Capolungo and Taupin Materials Theory

https://doi.org/10.1186/s41313-018-0013-9
(2019) 3:2
Materials Theory

ORIGINAL ARTICLE Open Access

GD3: generalized discrete defect dynamics

Laurent Capolungo1* and Vincent Taupin2

*Correspondence: [email protected]
1
Materials Science and Technology Abstract
Division, Los Alamos National A mesoscale model is introduced to study the dynamics of material defects lying at
Laboratory, Los Alamos NM87544,
USA
interface junctions. The proposed framework couples the dynamics of discrete
Full list of author information is dislocation and disclination lines. Disclinations are expected to be natural defects at
available at the end of the article interface junctions; their presence serving the purpose of accommodating
discontinuities in rotation fields at material interface junctions. Crystallography-based
rules are proposed to describe the kinematics of disclination motion. A discrete-
continuous couple-stress framework, in which discrete defect lines are introduced as
plastic eigenstrains and eigencurvatures, is proposed to explicitly
 follow
 the dynamics
of interfacial defects. The framework is then applied to study 101̄2 twin transverse
propagation and thickening in magnesium. Focusing first on the case of a twin domain,
It is shown that a disclination based representation of twin domains allows for an
appropriate mechanistic description of the kinematics of shear transformations. In what
concerns twin thickening, the stability of defects at twin interfaces is further studied. To
this end, a 3D crater lying on a twin interface is described as a dipole of disclination
loops. Upon self-relaxation, it is found that out of plane motion of disclinations
followed by the nucleation of twinning dislocations can be activated; thereby showing
that conservative non-planar motion of disclinations can be thermodynamically
favorable; mechanism that had been postulated some 50 years ago.
Keywords: Disclinations, Dislocations, Interfaces, Twinning, Magnesium

Introduction
In the context of multi-scale modeling, the current performance of discrete disloca-
tion dynamics (DDD) codes allows for a linkage between unit processes associated
with dislocation motion -as informed by atomistic simulations- and their collective
effects on both hardening and slip (Arsenlis et al. 2007; Bulatov et al. 1998; Devin-
cre and Kubin 1997; Fivel and Canova 1999; Ghoniem et al. 2000; Zbib et al. 1998).
With this, one can validate, select and calibrate, crystal plasticity based constitutive
laws. Examples of success include the quantification of: latent hardening effects result-
ing from both short and long-range interactions between dislocations (Bertin et al.
2014; Bulatov et al. 2006; Devincre 2013; Franciosi et al. 1980; Madec et al. 2003;
Kubin et al. 2008; Queyreau et al. 2009), the simultaneous effects of voids and self-
interstitial loops induced by irradiation on strength (Sobie et al. 2017a, b), etc. These
advances constitute critical steps towards the development of statistically representative,
microstructure and chemistry sensitive constitutive models in the very spirit of seam-
less multi-scale modeling as motivated in several seminal reviews in the field. Among
the several roadblocks that remain, focus is placed here on the treatment of material
© The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0
International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and
reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the
Creative Commons license, and indicate if changes were made.
Capolungo and Taupin Materials Theory (2019) 3:2 Page 2 of 21

interfaces (Bilby 1955; Bollmann 1970; Frank 1950; Priester 2013; Sutton and Baluffi
1995), and their effects on strength, hardening and eventually ductility (Capolungo
et al. 2007; Hoagland et al. 2002; Mompiou et al. 2009; Wang and Ma 2004; Bayerschen
et al. 2016; Spearot and Sangid 2014). One notes that in the case of dislocations, albeit with
limitations pertaining to the treatment of the system dynamics that disregards entropic
effects, one can trace an explicit linkage between early static and two dimensional descrip-
tions of dislocations (i.e. line tension, stress field, effect on diffraction line broadening),
molecular statics and dynamics studies of dislocation core structures and their mobility
-often aided by density functional theory and minimum energy pathway search methods
- and discrete dislocation dynamics. However, such linkage remains lacking in the case
of defects present at material interfaces. Indeed, early as well as more recent topological
descriptions of admissible material defects provide for a two-dimensional registry of likely
interfacial defects (ID) in terms of disconnection, disclination or dispiration content. Let
us note that other ID based material interface representations have been proposed in the
literature. Some of these account for the fine scale atomic structure of the interface while
other -such as those in the lineage of the Frank Bilby and Bollman’s O-lattice theory -
focus mostly on the material macroscopic interface degrees of freedom. These viewpoints
have largely been enhanced by further considering the relationship between the proposed
description of the interface and the resulting energy. Further, the appropriateness of these
constructs can be validated against atomistic simulations. This was done for example in
the case of hetero-interfaces which have been recently represented as networks of discrete
dislocation lines (Vattré and Demkowicz 2015).
Despite these advances, there is a clear disconnect between these interface con-
structs and crystal plasticity constitutive models in which interface mediated plasticity
is accounted for. To address these shortcomings and to depart from a simple treat-
ment of material interfaces as impenetrable obstacles, different classes of DDD models
with increasing level of complexity have been developed. A first class of models con-
sists of introducing dislocation sources in both bulk and material interfaces, which
allows modeling the decomposition of dislocations at grain boundaries and the asso-
ciated plastic straining by glide (grain boundary sliding) and climb (diffusion), as
well as slip transfer (Ahmed and Hartmaier 2011; Quek et al. 2014, 2016). A second
class of models proposes to add slip transmission and re-emission criteria at material
interfaces, which can invoke critical shear stress (Fan et al. 2015), interfacial energy
(Li et al. 2009; Hou et al. 2009) -possibly as obtained from atomistic simulations (Zheng
et al. 2017)-, or line-tension (de Koning et al. 2002, 2003; Zhou and LeSar 2012). A third
class of models consists of explicitly modeling ID as well as their evolution and interac-
tions with bulk dislocations. Discrete dislocation networks were recently used to model
the migration of low angle mixed grain boundaries (Lim et al. 2012). Networks of mis-
fit dislocations where used to study Ni-based superalloys (Gao et al. 2015). Discrete
intrinsic dislocation networks were also used to simulate their complex interactions with
extrinsic dislocations (Fan et al. 2016; Liu et al. 2011, 2012; Yashiro et al. 2006). Discrete
dislocation networks can be built from coupled kinematic/energetic considerations
(Vattré and Demkowicz 2013, 2015; Vattré et al. 2014b; Vattré 2017a, b), or they can be
atomistically informed (Burbery et al. 2017; Wang et al. 2014).
In general, ID can adopt the form of dislocations (Vattré et al. 2014b), disclinations
(Muntifering et al. 2016; Reinholz et al. 2016; Rösner et al. 2011) and disconnections
Capolungo and Taupin Materials Theory (2019) 3:2 Page 3 of 21

(Hirth and Pond 1996; Hirth and Pond 2006, 2016). Disconnections, consisting of a dislo-
cation and a step (Hirth and Pond 1996), were experimentally reported in several serrated
interfaces. They were shown to play a key role in mediating the motion of interfaces. This
is particularly true in the cases of twin boundaries and of interphase boundaries (Mom-
piou et al. 2009; Khater et al. 2012; Rajabzadeh et al. 2013; Hirth et al. 2016). Atomistic
simulations of the propagation of twins show that nucleation of serrations along the twin
boundary occurs concomitantly with the motion of material interface junctions located at
the twin tip (Xu et al. 2013). A body of work has suggested that these interface junctions
could be represented by disclination lines (Xu et al. 2013).
Focusing now on the case of disclinations. These rotational type line defects have been
found to be practical for describing material interfaces (Fressengeas et al. 2011; Romanov
and Kolesnikova 2009). This is the case for example of high-angle grain boundaries.
Alternatively, as previously stated disclinations have also been identified as the most likely
natural defects in specific regions of materials. This includes triple lines, twin tips, etc.
(Beausir and Fressengeas 2013; Bozhko et al. 2014; Cordier et al. 2014; Kolesnikova et al.
2016; Murayama et al. 2002; Rösner et al. 2011; Sun et al. 2016). Overall, whether disclina-
tions are a convenient mathematical object to describe material interfaces or are intrinsic
interfacial defects (similarly to dislocations) remains debated. One notes that such ques-
tion is complex as the diffraction signature of disclination fields cannot be uniquely
identified and as extinction conditions for disclinations cannot be derived for transmis-
sion electron microscopy. Further, while some have used disclinations as discrete static
objects, the detailed crystallography of disclination lines has not yet been postulated.
Thus far, networks of discrete wedge disclination lines were used to predict twin evolu-
tion in NiMnGa alloys (Reinholz et al. 2016).However a general and 3D consideration of
disclination motion was not proposed. This question includes that of the natural habit
planes of disclinations as well as of their out-of-plane motion. Remarkably, using a con-
tinuous description of incompatibilies in the elastic strain and curvature fields associated
with disclinations, it was shown that it can be kinematically admissible for a disclination
to act as a source/sink of dislocation (deWit 1970). However both the thermodynamics
and crystallography of this kinematic requirement have not been studied.
To address these questions, the present manuscript proposes a simple crystallographic
rule to identify potential habit planes of disclinations. Further, a generalized defect
dynamic framework is derived to predict both in plane and out of plane motion of discli-
nations and its relationship to dislocation emission. The original work of deWit (1970;
1973), who layed-out the foundations for the generalized treatment of line defects, is
therefore extended in this study to consider the full three-dimensional features associated
 
