Collective Variables for Crystallization Simulations─from Early Developments to Recent Advances

Neha, .; Tiwari, Vikas; Mondal, Soumya; Kumari, Nisha; Karmakar, Tarak

doi:10.1021/acsomega.2c06310

Cited by 17 publications

(21 citation statements)

References 102 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…the collective variables (CVs) -that in biased sampling are used to define the bias potential [147, 43? ]. We briefly discuss this point at the end of this section and for a comprehensive overview, we refer the interested reader to reviews on this topic from Giberti et al [84] and Neha et al [148]. These types of simulations are typically expensive, especially if multiple CVs are biased.…”

Section: Biased Enhanced Sampling Approachesmentioning

confidence: 99%

“…In two-step processes, a two-dimensional CV space representing the extent of the largest cluster and of the largest ordered domain in the nucleating phase have also emerged as good descriptors of the reaction coordinate [43,46,89,90], which also lend themselves to a theoretical description of two-step nucleation [45]. More recently, the application of Machine Learning methods and the data-driven identification of low-dimensional reaction coordinates for nucleation has emerged as a viable strategy to identify combinations of CVs that enable an effective, low-dimensional description of nucleation processes [43,148,180,181,182], that allows for the application of biased enhanced sampling by driving the polymorph-specific crystal nucleation. [167] The definition of effective CVs for describing and enhancing the sampling of complex nucleation processes in solution also hinges on our ability to define order parameters that can resolve well the atomic environments that are characteristic of specific crystalline structures.…”

Section: Collective Variablesmentioning

confidence: 99%

“…[167] The definition of effective CVs for describing and enhancing the sampling of complex nucleation processes in solution also hinges on our ability to define order parameters that can resolve well the atomic environments that are characteristic of specific crystalline structures. While this is routinely done for atomic or sing-particle-based crystals using bond-orientational OPs such as the Steinhardt order parameters [183,184,148,84] it remains a challenge for molecular crystals. Approaches based on the calculation of generalised pair distribution functions [185] or on the calculation of properties of the distributions of order parameters [186,187,188] demonstrate promise but still require significant improvements in terms of computational efficiency, generalisability to molecular systems with a significant degree of conformational complexity [189] and with hundreds of putative polymorphs [190,191].…”

Section: Collective Variablesmentioning

confidence: 99%

See 2 more Smart Citations

Theoretical and computational approaches to study crystal nucleation from solution

Finney

Salvalaglio

2023

Preprint

View full text Add to dashboard Cite

Nucleation is the initial step towards the formation of crystalline materials from solutions. Various factors, such as environmental conditions, additives, and external forces, can influence its outcomes and rates. Indeed, controlling this rate-determining step towards phase separation can affect the material structure and properties, and it is crucial in a range of scientific fields. In this regard, atomistic simulation methods can be exploited to gain insight into nucleation mechanisms - an aspect difficult to ascertain in experiments - and estimate nucleation rates. However, the microscopic nature of simulations affects the phase behaviour of nucleating solutions when compared to macroscopic systems. Additionally, a challenge in modelling nucleation from solution is associated with the inadequacy of standard molecular simulations to access the timescales necessary to observe crystal nucleation due to the inherent rareness of these events. In recent decades, simulation methods have emerged to circumvent length- and timescale limitations. However, which simulation method is most suitable for studying crystal nucleation from solution is not always obvious. This review summarises the recent advances in this field, providing an overview of the typical nucleation mechanisms and the suitability of different simulation methods to study them. By doing so, we aim to provide a deeper understanding of the complexities associated with modelling crystal nucleation from solution and identify areas for further research. Our review targets researchers across various scientific fields, including materials science, chemistry, physics and engineering, and will hopefully contribute to developing new strategies for understanding and controlling nucleation.

show abstract

Section: Biased Enhanced Sampling Approachesmentioning

confidence: 99%

Section: Collective Variablesmentioning

confidence: 99%

Section: Collective Variablesmentioning

confidence: 99%

See 1 more Smart Citation

Theoretical and computational approaches to study crystal nucleation from solution

Finney

Salvalaglio

2023

Preprint

View full text Add to dashboard Cite

show abstract

“…What variables can be considered informative? Although they may differ depending on the application, there are several requirements such variables should fulfill [7,27,28,30,33,45,46]:…”

Section: B Collective Variables and Target Mappingmentioning

confidence: 99%

Manifold Learning in Atomistic Simulations: A Conceptual Review

Rydzewski¹,

Chen²,

Valsson³

2023

Preprint

View full text Add to dashboard Cite

Analyzing large volumes of high-dimensional data requires dimensionality reduction: finding meaningful low-dimensional structures hidden in their high-dimensional observations. Such practice is needed in atomistic simulations of dynamical systems where even thousands of degrees of freedom are sampled. An abundance of such data makes it strenuous to gain insight into a specific physical problem. Our primary aim in this review is to focus on unsupervised machine learning methods that can be used on simulation data to find a low-dimensional manifold providing a collective and informative characterization of the studied process. Such manifolds can be used for sampling longtimescale processes and free-energy estimation. We describe methods that can work on data sets from standard and enhanced sampling atomistic simulations. Compared to recent reviews on manifold learning for atomistic simulations, we consider only methods that construct low-dimensional manifolds based on Markov transition probabilities between high-dimensional samples. We discuss these techniques from a conceptual point of view, including their underlying theoretical frameworks and possible limitations.

show abstract

“…A possible basis for the interpretable construction of high-dimensional representations in complex systems is the preservation of their time scale separation between slow and fast variables. The slow variables are intrinsically related to the kinetics of rare transitions between long-lived metastable states in configuration space, , which are essential in many processes, for instance, catalysis, crystallization, or conformational transitions. , The fast variables, however, are adiabatically slaved to the dynamics of the slow variables and correspond mainly to equilibration within metastable states. Therefore, we can consider different representations of the same system equivalent if the same time scale separation characterizes them.…”

mentioning

confidence: 99%

Selecting High-Dimensional Representations of Physical Systems by Reweighted Diffusion Maps

Rydzewski

2023

J. Phys. Chem. Lett.

View full text Add to dashboard Cite

Constructing reduced representations of high-dimensional systems is a fundamental problem in physical chemistry. Many unsupervised machine learning methods can automatically find such low-dimensional representations. However, an often overlooked problem is what high-dimensional representation should be used to describe systems before dimensionality reduction. Here, we address this issue using a recently developed method called the reweighted diffusion map [J. Chem. Theory Comput. 2022, 18, 7179−7192]. We show how high-dimensional representations can be quantitatively selected by exploring the spectral decomposition of Markov transition matrices built from data obtained from standard or enhanced sampling atomistic simulations. We demonstrate the performance of the method in several high-dimensional examples.

show abstract

Collective Variables for Crystallization Simulations─from Early Developments to Recent Advances

Cited by 17 publications

References 102 publications

Theoretical and computational approaches to study crystal nucleation from solution

Theoretical and computational approaches to study crystal nucleation from solution

Manifold Learning in Atomistic Simulations: A Conceptual Review

Selecting High-Dimensional Representations of Physical Systems by Reweighted Diffusion Maps

Contact Info

Product

Resources

About