Showing 1–50 of 75 results for author: Whitelam, S

Oscillatrons: neural units with time-dependent multifunctionality

Authors: Stephen Whitelam

Abstract: Several branches of computing use a system's physical dynamics to do computation. We show that the dynamics of an underdamped harmonic oscillator can perform multifunctional computation, solving distinct problems at distinct times within a dynamical trajectory. Oscillator computing usually focuses on the oscillator's phase as the information-carrying component. Here we focus on the time-resolved a… ▽ More Several branches of computing use a system's physical dynamics to do computation. We show that the dynamics of an underdamped harmonic oscillator can perform multifunctional computation, solving distinct problems at distinct times within a dynamical trajectory. Oscillator computing usually focuses on the oscillator's phase as the information-carrying component. Here we focus on the time-resolved amplitude of an oscillator whose inputs influence its frequency, which has a natural parallel as the activity of a time-dependent neural unit. We call this unit an oscillatron. The activity of an oscillatron at fixed time is a nonmonotonic function of the input, and so it can solve nonlinearly-separable problems such as XOR. The activity of the oscillatron at fixed input is a nonmonotonic function of time, and so it is multifunctional in a temporal sense, able to carry out distinct nonlinear computations at distinct times within the same dynamical trajectory. Time-resolved computing of this nature can be done in or out of equilibrium, with the natural time evolution of the system giving us multiple computations for the price of one. △ Less

Submitted 9 May, 2024; v1 submitted 23 April, 2024; originally announced April 2024.

Learning protocols for the fast and efficient control of active matter

Authors: Corneel Casert, Stephen Whitelam

Abstract: We show that it is possible to learn protocols that effect fast and efficient state-to-state transformations in simulation models of active particles. By encoding the protocol in the form of a neural network we use evolutionary methods to identify protocols that take active particles from one steady state to another, as quickly as possible or with as little energy expended as possible. Our results… ▽ More We show that it is possible to learn protocols that effect fast and efficient state-to-state transformations in simulation models of active particles. By encoding the protocol in the form of a neural network we use evolutionary methods to identify protocols that take active particles from one steady state to another, as quickly as possible or with as little energy expended as possible. Our results show that protocols identified by a flexible neural-network ansatz, which allows the optimization of multiple control parameters and the emergence of sharp features, are more efficient than protocols derived recently by constrained analytical methods. Our learning scheme is straightforward to use in experiment, suggesting a way of designing protocols for the efficient manipulation of active matter in the laboratory. △ Less

Submitted 28 March, 2024; v1 submitted 28 February, 2024; originally announced February 2024.

Nonequilibrium formulation of varying-temperature bit erasure

Authors: Stephen Whitelam

Abstract: Landauer's principle states that erasing a bit of information at fixed temperature T costs at least kT ln 2 units of work. Here we investigate erasure at varying temperature, to which Landauer's result does not apply. We formulate bit erasure as a stochastic nonequilibrium process involving a compression of configuration space, with physical and logical states associated in a symmetric way. Erasur… ▽ More Landauer's principle states that erasing a bit of information at fixed temperature T costs at least kT ln 2 units of work. Here we investigate erasure at varying temperature, to which Landauer's result does not apply. We formulate bit erasure as a stochastic nonequilibrium process involving a compression of configuration space, with physical and logical states associated in a symmetric way. Erasure starts and ends at temperature T, but temperature can otherwise vary with time in an arbitrary way. Defined in this way, erasure is governed by a set of nonequilibrium fluctuation relations that show that varying-temperature erasure can done with less work than k T ln 2. As a result, erasure and the complementary process of bit randomization can be combined to form a work-producing engine cycle. △ Less

Submitted 22 December, 2023; originally announced December 2023.

Free-energy estimates from nonequilibrium trajectories under varying-temperature protocols

Authors: Stephen Whitelam

Abstract: The Jarzynski equality allows the calculation of free-energy differences using values of work measured in nonequilibrium trajectories. The number of trajectories required to accurately estimate free-energy differences in this way grows sharply with the size of work fluctuations, motivating the search for protocols that perform desired transformations with minimum work. However, protocols of this n… ▽ More The Jarzynski equality allows the calculation of free-energy differences using values of work measured in nonequilibrium trajectories. The number of trajectories required to accurately estimate free-energy differences in this way grows sharply with the size of work fluctuations, motivating the search for protocols that perform desired transformations with minimum work. However, protocols of this nature can involve varying temperature, to which the Jarzynski equality does not apply. We derive a variant of the Jarzynski equality that applies to varying-temperature protocols, and show that it can have better convergence properties than the standard version of the equality. We derive this modified equality, and the associated fluctuation relation, within the framework of Markovian stochastic dynamics, complementing related derivations done within the framework of Hamiltonian dynamics. △ Less

Submitted 18 November, 2023; v1 submitted 12 November, 2023; originally announced November 2023.

How to train your demon to do fast information erasure without heat production

Authors: Stephen Whitelam

Abstract: Time-dependent protocols that perform irreversible logical operations, such as memory erasure, cost work and produce heat, placing bounds on the efficiency of computers. Here we use a prototypical computer model of a physical memory to show that it is possible to learn feedback-control protocols to do fast memory erasure without input of work or production of heat. These protocols, which are enact… ▽ More Time-dependent protocols that perform irreversible logical operations, such as memory erasure, cost work and produce heat, placing bounds on the efficiency of computers. Here we use a prototypical computer model of a physical memory to show that it is possible to learn feedback-control protocols to do fast memory erasure without input of work or production of heat. These protocols, which are enacted by a neural-network ``demon'', do not violate the second law of thermodynamics because the demon generates more heat than the memory absorbs. The result is a form of nonlocal heat exchange in which one computation is rendered energetically favorable while a compensating one produces heat elsewhere, a tactic that could be used to rationally design the flow of energy within a computer. △ Less

Submitted 12 November, 2023; v1 submitted 17 May, 2023; originally announced May 2023.

Remembering the work of Phillip L. Geissler: A coda to his scientific trajectory

Authors: Gregory R. Bowman, Stephen J. Cox, Christoph Dellago, Kateri H. DuBay, Joel D. Eaves, Daniel A. Fletcher, Layne B. Frechette, Michael Grünwald, Katherine Klymko, JiYeon Ku, Ahmad K. Omar, Eran Rabani, David R. Reichman, Julia R. Rogers, Andreana M. Rosnik, Grant M. Rotskoff, Anna R. Schneider, Nadine Schwierz, David A. Sivak, Suriyanarayanan Vaikuntanathan, Stephen Whitelam, Asaph Widmer-Cooper

Abstract: Phillip L. Geissler made important contributions to the statistical mechanics of biological polymers, heterogeneous materials, and chemical dynamics in aqueous environments. He devised analytical and computational methods that revealed the underlying organization of complex systems at the frontiers of biology, chemistry, and materials science. In this retrospective, we celebrate his work at these… ▽ More Phillip L. Geissler made important contributions to the statistical mechanics of biological polymers, heterogeneous materials, and chemical dynamics in aqueous environments. He devised analytical and computational methods that revealed the underlying organization of complex systems at the frontiers of biology, chemistry, and materials science. In this retrospective, we celebrate his work at these frontiers. △ Less

Submitted 24 February, 2023; originally announced February 2023.

Journal ref: Ann. Rev. Phys. Chem. 74, 11.1-11.27 (2023)

Demon in the machine: learning to extract work and absorb entropy from fluctuating nanosystems

Authors: Stephen Whitelam

Abstract: We use Monte Carlo and genetic algorithms to train neural-network feedback-control protocols for simulated fluctuating nanosystems. These protocols convert the information obtained by the feedback process into heat or work, allowing the extraction of work from a colloidal particle pulled by an optical trap and the absorption of entropy by an Ising model undergoing magnetization reversal. The learn… ▽ More We use Monte Carlo and genetic algorithms to train neural-network feedback-control protocols for simulated fluctuating nanosystems. These protocols convert the information obtained by the feedback process into heat or work, allowing the extraction of work from a colloidal particle pulled by an optical trap and the absorption of entropy by an Ising model undergoing magnetization reversal. The learning framework requires no prior knowledge of the system, depends only upon measurements that are accessible experimentally, and scales to systems of considerable complexity. It could be used in the laboratory to learn protocols for fluctuating nanosystems that convert measurement information into stored work or heat. △ Less

Submitted 10 April, 2023; v1 submitted 19 November, 2022; originally announced November 2022.

