international journal of hydrogen energy xxx (xxxx) xxx

Available online at www.sciencedirect.com

ScienceDirect

journal homepage: www.elsevier.com/locate/he

Phase field modeling of hydrogen embrittlement

Chuanshi Huang, Xiaosheng Gao*

Department of Mechanical Engineering, University of Akron, Akron, OH 44325, USA

highlights graphical abstract

Hydrogen transport and hydrogen

embrittlement mechanisms are
combined in a phase field model.

A hydrogen diffusion equation is
derived and a trapping density
function is proposed.

HELP is modeled by a reduction of
the yield stress with hydrogen
concentration.

HEDE is modeled by a reduction of
the critical energy release rate
with hydrogen concentration.

Simulation results are in qualita-
tive agreement with results of
previous investigations.

article info abstract

Article history: A numerical framework is developed in this paper to study hydrogen embrittlement. A
Received 20 January 2020 hydrogen diffusion equation is derived, and a trapping density function is proposed in the
Received in revised form framework of the phase field model. Effect of hydrogen enhanced localized plasticity (HELP)
1 May 2020 and hydrogen enhanced decohesion (HEDE) are modeled by reducing the yield stress and
Accepted 3 May 2020 decreasing the critical energy release rate respectively. Simulation results of a compact tension
Available online xxx specimen and a double notched tension specimen show that hydrogen accumulates at the
crack/notch tip region driven by positive hydrostatic stress as well as more traps produced by
Keywords: plastic deformation in this area. Both HELP and HEDE reduce the load carrying capacity of the
Hydrogen embrittlement specimen, and their effects depend on the model parameters. The proposed model provides a
Phase field numerical tool that can be used to comprehensively simulate hydrogen embrittlement and
Hydrogen transport predict the ductile to brittle transition of the material due to the presence of hydrogen.
Hydrogen enhanced localized © 2020 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.
plasticity (HELP)
Hydrogen enhanced decohesion
(HEDE)
Fracture

* Corresponding author.
E-mail address: [email protected] (X. Gao).
https://doi.org/10.1016/j.ijhydene.2020.05.015
0360-3199/© 2020 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.

Please cite this article as: Huang C, Gao X, Phase field modeling of hydrogen embrittlement, International Journal of Hydrogen Energy,
https://doi.org/10.1016/j.ijhydene.2020.05.015
2 international journal of hydrogen energy xxx (xxxx) xxx

investigated how the dissolved hydrogen influence the inter-

Introduction action of dislocations, and proposed a hydrogen shielding ef-
fect to interpret the hydrogen promoted mobility of
Dissolved hydrogen atoms in metals degrade the material’s
dislocations. Liang et al. [20] used an axisymmetric unit cell
mechanical properties, such as strength and ductility, result-
model to investigate the mutual effect between hydrogen and
ing in material failing at a lower level of load than it can
micro-void coalescence under a range of stress triaxialities,
normally bear. This phenomenon is known as hydrogen
and their results suggested that the trapped hydrogen
embrittlement. Johnson first observed the phenomenon of
strongly influence the void coalescence process. Huang et al.[
hydrogen embrittlement in 1875 when he immerged a piece of
16] used a 3D unit cell to investigate how ductile fracture is
iron in strong hydrochloric or dilute sulphuric acids for a few
affected by the HELP mechanism with different stress states
minutes and found a significant reduction in the ductility of
imposed on the representative material volume. Luo et al. [22]
the material [1]. Johnson’s experiment has initiated the study
considered the influence of the of the loading speed on the
of the hydrogen embrittlement of metals [2]. Hydrogen em-
hydrogen distribution in the material, and studied the HELP
brittles a variety of materials, and hydrogen-induced cracking
effect on ductile fracture under the influence of steady state
often leads to a sudden, catastrophic failure without warning.
hydrogen distribution. Yu et al. [23] studied hydrogen-
Hydrogen embrittlement is known as one of the most com-
microvoid interactions under the effect of HELP, and investi-
mon failure causes of the high strength steel components in
gated the effect of hydrogen on the fracture loci. Depover and
military aircraft. It is also one of the most critical corrosion
Verbeken experimentally observed that the HELP mechanism
failures in oil and gas industry as well as in construction.
played an important role in hydrogen embrittlement [24].
Hydrogen can be introduced into the material during
Previous studies suggest that hydrogen in metals lowers
manufacture or over time through environmental exposure.
the surface energy and promote cracking [25e27]. The HEDE
Hydrogen atoms can diffuse between interstitial lattice sites
mechanism is a result of dissolved hydrogen atoms reducing
(NILS) in metals. They can also be trapped by imperfections
the strength of atomic bonds of the parent material. Conse-
such as dislocation core. The diffusion process is dominated
quently, dissolved hydrogen may cause a transition of the
by the gradients of chemical potential which is characterized
fracture mechanism from ductile to brittle [28]. Troiano [29]
by the gradient of the hydrogen concentration in the lattice
introduced the concept that the cohesive strength of the iron
sites as well as the gradient of the hydrostatic stress. Oriani [3]
lattice is lowered by the dissolved hydrogen. Beachem argued
developed a formulation based on the assumed equilibrium
that the presence of sufficiently accumulated lattice hydrogen
relation that describe the population of the trapped hydrogen
ahead of the crack tip promotes whatever deformation pro-
and the diffusive hydrogen (lattice hydrogen). Using this the-
cesses the microstructure allows for and proposed a model to
ory, Sofronis and McMeeking [4] derived a non-linear
explain various fracture mechanisms [10]. Jiang et al. [30] used
hydrogen diffusion equation and studied the coupled tran-
the Born-Haber thermodynamic cycle to calculate the fracture
sient hydrogen diffusion and plastic straining around a
energy of hydrogen attacked metals (Fe and Al), and found
blunting crack tip. They also proposed a function of the
that fracture energy decreases with the increase of hydrogen
equivalent plastic strain for the trap density based on Kum-
concentration. Based on this calculation, Serebrinsky [31]
nick and Johnson’s experimental observation [5]. Lufrano and
proposed a quadratic relationship between the cohesive
Sofronis [6] further modified Sofronis and McMeeking’s
strength and the hydrogen concentration. Martı́nez-Pan ~ eda
hydrogen diffusion model by adding the hydrogen induced
et al. [32] modeled hydrogen embrittlement by considering the
dilatation to the material constitutive equations.
critical energy release rate as a function of hydrogen con-
To interpret hydrogen embrittlement, many mechanisms
centration. Wang et al. [33] conducted atomic simulations to
have been proposed by different researchers, such as
analyze hydrogen-induced decohesion of grain boundary
hydrogen enhanced decohesion (HEDE) [2,7e9] and hydrogen
under different hydrogen charging conditions, and observed
enhanced localized plasticity (HELP) [10e14]. The HELP effect
the occurrence of intergranular fracture when the grain
can reduce the ductility of metals by promoting plastic
boundary cohesive energy was reduced by 37% due to the
deformation localization in the ductile fracture process.
presence of hydrogen. They suggested that intergranular
Beachem [10] conducted torsion tests of the hydrogen charged
fracture is a result of the combination of HEDE and HELP
and hydrogen free steel pipes and observed that hydrogen
mechanisms.
decreased the local flow stress. Matsui et al.[15] and Moriya
To comprehensively model hydrogen embrittlement of
et al. [16] systematically studied the effect of hydrogen on
metals, the numerical method needs to be able to simulate
mechanical properties of purity iron under different temper-
brittle fracture, ductile fracture and the fracture mechanism
atures and hydrogen concentrations. They concluded that
transition between ductile and brittle. Many numerical
trapped hydrogen at the screw dislocation core promote screw
models according to different fracture criteria are proposed to
dislocations, and on the other hand, hydrogen impedes edge
simulate the fracture processes of different materials. Among
dislocations under a sufficiently low level of temperature.
them, the phase field model is not only flexible to implement
Ferreira et al. [17] observed that the elastic interaction be-
but also has the ability to simulate complex fracture process,
tween dislocations was reduced by hydrogen in stainless
such as crack merging and branching. The phase field model is
steel. Robertson [18] performed in-situ TEM deformation ex-
developed from the Griffith’s theory [34], but it does not need a
periments on different metals and the results showed
pre-defined crack or a fracture path [35]. The feasibility of this
persuasive evidence that hydrogen could enhance the
method was demonstrated by Bourdin et al. [36,37]. Miehe
mobility of dislocations. Birnbaum and Sofronis [19]
et al. [38,39] proposed an algorithm which splits the crack

