Journal Pre-proof

Integrated computation of corrosion: Modelling, simulation and

applications

Chaofang Dong , Yucheng Ji , Xin Wei , Aoni Xu , Dihao Chen ,

Ni Li , Decheng Kong , Xiejing Luo , Kui Xiao , Xiaogang Li

PII: S2667-2669(21)00008-6
DOI: https://doi.org/10.1016/j.corcom.2021.07.001
Reference: CORCOM 8

To appear in: Corrosion Communications

Received date: 6 April 2020

Revised date: 28 June 2020
Accepted date: 2 July 2020

Please cite this article as: Chaofang Dong , Yucheng Ji , Xin Wei , Aoni Xu , Dihao Chen ,
Ni Li , Decheng Kong , Xiejing Luo , Kui Xiao , Xiaogang Li , Integrated computation of
corrosion: Modelling, simulation and applications, Corrosion Communications (2021), doi:
https://doi.org/10.1016/j.corcom.2021.07.001

This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition
of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of
record. This version will undergo additional copyediting, typesetting and review before it is published
in its final form, but we are providing this version to give early visibility of the article. Please note that,
during the production process, errors may be discovered which could affect the content, and all legal
disclaimers that apply to the journal pertain.

© 2021 Published by Elsevier B.V. on behalf of Institute of Metal Research, Chinese Academy of
Sciences.
This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/)
Highlights
 Integrated computation of corrosion has been generally summarized.
 Cross-scale models of pitting or SCC are established from atom to macro-scale.
 Artificial intelligence methods can be used to design materials with good corrosion
resistance.
Review

Integrated computation of corrosion: Modelling, simulation and

applications
1,*
Chaofang Dong , Yucheng Ji 1, Xin Wei 2, Aoni Xu 1, Dihao Chen 1, Ni Li 1, Decheng Kong 1,
Xiejing Luo 1, Kui Xiao 1, Xiaogang Li 1
1
Corrosion and Protection Center, Beijing Advanced Innovation Center for Materials Genome
Engineering, University of Science and Technology Beijing, Beijing 100083, China
2
School of Physics and Materials Engineering, Hefei Normal University, Hefei 230601, China
* Corresponding author.

E-mail address: cfdong@ustb.edu.cn (C.F. Dong).

Article history:

Received 6 April 2020

Received in revised form 28 June 2020

Accepted 2 July 2020

Available online

Abstract:In the last decade, integrated computation of corrosion has made significant
progress towards the atomic-scale clarification of corrosion mechanisms and computer-aided
designing of advanced materials with excellent corrosion resistance. This review focuses on
the theoretical calculation methods and developing tendency in corrosion study, and three
specific applications are presented. First-principle techniques combined with molecular
dynamics method, peridynamic theory and finite element method provide multiscale models
to investigate micro-mechanisms of stress corrosion cracking and hydrogen-induced cracking.
Calculations of passivity and passive film breakdown are elaborated through point defects
diffusion and its correlation of the energy level degeneracy. By surveying publications, the
artificial intelligence technology is pointed out how the computer can pave the way of
predicting corrosion degrees as well as designing new corrosion resistant materials. To get
better and efficient development of integrated computation of corrosion, extensive
cooperation and powerful data infrastructure are needed by stronger collaboration in the
future.

Key words: Corrosion; Computation; Modelling; Degradation evaluation

1. Introduction

The driving force for development of modelling, calculation and simulation methods is
essentially the dynamic of exploring the laws of nature, from unknown to knowledge, then forward
to the further unknown in the coming future through creating and predicting [1-3]. Especially in the
research field of materials science, these techniques provide powerful tools to establish numerical
principles, analyze digital processes and discover advanced materials [4-8]. The joint use of
multiple systematic computing tools has been defined in many ways, the most famous of which are
Integrated Computational Materials Engineering (ICME) and Materials Genome Engineering
(MGE) [9]. These projects emphasize aims such as cost reduction, efficiency improvement, data
platform and engineering application for accelerating the innovation process of advanced materials
[10-14]. The evaluation of service performance (including corrosion, fatigue, cracking, creep,
high-temperature oxidation and erosion) is a vital part of the whole chain working in sync [15-18].

Corrosion, mentioned as one of the most challenging issues in materials computation owing to
complicated behavior [19, 20], is evident that the reactions and their influence are highly dependent
on the composition and structure of matrix metals or alloys [21-24], service loads [25, 26],
environmental parameters [4, 14, 27], interface condition [26, 28-33], of the comprehensive system
which exhibits “living” changes over time. To achieve better and more accurate results, research
team of corrosion modelling and computational data analysis must embrace superior computer
technology which has been developed with dramatic growth and rapid progress [34-39]. The
emerging computer-aided techniques in physics, chemistry or mechanics may also be accompanied
by the new generation methods and research hotspots of corrosion computing under certain
conditions [40-42]. The number of computing examples with interdisciplinary characteristics
available for corrosion investigating of metals and coatings has proliferated greatly during recent
years [43-47]. Many new software or network infrastructures are ICME and high-throughput
techniques which are suitable for fast and large scale computing of length ranging from nanometer
(an adsorption monolayer or passive film) to micrometer (pitting or cracks) [48]. The Questeck,
Materials Project, AFlow, MatNavi, Open Quantum Materials Database, etc. provide steps and
flow-tools for innovating materials composition and structure based on calculated E-pH diagrams,
work function value or band gap data [49-51]. These computational tools are still being improved
and updated to reach a higher degree of accuracy and reliability.

An online searching about the number of publications that used Corrosion, Computation and
Simulation as keywords to study the behavior of electrochemical reaction or materials failure has
been done here (Fig. 1), and the data is surveyed on www.sciencedirect.com from 1990 to present.
In total, 35,940 articles have been identified. From the histogram, the statistical analysis represents
the number of research articles employing main calculation methods, these methods are divided into
seven categories which are most noteworthy and widely spread.

The retrieval method of the subject words will cover some articles based on material
calculation, in which corrosion is only a performance parameter of materials, and the number of
articles will be greater than the literature based on corrosion. However, we can still see that the
research heat has an obvious upward trend. More precise time-space scale classification methods
and their applications will be summarized in the next part.

2. What you can do with integrated computation of corrosion

There are three main branches which can be launched by coupling of calculations and
experiments in corrosion science [52-56]. First, the multiscale computation is so difficult in many
cases, and its output parameter may not be logical and suitable as the input value of associated
scaling hierarchy. This is also the particular vital part in the MGE discussion of the transfer data
between various models above. By this feature, the thermodynamic and kinetic parameters must
therefore be carefully considered whether they have reasonable meaning and standard data interface
from a physical and chemical point of view [57]. Secondly, the electrochemical reaction dominates
the corrosion process, which occurs heterogeneously at a metal/environment interface. The
reactants and products are anisotropic distributed on the local area, the electric field and
concentration difference play a major role in the direction perpendicular to the matrix surface while
the two-dimensional adsorption and diffusion control the behavior in the parallel direction [58].
Comparing to the study of bulk materials, the experiment is more challenging to capture dynamic
changing of the monolayer or oxide film which is only around tens and hundreds of nanometers,
adding to its complexity [59-61]. On the contrary, it is a good object for theoretical modeling and
simulation which can solve the above problem well at the microscopic scale [62-64]. Finally, in the
case of engineering application such as the high-speed train or oversea bridge, their corrosion
behavior is frequently difficult to be evaluated in a changing environment, and the use of short time
or small sampling data of known outdoor exposure tests is open to question. This problem calls
forward a clear demand for monitoring, detection and computer-aided assessment of corrosion big
data [65, 66]. In particular, consideration of the blooming development of information technology
can still provide surprising theoretical finding out of experimental observation, and can often be
described comprehensively and predictively [67, 68]. Moreover, we can also use artificial
intelligence (AI) techniques to analyze and mine the data generating from the above research
branches, so as to develop new materials with better corrosion resistance [69-72].

Integrated computation of corrosion is a set of techniques for investigating materials

degradation problems on workstation or supercomputing center. In studying corrosion mechanisms
researchers have a selection of methods at their requirement. From Fig. 1, the main tools available
belong to six broad classes. The scale of passivation film, which plays an important role in metallic
materials, is located at the nanometer. To study the decomposition and repassivation of passivation
film, first-principle calculations (including ab initio and density function theory) are required.
First-principle calculation uses periodic conditions to construct surface or interface structure, which
scale can reach several nanometers to several dozen nanometers [73]. The stability of the
passivation film is estimated by calculating the atomic valence changes and the most stable
structure [74]. First-principle based on quantum mechanics is accurate in calculation, but it also
consumes huge computing resources, which in turn leads to a smaller calculation scale. Molecular
dynamics (MD) or Monte Carlo (MC) method has great advantages in studying the absorption of
polymers on the surface of materials, and it also has application in dislocation distribution and
phase transition [75-77]. The simulation scale of the MD/MC method is tens of nanometers or even
micrometers. With excellent parallel algorithm optimization, the computing resources of MD/MC
method are greatly reduced. For higher-scale corrosion problems, such as the effect of grains type,
pitting initiation and crack growth, the finite element method (FEM) and phase field methods are
more suitable. FEM sets different properties for meshes to simulate related problems. These
methods not only have high calculation accuracy but also can adapt to complex shapes above
micrometer scale [78, 79].