Journal ref: Phys. Rev. X 13, 021005 (2023)

Low-dissipation self-assembly protocols of active sticky particles

Authors: Stephen Whitelam, Jeremy D. Schmit

Abstract: We use neuroevolutionary learning to identify time-dependent protocols for low-dissipation self-assembly in a model of generic active particles with interactions. When the time allotted for assembly is sufficiently long, low-dissipation protocols use only interparticle attractions, producing an amount of entropy that scales as the number of particles. When time is too short to allow assembly to pr… ▽ More We use neuroevolutionary learning to identify time-dependent protocols for low-dissipation self-assembly in a model of generic active particles with interactions. When the time allotted for assembly is sufficiently long, low-dissipation protocols use only interparticle attractions, producing an amount of entropy that scales as the number of particles. When time is too short to allow assembly to proceed via diffusive motion, low-dissipation assembly protocols instead require particle self-propulsion, producing an amount of entropy that scales with the number of particles and the swim length required to cause assembly. Self-propulsion therefore provides an expensive but necessary mechanism for inducing assembly when time is of the essence. △ Less

Submitted 15 September, 2022; v1 submitted 9 June, 2022; originally announced June 2022.

Training neural networks using Metropolis Monte Carlo and an adaptive variant

Authors: Stephen Whitelam, Viktor Selin, Ian Benlolo, Corneel Casert, Isaac Tamblyn

Abstract: We examine the zero-temperature Metropolis Monte Carlo algorithm as a tool for training a neural network by minimizing a loss function. We find that, as expected on theoretical grounds and shown empirically by other authors, Metropolis Monte Carlo can train a neural net with an accuracy comparable to that of gradient descent, if not necessarily as quickly. The Metropolis algorithm does not fail au… ▽ More We examine the zero-temperature Metropolis Monte Carlo algorithm as a tool for training a neural network by minimizing a loss function. We find that, as expected on theoretical grounds and shown empirically by other authors, Metropolis Monte Carlo can train a neural net with an accuracy comparable to that of gradient descent, if not necessarily as quickly. The Metropolis algorithm does not fail automatically when the number of parameters of a neural network is large. It can fail when a neural network's structure or neuron activations are strongly heterogenous, and we introduce an adaptive Monte Carlo algorithm, aMC, to overcome these limitations. The intrinsic stochasticity and numerical stability of the Monte Carlo method allow aMC to train deep neural networks and recurrent neural networks in which the gradient is too small or too large to allow training by gradient descent. Monte Carlo methods offer a complement to gradient-based methods for training neural networks, allowing access to a distinct set of network architectures and principles. △ Less

Submitted 9 August, 2022; v1 submitted 15 May, 2022; originally announced May 2022.

Cellular automata can classify data by inducing trajectory phase coexistence

Authors: Stephen Whitelam, Isaac Tamblyn

Abstract: We show that cellular automata can classify data by inducing a form of dynamical phase coexistence. We use Monte Carlo methods to search for general two-dimensional deterministic automata that classify images on the basis of activity, the number of state changes that occur in a trajectory initiated from the image. When the number of timesteps of the automaton is a trainable parameter, the search s… ▽ More We show that cellular automata can classify data by inducing a form of dynamical phase coexistence. We use Monte Carlo methods to search for general two-dimensional deterministic automata that classify images on the basis of activity, the number of state changes that occur in a trajectory initiated from the image. When the number of timesteps of the automaton is a trainable parameter, the search scheme identifies automata that generate a population of dynamical trajectories displaying high or low activity, depending on initial conditions. Automata of this nature behave as nonlinear activation functions with an output that is effectively binary, resembling an emergent version of a spiking neuron. △ Less

Submitted 25 July, 2022; v1 submitted 10 March, 2022; originally announced March 2022.

Learning stochastic dynamics and predicting emergent behavior using transformers

Authors: Corneel Casert, Isaac Tamblyn, Stephen Whitelam

Abstract: We show that a neural network originally designed for language processing can learn the dynamical rules of a stochastic system by observation of a single dynamical trajectory of the system, and can accurately predict its emergent behavior under conditions not observed during training. We consider a lattice model of active matter undergoing continuous-time Monte Carlo dynamics, simulated at a densi… ▽ More We show that a neural network originally designed for language processing can learn the dynamical rules of a stochastic system by observation of a single dynamical trajectory of the system, and can accurately predict its emergent behavior under conditions not observed during training. We consider a lattice model of active matter undergoing continuous-time Monte Carlo dynamics, simulated at a density at which its steady state comprises small, dispersed clusters. We train a neural network called a transformer on a single trajectory of the model. The transformer, which we show has the capacity to represent dynamical rules that are numerous and nonlocal, learns that the dynamics of this model consists of a small number of processes. Forward-propagated trajectories of the trained transformer, at densities not encountered during training, exhibit motility-induced phase separation and so predict the existence of a nonequilibrium phase transition. Transformers have the flexibility to learn dynamical rules from observation without explicit enumeration of rates or coarse-graining of configuration space, and so the procedure used here can be applied to a wide range of physical systems, including those with large and complex dynamical generators. △ Less

Submitted 17 February, 2022; originally announced February 2022.

Comment on the literature definition(s) of "dynamical phase transition"

Authors: Stephen Whitelam

Abstract: Comment on the literature definition(s) of "dynamical phase transition" Comment on the literature definition(s) of "dynamical phase transition" △ Less

Submitted 16 December, 2021; originally announced December 2021.

Deposition control of model glasses with surface-mediated orientational order

Authors: Stephen Whitelam, Peter Harrowell

Abstract: We introduce a minimal model of solid-forming anisotropic molecules that displays, in thermal equilibrium, surface orientational order without bulk orientational order. The model reproduces the nonequilibrium behavior of recent experiments in that a bulk nonequilibrium structure grown by deposition contains regions of orientational order characteristic of the surface equilibrium. This order is dep… ▽ More We introduce a minimal model of solid-forming anisotropic molecules that displays, in thermal equilibrium, surface orientational order without bulk orientational order. The model reproduces the nonequilibrium behavior of recent experiments in that a bulk nonequilibrium structure grown by deposition contains regions of orientational order characteristic of the surface equilibrium. This order is deposited in general in a nonuniform way, because of the emergence of a growth-poisoning mechanism that causes equilibrated surfaces to grow slower than non-equilibrated surfaces. We use evolutionary methods to design oscillatory protocols able to grow nonequilibrium structures with uniform order, demonstrating the potential of protocol design for the fabrication of this class of materials. △ Less

Submitted 7 June, 2021; v1 submitted 9 May, 2021; originally announced May 2021.

Neuroevolutionary learning of particles and protocols for self-assembly

Authors: Stephen Whitelam, Isaac Tamblyn

Abstract: Within simulations of molecules deposited on a surface we show that neuroevolutionary learning can design particles and time-dependent protocols to promote self-assembly, without input from physical concepts such as thermal equilibrium or mechanical stability and without prior knowledge of candidate or competing structures. The learning algorithm is capable of both directed and exploratory design:… ▽ More Within simulations of molecules deposited on a surface we show that neuroevolutionary learning can design particles and time-dependent protocols to promote self-assembly, without input from physical concepts such as thermal equilibrium or mechanical stability and without prior knowledge of candidate or competing structures. The learning algorithm is capable of both directed and exploratory design: it can assemble a material with a user-defined property, or search for novelty in the space of specified order parameters. In the latter mode it explores the space of what can be made rather than the space of structures that are low in energy but not necessarily kinetically accessible. △ Less

Submitted 22 December, 2020; originally announced December 2020.

