Chemo-Mechanical Model of SEI Growth on Silicon Electrode Particles

†‡ †‡§ ∗†‡§
Lars von Kolzenberg Arnulf Latz Birger Horstmann
December 1, 2022

Abstract of the currently used graphite [1,2] . However, this capacity

gain comes at the cost of volume expansions up to 300% [3] .
arXiv:2108.04078v3 [physics.chem-ph] 30 Nov 2022

Silicon anodes promise high energy densities of next- These large expansions lead to high mechanical stresses,
generation lithium-ion batteries, but suffer from shorter which deteriorate the anode and lead to faster aging and
cycle life. The accelerated capacity fade stems from the shorter battery lifetime [3,4] .
repeated fracture and healing of the solid-electrolyte
The main aging mechanism in lithium-ion batteries with
interphase (SEI) on the silicon surface. This interplay
graphite or silicon anodes is the formation and growth
of chemical and mechanical effects in SEI on silicon
of the solid-electrolyte interphase (SEI) [5–15] . The SEI
electrodes causes a complex aging behavior. However, so
forms during the first battery cycle, when the anode po-
far, no model mechanistically captures the interrelation
tential is drawn below the electrolyte reduction poten-
between mechanical SEI deterioration and accelerated
tial [16–18] . This initiates reactions of electrolyte molecules
SEI growth. In this article, we present a thermodynami-
with lithium ions, which form organic compounds like
cally consistent continuum model of an electrode particle
lithium ethylene dicarbonate Li2 EDC and inorganic com-
surrounded by an SEI layer. The silicon particle model
pounds like LiF, Li2 CO3 , and Li2 O [7,18–32] . These prod-
consistently couples chemical reactions, physical trans-
ucts precipitate on the anode in a dual layer structure
port, and elastic deformation. The SEI model comprises
with a compact, inorganic inner layer and a porous, or-
elastic and plastic deformation, fracture, and growth.
ganic outer layer [33,34] and thus form a nanometer thick
Capacity fade measurements on graphite anodesand
and stable SEI at around 0.15 V vs. lithium metal [32] . In
in-situ mechanical SEI measurements on lithium thin
subsequent cycles, this SEI shields the electrolyte from the
films provide parametrization for our model. For the first
low anode potentials and thereby enables a stable battery
time, we model the influence of cycling rate on the long-
operation. However, the shielding effect is not perfect, so
term mechanical SEI deterioration and regrowth. Our
that the SEI continues to grow over time effectively low-
model predicts the experimentally observed transition in
ering the usable capacity [6,7] .
time dependence from square-root-of-time growth during
battery storage to linear-in-time growth during continued Battery storage experiments √ revealed that long-
cycling. Thereby our model unravels the mechanistic term SEI growth follows a t-time dependence point-
[35]
dependence of battery aging on operating conditions ing to a self-passivating process . As a pos-
and supports the efforts to prolong the battery life of sible long-term growth mechanism electrolyte diffu-
[30,36–46] [36,38,40,43,47–52]
next-generation lithium-ion batteries. sion , electron conduction , electron
tunneling [36,47,53] and the diffusion of neutral lithium in-
terstitial atoms [36,47,54,55] were proposed. However, only
the diffusion of electrons, e.g., via neutral lithium inter-
1 Introduction stitial atoms, yields the experimentally observed voltage
dependence of capacity fade [36,56] .
Lithium-ion batteries progressed to the benchmark bat-
Besides the open circuit voltage of the anode, also the
tery technology for mobile applications owing to their su-
operating conditions during battery cycling strongly af-
perior energy density as well as longevity. The use of sili-
fect SEI growth. In a recent experiment, Attia et al. [57]
con anodes would further increase the energy density, be-
showed the dependence of SEI growth on the magnitude
cause silicon has nearly the tenfold theoretical capacity
and direction of applied current. Two mechanistic mod-
∗ Corresponding Author: [email protected]
els describe this experimentally observed trend with good
† German Aerospace Center, Pfaffenwaldring 38-40, 70569
accuracy [47,52] . Implementing mechanistic SEI models to
Stuttgart, Germany
‡ Helmholtz Institute Ulm, Helmholtzstraße 11, 89081 Ulm, Ger- three-dimensional (3D) cell simulations gives further in-
many sights on the effect of battery operation on SEI growth.
§ Ulm University, Albert-Einstein-Allee 47, 89081 Ulm, Germany Using a purely reaction limited model for SEI growth

1
Franco and coworkers [58] evaluated SEI heterogeneity in
a 3D model, and discussed its dependence on electrode
mesostructure. By leaving out the coupling to transport
limitations this approach is only valid for the first few
cycles contrary to the fully coupled models of Bazant,
Horstmann, and Latz [47,52,59,60] .
Moreover, cycling experiments regularly reveal acceler-
ated SEI growth on high capacity anodes like silicon due to
Figure 1: Schematic representation of particle and SEI
large particle expansion and contraction (breathing) [34,61] . deformation. The compound deforms plastically Fpl , elastically
This geometrical change strains the SEI until it eventually Fel , and finally chemically Fch .
fractures, which leads to formation of new SEI upon direct
contact between electrolyte and electrode [31,62–64] . Several
groups developed mechanistic models to describe the me- for the dissipation R0 in the Lagrangian frame, [80,83–86]
chanical response of the SEI on battery cycling [14,64–79] . ~ Li,0 · ∇0 ·µLi + M : Lpl ≥ 0.
R0 = −N (1)
However, these models focus on SEI mechanics and incor-
porate at most simple SEI growth models [14,65–68,75,79] . Equation 1 restricts the choice of constitutive equations for
In this paper, we develop a detailed electrochemo- lithium flux N~ Li,0 and plastic flow Lpl = Ḟpl F−1 . They
pl
mechanical model to describe SEI mechanics and growth depend on the corresponding forces, namely the chemical
−T
on a deforming electrode particle. We describe the chemo- potential µLi and the Mandel stress M = JFT rev σFrev with
mechanics of the electrode particle with a thermody- [87,88]
the Cauchy stress σ . In the following, we present our
namically consistent model [80] . The electrochemical part electrode particle and SEI growth models.
of the SEI model relies on our previous works on SEI
growth [36,38,39,47] . The mechanical part of the SEI model 2.1 Electrode Particle Model
describes the SEI as porous dual-layer structure [33,34] ,
which deforms elastic-perfectly plastic [81] . We use the chemo-mechanical electrode particle model dis-
In the next section, we develop the model based on ir- cussed in our previous work [80] . The particle deformation
reversible thermodynamics and show the details of its im- F is completely reversible and consists of an elastic part
plementation in the subsequent section. Afterwards, we Fel due to mechanical stress and a chemical part Fch re-
parametrize the model chemistry and mechanics with re- sulting from changes in lithium concentration cLi,0 ,
cent experiments [56,81] . Based on this parametrization we
F = Frev = Fel Fch . (2)
analyze the electrochemical and mechanical predictions of
the model in the short- and long-term. Finally, we summa- The continuity equation 3,
rize our results and show possible extensions of our model.
~ Li,0
ċLi,0 = −∇0 · N with ~ Li,0 = −DLi ∂µLi ∇0 µLi , (3)
N ∂cLi,0

defines the change of lithium concentration with the diffu-

2 Theory sion constant DLi . The elastic deformation is determined
by the momentum balance
This section briefly presents our model for coupling chem-
istry and mechanics in electrode particles covered by SEI. 0 = ∇0 ·P, (4)
The reader is referred to section SI-1 in the supporting
information for further details. We summarize the system with the first Piola–Kirchhoff stress tensor P = JσF−T .
of differential-algebraic equation, which we implement, in
Equation SI-45 to Equation SI-51. 2.2 SEI Model
The deformations of the silicon electrode particle and Now, we derive a model for coupled SEI growth and me-
the SEI during lithium intercalation and deintercalation chanics. We model the SEI as porous medium consisting
is schematically depicted in Figure 1. These are captured of an incompressible SEI matrix and electrolyte inside its
with deformation gradient F = ∂~x/∂ X ~ 0 , which relates the
pores, as introduced by Single et al. [38,39] . The volume
Lagrangian domain Ω0 to the Eulerian domain Ω. The fraction of solid SEI, SEI = VSEI /V , characterizes the
volume expansion is given by its determinant det F = J = morphology at each point.
V /V0 . We divide the deformation into a reversible Frev The deformation tensor F̄ describes the overall volume
and an irreversible part Fpl [82] . deformation and consists of three parts,
We derive our model from non-equilibrium thermody-
namics in Section SI-1 and find the following inequality F̄ = F̄el F̄pl F̄ref . (5)

2
The first part is a reference deformation F̄ref , which we
introduce to set the stress free SEI configuration. Most
importantly, the SEI deforms plastically F̄pl and elasti-
cally F̄el .