Please cite this article as: Huang C, Gao X, Phase field modeling of hydrogen embrittlement, International Journal of Hydrogen Energy,
https://doi.org/10.1016/j.ijhydene.2020.05.015
 international journal of hydrogen energy xxx (xxxx) xxx 3

phase field and the displacement field and applied the phase
model to predict brittle fracture, and Miehe et al. [40] further  
Jeþ ¼ KtrðεÞ2þ þ 2m εedev : εedev ; (3)
developed their model by including the plastic contribution to
predict ductile fracture. Ambati et al. [41,42] introduced the where K represents the bulk modulus, the bracket x± ¼
plastic strain state into the degradation function to delay ðx ±jxjÞ =2, m represents the shear modulus, trðεÞ represents the
crack initiation, and compared the numerical results with trace of the elastic strain tensor, εedev ¼ εe  trðεÞI=3 is the
experimental observations. Borden et al. [43] introduced a elastic part of the deviatoric strain tensor, and I represents the
cubic degradation function and applied it to the yield surface second order identity tensor.
during the fracture process. They also considered the effect of The plastic adjustment function, A, is expressed as
the stress triaxiality on the crack driving force. Huang and Gao  p
aε
[44] modified the model of Miehe et al. [40] by proposing a new A ¼ exp ; (4)
εf
approach for material degradation and incorporating a plastic
adjustment factor into the crack driving force function. Nu- where a is a constant coefficient, εp represents the effective
merical examples show that this model is capable of simu- plastic strain, and εf is the failure strain.
lating the fracture mechanism transition between ductile and The degradation function used to describe the continu-
brittle. ously broken process of the material is defined as
This study incorporates hydrogen transport in metals and
2
the resulting HELP and HEDE mechanisms into the phase field gðdÞ ¼ ð1  dÞ ; (5)
model developed by Huang and Gao [44] to numerically model and both the Young’s modulus and the yield surface are
the phenomenon of hydrogen embrittlement. We first briefly assumed to deteriorate with the increasing phase field value.
review the phase field model, followed by the derivation of the Based on this assumption, the yield function is modified as
hydrogen diffusion equation and a proposed trapping density
function to include the effect of the crack phase field value. f ðs; d; εp Þ ¼ gðdÞs  gðdÞsy ðεp Þ; (6)
Hydrogen embrittlement modeling considers two mecha- where s represents the equivalent stress, and sy represents
nisms. The HELP mechanism is modeled by decreasing the
the current yield stress.
local flow stress with hydrogen concentration, and the HEDE
mechanism is modeled by reducing the critical energy release
rate with hydrogen concentration. Finally, a compact tension Hydrogen transport coupled with deformation
specimen and a flat specimen with a double notch are used to and phase field
demonstrate the numerical model, and the simulation results
are presented and discussed. Governing equation of hydrogen diffusion

The dissolved hydrogen atoms in most metals can either be

Phase field model trapped at the reversible trapping sites or diffuse between the
normal interstitial lattice sites (NILS). Thus, they are divided
The phase field model presented by Huang and Gao is modi- into two populations, lattice hydrogen and trapped hydrogen,
fied in this study. Detailed descriptions of this model can be and the total hydrogen concentration is the sum of two parts
found in Ref. [44]. In the framework of the phase field method,
the crack surface is modeled in a diffusive manner using a C ¼ CL þ CT ¼ bL qL NL þ bT qT NT ; (7)
field variable d2ð0; 1Þ, with d ¼ 0 indicating the unbroken
where CL represents the lattice hydrogen concentration, CT
state while d ¼ 1 indicating the broken state. The phase field
represents the trapped hydrogen concentration, bL is the
value is calculated by solving the weak form of the equilibrium
number of lattice sites per solvent atom, bT is the number of
equation
sites per trap, qL represents the occupancy of the lattice sites,
Z h i qT represents the occupancy of the trapping sites, NL repre-
 d þ l2 Dd þ 2ð1  dÞH dV ¼ 0; (1) sents the number of solvent atoms per unit lattice volume,
U
and NT represents the density of the trapping sites.
where D is the Laplacian operator, l indicates the length scale The lattice hydrogen concentration and trapped hydrogen
of the crack surface, H is the crack driving force defined as H ¼ concentration follow an equilibrium relation. The local equi-
~ sÞ in loading period s2½0; t, and
max Hðx; librium relation between lattice hydrogen and trapped
hydrogen is expressed as [3].
~¼ J ;
eþ
H (2)  
Gc =lA qT qL WB
¼ exp ; (8)
1  qT 1  qL RQ
where J is the positive part of the stored elastic energy
eþ

density, Gc is the critical energy release rate, and A is a func- where Q represents the absolute temperature, R represents
tion accounts for a plastic adjustment. The value of Hðx;~ sÞ is the gas constant, and WB represents the binding energy of the
calculated according to the current loading step time s and the trapping sites.
current position x. The positive part of the stored elastic en- Consider a domain U with surface vU, the local balance
equation of total hydrogen is
ergy density Jeþ is defined as

Please cite this article as: Huang C, Gao X, Phase field modeling of hydrogen embrittlement, International Journal of Hydrogen Energy,
https://doi.org/10.1016/j.ijhydene.2020.05.015
4 international journal of hydrogen energy xxx (xxxx) xxx

where the lattice diffusion coefficient DL ¼ ML RQ. Substituting

Z Z
Eqs. (14) and (18) into Eq. (9) results in
ðC_ L þ C_ T ÞdV þ J , ndS ¼ 0 ; (9)
8 Z   9
U vU
> CL þ CT ð1  qT Þ _ dNT >
>
> CL þ qT ε_ p dVþ >
>
>
> CL dεp >
>
where n is the outward unit normal of the domain surface and < U =
J is the hydrogen diffusion flux. Based on Eqs. (5) and (6) and > Z " #
>
¼0 : (19)
>
>
2
ð1  dÞ DL CL VH >
>
the fact that qL ≪ 1, the concentration of trapped hydrogen can >
>  ð1  dÞ
2
D VC þ Vs ,ndS >
>
: L L
RQ
H ;
be calculated by vU

b T NT Applying Gauss’s divergence theorem results in

CT ¼ ; (10)
1 þ KbLTNCLL CL þ CT ð1  qT Þ _ h i
2
CL  V , ð1  dÞ DL VCL
with KT ¼ expðWB =RQÞ. CL
" # (20)
2
Taking the time derivative of the trapped hydrogen, we can ð1  dÞ DL CL VH dNT
þV, VsH þ qT ε_ p ¼ 0:
get RQ dεp

