Three parallel algorithms for classical molecular dynamics are presented. The first assigns each processor a fixed subset of atoms; the second assigns each a fixed subset of inter-atomic forces to compute; the third assigns each a fixed spatial region. The algorithms are suitable for molecular dynamics models which can be difficult to parallelize efficiently—those with short-range forces where the neighbors of each atom change rapidly. They can be implemented on any distributed-memory parallel machine which allows for message-passing of data between independently executing processors. The algorithms are tested on a standard Lennard-Jones benchmark problem for system sizes ranging from 500 to 100,000,000 atoms on several parallel supercomputers--the nCUBE 2, Intel iPSC/860 and Paragon, and Cray T3D. Comparing the results to the fastest reported vectorized Cray Y-MP and C90 algorithm shows that the current generation of parallel machines is competitive with conventional vector supercomputers even for small problems. For large problems, the spatial algorithm achieves parallel efficiencies of 90% and a 1840-node Intel Paragon performs up to 165 faster than a single Cray C9O processor. Trade-offs between the three algorithms and guidelines for adapting them to more complex molecular dynamics simulations are also discussed.

Fast parallel algorithms for short-range molecular dynamics

The Mn+1AXn, or MAX, phases are nanolayered, hexagonal, machinable, early transition-metal carbides and nitrides, where n = 1, 2, or 3, M is an early transition metal, A is an A-group element (mostly groups 13 and 14), and X is C and/or N. These phases are characterized by a unique combination of both metallic and ceramic properties. The fact that these phases are precursors for MXenes and the dramatic increase in interest in the latter for a large host of applications render the former even more valuable. Herein we describe the structure of most, if not all, MAX phases known to date. This review covers ≈155 MAX compositions. Currently, 16 A elements and 14 M elements have been incorporated in these phases. The recent discovery of both quaternary in- and out-of-plane ordered MAX phases opens the door to the discovery of many more. The chemical diversity of the MAX phases holds the key to eventually optimizing properties for prospective applications. Since many of the newer quaternary (and higher) phases have yet to be characterized, much work remains to be done.

On the Chemical Diversity of the MAX Phases

https://ceramics.onlinelibrary.wiley.com/doi/pdf/10.1111/jace.17544

Processing of MAX phases: From synthesis to applications

The MAX phase borides Zr2SB and Hf2SB

Surface and bulk ripplocations in layered nanomaterials have recently attracted the attention of researchers because they possess the properties of topological solitons, which are capable of efficient transportof mass and energy and of mediating plastic deformation. In a ripplocation, one or a few layers at the surface or in the bulk of a material are bent or folded. So far, only the static properties of ripplocations have been analyzed. In the present study, the dynamics of graphene bubbles and folds on a graphite substrate are analyzed by full-atomic molecular dynamics and with the help of the two-dimensional chain model. It is demonstrated that such objects, classified as surface ripplocations, are robust solitary waves that propagate while practically radiating no energy. Energy and geometrical parameters of the ripplocations are calculated as the functions of their propagation velocity. In the presence of thermal fluctuations the ripplocations can be accelerated or decelerated, showing a random-walk-like dynamics. Collisions of ripplocations result in their merger. Overall, our results reveal that layered materials can support surface ripplocations that are highly mobile topological solitary waves efficiently transporting mass and energy.

Dynamics of surface graphene ripplocations on a flat graphite substrate

Layered materials and formations are ubiquitous in Nature and span the gamut from individual graphene layers in graphite, to layered composites, to geologic formations. And while the similarities in the deformation of the latter two have been recognized, that the same physics applies at the atomic scale has not. Using atomistic simulations on graphite, and simple instrumented cylindrical indentation experiments on decks of cards and thin steel sheets the authors show that, in all cases, confined buckling leads to the nucleation of multiple ripplocations that rapidly propagate away from under the indenter in a wavelike manner. Upon unloading, they disappear, after dissipating considerable frictional energy. In short, Nature's solution for the deformation of all layered solids, g20 orders of magnitude in scale, is as simple as it is universal: buckling. To be able to shed light on how an earthquake propagates from studying the deformation of graphite and vice versa is quite astonishing and remarkable indeed.

https://link.aps.org/accepted/10.1103/PhysRevMaterials.3.013602

Ripplocations: A universal deformation mechanism in layered solids

Polymeric foams are used for impact protection due to their ability to absorb large amounts of strain energy. In this work, the compressive response of an open cell polyurethane foam currently used as liner in the advanced combat helmet is examined across strain rates. A traditional load frame is used to investigate the quasi-static behavior, and two different modifications of a conventional Kolsky (split-Hopkinson) bar configuration are used to probe the dynamic response. A unique, independent method not relying on strain gage signals is presented that leverages high-speed full-field imaging to track the velocity on each side of the sample-bar interface and used to extract the dynamic stress-strain response; the results are compared against traditional strain gage measurements. X-ray tomography is used to examine the global morphological characteristics of the foam. The foam is found to be strongly rate dependent, where the characteristic properties vary logarithmically with strain rate. An analytical expression is presented to describe the rate dependency that collapses all stress-strain curves on a master curve. Full-field kinematic data from digital image correlation taken during loading is used to extract a nonlinear Poisson’s ratio as a function of strain, which is found to be strain rate insensitive. A tangent Poisson function is used to explore the foam’s auxetic behavior. These findings provide insight on physically-based constitutive modeling of foams, crucial to predictive brain injury simulations, as well as motivate the need to probe local heterogenous behavior across strain rates moving forward.

Strain Rate Dependent Compressive Response of Open Cell Polyurethane Foam

Digital image correlation (DIC) is a non-contact, optical, full-field displacement measurement technique used to map deformations on a body under an applied load by tracking surface features. It is a widely used method in experimental mechanics, owed in part to its ease of setup, and applicability across length scales and material systems. Most commonly, DIC surface patterns consist of contrasting black and white speckles applied by spray paint or an airbrush, which can lack control and consistency of the distribution of speckle sizes. In this paper, drop on demand (DOD) inkjet printing of colloidal suspensions is introduced as a means to apply a speckle pattern with precision on the micro- to mesoscale. A dot array and pseudo-random DOD pattern are applied to high-density polyethylene (HDPE) specimens, tested in compression following ASTM D695-15, and compared with a traditional airbrush speckle method. The results show that the DOD pseudo-random pattern provides a more consistent strain measurement than the other two methods. The array DOD and airbrush patterns introduce erroneous signals which can be explained on the basis of two different principles. A directional pattern, such as the array DOD, is known to introduce errors in the displacement vector identification. For the random airbrush pattern, a quantification of the speckle sizes revealed a large fraction of speckles that had a characteristic length scale of around 1 pixel, which is known to introduce bias errors, since the sampling of the speckle pattern does not satisfy the Nyquist-Shannon sampling theorem. These findings illustrate the utility of DOD speckling for certain DIC applications.

https://par.nsf.gov/servlets/purl/10166687

S. Koumlis

Papers

Ripplocations: A universal deformation mechanism in layered solids

Strain Rate Dependent Compressive Response of Open Cell Polyurethane Foam

Drop on Demand Colloidal Suspension Inkjet Patterning for DIC