with disclination dynamics. We propose a Generalized Discrete Defect Dynamics GD3
model which combines the dynamics of discrete dislocation and of disclination lines. The
approach can be perceived as an extension of a recently proposed discrete-continuous
Fast Fourier Transform (FFT) DDD model (Bertin et al. 2015), to account for disclinations
in the context of a higher-order continuum Field Disclination and Dislocation Mechanics
(FDDM) framework (Fressengeas et al. 2011; Taupin et al. 2013). We apply the GD3 model
 
to the problem of the transverse propagation of 101̄2 twin domains in hexagonal close
packed magnesium (Partridge 1967). Among others, it is shown that a disclination based
construction of 3-dimensional twin domains can render the kinematics of the shear trans-
formation. Following this, GD3 is used to assess the stability, and the potential role of
Capolungo and Taupin Materials Theory (2019) 3:2 Page 4 of 21

long prismatic-basal and pyramidal-pyramidal interfaces/steps on twin thickening. While

the existence of these large steps has been reported experimentally (Liu et al. 2016), to
date neither their stability nor their motion has been studied. In the context of a fully 3-
dimensional configuration, these defects are represented as a crater on the twin interface
which is formally described with a disclination dipole arrangement. As will be shown, the
simulations suggest a new mechanism whereby large steps along the twin interface could
decay by the emission of twinning dislocations. The results are qualitatively similar to
those predicted in Xu et al. (2013) via atomistic simulations and prove that the aforemen-
tioned dislocation source mechanism, proposed by DeWit (1970) is not only kinematically
admissible but also thermodynamically possible.

Generalized discrete defect dynamics

Useful mathematical notations are provided in the appendix. A small strain setting is
assumed. The material displacement field vector u is single-valued and continuous at any
point of the crystal. The distortion tensor, defined as the gradient of the displacement
U = grad u, is curl-free:
curl U = 0. (1)
The strain tensor  is the symmetric part of U, the rotation tensor ω its skew-symmetric
part and the associated rotation vector ω is:
1 1
ω = − ω : X = curl u. (2)
2 2
It is also a single-valued continuous field. The gradient of the rotation vector is the
second order curvature tensor κ:
κ = gradω. (3)
The total curvature tensor is a compatible curl-free tensor, like the distortion tensor.
The total distortion and curvature tensors are additively decomposed into elastic and
plastic parts:
U = Ue + U p (4)
κ = κe + κp (5)
By decomposing the elastic and plastic distortion into strains and rotations, one also has
the relations:
curl  e = κ te − tr(κ e )I (6)
curl  p = κ tp − tr(κ p )I. (7)
In the above two equations, the right part corresponds to the the curl of the elastic and
plastic spin tensors, respectively.
The mechanical framework proposed considers line defects (i.e. dislocations and discli-
nations) both as discrete and as continuous objects. As such the mechanical fields (i.e.
Cauchy stress, couple-stress, total strains, total curvatures) are obtained by first quantify-
ing the plastic strain and plastic curvatures associated with the motion of dislocations and
disclinations, respectively. In the following, when needed, the symbols  and ⊥ will be
used to differentiate between disclination and dislocation related quantities, respectively.
In the case of dislocations lying on specific slips systems, one writes the total plastic
distortion field within the medium as follows:
Capolungo and Taupin Materials Theory (2019) 3:2 Page 5 of 21

p⊥

Uij = bsi ns⊥ s
j γ . (8)
s

The symmetric part of the plastic distortion is the plastic strain

p⊥ 1   s s⊥ 
ij = bi nj + bsj ns⊥
i γ s, (9)
2 s
while the skew-symmetric part is the plastic spin
p⊥ 1   s s⊥ 
ωij = bi nj − bsj ns⊥
i γ s. (10)
2 s

In the above relations, bs and ns⊥ denote the unit Burgers vector and slip normal direc-
tion of slip system s. The local plastic shear γ s is obtained by multiplying the Burgers
vector of the dislocation considered by the area is has sheared, normalized by the simula-
tion volume. Importantly, Eq. (10) describes the crystal lattice rotation due to dislocation
glide. Lattice rotations are therefore accounted for as they will impact the local elastic
fields as a consequence of the elastic anisotropy of the crystal. Further, the evolution of
lattice curvatures (the gradient of lattice rotations) due to dislocation glide will also affect
the dynamics of disclinations.
We decompose the plastic curvature into disclination and dislocation parts:

κ p = κ p + κ p⊥ . (11)

The plastic curvature tensor due to dislocations is fully compatible as it is the gradient
of the plastic rotation. It can be written as:
p⊥ p⊥
κkl = eijk Uji,l . (12)

On the contrary, the plastic curvature due to disclinations κ p must contain both an
incompatible part κ pi and a compatible part κ pc . The compatible part will be defined
later on.
The net plastic rotation jump due to the presence of a disclination is denoted with the
Frank vector . Assuming known the planes on which disclinations can move, similarly
to the case of dislocations, one can thus write the local incompatible plastic curvature
field as follows:
pi

κij = si ns s
j κ , (13)
s

where  and ns denote the unit Frank vector and ’slip’ normal direction of system s.
s

Obviously the ’slip’ planes for disclinations are not necessarily the same as those of
dislocations. The plastic curvature κ s is obtained by multiplying the Frank vector of the
disclination by the area it has swept, normalized by the simulation volume. In doing so,
we impose that disclination lines and loops move in crystallographic planes, which is
physically realistic and original as compared to the Field Disclination and Dislocation
Mechanics continuous framework (Fressengeas et al. 2011). As shown in the next section,
it is important to note that the plastic curvature induced by disclinations also generates
plastic strains, which in turn affect the Cauchy stress field. Given plastic strain and cur-
vature fields due to a distribution of dislocation and disclination lines, we now search
for the associated internal stress and couple-stress fields. In the absence of significant
body forces, both momentum and moment of momentum are conserved (Cosserat and
Cosserat 1909):
Capolungo and Taupin Materials Theory (2019) 3:2 Page 6 of 21

div T = 0 (14)
div M − T : X = 0, (15)
where the stress tensor T is generally non-symmetric. Thus, the rotated stress vector
T : X balances the divergence of the couple-stress tensor M when T is non-symmetric.
Given a description of the specific free energy density function containing contributions
from elastic strains and curvatures, one can introduce the constitutive relationship for the
symmetric Cauchy stress Tsym and deviatoric couple-stress Mdev (Upadhyay et al. 2013):
Tsym = C :  e + D : κ e (16)
dev e e
M = A:κ +B: . (17)
Plastic strains and curvatures due to defects appear in the above balance equations,
through the additive decomposition  e = ( −  p ) and κ e = (κ − κ p ). Further note that
tensors A, B and D involve internal characteristic lengths that are the order of a fraction
of the Burgers vector magnitude (Taupin et al. 2014; Po et al. 2014; Seif et al. 2015).
Focus is now placed on the motion of discrete dislocation and disclination lines under
the action of internal stress and couple-stress fields. At any point on the dislocation, the
Peach-Koehler force is written as:
 
F⊥ = Tt .b × t⊥ , (18)
where b is the Burgers vector and t⊥ denotes the line unit vector. Similarly, the Peach-
Koehler-type force acting on a point lying on a disclination is expressed as follows:
 
F = Mt . × t , (19)
where  is the Frank vector and t denotes the line unit vector. It is a discrete version of
the driving force defined in earlier work (Fressengeas et al. 2011). Naturally, in the context
of discrete defect dynamics, forces (18) and (19) should be integrated along a discretized
segment in order to extract the forces acting on a node. From the known Peach-Koehler
forces, the local velocity of points lying on the dislocation or disclination lines can be
computed. In the present case, one assumes an overdamped equation of motion expressed
as follows:
F ⊥ = B ⊥ V⊥ , (20)
  