To find the logical clues and relationships between corrosion concepts and calculation
methods such as “Treasure map”, it is customary to use knowledge graph to obtain optimum path.
Owing to the limitation of information access, those methods are usually confined to published
literature. Using CiteSpace approach, the word analysis of literature has been done on Web of
Science (WOS) from 1950 to 2020 [80, 81]. This paper reveals the hot spots of different research
issues and the methods used, and a typical diagram is given in Fig. 2. The contents in the figure are
as follows: the gray area in the lower left corner is to study the atmospheric corrosion of weathering
steel by MC method; the red part is to study the corrosion initiation mechanism by quantum
chemistry and molecular structure calculation; the green area is to study the oxidation and
dissolution of metal ions by ab initio first-principle method; the pink area is to study the cross scale
fracture of cracks by MD. More information on corrosion types, computational models, theoretical
methods, and experimental tests are summarized in Table 1.

Integrated computation methods and applications are introduced respectively with some
examples in the following parts (Fig. 3). The behaviors of aluminum, copper, steel and their alloys
in atmospheric environment or simple chemical solutions will be considered here to explain the
differences in corrosion phenomenon and its mechanism [15, 28, 33, 140].

3. Multiscale computing of corrosion

3.1. Perspective

While true automatic process computing of multiscale corrosion problem is a key part of our
work in script programming, but in the real situation, most multiscale work on modelling and
simulation is accomplished by passing the parameter values between stand-alone modules or
software [150]. As mentioned by ICME, the means of data transferring contain inherent flaws at
each scale in a theoretical way (see Fig. 4). Besides, the errors and uncertainties become a major
issue when the boundary conditions of the model are assumed and simplified. The experimental
data can provide important support for calibration and certification of the calculated results which
may acquire various information from different groups or researchers [151]. Energy, which has
control over the direction of thermodynamic reaction and structural stability, is an important
parameter to string modules of different scales [16, 48]. Now, stress corrosion cracking (SCC) is
selected to illustrate the computation flow. As such SCC is very sophisticated relative to the pure
mechanical fracture of passive metals such as aluminum or high-strength stainless steel, and results
in a sudden loss of strength. The failure mode for SCC of Al alloys has been widely suggested that
hydrogen embrittlement (HE) increases its susceptibility and causes intergranular cracking. Serving
in atmospheric environment, corrosion rate is thereby rendered low when the metal keeps its
passivity, but it is worth noting that hydrogen-induced decohesion of grain boundaries (GBs) is not
easy to observe — it is inner damage [152-154]. From the adsorption, relaxing and oxidation of
water, oxygen and chloridion on the metal surface to the dissociated hydrogen atoms permeation,
diffusion and segregation, which eventually leads to the cracking of materials, the mechanism in
this process needs to be interpreted from the atomic scale to macroscale. Cross-scale calculations
combined with first-principles, discontinuous peridynamic model and finite element method were
performed to investigate the pitting and SCC in aluminum and its alloys [16, 48].

3.2. The oxide film formation and localized break down

Oxide film formation and its breaking down behavior were studied by the
single/multi-molecules adsorption of H2O, O2 and Cl-. The co-adsorption of H2O and O2 can
accelerate the generation of oxide film which is thicker by the O atoms migration from surface to
inner structure. As mentioned [33], an Al atom bonded with H2O can be pulled out of the surface in
all adsorbed structures because of the electrostatic repulsion between pre-adsorbed O atoms and the
O atoms of H2O (Fig. 5(a)). For the co-adsorption of H2O and O2 on Al(111) surface, H2O
molecules can dissociate into OH and H when both of the two O atoms of O2 can interact with H
atoms of H2O, and the dissociating energy of H2O decreased from 137.58 to 38.18 kJ/mol with the
aid of the pre-adsorbed O atom. For the dissociation adsorption of H2O, the redistribution of charge
occurs and the value of charge transfer from the surface and O2 to H2O is much larger than that of
their respective molecular adsorption (Fig. 5(b))

In most cases for engineering, the data of pure metals may not be directly quoted. By the
way of making the computational results to be close to the actual condition, the corrosion
mechanisms of 2xxx and 7xxx series Al alloys were analyzed to make a distinction between the
oxide films on matrix and that on intermetallic by DFT. The average charge transfer from Al2CuMg
and Al surfaces to adsorbates (OH-/H2O/O2) were 3.16 and 3.38 eV, respectively, indicating that the
electronic interaction between the Al surface and adsorbates was stronger than that between the
Al2CuMg surface and adsorbates. For the Al2CuMg (001) surface, the charge transfer from the
Al-terminated surface to adsorbates (3.3 eV) was larger than that from the mix (Cu, Mg)-terminated
surface to adsorbates (2.92 eV). The reason was that the presence of Cu and Mg atoms on the
termination induced the reduction of charge transfer. The average adsorption energy of O2 and H2O
molecules on the Al2CuMg surface (9.62 eV) was smaller than that on the Al surface (10.77 eV).
The results showed that O2 and H2O preferred to adsorb on the Al matrix surface, rather than on the
Al2CuMg surface, which was consistent with the AES result [45]. The differences of charge
densities of adsorbed structures of O atoms on Al (100), Al (111), Al (101) surfaces, Al2CuMg (101)
surfaces (S-1, S-2, S-3) and Al2CuMg (001) (J-1, J-2) surfaces are provided in Fig. 5(c–e), and the
geometrical parameters, adsorption energies and charge transfer of H2O and O2 co-adsorbed
structures on the Al2CuMg and Al surfaces are shown in Table 2.

The oxide film formed on the passive metal surface can isolate the substrate from the external
environment and obstruct the next step of corrosion. However, when the corrosive ions such as Cl‒
are existing, the oxide film turns to be unstable, then local weakening could lead to breaking down
and pitting initiation. Frankel [155] and Kruger [156] had successfully summarized the classic
mechanism of pitting initiation, and it can be divided into three categories: (I) ion migration and
permeation of oxide film; (II) ion adsorption/displacement in solution leads to the thinning of
passivation film; and (III) the thinning and repassivation of the film.

As to Al system, the corrosion mechanism of copper in a medium containing Cl− is well

established, while a detailed understanding regarding the interaction between Cl− and Al2O3,
especially at the atomic scale, remains a key issue. The scientific problem is mainly to focus on two
topics regarding the breakdown of the passive film grown on Al and its alloys at the atomic scale.
We have employed the first-principle calculations as the tool to study this topic. The computational
results show that the adsorption of Cl‒ on the surface of aluminum oxide film is an exothermic
reaction (Fig. 6(a), while the permeation process of Cl‒ to the sub-surface is an endothermic
reaction (Fig. 6(b)). With the increasing of Cl‒ concentration, the driving force decreases. It
provides a theoretical explanation that the pitting initiation mechanism tends to Cl‒ adsorption
induced by oxide thinning instead of permeation mechanism.

3.3. Hydrogen diffusion and GBs decohesion

HE, which plays a key role in SCC of Al alloys, leads to distinct mechanical degradation.
There are three fundamental mechanisms for HE, namely (i) hydride-induced embrittlement, (ii)
hydrogen-enhanced decohesion, and (iii) hydrogen-enhanced localized plasticity. Experimental and
theoretical efforts have been made using conventional mechanical and in situ tests to elaborate the
mechanism for HE. Based on the computation method in Section 3.2, the adsorption, solution, and
diffusion of H atom were performed along Al (111) surface to obtain the most stable
adsorption/solution site and the diffusion energy barrier. A 120-atoms grain boundary (GB) and
grain models were constructed to characterize the solution and diffusion of H atom (Fig. 7(a)). H
atom is energetic favorable to dissolve in sites of GB (Fig. 7(b)) and the ability of H diffusion in
GBs is stronger than that in grain inner. H atom in the interstitial site of grain will be easy to
migrate into GBs by addition of Al atom vacancy. Besides, the introduction Al atom vacancy can
decrease the energy barrier of H diffusion in grain inner, while the promotion of Al atom vacancy
on the H diffusion in GBs is related to its migrated direction (Fig. 7(c-f)).

The strength of GBs characterized by its cohesive energy was calculated by first-principle
method, which decreased with the increasing H concentration segregated in GBs. The cohesive
energy is defined as the reversible free energy change for the formation of free surfaces from the
GB. Approximating the free energy as the heat of formation from VASP, the cohesive energy can be
calculated based on the following equation:

GB
Ecoh   surf
1
  surf
2
  GB (1)
1 2
where 𝛾surf and 𝛾surf are the energies of two surfaces that form the GB and 𝛾GB is the GB
energy.

