Atomistic to continuum modeling of solidification microstructures

https://doi.org/10.1016/j.cossms.2015.09.001Get rights and content

Highlights

  • We review solidification modeling advances from the atomic to the continuum scale.

  • We discuss interface properties and coalescence in light of atomistic calculations.

  • At the continuum scale, we focus on phase-field modeling with selected examples.

  • At the grain scale, we highlight a new multiscale approach for dendritic growth.

  • We summarize some remaining challenges in solidification modeling across scales.

Abstract

We summarize recent advances in modeling of solidification microstructures using computational methods that bridge atomistic to continuum scales. We first discuss progress in atomistic modeling of equilibrium and non-equilibrium solid–liquid interface properties influencing microstructure formation, as well as interface coalescence phenomena influencing the late stages of solidification. The latter is relevant in the context of hot tearing reviewed in the article by M. Rappaz in this issue. We then discuss progress to model microstructures on a continuum scale using phase-field methods. We focus on selected examples in which modeling of 3D cellular and dendritic microstructures has been directly linked to experimental observations. Finally, we discuss a recently introduced coarse-grained dendritic needle network approach to simulate the formation of well-developed dendritic microstructures. This approach reliably bridges the well-separated scales traditionally simulated by phase-field and grain structure models, hence opening new avenues for quantitative modeling of complex intra- and inter-grain dynamical interactions on a grain scale.

Introduction

The quantitative prediction of solidification microstructures has been a long standing computational challenge [15], [5]. It generally requires accurate modeling of the complex dynamics of the solid–liquid interface during the entire solidification process. The past two decades have witnessed major progress in modeling this dynamics through atomistic- and continuum-scale simulations, and a handshake between those methods.

On the one hand, atomistic simulations have yielded accurate predictions of fundamental anisotropic properties of the solid–liquid interface for various metallic systems, such as the interface free-energy [65], [66], [5], [48] and kinetic coefficient [69], [66], [130], [53], [101] that affect microstructural pattern formation over a wide range of solidification conditions. Those simulations typically use a million atoms and interatomic potentials modeled by the embedded-atom method (EAM). They have also yielded useful quantitative insights into solute-trapping in rapidly solidified alloys [158], [70], as well as interface coalescence phenomena that control crystal cohesion during the late stages of solidification [67], [46], [111] and are relevant to hot tearing, reviewed in the article by Rappaz in this issue [118].

On the other hand, phase-field simulations have been extensively used to model the formation of monophase and polyphase solidification microstructures in a wide range of dendritic, eutectic and peritectic alloys [16], [5]. Progress has been made to formulate quantitative phase-field models of alloy solidification for a wider range of phase diagrams [113] and diffusive transport properties [19], [20] within the theoretical framework of the thin interface limit [79], [38], [77]. Results of atomistic modeling have been incorporated into phase-field simulations to model complex observations. For example, the quantitative characterization of interface free-energy anisotropy derived from atomistic simulations has been useful to model changes of dendrite growth orientation [63], [35]. Moreover, boosted by steady increases in computing speed and massively parallel code implementations on multicore architectures such as Graphic Processing Units (GPUs), phase-field simulations have become increasingly 3D and able to access experimentally relevant length and time scales [124], [12].

In the following section, we review recent advances in atomistic modeling of solid–liquid interface properties and coalescence, focusing on the prediction of key parameters influencing microstructure formation or hot tearing on a larger scale. In Section 3, the core section of this article, we review progress in phase-field modeling focusing on cellular and dendritic microstructures. We leave out two-phase microstructures, reviewed in the separate article by Plapp in this issue [114]. In Section 4, we discuss recent progress to bridge the gap between phase-field modeling on the microstructure scale and grain structure modeling [118]. This gap stems from the fact that, while 3D volumes up to about a mm3 are within reach of today’s large scale phase-field computations, those volumes are still minute on the scale of a casting. In contrast, Cellular Automata coupled with Finite Elements (CAFE) models [119], [51], [118] can access those much larger scales and predict grain structures of castings [26]. However, those models do not resolve dynamical interactions between branches of hierarchical dendritic networks, which can strongly influence both intra-grain microstructure selection and the growth competition of different grains [139]. Section 4 describes a Dendritic Needle Network (DNN) approach designed to quantitatively model those interactions on the grain scale, beyond the current reach of phase-field modeling. While this approach is not a substitute for grain structure models, it opens new avenues to test and improve those models by investigating complex branch interactions that shape the grain structure [138].