F =B V , (21)
where B⊥ and B respectively denote the viscous drag coefficient matrix for dislocation
and disclination motion, while V⊥ and V are the defect velocity vectors. Atomistic sim-
ulations can be used to estimate the drag coefficients for dislocation glide, depending on
temperature, strain rate, segment character etc. (Bitzek and Gumbsch 2004; Cho et al.
2017). However, they are not known for disclinations. We will assume in our simulations
that dislocation mobility is significantly higher than disclination mobility, meaning that
disclination-mediated interface motion and plastic relaxation mechanisms is then much
slower than dislocation glide in grains. Further for the sake of simplicity, the mobility of
dislocations and of disclination are taken as independent of the segment character. Here
as well, velocities (20) and (21) are then rewritten by introducing the interpolation func-
tions used to discretize each segment. The resulting equations are then integrated over
each segment in contact with a specific node such as to extract an equation of motion at
each node as is typically done in discrete dislocation dynamics.
Capolungo and Taupin Materials Theory (2019) 3:2 Page 7 of 21

Finally, from the viewpoint of kinematics alone, one can introduce a generalized
Burgers vector conservation law. It simply states that the flow of Burgers vector is a
conserved quantity. This type of expression is typically used to solve the kinematics of

junction formation and unzipping processes, i.e. bi = 0. However, in the present case
the equation of continuity is expressed dynamically and in the presence of disclinations,
such that non-curl-free plastic curvature rates can lead to the generation of new dislo-
cation lines. This was originally proposed by deWit (1970) and recently derived in the
Field Dislocation and Disclination Mechanics model (Fressengeas et al. 2011). In order to
express this dislocation source/sink term, we first define the Nye dislocation density ten-
sor component αij = S 1
nbi tj⊥ in the cartesian frame. The number n of dislocation lines
of Burgers vector component bi gives a total length of Burgers vector per unit resolution
surface S. With this areal tensor, the evolution of dislocation densities writes:

α̇ = −curl U̇p , (22)

Further, recalling the decomposition of the plastic distortion into strains and spins, one
has:

α̇ = −curl ˙p + κ̇ pt − tr(κ̇ p )I. (23)

The plastic curvature rates include contributions from dislocation glide (the gradient of
the plastic rotation) and from disclination motion, such that κ̇ p = κ̇ p⊥ + κ̇ p . The last
contribution is the source/sink term for dislocations (Fressengeas et al. 2011). Coming
back to the discrete version, this source/sink term can be written in component form as:
   
p
ṅbi tj = S κ̇ji − tr κ̇ p δij . (24)

In the above discrete form, ṅ is the number of generated dislocations. The vectors bi and
tj are expressed in the reference coordinate system. As such, they have to be projected on
possible crystallographic dislocation systems to nucleate discrete dislocations. A priori,
there is no unique solution (nucleation of dislocation partials for instance) and we shall
select the solution, based on energetic arguments such as dissipation, or on experimen-
tal observations and atomistic simulations. A first application of this source/sink term is
shown later on in this paper.

Numerical implementation, illustrations

Numerical implementation
In this section, we present the important numerical aspects related to the mesoscale
model. The field equations will be solved by using Fast Fourier Transform (FFT) algo-
rithms. First, we present the regularization method used to distribute the plastic fields
induced by discrete dislocation and disclination lines (Bertin et al. 2015) within the FFT
grid. Second, we show how to uniquely calculate the plastic strains arising from plastic
curvatures induced by disclinations (Fressengeas et al. 2011). Third, we provide details on
the spectral FFT code used to solve the balance equations.
We start with the calculation of plastic eigendistortions Up⊥ and plastic eigencurvatures
κ p due to discrete dislocation and disclination lines, respectively. The reader is referred
to the work by Bertin et al for more details (Bertin et al. 2015). The numerical method
is presented in the case of dislocation lines, then the exact same procedure applies for
Capolungo and Taupin Materials Theory (2019) 3:2 Page 8 of 21

disclination lines. Nabarro showed that dislocations can be described as a set of coher-
ent misfitting platelet inclusions (see for instance Mura (1987); Wang et al. (2001)). This
representation of dislocations was recently used to develop the so-called Discrete Contin-
uous Model (DCM) (Lemarchand et al. 2001; Vattré et al. 2014a). In this eigenstrain-based
numerical formulation, the elastic fields of dislocations is obtained by introducing dis-
locations as plate-like inclusions (see Fig. 1), bearing the appropriate plastic strain, in a
finite element solver. With this approach, a dislocation line – defined by its Burgers vector
b and its defect surface S with unit normal n – produces a plastic distortion:

p  
Uij (x) = − bi nj δ(S − x) (25)

where δ(S − x) denotes the three-dimensional Dirac delta function that is zero every-
where except on surface S, and which accounts for the displacement discontinuity [u] = b
across S. The plastic distortion U p (x), directly resulting from the motion of dislocation
lines, is numerically updated using the same regularization procedure. The increment of
plastic distortion generated at voxel xd from the motion of dislocation segments on all
slip systems is expressed as:

 
dU p (xd ) = b s ⊗ n s dγ s (xd ) (26)
s

where the summation is performed over all slip systems s with Burgers vector b s and unit
normal n s , and where dγ s (xd ) denotes the plastic shear increment resulting from crys-
tallographic slip on system s. Considering the motion of all dislocation segments across
voxel xd , increment dγ s (xd ) is expressed as:

Fig. 1 Dislocation eigenstrain formalism. a Schematic of a dislocation loop L defined as the boundary of a cut
introduced over a surface S within a continuous material. A dislocation with Burgers vector b is introduced
when the crystal in domain S+ above surface S is slipped by an amount b = b in the direction of b/b
with respect to the crystal in domain S− below surface S, thereby generating a displacement jump [u] = b
across surface S. b In the eigenstrain theory, dislocations are considered as plate-like Eshelbian inclusions of
thickness t. Surface S corresponding to the slip plane of the dislocation is defined by the plane formed by the
b×t
Burgers vector b and the line direction t such that n = b×t . For a dislocation, t corresponds to the
inter-atomic distance associated with its slip plane
Capolungo and Taupin Materials Theory (2019) 3:2 Page 9 of 21

 xd
dγ s (xd ) = dγij (27)
ij
x
where dγij d denotes the shear produced by the glide of dislocation segment ij at voxel
xd , and where the summation is performed over all segments ij gliding across voxel xd .
As depicted in Fig. 1, dislocations are represented as a plate-like inclusion of thickness
h = 2L, with L denoting the distance between voxels. With this, the shear produced by the
motion of a dislocation can be determined as follows. It is considered that an elementary
sheared area dS(x) centered in x swept by the glide of a portion of a dislocation segment ij
produces an elementary homogeneous plastic shear dγ (x) within an elementary spherical
volume dφ(x) of radius h/2 (Vattré et al. 2014a), such that:
b dS(x) 6b
dγ (x) = = c(d) 3 dS(x) (28)
dφ(x) πh
where b denotes the magnitude of the Burgers vector of the dislocation line. c(d) denotes a
correction factor, as defined in Bertin et al. (2015), which ensures that the predicted stress
fields remain correct regardless of the distance d between the segment and the center of
x
the voxel. Following Eq. 28, the plastic shear dγij d in expression (27) produced by the
glide of a dislocation segment ij is regularized at each grid point xd as:
xd 6b x b x
dγij = dS d = c(d) dSij d (29)
πh3 ij Ve
where Ve = πh3 /6 is the volume of the elementary spherical sheared volume dφ(xd ) of
x
radius h/2, and where quantity dSij d corresponds to the intersection between the area
dAij swept by dislocation segment ij and the elementary sphere of volume Ve centered
at voxel xd . Further details on this regularization procedure (e.g. choice for the regular-
x
ization parameter h, numerical procedure to compute dSij d , numerical implementation,
etc.) can be found in a recent paper (Bertin et al. 2015). The same procedure applies for
disclinations, the Frank vector substituting for the Burgers vector and using disclination
slip systems.
From the known incompatible plastic curvature due to disclinations κ pi , one must
quantify the associated incompatible plastic strain  p and compatible plastic curvature
κ pc fields. The reader is referred to the original paper introducing Field Disclination and
Dislocation Mechanics for more details (Fressengeas et al. 2011). Once  p is known, the
total plastic strain  p =  p +  p⊥ is obtained and the balance equations can be finally
solved. The plastic strain  p satisfies:
 t  t    
curl  p = κ pi + κ pc − tr κ pi I − tr κ pc I, (30)