The existence of H atom weakens the metal bond between Al atoms at the GBs, and the
cohesive energy of GBs decreases linearly with the increase of H concentration, that is to say, the
strength of GBs linearly decreases. Compared with intergranular, there are more H atoms trapped at
the GBs, which is a fast diffusion channel to H atoms.
 To expand the application scope and accuracy of the model method, the commonly used 7xxx
series high-strength aluminum alloy was selected to study the effect of precipitated η′ phase. Studies
have also shown that the segregation of Mg atoms at GBs is spontaneous, which weakens the grain
boundary strength [157]. The combination of HE and Mg segregation induced stress corrosion
cracking [158]. Our results have clearly shown the hydrogen atoms can be stably adsorbed on both
Al and MgZn2 surfaces (Fig. 8(a) and (b)) [16]. The adsorption capacity of the precipitate phase
with respect to the hydrogen atom is slightly stronger than that of the Al surface. The most stable
adsorption site on the Al surface is FCC site. The most stable adsorption site of MgZn 2 is the bridge
sites. Mg atoms exert significant influence on the adsorption of H atoms. The strongest adsorption
site is at the grain boundary (see Fig. 8(c) and (d)), with an adsorption energy as high as -3.16 eV.
This indicated that H atoms will concentrate at the grain boundary, resulting in significant
weakening of grain boundary cohesive energy, which is only 13.4% of the cohesive energy of pure
Al-Al GBs.

3.4. Stress concentration and intergranular cracking

Owing to the experimental test is difficult to observe the inner cracks caused by HE,
calculation and simulation methods have been proposed. FEM and damage mechanics model have
been applied to SCC [159]. It is also necessary to investigate the role of GBs damage in
intergranular cracking behavior. The cohesive element approach can be used to explicitly account
for the damage to GBs and has been reported [160, 161]. Simonovski and Cizelj had used a
cohesive element approach to model intergranular cracking at the grain level in which cohesive
elements with zero physical thickness were directly inserted between the adjacent grains [162]. The
fracture energy is one of the most important parameters for cohesive finite elements calculations,
which usually estimated by some approximate algorithms [163]. Cohesive energies, as fracture
energies, were input into cohesive finite element calculations to simulate the intergranular cracking
behavior, which was a cross-scale study. It is a very efficient way to investigate the evolution of
hydrogen induced intergranular cracking. This work provided a scheme through a cross-scale
calculation from atomic scale to macro-scale to investigate the behavior and mechanism of passivity
and hydrogen induced intergranular cracking in metals as shown in Fig. 9 [48].

Based on the stress distribution calculation, the effect of the additional stress caused by the
oxide film at the crack tip of SCC and the effect of pitting on the crack initiation were studied by
the FEM. The oxide film of AA2024-T351 exerts a powerful effect on distributions of stress and
strain at the Wedge opening loading (WOL) specimen notch. The stress and strain in the oxide film
and matrix at the notch increase with the oxide film thickening. Additional stress exists near the
interface between the oxide film and matrix due to the difference of the elastic modulus. This
additional stress increases the stress in the oxide film. On the other hand, it causes relaxation of the
stress and strain within the matrix, which leads to the result that the maximum stress and strain in
the matrix are both localized at the 10 μm ahead of the interface. However, the effect of the
additional stress could be neglected when the thickness of the oxide film is reduced to 5 μm.

The distribution of stress around the pit is uniform. The stress decreases with the distance away
from the pit. On the contrary, the strain near the pit mouth is greater than that at the pit base, which
indicates that the localized plastic strain exists near the pit mouth. Numerical simulation by
extended FEM indicates that the crack evolves from the pit predominantly near the pit mouth and
propagates along the specimen surface [164] (Fig. 10). It is implied that localized strain is the
dominant mechanical factor that affects the initiation and propagation stages of the crack evolving
from pit.

Cracks generation and growth is a dynamic process. The above work gives the stress
concentration and the trends of hydrogen-induced intergranular cracking. As a discontinuous
fracture mechanics problem, the peridynamic (PD) theory provides an algorithm considering the
multi-cracks system and time t (Fig. 11) [16]. The PD models demonstrate that Al-H-MgZn2 GB
can be destroyed faster than pure Al GBs. Under the same condition, the damage rate of
Al-H-MgZn2 GB exceeds 50% at 8.52 ns. However, the damage rate of pure Al GB is estimated to
be 20% at the end of the simulation.

4. Passivation modelling and electrochemical verification

4.1. Perspectives

Passivity, is a double-edged sword for metal anti-corrosion, providing a protective layer

covering the bare metal surface on one side, while inducing pitting corrosion as a result of local
breakdown of the protective layer, named passive film. Therefore, it is very important to understand
the relationship between the component, structure of passive films and their resistance ability, to
minimize the possibility of pitting occurring. However, in most cases, the exact information about
the component and structure of passive films is highly inaccessible since the film is usually
ultra-thin (1–100 nm) and cannot be separated from metal via current characterization techniques.
Atomic simulation, like DFT calculations, MD simulations, etc., thus, offers a feasible approach to
explore the physical properties of this kind of nanoscale film via constructing atomic models.

Passive film usually works as two main topics: ion migration barrier and electron transfer
barrier to prevent the decomposition of itself, as well as metal corrosion. Specifically, the content is
as follows: (I) Ion migration barrier — The reaction between bare metal and aggressive anions, like
Cl-, in service environment, results in the leaching phenomenon of metal atoms, in macroscale,
metal corrosion. The existence of passive film prevents the migration of anions from electrolyte to
metal surface, and hence the attachment between anion and metal, which produces a protective
effect. (II) Electron transfer barrier — Aggressive anions still can work on passive film, though it
cannot reach underlying metal, inducing a local wastage of the layer. Fortunately, passive films
usually exhibit semiconductor properties, which inhibit electron transfer across the film from one
side to another side under low driving force. This property promotes charge accelerating on film
surface, and thus, suppressing further decomposition process.

Many efforts on electrochemical theory contributed to the understanding of passivity.

Meanwhile, charged point defects from Point Defect Models are the only descriptor considering the
inhibition effect of passive film to both ions and electrons together [165]. We will introduce the
investigations of passive films, taking copper as an example, from these two aspects in the
following sections.

4.2 Ion migration barrier

We first investigated the ion migration ability in passive film on copper in non-aggressive
environment to test the accuracy of atomic calculations. Ab initio molecular dynamic simulations
were performed at 298 K, and 101 kPa, aiming at a defective Cu2O surface model (Fig. 12(a))
which was reported as the major content of copper passive film. After running 1.3 ps with a
timestep of 1.0 fs, we obtained the diffusion rate of target point defects, which is of the order of
10-17 cm2/s. To verify the suitability of the atomic migration dynamic simulation, we implemented
relative electrochemical experiments to analyze the average migration rate of point defects in the
copper passive film. Mott-Schottky and impedance were employed to characterize the concentration
of point defects and steady-state film thickness (Fig. 12(b)), then to derive the average diffusion rate
according to Eq. (3) [25]. Interestingly, the value of diffusivity from experiments is in the same
order of 10−17 cm2/s.

After the successful initial test of this atomic study approach, we then applied this method in
an aggressive serve environment to reveal the effect of anions on point defects migration. As shown
in Fig. 13(a), the Cl- absorbed on passive film surface accelerates the migration of cation defects
from 10-17 to 10-14 cm2/s, while the inserted Cl- just slightly increases the diffusivity to 10-16 cm2/s
[27]. Thus, an atomic conclusion could be obtained, where low concentration Cl - has very little
influence on the protection of passive film on copper. However, when there are S2- and Cl- in the
electrolyte, the diffusivity was improved to 10−12 cm2/s (Fig. 13(b) and (c)), assistance with the fact
where S2- works as a booster to enlarge the aggression of Cl- [15].

4.3 Electron transfer barrier

Similarly, we first studied the semiconductor properties of passive film in non-aggressive

environment, and then applied the method to aggressive solution after verified by experimental
results. Here we utilized bandgap as a descriptor for the charge resistance of passive film. We first
did the calculations for Cu2O with different kinds of point defects, as shown in Fig. 14, and found
that (1) oxygen vacancies are hard to create and almost have no effect on band gap or conductivity;
(2) VacancyCu-1 has the lowest formation enthalpies and induce an ultra-shallow acceptor state in
bandgap, which convert the original direct gap to indirect gap; (3) VacancyCu-2 does not change the
direct gap character but the acceptor energy level induced is a little higher than VacancyCu-1. Then
we derived approximate relation expressions between bandgap and vacancy type (V) and

concentration (C), Eg = − 5.4 ·CV +2.18 for VCu-1 and Eg = − 15.6·CV + 2.18 for VCu-2, and verified it

using previous reports. Finally, we applied derived linear relation to analyze our
photo-electrochemical results about the passive film formed on copper and concluded the
compositions of these films which cannot be approximately detected via current technologies [21].
 We then applied this strategy to reveal the effect of aggressive anions on passive film
semiconductor properties. When we detected the bandgap of copper passive film in various Cl -
concentration electrolytes, the results show a passive film with a narrow bandgap, in which the
valence band maximum (VBM) increased upon introducing chloride into the electrolyte (Fig.
15(a)).