Section snippets

Solid–liquid interface properties

The advent of microscopic solvability theory during the 1980s [86], [81] led to the understanding that anisotropic properties of the solid–liquid interface have a profound influence on dendrite growth. Those properties generally determine the selection of both the dendrite tip operating state and growth direction. Predictions of solvability theory were validated by phase-field simulations during the 1990s [78], [79] and the first decade of this millennium witnessed major progress in

Quantitative phase-field modeling

Due to their simplicity and potential in handling complex free-boundary problems, phase-field (PF) models have become ubiquitous in materials science [16], [29]. They are well adapted to study solid–liquid interface dynamics during solidification. Quantitative predictions have long been limited by the requirement of a diffuse interface width of comparable size to the physical interface, i.e. a few atomic planes. However, the development of asymptotic analysis of PF equations, now referred to as

Towards the grain scale

In concentrated alloys, solidification usually produces grains in the form of hierarchical networks of sharp needle-like branches, like in Fig. 5a. Thus, several orders of magnitude separate the scale of the tip radius ρ and the diffusion length D/V where D is the solute diffusivity and V the growth velocity. Such morphologies are challenging to model with PF simulations that need to resolve the solid–liquid interface of each branch. To overcome this limitation, we recently developed a

Outlook

The last two decades have witnessed major progress in atomistic simulations methods to compute key interfacial properties impacting solidification microstructures. However, quantitative predictions of those properties remain scarce for alloys. Atomistic modeling is needed to predict interfacial energy anisotropies in the Al–Zn system in order to link more directly continuum scale modeling to experimentally observed changes of dendrite growth directions as a function of Zn concentration. Current

Acknowledgements

A.K. acknowledges support of grant DEFG02-07ER46400 from the US Department of Energy, Office of Basic Energy Sciences, for support of atomistic-scale phase-field-crystal modeling of interfacial properties, and NASA for continuum-scale phase-field and dendritic-needle-network modeling of cellular/dendritic microstructures. D.T. is supported by Amy Clarke’s Early Career award from the U.S. DOE, Office of Basic Energy Sciences, Division of Materials Sciences and Engineering and Los Alamos National

References (165)

  • E.G. Castle et al.

    Evidence for an extended transition in growth orientation and novel dendritic seaweed structures in undercooled Cu–8.9 wt% Ni

    J. Alloy. Compd.

    (2014)
  • S.-W. Chan et al.

    Test for a possible “melting” transition in grain boundaries in aluminum near the melting point

    Scr. Metall.

    (1985)
  • A. Chernov

    Notes on interface growth kinetics 50 years after Burton, Cabrera and Frank

    J. Cryst. Growth

    (2004)
  • V. Fallah et al.

    Phase-field simulation of solidification morphology in laser powder deposition of Ti–Nb alloys

    Acta Mater.

    (2012)
  • J.-L. Fattebert et al.

    Phase-field modeling of coring during solidification of Au–Ni alloy using quaternions and CALPHAD input

    Acta Mater.

    (2014)
  • C.-A. Gandin

    Modeling of solidification: Grain structures and segregations in metallic alloys

    C.R. Phys.

    (2010)
  • Y.F. Gao et al.

    Molecular dynamics simulations of the crystal–melt interface mobility in hcp Mg and bcc Fe

    J. Cryst. Growth

    (2010)
  • W.L. George et al.

    A parallel 3d dendritic growth simulator using the phase-field method

    J. Comput. Phys.

    (2002)
  • J. Ghmadh et al.

    Directional solidification of inclined structures in thin samples

    Acta Mater.

    (2014)
  • V.K. Gupta et al.

    Thin intergranular films and solid-state activated sintering in nickel-doped tungsten

    Acta Mater.

    (2007)
  • S. Gurevich et al.

    Evolution of microstructural length scales during solidification of magnesium alloys

    Acta Mater.

    (2012)
  • J. Hoyt et al.

    Atomistic and continuum modeling of dendritic solidification

    Mater. Sci. Eng.: R: Rep.