The compatible curvature κ pc is the gradient of a continuous plastic rotation field. The
latter rotation and the plastic strain  p compose the plastic distortion Up , such that the
above incompatibility equation can be written as:
 t  
curl Up = κ pi − tr κ pi I. (31)

Note that Up cannot contain any gradient part to ensure uniqueness of the solution
(Acharya 2001). As such, it must be divergence-free. Hence, by taking the curl of above
equation, Up is solution of the higher-order equation:
 t   
div gradUp = Up = −curl κ pi − tr κ pi I . (32)
Capolungo and Taupin Materials Theory (2019) 3:2 Page 10 of 21

Note in the above equation that the symbol  in front of the plastic distortion indi-
cates the Laplace operator applied to the plastic distortion. No boundary condition
on Up applies here because of periodic boundary conditions in our FFT framework.
Equation (32) is solved in the Fourier space by using FFT algorithms and special rules
for evaluating spatial derivatives. The numerical method used is presented in a recent
paper (Berbenni et al. 2014). Finally, the plastic strain  p can be derived as the sym-
metric part of Up . The compatible plastic curvature κ pc can be calculated from the
skew-symmetric part of the distortion.
Once the above plastic strains and curvatures are known, we can solve for the stress
and couple-stress balance Eqs. (14, 15). Recall that the elastic moduli tensors A, B and D
involve internal characteristic lengths that are the order of a fraction of the Burgers vec-
tor magnitude. Hence, at a resolution scale sufficiently larger than interatomic distances,
which is the case in the proposed mesoscale model, the couple-stresses do not affect
the balance of the stress field and the skew-symmetric stress is negligible as compared
to the symmetric Cauchy stress. Hence, solving the above balance equations or simply
div(C :  e ) = 0 yields exactly the same stress and couple-stress fields. Obviously, it is not
true anymore when using a very small resolution scale to model the core elastic fields of
defects (Taupin et al. 2017). In the model, the elastic stiffness tensor C is heterogeneous
because of crystal orientation and elastic anisotropy. We use the accelerated FFT spec-
tral scheme formerly introduced by Eyre and Milton (Eyre and Milton 1999; Michel et al.
2001) and recently used in the DCM-FFT-DDD code (Bertin et al. 2015). The converged
solution of the balance equations provides the total strain tensor .
The last step consists of deriving the total curvature tensor κ, such that the elastic, total
minus plastic, curvature is known and the stresses and couple-stresses can be evaluated as
per constitutive elastic laws (16, 17). The strategy proposed to derive the curvature tensor
in the Fourier space is as follows. The Cauchy stress field satisfies the balance equation:

div Tsym = 0, (33)

which reads in component form

sym
Tij,j = 0. (34)

We now switch from the real space to the Fourier space. Let λ̂ denote the Fourier trans-
form of a function λ and ξ be the Fourier frequency vector with components ξi in a
reference cartesian frame. The imaginary number is denoted by i. The balance Eq. (34)
becomes
sym
iξj T̂ij = 0. (35)

The heterogeneous elastic moduli tensor is C = C0 +δC, where C0 is the stiffness tensor
of a homogeneous reference medium and δC is the spatial fluctuation of C. Expressing
sym  p
the Cauchy stress as Tij = Cijkl uk,l − kl allows rewritting Eq. (35) in the well-known
form
0
ξl ξj Cijkl ûk = iξj τ̂ij , (36)

where τ̂ij is the stress polarization tensor that includes both the fluctuation δC and the
plastic strain. The above equations set an algebraic linear system of the form

Ĝik ûk = fˆi , (37)

Capolungo and Taupin Materials Theory (2019) 3:2 Page 11 of 21

where Ĝ is associated with the homogeneous reference tensor C0 , and ûk is the material
displacement. The latter is given by
−1 ˆ
ûk = Ĝki fi . (38)
From the expression of the displacements, we now derive the curvature components.
We first express the rotation vector as half of the curl of the displacement vector:
1 1 −1 ˆ
ω̂i = i eilk ξl ûk = i eilk ξl Ĝkm fm . (39)
2 2
The curvatures thus read:
1 1 −1 ˆ
κ̂ij = − eilk ξj ξl ûk = − eilk ξj ξl Ĝkm fm . (40)
2 2
We finally cast the stress polarization tensor in f̂. Eq. (40) is transformed into :
1 −1
κ̂ij = − eilk ξj ξl Ĝkm iξn τ̂mn . (41)
2
By defining the following fourth-order tensor:
i −1
ˆ ijmn
κ
= − eilk ξj ξl ξn Ĝkm , (42)
2
The total curvature finally reads:
κ̂ij = ˆ ijmn
κ
τ̂mn . (43)
An inverse FFT can finally be used to obtain the curvature tensor in the real space.
Finally, one obtains the elastic curvature tensor by subtracting the plastic part. Note that
the material rotation vector is also known from Eq. (39), which allows calculating the
elastic rotation by subtracting the plastic rotation.

Illustrations
We now present two illustrations of the proposed model. A simple cubic crystal struc-
ture with lattice parameter a = 0.32nm is assumed. Isotropic elasticity is used, with shear
modulus μ = 17GPa and Poisson coefficient ν = 0.3. Further, we consider a periodically
reproduced cubic simulation box with dimension (1800a)3 . The critical distance for dis-
crete defect interactions (annihilation, junctions etc.) is set to be 10a = 3.2nm and the
defect remeshing distance to 20a = 6.4nm. The FFT grid is made of 1283 voxels, imply-
ing a spatial resolution for the elastoplastic fields of ∼ 14a ∼ 4.5nm. Regarding defect
velocities, the drag coefficient matrix is taken as diagonal and isotropic such that only one
drag coefficient is necessary to describe dislocation glide. Here we take B⊥ = 10−2 Pa.s.
The drag coefficient matrix for disclinations is also taken as diagonal and isotropic result-
ing in the need for only one drag coefficient. Its value is 100 times that of dislocations. In
these illustrations, we restrict ourselves to the sole consideration of disclination dynam-
ics. The glide planes of disclinations are limited to cube faces of the crystal cell; e.g. {001}.
The Frank vector of disclinations is aligned with the cube directions and its magnitude is
denoted by . In the first simulation, we propose to model an isolated circular disclination
loop. Figure 2 shows the self-relaxation of a circular disclination loop of radius r=100nm,
under the action of its own couple-stress field. The Frank vector is along [100] direction
with magnitude arbitrarily chosen as  = 1rad. The initial loop is shown on the top right
and top left of the figure. The loop can glide in the (001) plane. The evolution of the shear
stress T23 (respectively the couple-stress M12 ) during the self-relaxation of the disclina-
tion loop is also shown on the left (respectively right) column in the figure. The simulation
Capolungo and Taupin Materials Theory (2019) 3:2 Page 12 of 21

Fig. 2 Self-annihilation of a disclination loop. The disclination loop (red line) is let to self-relax under its own
elastic field. (a,b,c) Relaxation of the internal shear stress field. (d,e,f) Relaxation of the internal couple-stress
field

predicts the self-annihilation of the disclination loop under the action of its own couple-
stress field. In short, it is shown here that the planar motion of disclinations is similar
to those of dislocations. Although not shown here, applying a couple-stress M13 leads
to the expansion of the disclination loop. Focus is then placed on the case of short and
long range elastic interactions between twin disclinations. Consider a two-dimensional
scenario in which a matrix contains a stack of domains as depicted in Fig. 3. Further
assume that the crystal disorientation axis across each domain is parallel to X1 . Each
domain is bounded by a wedge disclination quadrupole, the wedge disclination lines being
infinitely straight. The Frank vector is along [100] direction with magnitude arbitrarily
chosen as  = 0.1rad and the elastic fields are invariant w.r.t. the X1 axis. This configu-
ration is very similar to that obtained in 2D simulations of twins in shape memory alloys
(Reinholz et al. 2016), except that it is more symmetric here. The initial in plane internal
shear stress field is shown in the Fig. 3 and resembles to that predicted in simulations of
shape memory alloys. To model twin thickening/thinning, the disclination slip planes are
now taken as (010). The system is then set to relax in its own elastic field. Figure 3 shows
the migration of twin boundaries in the normal [001] direction, i.e. thickening or thinning
of domains, until one recovers a defect free matrix. We note that the long and short range
(i.e. annihilation) interactions between the twin domains are fully accounted for.
Capolungo and Taupin Materials Theory (2019) 3:2 Page 13 of 21

Fig. 3 Self-relaxation of twin stacks. Initial twin stacks are let to self-relax by thickening/thinning of twins in
the normal direction. a: sketch of the simulated twin stacks using wedge disclination lines. b: Colour-coded is
shown the initial in plane shear stress field (MPa) due to disclination lines (red lines). c View in the (X1 , X3 )
plane showing the progressive growth and shrinkage of twins during self-relaxation until a defect-free matrix
is obtained

