Meccanica

https://doi.org/10.1007/s11012-020-01209-y (0123456789().,-volV)
( 01234567
89().,-volV)

RECENT ADVANCES IN COMPUTATIONAL MECHANICS AND INNOVATIVE MATERIALS

Modeling the chemo-mechanical behavior of all-solid-state

batteries: a review.
Donald Bistri . Arman Afshar . Claudio V. Di Leo

Received: 1 February 2020 / Accepted: 5 July 2020

Ó Springer Nature B.V. 2020

Abstract Solid-state-batteries (SSBs) present a pertaining to Li-metal anodes and dendrite initiation
promising technology for next-generation batteries and growth mechanisms are reviewed, making the
due to their superior properties including increased transition from traditional liquid electrolyte anodes to
energy density, wider electrochemical window and next generation all-solid-state anodes. Second, chemo-
safer electrolyte design. Commercialization of SSBs, mechanics modeling of the SSE is reviewed with a
however, will depend on the resolution of a number of particular focus on the formation of a thermodynam-
critical chemical and mechanical stability issues. The ically unstable interphase layer at the anode/SSE
resolution of these issues will in turn depend heavily interface. Finally, we conclude with a review of
on our ability to accurately model these systems such chemo-mechanics modeling efforts for solid-state
that appropriate material selection, microstructure composite cathodes. For each of these critical areas
design, and operational parameters may be deter- in a SSB we conclude by highlighting the key open
mined. In this article we review the current state-of-the areas for future research as it pertains to modeling the
art modeling tools with a focus on chemo-mechanics. chemo-mechanical behavior of these systems.
Some of the key chemo-mechanical problems in SSBs
involve dendrite growth through the solid-state elec- Keywords Solid-state battery  Chemo-mechanics
trolyte (SSE), interphase formation at the anode/SSE modeling  Dendrite growth  Interphase instability 
interface, and damage/decohesion of the various Composite cathodes
phases in the solid-state composite cathode. These
mechanical processes in turn lead to capacity fade,
impedance increase, and short-circuit of the battery,
ultimately compromising safety and reliability. The 1 Introduction
article is divided into the three natural components of
an all-solid-state architecture. First, modeling efforts Solid-state-batteries (SSBs) present a promising tech-
nology for next-generation batteries. In recent years,
commercialization of SSBs has attracted significant
In honor of Professor J. N. Reddy for his 75th Birthday. research attention driven by their promise of superior
performance compared to conventional Lithium-ion
D. Bistri  A. Afshar  C. V. Di Leo (&) batteries (LIBs). Transition to SSBs has significant
School of Aerospace Engineering, Georgia Institute of
Technology, 270 Ferst Drive, Atlanta,
advantages including enhanced current density, faster
GA 30332, USA charging time, wider electrochemical window, and
e-mail: cvdileo@gatech.edu

123
 Meccanica

improved safety. Contrary to conventional LIBs which properties, successful integration of SSEs in SSBs
make use of graphite intercalation anodes, incorpora- requires meeting particular chemical standards,
tion of Li metal anode enables for a 10-time increase in including negligible electronic conductivity, high
volumetric capacity, leading to a significant improve- ionic conductivity, good electrochemical compatibil-
ment in energy density (3860mAh g1 ) [1–3]. Owing ity as well as several mechanical standards including
to their wider electrochemical window (0-5V), SSBs appropriate stiffness and fracture toughness to pre-
also enable for coupling of Li metal anode with high- serve mechanical integrity. While liquid electrolytes
voltage cathode materials as demonstrated by several traditionally have a higher ionic conductivity and
recent studies [4–6]. From a safety perspective, allow for excellent wetting of electrode surface,
replacement of toxic and flammable organic elec- comparable ionic conductivity values have been
trolytes with highly conductive (  103 S cm1 ), achieved for several ceramic SSEs as reported in
mechanically superior inorganic solid-state elec- multiple studies [10–15]. Resolution of sluggish ionic
trolytes (SSEs) can alleviate concerns associated with conductivity, while a major improvement in research
failure of batteries due to short-circuit or ignition efforts to integrate SSEs in next-generation battery
[7–9]. A particularly attractive SSB architecture with designs, addresses one of the multiple challenges
high energy density features a lithium metal anode, a hindering commercialization of SSBs.
composite solid-state cathode and an inorganic SSE Research efforts have recently shifted towards
separating the electrodes. Figure 1 highlights these developing a comprehensive, multi-physics under-
three particular areas of interest in an all-solid-state standing of the complex nature of the solid-solid
architecture as well as the mechanical modeling issues interfaces, as critical challenges associated with
which must be addressed in order to understand and electrochemical and mechanical stability of SSEs
commercialize these systems. exist there [16]. Among these issues, dendrite growth
Traditionally, SSEs are categorized into two constitutes one of the most intricate challenges in
groups, pertaining to organic solid polymers and SSBs, resorting to the interplay of both mechanical
inorganic solids (oxide or sulfide based). Among other and chemical forces. Lithium dendrite infiltration
through SSEs eventually leading to short-circuit of
battery has been vastly reported for a multitude of
representative SSEs [17–21]. Contrary to dendrite
growth processes in liquid electrolyte systems, growth
in SSBs depends on a fundamentally different mech-
anism hypothesized to be centered on morphology of
SSE surface. Internal defects (pores, surface cracks
and grain boundaries), inherent to the microstructural
composition of inorganic SSEs serve as sites for non-
uniform redistribution and infiltration of Li metal
leading to potential dendrite growth. From a mechan-
ics perspective, the process of Li dendrite growth
depends on microstructural mechanical features
including density of surface imperfections, defect-
size distribution, local stress-state and viscoplastic
deformation of the Li metal. This complex problem
thus requires a comprehensive chemo-mechanical
theoretical/computational framework [22].
Unwanted electrochemical reactions between the
Li-metal anode and SSEs during cyclic charge/
Fig. 1 Architecture of an all-solid-state battery composed of discharge of the battery can produce compounds
Li-Metal anode, solid-state electrolyte separator and composite
which lead to the formation of a nonhomogeneous
cathode. Highlighted under each of these three main compo-
nents are some of the key mechanics phenomena affecting the interphase layer of low ionic conductivity. This in turn
electrochemical performance of these systems can significantly limit ion transfer kinetics and

123
Meccanica

increase interfacial resistance as reported in several necessary framework for material characterization, be
representative SSEs [5, 23–26]. Simultaneously, for- it for active electrode particles or SSEs.
mation of nonhomogeneous interphase regions is Further, design of SSBs is inherently a multiscale
accompanied by volumetric expansion, which in the and multiphysics problem by nature. A complete
context of densely packed all-solid-state architectures framework for modeling SSBs requires integration of
can induce high stresses resulting in fracture and models beyond continuum-level. From an optimiza-
degradation of the SSE [27]. tion perspective, both continuum and first-principles
Composite cathodes, as illustrated in Fig. 1, also models are integral components in development of
present another area where mechanics plays a critical hierarchical, data science frameworks. Here, first-
role. Composite cathodes in SSBs rely on a densely principle calculations provide the necessary material
packed architecture of active material, surrounded by properties, while continuum level frameworks solve
a conductive SSE to support ionic transport, in the coupled chemo-mechanics problem. Optimal
addition to complementary additives for electronic design of SSBs involves optimization of both material
conduction. The active material undergoes volumetric and geometry simultaneously, which in turn calls for
expansion/contraction during insertion/extraction of an integrated use of both frameworks.
ionic species, which, due to the confined nature of the Significant efforts have been invested towards
all-solid-state architecture, can lead to generation of reviewing the chemo-mechanical behavior of SSBs
high stresses. Batteries employing liquid electrolytes in a series of past works from both an experimental and
isolate volumetric strains at the active particle level, computational standpoint. Early review works [30, 31]
given the incapacitation of liquid electrolytes to focused exclusively on prospects of SSBs and assess-
transfer stresses. As a result, in current battery designs, ment of potential SSE candidates with high ionic
any morphological changes of the active particles do conductivity. However, no extensive connection to
not tend to translate to interfacial delamination at the interface instabilities present in SSBs was made from a
electrode-electrolyte interface, which continues to modeling perspective. Lewis et al. [32] studied
support ion transfer. The confined nature of the interfaces in SSBs from an experimental standpoint.
solid-state electrode allows for transmission of A special focus of their review centered on the
stresses through the stiff solid framework resulting potential of employing advanced characterization
in a complex coupled chemo-mechanical behavior techniques ( eg. tomographic imaging, spectroscopy)
which is prone to damage through failure of the active aside from conventional in-situ characterization meth-
particles, the SSE itself, or delamination of the ods to study interfaces in SSBs. A review of existing
interface between the two [28, 29]. characterization techniques for study of SSBs as it
To overcome the aforementioned mechanical chal- applies to interphase formation and anode/cathode-
lenges we require an understanding of the interplay of SSE interface was also presented in the work of Xu
chemo-mechanical phenomena at cell constituents. et al. [33]. In-situ operando techniques were reviewed
The objective of this review article is to provide a and potential use of non-destructive characterization
comprehensive overview of the present theoretical and techniques for future studies of interfaces were
computational frameworks relevant to the study of proposed. A design perspective on better engineering
electro-chemo-mechanical phenomena in SSBs. How- of interfaces in SSBs was introduced in the work of
ever, certain aspects of chemo-mechanical modeling Shen et al. [34]. However, the scope of the review was
in SSBs (ex. interphase formation, dendrite growth) only concerned with instabilities at the anode/SSE
require a formal treatment, which extends beyond the interface and issues associated with damage in com-
realm of continuum mechanics. Therefore, where posite cathodes were omitted. A departure from a
necessary, our review invokes additional modeling purely experimental treatment of issues in SSBs
frameworks including both molecular dynamics (MD) towards a combined effort on reviewing of both
and first-principles calculations. While not continuum experimental and theoretical/computational efforts is
in nature, the aforementioned computational frame- presented in the work of Wang et al. [35]. Here, major
works give valuable insights for study of instabilities computational efforts in study of electrode-SSE inter-
in SSBs in addition to helping establish a much face are summarized. However, no insight is provided
on limitations of current frameworks and potential

123
 Meccanica

areas for improvement in future modeling efforts for electrolytes. However, as discussed briefly in the
study of SSBs. The review covers current modeling introduction and elaborated in past studies, successful
efforts on dendrite growth and damage of active design and operation of SSBs is hampered by several
particles and SSEs, but no connection from a modeling challenges [30–32, 34, 35, 38, 39]. One major obstacle
perspective is made to the phenomena of interphase is tied to unstable deposition of Li at the Li anode/SSE
formation and associated chemo-mechanical issues. A interface, which can lead to the onset of dendrite
review on joint efforts, both experimental and theo- formation and growth. Dendrite protrusions can initi-
retical in studying of SSBs is also presented in the ate at perturbations of the interface or microstructural
work of Zhang et al. [36]. This review focuses heterogeneities and subsequently grow through the
primarily on mechanical damage in SSBs as it pertains SSE under continuous cycling, leading to short-circuit
to active particles, SSEs and dendrite growth. Model- of the battery. It is critical thus to understand the
ing efforts reviewed in this work are limited to a select mechanisms of Li plating/stripping, uneven deposition
few and no identification of major limitations of under non-uniform interface kinetics, and propagation
current employed models is made from a theoretical of Li dendrites through the SSE. This section provides
perspective. Zhao et al. [37] focused their review on a review of the modeling efforts which have been
the transfer of already established chemo-mechanical developed to capture the mechanisms for dendrite
frameworks in LIBs to model the behavior of SSBs. initiation and growth. We begin by reviewing the
The scope of our review paper intends to provide a research done in this area as it pertains to Li dendrite
comprehensive analysis of current computational growth in liquid electrolyte systems as it is critical to
framework for modeling the chemo-mechanical understanding current research in SSBs.
behavior of SSBs as it pertains to all its components
including Li-metal anodes, SSEs and composite 2.1 Modeling of dendrite growth in liquid-
cathodes. In summarizing current modeling efforts, electrolyte systems
the focus of this work is to not merely showcase
current efforts at the forefront of modeling of issues in One of the first models of dendrite growth was
SSBs. Here special attention is paid to identifying key proposed by Barton and Bockris [40], who investi-
assumptions/limitations upon which current modeling gated electrolytic growth of spherical silver deposi-
framework builds. In doing so, future research avenues tions in silver-nitrate electrolytes. Their formulation
for computational efforts in multi-physics modeling of constitutes one of the early works in a series of efforts
SSBs are identified. Additionally, the authors propose which model dendrite growth under pure diffusion
available modeling tools from study of phenomena control, disregarding the effect of migration in trans-
which share similarities with aforementioned issues in port of ionic species. A first latent stage prior to
SSBs and can serve as starting points in the develop- dendrite growth is envisioned, during which protru-
ment of future modeling frameworks for addressing sions emerging from the electrode surface were
instabilities in SSBs. This paper is organized follow- chemically controlled by the characteristic diffusion
ing the natural design of a SSB as shown in Fig. 1. We layer of the substrate. Having penetrated the diffusion
review the relevant chemo-mechanical phenomena layer, dendrite growth followed, as marked by a
and modeling for Li-Metal anodes and dendrite transition from linear to spherical diffusion at the tip.
growth in Sect. 2, for solid-state electrolytes and Additionally, surface tension forces were incorporated
electrode/SSE interfaces in Sect. 3, and for active to regulate kinetics at the tip. This first proposed
materials and composite cathodes in Sect. 4. theoretical framework made use of linear kinetics
(hence applicable to small overpotentials) and did not
account for variations in concentration and potential
2 Lithium metal anode near the vicinity of the dendrite tip during growth.
Diggle et al. [41], improved this model by accounting
Use of inorganic SSEs with improved chemo-me- for non-linear Tafel interface kinetics (higher overpo-
chanical properties has the potential to alleviate failure tentials but below limiting current) to calculate the
of the cell due to short-circuit driven by dendrite growth rate of zinc dendrites in alkaline zincate
growth or ignition as may be encountered in liquid solutions as a function of overpotential, concentration