Journal ref: Phys. Rev. Lett. 127, 018003 (2021)

Varied phenomenology of models displaying dynamical large-deviation singularities

Authors: Stephen Whitelam, Daniel Jacobson

Abstract: Singularities of dynamical large-deviation functions are often interpreted as the signal of a dynamical phase transition and the coexistence of distinct dynamical phases, by analogy with the correspondence between singularities of free energies and equilibrium phase behavior. Here we study models of driven random walkers on a lattice. These models display large-deviation singularities in the limit… ▽ More Singularities of dynamical large-deviation functions are often interpreted as the signal of a dynamical phase transition and the coexistence of distinct dynamical phases, by analogy with the correspondence between singularities of free energies and equilibrium phase behavior. Here we study models of driven random walkers on a lattice. These models display large-deviation singularities in the limit of large lattice size, but the extent to which each model's phenomenology resembles a phase transition depends on the details of the driving. We also compare the behavior of ergodic and non-ergodic models that present large-deviation singularities. We argue that dynamical large-deviation singularities indicate the divergence of a model timescale, but not necessarily one associated with cooperative behavior or the existence of distinct phases. △ Less

Submitted 12 April, 2021; v1 submitted 16 December, 2020; originally announced December 2020.

Journal ref: Phys. Rev. E 103, 032152 (2021)

Dynamical large deviations of two-dimensional kinetically constrained models using a neural-network state ansatz

Authors: Corneel Casert, Tom Vieijra, Stephen Whitelam, Isaac Tamblyn

Abstract: We use a neural network ansatz originally designed for the variational optimization of quantum systems to study dynamical large deviations in classical ones. We obtain the scaled cumulant-generating function for the dynamical activity of the Fredrickson-Andersen model, a prototypical kinetically constrained model, in one and two dimensions, and present the first size-scaling analysis of the dynami… ▽ More We use a neural network ansatz originally designed for the variational optimization of quantum systems to study dynamical large deviations in classical ones. We obtain the scaled cumulant-generating function for the dynamical activity of the Fredrickson-Andersen model, a prototypical kinetically constrained model, in one and two dimensions, and present the first size-scaling analysis of the dynamical activity in two dimensions. These results provide a new route to the study of dynamical large-deviation functions, and highlight the broad applicability of the neural-network state ansatz across domains in physics. △ Less

Submitted 17 November, 2020; originally announced November 2020.

Journal ref: Phys. Rev. Lett. 127, 120602 (2021)

Correspondence between neuroevolution and gradient descent

Authors: Stephen Whitelam, Viktor Selin, Sang-Won Park, Isaac Tamblyn

Abstract: We show analytically that training a neural network by conditioned stochastic mutation or neuroevolution of its weights is equivalent, in the limit of small mutations, to gradient descent on the loss function in the presence of Gaussian white noise. Averaged over independent realizations of the learning process, neuroevolution is equivalent to gradient descent on the loss function. We use numerica… ▽ More We show analytically that training a neural network by conditioned stochastic mutation or neuroevolution of its weights is equivalent, in the limit of small mutations, to gradient descent on the loss function in the presence of Gaussian white noise. Averaged over independent realizations of the learning process, neuroevolution is equivalent to gradient descent on the loss function. We use numerical simulation to show that this correspondence can be observed for finite mutations,for shallow and deep neural networks. Our results provide a connection between two families of neural-network training methods that are usually considered to be fundamentally different. △ Less

Submitted 10 September, 2021; v1 submitted 14 August, 2020; originally announced August 2020.

Hysteresis curves reveal the microscopic origin of cooperative CO$_2$ adsorption in diamine-appended metal-organic frameworks

Authors: John R. Edison, Rebecca L. Siegelman, Zdenek Preisler, Joyjit Kundu, Jeffrey R. Long, Stephen Whitelam

Abstract: Diamine-appended metal{organic frameworks (MOFs) of the form Mg2(dobpdc)(diamine)2 adsorb CO2 in a cooperative fashion, exhibiting an abrupt change in CO2 occupancy with pressure or temperature. This change is accompanied by hysteresis. While hysteresis is suggestive of a firstorder phase transition, we show that hysteretic temperature-occupancy curves associated with this material are qualitative… ▽ More Diamine-appended metal{organic frameworks (MOFs) of the form Mg2(dobpdc)(diamine)2 adsorb CO2 in a cooperative fashion, exhibiting an abrupt change in CO2 occupancy with pressure or temperature. This change is accompanied by hysteresis. While hysteresis is suggestive of a firstorder phase transition, we show that hysteretic temperature-occupancy curves associated with this material are qualitatively unlike the curves seen in the presence of a phase transition; they are instead consistent with CO2 chain polymerization, within one-dimensional channels in the MOF, in the absence of a phase transition. Our simulations of a microscopic model reproduce this dynamics, and point the way toward rational control, in and out of equilibrium, of cooperative adsorption in this industrially important class of materials. △ Less

Submitted 11 October, 2020; v1 submitted 25 April, 2020; originally announced April 2020.

Long-Range Exciton Diffusion in Two-Dimensional Assemblies of Cesium Lead Bromide Perovskite Nanocrystals

Authors: Erika Penzo, Anna Loiudice, Edward S. Barnard, Nicholas J. Borys, Matthew J. Jurow, Monica Lorenzon, Igor Rajzbaum, Edward K. Wong, Yi Liu, Adam M. Schwartzberg, Stefano Cabrini, Stephen Whitelam, Raffaella Buonsanti, Alexander Weber-Bargioni

Abstract: Förster Resonant Energy Transfer (FRET)-mediated exciton diffusion through artificial nanoscale building block assemblies could be used as a new optoelectronic design element to transport energy. However, so far nanocrystal (NC) systems supported only diffusion length of 30 nm, which are too small to be useful in devices. Here, we demonstrate a FRET-mediated exciton diffusion length of 200 nm with… ▽ More Förster Resonant Energy Transfer (FRET)-mediated exciton diffusion through artificial nanoscale building block assemblies could be used as a new optoelectronic design element to transport energy. However, so far nanocrystal (NC) systems supported only diffusion length of 30 nm, which are too small to be useful in devices. Here, we demonstrate a FRET-mediated exciton diffusion length of 200 nm with 0.5 cm2/s diffusivity through an ordered, two-dimensional assembly of cesium lead bromide perovskite nanocrystals (PNC). Exciton diffusion was directly measured via steady-state and time-resolved photoluminescence (PL) microscopy, with physical modeling providing deeper insight into the transport process. This exceptionally efficient exciton transport is facilitated by PNCs high PL quantum yield, large absorption cross-section, and high polarizability, together with minimal energetic and geometric disorder of the assembly. This FRET-mediated exciton diffusion length matches perovskites optical absorption depth, opening the possibility to design new optoelectronic device architectures with improved performances, and providing insight into the high conversion efficiencies of PNC-based optoelectronic devices. △ Less

Submitted 1 September, 2020; v1 submitted 9 March, 2020; originally announced March 2020.

Watch and learn -- a generalized approach for transferrable learning in deep neural networks via physical principles

Authors: Kyle Sprague, Juan Carrasquilla, Steve Whitelam, Isaac Tamblyn

Abstract: Transfer learning refers to the use of knowledge gained while solving a machine learning task and applying it to the solution of a closely related problem. Such an approach has enabled scientific breakthroughs in computer vision and natural language processing where the weights learned in state-of-the-art models can be used to initialize models for other tasks which dramatically improve their perf… ▽ More Transfer learning refers to the use of knowledge gained while solving a machine learning task and applying it to the solution of a closely related problem. Such an approach has enabled scientific breakthroughs in computer vision and natural language processing where the weights learned in state-of-the-art models can be used to initialize models for other tasks which dramatically improve their performance and save computational time. Here we demonstrate an unsupervised learning approach augmented with basic physical principles that achieves fully transferrable learning for problems in statistical physics across different physical regimes. By coupling a sequence model based on a recurrent neural network to an extensive deep neural network, we are able to learn the equilibrium probability distributions and inter-particle interaction models of classical statistical mechanical systems. Our approach, distribution-consistent learning, DCL, is a general strategy that works for a variety of canonical statistical mechanical models (Ising and Potts) as well as disordered (spin-glass) interaction potentials. Using data collected from a single set of observation conditions, DCL successfully extrapolates across all temperatures, thermodynamic phases, and can be applied to different length-scales. This constitutes a fully transferrable physics-based learning in a generalizable approach. △ Less

Submitted 3 March, 2020; originally announced March 2020.