SEI Growth
The SEI grows when electrons e− , lithium ions Li+ , and
electrolyte molecules react. We simplify the multitude of
possible SEI growth reactions to the formation of the most
prominent SEI component Li2 EDC from EC,

2Li0 + 2EC → Li2 EDC + R (6)

with neutral lithium interstitials Li0 = Li+ + e− , the sol-

vent EC, and gaseous byproducts R.
We model the transport of electrons through the SEI as
diffusion of localized electrons according to our previous Figure 2: Representation of elastic, plastic and broken regimes in
works [36,47] . A prominent example are the aforementioned mechanical model. Yield surface f = 0 for different degrees ξ of
damage. The elastic regime for the undamaged case ξ = 0 is
lithium interstitials Li0 [54,55] . The Li0 concentration in the
colored in green. With increasing damage ξ, the surface shrinks
SEI evolves according to until it converges to (0, 0) if SEI is broken ξ = 1.
d(SEI cLi0 ,0 ) ~ Li0 ,0 − 2Jr
¯ SEI ΓAV .
= −SEI ∇0 · N (7)
dt Based on the stress PSEI , we proceed to develop a model
0
Here, the diffusion of Li through the SEI is described by for plastic deformation and fracture of the SEI. For the
the flux density N~ Li0 , see subsection SI-1.3. The diffusing plastic deformation we introduce the yield function f ,
0
Li can react with the solvent inside the SEI pores and which tends to zero if the SEI reaches its yield criterion.
at the SEI-electrolyte interface. We model the available The fracture depends on the damage variable ξ, which
reaction surfaces with the specific surface area AV adapted describes the degree of deterioration and reaches from 0
from Single et al. [38,39] and shown detailedly in section SI- (intact) to 1 (broken). We couple the damage variable to
4. In Equation 7, Γ is the surface site density and rSEI is the yield function with the Gurson-Tvergaard-Needleman
the rate of the SEI formation reaction in Equation 6, approach [90–93]
3 2
rSEI = kc2Li0 . Mdev
 
(8) c 1 tr (Mc )
f= 2
2 + 2ξ cosh − 1 − ξ 2 ≤ 0. (11)
σY 2 σY
Here, cLi0 is the Li0 concentration inside the SEI and k is
a rate constant. Here Mdev c = Mc − 1/3 tr Mc is the deviatoric part of the
We capture formation of new SEI by tracking the SEI −T
adapted Mandel stress Mc = FT rev σFrev inside the SEI
volume fraction in the Lagrangian frame εSEI,0 = J ¯ SEI ,
and σY is the yield stress [87,88,94]
. The damage variable ξ
¯ SEI ΓAV V̄SEI depends on the SEI porosity  elyt = 1 − SEI by Equation
ε̇SEI,0 = Jr (9)
12 [93] .
with the average molar volume of SEI components V̄SEI . (
elyt if elyt < elyt,crit
In summary, our SEI growth model takes into account the ξ= elyt −elyt,crit

coupling of transport and reaction processes.  elyt + (1 −  elyt ) · elyt,frac −elyt,crit else.
(12)
SEI Mechanics The critical SEI porosity elyt,crit accounts for pore coa-
lescence, which accelerates mechanical degradation, if the
The SEI deforms elastoplastically until it eventually frac- SEI porosity is above the critical porosity elyt > elyt,crit .
tures as the electrode particle beneath expands and con- The fracture SEI porosity elyt,frac describes the poros-
tracts. We determine the elastic deformation with the ity at which the SEI ultimately breaks with ξ(elyt ≥
momentum balance inside the SEI elyt,frac ) = 1.
∇0 ·PSEI = 0. (10) In Figure 2, we show the yield surface f = 0, at which
the SEI flows plastically. We observe the classical von-
and determine PSEI from the micromechanical model for Mises yield surface for ξ = 0, which withstands arbitrary
porous solids of Danielsson et al. [89] , see Equation SI-37. large hydrostatic stress and only depends on the deviatoric

3
stress. Damage causes the yield surface to shrink until it conducted on full cells with a graphite anode during stor-
converges to the stress state (0, 0) for ξ = 1. age so that the influence of mechanics is negligible. Sec-
To describe the plastic flow upon reaching this yield ond, the experiments of Yoon et al. [81] parametrize the
surface, we rely on the maximum plastic dissipation pos- mechanical part of our model. In their experiments, they
tulate [87,88,95–98] as additional restriction to the principle grow SEI on a thin lithium film and then mechanically
of non-negative dissipation, Equation 1. This postulate strain it. Thus, the experiment is almost independent from
from plasticity theory constraints plastic flow to the nor- electrochemical effects. The parametrization with other
mal direction of the yield surface ∂f /∂Mc and thus leads substrates than silicon introduces additional uncertainties
to the following constitutive equation [87,88] , to our model. The choice of substrate determines the nu-
cleation properties, but the growth mechanism should be
∂f
Lpl = φ , (13) independent from the substrate. However, the evolving
∂Mc SEI composition and structure can, of course, be affected
by the innermost nucleation sites of the SEI.
where the plastic multiplier φ is non-negative, φ ≥ 0, guar-
anteeing non-negativity of the dissipation rate in Equation
1. The plastic multiplier and the yield function f addition- Electrochemical SEI Growth
ally obey the Karush-Kuhn-Tucker condition φf = 0 [88] . [56]
Thus, plastic flow is suppressed φ = 0 during elastic defor- We rely on the experiments of Keil et al. obtained with
mation, f < 0. For plastic deformation, f = 0, φ results graphite anodes to parametrize the chemical SEI growth
[56]
from the consistency condition f˙ = 0 . Note that the
[88] model. In their experiments, Keil et al. stored batteries
plastic flow is not trace-free and thus not volume preserv- at different states of charge and measured the capacity
ing, because the yield criterion, Equation 11 depends on fade after 9.5 months. In line with the approach of Single
[36]
the hydrostatic stress tr (Mc ) /3. et al. , we subdivide the measured irreversible capacity
fade into two parts. The first part ∆SoHlin is not SEI
related and linear in time, the second part stems from the
3 Parameterization SEI and is predicted by our model.
In Figure 3 we compare the results of our simulation
We assume a homogeneous electrode particle and a dual with the experimental storage data of Keil et al. [56] . We
layer SEI consisting of a dense, inorganic inner layer with a conclude that our model results in square-root-of-time
thickness Rin and a porous, organic outer layer [33,34,38,39]
. growth during storage and describes the experimentally
We introduce a thickness dependent minimum porosity observed SoC-dependence well. Based on these growth pa-
elyt,min (R̃) to enforce this morphology and stop the SEI rameters we now proceed to validate our mechanical SEI
reaction once this porosity is reached locally r̃SEI,0 (elyt ≤ model.
elyt,min ) = 0. To reduce the set of SEI parameters, we
further assume that this minimum porosity corresponds
SEI Mechanics
to the critical porosity for pore coalescence elyt,crit =
elyt,min , see Equation 12. Besides the minimum porosity We rely on recent membrane bulge measurements to vali-
elyt,min , also Young’s modulus ESEI and the yield strength date our mechanical SEI model [81] . Yoon et al. [81] grow an
σY vary between both layers. To continuously transition SEI from a thin lithium film located on a polymeric sup-
these SEI properties y, we use Equation 14, port. By applying pressure, the resulting SEI/polymer
film bulges. The pressure/bulge characteristics are then
y(Rin < R < Rin + Rtrans ) translated to stress-strain curves for the SEI in the cir-
3 2
 
