Excellent! Next, you can **embed** this page using
one of several
options.

To the site owner:

**Action required!** Mendeley is changing its
API. In order to keep using Mendeley with BibBase past April
14th, you need to:

- renew the authorization for BibBase on Mendeley, and
- update the BibBase URL in your page the same way you did when you initially set up this page.

2019
(3)

Enabling QM-accurate simulation of dislocation motion in \ensuremath\gamma?Ni and \ensuremath\alpha?Fe using a hybrid multiscale approach.
Bianchini, F.; Glielmo, A.; Kermode, J. R.; and Vita, A. D.
*Physical Review Materials*, 3(4). April 2019.

Paper bibtex abstract

Paper bibtex abstract

@article{wrap116331, volume = {3}, number = {4}, month = {April}, author = {F. Bianchini and A. Glielmo and James R. Kermode and A. De Vita}, title = {Enabling QM-accurate simulation of dislocation motion in {\ensuremath{\gamma}}?Ni and {\ensuremath{\alpha}}?Fe using a hybrid multiscale approach}, publisher = {American Physical Society}, journal = {Physical Review Materials}, year = {2019}, url = {http://wrap.warwick.ac.uk/116331/}, abstract = {We present an extension of the ?learn on the fly? method to the study of the motion of dislocations in metallic systems, developed with the aim of producing information-efficient force models that can be systematically validated against reference QM calculations. Nye tensor analysis is used to dynamically track the quantum region centered at the core of a dislocation, thus enabling quantum mechanics/molecular mechanics simulations. The technique is used to study the motion of screw dislocations in Ni-Al systems, relevant to plastic deformation in Ni-based alloys, at a variety of temperature/strain conditions. These simulations reveal only a moderate spacing ( {$\sim$} 5 {\rA} ) between Shockley partial dislocations, at variance with the predictions of traditional molecular dynamics (MD) simulation using interatomic potentials, which yields a much larger spacing in the high stress regime. The discrepancy can be rationalized in terms of the elastic properties of an hcp crystal, which influence the behavior of the stacking fault region between Shockley partial dislocations. The transferability of this technique to more challenging systems is addressed, focusing on the expected accuracy of such calculations. The bcc {\ensuremath{\alpha}} ? Fe phase is a prime example, as its magnetic properties at the open surfaces make it particularly challenging for embedding-based QM/MM techniques. Our tests reveal that high accuracy can still be obtained at the core of a dislocation, albeit at a significant computational cost for fully converged results. However, we find this cost can be reduced by using a machine learning approach to progressively reduce the rate of expensive QM calculations required during the dynamical simulations, as the size of the QM database increases.} }

We present an extension of the ?learn on the fly? method to the study of the motion of dislocations in metallic systems, developed with the aim of producing information-efficient force models that can be systematically validated against reference QM calculations. Nye tensor analysis is used to dynamically track the quantum region centered at the core of a dislocation, thus enabling quantum mechanics/molecular mechanics simulations. The technique is used to study the motion of screw dislocations in Ni-Al systems, relevant to plastic deformation in Ni-based alloys, at a variety of temperature/strain conditions. These simulations reveal only a moderate spacing ( $\sim$ 5 Å ) between Shockley partial dislocations, at variance with the predictions of traditional molecular dynamics (MD) simulation using interatomic potentials, which yields a much larger spacing in the high stress regime. The discrepancy can be rationalized in terms of the elastic properties of an hcp crystal, which influence the behavior of the stacking fault region between Shockley partial dislocations. The transferability of this technique to more challenging systems is addressed, focusing on the expected accuracy of such calculations. The bcc \ensuremath\alpha ? Fe phase is a prime example, as its magnetic properties at the open surfaces make it particularly challenging for embedding-based QM/MM techniques. Our tests reveal that high accuracy can still be obtained at the core of a dislocation, albeit at a significant computational cost for fully converged results. However, we find this cost can be reduced by using a machine learning approach to progressively reduce the rate of expensive QM calculations required during the dynamical simulations, as the size of the QM database increases.

Data for Enabling QM-accurate simulation of dislocation motion in gamma-Ni and alpha-Fe using a hybrid multiscale approach.
Bianchini, F.; Glielmo, A.; Kermode, J. R.; and Vita, A. D.
March 2019.

Paper bibtex abstract

Paper bibtex abstract

@misc{wrap115558, month = {March}, title = {Data for Enabling QM-accurate simulation of dislocation motion in gamma-Ni and alpha-Fe using a hybrid multiscale approach}, author = {Federico Bianchini and Aldo Glielmo and James R. Kermode and Alessandro De Vita}, publisher = {School of Engineering, University of Warwick}, year = {2019}, keywords = {rdocheck}, url = {http://wrap.warwick.ac.uk/115558/}, abstract = {Molecular dynamics trajectories supporting calculations reported in the accompanying article.} }

Molecular dynamics trajectories supporting calculations reported in the accompanying article.

A preconditioning scheme for minimum energy path finding methods.
Makri, S.; Ortner, C.; and Kermode, J. R.
*The Journal of Chemical Physics*, 150(9). 2019.

Paper bibtex abstract

Paper bibtex abstract

@article{wrap114695, volume = {150}, number = {9}, title = {A preconditioning scheme for minimum energy path finding methods}, author = {Stela Makri and Christoph Ortner and James R. Kermode}, publisher = {American Institute of Physics}, year = {2019}, journal = {The Journal of Chemical Physics}, url = {http://wrap.warwick.ac.uk/114695/}, abstract = {Popular methods for identifying transition paths between energy minima, such as the nudged elastic band and string methods, typically do not incorporate potential energy curvature information, leading to slow relaxation to the minimum energy path for typical potential energy surfaces encountered in molecular simulation. We propose a preconditioning scheme which, combined with a new adaptive time step selection algorithm, substantially reduces the computational cost of transition path finding algorithms. We demonstrate the improved performance of our approach in a range of examples including vacancy and dislocation migration modeled with both interatomic potentials and density functional theory.} }

Popular methods for identifying transition paths between energy minima, such as the nudged elastic band and string methods, typically do not incorporate potential energy curvature information, leading to slow relaxation to the minimum energy path for typical potential energy surfaces encountered in molecular simulation. We propose a preconditioning scheme which, combined with a new adaptive time step selection algorithm, substantially reduces the computational cost of transition path finding algorithms. We demonstrate the improved performance of our approach in a range of examples including vacancy and dislocation migration modeled with both interatomic potentials and density functional theory.

2018
(6)

Machine learning a general-purpose interatomic potential for silicon.
Bartók, A. P.; Kermode, J. R.; Bernstein, N.; and Csányi, G.
*Physical Review X*, 8(4). December 2018.

Paper bibtex abstract

Paper bibtex abstract

@article{wrap112164, volume = {8}, number = {4}, month = {December}, author = {Albert P. Bart{\'o}k and James R. Kermode and Noam Bernstein and G{\'a}bor Cs{\'a}nyi}, title = {Machine learning a general-purpose interatomic potential for silicon}, publisher = {American Physical Society}, journal = {Physical Review X}, year = {2018}, url = {http://wrap.warwick.ac.uk/112164/}, abstract = {The success of first-principles electronic-structure calculation for predictive modeling in chemistry, solid-state physics, and materials science is constrained by the limitations on simulated length scales and timescales due to the computational cost and its scaling. Techniques based on machine-learning ideas for interpolating the Born-Oppenheimer potential energy surface without explicitly describing electrons have recently shown great promise, but accurately and efficiently fitting the physically relevant space of configurations remains a challenging goal. Here, we present a Gaussian approximation potential for silicon that achieves this milestone, accurately reproducing density-functional-theory reference results for a wide range of observable properties, including crystal, liquid, and amorphous bulk phases, as well as point, line, and plane defects. We demonstrate that this new potential enables calculations such as finite-temperature phase-boundary lines, self-diffusivity in the liquid, formation of the amorphous by slow quench, and dynamic brittle fracture, all of which are very expensive with a first-principles method. We show that the uncertainty quantification inherent to the Gaussian process regression framework gives a qualitative estimate of the potential?s accuracy for a given atomic configuration. The success of this model shows that it is indeed possible to create a useful machine-learning-based interatomic potential that comprehensively describes a material on the atomic scale and serves as a template for the development of such models in the future. } }

The success of first-principles electronic-structure calculation for predictive modeling in chemistry, solid-state physics, and materials science is constrained by the limitations on simulated length scales and timescales due to the computational cost and its scaling. Techniques based on machine-learning ideas for interpolating the Born-Oppenheimer potential energy surface without explicitly describing electrons have recently shown great promise, but accurately and efficiently fitting the physically relevant space of configurations remains a challenging goal. Here, we present a Gaussian approximation potential for silicon that achieves this milestone, accurately reproducing density-functional-theory reference results for a wide range of observable properties, including crystal, liquid, and amorphous bulk phases, as well as point, line, and plane defects. We demonstrate that this new potential enables calculations such as finite-temperature phase-boundary lines, self-diffusivity in the liquid, formation of the amorphous by slow quench, and dynamic brittle fracture, all of which are very expensive with a first-principles method. We show that the uncertainty quantification inherent to the Gaussian process regression framework gives a qualitative estimate of the potential?s accuracy for a given atomic configuration. The success of this model shows that it is indeed possible to create a useful machine-learning-based interatomic potential that comprehensively describes a material on the atomic scale and serves as a template for the development of such models in the future.