Applications
Focus is now placed on the application of the disclination and dislocation dynamics
framework to study the kinematics and thermodynamics of twin propagation in Mg, as
well as twin thickening mediated by the motion of ledges and emission / propagation of
 
twinning dislocations at twin interfaces. We consider 101̄2 twins (Partridge 1967). The
 
twin plane normal is 101̄2 and coincides with the X3 axis, the shear direction is 101̄1

and coincides with the X2 axis. The normal to shear direction in the twin plane is 12̄10
and coincides with the X1 axis. The value of twinning shear is s = 0.131. Twinning dis-

locations gliding in the twin plane have a Burgers vector in the shear direction 101̄1
and the magnitude is b = 0.0446nm. The elementary step associated with disconnec-
tions and separating adjacent twin planes is h = 0.378nm. In the reference coordinate
p
system chosen, the shear transformation s leads to a distortion U23 = s. The plastic dis-
p p
tortion can therefore be decomposed into plastic strains, 23 = 32 = s/2, and spins
p p p
ω23 = −ω32 = s/2. In consequence the rotation vector is 1 = −s/2. Therefore, the mag-
nitude of the Frank vector of the elementary disclinations used to build the twins and twin
interface ledges is 0.0655rad=3.7°. The magnesium lattice parameters are a = 0.32nm and
c = 0.52nm. Isotropic elasticity is reasonably assumed (Capolungo et al. 2010), with shear
modulus μ = 17GPa and Poisson coefficient ν = 0.3.
We postulate that disclinations could be natural defects at the intersections of material
interfaces (i.e. junction of interfaces). This is the case triple lines, twin tips or ledges on
the twin interface. This has also been suggested in other body of works (Bollmann 1970;
Xu et al. 2013; Barrett and El Kadiri 2014). With regards to disclinations’ habit planes,
as suggested by atomistic simulations -albeit providing at best circumstantial evidence-
Capolungo and Taupin Materials Theory (2019) 3:2 Page 14 of 21

we further postulate that disclinations can naturally glide on minimum energy inter-
face planes. In the case of tensile twinning in magnesium, disclinations would therefore
 
glide on the 101̄2 twin plane. In addition, wedge disclinations which would bound the
tip of a twin domain could also glide/climb on the basal/prismatic interface. The lat-
ter was shown to be a minimum energy interface. In what concerns the ’dark side’ of
the twin (Liu et al. 2016), i.e. observing the twin domain along the twin shear direction,
twist disclinations would then potentially ’cross-slip’ on pyramidal/pyramidal boundaries





(i.e. 2̄110 // 21̄1̄0 or potentially 1̄010 // 101̄1̄ ).

 
transverse propagation of the 101̄2 twin
As stated in the above, disclination and their motion allow for an explicit description
of lattice rotations. As such they appear as ideal defects/constructs to track the kinetics
 
of propagation of twin domains. This is studied in the case of an isolated 101̄2 twin
domain which stability and transverse propagation is modeled as shown in Figs. 4–6. A
twin embryo is created simply by placing two disclination loops of opposite Frank vec-
p
tor 1 = e1 = −1 = ±s/2 above one another on parallel twin planes. The positive
(resp. negative) loop is placed on the bottom (resp. top) of the dipole, such as to delimit
a twin domain with negative plastic rotation −s/2 with respect to the matrix. Figure 4
shows the elastic fields within the initial twin domain. In the top of the figure are shown
the in plane shear stress field T23 and the elastic dilatation/contraction in the plane par-

allel to the twinning shear direction 101̄1 . Disclination lines tangent to that plane are of
pure wedge character. The location of elastic dilatations (resp. contractions) allows iden-
tifying negative (resp. positive) wedge disclinations. The bottom of Fig. 4 shows the out of
plane shear stress fields in the plane normal to the shear direction. The disclination lines

Fig. 4 Initial elastic fields in a (101̄2) twin domain. The twin is built with a dipole of disclination loops (red
lines). (a,b) Stress field T23 and elastic dilatation/contraction in a slice parallel to the shear direction, the frame
is shown on the left. (c,d) Stress fields T23 and T12 in a slice normal to the shear direction, the frame is shown
on the left
Capolungo and Taupin Materials Theory (2019) 3:2 Page 15 of 21

Fig. 5 Self-shrinkage in the transverse direction of a (101̄2) twin. The twin is modelled by a dipole of
disclination loops (red solid lines). From (a) to (d): Progressive self-shrinkage of the twin. In blue colour is
shown the plastic rotation around X1 . Only plastic rotation values smaller than -0.06 rad are shown

Fig. 6 Growth in the transverse direction of a (101̄2) twin. The twin is modelled by a dipole of disclination
loops (red solid lines) and gowth as per application of a shear strain 23 of 0.01. From (a) to (d): progressive
growth of the twin. The arrows show the applied shear in figure (a). In blue colour is shown the plastic
rotation around X1 . Only plastic rotation values smaller than -0.06 rad are shown
Capolungo and Taupin Materials Theory (2019) 3:2 Page 16 of 21

tangent to that plane are of pure twist character. One can see that the plastic eigenstrains
and eigencurvatures resulting from the generation of the twin domain is compensated
in the matrix phase such as to lead to the generation of large ’back stresses’ acting on
the disclinations. Importantly, one can note an asymmetry of the shear stress field T23 in
planes parallel (Fig. 4a) and normal (Fig. 4c) to the twinning shear direction. Strong pos-
itive stresses can be seen in front of the twin in the plane parallel to the shear direction,
while they are much less significant in the other plane. Hence, from a purely elastostatic
standpoint and considering only stresses, the latter suggest that twin propagation should
 
be favored in the shear direction 101̄1 as compared to the normal direction 12̄10 .
However, from a dynamic point of view, disclination motion is driven by couple-stresses
and the kinetics of propagation of the twin domain in shear and lateral directions depend
on the mobility of disclinations. As such, introducing a segment-character dependent
disclination drag will allow controlling the kinetics, together with internal elastic fields of
the twin. This will be the object of further studies. Figure 5 shows the initial twin domain
as well as the plastic rotation field (the matrix phase is taken as reference for comput-
ing orientations). It then shows the evolution of the disclination loops and the associated
rotation field during self-relaxation under the internal couple-stress field of the twin. The
twin domain is consistently observed to shrink, together with the rotation field, such
that one ends up with a single phase matrix. However and as expected, upon imposing a
macroscopic shear strain to the system, 23 = 0.01, it is found that the twin domain can
now expand. This is shown in Fig. 6. This applied shear strain yields a couple-stress as per
non local elasticity tensor B (Taupin et al. 2014, 2017), which is in turn responsible for
disclination loops expansion. Very importantly, not only the prediction of twin growth in
the transverse direction can be predicted by our 3D model, but so does the evolution of
the rotation field, as shown in Fig. 6. We note that the shape of the twin domain remains
almost perfectly circular during shrinking and growth.

Twin thickening
Thickening of a twin domain can be mediated by several processes which all lead to the
generation of ID. Focus is placed on the generation of twinning dislocations from ledges
present on a twin interface. For this first study, the intent is not to extract the kinetics
and overall contributions of these mechanism but solely to assess the thermodynamics
and kinematics of the mechanism. As shown in Fig. 7, we consider a twin plane con-
taining a ’crater’ in which the matrix phase locally penetrates into the twin domain. The
boundaries of the crater can therefore be modeled as a disclination loop dipole with
Frank vector corresponding to half of the twinning shear as in the previous case. For
the sake of simplicity a square loop configuration is taken and the initial loop sizes are
200a and 100a. The crater height is 50a. The four interfaces connecting the two square
disclination loops can be interpreted as Basal / Prismatic (BP) and Pyramidal/ Pyramidal
(PP≈ PP1 or PP2) interfaces. As in the previous case, disclination loops can glide in the
 
101̄2 twin plane. However, out of plane motion of disclination loops is also allowed
in the BP and PP interfaces. The choice of the plane of motion is simply based on the
driving forces magnitude acting on disclination lines. If disclinations experience out of
plane motion, following Eq. (24), twinning dislocation loops with Burgers vector b in
the twinning shear direction will be generated in adjacent twin planes once disclination
loops have moved by an elementary step distance h perpendicular to the twin plane.
Capolungo and Taupin Materials Theory (2019) 3:2 Page 17 of 21