2(R − Rin ) 3(R − Rin ) cumferential direction. Moreover, atomic force microscopy
= yin + (yout − yin ) · 3 − 2 +1 visualizes cracks inside the SEI depending on its expan-
Rtrans Rtrans
(14) sion. To mimic these experiments, we expand the SEI
continuously with a constant velocity ṙ1 at the inner-
with the transition thickness Rtrans . In Table SI-1 we list most SEI element and calculate R the mean circumferen-
[1,36,38,39,56,81,99–103] tial SEI stress σ̄ = 1/L
the parameters of our simulation . The RSEI,φ SEI σSEI,φ dr and expansion
further parametrization of the SEI on silicon is challeng- SEI,φ 
¯ = 1/L SEI F SEI,φ − 1dr.
ing due to the strong interrelation of electrochemical and Figure 4 shows the experimental results of Yoon et al. [81]
mechanical effects. To reduce this complexity, we adopt compared to our simulation results. We see that our model
two experiments, one for each submodel. First, we use the agrees well with the experimental stress-strain curve. Fur-
experiments of Keil et al. [56] to parametrize the electro- thermore, our SEI fracture model matches the experimen-
chemical part of the SEI model. These experiments were tally observed crack evolution.

4
 Figure 4: Comparison of mechanical experiments on SEI from
lithium thin films (crosses) [81] with the simulated results (line) to
parametrize the mechanical SEI model. Stress-strain curve in blue,
SEI crack formation in orange. The steps in both curves result
from the numerical discretization of the SEI.

Figure 3: Parametrization of aging model with SEI growth on

graphite electrodes. a) Comparison of the experimental aging
data [56] (crosses) with the simulated results for a graphite anode
(line). b) SEI growth over time for different states of charge.

Figure 5: Geometrical representation of our spherical symmetric

simulation setup, which consists of a spherical silicon particle
4 Results and Validation surrounded by a porous SEI. Varying state of charge during cycling
induces changes of particle size and SEI morphology.
In this section, we analyze the model outlined above. We
start by studying the particle-SEI geometry during one
cycle in section 4.1. Next, we take a closer look at the SEI responds to this volume change by thinning for high
mechanical response of silicon particle and SEI in section SoC and thickening for low SoC.
4.2. Subsequently, we analyze the SEI growth during one We further resolve the geometrical response of the SEI in
cycle in section 4.3 and finally look at the long-term SEI Figure 6. In Figure 6 a), we see the SEI breathing in each
growth 4.4. Unless otherwise specified, the particles were cycle. During delithiation, the SEI expands from 20 nm
cycled at 1C between Umax = 0.5 V and Umin = 0.05 V. to 30 nm. Moreover, the SEI densifies during delithiation
leading to a higher SEI volume fraction. The SEI expan-
4.1 Geometry sion and densification result from the volume conservation
of the SEI matrix. As the electrode particle shrinks, so
Lithiation and delithiation strongly affect particle and SEI does the inner radius of the SEI shell. The SEI compen-
geometry. In Figure 5, we show six distinct configurations sates this loss in volume by increasing its volume fraction
of our spherical symmetric simulation domain during a and thickness. This reversible short-term expansion of the
battery cycle. We see that the varying state of charge SEI overshadows the long-term SEI growth taking place
induces volume changes inside the electrode particle ac- in a time scale of months. In Figure 6, we thus isolate this
cording to the chemical expansion Fch . The surrounding long-term change in geometry by plotting the SEI thick-

5
 fully compressive during lithiation and fully tensile during
delithiation. We observe two interesting phenomena in the
stress response of the SEI. First, the SEI is initially not
stress free. Second, the stress magnitude inside the SEI
remains largely constant over the cycle and differs largely
between inner and outer SEI.
The stress profile inside the particle stems from con-
centration gradients. Initially, the concentration is homo-
geneous so that the stress vanishes. Upon lithiation, the
concentration in the particle center is lower than at its sur-
face leading to volume mismatches. To compensate this
mismatch, the particle center stretches elastically while
the outer particle compresses elastically. This causes the
observed tensile stress in the particle center and the com-
pressive stress in the outer particle. During charging, the
concentration gradient is inverted leading to the opposite
behavior [80] .
The SEI stress response in contrast is solely caused by
the particle breathing, because the lithium ion concen-
tration inside the SEI is constant [86] . The initial SEI
stress in Figure 5 a) results from the SEI deformation
from its reference configuration at Uref = 0.15 V [29–32] to
Umin = 0.05 V. Subsequently, the stress magnitude inside
the SEI remains largely constant due to plastic deforma-
tion. The observed stress is thus the yield causing stress
with f (σ) = 0, see Equation 11 and Figure 2. Along
the SEI, we observe a stress profile due to the prescribed
dual layer structure, see Equation 14. The lower limiting
porosity elyt,min and the higher Young’s modulus ESEI
Figure 6: SEI porosity evolution over time. a) Reversible and yield strength σY of the dense, inorganic inner layer
mechanical breathing of the SEI during one cycle. b) Irreversible lead to a higher stress magnitude compared to the porous,
chemical SEI growth over several months during storage. organic outer layer.
We further analyze the mechanical response of the SEI
ness during storage over several months. We observe, that by plotting the stress-strain curve inside the inner and the
the initial profile grows into a dual layer SEI structure withouter SEI in Figure 8. We observe a hysteresis in the stress
a sharp front to the electrolyte. The dual layer structure response of the SEI with tensile stress during lithiation and
agrees with the predictions of Single et al. [38,39] and is en-
compressive stress during delithiation. During lithiation,
forced here via the limiting porosity elyt,min . Our choice the SEI expands linear elastically in hoop direction until it
of a fast reaction rate, verified by the experiments in Fig- reaches an expansion of 5%. Then the plastic deformation
ure 3 yields the sharp reaction front. sets in and expands the inner SEI as much as 40% and
the outer SEI around 20%. The stress magnitude in the
outer SEI is approximately constant at 10 MPa, because
4.2 Mechanics
we assume perfect plasticity [81] . In contrast, the stress
Next, we investigate the mechanical response of particle magnitude in the inner SEI is not as constant, but varies
and SEI to the previously discussed geometrical changes. around a value of 40 MPa. This difference between inner
Figure 7 shows the stress state in the six different configu- and outer SEI results from the radial stress component,
rations of Figure 5. Initially, in Figure 7 a), the electrode see Figure 7. While the radial stress in the inner SEI
particle is stress free. The lithiation half-cycle, Figure 7 varies during the cycle due to the mechanical particle-SEI
b)-d), induces tensile stress in the inner part of the par- boundary condition the radial stress in the outer SEI is
ticle and compressive stress in the outer part. During negligible. During delithiation, the tensile stress releases
delithiation, Figure 7 e)-f), this behavior is inverted with elastically until a compression of 5% is reached. Then
tensile stress in the inner part and compressive stress in the plastic flow compresses the SEI in hoop direction with
the outer part. The stress inside the SEI in contrast is a similar stress and strain magnitude as in the lithiation