vCT _ vCT This is the governing equation for hydrogen transport that
C_ T ¼ CL þ ε_ p : (11) takes into account the effect of the crack phase field. From Eq.
vCL vεp
(20) we can see that the diffusion of lattice hydrogen is not
The partial derivative of the trapped hydrogen with respect
only affected by the gradient of lattice hydrogen concentra-
to the lattice hydrogen is
tion, but also affected by the trapped hydrogen concentration,
vCT CT ð1  qT Þ the gradient of hydrostatic stress, the plastic deformation, and
¼ : (12) the phase field value.
vCL CL
Taking the partial derivative of the trapped hydrogen with
Hydrogen trapping
respect to the effective plastic strain leads to

vCT vCT dNT dNT In Eq. (7), the trapped hydrogen concentration is a function of
¼ ¼ qT : (13) the occupancy of the trapping sites qT , the number of sites per
vεp vNL dεp dεp
trap bT , and the trapping density NT . Here bT is a material
Substituting Eqs. (12) and (13) into Eq. (11), we can obtain constant, and qT is a factor for the equilibrium relation be-
the time derivative of the trapped hydrogen as tween trapped hydrogen and lattice hydrogen. As for the
CT ð1  qT Þ _ dNT trapping density NT , Kumnick and Johnson [46] showed that in
C_ T ¼ CL þ qT ε_ p : (14) iron, it is associated with dislocations developed during
CL dεp
plastic deformation. Sofronis and McMeeking [4] suggested
The total hydrogen flux J is defined as that in BCC iron, the trapping density NT increases mono-
tonically with the increased plastic strain and it is indepen-
J ¼  ML CL VmL  MT CT VmT ; (15)
dent of temperature. They also assumed that the trapping
where ML and MT are the mobility and mL and mT are the chemical density saturates as the plastic strain becomes larger than 0.8.
potential of the lattice hydrogen and trapped hydrogen respectively. According to the experimental observations presented by
It is assumed that trapping sites are isolated and transport between Kumnick and Johnson [46], Taha and Sofronis [47] proposed
traps is by lattice diffusion [45]. Thus, the mobility of trapped that for iron and steels, the trapping density is a function of
hydrogen MT is assumed to be zero. Moreover, the mobility of the the plastic strain, and the relation is written as follows
hydrogen in lattice sites ML is assumed to degrade with the occur-
rence of facture. With these assumptions, the total hydrogen diffu- logðNT Þ ¼ 23:26  2:33 expð5:5εp Þ: (21)
sion flux J becomes Atomistic simulations [48] show that as a crack propagates
2 through a hydrogen-rich region, the hydrogen atoms are
J ¼  ð1  dÞ ML CL VmL : (16)
trapped along the newly created crack surfaces, and behind
Here the degradation function is incorporated into the the new crack tip hydrogen concentration does not increase.
hydrogen diffusion flux function, which means that when the However, in the phase field model, the crack topology is rep-
material is fully broken (d ¼ 1), there is no hydrogen diffusion resented in a diffusive manner using the phase field value, d,
along or cross the newly developed crack surfaces. and the plastic strain in the area of the diffusive crack still
The chemical potential of the lattice hydrogen is increases monotonically after fracture occurred, resulting in a
  monotonic increase in trapping density there according to Eq.
CL
mL ¼ m0L þ RQln  VH sH ; (17) (21). To overcome this problem, we propose a new trapping
NL
density function in an incremental form as follows
where sH is the hydrostatic stress defined as by sH ¼
logðNT Þ ¼ 23:26  2:33 expð5:5~εp Þ: (22)
ðs11 þs22 þs33 Þ=3. From Eqs. (16) and (17), we can obtain
p
2
Different from (21), here the plastic factor ~ε is not the
2 ð1  dÞ DL CL VH effective plastic strain, but a function of the effective plastic
J ¼  ð1  dÞ DL VCL þ VsH ; (18)
RQ

Please cite this article as: Huang C, Gao X, Phase field modeling of hydrogen embrittlement, International Journal of Hydrogen Energy,
https://doi.org/10.1016/j.ijhydene.2020.05.015
 international journal of hydrogen energy xxx (xxxx) xxx 5

strain and the crack phase field value. At time increment nþ

p
1，the plastic factor is defined as ~εnþ1 ¼ ~εpn þ ð1  dnþ1 Þ2 Dεp . ~¼ J :
eþ
H (27)
Gc =lA

Hydrogen embrittlement modeling

Constitutive equations
Various mechanisms contribute to hydrogen induced
cracking, among which the HELP and HEDE mechanisms are Huang and Gao [44] adopted the von Mises plasticity theory in
incorporated into the phase filed model in this study. Birn- their phase field model. They assumed that only the deviatoric
baum and Sofronis [14] suggested that dissolved hydrogen part of the total stress and the tensile part of the volumetric
atoms in steels promote the mobility of dislocation, thus stress are degraded as the phase field value increases.
decrease the local yield stress. Sofronis et al. [49] and Liang Therefore, the stress tensor can be calculated by
et al. [50] proposed a HELP model that assumed the yield 2 2
stress to be a function of the hydrogen concentration and the s ¼ ½1  Hðεkk Þd Kεkk I þ 2ð1  dÞ mεedev ; (28)
plastic strain to phenomenologically describe the HELP ef- where Hðεkk Þ represents the Heaviside function. And in time
fect as interval ½tn ; tnþ1 , the trial elastic strain tensor is calculated by
 
  εp N εenþ1
trial
¼ εen þ Dε; (29)
sy εp ; c ¼ s0 ðcÞ 1 þ ; (23)
ε0
where εen is the elastic strain tensor updated at the end of the
with increment n and Dε indicates the total strain increment tensor
8  at increment n þ 1. Then, the volumetric elastic trail stress
< ðw  1Þ c s0 s0 ðcÞ > hs0 tensor penþ1
trial
and the deviatoric elastic trial stress tensor senþ1
trial
s0 ðcÞ ¼ c0 ; (24)
: are expressed as
hs0 s0 ðcÞ  hs0
   2 e trial
where c represents the hydrogen concentration, s0 ðcÞ repre- penþ1
trial
¼ 1  H εekktrial
nþ1 d Kεkk nþ1 I;
(30)
sents the initial yield stress with the HELP effect, E is the senþ1
2
¼ 2ð1  dÞ mεedev nþ1 ;
trial trial

Young’s modulus, ε0 ¼ s0/E with s0 ¼ s0 (0) indicating the

initial yield condition with no presence of hydrogen, N rep- The return-mapping algorithm is used to integrate the
resents the strain hardening exponent, εp is the plastic strain, plastic rate equation, details can be found in Ref. [44].
w is a softening parameter, and hs0 represents the lower
bound of the yield stress with h having a constant value be- Yield function
tween 0 and 1, indicating the maximum HELP effect on the
yield stress. To incorporate the HELP effect in the phase field model, Eq.
The HEDE effect is a result of the reduced bonding energy (24) should be combined with Eq. (6). Therefore, the yield
between metal atoms and is often reflected by the reduction of condition becomes a function of the hydrogen concentration,
cohesive strength along the grain boundaries [51], the fracture the phase field value and the effective plastic strain
mechanism transition from ductile to brittle [28], and the f ðs; c; d; εp Þ ¼ sðdÞ  sy ðc; d; εp Þ: (31)
reduction of surface energy. Here a phenomenological model
is used to account for the HEDE effect, in which the critical The proposed form of the yield function is expressed as
energy release rate is assumed to be a decreasing function of
f ðs; c; d; εp Þ ¼ gðdÞs  sðcÞgðdÞsy ðεp Þ; (32)
the hydrogen concentration
where gðdÞ is the degradation function as defined in Eq. (5),
Gc ¼ Gc ðcÞ; (25) and sðcÞ is the softening function similar to Eq. (24)
and the function Gc ðcÞ is assumed to take the form of 8 
< ð1  wÞ c sðcÞ > h
8 
sðcÞ ¼ c0 : (33)
< ðz  1Þ c Gc Gc ðcÞ > xGc :
Gc ðcÞ ¼ c0 ; (26) h sðcÞ  h
:
xGc Gc ðcÞ  xGc Here w and h are the same as in Eq. (24). Noted that sðcÞ is a
where z and x are parameters controlling the reduction of the dimensionless softening factor.
critical energy release rate, with xGc being a lower bound
value. It is worth pointing out that Eq. (26) is not derived from
experimental data nor from atomistic simulations. It is rather
Numerical implementation
to provide a simple model to account for the effect of
hydrogen on the critical energy release rate phenomenologi-
A user subroutine UEL for a commercial finite element soft-
cally. Note that Eqs. (24) and (26) have similar forms. With the
ware ABAQUS is developed to implement the modified phase
definition of Gc , the crack driving force function given by Eq. field model, with the phase field value treated as the degrees
(2) now can be expressed as