123
Meccanica

and temperature. Successive thinning of the dendrite [49], Pritzker and Fahidy [50] and Sundstrom and
tip enabled for spherical diffusion to dominate prop- Bark [51].
agation, confirming the previous findings by Barton Following the first observation of dendrite growth
and Bockris [40]. in lithium-polymer systems [52], a general framework
In contrast to the aforementioned theories assuming describing dendrite growth in a parallel-electrode
diffusion-limited ionic transport, migration-limited lithium-polymer cell during galvanostatic charging
mass transfer formulations of dendrite growth was proposed by Monroe and Newman [44]. A major
emerged with the work of Chazalviel [42] and later improvement in their formulation centered on addition
Brissot et al. [43]. These formulations investigated of well-defined thermodynamic points to capture the
metallic electrodepositions in a high electric field (’ variations in concentration and potential ahead of the
10 V), commensurate with cell currents above the dendrite tip. Additionally, the assumption of a well-
limiting current density, a domain in which elec- supported electrolyte was invalidated when solving for
troneutrality is invalid. The mechanism for dendrite the transient concentration and potential profiles in the
initiation was stipulated to be a direct outcome of electrolyte. Henceforth, both diffusion and migration
anion depletion in the vicinity of the electrode leading ion transfer mechanisms were incorporated in the
to a build up of space charge. This theoretical formulation. Their findings concluded that dendrites
formulation also helped the development of dendrite accelerated across the cell at all conditions, with cell
growth modeling, however it is of limited use in the shortening occurring at current densities above 75% of
study of rechargeable batteries owing to their much the limiting current. A few limitations of this model
lower current densities, as pointed out by Monroe and are the assumption of a single emerging dendrite and
Newman [44]. unvarying transport properties. To address the afore-
An important technique in assessing the stability of mentioned limitations, a simplified diffusion-reaction
electrodepositions, theoretical linear stability analysis mathematical model for dendritic growth with con-
emerged following the work of Mullins and Sekerka centration dependent diffusion coefficient following
[45, 46]. Morphological stability of a spherical particle Stewart and Newman [53] was later proposed by
under linear diffusion control was investigated upon Akolkar [54]. This formulation was further extended
application of an initial harmonic perturbation. Result- in a later work to incorporate the effect of temperature
ing deformation was then modeled as a linear super- variations on Li-ion transport [55]. An Arrhenius-type
position of spherical harmonics and time dependence temperature dependence of the diffusion coefficient
of expansion coefficients was assessed under the was adopted. Given an operating current density, a
interplay of both concentration gradient and surface critical temperature was evaluated below which den-
tension to judge stability. While initially specialized drite formation proceeded in an unrestrained fashion.
for spherical morphologies, the methodology was Noteworthy research efforts at this time also include
subsequently extended to non-spherical geometries. In the work of Ely and Garcia [56], who adopted a
doing so, assumption of an isotropic interface was thermodynamic framework to devise a five-stage
employed, thus disregarding any crystallographic mechanism of dendrite initiation and early growth
factors. Following Mullins and Sekerka, Aogaki and for reaction-rate limited systems undergoing hetero-
Makino [47] extended previous efforts and modeled geneous depositions. A critical thermodynamic and
the stability of electrode systems under diffusion- kinetic radius was identified for growth of electrode-
control non-steady depositions. Representing interfa- positions, below which emerging protrusions merely
cial deformation as superposition of a series of two- diffused in the electrolyte. Aside from theoretical
dimensional waves, deposition growth and particle efforts centered on understanding the mechanisms for
size were investigated as a function of variations in dendrite initiation and growth in liquid electrolyte
current density, bulk concentration and temperature. systems, modeling efforts with an emphasis on
Theoretical predictions were subsequently compared morphological evolution of dendrites have also been
against SEM crystal photographs [48]. Subsequent developed. Phase field models were extensively
early contributions in study of stability of metal adopted for such purpose to different levels of
electrodepositions followed the work of Barkey et al. complexity [57–62].

123
 Meccanica

A common feature of all aforementioned theoret- in which surface kinetics are solely regulated by
ical frameworks is that they relied on surface tension surface tension,
as being the only mechanism to counter amplification 0 a0
of surface roughness. As a result, mechanical contri- Dla;a 
e ¼ 2cV M K
ð2Þ
butions were not accounted for in modeling of Accounting for this, the impact of elastic deformation
electrodeposition kinetics at interface [63]. on deposition kinetics was assessed through a modi-
fied Butler-Volmer equation of the following form:
2.2 Modeling of dendrite growth in solid-state    
batteries ð1  aa ÞDle aa Fgs
in ¼i0;ref  exp exp 
RT RT
Modeling dendrite growth in SSBs requires a com-   ð3Þ
ac Fgs
prehensive understanding of the interplay of mechan- exp 
RT
ics and electrodeposition kinetics. First modeling
efforts in this area related to incorporating the role where F is the Faraday constant, T is temperature, R is
of mechanics on interfacial kinetics, without applica- the universal gas constant, i0;ref is the exchange current
tion to any particular problem. A seminal contribution density pertaining to an undeformed configuration, aa ,
in this area is the work of Monroe and Newman [63]. ac are anodic and cathodic charge transfer coefficients
An extended electrodeposition framework which and gs is the surface overpotential.
incorporated the effects of external pressure, elastic A more general framework for coupling mechanics
deformation and viscous stresses was proposed. to Butler-Volmer kinetics was developed by Ganser
Thermodynamic phases for proper description of the et al. [64]. In this work, the mechanical component of
interface were accordingly introduced. Starting with stress influencing the interfacial reaction need not be
balance of momentum and accounting for equilibrium the hydrostatic stress but rather a measure of resolved
conditions across the electrode-electrolyte interface, a stress in the direction of interfacial growth. Emergence
relationship for the electrochemical potential differ- of a theoretical framework to model the role of
ence (Dle ) induced by local deformations in terms of mechanics on electrodeposition kinetics created an
surface tension (c), pressure (Dp), deviatoric (Dsd ) and avenue for further modeling efforts of dendrite growth
viscous ðDsv Þ stress differences across the interface in SSBs.
was developed, Monroe and Newman applied their framework [63]
! to study stability of Li-solid polymer electrolyte
b0
VM þ VMzþ n
a0
a;a0 interfaces [65]. A linear elastic theory was employed
Dle ¼  2cK þ ean  ½en a  ðDsa;b
d
2zþ to assess how variation in electrolyte properties
! modeled as a linear elastic material affected roughness
o  a0  b0
V  V zþ 0 0
at the interface. Small amplitude periodic displace-
þ Dsa;b
v Þ þ
M M
ÞðDpa;a þ Dpb;b
2zþ ments were applied at the interface, and analytical
ð1Þ solutions for displacement distribution were after-
wards computed by solving for a steady-state equation
0
Here (a0 ; b ) represent undeformed states of electrode of motion with proper boundary conditions. Compres-
and electrolyte while (a; b) represent the roughened sive stress, deviatoric stress, and surface tension at the
electrode surface and its neighbouring electrolyte. interface were subsequently calculated and the con-
Additionally, K is the mean curvature of the tribution of each component to stability of the
roughened surface, VM , VM zþ are the molar volumes interface was assessed. While promoting interface
of reacting metal in electrode and cation in electrolyte, stability at all conditions, surface tension was demon-
zþ is the cation valence, while en is the unit normal strated to have a much smaller impact to electrode
vector at the electrode-electrolyte interface. stability compared to deviatoric and hydrostatic
In the limiting case where mechanical contributions stresses acting along the interface. A destabilizing
are neglected, their work correctly predicted the effect, increasing with an increase in shear modulus of
earlier model proposed by Barton and Bockris [40], the electrolyte was found for deviatoric stresses. In

123
Meccanica

contrast, compressive stresses had a stabilizing effect use of solid electrolytes as a mechanical barrier for
with increase in shear modulus of the solid polymer preventing dendrite growth. However, several exper-
electrolyte. Monroe and Newman’s critical finding is imental observations have shown that inorganic SSEs
that a region of stability for dendrite growth preven- still suffered from dendrite growth—despite satisfying
tion existed, granted the shear modulus of solid the condition of having a shear modulus at least twice
polymer electrolyte is at least twice larger than that larger than that of Li-metal—leading to cell short-
of Li-metal. Their findings, assessing the contribution circuiting [17–21].
of each factor to interfacial stability parameter are An alternative mechanism for initiation and prop-
shown in Fig. 2a. Here, a state of Dl\0, whith Dl - a agation of dendrites through SSEs was proposed by
stability parameter, represents stable electrodeposition Porz et al. [21], following their study of four SSEs
and helps in turn mitigate surface roughening. In subjected to different surface defect densities and
principle, these findings provided great promise for the internal microstructures. Contrary to the findings of

Fig. 2 Microstructural aspects of dendrite growth in SSBs. normal to the grain boundary planes. A 50% reduction in elastic
a Variation in contribution to interfacial stability of surface constant is obtained at the grain boundary relative to the bulk
tension, deformation forces and compressive forces as a SSE. Reproduced with permission from [66]. d Phase-field
function of shear modulus of the solid polymer electrolyte. simulation results on the effect of trapped excess electrons on
Reproduced with permission from [65]. b Interconnected pores Li-electrodeposition illustrating morphological differences
in the vicinity of grain boundaries, providing pathways for between an ideal excess electron-free electrolyte configuration,
dendrite growth in dense SSEs. Reproduced with permission and a real electrolyte configuration with trapped excess
from [17]. c Calculated elastic constant from MD simulations electrons at the grain boundaries. Reproduced with permission
performed on a LLZO electrolyte as a function of position from [67]

123
 Meccanica

Monroe and Newman, their observations suggested The role of microstructural features in promoting
that Li infiltration through the SSE was independent of uneven Li deposition and subsequent metal penetra-
the shear modulus of the electrolyte and proceeded tion through SSEs was also studied by Yu and Siegel
through a different mechanism than protrusions in [66]. They postulated that nanoscale softening of the
liquid electrolytes. Surface morphology, particularly SSE in the vicinity of grain boundaries provided a
defect size and density were identified as the main mechanism for preferential plating of Li, thus leading
factors for Li metal to plate in pre-existing imperfec- to dendrite growth. An important novelty of their work
tions which then drove crack propagation and dendrite was accounting for heterogeneities at the SSE
growth through the SSE. A representative illustration microstructure. Molecular dynamics simulations were
of propagation of dendrites through surface imperfec- performed on a representative LLZO electrolyte on
tions in dense garnet-type SSEs is shown in Fig. 2b two low-energy tilt and twist grain boundaries to
reproduced from the work of Ren et al. [17]. measure the elastic properties in both the bulk of the
Following their experimental observations, Porz electrolyte and near the grain boundaries. A 50%
et al. [21] developed a simplified one-dimensional, reduction in elastic modulus was observed locally at
chemo-mechanical formulation to model the stress the grain boundaries as compared to the bulk. Repre-
build up at the crack tip as a function of Li-plating sentative data of the variation in elastic constants in
overpotential DU through the relationship the bulk and near the grain boundaries generated from
F MD simulations is shown in Fig. 2c. Tian et. [67]
r0;max ¼  D/; ð4Þ adopted a multiscale framework coupling density
VmLi
functional theory (DFT) with a phase-field model to
where F is the Faradey constant and VLi m is the molar investigate Li dendrite initiation and growth inside
volume of Lithium. Critical stresses at the tip for crack SSEs with microstructural and internal imperfections.
advancement through the SSE were evaluated as a DFT calculations were employed on four representa-
function of defect size following a linear elastic tive SSEs to evaluate the density of electronic surface
fracture mechanics framework. This in turn enabled states along with distribution of excess electrons in
studying the variation in required overpotential for internal defects present in SSEs. As postulated in their
crack propagation with defect size. However, no earlier work, existence of electronic surface states
formal treatment of the role of mechanics on elec- served as a trapping mechanism for excess electrons,
trodeposition kinetics is presented in this work. thus enabling the nucleation of Li metal depositions at
More recently, Klinsmann et al. have modeled surface defects [69]. Concentrations of surface excess
advancement of a lithiated crack adjacent to a Li metal electrons obtained from DFT results were then incor-
negative electrode [68] with a more complete contin- porated in a phase-field model to study morphology of
uum-mechanics framework. Conditions under which dendrite evolution. Representative results from their
pressure build up inside the crack due to Li plating can work are shown in Fig. 2d. An increased intergranular
either promote unstable crack propagation or hinder it dendrite growth rate and penetration depth was
by blocking lithium insertion were explored. The observed due to an increased electrochemical driving
critical pressure required for fracture propagation was force from excess surface electrons. Additionally, the
subsequently translated to an equivalent surface presence of trapped excess electrons in SSE defects
overpotential regime in which safe operation of the leads to an increased probability of isolated nucle-
battery is ensured, as described by the following ation. Their findings demonstrated a dominant effect
relationship of internal microstructural imperfections and their
electronic conducting properties (band gap, distribu-
KC XM
jgj  gc ¼ pffiffiffiffiffiffiffiffiffi : ð5Þ tion of excess electrons) on dendrite growth as
1:12F ðpaÞ opposed to shear modulus and fracture energy in the
Here gc defines the critical magnitude of the surface bulk of the SSE.
overpotential, KC is the fracture toughness of the solid Parallel to these efforts, studies have been per-
electrolyte, a is the initial crack size and XM is the formed on the role of mechanics on electrodeposition
molar volume of metal. kinetics. Earlier linear perturbation analysis in a Li-
solid polymer electrolyte by Monroe and Newman

123
Meccanica

[65] was followed by a series of works, which Ahmad and Viswanathan later postulated a general
employed a similar technique to study stability of stability criteria for electrodeposition at isotropic
electrodepositions at an interface. Barai et al. [70] solid-solid interfaces [71]. A major contribution of
considered an analogous morphological domain for their work was extending the previous formulation by
dendrite initiation as Monroe and Newman, but Monroe and Newman to account for the effect of
revised the initial configuration of Li metal to be density changes of Li metal on the stability of
stress free, removing externally applied tractions at electrodeposition at the interface. Their work special-
interface. This enabled for modeling of early dendrite ized the earlier theory of Monroe and Newman [63]
growth behavior from a relaxed state. A small-strain (refer to Eq. (1) for variable definitions)
finite element procedure was adopted to solve the 1   

mechanical equilibrium equation with a compressive Dle ¼  VM þ VM zþ  cK þ en  sad  sbd en
2z
distributed load on the electrolyte’s side applied as a
1  0 0
boundary condition. This allowed for more accurate þ VM  VM zþ Dpa;a þ Dpb;b
modeling of the compressive stresses induced in both 2z
electrode and electrolyte by new phases of Li depo- ð6Þ
sition. In contrast to the classical finding of Monroe and introduced an extra parameter, the volume ratio
and Newman, Barai et al. determined that an initial V
(m ¼ VMMzþ ) to formally account for the effect of density
relaxed state of Li prevented dendrite growth at low
current densities regardless of the shear modulus of the changes of Li in the electrode versus the SSE. The
electrolyte. They proposed current distribution in the stabilizing or destabilizing nature of the hydrostatic
electrolyte as a dominant factor for dendrite growth. A term was investigated as a function of the aforemen-
comparison of the stability window with variation in tioned parameter. Following this, a density-driven
shear modulus of solid electrolyte for both pre-stressed stability mechanism was postulated, in which case Li
and relaxed initial Li states is shown in Fig. 3a. A in solid electrolyte is more densely packed than Li in
more detailed framework of growth of electrode/SSE metal electrode. This provided an extension to the
interface in the presence of an elastic prestress was earlier work of Monroe and Newman [65] specialized
formulated by Natsiavas et al. [73]. Interfacial growth for solid-polymer electrolytes in which hydrostatic
was mathematically modeled under the interplay of pressure has a dominant effect over density changes on
numerous processes including the kinetics of ionic stability of electrodepositions. The interplay of both
transport at both SSE and interface region, elastic mechanisms for interfacial stability was investigated
stress due to built-in prestress and volumetric changes and a two-parameter stability plot including shear
in SSE, and electro-mechanical interactions captured modulus ratio and molar volume ratio was generated.
by Maxwell stress. A linear stability analysis was Two stability regimes were identified including a
employed, specialized for planar growth in a Li- pressure dominated regime at high molar volume ratio
LiPON system and conditions for stable electrodepo- (m [ 1), and a density-driven regime at low molar
sition were studied as a function of stress state, elastic volume ratio (m\1), see Fig. 3b. Resistance to den-
modulus, deposition kinetics and interfacial energy. drite growth of representative SSEs was assessed
The presence of prestress was determined to have a based on their position in the stability diagram.
stabilizing effect on the electrode-electrolyte inter- Contrary to most previous studies, given that conven-
face, enabling for a wider window of stable perturba- tional inorganic SSEs have a low molar volume, softer
tion wavenumbers (k) as compared to a state of no SSEs were determined to be better at stabilizing
prestress. This is schematically illustrated in Fig. 3d dendrite growth. However all linear stability analyses
produced from [73]. Simulations employing a finite studied at this point in time resorted to simplification
difference scheme also confirmed the advantageous of an isotropic elastic interface. Ahmad and Viswa-
effect of prestress on stability of electrodeposition nathan extended their previous study to account for
growth, as demonstrated by significant reduction in material anisotropy and assess its contribution to
amplitude of an initial perturbation upon application stability of electrodeposition at solid-solid interfaces
of a prestress, see Fig. 3e. [72]. A continuum mechanics framework was devel-
oped to solve for two-dimensional problems in

