Machine learning predictions of surface migration barriers in nucleation and non-equilibrium growth

Martynec, Thomas; Karapanagiotis, Christos; Klapp, Sabine H. L.; Kowarik, Stefan

doi:10.1038/s43246-021-00188-1

Cited by 8 publications

(3 citation statements)

References 41 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…It is important to stress that the process of diffusion on formed flat terraces, while considered an important factor for adatoms in inorganic crystal growth, [ 58 ] it is often overlooked in kinetic models of organic crystal growth (see, e.g., ref. [59]), probably because in the typical case of films of standing molecules, barriers are low [ 60 ] and the process is not the rate determining step ruling the growth mechanism.…”

Section: Resultsmentioning

confidence: 99%

Cocrystal Growth in Organic Semiconductor Thin Films: Simulation of Pentacene, Perfluoropentacene, and Their 1:1 Blend Deposited On Graphite

Lorini

Soprani

Muccioli

2023

Advcd Theory and Sims

View full text Add to dashboard Cite

The understanding of crystal formation in thin films and the precise knowledge of the relation between structure and surface diffusion are two important requirements for the efficient (nano)fabrication of organic electronic devices. Here a computational approach for simulating vapor-phase deposition is employed to obtain and investigate three types of crystalline thin films on graphite. All systems, namely pentacene, perfluoropentacene, and their 1:1 blend, which forms an alternate cocrystal, are constituted by recumbent molecules in accordance with experimental findings. The contributions of intermolecular interactions and of molecular rearrangements occurring during the deposition are analyzed to rationalize the final morphologies. Then, the generated structures are employed to evaluate the energy barriers that prevent molecular diffusion at terraces and step-edges, and to study the reorganization of the films upon high-temperature annealing. The broad agreement with experimental observations and the possibility of evaluating the potential energy surface at the molecular detail render the proposed approach a promising tool to make predictions for other systems.

show abstract

Section: Resultsmentioning

confidence: 99%

Cocrystal Growth in Organic Semiconductor Thin Films: Simulation of Pentacene, Perfluoropentacene, and Their 1:1 Blend Deposited On Graphite

Lorini

Soprani

Muccioli

2023

Advcd Theory and Sims

View full text Add to dashboard Cite

show abstract

“…Note that because of the existence of thermal fluctuation, the local composition and structure around the nucleus are varied during the nucleation process. Recent simulations show that the local compositional and structural fluctuation can play an important role in the nucleation process, [30,31] even for the stoichiometric composition. As the composition of residual liquid keeps changing during the off-stoichiometric nucleation process, its compositional fluctuation can be much more S1, Supporting Information).…”

Section: Introductionmentioning

confidence: 99%

The Crystallization of Disordered Materials under Shock Is Governed by Their Network Topology

et al. 2023

View full text Add to dashboard Cite

When the shock load is applied, materials experience incredibly high temperature and pressure conditions on picosecond timescales, usually accompanied by remarkable physical or chemical phenomena. Understanding the underlying physics that governs the kinetics of shocked materials is of great importance for both physics and materials science. Here, combining experiment and large‐scale molecular dynamics simulation, the ultrafast nanoscale crystal nucleation process in shocked soda‐lime silicate glass is investigated. By adopting topological constraints theory, this study finds that the propensity of nucleation is governed by the connectivity of the atomic network. The densification of local networks, which appears once the crystal starts to grow, results in the underconstrained shell around the crystal and prevents further crystallization. These results shed light on the nanoscale crystallization mechanism of shocked materials from the viewpoint of topological constraint theory.

show abstract

“…Pure top-down modeling can easily lead to models that are oversimplified, contain invalid assumptions, or can lead to model parameters that lose their intended physical interpretation. Experimental data is used to tune the remaining parameters of such rate models and one can make use of machine learning techniques [44,45] to optimize the parameter tuning. However, experimental data alone is often insufficient to determine the transition rates of all relevant elementary processes (free diffusion, edge diffusion, ascension, descension, dissociation, etc.…”

mentioning

confidence: 99%

Structure Formation of C₆₀ on Insulating CaF₂ Substrates: Matching Experiments with Simulations