Fig. 7 Simulation of a twin interface crater by using a dipole of disclination loops. a 2D Sketch of the initial
crater configuration (blue lines). Blue triangles denote wedge disclinations. b 2D sketch of the dynamics of
the crater. Dotted horizontal lines denote adjacent planes parallel to the twin plane and separated by
elementary steps h. Orange triangles show the new position of disclinations, leading to the nucleation of
steps and edge dislocations, i.e. disconnections. c Initial dipole of square disclination loops forming the crater
(blue lines). (d,e,f) Generation of dislocation loops (red lines) from the out of plane motion of disclination
loops of opposite sign (orange lines). The lower dislocation loop shrinks while the upper one expands and
reaches an equilibrium position

Upon self-relaxation, the couple-stress field induces an out of twin plane motion of the
disclination loops crossing onto the BP and PP1 planes. The loops attract each other, such
that the upper loop moves down while the lower loop moves up. Once both loops have
moved by a distance h perpendicular to the twin plane, a dipolar arrangement of square
twinning dislocation loops (red lines in the figure) is thus generated. As the mobility of
dislocations is higher than that of disclinations, the upper and lower dislocation loops
rapidly expand and shrink, respectively. Interestingly, the upper loop reaches an equilib-
rium elliptical shape, as shown in Fig. 7d, due to a balance between the repulsing stress
field of the crater and the line tension of the elliptical dislocation loop. The loop is more
elongated in the screw direction. As sketched in the Fig. 7, we shall interpret the observed
mechanism as the nucleation of disconnection loops from ledges in the twin interface.
The generated disconnection loops are indeed composed of a step and of a dislocation,
and they help propagating the twin interface. Such a mechanism is very similar to obser-
vations in twin boundary interfaces in Mg as obtained from MD simulations, and can be
invoked in other interfaces (Mompiou et al. 2009; Khater et al. 2012; Rajabzadeh et al.
2013; Hirth et al. 2016).

Conclusions
 
We have proposed a novel Generalized Discrete Defect Dynamics GD3 mesoscale
model. In addition to discrete dislocation lines, discrete disclination lines are introduced
to account for the incompatibilty of elastic curvatures and discontinuity of elastic rota-
tions in material interface defects, such as triple lines, twin tips, and interface ledges.
Apart from the numerical implementation, our work proposes that disclinations could
be intrinsic defects that are to be unambiguously found at junction between material
interfaces. Further, our work postulates that disclination motion is constrained to be
located on minimum energy interface planes (in the case of twinning the Coherent Twin
Capolungo and Taupin Materials Theory (2019) 3:2 Page 18 of 21

boundary, Basal Prismatic boundary etc.) Third, we evidence the possibility for disclina-
tions to change glide plane and emit dislocation in a fully conservative fashion. Applying
 
the model to 101̄2 twinning in magnesium, It is shown that twin domains can be appro-
priately described by using dipoles of disclination loops. Further, both shrinking and
propagation of twins can then be simulated. Importantly, the introduction of disclination
loops as plastic eigencurvatures allows to conveniently track the evolution of rotations
during twin propagation. The disclination based constructs of twins thus appears to be
practical as internal stress fields, rotations and propagation can be rendered simultane-
ously. Further, the model was used to study the stability of 3-dimensional defects, craters,
along the twin interface. Simulations suggest that large ledges along the twin boundaries
could decay by generating twinning dislocations; a phenomenon observed in atomistic
simulations that insofar had not been mechanistically described or understood. The pre-
diction of this mechanism is the sole consequence of the higher order kinematics and
thermodynamic framework of the model. Naturally, the proposed framework pauses sev-
eral interrogations pertaining to the mobility of disclinations, their rate of decay and the
short range interactions with dislocations. These will be pursued in further studies.

Appendix: mathematical notations

A bold symbol denotes a tensor. When there may be ambiguity, an arrow is superposed
to represent a vector: V. The symmetric part of tensor A is denoted Asym . Its skew-
symmetric and deviatoric parts are Askew and Adev respectively. The transpose of a tensor
A is denoted At . The tensor A · B, with rectangular Cartesian components Aik Bkj , results
from the dot product of tensors A and B, and A ⊗ B is their tensorial product, with com-
ponents Aij Bkl . The vector A · V, with rectangular Cartesian components Aij Vj , results
from the dot product of tensor A and vector V. A : represents the trace inner product of
the two second order tensors A : B = Aij Bij , in rectangular Cartesian components, or the
product of a higher order tensor with a second order tensor, e.g., A : B = Aijkl Bkl . The
cross product of a second-order tensor A and a vector V, the div and curl operations for
second-order tensors are defined row by row, in analogy with the vectorial case. For any
base vector ei of the reference frame:
 
(A × V)t · ei = At · ei × V (44)
 
(div A)t · ei = div At · ei (45)
 
(curl A)t · ei = curl At · ei . (46)

In rectangular Cartesian components:

(A × V)ij = ejkl Aik Vl (47)

(divA)i = Aij,j (48)
(curlA)ij = ejkl Ail,k . (49)

where ejkl is a component of the third-order alternating Levi-Civita tensor X. A vector A

is associated with tensor A by using its trace inner product with tensor X:

1 1
(A)k = − (A : X)k = − eijk Aij . (50)
2 2
Capolungo and Taupin Materials Theory (2019) 3:2 Page 19 of 21

In the component representation, the spatial derivative with respect to a Cartesian

coordinate is indicated by a comma followed by the component index. A superposed dot
represents a material time derivative.
Abbreviations
BP: Basal/Pyramidal; DCM: Discrete continuous model; DDD: Discrete dislocation dynamics; FFT: Fast fourier transform;
FDDM: Field disclination and dislocation mechanics; GD3 : Generalized discrete defect dynamics; ID: Interfacial defect; PP:
Pyramidal/Pyramidal; PP1: Pyramidal/Pyramidal 1; PP2: Pyramidal/Pyramidal 2

Acknowledgements
This work was fully funded by the U.S. Department of Energy under Project FWP 06SCPE401 from the office of Basic
Energy Sciences. The authors thank C.N. Tomé and D. Spearot for valuable discussions.

Availability of data and materials

All numerical data generated for this work can be access by contacting the corresponding author.

Authors’ contributions
LC and VT developed the model, ran/interpreted the simulations and wrote the manuscript. Both authors read and
approved the final manuscript.

Competing interests
The authors declare no competing interests.

Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Author details
1 Materials Science and Technology Division, Los Alamos National Laboratory, Los Alamos NM87544, USA. 2 Université de

Lorraine, CNRS, Arts et Métiers ParisTech, LEM3, F-57000 Metz, France.