6
Figure 7: Radial stress σr (dotted) and hoop stress σφ (dashed) inside particle and SEI during one cycle. The six snapshots correspond
to the six geometries shown in Figure 5. In terms of color, the electrode domain has the respective SoC color (see Figure 5) and the
SEI/electrolyte domain is subdivided into SEI (green) and electrolyte (blue) according to the porosity profile.

pressions might lead to other mechanical failure modes

like buckling or delamination, which we do not consider in
our reductionist model [62,64] . To further analyze SEI frac-
ture, we thus subject our SEI to larger tensile strains. We
therefore set the stress-free SEI configuration to a smaller
particle size with Uref = 0.3 V and cycle with C/100 to
increase the SoC swing.
With Uref = 0.3 V, we can now study SEI fraction within
a single cycle. Figure 9 shows the proceeding SEI frac-
ture for this simulation setup during lithiation. The crack
starts from the SEI surface and expands through the outer
SEI stopping as it approaches the inner SEI. We observe
that the SEI deteriorates much stronger once it is broken
leading to larger pore expansion compared to our standard
cycle shown in Figure 6. However, similar to the fully in-
Figure 8: Stress strain curves for the inner (blue) and outer
tact case in Figure 6, we see that this deformation is re-
(orange) SEI during cycling. The SEI first deforms linear elastic
and then flows perfectly plastic upon reaching the yield stress. versible and the crack closes again as the particle delithi-
Lithiation and delithiation induce opposing mechanical loads, ates and the SEI compresses. This accords well to the
which causes the observed hysteresis during one cycle. experimental findings of Kumar et al. [31] , who observed
SEI cracks only in the outer SEI, which close again upon
delithiation. However, our homogenized 1D model cannot
half-cycle. capture the precise shape of the cracks and whether the
So far, we analyzed elastic and plastic SEI deforma- same cracks would open again in the next cycle [31] . Over-
tion and observed no SEI fracture. This is because our all, this mechanism accelerates SEI growth by lowering the
model SEI was formed at a largely expanded particle with SEI thickness and increasing the pore volume. These re-
Uref = 0.15 V. Thereby, the SEI exhibits large compres- sults show that low potentials in the initial SEI formation
sive, but only small tensile strains in hoop direction, which cycle increase the battery lifetime by enhancing the me-
effectively prevents SEI fracture. Nevertheless, large com- chanical stability of the SEI. In the next two sections, we

7
 Figure 10: SEI thickness (blue) and capacity (orange) during
Figure 9: SEI porosity profile evolution under large mechanical charging (dark grey) and discharging (light grey) with 1C. The
stress. The SEI consists of intact (green) and broken (red) thickness shows the reversible mechanical SEI breathing during
domains. Darker colors indicate higher particle SoC. Accordingly, each cycle. The SEI capacity elucidates the asymmetric SEI
the mechanical SEI degradation increases with the particle state of growth, which is accelerated during lithiation and decelerated
charge. during discharging.

analyze the SEI growth during short-term and long-term 4.4 Long-Term SEI Growth
cycling.
Because the SEI hardly grows during a single cycle, we
analyze the long-term SEI growth after several battery
4.3 Short-Term SEI Growth
cycles in this subsection. In Figure 11 a), we illustrate the
Significant SEI growth typically occurs in a time span capacity QSEI bound in the SEI of a silicon particle cycled
of months to years. Nevertheless, our model allows for one year at C/10 with the standard potential of initial
us to visualize the small SEI growth during one cy- SEI formation Uref = 0.15 V. Additionally,
R we plot the
cle. We start by analyzing the SEI thickness LSEI = mean SEI volume fraction ¯SEI = SEI dr in Figure 11 b).
R(SEI >R0.05) and the SEI capacity consumption QSEI = This quantity enriches our analysis of morphological SEI
2F/V̄SEI εSEI 4πR2 dR during our standard cycle in Fig- changes, because it also captures the influence of reversible
ure 10. We observe a reversible thinning and thickening SEI densification/porosification during each cycle, which
of the SEI during the cycle, corresponding to our findings we observed in Figure 6 and 9. Moreover, our SEI growth
from Figure 6. The irreversible SEI growth only plays model, Equation 9, rather depends on the porosity profile
a minor role, so that the initial thickness approximately elyt (R) than the macroscopic SEI thickness LSEI .
corresponds to the final thickness. We thus resolve the In Figure 11 a), we observe the same trend as in Figure
irreversibly consumed SEI capacity QSEI during the cy- 10 during each cycle (see zoom inlet); the SEI thickness os-
cle on the second y-axis and observe asymmetric capacity cillates and the SEI capacity grows asymmetrically. Simi-
consumption during one cycle: Irreversible SEI growth is larly, the mean SEI volume fraction ¯SEI oscillates, shown
accelerated by lithiation and decelerated by delithiation. in the zoom inlet of Figure 11 b). Over the long-term,
Moreover, the SEI growth is fastest at the end of lithiation, the amplitude of oscillations of mean SEI volume ¯SEI in-
i.e., for high SoC. creases from 5 nm to 15 nm as SEI fracture progresses, see
The asymmetric SEI growth results from the exponen- Figure 9. The fracture in turn decreases the SEI passiva-
tial dependence of the lithium interstitial concentration on tion so that we observe a linear capacity fade in Figure 11
the anode OCV U0 and intercalation overpotential ηint [47] . a). In contrast, during
√ storage, we observe a self-limiting
The influence of anode OCV on SEI growth was first the- SEI growth QSEI ∝ t, see Figure 3 b). This accelerated
oretically described by Single et al. [36] in line with the growth results from the interplay of battery cycling and
storage experiments of Keil et al. [56] , see Figure 3. The SEI growth, i.e., from mechanical SEI deterioration.
influence of intercalation overpotential stems from our re- In Figure 12, we thus illustrate how the SEI capacity
cent electrochemical SEI model [47] . This model agrees well increases over time for different charging currents as com-
to the experiments of Attia et al. [57] , which revealed de- pared to battery storage. We observe that higher charg-
pendence of SEI growth on the current magnitude and ing currents lead to faster SEI growth. But the relation-
direction. ship between charging current and SEI growth seems to

8
 Figure 12: Irreversible SEI capacity consumption QSEI relative to
the maximum particle capacity Qel,max for silicon particles cycled
for one year with different charging currents.

Figure 11: SEI mechanics and growth on a silicon electrode during

one year of continuous cycling with C/10. (a) Irreversibly bound
SEI capacity relative to electrode capacity QSEI /Qel,max (blue)
compared to a linear growth law (dashed orange line). The zoom
inlet shows the asymmetric growth in each cycle,
R see Figure 10. (b)
Evolution of mean SEI volume fraction ¯SEI = SEI dr.
Figure 13: Evolution of the SEI volume fraction profile on a silicon
particle cycled with C/100 (green) or C/10 (red) for one year.

be more complex, as we observe only a small difference

between C/100 and C/50 opposed to the large difference cycles with C/100, the cycle time is sufficient for the in-
between C/50 and C/20. ner SEI to reform in the newly created pores and thereby
In Figure 13, we therefore plot the dynamic SEI profile reattain its self-passivating character. However, in batter-
for a particle charged with C/100 (green) and C/10 (red). ies charged with C/10, the cycle time is too short for the
We observe a fundamentally different SEI growth between SEI to reform. As a result, the inner SEI fully deteriorates
these two charging currents. The particle charged with and leaves the anode unshielded from the electrolyte. This
C/100 closely follows the prescribed limiting profile similar causes unlimited SEI growth leading to the observed linear
to the storage case, Figure 6 b). In contrast, charging with growth in the long-term.
C/10 leads to a spread out profile in which the shielding To sum up our long-term results for SEI growth shown
inner layer vanishes over time. in Figure 11, 12, and 13, we observe a fundamental transi-
The reason for this difference lies in the different time- tion
√ in time-dependence. Starting from the well-known
scales imposed by the different charging rates. In each cy- t-SEI-growth during storage, cycling with increasing
cle, the inner SEI undergoes large deformations as shown current accelerates growth and at C/10 the SEI grows lin-
in Figure 6 a). These large deformations are not com- ear with time. This growth acceleration stems from the
pletely reversible and the inner SEI layer flows plastically continuous pore expansion inside the inner SEI caused by
to a thicker and more porous geometry. If the battery large deformations during cycling. The faster the cycling