Multiscale simulations of critical interfacial failure in carbon nanotube-polymer composites.
Go\lębiowski, J. R.; Kermode, J. R.; Mostofi, A. A.; and Haynes, P. D.
*The Journal of Chemical Physics*, 149(22). December 2018.

Paper bibtex abstract

Paper bibtex abstract

@article{wrap111827, volume = {149}, number = {22}, month = {December}, author = {Jacek R. Go{\l}{\ke}biowski and James R. Kermode and Arash A. Mostofi and Peter D. Haynes}, title = {Multiscale simulations of critical interfacial failure in carbon nanotube-polymer composites}, publisher = {American Institute of Physics}, journal = {The Journal of Chemical Physics}, year = {2018}, keywords = {DataAyDn DataSy Datai}, url = {http://wrap.warwick.ac.uk/111827/}, abstract = {Computational investigation of interfacial failure in composite materials is challenging because it is inherently multi-scale: the bond-breaking processes that occur at the covalently bonded interface and initiate failure involve quantum mechanical phenomena, yet the mechanisms by which external stresses are transferred through the matrix occur on length and time scales far in excess of anything that can be simulated quantum mechanically. In this work, we demonstrate and validate an adaptive quantum mechanics (QM)/molecular mechanics simulation method that can be used to address these issues and apply it to study critical failure at a covalently bonded carbon nanotube (CNT)-polymer interface. In this hybrid approach, the majority of the system is simulated with a classical forcefield, while areas of particular interest are identified on-the-fly and atomic forces in those regions are updated based on QM calculations. We demonstrate that the hybrid method results are in excellent agreement with fully QM benchmark simulations and offers qualitative insights missing from classical simulations. We use the hybrid approach to show how the chemical structure at the CNT-polymer interface determines its strength, and we propose candidate chemistries to guide further experimental work in this area. } }

Computational investigation of interfacial failure in composite materials is challenging because it is inherently multi-scale: the bond-breaking processes that occur at the covalently bonded interface and initiate failure involve quantum mechanical phenomena, yet the mechanisms by which external stresses are transferred through the matrix occur on length and time scales far in excess of anything that can be simulated quantum mechanically. In this work, we demonstrate and validate an adaptive quantum mechanics (QM)/molecular mechanics simulation method that can be used to address these issues and apply it to study critical failure at a covalently bonded carbon nanotube (CNT)-polymer interface. In this hybrid approach, the majority of the system is simulated with a classical forcefield, while areas of particular interest are identified on-the-fly and atomic forces in those regions are updated based on QM calculations. We demonstrate that the hybrid method results are in excellent agreement with fully QM benchmark simulations and offers qualitative insights missing from classical simulations. We use the hybrid approach to show how the chemical structure at the CNT-polymer interface determines its strength, and we propose candidate chemistries to guide further experimental work in this area.

Accelerating multiscale modelling of fluids with on-the-fly Gaussian process regression.
Stephenson, D.; Kermode, J. R.; and Lockerby, D. A.
*Microfluidics and Nanofluidics*, 22. December 2018.

Paper bibtex abstract

Paper bibtex abstract

@article{wrap111053, volume = {22}, month = {December}, title = {Accelerating multiscale modelling of fluids with on-the-fly Gaussian process regression}, author = {David Stephenson and James R. Kermode and Duncan A. Lockerby}, publisher = {Springer}, year = {2018}, journal = {Microfluidics and Nanofluidics}, keywords = {DataAyDn DataSn}, url = {http://wrap.warwick.ac.uk/111053/}, abstract = {We present a scheme for accelerating hybrid continuum-atomistic models in multiscale fluidic systems by using Gaussian process regression as a surrogate model for computationally expensive molecular dynamics simulations. Using Gaussian process regression, we are able to accurately predict atomic-scale information purely by consideration of the macroscopic continuum-model inputs and outputs and judge on the fly whether the uncertainty of our prediction is at an acceptable level, else a new molecular simulation is performed to continually augment the database, which is never required to be complete. This provides a substantial improvement over the current generation of hybrid methods, which often require many similar atomistic simulations to be performed, discarding information after it is used once. We apply our hybrid scheme to nano-confined unsteady flow through a high-aspect-ratio converging?diverging channel, and make comparisons between the new scheme and full molecular dynamics simulations for a range of uncertainty thresholds and initial databases. For low thresholds, our hybrid solution is highly accurate{--}around that of thermal noise. As the uncertainty threshold is raised, the accuracy of our scheme decreases and the computational speed-up increases (relative to a full molecular simulation), enabling the compromise between accuracy and efficiency to be tuned. The speed-up of our hybrid solution ranges from an order of magnitude, with no initial database, to cases where an extensive initial database ensures no new MD simulations are required.} }

We present a scheme for accelerating hybrid continuum-atomistic models in multiscale fluidic systems by using Gaussian process regression as a surrogate model for computationally expensive molecular dynamics simulations. Using Gaussian process regression, we are able to accurately predict atomic-scale information purely by consideration of the macroscopic continuum-model inputs and outputs and judge on the fly whether the uncertainty of our prediction is at an acceptable level, else a new molecular simulation is performed to continually augment the database, which is never required to be complete. This provides a substantial improvement over the current generation of hybrid methods, which often require many similar atomistic simulations to be performed, discarding information after it is used once. We apply our hybrid scheme to nano-confined unsteady flow through a high-aspect-ratio converging?diverging channel, and make comparisons between the new scheme and full molecular dynamics simulations for a range of uncertainty thresholds and initial databases. For low thresholds, our hybrid solution is highly accurate–around that of thermal noise. As the uncertainty threshold is raised, the accuracy of our scheme decreases and the computational speed-up increases (relative to a full molecular simulation), enabling the compromise between accuracy and efficiency to be tuned. The speed-up of our hybrid solution ranges from an order of magnitude, with no initial database, to cases where an extensive initial database ensures no new MD simulations are required.

Accelerating multiscale modelling of fluids with on-the-fly Gaussian process regression.
Stephenson, D.; Kermode, J. R.; and Lockerby, D. A.
*Microfluidics and Nanofluidics*, 22(12). December 2018.

Paper bibtex

Paper bibtex

@article{wrap120435, volume = {22}, number = {12}, month = {December}, author = {David Stephenson and James R. Kermode and Duncan A. Lockerby}, title = {Accelerating multiscale modelling of fluids with on-the-fly Gaussian process regression}, publisher = {Springer}, journal = {Microfluidics and Nanofluidics}, year = {2018}, url = {http://wrap.warwick.ac.uk/120435/} }

Imeall : a computational framework for the calculation of the atomistic properties of grain boundaries.
Lambert, H.; Fekete, A.; Kermode, J. R.; and Vita, A. D.
*Computer Physics Communications*, 232: 256–263. November 2018.

Paper bibtex abstract

Paper bibtex abstract

@article{wrap106587, volume = {232}, month = {November}, author = {H. Lambert and Adam Fekete and James R. Kermode and A. De Vita}, title = {Imeall : a computational framework for the calculation of the atomistic properties of grain boundaries}, publisher = {Elsevier Science BV}, journal = {Computer Physics Communications}, pages = {256--263}, year = {2018}, keywords = {DataAyDn DataSy Datal}, url = {http://wrap.warwick.ac.uk/106587/}, abstract = {We describe the Imeall package for the calculation and indexing of atomistic properties of grain boundaries in materials. The package provides a structured database for the storage of atomistic structures and their associated properties, equipped with a programmable application interface to interatomic potential calculators. The database adopts a general indexing system that allows storing arbitrary grain boundary structures for any crystalline material. The usefulness of the Imeall package is demonstrated by computing, storing, and analysing relaxed grain boundary structures for a dense range of low index orientation axis symmetric tilt and twist boundaries in -iron for various interatomic potentials. The package?s capabilities are further demonstrated by carrying out automated structure generation, dislocation analysis, interstitial site detection, and impurity segregation energies across the grain boundary range. All computed atomistic properties are exposed via a web framework, providing open access to the grain boundary repository and the analytic tools suite.} }

We describe the Imeall package for the calculation and indexing of atomistic properties of grain boundaries in materials. The package provides a structured database for the storage of atomistic structures and their associated properties, equipped with a programmable application interface to interatomic potential calculators. The database adopts a general indexing system that allows storing arbitrary grain boundary structures for any crystalline material. The usefulness of the Imeall package is demonstrated by computing, storing, and analysing relaxed grain boundary structures for a dense range of low index orientation axis symmetric tilt and twist boundaries in -iron for various interatomic potentials. The package?s capabilities are further demonstrated by carrying out automated structure generation, dislocation analysis, interstitial site detection, and impurity segregation energies across the grain boundary range. All computed atomistic properties are exposed via a web framework, providing open access to the grain boundary repository and the analytic tools suite.

Understanding and mitigating hydrogen embrittlement of steels : a review of experimental, modelling and design progress from atomistic to continuum.
Barrera, O.; Bombac, D.; Chen, Y.; Daff, T. D.; Galindo-Nava, E.; Gong, P.; Haley, D.; Horton, R.; Katzarov, I.; Kermode, J. R.; Liverani, C.; Stopher, M.; and Sweeney, F.
*Journal of Materials Science*. February 2018.