Please cite this article as: Huang C, Gao X, Phase field modeling of hydrogen embrittlement, International Journal of Hydrogen Energy,
https://doi.org/10.1016/j.ijhydene.2020.05.015
6 international journal of hydrogen energy xxx (xxxx) xxx

of freedom #11. Details of the numerical procedure is

Table 1 e Material parameters.
described in Ref. [44]. The hydrogen diffusion governing
equation is solved using an ABAQUS user subroutine UMATHT Properties Values Units
[21]. A common block is used to pass variables between UEL Young’s modulus E ¼ 200 GPa
and UMATHT. Possion’s ratio y ¼ 0:3 e
The HELP effect is included in the phase field model by Yield stress sy ¼ 500 MPa
Hardening modulus h ¼ 100 MPa
taking into account the hydrogen effect on the yield stress as
Critical energy release rate Gc ¼ 20 mJ=mm2
described in Eqs. (32) and (33). The HEDE effect is included in
Fracture strain εf ¼ 0:05 e
the phase filed model by varying the critical energy release Coefficient a a ¼8 e
rate according to Eqs. (25) and (26). A previous study by Huang Diffusion coefficient DL ¼ 0:0127 mm2 =s
et al. [21] suggests that HELP embrittles ductile materials by Molar volume of iron VM ¼ 7160 mm3 =mol
accelerating the growth of micro-voids, and another study by Lattice site density NL ¼ 8:46  1019 atoms=mm3
Huang and Gao [44] suggests that lowering the critical energy Molar volume of hydrogen in solid VH ¼ 2000 mm3 =mol
solution
release rate reduces material’s strength as well as ductility. In
Binding energy WB ¼ 60 kJ=mol
the present hydrogen embrittlement model, both HELP and Number of NILS per solvent atom bL ¼ 6 e
HEDE are accounted for. To demonstrate the proposed model, Number of sites per trap bT ¼ 1 e
numerical simulations of a compact tension (CT) specimen
and a double notched flat specimen are conducted for four
hydrostatic stress moves with the crack tip [44]. Fig. 3 shows
cases: 1) no HELP or HEDE effect, 2) only HELP effect, 3) only
the distributions of crack phase field value and lattice
HEDE effect, and 4) with both HELP and HEDE effects. Both
hydrogen concentration on the mid-surface of the specimen
specimens are under displacement-control. The loading
after some amount of crack propagation. The red color in
speeds are set to be slow enough to ensure sufficient time for
Fig. 3(a) represents the new crack surfaces. Fig. 3(b) shows the
lattice hydrogen diffusion so that the hydrogen concentration
area of high lattice hydrogen distribution moves to the new
field remains steady state during the loading process. Assume
crack tip where positive hydrostatic stress exists. These re-
a uniformly initial hydrogen distribution in both specimens
sults are in accord with experimental observations by previ-
with the initial value CL ¼ 2.084  1012 atoms/mm3, and the
ous researchers [52,53].
trapped hydrogen concentration CT is calculated according to
Figs. 4 and 5 show the distributions of crack phase field
the equilibrium relation. The “no flux boundary condition” is
value, trapped hydrogen concentration and total hydrogen
imposed on the exterior surfaces of both specimens.
concentration prior to and after fracture initiation respec-
Hydrogen redistribution in the specimen is driven by the
tively. Here Figs. 4(a) and 5(a), showing the crack phase field
varying stress and deformation fields. The material’s me-
distributions, are the same as Figs. 2(a) and 3(a) respectively,
chanical properties and parameters for hydrogen diffusion are
whose purpose is to indicate the crack tip location. Prior to the
listed in Table 1 [21,44]. The length scale is assumed to be l ¼
onset of crack initiation, trapped hydrogen is concentrated in
1:0 mm. The numerical results of the CT specimen are pre-
a small region ahead of the crack tip where plastic deforma-
sented and discussed first in the next section, followed by the
tion takes place, Fig. 4(b). After crack starts propagating, more
results of the double notched specimen.
hydrogen is trapped along the newly created crack surfaces as
shown in Fig. 5(b). This agrees with the atomistic simulation
result [48] that trapped hydrogen atoms tend to stay on the
Compact tension specimen
newly created crack surfaces.
Fig. 4(c) shows the distribution of total hydrogen concen-
In this section, simulations of a CT specimen are conducted
tration (lattice hydrogen plus trapped hydrogen) prior to crack
and the numerical results are discussed. Fig. 1(a) shows the
initiation, where the hydrogen concentration is the highest at
dimensions (in mm) of the CT specimen having a thickness of
the crack tip region. Fig. 5(c) shows the distribution of total
2.5 mm. Fig. 1(b) displays the finite element mesh showing
hydrogen concentration ahead of the crack tip after some
refined mesh in the region where fracture is expected to occur.
amount of crack propagation. Note that in Fig. 5(c), fractured
The minimum element size is 0.25 mm. Results of hydrogen
elements with phase field value equal to one are removed
diffusion and hydrogen trapping of the case with no HELP or
from the picture, which is for the purpose of a better illus-
HEDE effect are presented and discussed first, followed by the
tration of the total hydrogen concentration in the material.
di scussion of the effects of HELP and HEDE on hydrogen
Fig. 5(c) indicates that hydrogen keeps transporting to and
embrittlement.
being trapped in the area ahead of the growing crack tip.
It is worth mentioning that by comparing the crack tip
Lattice hydrogen diffusion and hydrogen trapping
hydrogen concentrations shown in Figs. 4(c) and 5(c), it is
noticed that the total hydrogen concentration at the crack tip
Fig. 2 shows the distributions of crack phase field value and
region decreases as crack propagates. This is because the total
lattice hydrogen concentration on the mid-surface of the
amount of hydrogen is fixed in the simulation (no flux
specimen just before the onset of crack initiation. The lattice
boundary condition on the exterior surfaces) and more
hydrogen concentration is high in the area ahead of the crack
hydrogen is trapped along the newly created crack surfaces as
tip as a result of high positive hydrostatic stress in this area.
crack propagates. This result agrees with the atomistic
As the phase field values reaches one, crack starts to propa-
simulation result by Song and Curtin [37].
gate. During crack propagation, the area of high positive

Please cite this article as: Huang C, Gao X, Phase field modeling of hydrogen embrittlement, International Journal of Hydrogen Energy,
https://doi.org/10.1016/j.ijhydene.2020.05.015
 international journal of hydrogen energy xxx (xxxx) xxx 7

Fig. 1 e (a) Dimensions (in mm) of a CT specimen. (b) Finite element mesh of the CT specimen.