9
rate, the less time the SEI has to repair these pores. As Our model agrees qualitatively well with the experimen-
a result, the porosity profile in Figure 13 at C/100 corre- tally observed SEI cracking during lithiation and healing
sponds to the prescribed SEI profile, which we also observe during delithiation on silicon particles [31,64] . We obtain
during battery storage, see Figure 3 and 9. In contrast, the parametrization for our electrochemo-mechanical SEI
faster cycling with C/10 deteriorates the inner SEI over model from experiments on different substrates. The stor-
time, so that the particle is no longer passivated and the age experiments of Keil et al. [56] conducted on graphite
SEI grows rapidly. This finding rationalizes empirically particles parametrize the electrochemical part of the SEI
motivated SEI growth models, which obtain a linear SEI model. The mechanical measurements of Yoon et al. [81]
growth from prescribing a constant SEI fracture and re- conducted on SEI from lithium thin film parametrizes the
growth term for every cycle [40,53,104] . mechanical part of the SEI model.
Despite the qualitatively good predictions, our model is For the first time, our so-validated model showed the
only a first step towards better understanding SEI growth complex relationship between SEI mechanics and electro-
on silicon electrode particles. First, our parametrization chemical growth on silicon electrodes. Namely, mechan-
with SEI experiments on graphite [56] and lithium thin ical SEI pore expansion further accelerates SEI growth
film [81] neglects structural differences of the SEI on the dif- at high states of charge. Moreover, continuous pore cre-
ferent substrates and thereby leads to uncertainties. Novel ation during SEI expansion deteriorates the inner SEI in
capacity fade experiments on silicon particles will be help- the long-term. For cycling currents J > C/20, the cy-
ful to obtain better estimates for the model parameters. cle time is too short to repair the inner SEI. √ As a result,
Second, we rely on a 1D-spherical symmetrical single par- we observe a transition from self-passivating ( t-) to non-
ticle model instead of resolving the whole electrode. As passivating (t-time-dependent) SEI growth with increasing
such, our model is agnostic of the circumferentially hetero- cycling currents. These new insights extend our under-
geneous SEI structure and anisotropy within the electrode standing of the influence of battery operation on battery
particles. Third, we only consider a single particle in- life. This will aid in designing battery operation protocols
stead of analyzing the electrode microstructure as a whole. for next-generation lithium-ion batteries.
Thereby, our model neglects structural heterogeneity re- Future works can extend our model for additional me-
sulting from particle size distribution, as well as graphite, chanical SEI deterioration, SEI heterogeneity, and lithium
carbon and binder phases. This heterogeneity leads to plating. Implementing our model in two or three dimen-
complex electrochemo-mechanical interactions, especially sions allows for an in-depth analysis of further mechanical
in graphite/silicon blend electrodes as recently shown by SEI damaging like crack formation, spallation, or delam-
Liu et al. [105] . ination. Moreover, this approach paves the way to bet-
However, these limitations offer promising possibilities ter account for the heterogeneity and polycristallinity of
to extend and apply our model. With efficient numerical the SEI. As our model relies on lithium atoms as medi-
schemes, e.g., as presented in ref. 80, our model can be ators for SEI growth, lithium plating, i.e., the accumu-
analyzed in 3D to reveal heterogeneity of SEI growth on lation of lithium on the anode, could be implemented
individual particles. Moreover, incorporating a simplified in our model as additional degradation mode. Further-
0D-version of our model in 3D-resolved electrode mod- more, integrating our model in 3D full cell simulations,
els [83,84,105] is a promising approach to reveal interparticu- would capture the influence of heterogeneous electrodes
lar SEI heterogeneity as well as the complex electrochemo- on battery degradation. Especially graphite/silicon blend
mechanical interactions of the different constituents. In electrodes, which suffer from large mechanical differences,
this context, our SEI model can direcly be adapted to would profit from our degradation model. A solution to
describe the mechanics of binder shell covered electrode the accompanying computational challenges would be to
particles. simplify our SEI model to a 0D-version and implement it
into a fully integrated electrochemo-mechanical FEM cell
model [80,106,107] .
5 Conclusion
We have developed a thermodynamically consistent Acknowledgements
chemo-mechanical model of an electrode particle coated
with a solid-electrolyte interphase (SEI). The electrode We gratefully acknowledge funding and support by the
model is derived from a free energy functional and ac- German Research Foundation (DFG) within the re-
counts for chemical deformation and elastic stress [80] . The search training group SiMET under the project number
SEI model accounts for elastic and plastic deformation, 281041241/GRK2218. The support of the bwHPC initia-
fracture, and lithium atom mediated SEI growth [36,47] tive through the use of the JUSTUS HPC facility at Ulm
based on the SEI volume fraction as order parameter. University is acknowledged. This work contributes to the