Received: 25 May 2018 Accepted: 6 December 2018

References
A. Acharya, A model of crystal plasticity based on the theory of continuously distributed dislocations. J. Mech. Phys.
Solids. 49, 761 (2001)
N. Ahmed, A. Hartmaier, Mechanisms of grain boundary softening and strain-rate sensitivity in deformation of
ultrafine-grained metals at high temperatures. Acta Mater. 59, 4323–4334 (2011)
A. Arsenlis, W. Cai, M. Tang, M. Rhee, T. Oppelstrup, G. Hommes, T. G. Pierce, V. V. Bulatiov, Enabling strain hardening
simulations with dislocation dynamics. Modelling Simul. Mater. Sci. Eng. 15, 553–95 (2007)
C. D. Barrett, H. El Kadiri, The roles of grain boundary dislocations and disclinations in the nucleation of 101̄2 twinning.
Acta Mater. 63, 1–15 (2014)
E. Bayerschen, A. T. McBride, B. D. Reddy, T. Böhlke, Review on slip transmission criteria in experiments and crystal
plasticity models. J. Mater. Sci. 51, 2243–2258 (2016)
B. Beausir, C. Fressengeas, Disclination densities from EBSD orientation mapping. Int. J. Solids. Structures. 50, 137 (2013)
S. Berbenni, V. Taupin, K. S. Djaka, C. Fressengeas, A numerical spectral approach for solving elasto-static field dislocation
and g-disclination mechanics. Int. J. Solids Structures. 51, 4157 (2014)
N. Bertin, C. N. Tomé, I. J. Beyerlein, M. R. Barnett, L. Capolungo, On the strength of dislocation interactions and their effect
on latent hardening in pure magnesium. Int. J. Plasticity. 62, 72–92 (2014)
N. Bertin, M. V. Upadhyay, C. Pradalier, L. Capolungo, A FFT-based formulation for efficient mechanical fields computation
in isotropic and anisotropic periodic discrete dislocation dynamics. Modelling Simul. Mater. Sci. Eng. 23, 065009 (2015)
B. A. Bilby, in Bristol Conference Report on Defects in Crystalline Solids, Types of dislocation source (The Physical Society,
London, 1955), p. 124
E. Bitzek, P. Gumbsch, Atomistic study of drag, surface and inertial effects on edge dislocations in face-centered cubic
metals. Mater. Sci. Eng., A. 387-389, 11–15 (2004)
W. Bollmann, Crystal defects and crystalline interfaces. (Springer-Verlag, Berlin Heidelberg GmbH, 1970)
S. I. Bozhko, V. Taupin, M. Lebyodkin, C. Fressengeas, E. A. Levchenko, K. Radikan, O. Lübben, V. N. Semenov, I. V. Shvets,
Disclinations in molecular layers on WO2/W(110) surfaces. Phys. Rev. B. 90, 214106 (2014)
V. Bulatov, F. Abraham, F. Kubin, B. Devincre, S. Yip, Connecting atomistic and mesoscale simulations of crystal plasticity.
Nature. 391, 669–72 (1998)
V. V. Bulatov, L. L. Hsiung, M. Tang, A. Arsenlis, M. C. Bartelt, W. Cai, J. N. Florando, M. Hiratani, M. Rhee, G. Hommes, T. G.
Pierce, T. Diaz de la Rubia, Dislocation multi-junctions and strain hardening. Nature. 440, 1174–8 (2006)
N. B. Burbery, G. Po, R. Das, N. Ghoniem, W. G. Ferguson, J. Micromech, Dislocation dynamics in polycrystals with
atomistic-informed mechanisms of dislocation-grain boundary interactions. Mol. Phys. 02, 1750003 (2017)
L. Capolungo, D. E. Spearot, M. Cherkaoui, D. L. McDowell, J. Qu, K. I. Jacob, Dislocation nucleation from bicrystal interfaces
and grain boundary ledges: Relationship to nanocrystalline deformation. J. Mech. Phys. Solids. 55, 2300–2327 (2007)
L. Capolungo, I. J. Beyerlein, Z. Qwang, The role of elastic anisotropy on plasticity in hcp metals: a three-dimensional
dislocation dynamics study. Modelling Simul. Mater. Sci. Eng. 18, 085002 (2010)
J. Cho, J.-F. Molinari, G. Anciaux, Mobility law of dislocations with several character angles and temperatures in FCC
aluminum. Int. J. Plast. 90, 66–75 (2017)
Capolungo and Taupin Materials Theory (2019) 3:2 Page 20 of 21

P. Cordier, S. Demouchy, B. Beausir, V. Taupin, F. Barou, C. Fressengeas, Disclinations provide the missing mechanism for
deforming olivine-rich rocks in the mantle. Nature. 507, 51 (2014)
E. Cosserat, F. Cosserat, Théorie des corps déformables. (Hermann, Paris, 1909)
M. de Koning, R. Miller, V. V. Bulatov, F. F. Abraham, Modelling grain-boundary resistance in intergranular dislocation slip
transmission. Philos Mag A Physics of Condensed Matter Structure Defects and Mechanical Properties. 82, 2511–2527
(2002)
M. de Koning, R. J. Kurtz, V. V. Bulatov, C. H. Henager, R. G. Hoagland, W. Cai, M. Nomura, Modeling of dislocation-grain
boundary interactions in FCC metals. J. Nucl. Mater. 323, 281–289 (2003)
B. Devincre, L. P. Kubin, Mesoscopic simulations of dislocations and plasticity. Mater. Sci. Eng. Struct. Mater. Prop.
Microstruct. Process. 234, 8–14 (1997)
B. Devincre, Dislocation dynamics simulations of slip systems interactions and forest strengthening in ice crystals. Phil.
Mag. 93, 235–46 (2013)
R. deWit, in Fundamental aspects of dislocation theory. ed. by J. A. Simmons, R. deWit, and R. Bullough, Linear theory of
static disclinations. Nat. Bur. Stand. (US), Spec. Publ. 317, Vol. I, (1970), pp. 651–673
R. deWit, Theory of disclinations: IV. Straight disclinations. J Res Natl Bur Stand-A Phys Chem. 77A, 607 (1973)
D. J. Eyre, G. W. Milton, A fast numerical scheme for computing the response of composites using grid refinement. Eur
Phys J Appl Phys. 6, 41–47 (1999)
H. Fan, S. Aubry, A. Arsenlis, J. A. El-Awady, Orientation influence on grain size effects in ultrafine-grained magnesium.
Scripta Mater. 97, 25–28 (2015)
Fan, H., S. Aubry, A. Arsenlis, J. A. El-Awady, Grain size effects on dislocation and twinning mediated plasticity in
magnesium. Scripta Mater. 112, 50–53 (2016)
M. C. Fivel, G. R. Canova, Developing rigorous boundary conditions to simulations of discrete dislocation dynamics.
Modelling Simul. Mater. Sci. Eng. 7, 753–68 (1999)
P. Franciosi, M. Berveiller, A. Zaoui, Latent hardening in copper and aluminium single crystals. Acta Metall. 28, 273–283
(1980)
F. C. Frank, in Symposium on The Plastic Deformation of Crystalline Solids, The resultant content of dislocations in an
arbitrary intercrystalline boundary (Mellon Institute, Pittsburgh, 1950), pp. (NAVEXOS-P-834), 150
C. Fressengeas, V. Taupin, L. Capolungo, An elasto-plastic theory of dislocation and disclination fields. Int. J. Solids
Structures. 48, 3499–3509 (2011)
S. Gao, M. Fivel, A. Ma, A. Hartmaier, Influence of misfit stresses on dislocation glide in single crystal superalloys: A
three-dimensional discrete dislocation dynamics study. J. Mech. Phys. Solids. 76, 276–290 (2015)
N. M. Ghoniem, S. H. Tong, L. Z. Sun, Parametric dislocation dynamics: A thermodynamics-based approach to
investigations of mesoscopic plastic deformation. Phys. Rev. B. 61, 913–27 (2000)
J. P. Hirth, R. C. Pond, Steps, dislocations and disconnections as interface defects relating to structure and phase
transformations. Acta Mater. 44, 4749 (1996)
J. P. Hirth, R. C. Pond, J. Lothe, Disconnections in tilt walls. Acta Mater. 54, 4237 (2006)
J. P. Hirth, J. Wang, C. N. Tomé, Disconnections and other defects associated with twin interfaces. Progr. Mater. Sc. 83,
417–471 (2016)
R. G. Hoagland, T. E. Mitchell, J. P. Hirth, H. Kung, On the strengthening effects of interfaces in multilayer fee metallic
composites. Philos. Mag. a-Phys. Condens. Matter Struct. Defects Mech. Prop. 82, 643–664 (2002)
C. Hou, Z. Li, M. Huang, C. Ouyang, Cyclic Hardening Behavior of Polycrystals with Penetrable Grain Boundaries:
Two-Dimensional Discrete Dislocation Dynamics Simulation. Acta Mechanica Solida Sinica. 22, 295 (2009)
H. A. Khater, A. Serra, R. C. Pond, J. P. Hirth, The disconnection mechanism of coupled migration and shear at grain
boundaries. Acta Mater. 60, 2007 (2012)
A. L. Kolesnikova, M. Y. Gutkin, A. V. Proskura, N. F. Morozov, A. E. Romanov, Elastic fields of straight wedge disclinations
axially piercing bodies with spherical free surfaces. Int. J. Solids Structures. 99, 82–96 (2016)
L. Kubin, B. Devincre, T. Hoc, Modeling dislocation storage rates and mean free paths in face-centered cubic crystals. Acta
Mater. 56, 6040–6049 (2008)
C. Lemarchand, B. Devincre, L. P. Kubin, Homogenization method for a discrete-continuum simulation of dislocation
dynamics. J. Mech. Phys. Solids. 49, 1969–82 (2001)
Z. Li, C. Hou, M. Huang, C. Ouyang, Strengthening mechanism in micro-polycrystals with penetrable grain boundaries by
discrete dislocation dynamics simulation and Hall-Petch effect. Comput. Mater. Sci. 46, 1124–1134 (2009)
A. T. Lim, M. Haataja, W. Cai, D. J. Srolovitz, Stress-driven migration of simple low-angle mixed grain boundaries. Acta
Mater. 60, 1395–1407 (2012)
B. Liu, D. Raabe, P. Eisenlohr, F. Roters, A. Arsenlis, G. Hommes, Dislocation interactions and low-angle grain boundary
strengthening. Acta Mater. 59, 7125–7134 (2011)
B. Liu, P. Eisenlohr, F. Roters, D. Raabe, Simulation of dislocation penetration through a general low-angle grain boundary.
Acta Mater. 60, 5380–5390 (2012)
Y. Liu, N. Li, S. Shao, M. Gong, J. Wang, R. J. McCabe, Y. Jiang, C. N. Tomé, Characterizing the boundary lateral to the shear
direction of deformation twins in magnesium. Nat. Commun. 7, 11577 (2016)
R. Madec, B. Devincre, L. Kubin, T. Hoc, D. Rodney, The role of collinear interaction in dislocation-induced hardening.
Science. 301, 1879–1882 (2003)
J. C. Michel, H. Moulinec, P. Suquet, A computational scheme for linear and non-linear composites with arbitrary phase
contrast. Internat. J. Numer. Methods Engrg. 52, 139–160 (2001)
F. Mompiou, D. Caillard, M. Legros, Grain boundary shear-migration coupling–I. In situ TEM straining experiments in Al
polycrystals. Acta Materialia. 57, 2198 (2009)
B. Muntifering, L. Kovarik, N. D. Browning, R. C. Pond, W. B. Knowlton, P. Müllner, Stress-assisted removal of conjugation
boundaries in non-modulated Ni-Mn-Ga by coordinated secondary twinning. J. Mater. Sci. 51, 457–466 (2016)
T. Mura, Micromechanics of Defects in Solids. (Kluwer Academic Publishers, Dordrecht, 1987)
M. Murayama, J. M. Howe, H. Hidaka, S. Takaki, Atomic-level observation of disclination dipoles in mechanically milled,
nanocrystalline Fe. Science. 295, 2433 (2002)
P. G. Partridge, The crystallography and deformation modes of hexagonal close-packed metals. Metall. Rev. 12, 169–194
(1967)
Capolungo and Taupin Materials Theory (2019) 3:2 Page 21 of 21