123
 Meccanica

Fig. 3 Effect of mechanics on stability of electrodeposition controlled stability regime, while at high molar volumes (top-
kinetics. a Effective current density at the peak over the valley right) there is pressure-controlled stability regime. Reproduced
of dendritic protrusions plotted as a function of shear modulus of with permission from [71]. c Stability diagram for anisotropic Li
the solid electrolyte for both an initial relaxed Li-metal and a anode-isotropic electrolyte system, demonstrating the variabil-
pre-stressed Li-metal. At low-current densities, an initially ity in stability regions with crystallographic orientation of Li-
relaxed Li-metal is found to prevent dendrite growth irrespec- metal. Reproduced with permission from [72]. d Effect of
tive of the effective shear ratio. Reproduced with permission prestress on stability window of critical wavenumber of initial
from [70]. b Stability diagram for an isotropic Li anode/SSE perturbation compared against a system with no prestress.
interface with two regions of stable electrodepositions shaded in e Stabilization of an initial perturbation upon application of a
green. At low molar volumes (bottom-left) there is a density- prestress. Reproduced with permission from [73]

anisotropic linear elasticity in terms of a crystallo- leads to the emergence of locally non-uniform
graphic orientation-dependent elastic tensor. Similar stresses, however an average vanishing stress at the
to the case of an isotropic interface, two stability interface is maintained in the absence of external
domains were identified for an anisotropic electrode- tractions. An Airy stress function was employed to
isotropic electrolyte interface combination. However, derive the stress distribution at the interface, and
in the latter, critical shear modulus curves contouring results were incorporated in the theory of Ganser et al.
regions of stability varied depending on the crystal- [64] specialized for a single ion conducting solid
lographic orientation of Li-metal, see Fig. 3c. electrolyte to capture the effect of mechanics on
Recently, McMeeking et al. [74] performed a linear electrodeposition kinetics. Their findings also contra-
perturbation analysis to investigate surface roughen- dict the original model by Monroe and Newman,
ing of metal electrodes during plating. Their study demonstrating that long wavelength protrusions (L)
employed an interfacial morphology consisting of a always amplify regardless of the stiffness of the SSE,
sinusoidal (amplitude of the peaks negligible to as predicted by the following stability criterion
wavelength) stress-free electrode profile and a flat,
undeformed electrolyte configuration. Bonding of the
electrode to the electrolyte and successive plating

123
Meccanica

2 3
polymer electrolytes promoted stability of dendrite
6 ð1  2dectrode ÞI7 protrusions even at shear modulus two orders of
8pGC XM 41  qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 5½GM ð1  mC Þ þ GC ð1  mM Þ
ð2io Þ2 þ I2 magnitude lower than that of Li metal. The effect of
increasing the yield strength on the stability of
 FqIL½GC ð5  6mM  6mC þ 8mM mC Þ þ GM ð3  4mC Þ electrodeposition is shown in Fig. 4b. Recently,
ð7Þ Narayan and Anand [22] investigated dendrite growth
through the use of a large deformation isotropic
where GC , mC , GM , mM denote the shear modulus and
elastic-viscoplastic framework calibrated to the
Poisson’s ratio of electrolyte and electrode, XM is the
microindentation experiments of Wang and Cheng
metal molar volume, I is the current density, io is a
[76]. Figure 4c shows load versus indentation pre-
reference exchange current density, q is resistivity of
dicted by the calibrated model (solid lines) versus the
electrolyte and detrode is an influence coefficient which
experimental measurements (dashed lines) at various
captures the effect of electrode energy state on redox
indentation rates. Their formulation builds upon a
reaction energy barrier. Similar conclusions for
simple-power law to describe the effective plastic
amplification of surface roughness irrespective of the
shear strain-rate _p of the form
stiffness of the SSE were derived provided the current
density (I) is sufficiently large.  m1
r
A common critical feature of all of the aforemen- _p ¼ _0 ð8Þ
S
tioned modeling efforts is that they rely on an elastic
interface. The Li metal anode behaves plastically (and with flow resistance (S) evolving according to the
viscoplastically) and has a low homologous temper- following relationship
ature. Its plastic response has been the recent focus of a   
S a p
number of experimental investigations [76–81]. S_ ¼ H0 1  _ : ð9Þ
Ss
Development of a theoretical framework to incorpo-
rate the effect of plastic deformation on electrodepo- Here _0 denotes a reference strain rate, m is a strain
sition kinetics and dendrite growth spurred a new rate sensitity parameter, r is an equivalent tensile
avenue in computational research efforts. Early con- stress, while fH0 ; Ss ; ag are strain hardening param-
tributions towards this goal emerged with the work of eters with Ss defining a saturation flow strength. One
Ferrese and Newman [82]. Mechanical deformation of major assumption employed here resorted to neglect-
a Li-metal anode adjacent to a stiff solid-polymer ing dependence of saturation flow strength on loading
electrolyte was studied under a framework of elastic- rate.
plastic deformation. An isotropic-hardening J2 plas- Two representative case studies of interest for
ticity model was adopted utilizing the experimental modeling of SSBs were investigated including flat-
stress-strain data of Li metal by Schultz [81]. How- tening of an asperity in Li metal due to compressive
ever, viscoplastic effects were neglected. Similarly, loading from the bounding SSE, and infiltration of Li
Barai et al. [83] developed a computational framework metal into an imperfection on the surface of the SSE.
coupling mechanics, charge balance and mass con- Equivalent plastic strain and stress contours for the
servation equations to model stress evolution in both case of a pre-existing imperfection at the Li anode-
Li metal and polymer solid electrolytes. A nonlinear SSE interface undergoing compression from the
hardening law was adopted to account for plastic confining SSE are shown in Fig. 4d. Anand and
deformation. Plastic flow of Li metal was found to Narayan recently revisited their formulation [84]
promote stable deposition at interface, as evidenced by following the experiments of LePage et al. [78] who
a reduction in depth of penetration of dendrite provided stress strain data on polycrystalline Li foils.
protrusions as shown in Fig. 4a. This in turn enabled Here, earlier assumption of a constant saturation flow
slower electrodeposition kinetics at the tip relative to strength [refer to Eq. (9)] was modified to account for
the base of dendritic outgrowths. The effect of the experimentally observed strain hardening behavior
variation in yield strength of solid polymer elec- of Li metal at high loading rates,
trolytes on dendrite growth was assessed. Barai et al.
showed that increasing the yield strength in solid

123
 Meccanica

Fig. 4 Plastic modeling of Li metal and its role in dendrite Reproduced with permission from [75]. c Comparison between
growth. a Variation in amplitude of dendritic protrusions with numerically simulated load vs. indentation depth and experi-
shear modulus of solid electrolyte. For high shear modulus mental results by Wang and Cheng at different loading rates.
electrolytes, plastic deformation has a stabilizing effect, d Simulation domain for modeling lithium infiltration in a pre-
assisting in diminishing the height of dendritic protrusions. existing crack at the anode/SSE interface. Contour plots of
b Effect of increasing the yield strength of solid electrolyte on equivalent plastic strain and induced stress due to compression
stability of electrodeposition, captured through the variation in of the SSE onto the Li anode. Reproduced with permission from
ratio of current density at the peak over the valley of dendritic [22]
protrusions as a function shear modulus of electrolyte.

" #n
2.3 Summary on modeling Li metal anodes
 _p
Ss ¼ S Q
ð10Þ and dendrite growth
Aexpð RT Þ

where Ss denotes a saturation flow strength at a As reviewed above, understanding Li metal anodes
particular strain rate, S* and n are strain hardening and the process of dendrite formation and growth is
parameters, Q is an activation energy, A is a pre- critical in enabling the design of an all-solid-state
exponential factor, R is the gas constant and T is architecture which suppresses dendrite growth. We list
temperature. To validate their theory, simulations of here a few key areas of research the authors feel are of
inhomogeneous tension and compression experiments pressing relevance:
were conducted and compared against experimental – At present, modeling efforts on dendrite growth
results by Masias et al. [79], demonstrating consistent have resorted to a linear elastic, isotropic interface,
results. with modeling efforts which capture viscoplastic
deformation of Li metal and effect of plastic flow
on electrodeposition kinetics still at their infancy.

123
Meccanica

An experimentally validated constitutive theory 3.1 Interphase induced stress and fracture
for Li metal is required for modeling the manner in
which stresses develop near the dendrite growth As presented in the work of Lewis et al. [32], the
region and subsequently how these couple to the unwanted interphase has to be ionically conductive but
kinetics of growth. electronically insulating for the battery to operate
– While several theoretical studies on the interplay without failure, as an electronically conductive inter-
of chemistry and mechanics on electrodeposition phase allows for uncontrolled reaction of the pristine
kinetics of Li metal have been proposed, from a SSE and hence unstable growth of this undesirable
modeling perspective, a fully coupled chemo- mixed phase. In practical applications, electronically
mechanical theory for modeling of dendrite initi- insulating interphases form, however the low ionic
ation and growth within the solid-state electrolyte conductivity of the new mixture leads to high resis-
is still missing. tance of the cell, proving to be a challenging aspect of
– Development of a microstructure-resolved SSBs realization [24–26]. In addition to the undesir-
description of the process of dendrite growth able low ionic conductivity of the interphase, there are
through the SSE microstructure is also lacking critical mechanical issues accompanying its forma-
and necessary. In particular, this model should tion, particularly the volume expansion associated
include considerations for the presence of SSE with the reaction that causes the new phase. Occurring
defects, cracks, and grain boundaries which can in the confined all-solid-state architecture, this volume
serve as nucleation and growth sites for Li expansion introduces internal stresses, leading to
dendrites. possible fracture and degradation. Tippens et al. [87]
showed the formation of cracks in
Li1:4 Al0:4 Ge1:6 (PO4 )3 (LAGP) solid electrolyte using
X-ray tomography, interestingly arguing that it is the
3 Solid state electrolytes
presence of these cracks and the subsequent loss of
contact that cause high impedance in the battery
As overviewed in Sect. 1 and also discussed within the
operation, and not the transport properties of the
context of Li-metal anodes in Sect 2, solid-state
interphase itself. From a modeling perspective, the
electrolytes (SSEs) are a critical component in
same team also studied the stress distribution in the
enabling all-solid-state battery architectures. Pre-
aforementioned solid state cell using the chemo-
sented here is a review of the literature addressing
mechanics continuum theory of Di Leo et al. [88]. The
mechanics (and chemo-mechanics) phenomena in
finite element model and contours of reaction coordi-
SSEs with a particular focus on phenomena associated
nate are shown in Fig. 5b which by design agree well
with the formation and development of the interphase
with the experimental observation shown in Fig. 5a.
between the Li-metal anode and the electrolyte. As
From Tippens et al. [87], Fig. 5d illustrates the
will become evident in this section, research as it
evolution of a three-dimensional fracture network as
pertains to mechanics of SSEs is in its infancy.
the interphase forms during cell operation, along with
The interphase which forms at the anode/SSE
a proportional increase in impedance depicted in
interface is thermodynamically unstable, causing the
Fig. 5e. It should be noted, however, that the model
SSE itself to go under oxidation and reduction during
used to simulate the interphase formation and evolu-
charge and discharge [5, 6, 26, 85, 86]. Figure 5a,
tion relied on an ad-hoc description of the interphase
shows the formation of this interphase due to chemical
where its position was determined a-priori rather than
reactions. To summarize the current research in SSE
evolved according to the chemo-mechanical physics
behavior we divide this section into: (1) research
of the system. At present, capturing the phenomena of
addressing interphase formation and associated
interphase formation and evolution remains one of the
mechanical effects, (2) mechanical properties and
challenges for SSE realization from a modeling
failure of the SSE itself, and (3) delamination and
perspective. Other continuum formulations in
damage of the interface.
chemo-mechanics of solid interphases, such as the
one by Rejovitzky et al. [89], are more suited for
modeling of growth of solid interphase in

123
 Meccanica

Fig. 5 Mechanical effects of reaction induced interphase increase in the impedance. a–e are reproduced with permission
growth. a Interphase formation at the interface of Li anode from [87]. f Phase field model of fracture growth in an active
and LAGP SSE. b Finite element model of the swollen electrode LiMn2 O4 particle, starting from left to right: model
interphase showing contours of reaction coordinate. c Stress geometry, stress distribution and crack pattern, respectively.
distribution in the cell, leading to both radial and circumferential Reproduced with permission from [98]. g Delamination of a
cracks. d Network of extensive crack propagation in the SSE. SSE interface resulting in poor contact. Reproduced with
e Direct correlation between extent of crack propagation and permission from [99]

conventional LIBs, as opposed to structural and phase [95–97] might provide a potential starting point for
changes associated with interphase formation at the building more specific models of interface reactions in
anode-electrolyte interface in SSBs; treatment of SSBs. These models are particularly useful given the
growth models in LIBs are fundamentally different framework for oxidation/corrosion phenomena build
from a reaction model relevant to SSBs. There are, upon similar governing equations, in that species
however, models in the mechanics community from balance law obeys a diffusion reaction equation of the
which useful formulations for interphase growth can form
be drawn as they describe physical phenomena that
c_ ¼ mr2 l  R ð11Þ
share similarities with interface reactions in SSBs.
Particularly, the research conducted on modeling where fc; m; lg denote concentration, mobility and
oxidation in thermal barrier coatings [90–94], and on chemical potential. Here, R defines the species con-
modeling hydrogen assisted cracking in metals sumption rate, specific to the type of chemical