10
research performed at CELEST (Center for Electrochem- [10] N. Delpuech, N. Dupré, D. Mazouzi, J. Gaubicher,
ical Energy Storage Ulm-Karlsruhe). P. Moreau, J. S. Bridel, D. Guyomard, and
B. Lestriez. Correlation between irreversible capac-
ity and electrolyte solvents degradation probed by
Conflict of interest NMR in Si-based negative electrode of Li-ion cell.
Electrochem. commun., 33:72–75, 2013.
The authors declare no conflict of interest.
[11] B. Jerliu, E. Hüger, L. Dörrer, B. K. Seidlhofer,
R. Steitz, M. Horisberger, and H. Schmidt. Lithium
References insertion into silicon electrodes studied by cyclic
[1] C. K. Chan, H. Peng, G. Liu, K. McIlwrath, voltammetry and operando neutron reflectometry.
X. F. Zhang, R. A. Huggins, and Y. Cui. High- Phys. Chem. Chem. Phys., 20(36):23480–23491,
performance lithium battery anodes using silicon 2018.
nanowires. Nat. Nanotechnol., 3(1):31–35, 2008.
[12] D. Mazouzi, N. Delpuech, Y. Oumellal, M. Gau-
[2] Y. Sun, N. Liu, and Y. Cui. Promises and challenges thier, M. Cerbelaud, J. Gaubicher, N. Dupré,
of nanomaterials for lithium-based rechargeable bat- P. Moreau, D. Guyomard, L. Roué, and B. Lestriez.
teries. Nat. Energy, 1(7):1–12, 2016. New insights into the silicon-based electrode’s irre-
versibility along cycle life through simple gravimet-
[3] W. J. Zhang. A review of the electrochemical per- ric method. J. Power Sources, 220:180–184, 2012.
formance of alloy anodes for lithium-ion batteries.
J. Power Sources, 196(1):13–24, 2011. [13] H. G. Steinrück, C. Cao, G. M. Veith, and M. F.
Toney. Toward quantifying capacity losses due to
[4] H. Wu, G. Chan, J. W. Choi, I. Ryu, Y. Yao, M. T. solid electrolyte interphase evolution in silicon thin
Mcdowell, S. W. Lee, A. Jackson, Y. Yang, L. Hu, film batteries. J. Chem. Phys., 152(8), 2020.
and Y. Cui. Stable cycling of double-walled silicon
nanotube battery anodes through solid-electrolyte [14] J. Li, K. Adewuyi, N. Lotfi, R.G. Landers, and
interphase control. Nat. Nanotechnol., 7(5):310–315, J. Park. A single particle model with chemi-
2012. cal/mechanical degradation physics for lithium ion
battery state of health (soh) estimation. Appl. En-
[5] E. Peled. The Electrochemical Behavior of Alkali
ergy, 212:1178–1190, 2018.
and Alkaline Earth Metals in Nonaqueous Battery
Systems—The Solid Electrolyte Interphase Model. [15] K. Kalaga, M. T. F. Rodrigues, S. E. Trask, I. A.
J. Electrochem. Soc., 126(12):2047, 1979. Shkrob, and D. P. Abraham. Calendar-life ver-
[6] E. Peled, D. Golodnitsky, G. Ardel, and V. Eshke- sus cycle-life aging of lithium-ion cells with silicon-
nazy. The Sei Model - Application to Lithium Poly- graphite composite electrodes. Electrochim. Acta,
mer Electrolyte Batteries. Electrochim. Acta, 40(13- 280:221–228, 2018.
14):2197–2204, 1995.
[16] J. B. Goodenough and Y. Kim. Challenges for
[7] E. Peled. Advanced Model for Solid Electrolyte In- rechargeable Li batteries. Chem. Mat., 22(3):587–
terphase Electrodes in Liquid and Polymer Elec- 603, 2010.
trolytes. J. Electrochem. Soc., 144(8):L208–L210,
1997. [17] B. Horstmann, F. Single, and A. Latz. Review on
multi-scale models of solid-electrolyte interphase for-
[8] Y. Horowitz, H. G. Steinrück, H. L. Han, C. Cao, mation. Curr. Opin. Electrochem., 13:61–69, 2019.
I. I. Abate, Y. Tsao, M. F. Toney, and G. A. So-
morjai. Fluoroethylene Carbonate Induces Ordered [18] A. Wang, S. Kadam, H. Li, S. Shi, and Y. Qi. Review
Electrolyte Interface on Silicon and Sapphire Sur- on modeling of the anode solid electrolyte interphase
faces as Revealed by Sum Frequency Generation Vi- (SEI) for lithium-ion batteries. Npj Comput. Mater.,
brational Spectroscopy and X-ray Reflectivity. Nano 4(1):1–26, 2018.
Lett., 18(3):2105–2111, 2018.
[19] P. Lu, C. Li, E. W. Schneider, and S. J. Harris.
[9] Y. Li, W. Huang, Y. Li, A. Pei, D. T. Boyle, and Chemistry, impedance, and morphology evolution in
Y. Cui. Correlating Structure and Function of Bat- solid electrolyte interphase films during formation in
tery Interphases at Atomic Resolution Using Cryo- lithium ion batteries. J. Phys. Chem. C, 118(2):896–
electron Microscopy. Joule, 2(10):2167–2177, 2018. 903, 2014.

11
[20] D. Aurbach, B. Markovsky, M. D. Levi, E. Levi, Lithium Ion Batteries. Adv. Energy Mater., 6(8):1–
A. Schechter, M. Moshkovich, and Y. Cohen. New 12, 2016.
insights into the interactions between electrode ma-
terials and electrolyte solutions for advanced non- [31] R. Kumar, A. Tokranov, B. W. Sheldon, X. Xiao,
aqueous batteries. J. Power Sources, 81-82:95–111, Z. Huang, C. Li, and T. Mueller. In Situ and
1999. Operando Investigations of Failure Mechanisms of
the Solid Electrolyte Interphase on Silicon Elec-
[21] M. Winter. The solid electrolyte interphase - The trodes. ACS Energy Lett., 1(4):689–697, 2016.
most important and the least understood solid elec-
[32] G. M. Veith, M. Doucet, R. L. Sacci, B. Vacaliuc,
trolyte in rechargeable Li batteries. Z. Phys. Chem.,
J. K. Baldwin, and J. F. Browning. Determination
223(10-11):1395–1406, 2009.
of the Solid Electrolyte Interphase Structure Grown
[22] K. Xu. Nonaqueous liquid electrolytes for on a Silicon Electrode Using a Fluoroethylene Car-
lithium-based rechargeable batteries. Chem. Rev., bonate Additive. Sci. Rep., 7(1):1–15, 2017.
104(10):4303–4417, 2004.
[33] K. Edström, M. Herstedt, and D. P. Abraham.
[23] V. A. Agubra and J. W. Fergus. The formation A new look at the solid electrolyte interphase on
and stability of the solid electrolyte interface on graphite anodes in Li-ion batteries. J. Power
the graphite anode. J. Power Sources, 268:153–162, Sources, 153(2):380–384, 2006.
2014.
[34] E. Peled and S. Menkin. Review—SEI: Past, Present
[24] S. J. An, J. Li, C. Daniel, D. Mohanty, S. Nagpure, and Future. J. Electrochem. Soc., 164(7):A1703–
and D. L. Wood. The state of understanding of the A1719, 2017.
lithium-ion-battery graphite solid electrolyte inter- [35] M. Broussely, S. Herreyre, P. Biensan, P. Kasztejna,
phase (SEI) and its relationship to formation cy- K. Nechev, and R. J. Staniewicz. Aging mechanism
cling. Carbon, 105:52–76, 2016. in Li ion cells and calendar life predictions. J. Power
[25] P. Lu and S. J. Harris. Lithium transport within the Sources, 97-98:13–21, 2001.
solid electrolyte interphase. Electrochem. commun., [36] F. Single, A. Latz, and B. Horstmann. Iden-
13(10):1035–1037, 2011. tifying the Mechanism of Continued Growth of
[26] K. Xu. Electrolytes and interphases in Li-ion batter- the Solid–Electrolyte Interphase. ChemSusChem,
ies and beyond. Chem. Rev., 114(23):11503–11618, 11(12):1950–1955, 2018.
2014. [37] H. J. Ploehn, P. Ramadass, and R. E. White. Sol-
vent Diffusion Model for Aging of Lithium-Ion Bat-
[27] W. Huang, J. Wang, M. R. Braun, Z. Zhang, Y. Li,
tery Cells. J. Electrochem. Soc., 151(3):A456–A462,
D. T. Boyle, P. C. McIntyre, and Y. Cui. Dy-
2004.
namic Structure and Chemistry of the Silicon Solid-
Electrolyte Interphase Visualized by Cryogenic Elec- [38] F. Single, B. Horstmann, and A. Latz. Revealing SEI
tron Microscopy. Matter, 1(5):1232–1245, 2019. Morphology: In-Depth Analysis of a Modeling Ap-
proach. J. Electrochem. Soc., 164(11):E3132–E3145,
[28] M. Boniface, L. Quazuguel, J. Danet, D. Guyomard,
2017.
P. Moreau, and P. Bayle-Guillemaud. Nanoscale
Chemical Evolution of Silicon Negative Electrodes [39] F. Single, B. Horstmann, and A. Latz. Dynam-
Characterized by Low-Loss STEM-EELS. Nano ics and morphology of solid electrolyte interphase
Lett., 16(12):7381–7388, 2016. (SEI). Phys. Chem. Chem. Phys., 18(27):17810–
17814, 2016.
[29] A. Tokranov, B. W. Sheldon, C. Li, S. Minne, and
X. Xiao. In situ atomic force microscopy study of ini- [40] M. B. Pinson and M. Z. Bazant. Theory of SEI For-
tial solid electrolyte interphase formation on silicon mation in Rechargeable Batteries: Capacity Fade,
electrodes for Li-ion batteries. ACS Appl. Mater. Accelerated Aging and Lifetime Prediction. J. Elec-
Interfaces, 6(9):6672–6686, 2014. trochem. Soc., 160(2):A243–A250, 2012.
[30] A. Tokranov, R. Kumar, C. Li, S. Minne, X. Xiao, [41] M. Tang and J. Newman. Electrochemical Charac-
and B. W. Sheldon. Control and Optimization of the terization of SEI-Type Passivating Films Using Re-
Electrochemical and Mechanical Properties of the dox Shuttles. J. Electrochem. Soc., 158(5):A530–
Solid Electrolyte Interphase on Silicon Electrodes in A536, 2011.