G. Po, M. Lazar, D. Seif, N. Ghoniem, Singularity-free dislocation dynamics with strain gradient elasticity. J. Mech. Phys.
Solids. 68, 161–178 (2014)
L. Priester, Grain boundaries, from theory to engineering. Springer Ser. Mater. Sci. 172 (2013)
S. S. Quek, Z. Wu, Y. W. Zhang, D. J. Srolovitz, Polycrystal deformation in a discrete dislocation dynamics framework. Acta
Mater. 75, 92–105 (2014)
S. S. Quek, Z. H. Chooi, Z. Wu, Y. W. Zhang, D. J. Srolovitz, The inverse hall-petch relation in nanocrystalline metals: A
discrete dislocation dynamics analysis. J. Mech. Phys. Solids. 88, 252–266 (2016)
S. Queyreau, G. Monnet, B. Devincre, Slip systems interactions in α-iron determined by dislocation dynamics simulations.
Int. J. Plasticity. 25, 361–77 (2009)
A. Rajabzadeh, M. Legros, N. Combe, F. Mompiou, D. A. Molodov, Evidence of grain boundary dislocation step motion
associated to shear-coupled grain boundary migration. Phil. Mag. 93, 1299 (2013)
B. Reinholz, S. Brinckmann, A. Hartmaier, B. Muntifering, W. B. Knowlton, P. Müllner, Influence of the twin microstructure
on the mechanical properties in magnetic shape memory alloys. Acta Mater. 108, 197–206 (2016)
A. E. Romanov, A. L. Kolesnikova, Application of disclination concept to solid structures. Prog. Mat. Sci. 54, 740 (2009)
H. Rösner, C. Kübel, Y. Ivanisenko, L. Kurmanaeva, S. V. Divinski, M. Peterlechner, G. Wilde, Strain mapping of a triple
junction in nanocrystalline Pd. Acta Mater. 59, 7380 (2011)
D. Seif, G. Po, M. Mrovec, M. Lazar, C. Elsässer, P. Gumbsch, Atomistically enabled nonsingular anisotropic elastic
representation of near-core dislocation stress fields in α-iron. Phys. Rev. B. 91, 184102 (2015)
C. Sobie, L. Capolungo, D. L. McDowell, E. Martinez, Thermal activation of dislocations in large scale obstacle bypass.
J. Mech. Phys. Solids. 105, 150–160 (2017a)
C. Sobie, L. Capolungo, D. L. McDowell, E. Martinez, Scale transition using dislocation dynamics and the nudged elastic
band method. J. Mech. Phys. Solids. 105, 161–178 (2017b)
D. E. Spearot, M. D. Sangid, Insights on slip transmission at grain boundaries from atomistic simulations. Curr. Opinion
Solid State Mater. Sci. 18, 188–195 (2014)
X.-Y. Sun, V. Taupin, C. Fressengeas, P. Cordier, Continuous description of the atomic structure of grain boundaries using
dislocation and generalized-disclination density fields. Int. J. Plast. 77, 75 (2016)
A. P. Sutton, R. W. Baluffi, Interfaces in crystalline materials. (Clarendon Press, Oxford, 1995)
V. Taupin, L. Capolungo, C. Fressengeas, A. Das, M. Upadhyay, Grain boundary modeling using an elasto-plastic theory of
dislocation and disclination fields. J. Mech. Phys. Solids. 61, 370 (2013)
V. Taupin, L. Capolungo, C. Fressengeas, Disclination mediated plasticity in shear-coupled boundary migration. Int. J.
Plast. 53, 179–192 (2014)
V. Taupin, K. Gbemou, C. Fressengeas, L. Capolungo, Nonlocal elasticity tensors in dislocation and disclination cores. J.
Mech. Phys. Solids. 100, 62–84 (2017)
M. Upadhyay, L. Capolungo, V. Taupin, C. Fressengeas, Elastic constitutive laws for incompatible crystalline media: the
contributions of dislocations, disclinations and G-disclinations. Phil. Mag. 93, 794 (2013)
A. J. Vattré, M. J. Demkowicz, Determining the Burgers vectors and elastic strain energies of interface dislocation arrays
using anisotropic elasticity theory. Acta Mater. 61, 5172 (2013)
A. J. Vattré, B. Devincre, F. Feyel, R. Gatti, S. Groh, O. Jamond, A. Roos, Modelling crystal plasticity by 3D dislocation dynamics
and the finite element method: the discrete-continuous model revisited. J. Mech. Phys. Solids. 63, 491–505 (2014a)
A. J. Vattré, N. Abdolrahim, K. Kolluri, M. J. Demkowicz, Computational design of patterned interfaces using reduced order
models. Nature Sci. Reports. 4, 1 (2014b)
A. J. Vattré, M. J. Demkowicz, Partitioning of elastic distortions at a semicoherent heterophase interface between
anisotropic crystals. Acta Mater. 82, 234 (2015)
A. J. Vattré, Elastic strain relaxation in interfacial dislocation patterns: I. A parametric energy-based framework. J. Mech.
Phys. Solids. 105, 254–282 (2017a)
A. J. Vattré, Elastic strain relaxation in interfacial dislocation patterns: II. From long- and short-range interactions to local
reactions. J. Mech. Phys. Solids. 105, 283–305 (2017b)
Y. U. Wang, Y. M. Jin, A. M. Cutino, A. G. Khachaturyan, Nanoscale phase field microelasticity theory of dislocations: model
and 3D simulations. Acta Mater. 49, 1847–1857 (2001)
Y. M. Wang, E. Ma, Strain hardening, strain rate sensitivity, and ductility of nanostructured metals. Mat. Sci. Eng. A.
375–377, 46–52 (2004)
J. Wang, C. Zhou, I. J. Beyerlein, S. Shao, Modeling interface-dominated mechanical behavior of nanolayered crystalline
composites. J. Applied Mech. 66, 102–113 (2014)
B. Xu, L. Capolungo, D. Rodney, On the importance of prismatic/basal interface in the growth of (1012) twins in
hexagonal close packed crystals. Scripta Mater. 68, 901–904 (2013)
K. Yashiro, F. Kurose, Y. Nakashima, K. Kubo, Y. Tomita, H. M. Zbib, Discrete dislocation dynamics simulation of cutting of γ
precipitate and interfacial dislocation network in Ni-based superalloys. Int. J. Plasticity. 22, 713–723 (2006)
H. M. Zbib, M. Rhee, J. P. Hirth, On plastic deformation and the dynamics of 3D dislocations. Int. J. Mech. Sci. 40, 113–27
(1998)
Z. Zheng, D. S. Balint, F. P. E. Dunne, Investigation of slip transfer across HCP grain boundaries with application to cold
dwell facet fatigue. Acta Mater. 127, 43–53 (2017)
C. Zhou, R. LeSar, Dislocation dynamics simulations of plasticity in polycrystalline thin films. Int. J. Plasticity. 30-31,
185–201 (2012)