123
Meccanica

reaction, be it oxidation, SSE interphase formation, literature for fracture in SSEs. However, there are a
etc. number of studies dedicated to modeling of crack
A major difference in these existing formulations to propagation in electrode particles, which can be
modeling phenomena in SSBs is in the distinct adopted as a starting point for addressing chemo-
chemical driving forces controlling each reaction. mechanics of SSEs in the presence of cracks. The
Hence, enhancing current frameworks with models literature on modeling fracture of ceramics might also
specific to reaction kinetics of interphase formation in serve as a relevant foundation for the development of
SSEs presents an appealing avenue for future efforts. modeling techniques specific to SSEs [104–112].
To capture the bulk mechanical behavior of SSEs, Although the physical mechanisms governing crack
Ganser et al. [100] proposed a thermodynamically formation in SSEs are likely to differ from that of
consistent theory of transport in a host material using a active particles, in which diffusion and reaction
Helmholtz energy formulation which accounts for induced stresses govern, the methods developed for
species transport, elastic deformation and electrostat- active particles should serve as a suitable starting point
ics. They extended the concentrated solution theory of for modeling crack formation in SSEs.
Newman and Thomas-Alyea [101] to incorporate the Phase-field damage models have been developed
effects of mechanics, and then specialized their for damage in active particles. These models can
formulation to the case of binary electrolytes. Effect incorporate the crack degradation parameter as an
of deformation and mechanical stresses on reaction additional internal variable to the chemical variable
kinetics and ionic transport were assessed upon set, taking into account both its initiation and propa-
solving the coupled problem. Part of their study also gation. Particularly, Miehe et al. [98] presented a finite
focused on capturing the role of interfacial morpho- strain chemo-elasticity theory coupled with phase field
logical change on concentration profile and its poten- modeling to investigate fracture in Li-ion electrode
tial association with dendrite formation. A similar particles. Their regularized crack model uses a prin-
thermodynamically consistent treatment is also pre- cipal stress criterion for failure surface, and a degra-
sented in the work of Bucci et al. [102] who modeled dation function to ‘‘damage’’ both the stored energy
ionic transport in crystal lattices. Noteworthy to and the mobility of species across the crack faces.
mention here is also the analytical work of Grazioli They also regularized reaction kinetics at the fracture
et al. [103] on solid polymer electrolytes that surfaces, replacing it a with a source term in the
addressed the effects of mechanics. conservation equations. Using FEM to solve the
The aforementioned models provide a foundation resulting partial differential equations, they demon-
for modeling ionic transport in solids, however, these strated the versatility of the phase field method in
works, including those by the authors, neglect the simulating crack propagation in chemo-mechanics
fundamental physics of interphase formation and problems. Figure 5f shows a representative three-
growth, that is the reaction induced structural phase dimensional simulation of fracture in a coupled
transformations and the associated coupling to chemo-mechanics framework. Zhang et al. [113]
mechanical deformation and stress. It is the reaction extended these works to model diffusion induced
induced mechanical effects of the newly formed plasticity and fracture in both amorphous and crys-
interphase, and its associated chemical and structural talline silicone (Si) electrodes using a similar varia-
phase changes, which cause mechanical degradation tional based phase field damage model. Employing an
in the SSE [87]. A diffusion deformation reaction efficient time integration scheme, they successfully
theory capable of accommodating chemically induced simulated fracture of Si nano-particles and nano-
phase transformation is currently lacking in the pillars and investigated how energy release rate
mechanics literature and is a potential avenue for affected the electrode damage behavior. A similar
potential future research. work on effects of fracture propagation on mechanics
In addition to interphase formation induced stresses and subsequently on Li concentration and phase
in SSEs and their effect on driving forces for diffusion separation in polycrystalline particles of LiCoO2 was
and reaction, a critical mechanism for SSE failure is conducted by Yamakawa et al. [114], in which they
the formation and propagation of cracks. At present, estimated the effect of grain size and crystal orienta-
the authors could not identify an existing model in the tion on the extent of fracture. In Anand et al. [97] the

123
 Meccanica

Table 1 Compounds Compound Family

studied in Deng et al. [116]
using DFT to obtain the LiTi2 (PO4 )3 , NaZr2 (PO4 )3 NASICON
elasticity tensor
Li3 PO4 Phosphate
Li1=2 La1=2 TiO3 , Li1=8 La5=8 TiO3 Perovskite
Li7 La3 Zr2 O12 (LIZO), Li5 La3 M2 O12 Garnet
Li3 OCl, Li3 OBr, Na3 OCl, Na3 OBr Antiperovskite
Li3 PS4 , Li10 GeP2 S12 (LGPS), Li7 P3 S11 , Na3 PS4 Thiophosphate

authors developed an alternative thermodynamics fracture toughness and high stiffness one is better at
based damage model, in contrast to the aforemen- resisting lithium dendrite growth [116].
tioned variational based phase-field approaches. As the scope of this review is concerned with
Available phase-field models for treatment of modeling, here we summarize the reported mechanical
chemo-mechanical fracture in LIBs are transferable properties for a range of SSEs, all of which are
to address the reaction induced crack propagation in obtained from first-principle calculations. The results
SSE. Damage evolution in current phase-field models of these Ab-initio estimations can be used as inputs to
is governed by an equation of the form continuum models to optimize SSE designs, poten-
tially in a hierarchical multi-scale, multi-physics
g_ ¼ ð1  dÞH  ½d  l2 r2 d ð12Þ
paradigm. Deng et al. [116] used density functional
where fg; d; lg denote the viscous regularization theory (DFT) with projected augmented wave method
parameter, the damage variable and the characteristic [117] for ions interaction, to obtain a complete
length scale. Here H defines the damage driving force description of the elasticity tensor for the compounds
which depends on the underlying material and chem- shown in Table 1. They reported the following
istry of interest. Both stress and energy based criteria hierarchy of elastic moduli: thiophosphate \ antiper-
have been used in the literature to formulate H, and ovskite \ phosphate \NASICON \ garnet \ per-
may include elastic, inelastic, or chemo-mechanical ovskite. It was also determined that Na counterparts of
coupling terms. Damage evolution in SSEs can be the studied SSEs had smaller Young moduli compared
captured by a form similar to Eq. (12) provided the to the aforementioned Li compounds. From a
constitutive framework for damage driving force is microstructural standpoint, it was determined that
specialized to account for effect of reaction in addition tetragonal structures exhibited a higher elastic mod-
to capturing the specific fracture behavior of the SSE ulus compared to cubic ones among garnet based
material. SSEs. Regarding brittle vs ductile behavior, the DFT
calculations revealed thiophosphates as the most
3.2 Mechanical characteristics of solid-state ductile, with oxides being the most brittle compounds.
electrolytes It should be noted, however, that the employed metric
in assessing ductile vs brittle behavior resided solely
In the process of developing an understanding of the on simple shear over bulk ratio or Pugh’s number
mechanical behavior of SSEs, it is necessary to [118], and not on a complete stress strain constitutive
understand and model the mechanical behavior and curve.
properties of the SSE itself, as the elastic, plastic and Wu et al. [119] adopted a similar technique for
fracture properties of the electrolyte will determine characterization of antiperovskite Li3 OCl material.
how it responds to dendrite growth, induced stresses Upon calculation of three elastic constants of the cubic
due to interphase formation, crack propagation and material, it was determined that Li3 OCl is softer than
other relevant mechanical issues [115]. For instance, a LLTO and LIZO, but harder than LGPS. They also
ductile electrolyte is more desirable for accommodat- calculated a Pugh’s ratio of 1.35, indicating its
ing deformation and maintaining contact with the brittleness. As mentioned in their work, this inherent
lithium anode during battery operation, while a high brittleness makes this material susceptible to inter-
phase induced fracture during cell operation and can

123
Meccanica

severely limit its potential integration in SSBs as a computational complexity and uncertainties associ-
practical SSE. Their reported value of 0.22 for ated with quantum mechanical calculations. It has
Poisson’s ratio is also typical of ionic crystals [120]. been shown that mechanical phenomena during bat-
In a similar fashion, Wang et al. [121] studied LGPS tery operations heavily depend on the choice of the
along with Li4 P2 S7 and Li3 P7 S11 , which unlike the SSE material. While DFT simulations can help
tetragonal structure of the LGPS, have triclinic overcome the need for complex experimental charac-
formations. They reported that LGPS has higher terizations, they are still relatively time consuming
elastic constants compared to the two aforementioned given the vast combination of potential SSE com-
LISICON compounds. They also calculated a Pugh’s pounds. A possible avenue to accelerate the mechan-
number of 2.12 for LGPS, indicating its relative ical design of SSEs is to take advantage of data science
ductile nature. A detailed study on two b and c phases approaches, which by now have reached a relatively
of Li3 PS4 was performed by Yang et al. [122], mature state in the material science community
demonstrating that the c phase is a stiffer material [125–130]. This particular approach has previously
than its b counterpart. Both materials also showed been used for generic liquid electrolyte battery design
relative ductility, characterized by Pugh’s ratio of 2.26 [131, 132], but Ahmad et al. [133] recently extended
and 2.89 for c and b phases, respectively. Similarly, this work by employing machine learning techniques
compounds of Metal Borohydrides M2 B10 H10 (with M to scan a space of over 12000 inorganic SSE
being the place holder for different ions of Li, Na, K) compounds, identifying the ones that have the best
were studied in the work of Chen and Hong [123]. mechanical properties to suppress dendrite growth. It
Their study illustrated the Li compound to be stronger should be noted, however, that their data was only
that the sodium and potassium counterpart, observing trained on structural features of the compounds and
a decrease in stiffness as the ion becomes larger. While they considered elastic properties of the SSE to be the
referred works resort purely to material properties only dominant factor for dendrite suppression. As
obtained by first-principle calculations, experimental reviewed in Sect. 2, the phenomena of dendrite
efforts have also invested in characterizing properties growth is complex and governed not just by the
of SSEs. A review of such efforts has been provided in choice of elastic properties of the SSE. Addressing this
tabulated form by Wang et al. [35] and Zhang et al. phenomenon requires a continuum chemo-mechanical
[36] and are not replicated for conciseness in this model that accounts for microstructure, defects,
review. geometry, boundary conditions, etc. Adding the
Finally, Ahmad and Viswanathan [124] realized the results acquired from fully coupled chemo-mechani-
uncertainties associated with calculating energies cal continuum models as features to the training space
around the equilibrium state in DFT calculations, of these data science approaches constitutes one area
improving the previous works by quantifying this for potential improvement in pursuit of robust
uncertainty. They performed the calculations on an mechanical designs of SSBs.
ensemble of energies on four compounds of
Li10 GeP2 S12 , Na3 PS4 , LiI, and Na3 OCl, obtaining 3.3 Mechanical compatibility at the interface
probability distributions for mechanical properties
with quantified standard deviations. These probabilis- We conclude this section with another important
tic results can be put into a continuum model, leading mechanical problem in SSEs, namely void formation
to probabilistic design of SSBs. Also equally impor- and loss of contact at the lithium anode-SSE interface
tant to consider and currently lacking in the DFT [35, 99]. This phenomenon is depicted in Fig. 5g from
literature is the mechanical properties of the reacted the experimental work of Han et al. [99] on improving
interphase, as it seems natural to expect the properties resistance in LLCZN garnet based SSEs. This problem
of this mixed phase to be different from the bulk is specific to SSBs as traditional liquid electrolyte cells
unreacted material. Future DFT simulations can shed maintain excellent wetting properties during opera-
some light on how the properties of this structurally tion. In SSBs, on the other hand, continuous plating
crucial layer are changing during battery operation. and stripping of Li, accompanied by its volume
As appealing as these DFT results are for mechan- expansion during cycling, leads to formation of voids
ical design of SSEs, they suffer from the inherent and partial delamination of the interface. This partial

123
 Meccanica

loss of contact slows the transport of Li ions across the properties, microstructural characteristics, and
interface, significantly increasing the impedance operational requirements to prevent this form of
[99, 134]. In the authors knowledge, there seems to damage.
have been no attempts in the literature at modeling
void formation and loss of contact at the Li anode/SSE
interface. Potential models addressing this issue
4 Composite electrodes
should incorporate the micro-scale mechanisms of
void formation and growth, along with the macro-
Solid-state composite electrodes represent another
scale deformation and morphology evolution of the
critical component of an all-solid-state Li-Ion battery
interphase. Models should also aim to relate the void
where mechanical effects will play a crucial role, cf.
formation and morphology change to impedance
[32]. Both cathodes and anodes not composed of pure
increase at the interface in a coupled chemo-mechan-
solid Li metal will consist of a composite of active
ics framework.
particles surrounded by a solid-state electrolyte
matrix. During cycling, the active particles will
3.4 Summary on modeling solid-state electrolytes
undergo electrochemically induced expansion and
contraction while the SSE provides an ion pathway
As reviewed above, modeling of solid-state elec-
for Li diffusion. Developing models for these systems
trolytes is in its infancy. In particular modeling the
requires an understanding of three critical compo-
interphase which forms at the anode/SSE interface is
nents: (1) the behavior of the active particles them-
critical for understanding stress generation, fracture,
selves, (2) the behavior of the solid state electrolyte,
and delamination and there are currently no validated
and (3) the combined behavior of the composite. This
models for this phenomenon. We list here a few key
section will mainly focus on this third aspect of
areas of research the authors feel are of pressing
modeling composite behavior since it is in its infancy.
relevance:
Within this context, one must consider the composite
– At present, capturing the phenomena of interphase behavior from a pure chemo-mechanical perspective,
formation and evolution remains one of the the behavior of the interfaces and possible delamina-
challenges for SSE realization from a modeling tion, as well as the potential for fracture/damage
perspective. A diffusion deformation reaction within the active particles and SSE. Modeling of the
theory capable of accommodating chemically active particles will be briefly reviewed in Sect. 4.1 as
induced phase transformation is currently lacking significant work has been made in this area and models
in the mechanics literature and needs to be are fairly mature and predictive. The state of the art in
developed to address the issue of interphase modeling solid-state electrolytes was reviewed in
formation in SSEs. Sect. 3 and is not considered here.
– Damage and fracture modeling of SSEs is lacking
with no calibrated model currently available. In 4.1 Chemo-mechanical continuum modeling
particular, models should be capable of capturing of active particles
both the mechanical degradation and the electro-
chemical degradation in the form of an increase in With respect to the active particles, coupled chemo-
impedance. mechanical models have been developed for a number
– In the pursuit of robust mechanical SSEs it could of different active materials. To different degrees of
be beneficial to use fully coupled chemo-mechan- complexity and success, models have been developed
ical continuum models as features to the training both for intercalation electrodes [135–141], where
space of new data science approaches for material deformations are generally elastic and on the order of
discovery. 10% strain, and for conversion electrodes
– The process of void formation and loss of contact [88, 142–146], where deformations are elastic-inelas-
at the anode/SSE remains unaddressed. A cali- tic in nature and chemical strains are on the order of
brated model for this phenomena must be devel- 100% or more, including aspects of failure modes and
oped in order to understand the relevant material damage [147].