DFT calculations indicate that an increase in the copper vacancy concentration due to cation
extraction at the film/solution interface could increase the VBM, while oxygen vacancy generation
results in a decrease in the conduction band minimum (CBM) (Fig. 15(b)). The combination of
these results demonstrated that the aggressive role of chloride where an enhancement of the cation
vacancy concentration across the passive film occurs in response to the absorption of Cl- into
oxygen vacancies [28].

5. AI Assisted Corrosion simulation and materials design

5.1. Perspectives

Corrosion scientists have been using existing material and environmental data to predict
servicing performance, rather than a large number of repeated tests [166]. However, the most
difficult way to achieve this approach is the specific label definition of corrosion behaviors, which
may be used recorded in various types for the same sets of data [167]. The processing methods of
material scientists provide us with very good reference. Liu et al. had used machine learning to
identify the microstructure of the material, predicted the desired structure for performance and
designed the magnetoelastic Fe-Ga alloy to enhance its elastic, plastic and magnetostrictive
properties [168]. García and co-authors evaluated the accuracy of various machine learning
algorithms, such as multiple linear regression, clustering algorithm, decision tree and artificial
neural network in predicting Young's modulus. Due to the non-uniform characteristics of materials,
the error of machine learning results is too large to be accepted, which is also a big problem
unsolved in machine learning of material data [169]. Pilania et al. showed that if we could
effectively learn from the existing information or data, the material design cycle would be greatly
shortened, which is consistent with the "double halving" of MGE [170]. In recent years, data mining
and machine learning technology has been fleetly developed, and frequently-used algorithms
include K nearest neighbor, decision tree and its integrated algorithm, support vector machine,
genetic algorithm, and artificial neural network algorithm. Because of the advantages in dealing
with regression and classification problems, these algorithms and techniques have been employed in
the selection of process parameters, the classification of material states and the mining of
first-principle data. The AI aided decision-making has become an important and efficient tool for
data analysis, material performance prediction and optimization [171-173].

5.2. Corrosion data simulation with AI

Electrochemical methods, such as polarization curves and impedance spectroscopy, are the
most commonly used in corrosion studies. The difference of curve types and the discreteness of data
make it difficult to analyze and apply data automatically in complicated service environment.
Considering the time cost of a large number of repetitive tests, we have tried to adopt the machine
learning method to training model of electrochemical experimental data, and carried out the
simulation of electrochemical curves and corrosion performance prediction under different
environmental parameters [69]. Five machine learning methods were used in a Scikit-learn module
of Python for constructing of electrochemical polarization curves and impedance spectra. As an
application case, the corrosion data of copper used in simulated high-level nuclear waste disposal
environment were selected to do the simulation. The training sets included the curves measured in
different chloride ion concentrations, sulfide concentrations, and temperatures (Fig. 16). Based on
the input weight analysis, sulfide concentration had shown the maximum impact on polarization
curve. The weight analysis of this model agrees with the traditional experimental results. The
accuracy of the model after training is acceptable, and the electrochemical parameters can be drawn
from calculation in the same system when the detailed test data are not sufficient in other conditions
(Fig. 17).

5.3. Computing assisted designing for advanced corrosion resistant materials

The computational models and methods have been increasingly using to study the corrosion
mechanism and support the decision-making of degradation degree. One of the important
applications for integrated computation of corrosion is the computer-aided design and process
optimizing of advanced corrosion resistant materials after modelling and simulation of structure and
interface reaction process. For example, the MD calculation with an interface model including a
total of 5,500,000 atoms, which size is about 382 nm × 207 nm × 15 nm, had been performed to
simulate the Si segregation behavior of 3D printing AlSi10Mg alloy during the cooling process
combining with the first-principle calculations (Fig. 18). According to the calculation results, the
preparation parameters of the optimal structure for better corrosion resistance were determined, and
the additive manufacturing AlSi10Mg alloy was actually printed. A series of materials and corrosion
experiments were carried out to test the properties of the prepared coupons. The results show that
good corrosion resistance can be obtained under the parameters established by calculation [23].

With the continuous focusing on MGE, researchers can obtain a large amount of data through
high-throughput experiments or long-term accumulation. How to process these data and perform
data mining has gradually become a focus of attention in the research field. Furthermore, an attempt
to combine high-throughput computing data with machine learning is ongoing. Using
supercomputing platforms, we have obtained more than 200,000 first-principle calculation data.
However, it is difficult to get good results of some elements model, such as lanthanide and actinide,
in a short time. On the basis of the existing calculation results, the machine learning model is
established to predict the parameter range of complex elements and accelerate the DFT calculation
process. Utilizing a variety of correlation analysis methods, the input variables with high correlation
with the result were screened, and the variables with high correlation with the result were
synthesized, namely the atomic number of the element, electron layers, valence electrons, the
atomic radius, and the unit cell parameter. The prediction ability of the screened model could be
improved by about 15%. The gradient asymptotic decision tree algorithm has the best prediction
accuracy. It means the model is satisfying the requirements for qualitative prediction of complex
structure model, and this AI aided method may greatly decrease the computing time usage of
supercomputing platform.

6. Conclusions and outlook

Determination of the corrosion mechanism of materials or components can be carried out by

integrated computation of corrosion which has a long and stiff way to go. In the future, we need to
optimizing composition and manufacturing processes to increase the corrosion resistance for matrix
with lower cost of products or maintenance. Structure’s safety and lifetime in engineered service
environment depend on the reliable evaluation by which the modelling is established and the
multi-source data in which they are processed. It is worth noting that there are still a lot of
deficiencies in the programming of original code, which makes the development of corrosion
calculation limited by the technology in other fields, the development of multi-scale corrosion
computing system with robustness and high reliability needs more participation of physics,
mathematics, and computer experts.

Therefore, building efficient infrastructure including data, model and computing tools in
corrosion science and engineering is indispensable. Further cross-platform open-source work and
multi-team cooperation will contribute to the fast application and service of computer technology in
corrosion science and engineering.

Acknowledgements
The authors gratefully acknowledge the financial support from the National Key Research and
Development Program of China (No. 2017YFB 0702300), the National Natural Science Foundation
of China (No. 51871028) and the Fundamental Research Funds for the Central Universities (No.
FRF-TP-19-003B2).
References