12
[42] M. Tang and J. Newman. Transient Characteriza- [54] S. Shi, Y. Qi, H. Li, and L. G. Hector. Defect Ther-
tion of Solid-Electrolyte-Interphase Using Ferrocene. modynamics and Diffusion Mechanisms in Li2 CO3
J. Electrochem. Soc., 159(3):A281–A289, 2012. and Implications for the Solid Electrolyte Interphase
in Li-Ion Batteries. J. Phys. Chem. C, 117(17):8579–
[43] M. Tang, S. Lu, and J. Newman. Experimental
8593, 2013.
and Theoretical Investigation of Solid-Electrolyte-
Interphase Formation Mechanisms on Glassy Car- [55] F. A. Soto, Y. Ma, J. M. Martinez De La Hoz,
bon. J. Electrochem. Soc., 159(11):A1775–A1785, J. M. Seminario, and P. B. Balbuena. Formation
2012. and Growth Mechanisms of Solid-Electrolyte In-
[44] A. A. Tahmasbi, T. Kadyk, and M. H. Eikerling. terphase Layers in Rechargeable Batteries. Chem.
Statistical Physics-Based Model of Solid Electrolyte Mat., 27(23):7990–8000, 2015.
Interphase Growth in Lithium Ion Batteries. J. Elec- [56] P. Keil, S. F. Schuster, J. Wilhelm, J. Travi,
trochem. Soc., 164(6):A1307–A1313, 2017. A. Hauser, R. C. Karl, and A. Jossen. Calendar Ag-
[45] F. Hao, Z. Liu, P. B. Balbuena, and P. P. Mukher- ing of Lithium-Ion Batteries. J. Electrochem. Soc.,
jee. Mesoscale Elucidation of Solid Electrolyte In- 163(9):A1872–A1880, 2016.
terphase Layer Formation in Li-Ion Battery Anode.
[57] P. M. Attia, S. Das, S. J. Harris, M. Z. Bazant,
J. Phys. Chem. C, 121(47):26233–26240, 2017.
and W. C. Chueh. Electrochemical kinetics of sei
[46] F. Röder, R. D. Braatz, and U. Krewer. Multi-Scale growth on Carbon Black: Part I. experiments. J.
Modeling of Solid Electrolyte Interface Formation in Electrochem. Soc., 166(4):E97–E106, 2019.
Lithium-Ion Batteries. Comput. Aided Chem. Eng.,
38:157–162, 2016. [58] M. Chouchane, O. Arcelus, and A. A. Franco. Het-
erogeneous Solid-Electrolyte Interphase in Graphite
[47] L. von Kolzenberg, A. Latz, and B. Horstmann. Electrodes Assessed by 4D-Resolved Computational
Solid–Electrolyte Interphase During Battery Cy- Simulations. Batter. Supercaps, pages 1–8, 2021.
cling: Theory of Growth Regimes. ChemSusChem,
13(15):3901–3910, 2020. [59] T. Schmitt. Degradation Models and Simulation
Tools for Lithium and Zinc Batteries. PhD thesis,
[48] M. Broussely, P. Biensan, F. Bonhomme, P. Blan- Ulm University, 2019.
chard, S. Herreyre, K. Nechev, and R. J. Staniewicz.
Main aging mechanisms in Li ion batteries. J. Power [60] L. J. Bolay, T. Schmitt, O. S. Mendoza-Hernandez,
Sources, 146(1-2):90–96, 2005. Y. Sone, A. Latz, and B. Horstmann. Degradation of
lithium-ion batteries in aerospace. In 2019 European
[49] J. Christensen and J. Newman. A Mathemati- Space Power Conference (ESPC), pages 1–3, 2019.
cal Model for the Lithium-Ion Negative Electrode
Solid Electrolyte Interphase. J. Electrochem. Soc., [61] J. D. McBrayer, M. F. Rodrigues, M. C. Schulze,
151(11):A1977–A1988, 2004. D. P. Abraham, C. A. Apblett, I. Bloom, G. M.
Carroll, A. M. Colclasure, C. Fang, K. L. Harrison,
[50] A. M. Colclasure, K. A. Smith, and R. J. Kee.
G. Liu, S. D. Minteer, N. R. Neale, G. M. Veith, C. S.
Modeling detailed chemistry and transport for solid-
Johnson, J. T. Vaughey, A. K. Burrell, and B. Cun-
electrolyte-interface (SEI) films in Li-ion batteries.
ningham. Calendar aging of silicon-containing bat-
Electrochim. Acta, 58(1):33–43, 2011.
teries. Nat. Energy, 6(9):866–872, 2021.
[51] F. Röder, R. D. Braatz, and U. Krewer. Multi-Scale
Simulation of Heterogeneous Surface Film Growth [62] R. Kumar, P. Lu, X. Xiao, Z. Huang, and B. W.
Mechanisms in Lithium-Ion Batteries. J. Elec- Sheldon. Strain-Induced Lithium Losses in the Solid
trochem. Soc., 164(11):E3335–E3344, 2017. Electrolyte Interphase on Silicon Electrodes. ACS
Appl. Mater. Interfaces, 9(34):28406–28417, 2017.
[52] S. Das, P. M. Attia, W. C. Chueh, and M. Z.
Bazant. Electrochemical kinetics of SEI growth on [63] M. K. Jangid, F. J. Sonia, R. Kali, B. Ananthoju,
carbon black II: Modeling. J. Electrochem. Soc., and A. Mukhopadhyay. Insights into the effects of
166(4):E107–E118, 2019. multi-layered graphene as buffer/interlayer for a-Si
during lithiation/delithiation. Carbon, 111:602–616,
[53] D. Li, D. Danilov, Z. Zhang, H. Chen, Y. Yang, 2017.
and P. H. L. Notten. Modeling the SEI-Formation
on Graphite Electrodes in LiFePO4 Batteries. J. [64] K. Guo, R. Kumar, X. Xiao, B. W. Sheldon, and
Electrochem. Soc., 162(6):A858–A869, 2015. H. Gao. Failure progression in the solid electrolyte