Paper bibtex abstract

Paper bibtex abstract

@article{wrap98590, month = {February}, title = {Understanding and mitigating hydrogen embrittlement of steels : a review of experimental, modelling and design progress from atomistic to continuum}, author = {O. Barrera and D. Bombac and Y. Chen and T. D. Daff and E. Galindo-Nava and P. Gong and D. Haley and R. Horton and I. Katzarov and James R. Kermode and C. Liverani and M. Stopher and F. Sweeney}, publisher = {Springer}, year = {2018}, journal = {Journal of Materials Science}, url = {http://wrap.warwick.ac.uk/98590/}, abstract = {Hydrogen embrittlement is a complex phenomenon, involving several length- and timescales, that affects a large class of metals. It can significantly reduce the ductility and load-bearing capacity and cause cracking and catastrophic brittle failures at stresses below the yield stress of susceptible materials. Despite a large research effort in attempting to understand the mechanisms of failure and in developing potential mitigating solutions, hydrogen embrittlement mechanisms are still not completely understood. There are controversial opinions in the literature regarding the underlying mechanisms and related experimental evidence supporting each of these theories. The aim of this paper is to provide a detailed review up to the current state of the art on the effect of hydrogen on the degradation of metals, with a particular focus on steels. Here, we describe the effect of hydrogen in steels from the atomistic to the continuum scale by reporting theoretical evidence supported by quantum calculation and modern experimental characterisation methods, macroscopic effects that influence the mechanical properties of steels and established damaging mechanisms for the embrittlement of steels. Furthermore, we give an insight into current approaches and new mitigation strategies used to design new steels resistant to hydrogen embrittlement.} }

Hydrogen embrittlement is a complex phenomenon, involving several length- and timescales, that affects a large class of metals. It can significantly reduce the ductility and load-bearing capacity and cause cracking and catastrophic brittle failures at stresses below the yield stress of susceptible materials. Despite a large research effort in attempting to understand the mechanisms of failure and in developing potential mitigating solutions, hydrogen embrittlement mechanisms are still not completely understood. There are controversial opinions in the literature regarding the underlying mechanisms and related experimental evidence supporting each of these theories. The aim of this paper is to provide a detailed review up to the current state of the art on the effect of hydrogen on the degradation of metals, with a particular focus on steels. Here, we describe the effect of hydrogen in steels from the atomistic to the continuum scale by reporting theoretical evidence supported by quantum calculation and modern experimental characterisation methods, macroscopic effects that influence the mechanical properties of steels and established damaging mechanisms for the embrittlement of steels. Furthermore, we give an insight into current approaches and new mitigation strategies used to design new steels resistant to hydrogen embrittlement.

2017
(5)

Machine learning unifies the modeling of materials and molecules.
Bartók, A. P.; De, S.; Poelking, C.; Bernstein, N.; Kermode, J. R.; Csányi, G.; and Ceriotti, M.
*Science Advances*, 3(12). December 2017.

Paper bibtex abstract

Paper bibtex abstract

@article{wrap96346, volume = {3}, number = {12}, month = {December}, author = {Albert P. Bart{\'o}k and Sandip De and Carl Poelking and Noam Bernstein and James R. Kermode and G{\'a}bor Cs{\'a}nyi and Michele Ceriotti}, title = {Machine learning unifies the modeling of materials and molecules}, publisher = {American Association for the Advancement of Science}, journal = {Science Advances}, year = {2017}, keywords = {DataAyDn DataSy Datai Datar}, url = {http://wrap.warwick.ac.uk/96346/}, abstract = {Determining the stability of molecules and condensed phases is the cornerstone of atomistic modeling, underpinning our understanding of chemical and materials properties and transformations. We show that a machine-learning model, based on a local description of chemical environments and Bayesian statistical learning, provides a unified framework to predict atomic-scale properties. It captures the quantum mechanical effects governing the complex surface reconstructions of silicon, predicts the stability of different classes of molecules with chemical accuracy, and distinguishes active and inactive protein ligands with more than 99\% reliability. The universality and the systematic nature of our framework provide new insight into the potential energy surface of materials and molecules. } }

Determining the stability of molecules and condensed phases is the cornerstone of atomistic modeling, underpinning our understanding of chemical and materials properties and transformations. We show that a machine-learning model, based on a local description of chemical environments and Bayesian statistical learning, provides a unified framework to predict atomic-scale properties. It captures the quantum mechanical effects governing the complex surface reconstructions of silicon, predicts the stability of different classes of molecules with chemical accuracy, and distinguishes active and inactive protein ligands with more than 99% reliability. The universality and the systematic nature of our framework provide new insight into the potential energy surface of materials and molecules.

Computing energy barriers for rare events from hybrid quantum/classical simulations through the virtual work principle.
Swinburne, T.; and Kermode, J. R.
*Physical Review B (Condensed Matter and Materials Physics)*, 96. October 2017.

Paper bibtex abstract

Paper bibtex abstract

@article{wrap92999, volume = {96}, month = {October}, title = {Computing energy barriers for rare events from hybrid quantum/classical simulations through the virtual work principle}, author = {Thomas Swinburne and James R. Kermode}, publisher = {American Physical Society}, year = {2017}, journal = {Physical Review B (Condensed Matter and Materials Physics)}, url = {http://wrap.warwick.ac.uk/92999/}, abstract = {Hybrid quantum/classical techniques can flexibly couple ab initio simulations to an empirical or elastic medium to model materials systems that cannot be contained in small periodic supercells. However, due to electronic nonlocality, a total energy cannot be defined, meaning energy barriers cannot be calculated. We provide a general solution using the principle of virtual work in a modified nudged elastic band algorithm. Our method enables ab initio calculations of the kink formation energy for (100 ? edge dislocations in molybdenum and lattice trapping barriers to brittle fracture in silicon.} }

Hybrid quantum/classical techniques can flexibly couple ab initio simulations to an empirical or elastic medium to model materials systems that cannot be contained in small periodic supercells. However, due to electronic nonlocality, a total energy cannot be defined, meaning energy barriers cannot be calculated. We provide a general solution using the principle of virtual work in a modified nudged elastic band algorithm. Our method enables ab initio calculations of the kink formation energy for (100 ? edge dislocations in molybdenum and lattice trapping barriers to brittle fracture in silicon.

Data for Computing energy barriers for rare events from hybrid quantum/classical simulations through the virtual work principle.
Swinburne, T. D.; and Kermode, J. R.
September 2017.

Paper bibtex abstract

Paper bibtex abstract

@misc{wrap91747, month = {September}, title = {Data for Computing energy barriers for rare events from hybrid quantum/classical simulations through the virtual work principle}, author = {T. D. Swinburne and James R. Kermode}, publisher = {School of Engineering, University of Warwick}, year = {2017}, url = {http://wrap.warwick.ac.uk/91747/}, abstract = {Hybrid quantum/classical techniques can flexibly couple ab initio simulations to an empirical or elastic medium to model materials systems that cannot be contained in small periodic supercells. However, due to electronic nonlocality, a total energy cannot be defined, meaning energy barriers cannot be calculated. We provide a general solution using the principle of virtual work in a modified nudged elastic band algorithm. Our method enables ab initio calculations of the kink formation energy for (100 ? edge dislocations in molybdenum and lattice trapping barriers to brittle fracture in silicon.} }

Hybrid quantum/classical techniques can flexibly couple ab initio simulations to an empirical or elastic medium to model materials systems that cannot be contained in small periodic supercells. However, due to electronic nonlocality, a total energy cannot be defined, meaning energy barriers cannot be calculated. We provide a general solution using the principle of virtual work in a modified nudged elastic band algorithm. Our method enables ab initio calculations of the kink formation energy for (100 ? edge dislocations in molybdenum and lattice trapping barriers to brittle fracture in silicon.

In situ stable crack growth at the micron scale.
Sernicola, G.; Giovannini, T.; Patel, P.; Kermode, J. R.; Balint, D. S.; Britton, T. B.; and Giuliani, F.
*Nature Communications*, 8(1). July 2017.

Paper bibtex abstract

Paper bibtex abstract

@article{wrap90399, volume = {8}, number = {1}, month = {July}, author = {Giorgio Sernicola and Tommaso Giovannini and Punitbhai Patel and James R. Kermode and Daniel S. Balint and T. Ben Britton and Finn Giuliani}, title = {In situ stable crack growth at the micron scale}, publisher = {Nature Publishing Group}, journal = {Nature Communications}, year = {2017}, keywords = {Dataaydn datasy datal}, url = {http://wrap.warwick.ac.uk/90399/}, abstract = {Grain boundaries typically dominate fracture toughness, strength and slow crack growth in ceramics. To improve these properties through mechanistically informed grain boundary engineering, precise measurement of the mechanical properties of individual boundaries is essential, although it is rarely achieved due to the complexity of the task. Here we present an approach to characterize fracture energy at the lengthscale of individual grain boundaries and demonstrate this capability with measurement of the surface energy of silicon carbide single crystals. We perform experiments using an in situ scanning electron microscopy-based double cantilever beam test, thus enabling viewing and measurement of stable crack growth directly. These experiments correlate well with our density functional theory calculations of the surface energy of the same silicon carbide plane. Subsequently, we measure the fracture energy for a bi-crystal of silicon carbide, diffusion bonded with a thin glassy layer. } }

Grain boundaries typically dominate fracture toughness, strength and slow crack growth in ceramics. To improve these properties through mechanistically informed grain boundary engineering, precise measurement of the mechanical properties of individual boundaries is essential, although it is rarely achieved due to the complexity of the task. Here we present an approach to characterize fracture energy at the lengthscale of individual grain boundaries and demonstrate this capability with measurement of the surface energy of silicon carbide single crystals. We perform experiments using an in situ scanning electron microscopy-based double cantilever beam test, thus enabling viewing and measurement of stable crack growth directly. These experiments correlate well with our density functional theory calculations of the surface energy of the same silicon carbide plane. Subsequently, we measure the fracture energy for a bi-crystal of silicon carbide, diffusion bonded with a thin glassy layer.