    (2003)
  • J. Hoyt et al.

    Kinetic phase field parameters for the Cu–Ni system derived from atomistic computations

    Acta Mater.

    (1999)
  • T. Hsieh et al.

    Observations of roughening/de-faceting phase transitions in grain boundaries

    Acta Metall.

    (1989)
  • K.A. Jackson et al.

    An analytical model for non-equilibrium segregation during crystallization

    J. Cryst. Growth

    (2004)
  • R. Kobayashi et al.

    A continuum model of grain boundaries

    Physica D

    (2000)
  • G. Kurtuldu et al.

    Influence of Cr on the nucleation of primary Al and formation of twinned dendrites in Al–Zn–Cr alloys: Can icosahedral solid clusters play a role?

    Acta Mater.

    (2013)
  • G. Kurtuldu et al.

    Icosahedral quasicrystal-enhanced nucleation of the fcc phase in liquid gold alloys

    Acta Mater.

    (2014)
  • J. Li et al.

    Phase-field study of competitive dendritic growth of converging grains during directional solidification

    Acta Mater.

    (2012)
  • Q. Li et al.

    Evolution of the sidebranch structure in free dendritic growth

    Acta Mater.

    (1999)
  • A.E. Lobkovsky et al.

    Phase field model of premelting of grain boundaries

    Physica D

    (2002)
  • R. Masumura et al.

    Absolute solid–liquid and grain boundary energies of bismuth

    Scr. Metall.

    (1972)
  • A. Melendez et al.

    Measurements of dendrite tip growth and sidebranching in succinonitrile–acetone alloys

    J. Cryst. Growth

    (2012)
  • L. Mikheev et al.

    Mobility of a diffuse simple crystal–melt interface

    J. Cryst. Growth

    (1991)
  • Y. Mishin et al.

    Thermodynamics of grain boundary premelting in alloys. I. Phase-field modeling

    Acta Mater.

    (2009)
  • N. Moelans

    A quantitative and thermodynamically consistent phase-field interpolation function for multi-phase systems

    Acta Mater.

    (2011)
  • A. Adland et al.

    Phase-field-crystal study of grain boundary premelting and shearing in bcc iron

    Phys. Rev. B

    (2013)
  • N. Ahmad et al.

    Solute trapping and solute drag in a phase-field model of rapid solidification

    Phys. Rev. E

    (1998)
  • B.P. Athreya et al.

    Adaptive mesh computation of polycrystalline pattern formation using a renormalization-group reduction of the phase-field crystal model

    Phys. Rev. E

    (2007)
  • A. Barbieri et al.

    Predictions of dendritic growth rates in the linearized solvability theory

    Phys. Rev. A

    (1989)
  • C. Becker et al.

    Atomistic underpinnings for orientation selection in alloy dendritic growth

    Phys. Rev. Lett.

    (2007)
  • N. Bergeon et al.

    Spatiotemporal dynamics of oscillatory cellular patterns in three-dimensional directional solidification

    Phys. Rev. Lett.

    (2013)
  • M. Berghoff et al.

    Efficient techniques for bridging from atomic to mesoscopic scale in phase-field simulations

    J. Comput. Methods Sci. Eng.

    (2013)
  • J. Berry et al.

    Melting at dislocations and grain boundaries: a phase field crystal study

    Phys. Rev. B

    (2008)
  • W. Boettinger et al.

    Phase-field simulation of solidification

    Annu. Rev. Mater. Res.

    (2002)
  • G. Boussinot et al.

    Achieving realistic interface kinetics in phase-field models with a diffusional contrast

    Phys. Rev. E

    (2014)
  • G. Boussinot, E.A. Brener, C. Hueter, R. Spatschek, Elimination of Surface Diffusion in the Non-diagonal Phase Field...
  • J. Bragard et al.

    Linking phase-field and atomistic simulations to model dendritic solidification in highly undercooled melts

    Interface Sci.

    (2002)
  • E. Brener et al.

    Pattern selection in two-dimensional dendritic growth

    Adv. Phys.

    (1991)
  • E.A. Brener et al.

    Kinetic cross coupling between nonconserved and conserved fields in phase field models

    Phys. Rev. E

    (2012)
  • Cited by (106)

    View all citing articles on Scopus
    View full text