[1] X. Li, D. Zhang, Z. Liu, Z. Li, C. Du, C. Dong, Materials science: Share corrosion data, Nature
527 (2015) 441-442.
[2] Y. Wang, J. Lv, L. Zhu, Y. Ma, Crystal structure prediction via particle-swarm optimization,
Phys. Rev. B 82 (2010) 094116.
[3] M. Spencer, J. Eickholt, J. Cheng, A Deep Learning Network Approach to ab initio Protein
Secondary Structure Prediction, IEEE-ACM Trans. Comput. Biol. Bioinform. 12 (2015) 103-112.
[4] P. Lu, J.E. Saal, G.B. Olson, T. Li, O.J. Swanson, G.S. Frankel, A.Y. Gerard, K.F. Quiambao, J.R.
Scully, Computational materials design of a corrosion resistant high entropy alloy for harsh
environments, Scr. Mater. 153 (2018) 19-22.
[5] G.I. Csonka, J.P. Perdew, A. Ruzsinszky, P.H.T. Philipsen, S. Lebegue, J. Paier, O.A. Vydrov,
J.G. Angyan, Assessing the performance of recent density functionals for bulk solids, Phys. Rev. B
79 (2009) 155107.
[6] G. Kresse, J. Furthmuller, Efficient iterative schemes for ab initio total-energy calculations using
a plane-wave basis set, Phys. Rev. B 54 (1996) 11169-11186.
[7] S. Plimpton, Fast parallel algorithms for short-range molecular dynamics, J. Comput. Phys. 117
(1995) 1-19.
[8] S. Plimpton, B. Hendrickson, A new parallel method for molecular dynamics simulation of
macromolecular systems, J. Comput. Chem. 17 (1996) 326-337.
[9] M.F. Horstemeyer, Integrated Computational Materials Engineering (ICME) for metals: Using
Uultiscale Uodeling to Invigorate Engineering Design with Science, John Wiley & Sons, Hoboken,
New Jersey, 2012.
[10] D. D'Agostino, D. Parker, A framework for the cost-optimal design of nearly zero energy
buildings (NZEBs) in representative climates across Europe, Energy 149 (2018) 814-829.
[11] J. Sun, R.C. Remsing, Y. Zhang, Z. Sun, A. Ruzsinszky, H. Peng, Z. Yang, A. Paul, U.
Waghmare, X. Wu, M.L. Klein, J.P. Perdew, Accurate first-principles structures and energies of
diversely bonded systems from an efficient density functional, Nat. Chem. 8 (2016) 831-836.
[12] S. Canneaux, F. Bohr, E. Henon, KiSThelP: A Program to Predict Thermodynamic Properties
and Rate Constants from Quantum Chemistry Results, J. Comput. Chem. 35 (2014) 82-93.
[13] R. Liu, L. Liu, W. Tian, Y. Cui, F. Wang, Finite element analysis of effect of interfacial bubbles
on performance of epoxy coatings under alternating hydrostatic pressure, J. Mater. Sci. Technol. 64
(2019) 233-240.
[14] W. Qi, J. Wang, X. Li, Y. Cui, Y. Zhao, J. Xie, G. Zeng, Q. Gao, T. Zhang, F. Wang, Effect of
oxide scale on corrosion behavior of HP-13Cr stainless steel during well completion process, J.
Mater. Sci. Technol. 64 (2019) 153-164.
[15] D. Kong, A. Xu, C. Dong, F. Mao, K. Xiao, X. Li, D.D. Macdonald, Electrochemical
investigation and ab initio computation of passive film properties on copper in anaerobic sulphide
solutions, Corros. Sci. 116 (2017) 34-43.
[16] Y. Ji, C. Dong, X. Wei, S. Wang, Z. Chen, X. Li, Discontinuous model combined with an
atomic mechanism simulates the precipitated eta' phase effect in intergranular cracking of 7-series
aluminum alloys, Comput. Mater. Sci. 166 (2019) 282-292.
[17] W. Zhang, P.K. Liaw, Y. Zhang, Science and technology in high-entropy alloys, Sci.
China-Mater. 61 (2018) 2-22.
[18] X. Wei, C. Dong, A. Xu, X. Li, D.D. MacDonald, Oxygen-induced degradation of the
electronic properties of thin-layer InSe, Phys. Chem. Chem. Phys. 20 (2018) 2238-2250.
[19] W. Setyawan, S. Curtarolo, High-throughput electronic band structure calculations: Challenges
and tools, Comput. Mater. Sci. 49 (2010) 299-312.
[20] J. Carrasco, A. Hodgson, A. Michaelides, A molecular perspective of water at metal interfaces,
Nat. Mater. 11 (2012) 667-674.
[21] A. Xu, C. Dong, X. Wei, X. Li, D.D. Macdonald, DFT and photoelectrochemical studies of
point defects in passive films on copper, J. Electroanal. Chem. 834 (2019) 216-222.
[22] X. Zhao, C. Chen, Q. Sun, Y. Li, H. Yu, Molecular structure optimization design of inhibitors
based on frontier orbitals theory, Appl. Surf. Sci. 494 (2019) 895-907.
[23] Y. Ji, C. Dong, D. Kong, X. Li, Design materials based on simulation results of silicon induced
segregation at AlSi10Mg interface fabricated by selective laser melting, J. Mater. Sci. Technol. 46
(2020) 145-155.
[24] A. Xu, C. Dong, A. Wu, R. Li, L. Wang, D.D. Macdonald, X. Li, Plasma-modified C-doped
Co3O4 nanosheets for the oxygen evolution reaction designed by Butler-Volmer and first-principle
calculations, J. Mater. Chem. A 7 (2019) 4581-4595.
[25] A. Xu, C. Dong, X. Wei, F. Mao, X. Li, D.D. Macdonald, Ab initio calculation and
electrochemical verification of a passivated surface on copper with defects in 0.1 M NaOH,
Electrochem. Commun. 68 (2016) 62-66.
[26] Y. Zhu, J.D. Poplawsky, S. Li, R.R. Unocic, L.G. Bland, C.D. Taylor, J.S. Locke, E.A. Marquis,
G.S. Frankel, Localized corrosion at nm-scale hardening precipitates in Al-Cu-Li alloys, Acta Mater.
189 (2020) 204-213.
[27] X. Wei, C. Dong, P. Yi, A. Xu, Z. Chen, X. Li, Electrochemical measurements and atomistic
simulations of Cl--induced passivity breakdown on a Cu2O film, Corros. Sci. 136 (2018) 119-128.
[28] A. Xu, C. Dong, X. Wei, X. Li, D.D. Macdonald, The aggression behavior study of Cl - on the
defect structure of passive films on copper, RSC Adv. 9 (2019) 15772-15779.
[29] D. Costa, P. Marcus, Electronic core levels of hydroxyls at the surface of chromia related to
their XPS O 1s signature: A DFT+U study, Surf. Sci. 604 (2010) 932-938.
[30] M.M. Islam, B. Diawara, V. Maurice, P. Marcus, First principles investigation on the
stabilization mechanisms of the polar copper terminated Cu2O(111) surface, Surf. Sci. 603 (2009)
2087-2095.
[31] X. Wei, C. Dong, A. Xu, X. Li, First-principles study of the surface reparation of ultrathin InSe
with Se-atom vacancies by thiol chemistry, Appl. Surf. Sci. 475 (2019) 487-493.
[32] X. Wei, C. Dong, Z. Chen, K. Xiao, X. Li, A DFT study of the adsorption of O2 and H2O on
Al(111) surfaces, RSC Adv. 6 (2016) 56303-56312.
[33] X. Wei, C. Dong, Z. Chen, K. Xiao, X. Li, Co-adsorption of O2 and H2O on Al(111) surface: a
vdW-DFT study, RSC Adv. 6 (2016) 79836-79843.
[34] M. Armbrust, A. Fox, R. Griffith, A.D. Joseph, R. Katz, A. Konwinski, G. Lee, D. Patterson, A.
Rabkin, I. Stoica, M. Zaharia, A view of cloud computing, Commun. ACM 53 (2010) 50-58.
[35] T.D. Ladd, F. Jelezko, R. Laflamme, Y. Nakamura, C. Monroe, J.L. O'Brien, Quantum
computers, Nature 464 (2010) 45-53.
[36] A. Beloglazov, J. Abawajy, R. Buyya, Energy-aware resource allocation heuristics for efficient
management of data centers for Cloud computing, Future Gener. Comp. Sy. 28 (2012) 755-768.
[37] F. Feroz, M.P. Hobson, M. Bridges, MultiNest: an efficient and robust Bayesian inference tool
for cosmology and particle physics, Mon. Not. R. Astron. Soc. 398 (2009) 1601-1614.
[38] D. Connolly, H. Lund, B.V. Mathiesen, M. Leahy, A review of computer tools for analysing the
integration of renewable energy into various energy systems, Appl. Energy 87 (2010) 1059-1082.
[39] A.W. Goetz, M.J. Williamson, D. Xu, D. Poole, S. Le Grand, R.C. Walker, Routine
microsecond molecular dynamics simulations with AMBER on GPUs. 1. generalized born, J. Chem.
Theory Comput. 8 (2012) 1542-1555.
[40] Z. Li, N. Miao, J. Zhou, Z. Sun, Z. Liu, H. Xu, High thermoelectric performance of
few-quintuple Sb2Te3 nanofilms, Nano Energy 43 (2018) 285-290.
[41] W. Wang, X. Chen, G. Liu, Z. Shen, D. Xia, P.K. Wong, J.C. Yu, Monoclinic dibismuth
tetraoxide: A new visible-light-driven photocatalyst for environmental remediation, Appl. Catal. B
Environ. 176-177 (2015) 444-453.
[42] A. Xu, C. Dong, X. Wei, Y. Zhang, X. Li, The effect of surface electronic structure on the