The atomic simulation environment – a python library for working with atoms.
Larsen, A.; Mortensen, J.; Blomqvist, J.; Castelli, I.; Christensen, R.; Dulak, M.; Friis, J.; Groves, M.; Hammer, B.; Hargus, C.; Hermes, E.; Jennings, P. A.; Jensen, P.; Kermode, J. R.; Kitchin, J.; Kolsbjerg, E.; Kubal, J.; Kaasbjerg, K.; Lysgaard, S.; Maronsson, J.; Maxson, T.; Olsen, T.; Pastewka, L.; Peterson, A.; Rostgaard, C.; Schi\otz, J.; Schütt, O.; Strange, M.; Thygesen, K.; Vegge, T.; Vilhelmsen, L.; Walter, M.; Zeng, Z.; and Jacobsen, K. W.
*Journal of Physics: Condensed Matter*. 2017.

Paper bibtex abstract

Paper bibtex abstract

@article{wrap87141, title = {The atomic simulation environment {--} a python library for working with atoms}, author = {Ask Larsen and Jens Mortensen and Jakob Blomqvist and Ivano Castelli and Rune Christensen and Marcin Dulak and Jesper Friis and Michael Groves and Bjork Hammer and Cory Hargus and Eric Hermes and Paul A. Jennings and Peter Jensen and James R. Kermode and John Kitchin and Esben Kolsbjerg and Joseph Kubal and Kristen Kaasbjerg and Steen Lysgaard and Jon Maronsson and Tristan Maxson and Thomas Olsen and Lars Pastewka and Andrew Peterson and Carsten Rostgaard and Jakob Schi{\o}tz and Ole Sch{\"u}tt and Mikkel Strange and Kristian Thygesen and Tejs Vegge and Lasse Vilhelmsen and Michael Walter and Zhenhua Zeng and Karsten Wedel Jacobsen}, publisher = {Institute of Physics Publishing Ltd.}, year = {2017}, journal = {Journal of Physics: Condensed Matter}, keywords = {DataAyDn DataSx}, url = {http://wrap.warwick.ac.uk/87141/}, abstract = {The Atomic Simulation Environment (ASE) is a software package written in the Python programming language with the aim of setting up, steering, and analyzing atomistic simula- tions. In ASE, tasks are fully scripted in Python. The powerful syntax of Python combined with the NumPy array library make it possible to perform very complex simulation tasks. For example, a sequence of calculations may be performed with the use of a simple "for-loop" construction. Calculations of energy, forces, stresses and other quantities are performed through interfaces to many external electronic structure codes or force fields using a uniform interface. On top of this calculator interface, ASE provides modules for performing many standard simulation tasks such as structure optimization, molecular dynamics, handling of constraints and performing nudged elastic band calculations. } }

The Atomic Simulation Environment (ASE) is a software package written in the Python programming language with the aim of setting up, steering, and analyzing atomistic simula- tions. In ASE, tasks are fully scripted in Python. The powerful syntax of Python combined with the NumPy array library make it possible to perform very complex simulation tasks. For example, a sequence of calculations may be performed with the use of a simple "for-loop" construction. Calculations of energy, forces, stresses and other quantities are performed through interfaces to many external electronic structure codes or force fields using a uniform interface. On top of this calculator interface, ASE provides modules for performing many standard simulation tasks such as structure optimization, molecular dynamics, handling of constraints and performing nudged elastic band calculations.

2016
(6)

A framework for machine-learning-augmented multiscale atomistic simulations on parallel supercomputers.
Caccin, M.; Li, Z.; Kermode, J. R.; and Vita, A. D.
*International Journal of Quantum Chemistry*, 115(16): 1129–1139. August 2016.

Paper bibtex abstract

Paper bibtex abstract

@article{wrap68012, volume = {115}, number = {16}, month = {August}, author = {Marco Caccin and Zhenwei Li and James R. Kermode and Alessandro De Vita}, title = {A framework for machine-learning-augmented multiscale atomistic simulations on parallel supercomputers}, publisher = {John Wiley \& Sons}, year = {2016}, journal = {International Journal of Quantum Chemistry}, pages = {1129--1139}, keywords = {machine learning; quantum mechanics/molecular mechanics; HPC; fracture; partitioning}, url = {http://wrap.warwick.ac.uk/68012/}, abstract = {Recent advances in quantum mechanical (QM)-based molecular dynamics (MD) simulations have used machine-learning (ML) to predict, rather than recalculate, QM-accurate forces in atomic configurations sufficiently similar to previously encountered ones. Here, we discuss how ML approaches can be deployed within large-scale QM/MM materials simulations on massively parallel supercomputers, making QM zones of {$\gtrsim$}1000 atoms routinely attainable. We argue that the ML approach allows computational effort to be concentrated on the most chemically active subregions of the QM zone, significantly improving the overall efficiency of the simulation. We thus propose a novel method to partition large QM regions into multiple subregions, which can be computed in parallel to achieve optimal scaling. Then we review a recently proposed QM/ML MD scheme (Z. Li, J.R. Kermode, A. De Vita Phys. Rev. Lett., 2015, 114, 096405), discussing how this could be efficiently combined with QM-zone partitioning.} }

Recent advances in quantum mechanical (QM)-based molecular dynamics (MD) simulations have used machine-learning (ML) to predict, rather than recalculate, QM-accurate forces in atomic configurations sufficiently similar to previously encountered ones. Here, we discuss how ML approaches can be deployed within large-scale QM/MM materials simulations on massively parallel supercomputers, making QM zones of $\gtrsim$1000 atoms routinely attainable. We argue that the ML approach allows computational effort to be concentrated on the most chemically active subregions of the QM zone, significantly improving the overall efficiency of the simulation. We thus propose a novel method to partition large QM regions into multiple subregions, which can be computed in parallel to achieve optimal scaling. Then we review a recently proposed QM/ML MD scheme (Z. Li, J.R. Kermode, A. De Vita Phys. Rev. Lett., 2015, 114, 096405), discussing how this could be efficiently combined with QM-zone partitioning.

Data for A universal preconditioner for simulating condensed phase materials.
Packwood, D.; Kermode, J. R.; Mones, L.; Bernstein, N.; Woolley, J.; Gould, N.; Ortner, C.; and Csanyi, G.
April 2016.

Paper bibtex abstract

Paper bibtex abstract

@misc{wrap78579, month = {April}, title = {Data for A universal preconditioner for simulating condensed phase materials}, author = {David Packwood and James R. Kermode and Letif Mones and Noam Bernstein and John Woolley and Nicholas Gould and Christoph Ortner and Gabor Csanyi}, publisher = {School of Engineering, University of Warwick}, year = {2016}, url = {http://wrap.warwick.ac.uk/78579/}, abstract = {We introduce a universal sparse preconditioner that accelerates geometry optimisation and saddle point search tasks that are common in the atomic scale simulation of materials. Our preconditioner is based on the neighbourhood structure and we demonstrate the gain in computational efficiency in a wide range of materials that include metals, insulators and molecular solids. The simple structure of the preconditioner means that the gains can be realised in practice not only when using expensive electronic structure models but also for fast empirical potentials. Even for relatively small systems of a few hundred atoms, we observe speedups of a factor of two or more, and the gain grows with system size. An open source Python implementation within the Atomic Simulation Environment is available, offering interfaces to a wide range of atomistic codes.} }

We introduce a universal sparse preconditioner that accelerates geometry optimisation and saddle point search tasks that are common in the atomic scale simulation of materials. Our preconditioner is based on the neighbourhood structure and we demonstrate the gain in computational efficiency in a wide range of materials that include metals, insulators and molecular solids. The simple structure of the preconditioner means that the gains can be realised in practice not only when using expensive electronic structure models but also for fast empirical potentials. Even for relatively small systems of a few hundred atoms, we observe speedups of a factor of two or more, and the gain grows with system size. An open source Python implementation within the Atomic Simulation Environment is available, offering interfaces to a wide range of atomistic codes.

Modelling defects in Ni-Al with EAM and DFT calculations.
Bianchini, F.; Kermode, J. R.; and Vita, A. D.
*Modelling and Simulation in Materials Science and Engineering*, 24(4). April 2016.