123
Meccanica

Within the context of small deformations and deformations with a sharp propagating phase bound-
intercalation electrodes, some of the earliest works ary. A number of studies use simplified geometries
considered isotropic elastic intercalation electrodes (i.e. spherical or radially symmetric) to simplify to a
and within this restriction studied the critical geomet- one-dimensional problem which may be solved
rical, material, and electrochemical parameters to numerically [142, 154]. Huang et al. [154], for
prevent fracture using a linear elastic fracture mechan- example, demonstrated the manner in which stresses
ics approach [135, 136]. For certain types of electrodes develop in spherical particles undergoing large chem-
such as Graphite and LiFePO4 , phase-segregation (i.e. ically induced elastic-plastic deformations with or
staging) occurs during lithiation and models have been without the presence of a sharp interface. Figure6a–c
developed based largely on the Cahn–Hilliard phase- shows results from the work of Huang et al. [154],
field model to capture these phenomena, cf. where (a) shows the single-phase and two-phase
[137–139]. These models have been shown to accu- configurations. Figure 6b shows the distribution of
rately predict the electrochemical performance of equivalent re , radial rr , and hoop rh stresses across
these individual particles as well as the manner in the normalized particle radius R/A for the single-phase
which mechanical strains can affect the morphology of simulation while Fig. 6c shows the same result for the
the phase-segregation [138, 148, 149] or even suppress two-phase simulation. One particularly important
it entirely [137, 150]. Models have also been extended finding of this work is the manner in which the hoop
to include large elastic deformations and Cahn– stress may be either tensile or compressive on the
Hilliard-type phase-segregation [141, 144, 151]. With surface due to the presence of inelastic deformations
an accompanying finite-element implementation of and two-phase lithiation. This has important implica-
the coupled chemo-mechanics theoretical framework, tions in modeling fracture of these materials.
researchers have demonstrated in two and three- Amorphous Silicon (a-Si) has been widely studied
dimensional simulations the manner in which lithia- in the literature for its potential as a high-capacity
tion morphology is critically influenced by mechanical anode. Bucci et al. [146] and Di Leo et al. [88] have
deformation, e.g. [141, 152]. Although the aforemen- both calibrated continuum mechanics theories and
tioned models have been developed within the context demonstrated an ability to predict the Voltage vs.
of LIBs, they are particularly focused on the chemo- Charge behavior of simple a-Si electrodes coupled to
mechanical behavior of active particles and should be their mechanical behavior. Figure 6d, e, reproduced
equally applicable to modeling of these active mate- from the work of Bucci et al. [146], show the evolution
rials in the context of composite cathodes in SSBs. of cell potential and film stress in a thin-film a-Si
There are potential chemical interface stability issues electrode undergoing potentiostatic cycling. The
that might have to be considered for particular active model developed by Bucci et al. is calibrated and
particle/SSE combinations which might require dif- contrasted against the experimental results and
ferent theoretical models for capturing interfacial demonstrated the ability of these models to accurately
kinetics. However, the bulk behavior of the active predict both stress evolution and cell performance.
particle in a SSB composite cathode should still be Examples from Di Leo et al. [88] are shown in Fig. 6f–
well described through the above models. i where a large-deformation elastic-plastic chemo-
Conversion electrodes—such as Silicon which has mechanical theory and accompanying finite-element
been extensively studied—generally undergo very simulation is used to predict the behavior of hollow
large elastic-inelastic deformations with chemical double-walled nano-tubes experimentally realized by
strains on the order of hundreds of percent [153]. Wu et al. [156]. Figure 6f shows the finite-element
Continuum models have been developed for these discretization of the geometry and (h) the resulting
systems which account for large elastic-inelastic prediction of voltage vs. state-of-charge (SOC) for two
deformations accompanied by species diffusion and different C-rates. It is important to note that no
chemical strains [142–144, 146]. Closed form analyt- parameters are calibrated in this prediction. An
ical solutions have been developed for a few simple important aspect of this and similar studies is on the
cases. For example, Bower et al. [155] developed role of inelasticity as demonstrated in Fig. 6i which
analytical solutions for the problem of an elastic- shows Voltage vs. SOC for two simulations with and
plastic spherical particle undergoing diffusion induced without plastic deformation. (g) Shows contours of the

123
 Meccanica

Fig. 6 Representative simulations of active particles. a–c thin-film electrode. Reproduced with permission from [146]. f–i
Demonstrate a simulation using spherical symmetry considering Modeling of a hollow double-walled nanotube. f Shows the
both single-phase and two-phase lithiation within an elastic- finite-element discretization with h showing the resulting
plastic electrode. b Shows the distribution of equivalent re , simulation prediction against experimental measurements of
radial rr , and hoop rh stresses across the normalized particle Voltage vs. State-of-Charge (SOC). i Shows the effect on V vs.
radius R/A for the single-phase simulation while c shows the SOC of ignoring plastic deformation and g shows contours of
same result for the two-phase simulation. Reproduced with equivalent plastic strain, p after three half-cycles. Reproduced
permission from [153] and [154]. Evolution of d cell potential with permission from [88]
and e thin-film stress during the potentiostatic cycling of an a-Si

equivalent plastic strain p developed during three [159, 161]. The technique however is more complex
half-cycles of the hollow double-walled nanotube. when considering composite materials, such as the
Critical in the calibration of these theories is composite solid-state electrodes under consideration,
experimental data which shows both the voltage/ since the stress measured is an average over the
charge and stress/charge behavior of the system. In various phases and local stress state cannot be
recent years substrate curvature techniques [157–160] resolved. Researchers have recently begun investigat-
have been used to measure the mechanical stress in ing the use of substrate curvature techniques for
thin films of active materials during lithiation. These composite materials but there is at present no clear
experiments provide crucial information for modelers route for how these experiments will aid in the
as they quantify both the electrochemical and mechan- calibration and validation of composite electrode
ical performance of the material. Fig. 6d, e show an models [162–164]. It must also be noted that it is
example of the cell potential and film stress which can possible to inform aspects of the continuum mechanics
be measured during potentiostatic cycling of a thin- models through molecular dynamics or first-principles
film a-Si electrode. Experiments using substrate [165–169]. Zhao et al. [165] for example performed
curvature techniques have elucidated phenomena first-principle studies based on density functional
including the stress-potential coupling [157], strain- theory to elucidate the manner in which lithium
rate sensitivity [160], and fracture properties insertion in Silicon leads to a decreased stiffness,

123
Meccanica

strength, and a brittle-to-ductile transition. Sharma good agreement with experiments also published in
and co-workers have developed techniques based on their work. These chemo-mechanical continuum mod-
enhancing molecular dynamics with autonomous els introduce important aspects of the role of mechan-
basin climbing, kinetic Monte Carlo, and transition ics on interfacial electrochemistry but are still limited
state theory to gain insight into the diffusivity of Li in to solid electrodes.
a-Si [167], its inelastic response [168, 169]. Contin- An enhancement to these types of full-cell models
uum chemo-mechanics models for active materials are which accounts for the composite nature of the
relatively mature and a number of theoretical frame- electrodes has been developed around porous elec-
works and numerical techniques have been developed, trode theory, e.g. [176–179]. In such frameworks, the
calibrated, and validated which may serve as the electrodes are treated as a homogenized porous
foundation for the more pressing work of understand- medium. Renganathan et al. [176] developed a full-
ing all-solid-state composite cathodes. cell model with porous electrode theory used for both
electrodes. In this work the authors enhance traditional
4.2 Modeling of composite cathodes porous electrode theory to account for stress effects on
diffusion. Christensen [177] performs a similar anal-
Composite electrodes, both anodes and cathodes, are ysis including the role of fracture by assuming some
critical in the function of any Li-Ion battery. As degree of pre-existing fractures in the electrode.
reviewed in Sect. 4.1 above, the lithiation/delithiation Significant efforts have been developed in the theory
of active particles can cause significant deformations. of porous electrodes by Bazant and co-workers. In
Beyond the generation of stresses within the particles, Ferguson et al. [178], porous electrode theory is
these deformations can lead to significant particle-to- developed with a detailed Cahn–Hilliard model to
particle interactions as well as interactions with the account for phase-separation in active particles. The
surrounding matrix. In a solid-state electrode, in role of strain is introduced in the chemical potential by
contrast to a liquid electrolyte electrode, these inter- considering variations of the local concentration from
actions are enhanced as the high stiffness of the its volume averaged quantity. Through this analysis
electrodes leads to further confinement of the particles. the authors were able to demonstrate the importance of
Stresses exerted on the SSE can lead to a loss in mechanics in suppression of phase-separation and
electrochemical performance through damage/frac- promotion of homogeneous filling. Porous electrode
ture of the SSE or delamination of the active particles theory has the distinct advantage of producing a
from the SSE resulting in a loss of ionic conductivity, computationally efficient solution while still main-
e.g. [170, 171]. Further, it is possible that all-solid- taining a fairly rich description of the porous nature of
state composite electrodes will exhibit longer-range the microstructure. However, as with any homoge-
effects than liquid electrolyte systems where particle- nization technique, it is impossible to retain all details
to-particle mechanical interactions are enhanced due of the behavior at the microstructure and one loses for
to the presence of the stiff electrolyte matrix, e.g. example an understanding of stress distribution at the
[172]. active particle scale and the manner in which damage
Building on the theoretical frameworks developed might occur either through active particle failure,
for single active particles as reviewed in Sect. 4.1, the matrix failure, or interface delamination.
first progression towards modeling full battery behav- An alternative approach is to fully-discretize the
ior has been to consider half- or full-cell models with microstructure or a representative volume thereof, e.g.
solid electrodes e.g. [151, 173]. Bower et al. [151] [174, 175, 180]. Garcia et al. [174] developed a two-
developed a half-cell continuum chemo-mechanics dimensional finite element description of a detailed
model including an accounting for mechanical defor- composite cathode microstructure with cylindrical
mation and stress on the anode, cathode, electrostatic particles coupled to a simplified model for the
field and electrochemical reaction. Figure 7a shows a separator and anode behavior. Mechanical deforma-
schematic of the model under consideration, which tion was captured through the use of isotropic elastic
was applied to modeling a one-dimensional thin-film active particles with chemically-induced strains. Fig-
half-cell. The potential vs. time and capacity time ure 7b shows an example result using their simulation
plots are derived from the full-cell behavior and show tool for an electrode with graded porosity. Using this

123
 Meccanica

Fig. 7 Representative simulations of traditional electrode contours of normalized concentration, voltage, and hydrostatic
architectures. a Shows a schematic of a continuum chemo- stress. Reproduced with permission from [174]. c Results of a
mechanics representation of a half-cell Li-Ion system with the 3D numerical simulation of a composite NMC cathode informed
accompanying predictions for potential and capacity vs. time. by X-Ray tomography. The first two contours are of normalized
Reproduced with permission from [151]. b Results of a 2D concentration at varying charging voltages while the second two
numerical simulation of a composite cathode microstructure contours show equivalent stress at the same charging conditions.
with a liquid electrolyte. The bottom three images in (b) show Reproduced with permission from [175]

tool researchers studied the role of particle size and first two results show contours of normalized concen-
distribution on electrochemical and mechanical per- tration at two different voltages during charging as
formance. More recently, Xu et al. [175] performed well as a quiver plot of flux. The third and fourth
three-dimensional finite element simulations to results in Fig. 7c show contours of equivalent stress at
numerically resolve the microstructure of the same charging conditions. Using this detailed
LiNix Mny Coz O2 (NMC) composite cathodes. Syn- simulation tool, the authors quantified the manner in
chrotron X-ray tomography was used to inform the which mechanical damage influences battery perfor-
microstructural modeling and to validate model pre- mance. For example they quantified the manner in
dictions. Here, the composite cathodes modeled used a which electrochemical impedance is affected by
homogenized matrix consisting of the electrolyte and interfacial debonding. Using experimentally relevant
the carbon-binder matrix while active particles were microstructures is critical and recent trends, as
fully resolved. Although modeling a traditional liquid demonstrated in Xu et al. [175], have seen the use of
electrolyte material, a small stiffness (4GPa) is X-Ray tomography to develop experimentally accu-
attributed to the homogenized matrix. Damage was rate simulation domains. This of course adds compu-
modeled both in the active NMS particles through tational and experimental cost. Recently, Hofmann
cohesive zone modeling, and at the particle interfaces et al. [180] performed numerical simulations similar to
by penalizing interfacial resistance as a function of the those referenced above but importantly provided a
normal displacement of the interface. Figure 7c shows comparison on the use of synthetically generated
an example result of their numerical model where the microstructures vs. experimentally ‘‘exact’’

123
Meccanica

descriptions. In this approach one still needs to maximum capacity for the composite electrode, see
understand the key experimental descriptors of the Fig. 8b. Below this critical number, one is under-
microstructure such as a particle size and shape utilizing the available space, and above this number
distributions, but need not simulate the exact exper- one is generating excessive stresses due to particle-
imental microstructure. particle interactions. Similar to the work of Bucci et al.
The aforementioned research efforts focused on [172], existence of a critical volume fraction of active
LIBs, and although the results are not directly material for cathode compositions was investigated in
applicable as the higher stiffness of the SSE is a the work of Al-Siraj et al. [187]. Here, a linear decay in
critical component, they have served to pave the way effective diffusivity of SSE with increase in active
for current models developed specifically for all-solid- particle volume fraction was determined. Above a
state-batteries and solid-state composite cathodes in critical volume fraction, effective diffusivity of SSE
particular. Specifically, models for LIBs using a fully- becomes negligible in magnitude, thus limiting per-
discretized microstructure contain suitable models for formance. Modeling damage in composite solid-state
the behavior of active particles and an existing cathodes is critical. Bucci et al. [182] present arguably
framework for constituve modeling of the matrix (be the first quantitative study on the role of mechanical
it liquid or SSE) and an accounting for potential damage in composite all-solid-state cathodes. In this
damage through delamination using cohesive ele- work, cohesive zone modeling is used to capture
ments. With the development of proper chemo- damage in the SSE during cycling of the composite
mechanical constitutive theories for the SSE material cathode. Figure 8d shows the simulation domain used
and the active particle/SSE interfaces, these works can by the authors while (e) shows contours of normalized
serve as a starting point for modeling composite concentration and hydrostatic stress with cracks
cathodes in SSBs. From a purely electrochemical shown in black during lithiation to a state-of-charge
perspective Mykhaylov et al. [184] considered an all- of 0.7. Although the active particles are expanding due
solid-state construction with the use of porous elec- to chemically-induced strains, regions of tensile
trode theory, without consideration for mechanical stresses develop in the SSE resulting in the micro-
deformation and stress, and achieved closed-form or cracks shown in Fig. 8e. Critically, the authors
semi-closed form solutions for the performance of the demonstrated that compliant SSEs are more prone to
system. Bielefeld et al. [185] used a three-dimensional micro-cracking. Delamination of the active particles
description of the composite cathode microstructure to from the SSE was not considered in [182] and was
investigate its percolation characteristics. studied in a separate investigation by the same authors
Bucci and co-workers have done arguably the most [183]. There the interface between the active particle
extensive work on modeling the chemo-mechanics of and the SSE is studied through a one-dimensional
solid-state composite cathodes without [172] and with radially symmetric model with the microstructure
[182, 183, 186] the inclusion of damage modeling. In approximated by stacking of truncated octahedrons,
[172], the authors focus on how constraint induced see Fig. 8f. The authors found that very compliant
stresses modify cell performance using both an electrodes and high fracture energies are required to
analytical approach and numerical finite element delay the onset of delamination. The phase-map in
simulations. Figure 8a shows analytical results for Fig. 8f shows the necessary volume contraction
the open-circuit potential for Silicon vs Li with required to initiate delamination as a function of
varying stiffness of the electrolyte matrix. As is clear SSE stiffness and fracture energy. Even at small
from the figure, increasing matrix stiffness results in volume changes, delamination is expected for a broad
larger compressive stresses which have the effect of range of engineering relevant properties. Inelastic
shifting the open-circuit potential and reducing stor- deformation of the SSE was considered in this
age capacity. A two-dimensional finite element simplified geometry and can contribute significantly
description of a representative microstructure (similar to potentially delaying the onset of delamination by
to that shown in Fig. 8d) was used to investigate this releasing energy through inelastic deformation of the
behavior further. Through this analysis, Bucci and co- SSE. Clearly the coupled chemo-mechanical behavior
workers found that there is a critical volume ratio of of composite cathodes is complex and there is at
active material, Silicon to SSE at which one achieves present no clear understanding of the desired