bioactivity of neutral dopant Si, Ge, and Sn on TiO2 (110): a DFT study, Phys. Status Solidi B 255
(2018) 109492.
[43] Q. Hu, Y. Liu, T. Zhang, F. Wang, Corrosion failure analysis on the copper alloy flange by
experimental and numerical simulation, Eng. Fail. Anal. 109 (2020) 104276.
[44] M. Poberžnik, F. Chiter, I. Milošev, P. Marcus, D. Costa, A. Kokalj, DFT study of n-alkyl
carboxylic acids on oxidized aluminum surfaces: from standalone molecules to
self-assembled-monolayers, Appl. Surf. Sci. 525 (2020) 146156.
[45] N. Li, C. Dong, C. Man, J. Yao, In situ electrochemical atomic force microscopy and Auger
electro spectroscopy study on the passive film structure of 2024-T3 aluminum alloy combined with
a density functional theory calculation, Adv. Eng. Mater. 21 (2019) 1900386 .
[46] S. Hu, R. Liu, L. Liu, Y. Cui, E.E. Oguzie, F. Wang, Effect of hydrostatic pressure on the
galvanic corrosion of 90/10 Cu-Ni alloy coupled to Ti6Al4V alloy, Corros. Sci. 163 (2020) 108242.
[47] R. Liu, L. Liu, F. Meng, W. Tian, Y. Liu, Y. Li, F. Wang, Finite element analysis of the water
diffusion behaviour in pigmented epoxy coatings under alternating hydrostatic pressure, Prog. Org.
Coat. 123 (2018) 168-175.
[48] X. Wei, C. Dong, Z. Chen, K. Xiao, X. Li, The effect of hydrogen on the evolution of
intergranular cracking: a cross-scale study using first-principles and cohesive finite element
methods, RSC Adv. 6 (2016) 27282-27292.
[49] A. Jain, S.P. Ong, G. Hautier, C. Wei, K.A. Persson, Commentary: The Materials Project: A
materials genome approach to accelerating materials innovation, APL Mater. 1 (2013) 011002.
[50] J.E. Saal, S. Kirklin, M. Aykol, B. Meredig, C. Wolverton, Materials design and discovery with
high-throughput density functional theory: the open quantum materials database (OQMD), JOM 65
(2013) 1501-1509.
[51] C.E. Calderon, J.J. Plata, C. Toher, C. Oses, O. Levy, M. Fornari, A. Natan, M.J. Mehl, G. Hart,
M. Buongiorno Nardelli, S. Curtarolo, The AFLOW standard for high-throughput materials science
calculations, Comput. Mater. Sci. 108 (2015) 233-238.
[52] G. Kresse, J. Furthmüller, Efficiency of ab-initio total energy calculations for metals and
semiconductors using a plane-wave basis set, Comput. Mater. Sci. 6 (1996) 15-50.
[53] H.J.C. Berendsen, J.P.M. Postma, W.F. van Gunsteren, A. DiNola, J.R. Haak, Molecular
dynamics with coupling to an external bath, J. Chem. Phys. 81 (1984) 3684-90.
[54] G. Kresse, H. Jurgen, Ab initio molecular dynamics for liquid metals, Phys. Rev. B 47 (1993)
558-561.
[55] C. Lee, W. Yang, R.G. Parr, Development of the Colle-Salvetti correlation-energy formula into
a functional of the electron density, Phys. Rev. B 37 (1988) 785-789.
[56] J.P. Perdew, K. Burke, M. Ernzerhof, Generalized gradient approximation made simple, Phys.
Rev. Lett. 77 (1996) 3865-3868.
[57] X. Wei, C. Dong, Z. Chen, K. Xiao, X. Li, Density functional theory study of SO 2-adsorbed
Ni(111) and hydroxylated NiO(111) surface, Appl. Surf. Sci. 355 (2015) 429-435.
[58] X. Wei, C.F. Dong, Z.H. Chen, J.Y. Huang, K. Xiao, X.G. Li, Insights into SO2 and H2O
co-adsorption on Cu (100) surface with calculations of density functional theory, Trans. Nonferrous
Met. Soc. China 25 (2015) 4102-4109.
[59] N. Li, C. Dong, X. Wei, C. Man, J. Yao, J. Cao, X. Li, Scanning Kelvin probe force
microscopy and density functional theory studies on the surface potential of the intermetallics in
AA7075-T6 Alloys, J. Mater. Eng. Perform. 28 (2019) 4289-4301.
[60] A. Bouzoubaa, B. Diawara, V. Maurice, C. Minot, P. Marcus, Ab initio modelling of localized
corrosion: study of the role of surface steps in the interaction of chlorides with passivated nickel
surfaces, Corros. Sci. 51 (2009) 2174-2182.
[61] V. Maurice, P. Marcus, Passive films at the nanoscale, Electrochim. Acta 84 (2012) 129-138.
[62] J.D.V.D. Waals, The thermodynamic theory of capillarity under the hypothesis of a continuous
variation of density, J. Stat. Phys. 20 (1979) 200-244.
[63] J. Li, G. Xie, C.C. Lin, J. Liu, 2.5 dimensional GILD electromagnetic modeling and
applications, SEG Technical Program Expanded Abstracts (2002) 692-695.
[64] S. Wolfram, Statistical mechanics of cellular automata, Rev. Mod. Phys. 55 (1983) 601-44.
[65] Y. Feng, M. Liu, Y. Shi, H. Ma, D. Li, Y. Li, L. Lu, X. Chen, High-throughput modeling of
atomic diffusion migration energy barrier of fcc metals, Prog. Nat. Sci. 29 (2019) 341-348.
[66] G. Gabetta, Thirty-five years in environmentally assisted cracking in Italy: a point of view,
Corros. Rev. 33 (2015) 293-300.
[67] A. Bouzoubaa, B. Diawara, V. Maurice, C. Minot, P. Marcus, Ab initio study of the interaction
of chlorides with defect-free hydroxylated NiO surfaces, Corros. Sci. 51 (2009) 941-948.
[68] J.E. Saal, I.S. Berglund, J.T. Sebastian, P.K. Liaw, G.B. Olson, Equilibrium high entropy alloy
phase stability from experiments and thermodynamic modeling, Scr. Mater. 146 (2018) 5-8.
[69] X. Gong, C. Dong, J. Xu, L. Wang, X. Li, Machine learning assistance for electrochemical
curve simulation of corrosion and its application, Mater. Corros. 71 (2019) 474-484.
[70] Y. Shi, D. Fu, X. Zhou, T. Yang, Y. Zhi, Z. Pei, D. Zhang, L. Shao, Data mining to online
galvanic current of zinc/copper Internet atmospheric corrosion monitor, Corros. Sci. 133 (2018)
443-450.
[71] Y. Zhi, T. Yang, D. Fu, An improved deep forest model for forecast the outdoor atmospheric
corrosion rate of low-alloy steels, J. Mater. Sci. Technol. 49 (2020) 202-210.
[72] C.T. Ser, P. Žuvela, M.W. Wong, Prediction of corrosion inhibition efficiency of pyridines and
quinolines on an iron surface using machine learning-powered quantitative structure-property
relationships, Appl. Surf. Sci. 512 (2020) 145612.
[73] H. Ke, C.D. Taylor, Density functional theory: an essential partner in the integrated
computational materials engineering approach to corrosion, Corrosion 75 (2019) 708-726.
[74] V. Maurice, P. Marcus, Progress in corrosion science at atomic and nanometric scales, Prog.
Mater Sci. 95 (2018) 132-171.
[75] N. Ammouchi, H. Allal, Y. Belhocine, S. Bettaz, E. Zouaoui, DFT computations and molecular
dynamics investigations on conformers of some pyrazinamide derivatives as corrosion inhibitors for
aluminum, J. Mol. Liq. 300 (2020) 112309.
[76] C. Verma, H. Lgaz, D.K. Verma, E.E. Ebenso, I. Bahadur, M.A. Quraishi, Molecular dynamics
and Monte Carlo simulations as powerful tools for study of interfacial adsorption behavior of
corrosion inhibitors in aqueous phase: A review, J. Mol. Liq. 260 (2018) 99-120.
[77] K.D. Hammond, Parallel point defect identification in molecular dynamics simulations without
post-processing: A compute and dump style for LAMMPS, Comput. Phys. Commun. 247 (2020)
106862.
[78] G.L. Song, Recent progress in corrosion and protection of magnesium alloys, Adv. Eng. Mater.
7 (2005) 563-586.
[79] G.L. Song, B. Johannesson, S. Hapugoda, D. StJohn, Galvanic corrosion of magnesium alloy
AZ91D in contact with an aluminium alloy, steel and zinc, Corros. Sci. 46 (2004) 955-977.
[80] C. Chen, Expert Review. Science Mapping: A Systematic Review of the Literature, J. Data
Infor. Sci. 2 (2017) 1-40.
[81] C. Chen, CiteSpace: A Practical Guide for Mapping Scientific Literature, NY: Nova Science
Publishers, Hauppauge, 2016.
[82] K.S. Exner, J. Anton, T. Jacob, H. Over, Chlorine evolution reaction on RuO2(110): ab initio
atomistic thermodynamics study - Pourbaix diagrams, Electrochim. Acta 120 (2014) 460-466.
[83] L.F. Huang, J.M. Rondinelli, Accurate First-Principles Electrochemical Phase Diagrams for Ti