Paper bibtex abstract

Paper bibtex abstract

@article{wrap78334, volume = {24}, number = {4}, month = {April}, author = {Federico Bianchini and James R. Kermode and Alessandro De Vita}, title = {Modelling defects in Ni-Al with EAM and DFT calculations}, publisher = {Institute of Physics Publishing Ltd.}, journal = {Modelling and Simulation in Materials Science and Engineering}, year = {2016}, url = {http://wrap.warwick.ac.uk/78334/}, abstract = {We present detailed comparisons between the results of embedded atom model (EAM) and density functional theory (DFT) calculations on defected Ni alloy systems. We find that the EAM interatomic potentials reproduce low-temperature structural properties in both the {\ensuremath{\gamma}} and \$\{\{{$\backslash$}gamma\}{\^{ }}\{{$\backslash$}prime\}\}\$ phases, and yield accurate atomic forces in bulk-like configurations even at temperatures as high as {\texttt{\char126}}1200 K. However, they fail to describe more complex chemical bonding, in configurations including defects such as vacancies or dislocations, for which we observe significant deviations between the EAM and DFT forces, suggesting that derived properties such as (free) energy barriers to vacancy migration and dislocation glide may also be inaccurate. Testing against full DFT calculations further reveals that these deviations have a local character, and are typically severe only up to the first or second neighbours of the defect. This suggests that a QM/MM approach can be used to accurately reproduce QM observables, fully exploiting the EAM potential efficiency in the MM zone. This approach could be easily extended to ternary systems for which developing a reliable and fully transferable EAM parameterisation would be extremely challenging e.g. Ni alloy model systems with a W or Re-containing QM zone. } }

We present detailed comparisons between the results of embedded atom model (EAM) and density functional theory (DFT) calculations on defected Ni alloy systems. We find that the EAM interatomic potentials reproduce low-temperature structural properties in both the \ensuremath\gamma and \$\{\{{$\backslash$}gamma\}{\^{ }}\{{$\backslash$}prime\}\}\$ phases, and yield accurate atomic forces in bulk-like configurations even at temperatures as high as \textttḩar1261200 K. However, they fail to describe more complex chemical bonding, in configurations including defects such as vacancies or dislocations, for which we observe significant deviations between the EAM and DFT forces, suggesting that derived properties such as (free) energy barriers to vacancy migration and dislocation glide may also be inaccurate. Testing against full DFT calculations further reveals that these deviations have a local character, and are typically severe only up to the first or second neighbours of the defect. This suggests that a QM/MM approach can be used to accurately reproduce QM observables, fully exploiting the EAM potential efficiency in the MM zone. This approach could be easily extended to ternary systems for which developing a reliable and fully transferable EAM parameterisation would be extremely challenging e.g. Ni alloy model systems with a W or Re-containing QM zone.

Development of an exchange?correlation functional with uncertainty quantification capabilities for density functional theory.
Aldegunde, M.; Kermode, J. R.; and Zabaras, N.
*Journal of Computational Physics*, 311: 173–195. April 2016.

Paper bibtex abstract

Paper bibtex abstract

@article{wrap77127, volume = {311}, month = {April}, author = {Manuel Aldegunde and James R. Kermode and Nicholas Zabaras }, title = {Development of an exchange?correlation functional with uncertainty quantification capabilities for density functional theory}, publisher = {Academic Press Inc. Elsevier Science}, journal = {Journal of Computational Physics}, pages = {173--195}, year = {2016}, url = {http://wrap.warwick.ac.uk/77127/}, abstract = {This paper presents the development of a new exchange?correlation functional from the point of view of machine learning. Using atomization energies of solids and small molecules, we train a linear model for the exchange enhancement factor using a Bayesian approach which allows for the quantification of uncertainties in the predictions. A relevance vector machine is used to automatically select the most relevant terms of the model. We then test this model on atomization energies and also on bulk properties. The average model provides a mean absolute error of only 0.116 eV for the test points of the G2/97 set but a larger 0.314 eV for the test solids. In terms of bulk properties, the prediction for transition metals and monovalent semiconductors has a very low test error. However, as expected, predictions for types of materials not represented in the training set such as ionic solids show much larger errors.} }

This paper presents the development of a new exchange?correlation functional from the point of view of machine learning. Using atomization energies of solids and small molecules, we train a linear model for the exchange enhancement factor using a Bayesian approach which allows for the quantification of uncertainties in the predictions. A relevance vector machine is used to automatically select the most relevant terms of the model. We then test this model on atomization energies and also on bulk properties. The average model provides a mean absolute error of only 0.116 eV for the test points of the G2/97 set but a larger 0.314 eV for the test solids. In terms of bulk properties, the prediction for transition metals and monovalent semiconductors has a very low test error. However, as expected, predictions for types of materials not represented in the training set such as ionic solids show much larger errors.

Data for Modelling defects in Ni-Al with EAM and DFT calculations .
Bianchini, F.; Kermode, J. R.; and Vita, A. D.
April 2016.

Paper bibtex abstract

Paper bibtex abstract

@misc{wrap78256, month = {April}, title = {Data for Modelling defects in Ni-Al with EAM and DFT calculations }, author = {Federico Bianchini and James R. Kermode and Alessandro De Vita}, publisher = {School of Engineering, University of Warwick}, year = {2016}, url = {http://wrap.warwick.ac.uk/78256/}, abstract = {We present detailed comparisons between the results of embedded atom model (EAM) and density functional theory (DFT) calculations on defected Ni alloy systems. We find that the EAM interatomic potentials reproduce low-temperature structural properties in both the {\ensuremath{\gamma}} and \$\{\{{$\backslash$}gamma\}{\^{ }}\{{$\backslash$}prime\}\}\$ phases, and yield accurate atomic forces in bulk-like configurations even at temperatures as high as {\texttt{\char126}}1200 K. However, they fail to describe more complex chemical bonding, in configurations including defects such as vacancies or dislocations, for which we observe significant deviations between the EAM and DFT forces, suggesting that derived properties such as (free) energy barriers to vacancy migration and dislocation glide may also be inaccurate. Testing against full DFT calculations further reveals that these deviations have a local character, and are typically severe only up to the first or second neighbours of the defect. This suggests that a QM/MM approach can be used to accurately reproduce QM observables, fully exploiting the EAM potential efficiency in the MM zone. This approach could be easily extended to ternary systems for which developing a reliable and fully transferable EAM parameterisation would be extremely challenging e.g. Ni alloy model systems with a W or Re-containing QM zone.} }

We present detailed comparisons between the results of embedded atom model (EAM) and density functional theory (DFT) calculations on defected Ni alloy systems. We find that the EAM interatomic potentials reproduce low-temperature structural properties in both the \ensuremath\gamma and \$\{\{{$\backslash$}gamma\}{\^{ }}\{{$\backslash$}prime\}\}\$ phases, and yield accurate atomic forces in bulk-like configurations even at temperatures as high as \textttḩar1261200 K. However, they fail to describe more complex chemical bonding, in configurations including defects such as vacancies or dislocations, for which we observe significant deviations between the EAM and DFT forces, suggesting that derived properties such as (free) energy barriers to vacancy migration and dislocation glide may also be inaccurate. Testing against full DFT calculations further reveals that these deviations have a local character, and are typically severe only up to the first or second neighbours of the defect. This suggests that a QM/MM approach can be used to accurately reproduce QM observables, fully exploiting the EAM potential efficiency in the MM zone. This approach could be easily extended to ternary systems for which developing a reliable and fully transferable EAM parameterisation would be extremely challenging e.g. Ni alloy model systems with a W or Re-containing QM zone.

A universal preconditioner for simulating condensed phase materials.
Packwood, D.; Kermode, J. R.; Mones, L.; Bernstein, N.; Woolley, J.; Gould, N.; Ortner, C.; and Csanyi, G.
*Journal of Chemical Physics*, 144(16). April 2016.

Paper bibtex abstract

Paper bibtex abstract

@article{wrap78515, volume = {144}, number = {16}, month = {April}, author = {David Packwood and James R. Kermode and Letif Mones and Noam Bernstein and John Woolley and Nicholas Gould and Christoph Ortner and Gabor Csanyi}, title = {A universal preconditioner for simulating condensed phase materials}, publisher = {American Institute of Physics}, journal = {Journal of Chemical Physics}, year = {2016}, url = {http://wrap.warwick.ac.uk/78515/}, abstract = {We introduce a universal sparse preconditioner that accelerates geometry optimisation and saddle point search tasks that are common in the atomic scale simulation of materials. Our preconditioner is based on the neighbourhood structure and we demonstrate the gain in computational efficiency in a wide range of materials that include metals, insulators and molecular solids. The simple structure of the preconditioner means that the gains can be realised in practice not only when using expensive electronic structure models but also for fast empirical potentials. Even for relatively small systems of a few hundred atoms, we observe speedups of a factor of two or more, and the gain grows with system size. An open source Python implementation within the Atomic Simulation Environment is available, offering interfaces to a wide range of atomistic codes. } }

We introduce a universal sparse preconditioner that accelerates geometry optimisation and saddle point search tasks that are common in the atomic scale simulation of materials. Our preconditioner is based on the neighbourhood structure and we demonstrate the gain in computational efficiency in a wide range of materials that include metals, insulators and molecular solids. The simple structure of the preconditioner means that the gains can be realised in practice not only when using expensive electronic structure models but also for fast empirical potentials. Even for relatively small systems of a few hundred atoms, we observe speedups of a factor of two or more, and the gain grows with system size. An open source Python implementation within the Atomic Simulation Environment is available, offering interfaces to a wide range of atomistic codes.

2015
(5)

Low speed crack propagation via kink formation and advance on the silicon (110) cleavage plane.
Kermode, J. R.; Gleizer, A.; Kovel, G.; Pastewka, L.; Csanyi, G.; Sherman, D.; and Vita, A. D.
*Physical Review Letters*, 115(13): 1–5. September 2015.