13
 interphase (SEI) on silicon electrodes. Nano Energy, [75] I. Laresgoiti, S. Käbitz, M. Ecker, and D. U. Sauer.
68:104257, 2020. Modeling mechanical degradation in lithium ion bat-
teries during cycling: Solid electrolyte interphase
[65] R. Deshpande, M. Verbrugge, Y. Cheng, J. Wang, fracture. J. Power Sources, 300:112–122, 2015.
and P. Liu. Battery cycle life prediction with cou-
pled chemical degradation and fatigue mechanics. J. [76] Q. Deng, R. Hu, C. Xu, B. Chen, and J. Zhou.
Electrochem. Soc., 159(10):1730–1738, 2012. Modeling fracture of solid electrolyte interphase in
lithium-ion batteries during cycling. J. Solid State
[66] J. Xu, R. D. Deshpande, J. Pan, Y. Cheng, and Electrochem., 23(11):2999–3008, 2019.
V. Battaglia. Electrode side reactions, capacity loss
and mechanical degradation in lithium-ion batteries. [77] K. Zhao, M. Pharr, L. Hartle, J. J. Vlassak, and
J. Electrochem. Soc., 162(10):A2026–A2035, 2015. Z. Suo. Fracture and debonding in lithium-ion bat-
teries with electrodes of hollow core-shell nanostruc-
[67] R. D. Deshpande and D. M. Bernardi. Mod- tures. J. Power Sources, 218:6–14, 2012.
eling Solid-Electrolyte Interphase (SEI) Fracture:
Coupled Mechanical/Chemical Degradation of the [78] M. Tanaka, J. B. Hooper, and D. Bedrov. Role of
Lithium Ion Battery. J. Electrochem. Soc., Plasticity in Mechanical Failure of Solid Electrolyte
164(2):A461–A474, 2017. Interphases on Nanostructured Silicon Electrode:
Insight from Continuum Level Modeling. ACS Appl.
[68] E. Rejovitzky, C. V. Di Leo, and L. Anand. A theory Energy Mater., 1(5):1858–1863, 2018.
and a simulation capability for the growth of a solid
[79] N. Kotak, P. Barai, A. Verma, A. Mistry, and P. P.
electrolyte interphase layer at an anode particle in
Mukherjee. Electrochemistry-Mechanics Coupling
a Li-ion battery. J. Mech. Phys. Solids, 78:210–230,
in Intercalation Electrodes. J. Electrochem. Soc.,
2015.
165(5):A1064–A1083, 2018.
[69] Y. Kamikawa, K. Amezawa, and K. Terada. Chemo-
[80] G. F. Castelli, L. von Kolzenberg, B. Horstmann,
mechanical simulation of lif-rich solid–electrolyte in-
A. Latz, and W. Dörfler. Efficient simulation of
terphase formed from fluoroethylene carbonate on a
chemical-mechanical coupling in battery active par-
silicon anode. ACS Appl. Energy Mater., 4(4):3231–
ticles. Energy Technol., 2021.
3239, 2021.
[81] I. Yoon, S. Jurng, D. P. Abraham, B. L. Lucht,
[70] A. Verma and P. P. Mukherjee. Mechanistic Anal- and P. R. Guduru. Measurement of mechanical and
ysis of Mechano-Electrochemical Interaction in Sil- fracture properties of solid electrolyte interphase on
icon Electrodes with Surface Film. J. Electrochem. lithium metal anodes in lithium ion batteries. En-
Soc., 164(14):A3570–A3581, 2017. ergy Storage Mater., 25:296–304, 2020.
[71] D. E. Galvez-Aranda, A. Verma, K. Hankins, J. M. [82] G. A. Holzapfel. Nonlinear Solid Mechanics. John
Seminario, P. P. Mukherjee, and P. B. Balbuena. Wiley & Sons, Ltd., Chichester, 2000.
Chemical and mechanical degradation and mitiga-
tion strategies for Si anodes. J. Power Sources, [83] A. Latz and J. Zausch. Thermodynamic consis-
419(February):208–218, 2019. tent transport theory of Li-ion batteries. J. Power
Sources, 196(6):3296–3302, 2011.
[72] M. W. Verbrugge, Y. Qi, D. R. Baker, and Y. T.
Cheng. Diffusion-Induced Stress within Core-Shell [84] A. Latz and J. Zausch. Multiscale modeling of
Structures and Implications for Robust Electrode lithium ion batteries: Thermal aspects. Beilstein
Design and Materials Selection. Adv. Electrochem. J. Nanotechnol., 6(1):987–1007, apr 2015.
Sci. Eng., 15:193–225, 2015. [85] M. Schammer, B. Horstmann, and A. Latz. Theory
of Transport in Highly Concentrated Electrolytes. J.
[73] Y. He and H. Hu. Analysis of lithium ion concen-
Electrochem. Soc., 2021.
tration and stress in the solid electrolyte interphase
on the graphite anode. Phys. Chem. Chem. Phys., [86] F. Single, B. Horstmann, and A. Latz. Theory of
17(36):23565–23572, 2015. Impedance Spectroscopy for Lithium Batteries. J.
Phys. Chem. C, 2019.
[74] Y. He, H. Hu, K. Zhang, S. Li, and J. Chen. Me-
chanical insights into the stability of heterogeneous [87] J. Mandel. Plasticite classique et viscoplasticite:.
solid electrolyte interphase on an electrode particle. CISM International Centre for Mechanical Sciences.
J. Mater. Sci., 52(5):2836–2848, 2017. Springer, 1972.

14
 [88] J. Lubliner. Plasticity Revised Edition. J. Appl. [102] R. Naejus, D. Lemordant, R. Coudert, and P. Will-
Mech., 59(1):540, 2006. mann. Excess thermodynamic properties of binary
mixtures containing linear or cyclic carbonates as
[89] M. Danielsson, D. M. Parks, and M. C. Boyce. Con- solvents at the temperatures 298.15 K and 315.15
stitutive modeling of porous hyperelastic materials. K. J. Chem. Thermodyn., 29(12):1503–1515, 1997.
Mech. Mater., 36(4):347–358, 2004.
[103] M. N. Obrovac and L. J. Krause. Reversible Cycling
[90] A. L. Gurson. Continuum theory of ductile rupture of Crystalline Silicon Powder. J. Electrochem. Soc.,
by void nucleation and growth: Part I—Yield crite- 154(2):A103, 2007.
ria and flow rules for porous ductile media. J. Eng.
Mater. Technol. Trans. ASME, 99:2–15, 1977. [104] E. M. Perassi and E. P. Leiva. Capacity fading model
for a solid electrolyte interface with surface growth.
[91] V. Tvergaard. Influence of voids on shear band insta- Electrochim. Acta, 308:418–425, 2019.
bilities under plane strain conditions. Int. J. Fract.,
17(4):389–407, 1981. [105] C. Liu, O. Arcelus, T. Lombardo, H. Oularbi, and
Alejandro A Franco. Towards a 3d-resolved model
[92] V. Tvergaard. Influence of void nucleation on duc- of si/graphite composite electrodes from manufac-
tile shear fracture at a free surface. J. Mech. Phys. turing simulations. J. Power Sources, 512:230486,
Solids, 30(6):399–425, 1982. 2021.
[93] V. Tvergaard and A. Needleman. Analysis of the [106] R. Fang, M. Kronbichler, M. Wurzer, and W. A.
cup-cone fracture in a round tensile bar. Acta Met- Wall. Parallel, physics-oriented, monolithic solvers
all., 32(1):157–169, 1984. for three-dimensional, coupled finite element models
of lithium-ion cells. Comput. Methods Appl. Mech.
[94] F. Reusch, B. Svendsen, and D. Klingbeil. Local and
Eng., 350:803–835, 2019.
non-local Gurson-based ductile damage and failure
modelling at large deformation. Eur. J. Mech. A [107] G. F. Castelli and W. Dörfler. The numerical study
Solids, 22(6):779–792, 2003. of a microscale model for lithium-ion batteries. Com-
put. Math. with Appl., 77(6):1527–1540, 2019.
[95] R. von Mises. Mechanik der plastischen
Formänderung von Kristallen. J. Appl. Math.
Mech., 8(3):161–185, 1928.

[96] R. Hill. A variational principle of maximum plastic

work in classical plasticity. Q. J. Mech. Appl. Math.,
1(October):18–28, 1947.

[97] R. Hill. On the state of stress in a plastic-rigid body

at the yield point. Philos. Mag., 42(331):868–875,
1951.

[98] G. Taylor. A connexion between the criterion of

yield and the strain ratio. Proc. R. Soc. Lond. A
Math. Phys. Sci., 191(1027):441–446, 1947.

[99] A. Verma, A. A. Franco, and P. P. Mukherjee.

Mechanistic Elucidation of Si Particle Morphology
on Electrode Performance. J. Electrochem. Soc.,
166(15):A3852–A3860, 2019.

[100] V. B. Shenoy, P. Johari, and Y. Qi. Elastic softening

of amorphous and crystalline Li-Si Phases with in-
creasing Li concentration: A first-principles study.
J. Power Sources, 195(19):6825–6830, 2010.

[101] O. Borodin, G. D. Smith, and P. Fan. Molecular

dynamics simulations of lithium alkyl carbonates.
J. Phys. Chem. B, 110(45):22773–22779, 2006.

15