Oxides from Density Functional Calculations, Phys. Rev. B 92 (2015) 245126.
[84] A. Kokalj, H. Behzadi, R. Farahati, DFT study of aqueous-phase adsorption of cysteine and
penicillamine on Fe(110): Role of bond-breaking upon adsorption, Appl. Surf. Sci. 514 (2020)
145896.
[85] J. Liu, X. Zhang, M. Chen, L. Li, Z. Bing, j. Tang, S. Liu, DFT study on surface properties and
dissolution trends of Al (100) surfaces doped with Zn, Ga, In, Sn and Pb, Appl. Surf. Sci. 257 (2011)
4004-4009.
[86] H. Ma, X.-Q. Chen, R. Li, S. Wang, J. Dong, W. Ke, First-principles modeling of anisotropic
anodic dissolution of metals and alloys in corrosive environments, Acta Mater. 130 (2017) 137-146.
[87] X. Wei, C. Dong, P. Yi, A. Xu, Z. Chen, X. Li, Electrochemical measurements and atomistic
simulations of Cl−-induced passivity breakdown on a Cu2O film, Corros. Sci. 136 (2018) 119-128.
[88] A. Bouzoubaa, D. Costa, B. Diawara, N. Audiffren, P. Marcus, Insight of DFT and atomistic
thermodynamics on the adsorption and insertion of halides onto the hydroxylated NiO(111) surface,
Corros. Sci. 52 (2010) 2643-2652.
[89] C.H. Zhang, M. Liu, Y. Jin, D.B. Sun, The corrosive influence of chloride ions preference
adsorption on α-Al2O3 (0001) surface, Appl. Surf. Sci. 347 (2015) 386-391.
[90] C. Cui, R. Ma, A. Chen, Z. Pan, H. Tian, Experimental study and 3D cellular automata
simulation of corrosion pits on Q345 steel surface under salt-spray environment, Corros. Sci. 154
(2019) 80-89.
[91] H. Wang, E.H. Han, Computational simulation of corrosion pit interactions under
mechanochemical effects using a cellular automaton/finite element model, Corros. Sci. 103 (2016)
305-311.
[92] L. Li, X. Li, C. Dong, K. Xiao, L. Lu, Cellular automata modeling on pitting current transients,
Electrochem. Commun. 11 (2009) 1826-1829.
[93] M. Madrigal-Cano, J.M. Hallen, E.M. Arce-Estrada, T. Le Manh, Effect of crystallographic
texture and microstructure on pitting corrosion behavior of low carbon steels: A Monte Carlo model,
Comput. Mater. Sci. 161 (2019) 394-402.
[94] N. Murer, R.G. Buchheit, Stochastic modeling of pitting corrosion in aluminum alloys, Corros.
Sci. 69 (2013) 139-148.
[95] M. Strobel, K.-H. Heinig, T. Michely, Mechanisms of pit coarsening in ion erosion of fcc(111)
surfaces: a kinetic 3D lattice Monte-Carlo study, Surf. Sci. 486 (2001) 136-156.
[96] J. Song, W.A. Curtin, Atomic mechanism and prediction of hydrogen embrittlement in iron,
Nat. Mater. 12 (2013) 145-151.
[97] P.D. White, S.A. Barter, N. Medhekar, Hydrogen induced amorphisation around nanocracks in
aluminium, Eng. Fract. Mech. 161 (2016) 40-54.
[98] H.Y. Song, L. Zhang, M.X. Xiao, Molecular dynamics simulation of effect of hydrogen atoms
on crack propagation behavior of α-Fe, Phys. Lett. A 380 (2016) 4049-4056.
[99] A. Tehranchi, X. Zhou, W.A. Curtin, A decohesion pathway for hydrogen embrittlement in
nickel: Mechanism and quantitative prediction, Acta Mater. 185 (2020) 98-109.
[100] L. Zhang, B. Sun, Q. Zhang, H. Liu, K. Liu, H. Song, First-principles study of hydrogen
behavior in α-Pu2O3, Comput. Mater. Sci. 179 (2020) 109688.
[101] E. Torres, J. Pencer, D.D. Radford, Atomistic simulation study of the hydrogen diffusion in
nickel, Comput. Mater. Sci. 152 (2018) 374-380.
[102] X.H. Guo, S.Q. Shi, L.J. Qiao, Simulation of hydrogen diffusion and initiation of hydrogen‐
induced cracking in PZT ferroelectric ceramics using a phase field model, J. Am. Ceram. Soc. 90
(2007) 2868-2872.
[103] E. Martínez-Pañeda, A. Golahmar, C.F. Niordson, A phase field formulation for hydrogen
assisted cracking, Comput. Method Appl. Mech. Eng. 342 (2018) 742-761.
[104] T.-T. Nguyen, J. Bolivar, J. Réthoré, M.-C. Baietto, M. Fregonese, A phase field method for
modeling stress corrosion crack propagation in a nickel base alloy, Int. J. Solids Struct. 112 (2017)
65-82.
[105] T.T. Nguyen, J. Bolivar, Y. Shi, J. Réthoré, A. King, M. Fregonese, J. Adrien, J.Y. Buffiere,
M.C. Baietto, A phase field method for modeling anodic dissolution induced stress corrosion crack
propagation, Corros. Sci. 132 (2018) 146-160.
[106] W. Mai, S. Soghrati, A phase field model for simulating the stress corrosion cracking initiated
from pits, Corros. Sci. 125 (2017) 87-98.
[107] H.H. Kart, M. Uludogan, T. Cagin, DFT studies of sulfur induced stress corrosion cracking in
nickel, Comput. Mater. Sci. 44 (2009) 1236-1242.
[108] E. Bitzek, J.R. Kermode, P. Gumbsch, Atomistic aspects of fracture, Int. J. Fract. 191 (2015)
13-30.
[109] P. Arnoux, SCC of austenitic alloys in PWR water simulated with molecular dynamics, Acta
Neuropathol. 110 (2007) 563-578.
[110] K. Tohgo, H. Suzuki, Y. Shimamura, G. Nakayama, T. Hirano, Monte Carlo simulation of
stress corrosion cracking on a smooth surface of sensitized stainless steel type 304, Corros. Sci. 51
(2009) 2208-2217.
[111] Q.K. Li, Y. Zhang, W.Y. Chu, Molecular dynamics simulation of stress corrosion cracking in
Cu3Au, Comput. Mater. Sci. 25 (2002) 510-518.
[112] N.K. Das, K. Suzuki, Y. Takeda, K. Ogawa, T. Shoji, Quantum chemical molecular dynamics
study of stress corrosion cracking behavior for fcc Fe and Fe-Cr surfaces, Corros. Sci. 50 (2008)
1701-1706.
[113] N.K. Das, K. Suzuki, K. Ogawa, T. Shoji, Early stage SCC initiation analysis of fcc Fe–Cr–Ni
ternary alloy at 288 °C: A quantum chemical molecular dynamics approach, Corros. Sci. 51 (2009)
908-913.
[114] K.F. Khaled, Monte Carlo simulations of corrosion inhibition of mild steel in 0.5 M sulphuric
acid by some green corrosion inhibitors, J. Solid State Electrochem. 13 (2009) 1743-1756.
[115] Y. Sasikumar, A.S. Adekunle, L.O. Olasunkanmi, I. Bahadur, R. Baskar, M.M. Kabanda, I.B.
Obot, E.E. Ebenso, Experimental, quantum chemical and Monte Carlo simulation studies on the
corrosion inhibition of some alkyl imidazolium ionic liquids containing tetrafluoroborate anion on
mild steel in acidic medium, J. Mol. Liq. 211 (2015) 105-118.
[116] E. Javadi, M. Ghaffari, G. Bahlakeh, P. Taheri, Photocatalytic, corrosion protection and
adhesion properties of acrylic nanocomposite coating containing silane treated nano zinc oxide: A
combined experimental and simulation study, Prog. Org. Coat. 135 (2019) 496-509.
[117] D. Ravnikar, R.S. Rajamure, U. Trdan, N.B. Dahotre, J. Grum, Electrochemical and DFT
studies of laser-alloyed TiB2/TiC/Al coatings on aluminium alloy, Corros. Sci. 136 (2018) 18-27.
[118] B.R. Hinderliter, S.G. Croll, Simulations of nanoscale and macroscopic property changes on
coatings with weathering, J. Coat. Technol. 3 (2006) 203-212.
[119] D.P. Bentz, T. Nguyen, Simulation of diffusion in pigmented coatings on metals using
Monte-Carlo methods, J. Coat. Technol. 62 (1990) 57-63.
[120] S. Croll, B. Hinderliter, A framework for predicting the service lifetime of composite
polymeric coatings, J. Mater. Sci. 43 (2008) 6630-6641.
[121] Y. Hua, Application of numerical simulation calculation in cathodic protection system design
for underground pipeline and storage tank, Dev. Appl. Mater. 3 (2008) 24-28.
[122] I.A. Metwally, H.M. Al-Mandhari, A. Gastli, Z. Nadir, Factors affecting cathodic-protection
interference, Eng. Anal. Bound. Elem. 31 (2007) 485-493.
[123] G. Cui, Z. Li, X. Wei, X. Bai, Cathodic protection design of station area based on boundary
element method, J. China Univ. Pet. 38 (2014) 161-166.
[124] R. Montoya, V. Nagel, J.C. Galván, J.M. Bastidas, Influence of irregularities in the electrolyte
on the cathodic protection of steel: A numerical and experimental study, Mater. Corros. Prot. 64
(2013) 1055-1065.
[125] M. Zhang, J. Zhao, L.L. Zou, G. Wang, Numerical simulation and optimization of steel pipe