Paper bibtex abstract

Paper bibtex abstract

@article{wrap72523, volume = {115}, number = {13}, month = {September}, author = {James R. Kermode and Anna Gleizer and Guy Kovel and Lars Pastewka and Gabor Csanyi and Dov Sherman and Alessandro De Vita}, title = {Low speed crack propagation via kink formation and advance on the silicon (110) cleavage plane}, publisher = {American Physical Society}, year = {2015}, journal = {Physical Review Letters}, pages = {1--5}, url = {http://wrap.warwick.ac.uk/72523/}, abstract = {We present density functional theory based atomistic calculations predicting that slow fracturing of silicon is possible at any chosen crack propagation speed under suitable temperature and load conditions. We also present experiments demonstrating fracture propagation on the Si(110) cleavage plane in the {\texttt{\char126}}100 m/s speed range, consistent with our predictions. These results suggest that many other brittle crystals could be broken arbitrarily slowly in controlled experiments.} }

We present density functional theory based atomistic calculations predicting that slow fracturing of silicon is possible at any chosen crack propagation speed under suitable temperature and load conditions. We also present experiments demonstrating fracture propagation on the Si(110) cleavage plane in the \textttḩar126100 m/s speed range, consistent with our predictions. These results suggest that many other brittle crystals could be broken arbitrarily slowly in controlled experiments.

Molecular dynamics with on-the-fly machine learning of quantum-mechanical forces.
Li, Z.; Kermode, J. R.; and Vita, A. D.
*Physical Review Letters*, Volume 114. March 2015.

Paper bibtex abstract

Paper bibtex abstract

@article{wrap66611, volume = {Volume 114}, month = {March}, title = {Molecular dynamics with on-the-fly machine learning of quantum-mechanical forces}, author = {Zhenwei Li and James R. Kermode and Alessandro De Vita}, publisher = {American Physical Society}, year = {2015}, journal = {Physical Review Letters}, url = {http://wrap.warwick.ac.uk/66611/}, abstract = {We present a molecular dynamics scheme which combines first-principles and machine-learning (ML) techniques in a single information-efficient approach. Forces on atoms are either predicted by Bayesian inference or, if necessary, computed by on-the-fly quantum-mechanical (QM) calculations and added to a growing ML database, whose completeness is, thus, never required. As a result, the scheme is accurate and general, while progressively fewer QM calls are needed when a new chemical process is encountered for the second and subsequent times, as demonstrated by tests on crystalline and molten silicon.} }

We present a molecular dynamics scheme which combines first-principles and machine-learning (ML) techniques in a single information-efficient approach. Forces on atoms are either predicted by Bayesian inference or, if necessary, computed by on-the-fly quantum-mechanical (QM) calculations and added to a growing ML database, whose completeness is, thus, never required. As a result, the scheme is accurate and general, while progressively fewer QM calls are needed when a new chemical process is encountered for the second and subsequent times, as demonstrated by tests on crystalline and molten silicon.

Accuracy of buffered-force QM/MM simulations of silica.
Peguiron, A.; Ciacchi, L. C.; Vita, A. D.; Kermode, J. R.; and Moras, G.
*Journal of chemical physics*, Volume 142. February 2015.

Paper bibtex abstract

Paper bibtex abstract

@article{wrap66305, volume = {Volume 142}, month = {February}, title = {Accuracy of buffered-force QM/MM simulations of silica}, author = {Anke Peguiron and Lucio Colombi Ciacchi and Alessandro De Vita and James R. Kermode and Gianpietro Moras}, publisher = {American Institute of Physics}, year = {2015}, journal = {Journal of chemical physics}, url = {http://wrap.warwick.ac.uk/66305/}, abstract = {We report comparisons between energy-based quantum mechanics/molecular mechanics (QM/MM) and buffered force-based QM/MM simulations in silica. Local quantities{--}such as density of states, charges, forces, and geometries{--}calculated with both QM/MM approaches are compared to the results of full QM simulations. We find the length scale over which forces computed using a finite QM region converge to reference values obtained in full quantum-mechanical calculations is {$\sim$}10 {\rA} rather than the {$\sim$}5 {\rA} previously reported for covalent materials such as silicon. Electrostatic embedding of the QM region in the surrounding classical point charges gives only a minor contribution to the force convergence. While the energy-based approach provides accurate results in geometry optimizations of point defects, we find that the removal of large force errors at the QM/MM boundary provided by the buffered force-based scheme is necessary for accurate constrained geometry optimizations where Si?O bonds are elongated and for finite-temperature molecular dynamics simulations of crack propagation. Moreover, the buffered approach allows for more flexibility, since special-purpose QM/MM coupling terms that link QM and MM atoms are not required and the region that is treated at the QM level can be adaptively redefined during the course of a dynamical simulation.} }

We report comparisons between energy-based quantum mechanics/molecular mechanics (QM/MM) and buffered force-based QM/MM simulations in silica. Local quantities–such as density of states, charges, forces, and geometries–calculated with both QM/MM approaches are compared to the results of full QM simulations. We find the length scale over which forces computed using a finite QM region converge to reference values obtained in full quantum-mechanical calculations is $\sim$10 Å rather than the $\sim$5 Å previously reported for covalent materials such as silicon. Electrostatic embedding of the QM region in the surrounding classical point charges gives only a minor contribution to the force convergence. While the energy-based approach provides accurate results in geometry optimizations of point defects, we find that the removal of large force errors at the QM/MM boundary provided by the buffered force-based scheme is necessary for accurate constrained geometry optimizations where Si?O bonds are elongated and for finite-temperature molecular dynamics simulations of crack propagation. Moreover, the buffered approach allows for more flexibility, since special-purpose QM/MM coupling terms that link QM and MM atoms are not required and the region that is treated at the QM level can be adaptively redefined during the course of a dynamical simulation.

Atomistic aspects of fracture.
Bitzek, E.; Kermode, J. R.; and Gumbsch, P.
*International Journal of Fracture*, 191(1): 13–30. February 2015.

Paper bibtex abstract

Paper bibtex abstract

@article{wrap66441, volume = {191}, number = {1}, month = {February}, author = {Erik Bitzek and James R. Kermode and Peter Gumbsch}, title = {Atomistic aspects of fracture}, publisher = {Springer}, year = {2015}, journal = {International Journal of Fracture}, pages = {13--30}, url = {http://wrap.warwick.ac.uk/66441/}, abstract = {Any fracture process ultimately involves the rupture of atomic bonds. Processes at the atomic scale therefore critically influence the toughness and overall fracture behavior of materials. Atomistic simulation methods including large-scale molecular dynamics simulations with classical potentials, density functional theory calculations and advanced concurrent multiscale methods have led to new insights e.g. on the role of bond trapping, dynamic effects, crack- microstructure interactions and chemical aspects on the fracture toughness and crack propagation patterns in metals and ceramics. This review focuses on atomistic aspects of fracture in crystalline materials where significant advances have been achieved over the last ten years and provides an outlook on future perspectives for atomistic modelling of fracture.} }

Any fracture process ultimately involves the rupture of atomic bonds. Processes at the atomic scale therefore critically influence the toughness and overall fracture behavior of materials. Atomistic simulation methods including large-scale molecular dynamics simulations with classical potentials, density functional theory calculations and advanced concurrent multiscale methods have led to new insights e.g. on the role of bond trapping, dynamic effects, crack- microstructure interactions and chemical aspects on the fracture toughness and crack propagation patterns in metals and ceramics. This review focuses on atomistic aspects of fracture in crystalline materials where significant advances have been achieved over the last ten years and provides an outlook on future perspectives for atomistic modelling of fracture.

A Python interface to CASTEP.
Corbett, G.; Kermode, J. R.; Jochym, D. B.; and Refson, K.
2015.

Paper bibtex abstract

Paper bibtex abstract

@misc{wrap67696, title = {A Python interface to CASTEP}, author = {Greg Corbett and James R. Kermode and Dominik Bogdan Jochym and Keith Refson}, address = {Rutherford Appleton Laboratory Technical Reports}, publisher = {Rutherford Appleton Laboratory}, year = {2015}, keywords = {fortran, python, castep}, url = {http://wrap.warwick.ac.uk/67696/}, abstract = {This report documents a successful pilot project and feasibility study for adding a Python interface to the CASTEP first principles materials modelling code. Such an interface will allow the growing Python community within the scientific computing field access to CASTEP functionality, without the requirement of learning Fortran. To achieve this, changes have been made to the CASTEP source code to allow: - Serially re-entrant calling of a major task routine, specifically electronic\_minimisation(). - Automated generation of a Python interface. The reasoning behind these changes has been documented and coding practices that may hinder a full move to serial re-entrancy in future have been noted. To demonstrate the success of the project, top-level task control logic has been written in Python -- using the Fortran 2003 computational core to perform multiple calls to electronic\_minimisation().} }

This report documents a successful pilot project and feasibility study for adding a Python interface to the CASTEP first principles materials modelling code. Such an interface will allow the growing Python community within the scientific computing field access to CASTEP functionality, without the requirement of learning Fortran. To achieve this, changes have been made to the CASTEP source code to allow: - Serially re-entrant calling of a major task routine, specifically electronic_minimisation(). - Automated generation of a Python interface. The reasoning behind these changes has been documented and coding practices that may hinder a full move to serial re-entrancy in future have been noted. To demonstrate the success of the project, top-level task control logic has been written in Python – using the Fortran 2003 computational core to perform multiple calls to electronic_minimisation().