Fig. 2 e Distributions of crack phase field value (a) and Fig. 3 e Distributions of rack phase field values (a) and
lattice hydrogen concentration (b) on the mid-surface of the lattice hydrogen concentration (b) on the mid-surface of the
specimen before the onset of crack initiation. specimen after some amount of crack propagation.

Hydrogen embrittlement mechanisms

The simulation results presented above confirms that the
numerical model is capable of capturing the hydrogen diffu-
To demonstrate the influence of HELP and HEDE on hydrogen
sion and hydrogen trapping mechanisms and correctly
embrittlement, four cases, 1) no HELP or HEDE effect, 2) only
simulating the hydrogen transport phenomenon. Coupling
HELP effect, 3) only HEDE effect, and 4) with both HELP and
the hydrogen concentration with the HELP and HEDE mech-
HEDE effects, are analyzed and compared. Case 1 serves as a
anisms, parameters affecting hydrogen embrittlement can be
baseline for comparison. The HELP effect on the yield stress is
analyzed and discussed.

Please cite this article as: Huang C, Gao X, Phase field modeling of hydrogen embrittlement, International Journal of Hydrogen Energy,
https://doi.org/10.1016/j.ijhydene.2020.05.015
8 international journal of hydrogen energy xxx (xxxx) xxx

Fig. 5 e Distributions of crack phase field value (a), trapped

hydrogen concentration (b), and total hydrogen
concentration (c) on the mid-surface of the specimen after
Fig. 4 e Distributions of crack phase field value (a), trapped
some amount of crack propagation.
hydrogen concentration (b), and total hydrogen
concentration (c) on the mid-surface of the specimen prior
to fracture initiation.
Comparing the result of case 2, which includes only the
HELP effect, to the baseline case, where neither HELP nor
HEDE effect is considered, the results show that HELP reduces
defined by Eq. (32), and the values of w and h are set to be 0.9 the load carrying capacity of the specimen, i.e., it reduces the
and 0.5 respectively. The HEDE effect on the critical energy strength and fracture toughness of the material. Moreover,
release rate is defined by Eq. (26), and the values z and x are HELP leads to more plastic deformation in the specimen as
also set to be 0.9 and 0.5 respectively. Fig. 6 compares the load- reflected by a higher value of plastic strain at the same applied
displacement curves (solid lines) for the four cases. The vari- displacement in the maximum plastic strain-displacement
ations of the maximum plastic strain during the loading his- curve. These results agree with previous findings of the
tory (dashed lines) for the four cases are also shown in Fig. 6. HELP effect on ductile fracture [19,21]. Comparing the result of

Please cite this article as: Huang C, Gao X, Phase field modeling of hydrogen embrittlement, International Journal of Hydrogen Energy,
https://doi.org/10.1016/j.ijhydene.2020.05.015
 international journal of hydrogen energy xxx (xxxx) xxx 9

Fig. 6 e Load-displacement curves (solid lines) and maximum plastic strain-displacement curves (dashed lines) for different
cases, where the HELP parameters are w ¼ 0:9 and h ¼ 0:5, and the HEDE parameters are z ¼ 0:9 and x ¼ 0:5.

Fig. 7 e Load-displacement curves (solid lines) and maximum plastic strain-displacement curves (dashed lines) obtained
with different HELP parameters.

case 3, which includes only the HEDE effect, with the baseline carrying capacity decreases as the value of w decreases. In
case, it is seen that the HEDE effect also lowers the strength addition, the maximum value of plastic strain at fracture
and ductility of the material, however there is no noticeable initiation is higher when the HELP effect is considered, and it
difference between the maximum plastic strain-displacement increases as the value of w decreases.
curves of the two cases before crack initiation. With both HELP Similarly, to further discuss the HEDE effect, simulations
and HEDE mechanisms included in the simulation, case 4 are conducted with different values of parameter z. Fig. 8
exhibits the lowest load carrying capacity. compares the results obtained from three cases: (z ¼ 0:9; x ¼
To further discuss the HELP effect, simulations are 0:5), (z ¼ 0:8; x ¼ 0:5), and the baseline case. With only HEDE
conducted with different values of parameter w. Fig. 7 com- effect, the load carrying capacity decreases as the value of z
pares the results obtained from three cases: (w ¼ 0:9; h ¼ 0:5), decreases. In addition, the maximum value of plastic strain at
(w ¼ 0:8; h ¼ 0:5), and the baseline case. The sudden drop of fracture initiation is lower when the HEDE effect is consid-
the load-displacement curve indicates the onset of crack ered, and it decreases as the value of z decreases.
propagation. The red dots on the dashed lines represent the Fig. 9 compares the hydrogen embrittlement effects (HELP
maximum plastic strain values when the crack start to prop- and HEDE mechanisms) at different loading stages. The HELP
agate. Fig. 7 shows that with only HELP effect, the load and HEDE effects are evaluated by the values of s0 ðcÞ=s0 and

Please cite this article as: Huang C, Gao X, Phase field modeling of hydrogen embrittlement, International Journal of Hydrogen Energy,
https://doi.org/10.1016/j.ijhydene.2020.05.015
10 international journal of hydrogen energy xxx (xxxx) xxx

Fig. 8 e Load-displacement curves (solid lines) and maximum plastic strain-displacement curves (dashed lines) obtained
with different HEDE parameters.

Gc ðcÞ=Gc respectively. The values of s0 ðcÞ=s0 and Gc ðcÞ= Gc are the red color represents the newly created crack. Fig. 9(b1eb4)
the same because the values of the parameters used to show the distribution of the total hydrogen concentration at
describe the HELP and HEDE effects are the same (w ¼ z ¼ 0:9 the four loading stages. The peak hydrogen concentration is at
and h ¼ x ¼ 0:5). Fig. 9(a1ea4) show the distribution of the the crack tip and its value is decreasing from loading stage 1 to
crack phase field value at four different loading stages, with loading stage 4 because of the no hydrogen diffusion flux

Fig. 9 e Distributions of crack phase field value (a1ea4), total hydrogen concentration (b1eb4), and hydrogen embrittlement
effects (c1ec4) on the mid-surface of the specimen at different loading stages (number 1e4 in each row).

Please cite this article as: Huang C, Gao X, Phase field modeling of hydrogen embrittlement, International Journal of Hydrogen Energy,
https://doi.org/10.1016/j.ijhydene.2020.05.015
 international journal of hydrogen energy xxx (xxxx) xxx 11

further validate the proposed numerical model. Fig. 10(a)

shows the geometry of the double notched specimen having a
thickness of 3 mm, and Fig. 10(b) shows the finite element
mesh with the minimum element size of 0.4 mm. The mate-
rial properties and hydrogen diffusion parameters are the
same as those listed in Table 1, expect for a higher critical
energy release rate, Gc ¼ 60 mJ=mm3 , being used in simula-
tions conducted in this section. Four cases, 1) no HELP or HEDE
effect, 2) only HELP effect, 3) only HEDE effect, and 4) with both
HELP and HEDE effects, are analyzed and compared.