123
 Meccanica

Fig. 8 Representative simulations for composite solid-state permission from [181]. d Simulation domain for capturing
cathodes. a Shows analytical results for the open-circuit fracture of the SSE during lithiation, note that the SSE domain is
potential for Silicon vs Li with varying stiffness of the discretized with cohesive elements to simulate fracture. e Shows
electrolyte matrix. b Capacity (percent dilute vs. dense packing) contours of normalized concentration and hydrostatic stress
of Si in a representative two-dimensional microstructure as a with cracks shown in black during lithiation to a state-of-charge
function of volume ratio of Silicon vs. SSE. The highest capacity of 0.7. Reproduced with permission from [182]. f Shows a
is achieved due to a tradeoff between packing and stress schematic of the one-dimensional radially symmetric model
generation. Reproduced from [172] under creative commons with the microstructure approximated by stacking of truncated
license CC BY-NC-ND 4.0. c Shows contours of equivalent octahedrons used to study delamination. The phase-map in
(von Mises) stress generated due to thermal contraction (top) f shows the necessary volume contraction required to initiate
and delithiation (bottom) for a simulation performed using the delamination as a function of SSE stiffness and fracture energy.
smoothed boundary method (SBM). Reproduced with Reproduced with permission from [183]

properties of SSEs for improved performance. Recent which deserves attention in future investigations is the
studies suggest that compliant SSEs are beneficial to use of phase-field continuum damage models, also
prevent delamination while stiff SSEs are beneficial to discussed briefly in Sect. 3.1 above. In this approach,
delay microcraking. Further, the SSE material will damage can initiate at any given location and proceed
also serve a role at the anode side of the cell in dendrite in any direction without the need for a pre-defined
suppression as reviewed in Sect. 2, and the effect of its crack path as is the case for intrinsic cohesive zone
mechanical behavior there must also be considered. models. The general theory of phase-field continuum
Damage, either of the active particles, the SSE, or damage models is relatively mature [190, 191], and
the interface is critical to understanding composite there have been a few recent examples of its applica-
cathodes, and its investigation remains in its infancy. tion mainly towards modeling damage of active
Use of cohesive elements to discretize interfaces particles, e.g. [98, 113, 192–195]. It will be relevant
provides one potential mechanisms for understanding as a future research direction to investigate the use of
interfacial failure. This has been studied for electro- these techniques also in the context of modeling
chemical systems but outside of the context of solid- damage of SSEs.
state batteries, e.g. [143, 188, 189]. Another method

123
Meccanica

Finally, we conclude by noting that direct numer- performance of solid-state cathodes is critical.
ical simulation by finite element analysis is not There is first a need to develop simulations with the
exclusively the only technique for modeling the combined effects of possible active particle failure,
behavior of solid-state composite cathodes. Recently, SSE failure, and delamination of the two. Three-
Hao et al. [196] developed a Kinetic Monte Carlo dimensional simulations, although computation-
(KMC) model to solve the electrochemistry formula- ally expensive, should be realized in order to get an
tion and coupled it to a finite element solution where experimentally relevant understanding of 3D crack
each cell in the KMC corresponds to an FEA model to formation and propagation. Within this context,
account for chemically induced volume changes of the research needs to be performed to assess what are
active particles. Yu et al. [181], used smoothed the most promising damage modeling techniques
boundary method (SBM) to solve mechanical equi- for each of the aforementioned phenomena.
librium and compute stresses due to thermal dilatation – Finally, we began our review with a discussion of
as well as chemically induced deformations. Figure 8c porous electrode theory, a form of homogeniza-
shows representative results from Yu et al. [181] for tion. Once there is a solid fundamental under-
equivalent (von Mises) stress generated due to thermal standing of the detailed microstructural behavior
contraction (top) and delithiation (bottom). This study of solid-state composite cathodes there is a need to
is unique as one of the only ones currently investigat- develop efficient homogenized models. These
ing the role of stress generation due to thermal misfit computationally efficient models, if predictive,
between components of the composite cathode. can subsequently be used to explore the design
space of all-solid-state batteries and identify
4.3 Summary on modeling composite cathodes promising candidates for next-generation materi-
als. There is a strong possibility that data analytics
As reviewed above, composite cathodes will play a (i.e. machine-learning) can play a role in the
critical role in all-solid-state Li-Ion battery architec- development of these new homogenization tools.
tures. Study of their coupled mechanical-chemical
behavior is in its infancy, and a number of research
avenues remain to be explored. We list here a few key
5 Conclusion
areas of research the authors feel are of pressing
relevance:
The commercialization of next-generation all-solid-
– There is a need for detailed continuum modeling of state battery architectures will rely on the development
experimentally relevant, three-dimensional solid- of predictive models for the various components of the
state cathode microstructures. Research along this cell capable of capturing the complex interplay
vein has been performed for liquid electrolytes (cf. between chemistry and mechanics. The shear number
Xu et al. [175]), but has not yet been realized for of possible material combinations as well as the design
solid-state composite cathodes. The inclusion of of the microstructure limit the possibility of studying
damage, discussed in the third bullet point, is these systems from a purely experimental perspective.
critical. Modeling techniques are required to identify key
– Within the context of direct simulation of exper- possible candidates for materials and design at all
imentally relevant microstructures, there is a need levels of the SSB. In this review article we have
for improved constitutive descriptions of the summarized the state-of-the-art in chemo-mechanical
relevant phases of the microstructure. Anisotropic modeling for (1) Li-metal anodes and dendrite growth,
properties, thermo-mechanical behavior, inelastic (2) solid-state electrolytes, and (3) composite cath-
deformations all need to be rigorously accounted odes. In doing so one key takeaway is that modeling of
for. Further, appropriate boundary conditions for these systems is very much at its infancy. A few key
simulating these representative volume elements areas have been researched extensively. Driven in
have not been extensively studied. large part by the objective of improving active
– As elucidated by the recent work of Bucci and co- materials for traditional liquid electrolyte cell designs,
workers [182, 183, 186], the role of damage on there have been sustained efforts in the areas of

123
 Meccanica

modeling Li-dendrite growth through stability analy- accurate predictions on what the range of SSE/active
sis as well as in modeling the chemo-mechanical material properties need to be in order to mitigate
behavior of active intercalation and reaction electrode damage. Further, one may study the design of the
particles. However, as illustrated in this review, there microstructure to understand how key descriptors such
are significant new challenges in modeling the all- as particle and shape size distribution impact mechan-
solid-state battery architecture. ical degradation. Other future avenues for research,
A major challenge is in modeling the solid-state including development of new homogenization tech-
electrolyte and in particular the unstable formation and niques, are presented in Sect. 4.3
propagation of thermodynamically interphases at the
anode/SSE boundary. This is a fully coupled problem Acknowledgements C.V.D.L. acknowledges funding from
the National Science Foundation under Award No. CMMI-
in that the driving force for reaction at these interfaces
1825132.
is coupled to the local stress state and the stresses are
significantly influenced by the volumetric expansion Compliance with ethical standards
resulting from the formation of the new interphase.
Modeling SSE behavior is perhaps the least well Conflict of interest The authors declare that they have no
conflict of interest.
understood component of SSBs. A number of impor-
tant research avenues in this area are presented in
Sect. 3.4.
References
With respect to Li-metal anodes, significant
research has gone into predicting the onset and growth 1. Armand M, Tarascon JM (2008) Building better batteries.
of dendrites at the anode/SSE interface. A number of Nature 451:652
contradicting models have been developed for condi- 2. Bruce PG, Freunberger SA, Hardwick LJ, Tarascon JM
(2015) Li-O2 and Li-S batteries with high energy storage.
tions in which dendrite growth is suppressed. How-
Nat Mater 11:19
ever, at present, most models have been proven 3. Yao X, Liu D, Wang C, Long P, Peng G, Hu YS, Li H,
inaccurate by experimental observations. There is a Chen L, Xu X (2016) High-energy all-solid-state lithium
critical need in this area to develop fully-coupled, batteries with ultralong cycle life. Nano Lett 16(11):7148
4. Li J, Ma C, Chi M, Liang C, Dudney NJ (2015) Solid
high-fidelity chemo-mechanical models for dendrite
electrolyte: the key for high-voltage lithium batteries. Adv
growth problem with a detailed accounting for the role Energy Mater 5:1401408
of surface imperfections and grain boundaries on the 5. Zhu Y, He X, Mo Y (2015) Origin of outstanding stability
effect of Li dendrite growth at the anode/SSE inter- in the lithium solid electrolyte materials: insights from
thermodynamic analyses based on first-principles calcu-
face. Further, better constitutive models for the
lations. ACS Appl Mater Interfaces 7(42):23685
behavior of Li metal itself are only now being 6. Richards WD, Miara LJ, Wang Y, Kim JC, Ceder G (2015)
developed and these still require significant attention. Interface stability in solid-state batteries. Chem Mater
Future research avenues with respect to Li metal 28(1):266
7. Roth EP, Orendorff CJ (2012) How electrolytes influence
anodes and dendrite growth are presented in Sect. 2.3.
battery safety. Electrochem Soc Interface 21(2):45
Finally, modeling efforts were with respect to solid- 8. Yang C, Fu K, Zhang Y, Hitz E, Hu L (2017) Protected
state composite cathodes. Unlike the other portions of lithium-metal anodes in batteries: from liquid to solid. Adv
the battery, here significant research exists for the Mater 29(36):1701169
9. Zhang W, Nie J, Li F, Wang ZL, Sun C (2018) A durable
behavior of the individual active particles which are
and safe solid-state lithium battery with a hybrid elec-
expected to be deployed in these composite materials. trolyte membrane. Nano Energy 45:413
A number of the main mechanical modeling chal- 10. Hayashi A, Ohtomo T, Mizuno F, Tadanaga K, Tatsum-
lenges thus lie in understanding the damage and failure isago M (2003) All-solid-state Li/S batteries with highly
conductive glass-ceramic electrolytes. Electrochem
in these materials under cycling. There is a need here
Commun 5(8):701
for experimentally relevant (or exact) numerical 11. Murugan R, Thangadurai V, Weppner W (2007) Fast
descriptions of the composite cathode microstructure lithium ion conduction in garnet-type Li7La3Zr2O12.
integrated with fully-coupled chemo-mechanical Angew Chem Int Ed 46(41):7778
12. Thokchom JS, Gupta N, Kumar B (2008) Superionic
models for the individual constituents. Further, these
conductivity in a lithium aluminum germanium phosphate
models should include all possible mechanisms of glass-ceramic. J Electrochem Soc 155(12):A915
damage. Through such a detailed model one can make

123
Meccanica

13. Allen J, Wolfenstine J, Rangasamy E, Sakamoto J (2012) 28. Koerver R, Zhang W, de Biasi L, Schweidler S, Kon-
Effect of substitution (Ta, Al, Ga) on the conductivity of drakov AO, Kolling S, Brezesinski T, Hartmann P, Zeier
Li7La3Zr2O12. J Power Sources 206:315 WG, Janek J (2018) Chemo-mechanical expansion of
14. Seino Y, Ota T, Takada K, Hayashi A, Tatsumisago M lithium electrode materials-on the route to mechanically
(2014) A sulphide lithium super ion conductor is superior optimized all-solid-state batteries. Energy Environ Sci
to liquid ion conductors for use in rechargeable batteries. 11(8):2142
Energy Environ Sci 7:627 29. Zhang W, Richter FH, Culver SP, Leichtweiss T, Lozano
15. Thangadurai V, Pinzaru D, Narayanan S, Baral A Kumar JG, Dietrich C, Bruce PG, Zeier WG, Janek J (2018)
(2015) Correction to ‘‘Fast solid-state li ion conducting Degradation mechanisms at the Li10GeP2S12/LiCoO2
garnet-type structure metal oxides for energy storage’’. cathode interface in an all-solid-state lithium-ion battery.
J Phys Chem Lett 6(3):347 ACS Appl Mater Interfaces 10(26):22226
16. Song X, Lu Y, Wang F, Zhao X, Chen H (2020) A coupled 30. Goodenough JB, Kim Y (2010) Challenges for recharge-
electro-chemo-mechanical model for all-solid-state thin able li batteries. Chem Mater 22(3):587
film Li-ion batteries: the effects of bending on battery 31. Takada K (2013) Progress and prospective of solid-state
performances. J Power Sources 452:227803 lithium batteries. Acta Mater 61(3):759
17. Ren Y, Shen Y, Lin Y, Nan CW (2015) Direct observation 32. Lewis JA, Tippens J, Cortes FJQ, McDowell MT (2019)
of lithium dendrites inside garnet-type lithium-ion solid Chemo-mechanical challenges in solid-state batteries.
electrolyte. Electrochem Commun 57:27 Trends Chem 1:845–857
18. Sharafi A, Meyer HM, Nanda J, Wolfenstine J, Sakamoto J 33. Xu L, Tang S, Cheng Y, Wang K, Liang J, Liu C, Cao YC,
(2016) Characterizing the Li-Li7La3Zr2O12 interface Wei F, Mai L (2018) Interfaces in solid-state lithium bat-
stability and kinetics as a function of temperature and teries. Joule 2(10):1991
current density. J Power Sources 302:135 34. Shen Z, Zhang W, Zhu G, Huang Y, Feng Q, Lu Y (2020)
19. Cheng E, Sharafi A, Sakamoto J (2016) Intergranular Li Design principles of the anode-electrolyte interface for all
metal propagation through polycrystalline Li6.25Al0.25- solid-state lithium metal batteries. Small Methods
La3Zr2O12 ceramic electrolyte. Electrochim Acta 223:85 4:1900592
20. Aguesse F, Manalastas W, Buannic L, Lopez del Amo JM, 35. Wang P, Qu W, Song WL, Chen H, Chen R, Fang D (2019)
Singh G, Llords A, Kilner J (2017) Investigating the Electro-chemo-mechanical issues at the interfaces in solid-
dendritic growth during full cell cycling of garnet elec- state lithium metal batteries. Adv Funct Mater
trolyte in direct contact with li Metal. ACS Appl Mater 29(27):1900950
Interfaces 9(4):3808 36. Zhang F, Huang QA, Tang Z, Li A, Shao Q, Zhang L, Li X,
21. Porz L, Swamy T, Sheldon BW, Rettenwander D, Frmling Zhang J (2020) A review of mechanics-related material
T, Thaman HL, Berendts S, Uecker R, Carter WC, Chiang damages in all-solid-state batteries: mechanisms, perfor-
YM (2017) Mechanism of lithium metal penetration mance impacts and mitigation strategies. Nano Energy
through inorganic solid electrolytes. Adv Energy Mater 70:104545
7(20):1701003 37. Zhao Y, Stein P, Bai Y, Al-Siraj M, Yang Y, Xu BX (2019)
22. Narayan S, Anand L (2018) A large deformation elastic- A review on modeling of electro-chemo-mechanics in
viscoplastic model for lithium. Extreme Mech Lett 24:21 lithium-ion batteries. J Power Sources 413:259
23. Kobayashi T, Yamada A, Kanno R (2008) Interfacial 38. Albertus P, Babinec S, Litzelman S, Newman A (2018)
reactions at electrode/electrolyte boundary in all solid- Status and challenges in enabling the lithium metal elec-
state lithium battery using inorganic solid electrolyte, thio- trode for high-energy and low-cost rechargeable batteries.
LISICON. Electrochim Acta 53(15):5045 Nat Energy 3:16
24. Sakuda A, Hayashi A, Tatsumisago M (2010) Interfacial 39. Kerman K, Luntz A, Viswanathan V, Chiang YM, Chen Z
observation between LiCoO2 electrode and Li2S- P2S5 (2017) practical challenges hindering the development of
solid electrolytes of all-solid-state lithium secondary bat- solid state Li ion batteries. J Electrochem Soc
teries using transmission electron microscopy. Chem 164(7):A1731
Mater 22(3):949 40. Barton JL, Bockris JO, Ubbelohde ARJP (1962) The
25. Santhanagopalan D, Qian D, McGilvray T, Wang Z, Wang electrolytic growth of dendrites from ionic solutions. Proc
F, Camino F, Graetz J, Dudney N, Meng YS (2014) R Soc Lond A 268(1335):485
Interface limited lithium transport in solid-state batteries. 41. Diggle JW, Despic AR, Bockris JO (1969) The mechanism
J Phys Chem Lett 5(2):298 of the dendritic electrocrystallization of zinc. J Elec-
26. Wenzel S, Randau S, Leichtweiß T, Weber DA, Sann J, trochem Soc 116(11):1503
Zeier WG, Janek J (2016) Direct observation of the 42. Chazalviel JN (1990) Electrochemical aspects of the
interfacial instability of the fast ionic conductor generation of ramified metallic electrodeposits. Phys Rev
Li10GeP2S12 at the lithium metal anode. Chem Mater A 42:7355
28(7):2400 43. Brissot C, Rosso M, Chazalviel JN, Baudry P, Lascaud S
27. Lewis JA, Cortes FJQ, Boebinger MG, Tippens J, (1998) In situ study of dendritic growth inlithium/PEO-
Marchese TS, Kondekar N, Liu X, Chi M, McDowell MT salt/lithium cells. Electrochim Acta 43(10):1569
(2019) Interphase morphology between a solid-state 44. Monroe C, Newman J (2003) Dendrite growth in lithium/
electrolyte and lithium controls cell failure. ACS Energy polymer systems: a propagation model for liquid elec-
Lett 4(2):591 trolytes under galvanostatic conditions. J Electrochem Soc
150(10):A1377

