Phys.

[2] A.Z. Guo, A.M. Fluitt, J.J. de Pablo "Early-stage human islet amyloid polypeptide aggregation: Mechanisms behind dimer formation," (2018). J Chem Phys.

[3] H. Sidky, Y.J. Colón, J. Helfferich, B.J. Sikora, C. Bezik, W. Chu, F. Giberti, A.Z. Guo, X. Jiang, J. Lequieu, J. Li, J. Moller, M.J. Quevillon, M. Rahimi, H. Ramezani-Dakhel, V.S. Rathee, D.R. Reid, E. Sevgen, V. Thapar, M.A. Webb, J.K. Whitmer, J.J. de Pablo, "SSAGES: software suite for advanced general ensemble simulations," (2018). J Chem Phys.

[4] A.Z. Guo, E. Sevgen, H. Sidky, J.K. Whitmer, J.A. Hubbell, J.J. de Pablo, "Adaptive enhanced sampling by force-biasing using neural networks," (2018). J Chem Phys.

[5] E. Sevgen, A.Z.Guo, H. Sidky, J.K. Whitmer, J.J. de Pablo, "Combined force-frequency sampling for simulation of systems having rugged free energy landscapes," (2020). J Chem Theory Comput.

[6] A.Z. Guo, J. Lequieu, J.J. de Pablo, "Extracting collective motions underlying nucleosome dynamics via nonlinear manifold learning," (2019). J Chem Phys.

[7] A.Z Guo, S. Wilken, D. Levine, P.M. Chaikin, “2D Melting in Biased Random Organization,” In Preparation.