Lattice hydrogen diffusion and hydrogen trapping

The results of hydrogen diffusion and hydrogen trapping for

the baseline the case 1 (no HELP or HEDE effect) are displayed
in Fig. 11. Pictures a1, b1, c1 and d1 show the distributions of
the crack phase field value, the lattice hydrogen, the trapped
hydrogen and the total hydrogen prior to crack propagation
respectively. Pictures a2, b2, c2 and d2 show the correspond-
ing contours after some amount of crack propagation. Similar
to the results of presented in the previous section for the CT
specimen, the lattice hydrogen accumulates at the notch tips
prior to crack propagation, Fig. 11(b1), and the new crack tips
after crack propagation, Fig. 11(b2). More hydrogen is trapped
at the notch tips before crack propagation, Fig. 11(c1), and the
trapped hydrogen remains on the newly created crack sur-
faces after crack propagation, Fig. 11(c2). The total hydrogen
concentration is highest at the notch tips before fracture,
Fig. 10 e (a) Dimensions of a flat specimen with double
Fig. 11(d1), and at the new crack tips after crack propagation,
notches (in mm); (b) finite element mesh.
Fig. 11(d2). The peak value of total hydrogen concentration
reduces as the crack propagates.
boundary condition and more hydrogen being trapped along
the newly created crack surfaces. As a result of the decrease of Hydrogen embrittlement mechanisms
the total hydrogen, the embrittlement effect at the crack tip is
weakening as the loading process continues, Fig. 9(c1ec4). The load-displacement curves (solid lines) and the variations of
The strongest embrittlement effect takes place at the crack the maximum plastic strain curves (dashed lines) during the
tip before crack starts to initiate, Fig. 9(c1), where the values of loading process of the four cases are plotted in Fig. 12, where
s0 ðcÞ=s0 and Gc ðcÞ=Gc are the lowest (~0.86). At this stage, the reductions in strength and ductility caused by the HELP and
embrittlement effects are controlled by the accumulated HEDE effects can be observed. One difference that worth
hydrogen (lattice hydrogen and trapped hydrogen) at the mentioning is that, in Fig. 12, the load-displacement curve of
crack tip as well as the embrittlement parameters w and z. As case 2 is lower and the sudden drop occurs earlier than case 3,
loading continues and crack propagates, the values of s0 ðcÞ= s0 while Fig. 6 shows an opposite trend for the CT specimen. In
and Gc ðcÞ=Gc gradually increase to close to the initial value of Fig. 12, the displacements at which the load curves drop rapidly
0.9 due to diminished hydrogen accumulation at the crack tip. are marked by two vertical dashed lines labeled as “loading
From this stage on, the embrittlement effects are controlled stage 1”, “loading stage 2” and “loading stage 3” for case 2 (with
only by the embrittlement parameters w and z. HELP effect), case 3 (with HEDE effect) and case 4 (with both
Results presented above suggest that the proposed phase HELP and HEDE effects) respectively. This result suggests that,
field model has the ability to predict hydrogen embrittlement for the double notched specimen, the HELP effect on the
resulted from the HELP and HEDE mechanisms. HELP pro- strength and ductility is stronger than the HEDE effect, while it
motes the localization of plastic strain and accelerates mate- is opposite for the CT specimen. This difference may result
rial failure. HEDE reduces the critical energy release rate and from not only the geometrical disparity of the notch tips in two
facilitates crack propagation. The simulation results are in specimens but also the higher critical energy release rate used
qualitative agreement with the previous investigations for the simulations of a flat specimen with a double notch.
[19,21,44]. To analyze how the HELP mechanism embrittles the ma-
terial, Fig. 13 compares the simulation results for case 1 (no
HELP or HEDE) and case 2 (with HELP) at the loading stage 1
Double notched flat specimen indicated in Fig. 12. The HELP effect results in higher plastic
strain at the notch tip, Fig. 13(a1,a2), and higher plastic strain
In this section, simulations of a flat specimen with double results in more trapped hydrogen, Fig. 13(b1,b2), therefore,
notches are conducted, and the results are presented to leading to higher total hydrogen concentration at the notch tip

Please cite this article as: Huang C, Gao X, Phase field modeling of hydrogen embrittlement, International Journal of Hydrogen Energy,
https://doi.org/10.1016/j.ijhydene.2020.05.015
12 international journal of hydrogen energy xxx (xxxx) xxx

Fig. 11 e Distributions of crack phase field value (a1, a2), the lattice hydrogen (b1, b2), the trapped hydrogen (c1, c2), and the
total hydrogen (d1, d2) on the mid-surface of the specimen before (a1, b1, c1, d1) and after (a2, b2, c2, d2) the onset of crack
initiation.

area, Fig. 13(c1,c2). The accumulated hydrogen at the notch tip Fig. 14 examines the HEDE mechanism by comparing the
further promotes local plastic straining in this area, which in simulation results for case 1 (no HELP or HEDE) and case 3
turn facilitates more hydrogen accumulation. As a conse- (with HEDE) at the loading stage 2 indicated in Fig. 12. As
quence of this effect, crack initiation occurs faster for case 2 shown in Fig. 14(a1,a2,b1,b2), the HEDE mechanism also fa-
than case 1, Fig. 13(d1,d2). cilitates plastic straining and promotes hydrogen trapping at

Fig. 12 e Load-displacement curves (solid lines) and maximum plastic strain-displacement curves (dashed lines) for
different cases, where the HELP parameters are w ¼ 0:9 and h ¼ 0:5, and the HEDE parameters are z ¼ 0:9 and x ¼ 0:5.

Please cite this article as: Huang C, Gao X, Phase field modeling of hydrogen embrittlement, International Journal of Hydrogen Energy,
https://doi.org/10.1016/j.ijhydene.2020.05.015
 international journal of hydrogen energy xxx (xxxx) xxx 13

Fig. 14 e Distributions of the effective plastic strain (a1, a2),

the trapped hydrogen (b1, b2), the total hydrogen (c1, c2)
and the crack phase field value (d1, d2) on the mid-surface
of the specimen for case 1 (a1, b1, c1, d1) and case 3 (a2, b2,
Fig. 13 e Distributions of the effective plastic strain (a1, a2), c2, d2) at the loading stage 2 indicated in Fig. 12.
the trapped hydrogen (b1, b2), the total hydrogen (c1, c2)
and the crack phase field value (d1, d2) on the mid-surface
of the specimen for case 1 (a1, b1, c1, d1) and case 2 (a2, b2, Results in Figs. 13 and 14 suggest that the HELP mecha-
c2, d2) at the loading stage 1 indicated in Fig. 12. nism embrittles the material by promoting local plastic
deformation, and it has a strong effect on the hydrogen
distribution at the notch tip. On the other hand, the HEDE
the notch tip. But comparing to Fig. 13(a1,a2,b1,b2), the effect mechanism does not show a strong influence on the
of HEDE on plastic straining and hydrogen trapping is not as hydrogen redistribution prior to crack initiation, but it em-
significant as the effect of HELP. Thus, the difference of the brittles the material by reducing the critical energy release
total hydrogen concentration in Fig. 14(c1,c2) is hardly rate.
noticeable. However, as a result of the reduction of the critical Fig. 15 examines how HELP and HEDE mechanisms work
energy release rate caused by the HEDE effect, the crack together by comparing the simulation results of 4 cases at the
initiation speed is still faster than the baseline case, loading stage 3 indicated in Fig. 12. Both HELP (a2, b2, c2, d2)
Fig. 14(d1,d2). and HEDE (a3, b3, c3, d3) facilitate plastic straining and

Please cite this article as: Huang C, Gao X, Phase field modeling of hydrogen embrittlement, International Journal of Hydrogen Energy,
https://doi.org/10.1016/j.ijhydene.2020.05.015
14 international journal of hydrogen energy xxx (xxxx) xxx

Fig. 15 e Distributions of the effective plastic strain (a1ea4), the trapped hydrogen (b1eb4), the total hydrogen (c1ec4) and
the crack phase field value (d1ed4) on the mid-surface of the specimen for case 1 (a1, b1, c1, d1), case 2 (a2, b2, c2, d2), case 3
(a3, b3, c3, d3) and case 4 (a4, b4, c4, d4), at the loading stage 3 indicated in Fig. 12.