123
 Meccanica

45. Mullins W, Sekerka R (1963) Morphological stability of a formulation of butler-volmer electrochemical reaction
particle growing by difusion or heat flow. J Appl Phys kinetics including the influence of mechanics. J Elec-
34:323 trochem Soc 166(4):H167
46. Mullins WW, Sekerka RF (1964) Stability of a planar 65. Monroe C, Newman J (2005) The impact of elastic
interface during solidification of a dilute binary alloy. deformation on deposition kinetics at lithium/polymer
J Appl Phys 35(2):444 interfaces. J Electrochem Soc 152(2):A396
47. Aogaki R, Makino T (1981) Theory of powdered metal 66. Yu S, Siegel DJ (2018) Grain boundary softening: a
formation in electrochemistry-morphological instability in potential mechanism for lithium metal penetration through
galvanostatic crystal growth under diffusion control. stiff solid electrolytes. ACS Appl Mater Interfaces
Electrochim Acta 26(11):1509 10(44):38151
48. Aogaki R (1982) Image analysis of morphological insta- 67. Tian HK, Liu Z, Ji Y, Chen LQ, Qi Y (2019) Interfacial
bility in galvanostatic electrocrystallization: I. General electronic properties dictate Li dendrite growth in solid
expression for the growth mode of surface irregularities. electrolytes. Chem Mater 31(18):7351
J Electrochem Soc 129(11):2442 68. Klinsmann M, Hildebrand FE, Ganser M, McMeeking RM
49. Barkey DP, Muller RH, Tobias CW (1989) Roughness (2019) Dendritic cracking in solid electrolytes driven by
development in metal electrodeposition: II. Stability the- lithium insertion. J Power Sources 442:227226
ory. J Electrochem Soc 136(8):2207 69. Tian HK, Xu B, Qi Y (2018) Computational study of
50. Pritzker MD, Fahidy TZ (1992) Morphological stability of lithium nucleation tendency in Li7La3Zr2O12 (LLZO)
a planar metal electrode during potentiostatic electrode- and rational design of interlayer materials to prevent
position and electrodissolution. Electrochim Acta lithium dendrites. J Power Sources 392:79
37(1):103 70. Barai P, Higa K, Srinivasan V (2017) Effect of initial state
51. Sundstrm LG, Bark FH (1995) On morphological insta- of lithium on the propensity for dendrite formation: a
bility during electrodeposition with a stagnant binary theoretical study. J Electrochem Soc 164(2):A180
electrolyte. Electrochim Acta 40(5):599 71. Ahmad Z, Viswanathan V (2017) Stability of electrode-
52. Selim R, Bro P (1974) Some observations on rechargeable position at solid-solid interfaces and implications for metal
lithium electrodes in a propylene carbonate electrolyte. anodes. Phys Rev Lett 119:056003
J Electrochem Soc 121(11):1457 72. Ahmad Z, Viswanathan V (2017) Role of anisotropy in
53. Stewart SG, Newman J (2008) The use of UV/vis determining stability of electrodeposition at solid-solid
absorption to measure diffusion coefficients in LiPF6 interfaces. Phys Rev Mater 1:055403
electrolytic solutions. J Electrochem Soc 155(1):F13 73. Natsiavas P, Weinberg K, Rosato D, Ortiz M (2016) Effect
54. Akolkar R (2013) Mathematical model of the dendritic of prestress on the stability of electrode-electrolyte inter-
growth during lithium electrodeposition. J Power Sources faces during charging in lithium batteries. J Mech Phys
232:23 Solids 95:92
55. Akolkar R (2014) Modeling dendrite growth during 74. McMeeking RM, Ganser M, Klinsmann M, Hildebrand FE
lithium electrodeposition at sub-ambient temperature. (2019) Metal electrode surfaces can roughen despite the
J Power Sources 246:84 constraint of a stiff electrolyte. J Electrochem Soc
56. Ely DR, Garca RE (2013) Heterogeneous nucleation and 166(6):A984. https://doi.org/10.1149/2.0221906jes
growth of lithium electrodeposits on negative electrodes. 75. Barai P, Higa K, Srinivasan V (2017) Lithium dendrite
J Electrochem Soc 160(4):A662 growth mechanisms in polymer electrolytes and preven-
57. Guyer JE, Boettinger WJ, Warren JA, McFadden GB tion strategies. Phys Chem Chem Phys 19:20493
(2004) Phase field modeling of electrochemistry. II. 76. Wang Y, Cheng YT (2017) A nanoindentation study of the
Kinetics. Phys Rev E 69(2):021604 viscoplastic behavior of pure lithium. Scr Mater 130:191
58. Shibuta Y, Okajima Y, Suzuki T (2007) Phase-field 77. Xu C, Ahmad Z, Aryanfar A, Viswanathan V, Greer JR
modeling for electrodeposition process. Sci Technol Adv (2017) Enhanced strength and temperature dependence of
Mater 8(6):511 mechanical properties of Li at small scales and its impli-
59. Liang L, Qi Y, Xue F, Bhattacharya S, Harris SJ, Chen LQ cations for Li metal anodes. Proc Nat Acad Sci 114(1):57
(2012) Nonlinear phase-field model for electrode-elec- 78. LePage WS, Chen Y, Kazyak E, Chen KH, Sanchez AJ,
trolyte interface evolution. Phys Rev E 86(5):051609 Poli A, Arruda EM, Thouless MD, Dasgupta NP (2019)
60. Ely DR, Jana A, Garcı́a RE (2014) Phase field kinetics of Lithium mechanics: roles of strain rate and temperature
lithium electrodeposits. J Power Sources 272:581 and implications for lithium metal batteries. J Electrochem
61. Chen L, Zhang HW, Liang LY, Liu Z, Qi Y, Lu P, Chen J, Soc 166(2):A89
Chen LQ (2015) Modulation of dendritic patterns during 79. Masias A, Felten N, Garcia-Mendez R, Wolfenstine J,
electrodeposition: a nonlinear phase-field model. J Power Sakamoto J (2019) Elastic, plastic, and creep mechanical
Sources 300:376 properties of lithium metal. J Mater Sci 54(3):2585
62. Cogswell DA (2015) Quantitative phase-field modeling of 80. Fincher CD, Ojeda D, Zhang Y, Pharr GM, Pharr M (2020)
dendritic electrodeposition. Phys Rev E 92(1):011301 Mechanical properties of metallic lithium: from nano to
63. Monroe C, Newman J (2004) The effect of interfacial bulk scales. Acta Mater 186:215
deformation on electrodeposition kinetics. J Electrochem 81. Schultz RP (2002) Lithium: measurement of young’s
Soc 151(6):A880 modulus and yield strength. Tech. rep. Fermi National
64. Ganser M, Hildebrand FE, Klinsmann M, Hanauer M, Accelerator Lab., Batavia, IL (US)
Kamlah M, McMeeking RM (2019) An extended

123
Meccanica

82. Ferrese A, Newman J (2014) Mechanical deformation of a 99. Han X, Gong Y, Fu KK, He X, Hitz GT, Dai J, Pearse A,
lithium-metal anode due to a very stiff separator. J Elec- Liu B, Wang H, Rubloff G et al (2017) Negating interfacial
trochem Soc 161(9):A1350 impedance in garnet-based solid-state Li metal batteries.
83. Barai P, Higa K, Srinivasan V (2017) Lithium dendrite Nat Mater 16(5):572
growth mechanisms in polymer electrolytes and preven- 100. Ganser M, Hildebrand FE, Kamlah M, McMeeking RM
tion strategies. Phys Chem Chem Phys 19(31):20493 (2019) A finite strain electro-chemo-mechanical theory for
84. Anand L, Narayan S (2019) An elastic-viscoplastic model ion transport with application to binary solid electrolytes.
for lithium. J Electrochem Soc 166(6):A1092 J Mech Phys Solids 125:681
85. Wenzel S, Sedlmaier SJ, Dietrich C, Zeier WG, Janek J 101. Newman J, Thomas-Alyea KE (2012) Electrochemical
(2018) Interfacial reactivity and interphase growth of systems. Wiley, New York
argyrodite solid electrolytes at lithium metal electrodes. 102. Bucci G, Chiang YM, Carter WC (2016) Formulation of
Solid State Ion 318:102 the coupled electrochemical-mechanical boundary-value
86. Wang S, Xu H, Li W, Dolocan A, Manthiram A (2017) problem, with applications to transport of multiple charged
Interfacial chemistry in solid-state batteries: formation of species. Acta Mater 104:33
interphase and its consequences. J Am Chem Soc 103. Grazioli D, Zadin V, Brandell D, Simone A (2019) Elec-
140(1):250 trochemical-mechanical modeling of solid polymer elec-
87. Tippens J, Miers J, Afshar A, Lewis J, Cortes FJQ, Qiao H, trolytes: stress development and non-uniform electric
Marchese TS, Di Leo CV, Saldana C, McDowell MT current density in trench geometry microbatteries. Elec-
(2019) Visualizing chemo-mechanical degradation of a trochim Acta 296:1142
solid-state battery electrolyte. ACS Energy Lett 104. Zavattieri P, Raghuram P, Espinosa H (2001) A compu-
4:1475–1483 tational model of ceramic microstructures subjected to
88. Leo CVD, Rejovitzky E, Anand L (2015) Diffusion multi-axial dynamic loading. J Mech Phys Solids 49(1):27
deformation theory for amorphous silicon anodes: the role 105. Templeton DW, Holmquist TJ, MEYER HW, Grove DJ,
of plastic deformation on electrochemical performance. Int Leavy B (2002) A comparison of ceramic material models.
J Solids Struct 67–68:283 Ceram Trans 134:299
89. Rejovitzky E, Di Leo CV, Anand L (2015) A theory and a 106. Deshpande VS, Gamble EN, Compton BG, McMeeking
simulation capability for the growth of a solid electrolyte RM, Evans AG, Zok FW (2011) A constitutive description
interphase layer at an anode particle in a Li-ion battery. of the inelastic response of ceramics. J Am Ceram Soc
J Mech Phys Solids 78:210 94:s204
90. Loeffel K, Anand L (2011) A chemo-thermo-mechanically 107. Deshpande V, Evans A (2008) Inelastic deformation and
coupled theory for elastic-viscoplastic deformation, dif- energy dissipation in ceramics: a mechanism-based con-
fusion, and volumetric swelling due to a chemical reaction. stitutive model. J Mech Phys Solids 56(10):3077
Int J Plast 27(9):1409 108. Holland CC, McMeeking RM (2015) The influence of
91. Busso EP, Qian ZQ (2006) A mechanistic study of mechanical and microstructural properties on the rate-de-
microcracking in transversely isotropic ceramic-metal pendent fracture strength of ceramics in uniaxial com-
systems. Acta Mater 54(2):325 pression. Int J Impact Eng 81:34
92. Karlsson AM, Xu T, Evans A (2002) The effect of the 109. Espinosa HD, Zavattieri PD, Dwivedi SK (1998) A finite
thermal barrier coating on the displacement instability in deformation continuumndiscrete model for the description
thermal barrier systems. Acta Mater 50(5):1211 of fragmentation and damage in brittle materials. J Mech
93. Loeffel K, Anand L, Gasem ZM (2013) On modeling the Phys Solids 46(10):1909
oxidation of high-temperature alloys. Acta Mater 110. Ashby M, Sammis C (1990) The damage mechanics of
61(2):399 brittle solids in compression. Pure Appl Geophys
94. Zhao Y, Chen Y, Ai S, Fang D (2019) A diffusion, oxi- 133(3):489
dation reaction and large viscoelastic deformation coupled 111. Kimberley J, Ramesh K, Daphalapurkar N (2013) A
model with applications to SiC fiber oxidation. Int J Plast scaling law for the dynamic strength of brittle solids. Acta
118:173 Mater 61(9):3509
95. Salvadori A, McMeeking R, Grazioli D, Magri M (2018) A 112. Rice RW (1977) Microstructure dependence of mechani-
coupled model of transport-reaction-mechanics with cal behavior of ceramics. In: Treatise on materials science
trapping. Part I—small strain analysis. J Mech Phys Solids & technology, vol 11. Elsevier, pp 199–381
114:1 113. Zhang X, Krischok A, Linder C (2016) A variational
96. Di Leo CV, Anand L (2013) Hydrogen in metals: a coupled framework to model diffusion induced large plastic
theory for species diffusion and large elastic-plastic deformation and phase field fracture during initial two-
deformations. Int J Plast 43:42 phase lithiation of silicon electrodes. Comput Methods
97. Anand L, Mao Y, Talamini B (2019) On modeling fracture Appl Mech Eng 312:51
of ferritic steels due to hydrogen embrittlement. J Mech 114. Yamakawa S, Nagasako N, Yamasaki H, Koyama T, Asahi
Phys Solids 122:280 R (2018) Phase-field modeling of stress generation in
98. Miehe C, Dal H, Schänzel LM, Raina A (2016) A phase- polycrystalline LiCoO2. Solid State Ion 319:209
field model for chemo-mechanical induced fracture in 115. Ke X, Wang Y, Ren G, Yuan C (2019) Towards rationally
lithium-ion battery electrode particles. Int J Numer mechanical design of inorganic solid electrolytes for all-
Methods Eng 106(9):683 solid-state lithium ion batteries. Energy Storage Mater
26:313–324