2014
(3)

Validity of linear elasticity in the crack-tip region of ideal brittle solids.
Singh, G.; Kermode, J. R.; Vita, A. D.; and Zimmerman, R. W.
*International Journal of Fracture*, Volume 189(Number 1): 103–110. September 2014.

Paper bibtex abstract

Paper bibtex abstract

@article{wrap64442, volume = {Volume 189}, number = {Number 1}, month = {September}, author = {Gaurav Singh and James R. Kermode and Alessandro De Vita and Robert W. Zimmerman}, title = {Validity of linear elasticity in the crack-tip region of ideal brittle solids}, publisher = {Springer}, year = {2014}, journal = {International Journal of Fracture}, pages = {103--110}, keywords = {brittle,crack,elasticity,singularity}, url = {http://wrap.warwick.ac.uk/64442/}, abstract = {It is a well known that, according to classical elasticity, the stress in the crack-tip region is singular, which has led to a debate over the validity of linear elasticity in this region. In this work, comparisons of finite and small strain theories have been made in the crack-tip region of a brittle crystal to comment on the validity of linear elasticity in the crack tip region. We find that linear elasticity is capable of accurately defining the state of stress very close ({$\sim$}1nm) to a static crack tip.} }

It is a well known that, according to classical elasticity, the stress in the crack-tip region is singular, which has led to a debate over the validity of linear elasticity in this region. In this work, comparisons of finite and small strain theories have been made in the crack-tip region of a brittle crystal to comment on the validity of linear elasticity in the crack tip region. We find that linear elasticity is capable of accurately defining the state of stress very close ($\sim$1nm) to a static crack tip.

Multiscale modelling of materials chemomechanics : brittle fracture of oxides and semiconductors.
Kermode, J. R.; Peralta, G.; Li, Z.; and Vita, A. D.
*Procedia Materials Science*, Volume 3: 1681–1686. June 2014.

Paper bibtex abstract

Paper bibtex abstract

@article{wrap64443, volume = {Volume 3}, month = {June}, author = {James R. Kermode and Giovanni Peralta and Zhenwei Li and Alessandro De Vita}, booktitle = {20th European Conference on Fracture (ECF20)}, title = {Multiscale modelling of materials chemomechanics : brittle fracture of oxides and semiconductors}, publisher = {Elsevier B.V.}, year = {2014}, journal = {Procedia Materials Science}, pages = {1681--1686}, keywords = {atomistic,hybrid,mm simulation,multiscale,qm}, url = {http://wrap.warwick.ac.uk/64443/}, abstract = {Fracture is one of the most challenging ?multi-scale? problems to model: since crack propagation is driven by the concentration of a long-range stress field at an atomically sharp crack tip, an accurate description of the chemical processes occurring in the small crack tip region is therefore essential, as is the inclusion of a much larger region in the model systems. Both these requirements can be met by combining a quantum mechanical description of the crack tip with a classical atomistic model that captures the long-range elastic behaviour of the surrounding crystal matrix. Examples of the application of these techniques to fracture problems include: low-speed dynamical fracture instabilities in silicon; interactions between moving cracks and material defects such as dislocations or impurities; the crossover from thermally activated to catastrophic fracture; very slow crack propagation via kink formation and migration; and chemically activated fracture, where cracks advance under the concerted action of stress and corrosion by chemical species such as oxygen or water.} }

Fracture is one of the most challenging ?multi-scale? problems to model: since crack propagation is driven by the concentration of a long-range stress field at an atomically sharp crack tip, an accurate description of the chemical processes occurring in the small crack tip region is therefore essential, as is the inclusion of a much larger region in the model systems. Both these requirements can be met by combining a quantum mechanical description of the crack tip with a classical atomistic model that captures the long-range elastic behaviour of the surrounding crystal matrix. Examples of the application of these techniques to fracture problems include: low-speed dynamical fracture instabilities in silicon; interactions between moving cracks and material defects such as dislocations or impurities; the crossover from thermally activated to catastrophic fracture; very slow crack propagation via kink formation and migration; and chemically activated fracture, where cracks advance under the concerted action of stress and corrosion by chemical species such as oxygen or water.

Dissociative chemisorption of O2 inducing stress corrosion cracking in silicon crystals.
Gleizer, A.; Peralta, G.; Kermode, J. R.; Vita, A. D.; and Sherman, D.
*Physical Review Letters*, Volume 112(Number 11). March 2014.

Paper bibtex abstract

Paper bibtex abstract

@article{wrap64444, volume = {Volume 112}, number = {Number 11}, month = {March}, author = {Anna Gleizer and Giovanni Peralta and James R. Kermode and Alessandro De Vita and Dov Sherman}, title = {Dissociative chemisorption of O2 inducing stress corrosion cracking in silicon crystals}, publisher = {American Physical Society}, journal = {Physical Review Letters}, year = {2014}, url = {http://wrap.warwick.ac.uk/64444/}, abstract = {Fracture experiments to evaluate the cleavage energy of the (110)[1-10] and (111)[11-2] cleavage systems in silicon at room temperature and humidity give 2.7 +/- 0.3 and 2.2 +/-0.2 J/m{\^{ }}2, respectively, lower than any previous measurement and inconsistent with density functional theory (DFT) surface energy calculations of 3.46 and 2.88 J/m{\^{ }}2. However, in an inert gas environment, we measure values of 3.5 +/- 0.2 and 2.9 +/- 0.2 J/m{\^{ }}2, consistent with DFT, that suggest a previously undetected stress corrosion cracking scenario for Si crack initiation in room conditions. This is fully confirmed by hybrid quantum-mechanics?molecular-mechanics calculations.} }

Fracture experiments to evaluate the cleavage energy of the (110)[1-10] and (111)[11-2] cleavage systems in silicon at room temperature and humidity give 2.7 +/- 0.3 and 2.2 +/-0.2 J/m\^ 2, respectively, lower than any previous measurement and inconsistent with density functional theory (DFT) surface energy calculations of 3.46 and 2.88 J/m\^ 2. However, in an inert gas environment, we measure values of 3.5 +/- 0.2 and 2.9 +/- 0.2 J/m\^ 2, consistent with DFT, that suggest a previously undetected stress corrosion cracking scenario for Si crack initiation in room conditions. This is fully confirmed by hybrid quantum-mechanics?molecular-mechanics calculations.

2013
(1)

Macroscopic scattering of cracks initiated at single impurity atoms.
Kermode, J. R.; Ben-Bashat, L.; Atrash, F.; Cilliers, J. J.; Sherman, D.; and Vita, A. D.
*Nature Communications*, Volume 4. 2013.

Paper bibtex abstract

Paper bibtex abstract

@article{wrap64446, volume = {Volume 4}, title = {Macroscopic scattering of cracks initiated at single impurity atoms}, author = {James R. Kermode and L. Ben-Bashat and F. Atrash and J. J. Cilliers and D. Sherman and Alessandro De Vita}, publisher = {Nature Publishing Group}, year = {2013}, journal = {Nature Communications}, url = {http://wrap.warwick.ac.uk/64446/}, abstract = {Brittle crystals, such as coloured gems, have long been known to cleave with atomically smooth fracture surfaces, despite being impurity laden, suggesting that isolated atomic impurities do not generally cause cracks to deflect. Whether cracks can ever deviate when hitting an atomic defect, and if so how they can go straight in real brittle crystals, which always contain many such defects, is still an open question. Here we carry out multiscale molecular dynamics simulations and high-resolution experiments on boron-doped silicon, revealing that cracks can be deflected by individual boron atoms. The process, however, requires a characteristic minimum time, which must be less than the time spent by the crack front at the impurity site. Deflection therefore occurs at low crack speeds, leading to surface ridges which intensify when the boron-dopage level is increased, whereas fast-moving cracks are dynamically steered away from being deflected, yielding smooth cleavage surfaces.} }

Brittle crystals, such as coloured gems, have long been known to cleave with atomically smooth fracture surfaces, despite being impurity laden, suggesting that isolated atomic impurities do not generally cause cracks to deflect. Whether cracks can ever deviate when hitting an atomic defect, and if so how they can go straight in real brittle crystals, which always contain many such defects, is still an open question. Here we carry out multiscale molecular dynamics simulations and high-resolution experiments on boron-doped silicon, revealing that cracks can be deflected by individual boron atoms. The process, however, requires a characteristic minimum time, which must be less than the time spent by the crack front at the impurity site. Deflection therefore occurs at low crack speeds, leading to surface ridges which intensify when the boron-dopage level is increased, whereas fast-moving cracks are dynamically steered away from being deflected, yielding smooth cleavage surfaces.

2010
(2)

Hybrid quantum/classical modeling of material systems : the learn on the fly molecular dynamics scheme.
Moras, G.; Choudhury, R.; Kermode, J. R.; Csanyi, G.; Payne, M. C.; and Vita, A. D.
In Dumitrica, T., editor(s), *Trends in Computational Nanomechanics : Transcending Length and Time Scales*, volume Volume 9, pages 1–23. Springer, November 2010.