Janke

Höltkemeier

Kühnle

et al. 2022

Adv Materials Inter

View full text Add to dashboard Cite

An extensively studied organic molecule that has become a paradigmatic choice as an adsorbate for deposition experiments is the fullerene C 60 , which has been shown to produce a vast assortment of resulting cluster morphologies on metallic [10][11][12][13][14][15] and insulating [16][17][18][19][20][21][22][23][24][25][26][27][28][29][30][31][32][33][34] substrates. However, even for the self-assembly of such a well-studied molecule, reliable structure predictions and design principles are mostly lacking. Computer simulations can be used to fill this gap and to explore the parameter space efficiently. While first principle and atomistic simulations can account for molecular details, they are prohibitively expensive to access the length and time scales needed to determine cluster morphologies. On large lengths, continuum approaches [35] including phase-field modeling [36] require the input of effective parameters (mobilities, interfacial tensions, etc.). An intermediate particle-resolved technique is the kinetic Monte Carlo (KMC) method, which has been shown to be able to achieve the necessary length and time scales to simulate the cluster growth of several deposition experiments. [37][38][39][40][41][42][43] KMC stochastically advances a number of molecules moving between discrete lattice sites through elementary transitions that depend on the interactions with neighboring molecules.A major issue one faces with KMC simulations is the modeling of the elementary transition rates of the system, which can involve a large amount of free parameters. Typically one reduces that amount by assuming an Arrhenius law for the transition rates and by applying some constraints (like bond counting approaches) on the remaining energy barrier and attempt rate parameters. Pure top-down modeling can easily lead to models that are oversimplified, contain invalid assumptions, or can lead to model parameters that lose their intended physical interpretation. Experimental data is used to tune the remaining parameters of such rate models and one can make use of machine learning techniques [44,45] to optimize the parameter tuning. However, experimental data alone is often insufficient to determine the transition rates of all relevant elementary processes (free diffusion, edge diffusion, ascension, descension, dissociation, etc.).A bottom-up approach to help with the determination of KMC rate models is the use of molecular dynamics (MD) simulations. By assuming interaction potentials for the adsorbate and substrate interactions, one can set up systems in which one can directly measure the transition rates of interest. [37] In refs. [46,47] we have recently developed such an approach and gathered an extensive amount of data from MD simulations to inform KMC rate modelsThe epitaxial growth of metallic thin films has been studied intensively, leading to computational models that can predict diverse morphologies depending on thermodynamic and kinetic growth parameters. Much less is known about thin films of organic molecules. Here kinetic Mont...

show abstract

Machine learning predictions of surface migration barriers in nucleation and non-equilibrium growth

Cited by 8 publications

References 41 publications

Cocrystal Growth in Organic Semiconductor Thin Films: Simulation of Pentacene, Perfluoropentacene, and Their 1:1 Blend Deposited On Graphite

Cocrystal Growth in Organic Semiconductor Thin Films: Simulation of Pentacene, Perfluoropentacene, and Their 1:1 Blend Deposited On Graphite

The Crystallization of Disordered Materials under Shock Is Governed by Their Network Topology

Structure Formation of C₆₀ on Insulating CaF₂ Substrates: Matching Experiments with Simulations

Contact Info

Product

Resources

About

Machine learning predictions of surface migration barriers in nucleation and non-equilibrium growth

Cited by 8 publications

References 41 publications

Cocrystal Growth in Organic Semiconductor Thin Films: Simulation of Pentacene, Perfluoropentacene, and Their 1:1 Blend Deposited On Graphite

Cocrystal Growth in Organic Semiconductor Thin Films: Simulation of Pentacene, Perfluoropentacene, and Their 1:1 Blend Deposited On Graphite

The Crystallization of Disordered Materials under Shock Is Governed by Their Network Topology

Structure Formation of C60 on Insulating CaF2 Substrates: Matching Experiments with Simulations

Contact Info

Product

Resources

About

Structure Formation of C₆₀ on Insulating CaF₂ Substrates: Matching Experiments with Simulations