Learning to grow: control of material self-assembly using evolutionary reinforcement learning

Authors: Stephen Whitelam, Isaac Tamblyn

Abstract: We show that neural networks trained by evolutionary reinforcement learning can enact efficient molecular self-assembly protocols. Presented with molecular simulation trajectories, networks learn to change temperature and chemical potential in order to promote the assembly of desired structures or choose between competing polymorphs. In the first case, networks reproduce in a qualitative sense the… ▽ More We show that neural networks trained by evolutionary reinforcement learning can enact efficient molecular self-assembly protocols. Presented with molecular simulation trajectories, networks learn to change temperature and chemical potential in order to promote the assembly of desired structures or choose between competing polymorphs. In the first case, networks reproduce in a qualitative sense the results of previously-known protocols, but faster and with higher fidelity; in the second case they identify strategies previously unknown, from which we can extract physical insight. Networks that take as input the elapsed time of the simulation or microscopic information from the system are both effective, the latter more so. The evolutionary scheme we have used is simple to implement and can be applied to a broad range of examples of experimental self-assembly, whether or not one can monitor the experiment as it proceeds. Our results have been achieved with no human input beyond the specification of which order parameter to promote, pointing the way to the design of synthesis protocols by artificial intelligence. △ Less

Submitted 28 May, 2020; v1 submitted 17 December, 2019; originally announced December 2019.

Journal ref: Phys. Rev. E 101, 052604 (2020)

Decorated Protein Networks: Functional Nanomaterials with Tunable Domain Size

Authors: Ioatzin Rıos de Anda, Angelique Coutable-Pennarun, Chris Brasnett, Stephen Whitelam, Annela Seddon, John Russo, J. L. Ross Anderson, C. Patrick Royall1

Abstract: The implementation of natural and artificial proteins with designer properties and functionalities offers unparalleled opportunity for functional nanoarchitectures formed through self-assembly. However, to exploit the opportunities offered we require the ability to control protein assembly into the desired architecture while avoiding denaturation and therefore retaining protein functionality. Here… ▽ More The implementation of natural and artificial proteins with designer properties and functionalities offers unparalleled opportunity for functional nanoarchitectures formed through self-assembly. However, to exploit the opportunities offered we require the ability to control protein assembly into the desired architecture while avoiding denaturation and therefore retaining protein functionality. Here we address this challenge with a model system of fluorescent proteins. Using techniques of self-assembly manipulation inspired by soft matter where interactions between components are controlled to yield the desired structure, we show that it is possible to assemble networks of proteins of one species which we can decorate with another, whose coverage we can tune. Consequently, the interfaces between domains of each component can also be tuned, with applications for example in energy transfer. Our model system of fluorescent proteins eGFP and mCherry retain their fluorescence throughout the assembly process, thus demonstrating that functionality is preserved. △ Less

Submitted 13 November, 2019; originally announced November 2019.

Evolutionary reinforcement learning of dynamical large deviations

Authors: Stephen Whitelam, Daniel Jacobson, Isaac Tamblyn

Abstract: We show how to calculate the likelihood of dynamical large deviations using evolutionary reinforcement learning. An agent, a stochastic model, propagates a continuous-time Monte Carlo trajectory and receives a reward conditioned upon the values of certain path-extensive quantities. Evolution produces progressively fitter agents, eventually allowing the calculation of a piece of a large-deviation r… ▽ More We show how to calculate the likelihood of dynamical large deviations using evolutionary reinforcement learning. An agent, a stochastic model, propagates a continuous-time Monte Carlo trajectory and receives a reward conditioned upon the values of certain path-extensive quantities. Evolution produces progressively fitter agents, eventually allowing the calculation of a piece of a large-deviation rate function for a particular model and path-extensive quantity. For models with small state spaces the evolutionary process acts directly on rates, and for models with large state spaces the process acts on the weights of a neural network that parameterizes the model's rates. This approach shows how path-extensive physics problems can be considered within a framework widely used in machine learning. △ Less

Submitted 21 February, 2020; v1 submitted 2 September, 2019; originally announced September 2019.

Optimizing thermodynamic trajectories using evolutionary and gradient-based reinforcement learning

Authors: Chris Beeler, Uladzimir Yahorau, Rory Coles, Kyle Mills, Stephen Whitelam, Isaac Tamblyn

Abstract: Using a model heat engine, we show that neural network-based reinforcement learning can identify thermodynamic trajectories of maximal efficiency. We consider both gradient and gradient-free reinforcement learning. We use an evolutionary learning algorithm to evolve a population of neural networks, subject to a directive to maximize the efficiency of a trajectory composed of a set of elementary th… ▽ More Using a model heat engine, we show that neural network-based reinforcement learning can identify thermodynamic trajectories of maximal efficiency. We consider both gradient and gradient-free reinforcement learning. We use an evolutionary learning algorithm to evolve a population of neural networks, subject to a directive to maximize the efficiency of a trajectory composed of a set of elementary thermodynamic processes; the resulting networks learn to carry out the maximally-efficient Carnot, Stirling, or Otto cycles. When given an additional irreversible process, this evolutionary scheme learns a previously unknown thermodynamic cycle. Gradient-based reinforcement learning is able to learn the Stirling cycle, whereas an evolutionary approach achieves the optimal Carnot cycle. Our results show how the reinforcement learning strategies developed for game playing can be applied to solve physical problems conditioned upon path-extensive order parameters. △ Less

Submitted 22 November, 2021; v1 submitted 20 March, 2019; originally announced March 2019.

Comments: 11 pages, 5 figures

Journal ref: Phys. Rev. E 104, 064128 (2021)

Direct evaluation of dynamical large-deviation rate functions using a variational ansatz

Authors: Daniel Jacobson, Stephen Whitelam

Abstract: We describe a simple form of importance sampling designed to bound and compute large-deviation rate functions for time-extensive dynamical observables in continuous-time Markov chains. We start with a model, defined by a set of rates, and a time-extensive dynamical observable. We construct a reference model, a variational ansatz for the behavior of the original model conditioned on atypical values… ▽ More We describe a simple form of importance sampling designed to bound and compute large-deviation rate functions for time-extensive dynamical observables in continuous-time Markov chains. We start with a model, defined by a set of rates, and a time-extensive dynamical observable. We construct a reference model, a variational ansatz for the behavior of the original model conditioned on atypical values of the observable. Direct simulation of the reference model provides an upper bound on the large-deviation rate function associated with the original model, an estimate of the tightness of the bound, and, if the ansatz is chosen well, the exact rate function. The exact rare behavior of the original model does not need to be known in advance. We use this method to calculate rate functions for currents and counting observables in a set of network- and lattice models taken from the literature. Straightforward ansatze yield bounds that are tighter than bounds obtained from Level 2.5 of large deviations via approximations that involve uniform scalings of rates. We show how to correct these bounds in order to recover the rate functions exactly. Our approach is complementary to more specialized methods, and offers a physically transparent framework for approximating and calculating the likelihood of dynamical large deviations. △ Less

Submitted 18 September, 2019; v1 submitted 14 March, 2019; originally announced March 2019.

Journal ref: Phys. Rev. E 100, 052139 (2019)

Improving the accuracy of nearest-neighbor classification using principled construction and stochastic sampling of training-set centroids

Authors: Stephen Whitelam

Abstract: A conceptually simple way to classify images is to directly compare test-set data and training-set data. The accuracy of this approach is limited by the method of comparison used, and by the extent to which the training-set data cover configuration space. Here we show that this coverage can be substantially increased using coarse graining (replacing groups of images by their centroids) and stochas… ▽ More A conceptually simple way to classify images is to directly compare test-set data and training-set data. The accuracy of this approach is limited by the method of comparison used, and by the extent to which the training-set data cover configuration space. Here we show that this coverage can be substantially increased using coarse graining (replacing groups of images by their centroids) and stochastic sampling (using distinct sets of centroids in combination). We use the MNIST and Fashion-MNIST data sets to show that a principled coarse-graining algorithm can convert training images into fewer image centroids without loss of accuracy of classification of test-set images by nearest-neighbor classification. Distinct batches of centroids can be used in combination as a means of stochastically sampling configuration space, and can classify test-set data more accurately than can the unaltered training set. On the MNIST and Fashion-MNIST data sets this approach converts nearest-neighbor classification from a mid-ranking- to an upper-ranking member of the set of classical machine-learning techniques. △ Less

Submitted 4 February, 2021; v1 submitted 7 September, 2018; originally announced September 2018.