Paper bibtex abstract

Paper bibtex abstract

@incollection{wrap64453, volume = {Volume 9}, month = {November}, author = {Gianpietro Moras and Rathin Choudhury and James R. Kermode and Gabor Csanyi and Mike C. Payne and Alessandro De Vita}, booktitle = {Trends in Computational Nanomechanics : Transcending Length and Time Scales}, editor = {Traian Dumitrica}, title = {Hybrid quantum/classical modeling of material systems : the learn on the fly molecular dynamics scheme}, publisher = {Springer}, year = {2010}, journal = {Trends Comput. Nanomechanics Transcending Length Time Scales}, pages = {1--23}, url = {http://wrap.warwick.ac.uk/64453/}, abstract = {The atomistic simulation of many processes in materials involves large-size model systems where different levels of complexity need to be described simultaneously. While accurate quantum mechanical simulations of large-size systems are usually not affordable, less computationally intensive classical models are not suitable for the description of many chemical processes. Hybrid (quantum/classical) modelling schemes are required in these circumstances. Here, we describe the ?Learn on the fly? (LOTF) hybrid molecular dynamics scheme. Some technical aspects of this technique are illustrated through a series of examples of its applications to multiscale processes in silicon } }

The atomistic simulation of many processes in materials involves large-size model systems where different levels of complexity need to be described simultaneously. While accurate quantum mechanical simulations of large-size systems are usually not affordable, less computationally intensive classical models are not suitable for the description of many chemical processes. Hybrid (quantum/classical) modelling schemes are required in these circumstances. Here, we describe the ?Learn on the fly? (LOTF) hybrid molecular dynamics scheme. Some technical aspects of this technique are illustrated through a series of examples of its applications to multiscale processes in silicon

A first principles based polarizable O(N) interatomic force field for bulk silica.
Kermode, J. R.; Cereda, S.; Tangney, P.; and Vita, A. D.
*Journal of chemical physics*, Volume 133(Number 9). 2010.

Paper bibtex abstract

Paper bibtex abstract

@article{wrap64448, volume = {Volume 133}, number = {Number 9}, title = {A first principles based polarizable O(N) interatomic force field for bulk silica}, author = {James R. Kermode and Silva Cereda and P. Tangney and Alessandro De Vita}, publisher = {American Institute of Physics}, year = {2010}, journal = {Journal of chemical physics}, url = {http://wrap.warwick.ac.uk/64448/}, abstract = {We present a reformulation of the Tangney-Scandolo interatomic force field for silica J. Chem. Phys. 117, 8898 (2002), which removes the requirement to perform an Ewald summation. We use a Yukawa factor to screen electrostatic interactions and a cutoff distance to limit the interatomic potential range to around 10 A?. A reparametrization of the potential is carried out, fitting to data from density functional theory calculations. These calculations were performed within the local density approximation since we find that this choice of functional leads to a better match to the experimental structural and elastic properties of quartz and amorphous silica than the generalized gradient approximation approach used to parametrize the original Tangney-Scandolo force field. The resulting O(N) scheme makes it possible to model hundreds of thousands of atoms with modest computational resources, without compromising the force field accuracy. The new potential is validated by calculating structural, elastic, vibrational, and thermodynamic properties of \${$\backslash$}alpha\$-quartz and amorphous silica.} }

We present a reformulation of the Tangney-Scandolo interatomic force field for silica J. Chem. Phys. 117, 8898 (2002), which removes the requirement to perform an Ewald summation. We use a Yukawa factor to screen electrostatic interactions and a cutoff distance to limit the interatomic potential range to around 10 A?. A reparametrization of the potential is carried out, fitting to data from density functional theory calculations. These calculations were performed within the local density approximation since we find that this choice of functional leads to a better match to the experimental structural and elastic properties of quartz and amorphous silica than the generalized gradient approximation approach used to parametrize the original Tangney-Scandolo force field. The resulting O(N) scheme makes it possible to model hundreds of thousands of atoms with modest computational resources, without compromising the force field accuracy. The new potential is validated by calculating structural, elastic, vibrational, and thermodynamic properties of \${$\backslash$}alpha\$-quartz and amorphous silica.

2009
(1)

Hybrid atomistic simulation methods for materials systems.
Bernstein, N.; Kermode, J. R.; and Csanyi, G.
*Reports on Progress in Physics*, Volume 72(Number 2). 2009.

Paper bibtex abstract

Paper bibtex abstract

@article{wrap64450, volume = {Volume 72}, number = {Number 2}, title = {Hybrid atomistic simulation methods for materials systems}, author = {Noam Bernstein and James R. Kermode and Gabor Csanyi}, publisher = {Institute of Physics Publishing Ltd.}, year = {2009}, journal = {Reports on Progress in Physics}, keywords = {p05561}, url = {http://wrap.warwick.ac.uk/64450/}, abstract = {We review recent progress in the methodology of hybrid quantum/classical (QM/MM) atomistic simulations for solid-state systems, from the earliest reports in 1993 up to the latest results. A unified terminology is defined into which the various and disparate schemes fit, based on whether the information from the QM andMMcalculations is combined at the level of energies or forces. We discuss the pertinent issues for achieving ?seamless? coupling, the advantages and disadvantages of the proposed schemes and summarize the applications and scientific results that have been obtained to date.} }

We review recent progress in the methodology of hybrid quantum/classical (QM/MM) atomistic simulations for solid-state systems, from the earliest reports in 1993 up to the latest results. A unified terminology is defined into which the various and disparate schemes fit, based on whether the information from the QM andMMcalculations is combined at the level of energies or forces. We discuss the pertinent issues for achieving ?seamless? coupling, the advantages and disadvantages of the proposed schemes and summarize the applications and scientific results that have been obtained to date.

2008
(1)

Low-speed fracture instabilities in a brittle crystal.
Kermode, J. R.; Albaret, T.; Sherman, D.; Bernstein, N.; Gumbsch, P.; Payne, M. C.; Csányi, A.; and Vita, A. D.
*Nature*, Volume 455(Number 7217): 1224–1227. July 2008.

Paper bibtex

Paper bibtex

@article{wrap64456, volume = {Volume 455}, number = {Number 7217}, month = {July}, author = {James R. Kermode and T. Albaret and Dov Sherman and Noam Bernstein and P. Gumbsch and Mike C. Payne and A. Cs{\'a}nyi and Alessandro De Vita}, title = {Low-speed fracture instabilities in a brittle crystal}, publisher = {Nature Publishing}, year = {2008}, journal = {Nature}, pages = {1224--1227}, keywords = {p00064}, url = {http://wrap.warwick.ac.uk/64456/} }

2007
(1)

Multiscale modeling of defects in semiconductors : a novel molecular-dynamics scheme.
Csanyi, G.; Moras, G.; Kermode, J. R.; Payne, M. C.; Mainwood, A.; and Vita, A. D.
In Drabold, D. A.; and Estreicher, S. K., editor(s), *Theory of Defects in Semiconductors*, volume Volume 104, of Topics in Applied Physics, pages 193–212. Springer Berlin Heidelberg, Berlin Heidelberg, 2007.

Paper bibtex abstract

Paper bibtex abstract

@incollection{wrap64451, volume = {Volume 104}, author = {Gabor Csanyi and Gianpietro Moras and James R. Kermode and Mike C. Payne and Alison Mainwood and Alessandro De Vita}, series = {Topics in Applied Physics}, booktitle = {Theory of Defects in Semiconductors}, editor = {David A. Drabold and Stefan K. Estreicher }, title = {Multiscale modeling of defects in semiconductors : a novel molecular-dynamics scheme}, address = {Berlin Heidelberg}, publisher = {Springer Berlin Heidelberg}, year = {2007}, journal = {Topics in Applied Physics}, pages = {193--212}, keywords = {p00102}, url = {http://wrap.warwick.ac.uk/64451/}, abstract = {Now that the modeling of simple semiconductor systems has become reliable, accurate and routine, attention is focusing on larger scale, more complex simulations. Many of these necessarily involve multiscale aspects and can only be tackled by addressing the different length scales simultaneously. We discuss some of the types of problems that require multiscale approaches. Finally we describe the LOTF (learn-on-the-fly) hybrid scheme with a series of examples to show its versatility and power.} }

Now that the modeling of simple semiconductor systems has become reliable, accurate and routine, attention is focusing on larger scale, more complex simulations. Many of these necessarily involve multiscale aspects and can only be tackled by addressing the different length scales simultaneously. We discuss some of the types of problems that require multiscale approaches. Finally we describe the LOTF (learn-on-the-fly) hybrid scheme with a series of examples to show its versatility and power.

Copy&paste any of the following snippets into an existing page to embed this page. For more details see the documention.

```
<script src="https://bibbase.org/show?bib=http%3A%2F%2Fwrap.warwick.ac.uk%2Fcgi%2Fexportview%2Fauthor_id%2F39508%2FBibTeX%2F39508.bib&jsonp=1"></script>
```

```
<?php
$contents = file_get_contents("https://bibbase.org/show?bib=http%3A%2F%2Fwrap.warwick.ac.uk%2Fcgi%2Fexportview%2Fauthor_id%2F39508%2FBibTeX%2F39508.bib");
print_r($contents);
?>
```

```
<iframe src="https://bibbase.org/show?bib=http%3A%2F%2Fwrap.warwick.ac.uk%2Fcgi%2Fexportview%2Fauthor_id%2F39508%2FBibTeX%2F39508.bib"></iframe>
```