123
 Meccanica

116. Deng Z, Wang Z, Chu IH, Luo J, Ong SP (2016) Elastic experiments and generative deep learning. Energy Storage
properties of alkali superionic conductor electrolytes from Mater 21:446–456
first principles calculations. J Electrochem Soc 133. Ahmad Z, Xie T, Maheshwari C, Grossman JC, Viswa-
163(2):A67 nathan V (2018) Machine learning enabled computational
117. Kresse G, Joubert D (1999) From ultrasoft pseudopoten- screening of inorganic solid electrolytes for suppression of
tials to the projector augmented-wave method. Phys Rev B dendrite formation in lithium metal anodes. ACS Central
59(3):1758 Sci 4(8):996
118. Pugh S (1954) XCII. Relations between the elastic moduli 134. Tsai CL, Roddatis V, Chandran CV, Ma Q, Uhlenbruck S,
and the plastic properties of polycrystalline pure metals. Bram M, Heitjans P, Guillon O (2016) Li7La3Zr2O12
Lond Edinb Dublin Philos Mag J Sci 45(367):823 interface modification for Li dendrite prevention. ACS
119. Wu M, Xu B, Lei X, Huang K, Ouyang C (2018) Bulk Appl Mater Interfaces 8(16):10617
properties and transport mechanisms of a solid state 135. Woodford W, Chiang Y, Carter W (2010) ‘‘Electrochem-
antiperovskite Li-ion conductor Li 3 OCl: insights from ical shock’’ of intercalation electrodes: a fracture
first principles calculations. J Mater Chem A 6(3):1150 mechanics analysis. J Electrochem Soc 157(10):A1052
120. Mattesini M, Soler JM, Ynduráin F (2006) Ab initio study 136. Zhao K, Pharr M, Vlassak JJ, Suo Z (2010) Fracture of
of metal-organic framework-5 Zn4 O(1; 4 benzenedi- electrodes in lithium-ion batteries caused by fast charging.
carboxylate )3 : an assessment of mechanical and spectro- J Appl Phys 108(7):073517
scopic properties. Phys Rev B 73:094111 137. Bai P, Cogswell D, Bazant M (2011) Suppression of phase
121. Wang Z, Wu M, Liu G, Lei X, Xu B, Ouyang C (2014) separation in LiFePO4 nanoparticles during battery dis-
Elastic properties of new solid state electrolyte material charge. Nano Lett 11(11):4890–4896
Li10GeP2S12: a study from first-principles calculations. 138. Cogswell DA, Bazant MZ (2013) Theory of coherent
Int J Electrochem Sci 9(2):562 nucleation in phase-separating nanoparticles. Nano Lett
122. Yang Y, Wu Q, Cui Y, Chen Y, Shi S, Wang RZ, Yan H 13(7):3036
(2016) Elastic properties, defect thermodynamics, elec- 139. Bohn E, Eckl T, Kamlah M, McMeeking R (2013) A
trochemical window, phase stability, and Li? mobility of model for lithium diffusion and stress generation in an
Li3PS4: insights from first-principles calculations. ACS intercalation storage particle with phase change. J Elec-
Appl Mater Interfaces 8(38):25229 trochem Soc 160(10):A1638
123. Chen H, Hong T (2019) First-principles investigation of 140. Purkayastha R, McMeeking R (2016) Stress due to the
the mechanical and thermodynamic properties of the intercalation of lithium in cubic-shaped particles: a
metal-borohydrides as electrolytes for solid-state batteries. parameter study. Meccanica 51(12):3081
J Electrochem Soc 166(4):A493 141. Di Leo C, Rejovitzky E, Anand L (2014) A Cahn–Hilliard-
124. Ahmad Z, Viswanathan V (2016) Quantification of type phase-field theory for species diffusion coupled with
uncertainty in first-principles predicted mechanical prop- large elastic deformations: application to phase-separating
erties of solids: application to solid ion conductors. Phys Li-ion electrode materials. J Mech Phys Solids 70:1
Rev B 94(6):064105 142. Zhao K, Pharr M, Cai S, Vlassak J, Suo Z (2011) Large
125. White A (2012) The materials genome initiative: one year plastic deformation in high-capacity lithium-ion batteries
on. MRS Bull 37(8):715 caused by charge and discharge. J Am Ceram Soc
126. Kaufman L, Ågren J (2014) CALPHAD, first and second 94(s1):s226–s235
generation—birth of the materials genome. Scr Mater 70:3 143. Bower AF, Guduru P (2012) A simple finite element model
127. Jain A, Ong SP, Hautier G, Chen W, Richards WD, Dacek of diffusion, finite deformation, plasticity and fracture in
S, Cholia S, Gunter D, Skinner D, Ceder G et al (2013) lithium ion insertion electrode materials. Modell Simul
Commentary: the materials project: a materials genome Mater Sci Eng 20(4):045004
approach to accelerating materials innovation. Apl Mater 144. Anand L (2012) A Cahn–Hilliard-type theory for species
1(1):011002 diffusion coupled with large elastic-plastic deformations.
128. Rajan K (2005) Materials informatics. Mater Today J Mech Phys Solids 60(12):1983
8(10):38 145. Cui Z, Gao F, Qu J (2012) A finite deformation stress-
129. Ramprasad R, Batra R, Pilania G, Mannodi-Kanakkithodi dependent chemical potential and its applications to
A, Kim C (2017) Machine learning in materials infor- lithium ion batteries. J Mech Phys Solids 60(7):1280
matics: recent applications and prospects. NPJ Comput 146. Bucci G, Nadimpalli S, Sethuraman V, Bower A, Guduru P
Mater 3(1):54 (2014) Measurement and modeling of the mechanical and
130. Agrawal A, Choudhary A (2016) Perspective: materials electrochemical response of amorphous Si thin film elec-
informatics and big data: realization of the ‘‘fourth para- trodes during cyclic lithiation. J Mech Phys Solids 62:276
digm’’ of science in materials science. Apl Mater 147. Lyu D, Ren B, Li S (2019) Failure modes and mechanisms
4(5):053208 for rechargeable Lithium-based batteries: a state-of-the-art
131. Makeev MA, Rajput NN (2019) Computational screening review. Acta Mech 230(3):701
of electrolyte materials: status quo and open problems. 148. Cogswell DA, Bazant MZ (2018) Size-dependent phase
Curr Opin Chem Eng 23:58 morphologies in LiFePO4 battery particles. Electrochem
132. Bhowmik A, Castelli IE, Garcia-Lastra JM, Jørgensen PB, Commun 95:33
Winther O, Vegge T (2019) A perspective on inverse 149. Nadkarni N, Rejovitsky E, Fraggedakis D, Di Leo CV,
design of battery interphases using multi-scale modelling, Smith RB, Bai P, Bazant MZ (2018) Interplay of phase
boundary anisotropy and electro-auto-catalytic surface

123
Meccanica

reactions on the lithium intercalation dynamics in Li X evolution in composite sulfur cathodes. Energy Storage
FePO 4 plateletlike nanoparticles. Phys Rev Mater Mater 16:491
2(8):085406 165. Zhao K, Wang WL, Gregoire J, Pharr M, Suo Z, Vlassak
150. Zhang T, Kamlah M (2019) Phase-field modeling of the JJ, Kaxiras E (2011) Lithium-assisted plastic deformation
particle size and average concentration dependent misci- of silicon electrodes in lithium-ion batteries: a first-prin-
bility gap in nanoparticles of LixMn2O4, LixFePO4, and ciples theoretical study. Nano Lett 11(7):2962
NaxFePO4 during insertion. Electrochim Acta 298:31 166. Huang S, Zhu T (2011) Atomistic mechanisms of lithium
151. Bower A, Guduru P, Sethuraman V (2011) A finite strain insertion in amorphous silicon. J Power Sources
model of stress, diffusion, plastic flow, and electrochem- 196(7):3664
ical reactions in a lithium-ion half-cell. J Mech Phys Solids 167. Yan X, Gouissem A, Sharma P (2015) Atomistic insights
59(4):804 into Li-ion diffusion in amorphous silicon. Mech Mater
152. Zhang T, Kamlah M (2020) Mechanically Coupled Phase- 91:306
Field Modeling of Microstructure Evolution in Sodium Ion 168. Yan X, Gouissem A, Guduru PR, Sharma P (2017) Elu-
Batteries Particles of Na x FePO 4. J Electrochem Soc cidating the atomistic mechanisms underpinning plasticity
167(2):020508 in Li-Si nanostructures. Phys Rev Mater 1(5):055401
153. McDowell MT, Xia S, Zhu T (2016) The mechanics of 169. Darbaniyan F, Yan X, Sharma P (2020) An atomistic
large-volume-change transformations in high-capacity perspective on the effect of strain rate and lithium fraction
battery materials. Extreme Mech Lett 9:480 on the mechanical behavior of silicon electrodes. J Appl
154. Huang S, Fan F, Li J, Zhang S, Zhu T (2013) Stress gen- Mech 87(3):031011
eration during lithiation of high-capacity electrode parti- 170. Koerver R, Aygn I, Leichtweiß T, Dietrich C, Zhang W,
cles in lithium ion batteries. Acta Mater 61(12):4354 Binder JO, Hartmann P, Zeier WG, Janek J (2017)
155. Bower AF, Guduru PR, Chason E (2015) Analytical Capacity fade in solid-state batteries: interphase formation
solutions for composition and stress in spherical elastic- and chemomechanical processes in nickel-rich layered
plastic lithium-ion electrode particles containing a propa- oxide cathodes and lithium thiophosphate solid elec-
gating phase boundary. Int J Solids Struct 69:328 trolytes. Chem Mater 29(13):5574
156. Wu H, Chan G, Choi JW, Ryu I, Yao Y, McDowell MT, 171. Zhang W, Schröder D, Arlt T, Manke I, Koerver R, Pinedo
Lee SW, Jackson A, Yang Y, Hu L et al (2012) R, Weber DA, Sann J, Zeier WG, Janek J (2017) (Electro)
Stable cycling of double-walled silicon nanotube battery chemical expansion during cycling: monitoring the pres-
anodes through solid-electrolyte interphase control. Nat sure changes in operating solid-state lithium batteries.
Nanotechnol 7(5):310 J Mater Chem A 5(20):9929
157. Sethuraman VA, Chon MJ, Shimshak M, Srinivasan V, 172. Bucci G, Swamy T, Bishop S, Sheldon BW, Chiang YM,
Guduru PR (2010) In situ measurements of stress evolution Carter WC (2017) The effect of stress on battery-electrode
in silicon thin films during electrochemical lithiation and capacity. J Electrochem Soc 164(4):A645
delithiation. J Power Sources 195(15):5062 173. Wan TH, Ciucci F (2020) Electro-chemo-mechanical
158. Sethuraman VA, Van Winkle N, Abraham DP, Bower AF, modeling of solid-state batteries. Electrochim Acta
Guduru PR (2012) Real-time stress measurements in 331:135355
lithium-ion battery negative-electrodes. J Power Sources 174. Garcia R, Chiang Y, Carter W, Limthongkul P, Bishop C
206:334 (2005) Microstructural modeling and design of recharge-
159. Nadimpalli SP, Sethuraman VA, Bucci G, Srinivasan V, able lithium-ion batterie. J Electrochem Soc 152(1):A255
Bower AF, Guduru PR (2013) On plastic deformation and 175. Xu R, Yang Y, Yin F, Liu P, Cloetens P, Liu Y, Lin F, Zhao
fracture in Si films during electrochemical lithiation/ K (2019) Heterogeneous damage in Li-ion batteries:
delithiation cycling. J Electrochem Soc 160(10):A1885 experimental analysis and theoretical modeling. J Mech
160. Pharr M, Suo Z, Vlassak JJ (2014) Variation of stress with Phys Solids 129:160
charging rate due to strain-rate sensitivity of silicon elec- 176. Renganathan S, Sikha G, Santhanagopalan S, White R
trodes of Li-ion batteries. J Power Sources 270:569 (2010) Theoretical analysis of stresses in a lithium ion cel.
161. Pharr M, Suo Z, Vlassak JJ (2013) Measurements of the J Electrochem Soc 157(2):A155
fracture energy of lithiated silicon electrodes of Li-ion 177. Christensen J (2010) Modeling diffusion-induced stress in
batteries. Nano Lett 13(11):5570 Li-ion cells with porous electrodes. J Electrochem Soc
162. Sethuraman VA, Nguyen A, Chon MJ, Nadimpalli SP, 157(3):A366
Wang H, Abraham DP, Bower AF, Shenoy VB, Guduru 178. Ferguson T, Bazant M (2012) Nonequilibrium thermody-
PR (2013) Stress evolution in composite silicon electrodes namics of porous electrodes. J Electrochem Soc
during lithiation/delithiation. J Electrochem Soc 159(12):A1967
160(4):A739 179. Purkayastha R, McMeeking R (2012) A linearized model
163. Nadimpalli SP, Sethuraman VA, Abraham DP, Bower AF, for lithium ion batteries and maps for their performance
Guduru PR (2015) Stress evolution in lithium-ion com- and failure. J Appl Mech 79(3):031021
posite electrodes during electrochemical cycling and 180. Hofmann T, Westhoff D, Feinauer J, Andrä H, Zausch J,
resulting internal pressures on the cell casing. J Elec- Schmidt V, Müller R (2020) Electro-chemo-mechanical
trochem Soc 162(14):A2656 simulation for lithium ion batteries across the scales. Int J
164. Zhang Y, Luo Y, Fincher C, McProuty S, Swenson G, Solids Struct 184:24
Banerjee S, Pharr M (2019) In-situ measurements of stress 181. Yu HC, Taha D, Thompson T, Taylor NJ, Drews A,
Sakamoto J, Thornton K (2019) Deformation and stresses

123
 Meccanica

in solid-state composite battery cathodes. J Power Sources diffusion induced stresses. Int J Solids Struct
440:227116 48(16–17):2304
182. Bucci G, Swamy T, Chiang YM, Carter WC (2017) 190. Miehe C, Welschinger F, Hofacker M (2010) Thermody-
Modeling of internal mechanical failure of all-solid-state namically consistent phase-field models of fracture: vari-
batteries during electrochemical cycling, and implications ational principles and multi-field FE implementations. Int J
for battery design. J Mater Chem A 5(36):19422 Numer Methods Eng 83(10):1273
183. Bucci G, Talamini B, Balakrishna A, Chiang Y, Carter W 191. Borden M, Verhoosel C, Scott M, Hughes R, Landis C
(2018) Mechanical instability of electrode-electrolyte (2012) A phase-field description of dynamic brittle frac-
interfaces in solid-state batteries. Phys Rev Mater ture. Comput Methods Appl Mech Eng 217:77
2(10):105407 192. Miehe C, Welschinger F, Hofacker M (2010) A phase field
184. Mykhaylov M, Ganser M, Klinsmann M, Hildebrand F, model of electromechanical fracture. J Mech Phys Solids
Guz I, McMeeking R (2019) An elementary 1-dimensional 58(10):1716
model for a solid state lithium-ion battery with a single ion 193. Klinsmann M, Rosato D, Kamlah M, McMeeking R
conductor electrolyte and a lithium metal negative elec- (2016) Modeling crack growth during Li insertion in
trode. J Mech Phys Solids 123:207 storage particles using a fracture phase field approach.
185. Bielefeld A, Weber DA, Janek J (2018) Microstructural J Mech Phys Solids 92:313–344
modeling of composite cathodes for all-solid-state batter- 194. Klinsmann M, Rosato D, Kamlah M, McMeeking RM
ies. J Phys Chem C 123(3):1626 (2016) Modeling crack growth during Li extraction in
186. Bucci G, Swamy T, Chiang YM, Carter WC (2017) Ran- storage particles using a fracture phase field approach.
dom walk analysis of the effect of mechanical degradation J Electrochem Soc 163(2):A102
on all-solid-state battery power. J Electrochem Soc 195. Klinsmann M, Rosato D, Kamlah M, McMeeking RM
164(12):A2660 (2016) Modeling crack growth during Li extraction and
187. Al-Siraj M, Stein P, Xu BX (2020) The effect of mor- insertion within the second half cycle. J Power Sources
phology changes and mechanical stresses on the effective 331:32
diffusivity of solid electrolyte for lithium ion batteries. 196. Hao F, Mukherjee PP (2018) Mesoscale analysis of the
J Electrochem Soc 167(2):020535 electrolyte-electrode interface in all-solid-state Li-ion
188. Bhandakkar T, Gao H (2010) Cohesive modeling of crack batteries. J Electrochem Soc 165:A1857
nucleation under diffusion induced stresses in a thin strip:
Implications on the critical size for flaw tolerant battery
Publisher’s Note Springer Nature remains neutral with
electrodes. Int J Solids Struct 47(10):1424
regard to jurisdictional claims in published maps and
189. Bhandakkar T, Gao H (2011) Cohesive modeling of crack
institutional affiliations.
nucleation in a cylindrical electrode under axisymmetric

123