Journal ref: Entropy 23, 149 (2021)

Multi-point nonequilibrium umbrella sampling and associated fluctuation relations

Authors: Stephen Whitelam

Abstract: We describe a simple method of umbrella trajectory sampling for Markov chains. The method allows the estimation of large-deviation rate functions, for path-extensive dynamic observables, for an arbitrary number of models within a certain family. The general relationship between probability distributions of dynamic observables of members of this family is an extended fluctuation relation. When the… ▽ More We describe a simple method of umbrella trajectory sampling for Markov chains. The method allows the estimation of large-deviation rate functions, for path-extensive dynamic observables, for an arbitrary number of models within a certain family. The general relationship between probability distributions of dynamic observables of members of this family is an extended fluctuation relation. When the dynamic observable is chosen to be entropy production, members of this family include the forward Markov chain and its time reverse, whose probability distributions are related by the expected simple fluctuation relation. △ Less

Submitted 23 May, 2018; v1 submitted 12 March, 2018; originally announced March 2018.

Cooperative gas adsorption without a phase transition in metal-organic frameworks

Authors: Joyjit Kundu, Jurgen. F. Stilck, Jung-Hoon Lee, Jeffrey B. Neaton, David Prendergast, Stephen Whitelam

Abstract: Cooperative adsorption of gases by porous frameworks permits more efficient uptake and removal than does the more usual non-cooperative (Langmuir-type) adsorption. Cooperativity, signaled by a step-like isotherm, is usually attributed to a phase transition of the framework. However, the class of metal-organic frameworks mmen-M$_2$(dobpdc) exhibit cooperative adsorption of CO2 but show no evidence… ▽ More Cooperative adsorption of gases by porous frameworks permits more efficient uptake and removal than does the more usual non-cooperative (Langmuir-type) adsorption. Cooperativity, signaled by a step-like isotherm, is usually attributed to a phase transition of the framework. However, the class of metal-organic frameworks mmen-M$_2$(dobpdc) exhibit cooperative adsorption of CO2 but show no evidence of a phase transition. Here we show how cooperativity emerges in these frameworks in the absence of a phase transition. We use a combination of quantum and statistical mechanics to show that cooperativity results from a sharp but finite increase, with pressure, of the mean length of chains of CO2 molecules that polymerize within the framework. Our study provides microscopic understanding of the emergent features of cooperative binding, including the position, slope and height of the isotherm step, and indicates how to optimize gas storage and separation in these materials. △ Less

Submitted 13 December, 2017; originally announced December 2017.

Comments: 18 pages, 11 figures

Journal ref: Phys. Rev. Lett. 121, 015701 (2018)

Large deviations in the presence of cooperativity and slow dynamics

Authors: Stephen Whitelam

Abstract: We study simple models of intermittency, involving switching between two states, within the dynamical large-deviation formalism. Singularities appear in the formalism when switching is cooperative, or when its basic timescale diverges. In the first case the unbiased trajectory distribution undergoes a symmetry breaking, leading to a change of shape of the large-deviation rate function for a partic… ▽ More We study simple models of intermittency, involving switching between two states, within the dynamical large-deviation formalism. Singularities appear in the formalism when switching is cooperative, or when its basic timescale diverges. In the first case the unbiased trajectory distribution undergoes a symmetry breaking, leading to a change of shape of the large-deviation rate function for a particular dynamical observable. In the second case the symmetry of the unbiased trajectory distribution remains unbroken. Comparison of these models suggests that singularities of the dynamical large-deviation formalism can signal the dynamical equivalent of an equilibrium phase transition, but do not necessarily do so. △ Less

Submitted 25 May, 2018; v1 submitted 5 December, 2017; originally announced December 2017.

Journal ref: Phys. Rev. E 97, 062109 (2018)

Sampling rare fluctuations of discrete-time Markov chains

Authors: Stephen Whitelam

Abstract: We describe a simple method that can be used to sample the rare fluctuations of discrete-time Markov chains. We focus on the case of Markov chains with well-defined steady-state measures, and derive expressions for the large-deviation rate functions (and upper bounds on such functions) for dynamical quantities extensive in the length of the Markov chain. We illustrate the method using a series of… ▽ More We describe a simple method that can be used to sample the rare fluctuations of discrete-time Markov chains. We focus on the case of Markov chains with well-defined steady-state measures, and derive expressions for the large-deviation rate functions (and upper bounds on such functions) for dynamical quantities extensive in the length of the Markov chain. We illustrate the method using a series of simple examples, and use it to study the fluctuations of a lattice-based model of active matter that can undergo motility-induced phase separation. △ Less

Submitted 11 March, 2018; v1 submitted 12 September, 2017; originally announced September 2017.

Comments: Submitted along with arXiv:1709.03951 as a joint work

Journal ref: Phys. Rev. E 97, 032122 (2018)

Phase separation and large deviations of lattice active matter

Authors: Stephen Whitelam, Katherine Klymko, Dibyendu Mandal

Abstract: Off-lattice active Brownian particles form clusters and undergo phase separation even in the absence of attractions or velocity-alignment mechanisms. Arguments that explain this phenomenon appeal only to the ability of particles to move persistently in a direction that fluctuates, but existing lattice models of hard particles that account for this behavior do not exhibit phase separation. Here we… ▽ More Off-lattice active Brownian particles form clusters and undergo phase separation even in the absence of attractions or velocity-alignment mechanisms. Arguments that explain this phenomenon appeal only to the ability of particles to move persistently in a direction that fluctuates, but existing lattice models of hard particles that account for this behavior do not exhibit phase separation. Here we present a lattice model of active matter that exhibits motility-induced phase separation in the absence of velocity alignment. Using direct and rare-event sampling of dynamical trajectories we show that clustering and phase separation are accompanied by pronounced fluctuations of static and dynamic order parameters. This model provides a complement to off-lattice models for the study of motility-induced phase separation. △ Less

Submitted 16 November, 2017; v1 submitted 12 September, 2017; originally announced September 2017.

Comments: Submitted along with arXiv:1709.03951 as a joint work to PRE and PRL

Journal ref: J. Chem. Phys. 148, 154902 (2018)

Strong bonds and far-from-equilibrium conditions minimize errors in lattice-gas growth

Authors: Stephen Whitelam

Abstract: We use computer simulation to study the layer-by-layer growth of particle structures in a lattice gas, taking the number of incorporated vacancies as a measure of the quality of the grown structure. By exploiting a dynamic scaling relation between structure quality in and out of equilibrium, we determine that the best quality of structure is obtained, for fixed observation time, with strong intera… ▽ More We use computer simulation to study the layer-by-layer growth of particle structures in a lattice gas, taking the number of incorporated vacancies as a measure of the quality of the grown structure. By exploiting a dynamic scaling relation between structure quality in and out of equilibrium, we determine that the best quality of structure is obtained, for fixed observation time, with strong interactions and far-from-equilibrium growth conditions. This result contrasts with the usual assumption that weak interactions and mild nonequilibrium conditions are the best way to minimize errors during assembly. △ Less

Submitted 4 August, 2017; originally announced August 2017.

Irregular model DNA particles self-assemble into a regular structure

Authors: Zdenek Preisler, Barbara Sacca, Stephen Whitelam

Abstract: DNA nanoparticles with three-fold coordination have been observed to self-assemble in experiment into a network equivalent to the hexagonal (6.6.6) tiling, and a network equivalent to the 4.8.8 Archimedean tiling. Both networks are built from a single type of vertex. Here we use analytic theory and equilibrium and dynamic simulation to show that a model particle, whose rotational properties lie be… ▽ More DNA nanoparticles with three-fold coordination have been observed to self-assemble in experiment into a network equivalent to the hexagonal (6.6.6) tiling, and a network equivalent to the 4.8.8 Archimedean tiling. Both networks are built from a single type of vertex. Here we use analytic theory and equilibrium and dynamic simulation to show that a model particle, whose rotational properties lie between those those of the vertices of the 6.6.6 and 4.8.8 networks, can self-assemble into a network built from three types of vertex. Important in forming this network is the ability of the particle to rotate when bound, thereby allowing the formation of more than one type of binding motif. The network in question is equivalent to a false tiling, a periodic structure built from irregular polygons, and possesses 40 particles in its unit cell. The emergence of this complex structure, whose symmetry properties are not obviously related to those of its constituent particles, highlights the potential for creating new structures from simple variants of existing nanoparticles. △ Less

Submitted 25 July, 2017; originally announced July 2017.