promote hydrogen trapping at the notch tip. But the effect of HEDE mechanisms, the crack initiation speed is faster than
HEDE is not as significant as the effect of HELP. Thus, the the cases with only one mechanism is considered (d2, d3, d4).
difference of the total hydrogen concentration in c1 and c3 is The above results indicate that the mechanisms of HELP
hardly noticeable. It is worth noting that neither HELP nor and HEDE, when considered separately, each can promote
HEDE alone significantly promotes crack initiation at loading crack initiation and lead to material embrittlement. When
stage 3 (d1, d2, d3). When HELP and HEDE mechanisms work both mechanisms are considered, the combined effect on
together, leading to more plastic straining and hydrogen hydrogen embrittlement is more significant than when only
trapping at the notch tip (a4, b4), the total hydrogen concen- one mechanism is considered. HELP reduces the local yield
tration is close to the result when only HELP mechanism is stress and HEDE reduces the critical energy release rate, and
concerned (c2, c4). As a result of the combination of HELP and the two mechanisms work concurrently after plastic

Please cite this article as: Huang C, Gao X, Phase field modeling of hydrogen embrittlement, International Journal of Hydrogen Energy,
https://doi.org/10.1016/j.ijhydene.2020.05.015
 international journal of hydrogen energy xxx (xxxx) xxx 15

deformation develops. The correlation of both mechanisms is plastic strain and hydrostatic stress. Acta Mater
through the material response influenced by them. 1998;46:1519e26. https://doi.org/10.1016/S1359-6454(97)
00364-9.
[7] Oriani RA. A mechanistic theory of hydrogen embrittlement
of steels. Ber Bunsen Ges Phys Chem 1972;76:848e57. https://
Conclusions doi.org/10.1002/BBPC.19720760864.
[8] Oriani RA. Hydrogen - the versatile embrittler. Corrosion
This study presents a numerical model for hydrogen embrit- 1987;43:390e7. https://doi.org/10.5006/1.3583875.
tlement at the macroscopic scale. A previously proposed [9] Dwivedi SK, Vishwakarma M. Hydrogen embrittlement in
phase field method is extended by incorporating hydrogen different materials: a review. Int J Hydrogen Energy
transport and the resulting HELP and HEDE effects into the 2018;43:21603e16. https://doi.org/10.1016/
j.ijhydene.2018.09.201.
governing equations. The equation controlling hydrogen
[10] Beachem CD. A new model for hydrogen-assisted cracking
diffusion through lattice sites is modified to include the effect (hydrogen “embrittlement”). Metall Mater Trans
of the crack phase field. A new trapping density function is 1972;3:441e55. https://doi.org/10.1007/BF02642048.
proposed so that the trapped hydrogen remains on the newly [11] Robertson IM, Birnbaum HK. An HVEM study of hydrogen
created crack surfaces but the concentration does not increase effects on the deformation and fracture of nickel. Acta
during crack propagation. The HELP effect is modeled by Metall 1986;34:353e66. https://doi.org/10.1016/0001-
6160(86)90071-4.
reducing the yield stress with the presence of hydrogen and
[12] Birnbaum HK, Sofronis P. Hydrogen-enhanced localized
the HEDE effect is modeled by reducing the critical energy
plasticity-a mechanism for hydrogen-related fracture. Mater
release rate with hydrogen concentration. A compact tension Sci Eng, A 1994;176:191e202. https://doi.org/10.1016/0921-
specimen and a flat specimen with a double notch are used to 5093(94)90975-X.
demonstrate the numerical model. Results of a series of nu- [13] Ferreira PJ, Robertson IM, Birnbaum HK. Hydrogen effects on the
merical simulations suggest that: interaction between dislocations. Acta Mater 1998;46:1749e57.
https://doi.org/10.1016/S1359-6454(97)00349-2.
[14] Sofronis P, Robertson IM. Transmission electron microscopy

Hydrogen atoms tend to accumulate at the crack/notch tip
observations and micromechanical/continuum models for
region since the positive hydrostatic stress in this area the effect of hydrogen on the mechanical behaviour of
promotes lattice hydrogen cumulation and the plastic metals. Philos Mag A 2002;82:3405e13. https://doi.org/
deformation in this area generates more traps. 10.1080/01418610208240451.

HELP promotes the localization of plastic strain and ac- [15] Matsui H, Kimura H, Moriya S. The effect of hydrogen on the
celerates material failure. The load carrying capacity of the mechanical properties of high purity iron I. Softening and
specimen decreases as the value of the HELP parameter w hardening of high purity iron by hydrogen charging during
tensile deformation. Mater Sci Eng 1979;40:207e16. https://
decreases.
doi.org/10.1016/0025-5416(79)90191-5.

HEDE reduces the critical energy release rate and facilitates [16] Moriya S, Matsui H, Kimura H. The effect of hydrogen on the
crack propagation. The load carrying capacity of the spec- mechanical properties of high purity iron II. Effect of quenched-
imen decreases as the value of the HEDE parameter z in hydrogen below room temperature. Mater Sci Eng
decreases. 1979;40:217e25. https://doi.org/10.1016/0025-5416(79)90192-7.

The proposed numerical model, which combines the HELP [17] Ferreira PJ, Robertson IM, Birnbaum HK. Hydrogen effects on
the interaction between dislocations. Acta Mater
and HEDE effects, can comprehensively simulate hydrogen
1998;46:1749e57. https://doi.org/10.1016/S1359-6454(97)
embrittlement and predict the transition from ductile to
00349-2.
brittle caused by the presence of hydrogen. [18] Robertson IM. The effect of hydrogen on dislocation
dynamics. Eng Fract Mech 1999;64:649e73. https://doi.org/
references 10.1016/S0013-7944(99)00094-6.
[19] Birnbaum HK, Sofronis P. Hydrogen-enhanced localized
plasticity-a mechanism for hydrogen-related fracture. Mater
Sci Eng, A 1994;176:191e202. https://doi.org/10.1016/0921-
[1] Johnson WH. On some remarkable changes produced in iron 5093(94)90975-X.
and steel by the action of hydrogen and acids. Nature [20] Liang Y, Ahn DC, Sofronis P, Dodds RH, Bammann D. Effect of
1875;11:393. https://doi.org/10.1038/011393a0. hydrogen trapping on void growth and coalescence in metals
[2] Robertson IM, Sofronis P, Nagao A, Martin ML, Wang S, and alloys. Mech Mater 2008;40:115e32. https://doi.org/
Gross DW, et al. Hydrogen embrittlement understood. Metall 10.1016/j.mechmat.2007.07.001.
Mater Trans Phys Metall Mater Sci 2015;46:2323e41. https:// [21] Huang C, Luo T, Gao X, Graham SM. Modeling the effect of
doi.org/10.1007/s11661-015-2836-1. hydrogen on ductile fracture. Mater Perform Char
[3] Oriani R. The diffusion and trapping of hydrogen in steel. 2018;7:101e18. https://doi.org/10.1520/mpc20170073.
Acta Metall 1970;18:147e57. https://doi.org/10.1016/0001- [22] Luo T, Huang C, Gao X. An investigation of the effect of
6160(70)90078-7. hydrogen on ductile fracture using a unit cell model. Int J
[4] Sofronis P, McMeeking RM. Numerical analysis of hydrogen Hydrogen Energy 2019;44(16):8627e40. https://doi.org/
transport near a blunting crack tip. J Mech Phys Solid 10.1016/j.ijhydene.2019.02.069.
1989;37:317e50. https://doi.org/10.1016/0022-5096(89)90002- [23] Yu H, Olsen JS, He J, Zhang Z. Hydrogen-microvoid
1. interactions at continuum scale. Int J Hydrogen Energy
[5] Kumnick AJ, Johnson HH. Deep trapping states for hydrogen 2018;43:10104e28. https://doi.org/10.1016/
in deformed iron. Acta Metall 1980;28:33e9. https://doi.org/ j.ijhydene.2018.04.064.
10.1016/0001-6160(80)90038-3. [24] Depover T, Verbeken K. The detrimental effect of hydrogen
[6] Lufrano J, Sofronis P. Enhanced hydrogen concentrations at dislocations on the hydrogen embrittlement susceptibility
ahead of rounded notches and cracks-competition between