Journal ref: Soft Matter, 13, 47, 8894 (2017)

Rare behavior of growth processes via umbrella sampling of trajectories

Authors: Katherine Klymko, Phillip L. Geissler, Juan P. Garrahan, Stephen Whitelam

Abstract: We compute probability distributions of trajectory observables for reversible and irreversible growth processes. These results reveal a correspondence between reversible and irreversible processes, at particular points in parameter space, in terms of their typical and atypical trajectories. Thus key features of growth processes can be insensitive to the precise form of the rate constants used to g… ▽ More We compute probability distributions of trajectory observables for reversible and irreversible growth processes. These results reveal a correspondence between reversible and irreversible processes, at particular points in parameter space, in terms of their typical and atypical trajectories. Thus key features of growth processes can be insensitive to the precise form of the rate constants used to generate them, recalling the insensitivity to microscopic details of certain equilibrium behavior. We obtained these results using a sampling method, inspired by the "s-ensemble" large-deviation formalism, that amounts to umbrella sampling in trajectory space. The method is a simple variant of existing approaches, and applies to ensembles of trajectories controlled by the total number of events. It can be used to determine large-deviation rate functions for trajectory observables in or out of equilibrium. △ Less

Submitted 8 February, 2018; v1 submitted 3 July, 2017; originally announced July 2017.

Journal ref: Phys. Rev. E 97, 032123 (2018)

Minimal positive design for self-assembly of the Archimedean tilings

Authors: Stephen Whitelam

Abstract: A challenge of molecular self-assembly is to understand how to design particles that self-assemble into a desired structure and not any of a potentially large number of undesired structures. Here we use simulation to show that a strategy of minimal positive design allows the self-assembly of networks equivalent to the 8 semiregular Archimedean tilings of the plane, structures not previously realiz… ▽ More A challenge of molecular self-assembly is to understand how to design particles that self-assemble into a desired structure and not any of a potentially large number of undesired structures. Here we use simulation to show that a strategy of minimal positive design allows the self-assembly of networks equivalent to the 8 semiregular Archimedean tilings of the plane, structures not previously realized in simulation. This strategy consists of identifying the fewest distinct types of interparticle interaction that appear in the desired structure, and does not require enumeration of the many possible undesired structures. The resulting particles, which self-assemble into the desired networks, possess DNA-like selectivity of their interactions. Assembly of certain molecular networks may therefore require such selectivity. △ Less

Submitted 23 February, 2017; v1 submitted 1 June, 2016; originally announced June 2016.

Journal ref: Phys. Rev. Lett. 117, 228003 (2016)

Selective gas capture via kinetic trap**

Authors: Joyjit Kundu, Tod Pascal, David Prendergast, Stephen Whitelam

Abstract: Conventional approaches to the capture of CO_2 by metal-organic frameworks focus on equilibrium conditions, and frameworks that contain little CO_2 in equilibrium are often rejected as carbon-capture materials. Here we use a statistical mechanical model, parameterized by quantum mechanical data, to suggest that metal-organic frameworks can be used to separate CO_2 from a typical flue gas mixture w… ▽ More Conventional approaches to the capture of CO_2 by metal-organic frameworks focus on equilibrium conditions, and frameworks that contain little CO_2 in equilibrium are often rejected as carbon-capture materials. Here we use a statistical mechanical model, parameterized by quantum mechanical data, to suggest that metal-organic frameworks can be used to separate CO_2 from a typical flue gas mixture when used under {\em nonequilibrium} conditions. The origin of this selectivity is an emergent gas-separation mechanism that results from the acquisition by different gas types of different mobilities within a crowded framework. The resulting distribution of gas types within the framework is in general spatially and dynamically heterogeneous. Our results suggest that relaxing the requirement of equilibrium can substantially increase the parameter space of conditions and materials for which selective gas capture can be effected. △ Less

Submitted 31 May, 2016; originally announced June 2016.

Comments: 12 pages, 10 figures

Similarity of ensembles of trajectories of reversible and irreversible growth processes

Authors: Katherine Klymko, Juan P. Garrahan, Stephen Whitelam

Abstract: Models of bacterial growth tend to be `irreversible', allowing for the number of bacteria in a colony to increase but not to decrease. By contrast, models of molecular self-assembly are usually `reversible', allowing for addition and removal of particles to a structure. Such processes differ in a fundamental way because only reversible processes possess an equilibrium. Here we show at mean-field l… ▽ More Models of bacterial growth tend to be `irreversible', allowing for the number of bacteria in a colony to increase but not to decrease. By contrast, models of molecular self-assembly are usually `reversible', allowing for addition and removal of particles to a structure. Such processes differ in a fundamental way because only reversible processes possess an equilibrium. Here we show at mean-field level that dynamic trajectories of reversible and irreversible growth processes are similar in that both feel the influence of attractors, at which growth proceeds without limit but the intensive properties of the system are invariant. Attractors of both processes undergo nonequilibrium phase transitions as model parameters are varied, suggesting a unified way of describing reversible and irreversible growth. We also establish a connection at mean-field level between an irreversible model of growth (the magnetic Eden model) and the equilibrium Ising model, supporting the findings made by other authors using numerical simulations. △ Less

Submitted 18 March, 2016; originally announced March 2016.

Journal ref: Phys. Rev. E 96, 042126 (2017)

On the microscopic origin and macroscopic implications of lane formation in mixtures of oppositely-driven particles

Authors: Katherine Klymko, Phillip L. Geissler, Stephen Whitelam

Abstract: Colloidal particles of two types, driven in opposite directions, can segregate into lanes [Vissers et al. Soft Matter 7, 2352 (2011)]. This phenomenon can be reproduced by two-dimensional Brownian dynamics simulations of model particles [Dzubiella et al. Phys. Rev. E 65, 021402 (2002)]. Here we use computer simulation to assess the generality of lane formation with respect to variation of particle… ▽ More Colloidal particles of two types, driven in opposite directions, can segregate into lanes [Vissers et al. Soft Matter 7, 2352 (2011)]. This phenomenon can be reproduced by two-dimensional Brownian dynamics simulations of model particles [Dzubiella et al. Phys. Rev. E 65, 021402 (2002)]. Here we use computer simulation to assess the generality of lane formation with respect to variation of particle type and dynamical protocol. We find that laning results from rectification of diffusion on the scale of a particle diameter: oppositely-driven particles must, in the time taken to encounter each other in the direction of the drive, diffuse in the perpendicular direction by about one particle diameter. This geometric constraint implies that the diffusion constant of a particle, in the presence of those of the opposite type, grows approximately linearly with Peclet number, a prediction confirmed by our numerics over a range of model parameters. Such environment-dependent diffusion is statistically similar to an effective interparticle attraction; consistent with this observation, we find that oppositely-driven non-attractive colloids display features characteristic of the simplest model system possessing both interparticle attractions and persistent motion, the driven Ising lattice gas [Katz, Leibowitz, Spohn, J. Stat. Phys. 34, 497 (1984)]. These features include long-ranged correlations in the disordered regime, and a critical regime characterized by a change in slope of the particle current with Peclet number and by fluctuations that grow with system size. By analogy, we suggest that lane formation in the driven colloid system is in the macroscopic limit a phase transition, but that macroscopic phase separation would not occur in finite time upon starting from disordered initial conditions. △ Less

Submitted 29 August, 2016; v1 submitted 4 February, 2016; originally announced February 2016.

Journal ref: Phys. Rev. E 94, 022608 (2016)

The Ramachandran number: an order parameter for protein geometry

Authors: Ranjan V. Mannige, Joyjit Kundu, Stephen Whitelam

Abstract: Three-dimensional protein structures usually contain regions of local order, called secondary structure, such as $α$-helices and $β$-sheets. Secondary structure is characterized by the local rotational state of the protein backbone, quantified by two dihedral angles called $φ$ and $ψ$. Particular types of secondary structure can generally be described by a single (diffuse) location on a two-dimens… ▽ More Three-dimensional protein structures usually contain regions of local order, called secondary structure, such as $α$-helices and $β$-sheets. Secondary structure is characterized by the local rotational state of the protein backbone, quantified by two dihedral angles called $φ$ and $ψ$. Particular types of secondary structure can generally be described by a single (diffuse) location on a two-dimensional plot drawn in the space of the angles $φ$ and $ψ$, called a Ramachandran plot. By contrast, a recently-discovered nanomaterial made from peptoids, structural isomers of peptides, displays a secondary-structure motif corresponding to two regions on the Ramachandran plot [Mannige et al., Nature 526, 415 (2015)]. In order to describe such `higher-order' secondary structure in a compact way we introduce here a means of describing regions on the Ramachandran plot in terms of a single Ramachandran number, ${\mathcal{R}}$, which is a structurally meaningful combination of $φ$ and $ψ$. We show that the potential applications of ${\mathcal{R}}$ are numerous: it can be used to describe the geometric content of protein structures, and can be used to draw diagrams that reveal, at a glance, the frequency of occurrence of regular secondary structures and disordered regions in large protein datasets. We propose that ${\mathcal{R}}$ might be used as an order parameter for protein geometry for a wide range of applications. △ Less

Submitted 24 November, 2015; v1 submitted 10 November, 2015; originally announced November 2015.

Minimal physical requirements for crystal growth self-poisoning

Authors: Stephen Whitelam, Yuba Raj Dahal, Jeremy D. Schmit

Abstract: Self-poisoning is a kinetic trap that can impair or prevent crystal growth in a wide variety of physical settings. Here we use dynamic mean-field theory and computer simulation to argue that poisoning is ubiquitous because its emergence requires only the notion that a molecule can bind in two (or more) ways to a crystal; that those ways are not energetically equivalent; and that the associated bin… ▽ More Self-poisoning is a kinetic trap that can impair or prevent crystal growth in a wide variety of physical settings. Here we use dynamic mean-field theory and computer simulation to argue that poisoning is ubiquitous because its emergence requires only the notion that a molecule can bind in two (or more) ways to a crystal; that those ways are not energetically equivalent; and that the associated binding events occur with sufficiently unequal probability. If these conditions are met then the steady-state growth rate is in general a non-monotonic function of the thermodynamic driving force for crystal growth, which is the characteristic of poisoning. Our results also indicate that relatively small changes of system parameters could be used to induce recovery from poisoning. △ Less

Submitted 4 January, 2016; v1 submitted 19 October, 2015; originally announced October 2015.

Journal ref: J. Chem. Phys. 144, 064903 (2016)

Crystallization and arrest mechanisms of model colloids

Authors: Thomas K. Haxton, Lester O. Hedges, Stephen Whitelam

Abstract: We performed dynamic simulations of spheres with short-range attractive interactions for many values of interaction strength and range. Fast crystallization occurs in a localized region of this parameter space, but the character of crystallization pathways is not uniform within this region. Pathways range from one-step, in which a crystal nucleates directly from a gas, to two-step, in which substa… ▽ More We performed dynamic simulations of spheres with short-range attractive interactions for many values of interaction strength and range. Fast crystallization occurs in a localized region of this parameter space, but the character of crystallization pathways is not uniform within this region. Pathways range from one-step, in which a crystal nucleates directly from a gas, to two-step, in which substantial liquid-like clusters form and only subsequently become crystalline. Crystallization can fail because of slow nucleation from either gas or liquid, or because of dynamic arrest caused by strong interactions. Arrested states are characterized by the formation of networks of face-sharing tetrahedra that can be detected by a local common neighbor analysis. △ Less

Submitted 24 July, 2015; originally announced July 2015.

Comments: 17 pages (including appendix), 13 figures; open-source C++ library for implementing the virtual-move Monte Carlo algorithm is available at http://vmmc.xyz

Predicting the outcome of the growth of binary solids far from equilibrium

Authors: Ranjan V. Mannige, Stephen Whitelam

Abstract: The growth of multicomponent structures in simulations and experiments often results in kinetically trapped, nonequilibrium objects. In such cases we have no general theoretical framework for predicting the outcome of the growth process. Here we use computer simulations to study the growth of two-component structures within a simple lattice model. We show that kinetic trap** happens for many cho… ▽ More The growth of multicomponent structures in simulations and experiments often results in kinetically trapped, nonequilibrium objects. In such cases we have no general theoretical framework for predicting the outcome of the growth process. Here we use computer simulations to study the growth of two-component structures within a simple lattice model. We show that kinetic trap** happens for many choices of growth rate and inter-component interaction energies, and that qualitatively distinct kinds of kinetic trap** are found in different regions of parameter space. In a region in which the low-energy structure is an `antiferromagnet' or `checkerboard', we show that the grown nonequilibrium structure displays a component-type stoichiometry that is different to the equilibrium one but is insensitive to growth rate and solution conditions. This robust nonequilibrium stoichiometry can be predicted via a map** to the jammed random tiling of dimers studied by Flory, a finding that suggests a way of making defined nonequilibrium structures in experiment. △ Less

Submitted 22 July, 2015; originally announced July 2015.

Hierarchical assembly may be a way to make large information-rich structures

Authors: Stephen Whitelam

Abstract: Self-assembly in the laboratory can now yield `information-rich' nanostructures in which each component is of a distinct type and has a defined spatial position. Ensuring the thermodynamic stability of such structures requires inter-component interaction energies to increase logarithmically with structure size, in order to counter the entropy gained upon mixing component types in solution. However… ▽ More Self-assembly in the laboratory can now yield `information-rich' nanostructures in which each component is of a distinct type and has a defined spatial position. Ensuring the thermodynamic stability of such structures requires inter-component interaction energies to increase logarithmically with structure size, in order to counter the entropy gained upon mixing component types in solution. However, self-assembly in the presence of strong interactions results in general in kinetic trap**, so suggesting a limit to the size of an (equilibrium) structure that can be self-assembled from distinguishable components. Here we study numerically a two-dimensional hierarchical assembly scheme already considered in experiment. We show that this scheme is immune to the kinetic traps associated with strong `native' interactions (interactions designed to stabilize the intended structure), and so, in principle, offers a way to make large information-rich structures. In this scheme the size of an assembled structure scales exponentially with the stage of assembly, and assembly can continue as long as random motion is able to bring structures into contact. The resulting superstructure could provide a template for building in the third dimension. The chief drawback of this scheme is that it is particularly susceptible to kinetic traps that result from `non-native' interactions (interactions not required to stabilize the intended structure); the scale on which such a scheme can be realized therefore depends upon how effectively this latter kind of interaction can be suppressed. △ Less

Submitted 3 September, 2015; v1 submitted 27 May, 2015; originally announced May 2015.

Journal ref: Soft Matter 11, 8225 (2015)

Examples of molecular self-assembly at surfaces

Authors: Stephen Whitelam

Abstract: The self-assembly of molecules at surfaces can be caused by a range of physical mechanisms. Assembly can be driven by intermolecular forces, or molecule-surface forces, or both; it can result in structures that are in equilibrium or that are kinetically trapped. Here we review examples of self-assembly at surfaces that have been studied within the User program of the Molecular Foundry at Lawrence… ▽ More The self-assembly of molecules at surfaces can be caused by a range of physical mechanisms. Assembly can be driven by intermolecular forces, or molecule-surface forces, or both; it can result in structures that are in equilibrium or that are kinetically trapped. Here we review examples of self-assembly at surfaces that have been studied within the User program of the Molecular Foundry at Lawrence Berkeley National Laboratory, focusing on a physical understanding of what causes patterns seen in experiment. Some apparently disparate systems can be described in similar physical terms, indicating that simple factors -- such as the geometry and energy scale of intermolecular binding -- are key to understanding the self-assembly of those systems. △ Less

Submitted 5 December, 2014; originally announced December 2014.

Comments: Mini-review of arXiv:1109.1450, arXiv:1311.2877, and arXiv:1311.7417

Journal ref: Advanced Materials 27, 38, 5720 (2015)

Emergent rhombus tilings from molecular interactions with $M$-fold rotational symmetry

Authors: Stephen Whitelam, Isaac Tamblyn, Juan P. Garrahan, Peter H. Beton

Abstract: We show that model molecules with particular rotational symmetries can self-assemble into network structures equivalent to rhombus tilings. This assembly happens in an emergent way, in the sense that molecules spontaneously select irregular 4-fold local coordination from a larger set of possible local binding geometries. The existence of such networks can be rationalized by simple geometrical argu… ▽ More We show that model molecules with particular rotational symmetries can self-assemble into network structures equivalent to rhombus tilings. This assembly happens in an emergent way, in the sense that molecules spontaneously select irregular 4-fold local coordination from a larger set of possible local binding geometries. The existence of such networks can be rationalized by simple geometrical arguments, but the same arguments do not guarantee networks' spontaneous self-assembly. This class of structures must in certain regimes of parameter space be able to reconfigure into networks equivalent to triangular tilings. △ Less

Submitted 13 November, 2014; originally announced November 2014.