Please cite this article as: Huang C, Gao X, Phase field modeling of hydrogen embrittlement, International Journal of Hydrogen Energy,
https://doi.org/10.1016/j.ijhydene.2020.05.015
16 international journal of hydrogen energy xxx (xxxx) xxx

of Fe-C-X alloys: an experimental proof of the HELP [39] Miehe C, Hofacker M, Welschinger F. A phase field model for
mechanism. Int J Hydrogen Energy 2018;43:3050e61. https:// rate-independent crack propagation: robust algorithmic
doi.org/10.1016/j.ijhydene.2017.12.109. implementation based on operator splits. Comput Methods
[25] Petch NJ, Stables P. Delayed fracture of metals under static Appl Mech Eng 2010;199:2765e78. https://doi.org/10.1016/
load. Nature 1952;169:842e3. https://doi.org/10.1038/ j.cma.2010.04.011.
169842a0. [40] Miehe C, Hofacker M, Scha € nzel LM, Aldakheel F. Phase field
[26] Petch NJ. The lowering of fracture-stress due to surface modeling of fracture in multi-physics problems. Part II.
adsorption. Philos Mag A 1956;1:331e7. https://doi.org/ Coupled brittle-to-ductile failure criteria and crack
10.1080/14786435608238106. propagation in thermo-elastic-plastic solids. Comput
[27] Tromans D. On surface energy and the hydrogen Methods Appl Mech Eng 2015;294:486e522. https://doi.org/
embrittlement of iron and steels. Acta Metall Mater 10.1016/j.cma.2014.11.017.
1994;42:2043e9. https://doi.org/10.1016/0956-7151(94)90029- [41] Ambati M, Gerasimov T, De Lorenzis L. Phase-field modeling
9. of ductile fracture. Comput Mech 2015;55:1017e40. https://
[28] Depover T, Wallaert E, Verbeken K. Fractographic analysis of doi.org/10.1007/s00466-015-1151-4.
the role of hydrogen diffusion on the hydrogen [42] Ambati M, Kruse R, De Lorenzis L. A phase-field model for
embrittlement susceptibility of DP steel. Mater Sci Eng, A ductile fracture at finite strains and its experimental
2016;649:201e8. https://doi.org/10.1016/j.msea.2015.09.124. verification. Comput Mech 2016;57:149e67. https://doi.org/
[29] Troiano RA. The role of hydrogen and other interstitials in 10.1007/s00466-015-1225-3.
the mechanical behavior of metals. Trans ASM [43] Borden MJ, Hughes TJR, Landis CM, Anvari A, Lee IJ. A phase-
1960;52:54e80. field formulation for fracture in ductile materials: finite
[30] Jiang DE, Carter EA. First principles assessment of ideal deformation balance law derivation, plastic degradation, and
fracture energies of materials with mobile impurities: stress triaxiality effects. Comput Methods Appl Mech Eng
implications for hydrogen embrittlement of metals. Acta 2016;312:130e66. https://doi.org/10.1016/j.cma.2016.09.005.
Mater 2004;52:4801e7. https://doi.org/10.1016/ [44] Huang C, Gao X. Development of a phase field method for
J.ACTAMAT.2004.06.037. modeling brittle and ductile fracture. Comput Mater Sci
[31] Serebrinsky S, Carter EA, Ortiz M. A quantum-mechanically 2019;169:109089. https://doi.org/10.1016/
informed continuum model of hydrogen embrittlement. J J.COMMATSCI.2019.109089.
Mech Phys Solid 2004;52:2403e30. https://doi.org/10.1016/ [45] Krom AHM. Numerical modelling of hydrogen transport in
J.JMPS.2004.02.010. steel. 1998.
[32] Martı́nez-Pan ~ eda E, Golahmar A, Niordson CF. A phase field [46] Kumnick AJ, Johnson HH. Deep trapping states for hydrogen
formulation for hydrogen assisted cracking. Comput in deformed iron. Acta Metall 1980;28:33e9. https://doi.org/
Methods Appl Mech Eng 2018;342:742e61. https://doi.org/ 10.1016/0001-6160(80)90038-3.
10.1016/j.cma.2018.07.021. [47] Taha A, Sofronis P. A micromechanics approach to the study
[33] Wang S, Martin ML, Robertson IM, Sofronis P. Effect of of hydrogen transport and embrittlement. Eng Fract Mech
hydrogen environment on the separation of Fe grain 2001;68:803e37. https://doi.org/10.1016/S0013-7944(00)00126-
boundaries. Acta Mater 2016;107:279e88. https://doi.org/ 0.
10.1016/j.actamat.2016.01.067. [48] Song J, Curtin WA. Atomic mechanism and prediction of
[34] Griffith AA. The phenomena of rupture and flow in solids. hydrogen embrittlement in iron. Nat Mater 2013;12:145e51.
Philos Trans R Soc Lond A Math Phys Sci 1921;221:163e98. https://doi.org/10.1038/nmat3479.
https://doi.org/10.1098/rsta.1921.0006. [49] Sofronis P, Liang Y, Aravas N. Hydrogen induced shear
[35] Francfort GA, Marigo JJ. Revisiting brittle fracture as an localization of the plastic flow in metals and alloys. Eur J
energy minimization problem. J Mech Phys Solid Mech Solid 2001;20:857e72. https://doi.org/10.1016/S0997-
1998;46:1319e42. https://doi.org/10.1016/S0022-5096(98) 7538(01)01179-2.
00034-9. [50] Liang Y, Sofronis P, Aravas N. On the effect of hydrogen on
[36] Bourdin B, Francfort GA, Marigo JJ. Numerical experiments in plastic instabilities in metals. Acta Mater 2003;51:2717e30.
revisited brittle fracture. J Mech Phys Solid 2000;48:797e826. https://doi.org/10.1016/S1359-6454(03)00081-8.
https://doi.org/10.1016/S0022-5096(99)00028-9. [51] Oriani RA, Josephic PH. Equilibrium aspects of hydrogen-
[37] Bourdin B, Francfort GA, Marigo J-J. The variational approach induced cracking of steels. Acta Metall 1974;22:1065e74.
to fracture. J Elasticity 2008;91:5e148. https://doi.org/10.1007/ https://doi.org/10.1016/0001-6160(74)90061-3.
s10659-007-9107-3. [52] Frohmberg RP, Barnett WJ, Troiano AR. Delayed failure and
[38] Miehe C, Welschinger F, Hofacker M. Thermodynamically hydrogen embrittlement in steel. WADC Technical Report
consistent phase-field models of fracture: variational 54-320. 1954.
principles and multi-field FE implementations. Int J Numer [53] Johnson H, Army F. Hydrogen , crack initiation , and delayed
Methods Eng 2010;83:1273e311. https://doi.org/10.1002/ failure in steel. OstiGov 1956;3.
nme.2861.

Please cite this article as: Huang C, Gao X, Phase field modeling of hydrogen embrittlement, International Journal of Hydrogen Energy,
https://doi.org/10.1016/j.ijhydene.2020.05.015