Journal ref: Phys. Rev. Lett. 114, 115702 (2015)

Modeling sequence-specific polymers using anisotropic coarse-grained sites allows quantitative comparison with experiment

Authors: Thomas K. Haxton, Ranjan V. Mannige, Ronald N. Zuckermann, Stephen Whitelam

Abstract: Certain sequences of peptoid polymers (synthetic analogs of peptides) assemble into bilayer nanosheets via a nonequilibrium assembly pathway of adsorption, compression, and collapse at an air-water interface. As with other large-scale dynamic processes in biology and materials science, understanding the details of this supramolecular assembly process requires a modeling approach that captures beha… ▽ More Certain sequences of peptoid polymers (synthetic analogs of peptides) assemble into bilayer nanosheets via a nonequilibrium assembly pathway of adsorption, compression, and collapse at an air-water interface. As with other large-scale dynamic processes in biology and materials science, understanding the details of this supramolecular assembly process requires a modeling approach that captures behavior on a wide range of length and time scales, from those on which individual sidechains fluctuate to those on which assemblies of polymers evolve. Here we demonstrate that a new coarse-grained modeling approach is accurate and computationally efficient enough to do so. Our approach uses only a minimal number of coarse-grained sites, but retains independently fluctuating orientational degrees of freedom for each site. These orientational degrees of freedom allow us to accurately parameterize both bonded and nonbonded interactions, and to generate all-atom configurations with sufficient accuracy to perform atomic scattering calculations and to interface with all-atom simulations. We have used this approach to reproduce all available experimental X-ray scattering spectra (for stacked nanosheets, and for peptoids adsorbed at air-water interfaces and in solution), in order to resolve the microscopic, real-space structures responsible for these Fourier-space features. By interfacing with all-atom simulations, we have also laid the foundations for future multiscale simulations of sequence-specific polymers that communicate in both directions across scales. △ Less

Submitted 30 September, 2014; originally announced September 2014.

The Statistical Mechanics of Dynamic Pathways to Self-assembly

Authors: Stephen Whitelam, Robert L. Jack

Abstract: We describe some of the important physical characteristics of the `pathways', i.e. dynamical processes, by which molecular, nanoscale and micron-scale self-assembly occurs. We highlight the fact that there exist features of self-assembly pathways that are common to a wide range of physical systems, even though those systems may be different in respect of their microscopic details. We summarize som… ▽ More We describe some of the important physical characteristics of the `pathways', i.e. dynamical processes, by which molecular, nanoscale and micron-scale self-assembly occurs. We highlight the fact that there exist features of self-assembly pathways that are common to a wide range of physical systems, even though those systems may be different in respect of their microscopic details. We summarize some existing theoretical descriptions of self-assembly pathways, and highlight areas -- notably, the description of self-assembly pathways that occur `far' from equilibrium -- that are likely to become increasingly important. △ Less

Submitted 21 October, 2014; v1 submitted 9 July, 2014; originally announced July 2014.

Comments: To appear in Annual Review of Physical Chemistry

Journal ref: Annual Review of Physical Chemistry, 66: 143 (2015)

Growth of equilibrium structures built from a large number of distinct component types

Authors: Lester O. Hedges, Ranjan V. Mannige, Stephen Whitelam

Abstract: We use simple analytic arguments and lattice-based computer simulations to study the growth of structures made from a large number of distinct component types. Components possess 'designed' interactions, chosen to stabilize an equilibrium target structure in which each component type has a defined spatial position, and 'undesigned' interactions that allow components to bind in a compositionally-di… ▽ More We use simple analytic arguments and lattice-based computer simulations to study the growth of structures made from a large number of distinct component types. Components possess 'designed' interactions, chosen to stabilize an equilibrium target structure in which each component type has a defined spatial position, and 'undesigned' interactions that allow components to bind in a compositionally-disordered way. We find that high-fidelity growth of the equilibrium target structure can happen in the presence of substantial attractive undesigned interactions, as long as the energy scale of the set of designed interactions is chosen appropriately. This observation may help explain why equilibrium DNA 'brick' structures self-assemble even if undesigned interactions are not suppressed [Ke et al. Science 338, 1177 (2012)]. We also find that high-fidelity growth of the target structure is most probable when designed interactions are drawn from a distribution that is as narrow as possible. We use this result to suggest how to choose complementary DNA sequences in order to maximize the fidelity of multicomponent self-assembly mediated by complementary DNA interactions. We also comment on the prospect of growing macroscopic structures in this manner △ Less

Submitted 9 May, 2014; originally announced May 2014.

Comments: 13 pages, 10 figures

Journal ref: Soft Matter, 2014,10, 6404

Competing thermodynamic and dynamic factors select molecular assemblies on a gold surface

Authors: Thomas K. Haxton, Hui Zhou, Isaac Tamblyn, Dae** Eom, Zonghai Hu, Jeffrey B. Neaton, Tony F. Heinz, Stephen Whitelam

Abstract: Controlling the self-assembly of surface-adsorbed molecules into nanostructures requires understanding physical mechanisms that act across multiple length and time scales. By combining scanning tunneling microscopy with hierarchical ab initio and statistical mechanical modeling of 1,4-substituted benzenediamine (BDA) molecules adsorbed on a gold (111) surface, we demonstrate that apparently simple… ▽ More Controlling the self-assembly of surface-adsorbed molecules into nanostructures requires understanding physical mechanisms that act across multiple length and time scales. By combining scanning tunneling microscopy with hierarchical ab initio and statistical mechanical modeling of 1,4-substituted benzenediamine (BDA) molecules adsorbed on a gold (111) surface, we demonstrate that apparently simple nanostructures are selected by a subtle competition of thermodynamics and dynamics. Of the collection of possible BDA nanostructures mechanically stabilized by hydrogen bonding, the interplay of intermolecular forces, surface modulation, and assembly dynamics select at low temperature a particular subset: low free energy oriented linear chains of monomers, and high free energy branched chains. △ Less

Submitted 28 November, 2013; originally announced November 2013.

Comments: 6 pages, 4 figures

Journal ref: Phys. Rev. Lett. 111 265701 (2013)

Common physical framework explains phase behavior and dynamics of atomic, molecular and polymeric network-formers

Authors: Stephen Whitelam, Isaac Tamblyn, Thomas K. Haxton, Maria B. Wieland, Neil R. Champness, Juan P. Garrahan, Peter H. Beton

Abstract: We show that the self-assembly of a diverse collection of building blocks can be understood within a common physical framework. These building blocks, which form periodic honeycomb networks and nonperiodic variants thereof, range in size from atoms to micron-scale polymers, and interact through mechanisms as different as hydrogen bonds and covalent forces. A combination of statistical mechanics an… ▽ More We show that the self-assembly of a diverse collection of building blocks can be understood within a common physical framework. These building blocks, which form periodic honeycomb networks and nonperiodic variants thereof, range in size from atoms to micron-scale polymers, and interact through mechanisms as different as hydrogen bonds and covalent forces. A combination of statistical mechanics and quantum mechanics shows that one can capture the physics that governs assembly of these networks by resolving only the geometry and strength of building block interactions. The resulting framework reproduces a broad range of phenomena seen experimentally, including periodic and nonperiodic networks in thermal equilibrium, and nonperiodic supercooled and glassy networks away from equilibrium. Our results show how simple `design criteria' control assembly of a wide variety of networks, and suggest that kinetic trap** can be a useful way of making functional assemblies. △ Less

Submitted 24 March, 2014; v1 submitted 12 November, 2013; originally announced November 2013.

Comments: Supplemental Info at http://nanotheory.lbl.gov/people/polygons/polygon_supp.pdf

Journal ref: Phys. Rev. X 4 , 011044 (2014)