cathodic protection under the condition of 3.5% brine, Adv. Mater. Res. 690-693 (2013) 2751-2756.
[126] W.H.A. Peelen, R.B. Polder, E. Redaelli, L. Bertolini, Qualitative model of concrete
acidification due to cathodic protection, Mater. Corros. 59 (2008) 81-89.
[127] Y. Guo, A.K. Hu, H.T. Dong, Numerical simulation of cathodic protection system, Adv. Mater.
Res. 548 (2012) 682-685.
[128] L.Y. Pan, B.G. Lei, Z. Fan, W.-H. Li, Application of ANSYS finite element method in
cathodic protection of pipelines, Mater. Prot. 47 (2014) 45-47.
[129] G. Hu, B. Xiang, S. Zhang, An application of Matlab to calculating the cathodic protection
potential of ocean pipelines, Mar. Sci. 31 (2007) 34-37.
[130] D. Kong, C. Dong, X. Ni, C. Man, K. Xiao, X. Li, Insight into the mechanism of alloying
elements (Sn, Be) effect on copper corrosion during long-term degradation in harsh marine
environment, Appl. Surf. Sci. 455 (2018) 543-553.
[131] D. Kong, C. Dong, X. Ni, L. Zhang, H. Luo, R. Li, L. Wang, C. Man, X. Li, The passivity of
selective laser melted 316L stainless steel, Appl. Surf. Sci. 504 (2020) 144495.
[132] D. Kong, C. Dong, Z. Zheng, F. Mao, A. Xu, X. Ni, C. Man, J. Yao, K. Xiao, X. Li, Surface
monitoring for pitting evolution into uniform corrosion on Cu-Ni-Zn ternary alloy in alkaline
chloride solution: ex-situ LCM and in-situ SECM, Appl. Surf. Sci. 440 (2018) 245-257.
[133] C. Man, C. Dong, Z. Cui, K. Xiao, Q. Yu, X. Li, A comparative study of primary and
secondary passive films formed on AM355 stainless steel in 0.1M NaOH, Appl. Surf. Sci. 427
(2018) 763-773.
[134] C. Man, C. Dong, T. Liu, D. Kong, D. Wang, X. Li, The enhancement of microstructure on
the passive and pitting behaviors of selective laser melting 316L SS in simulated body fluid, Appl.
Surf. Sci. 467-468 (2019) 193-205.
[135] Z. Wang, C. Dong, S. Yang, D. Zhang, K. Xiao, X. Li, Facile incorporation of hydroxyapatite
onto an anodized Ti surface via a mussel inspired polydopamine coating, Appl. Surf. Sci. 378 (2016)
496-503.
[136] D. Kong, C. Dong, X. Ni, L. Zhang, H. Luo, R. Li, L. Wang, C. Man, X. Li, Superior
resistance to hydrogen damage for selective laser melted 316L stainless steel in a proton exchange
membrane fuel cell environment, Corros. Sci. 166 (2020) 108425.
[137] H. Luo, Q. Yu, C. Dong, G. Sha, Z. Liu, J. Liang, L. Wang, G. Han, X. Li, Influence of the
aging time on the microstructure and electrochemical behaviour of a 15-5PH ultra-high strength
stainless steel, Corros. Sci. 139 (2018) 185-196.
[138] C. Man, C. Dong, D. Kong, L. Wang, X. Li, Beneficial effect of reversed austenite on the
intergranular corrosion resistance of martensitic stainless steel, Corros. Sci. 151 (2019) 108-121.
[139] C. Man, C. Dong, L. Wang, D. Kong, X. Li, Long-term corrosion kinetics and mechanism of
magnesium alloy AZ31 exposed to a dry tropical desert environment, Corros. Sci. 163 (2020)
108274.
[140] L. Wang, C. Dong, C. Man, D. Kong, K. Xiao, X. Li, Enhancing the corrosion resistance of
selective laser melted 15-5PH martensite stainless steel via heat treatment, Corros. Sci. 166 (2020)
108427.
[141] L. Wang, C. Dong, J. Yao, Z. Dai, C. Man, Y. Yin, K. Xiao, X. Li, The effect of ɳ-Ni3Ti
precipitates and reversed austenite on the passive film stability of nickel-rich Custom 465 steel,
Corros. Sci. 154 (2019) 178-190.
[142] J. Yao, D.D. Macdonald, C. Dong, Passive film on 2205 duplex stainless steel studied by
photo-electrochemistry and ARXPS methods, Corros. Sci. 146 (2019) 221-232.
[143] S. Gao, C. Dong, H. Luo, K. Xiao, X. Pan, X. Li, Scanning electrochemical microscopy study
on the electrochemical behavior of CrN film formed on 304 stainless steel by magnetron sputtering,
Electrochim. Acta 114 (2013) 233-241.
[144] D. Kong, C. Dong, X. Wei, C. Man, X. Lei, F. Mao, X. Li, Size matching effect between
anion vacancies and halide ions in passive film breakdown on copper, Electrochim. Acta 292 (2018)
817-827.
[145] D. Kong, X. Ni, C. Dong, L. Zhang, C. Man, J. Yao, K. Xiao, X. Li, Heat treatment effect on
the microstructure and corrosion behavior of 316L stainless steel fabricated by selective laser
melting for proton exchange membrane fuel cells, Electrochim. Acta 276 (2018) 293-303.
[146] L. Li, X. Li, C. Dong, Y. Huang, Computational simulation of metastable pitting of stainless
steel, Electrochim. Acta 54 (2009) 6389-6395.
[147] D. Kong, C. Dong, X. Ni, L. Zhang, C. Man, J. Yao, Y. Ji, Y. Ying, K. Xiao, X. Cheng, X. Li,
High-throughput fabrication of nickel-based alloys with different Nb contents via a dual-feed
additive manufacturing system: Effect of Nb content on microstructural and mechanical properties,
J. Alloy. Compd. 785 (2019) 826-837.
[148] D. Kong, C. Dong, X. Ni, L. Zhang, C. Man, G. Zhu, J. Yao, J. Yao, L. Wang, X. Cheng, X. Li,
Effect of TiC content on the mechanical and corrosion properties of Inconel 718 alloy fabricated by
a high-throughput dual-feed laser metal deposition system, J. Alloy. Compd. 803 (2019) 637-648.
[149] D. Kong, C. Dong, X. Ni, L. Zhang, J. Yao, C. Man, X. Cheng, K. Xiao, X. Li, Mechanical
properties and corrosion behavior of selective laser melted 316L stainless steel after different heat
treatment processes, J. Mater. Sci. Technol. 35 (2019) 1499-1507.
[150] P. Priya, X. Yan, S. Chaudhuri, Study of intermetallics for corrosion and creep resistant
microstructure in Mg-RE and Mg-Al-RE alloys through a data-centric high-throughput DFT
framework, Comput. Mater. Sci. 175 (2020) 109541.
[151] B. Diawara, L. Joubert, D. Costa, P. Marcus, C. Adamo, Ammonia on Ni(111) surface studied
by first principles: Bonding, multilayers structure and comparison with experimental IR and XPS
data, Surf. Sci. 603 (2009) 3025-3034.
[152] S.P. Knight, N. Birbilis, B.C. Muddle, A.R. Trueman, S.P. Lynch, Correlations between
intergranular stress corrosion cracking, grain-boundary microchemistry, and grain-boundary
electrochemistry for Al-Zn-Mg-Cu alloys, Corros. Sci. 52 (2010) 4073-4080.
[153] R.G. Song, W. Dietzel, B.J. Zhang, W.J. Liu, M.K. Tseng, A. Atrens, Stress corrosion
cracking and hydrogen embrittlement of an Al-Zn-Mg-Cu alloy, Acta Mater. 52 (2004) 4727-4743.
[154] Y.P. Xiao, Q.L. Pan, W.B. Li, X.Y. Liu, Y.B. He, Influence of retrogression and re-aging
treatment on corrosion behaviour of an Al-Zn-Mg-Cu alloy, Mater. Des. 32 (2011) 2149-2156.
[155] G.S. Frankel, Pitting corrosion of metals: a review of the critical factors, J. Electrochem. Soc.
145 (1998) 2186.
[156] J. Kruger, Passivity of metals-a materials science perspective, Int. Mater. Rev. 33 (1988)
113-130.
[157] H. Toda, K. Minami, K. Koyama, K. Ichitani, M. Kobayashi, K. Uesugi, Y. Suzuki, Healing
behavior of preexisting hydrogen micropores in aluminum alloys during plastic deformation, Acta
Mater. 57 (2009) 4391-4403.
[158] R.K. Viswanadham, T.S. Sun, J.A.S. Green, Grain boundary segregation in Al-Zn-Mg
alloys-implications to stress corrosion cracking, Metall. Mater. Trans. A 11A (1980) 85-89.
[159] S. Xu, Y. Zhao, Using FEM to determine the thermo-mechanical stress in tube to tube-sheet
joint for the SCC failure analysis, Eng. Fail. Anal. 34 (2013) 24-34.
[160] T. Luther, C. Koenke, Polycrystal models for the analysis of intergranular crack growth in
metallic materials, Eng. Fract. Mech. 76 (2009) 2332-2343.
[161] P. Zhang, M. Karimpour, D. Balint, J. Lin, D. Farrugia, A controlled Poisson Voronoi
tessellation for grain and cohesive boundary generation applied to crystal plasticity analysis,
Comput. Mater. Sci. 64 (2012) 84-89.
[162] I. Simonovski, L. Cizelj, Cohesive element approach to grain level modelling of intergranular
cracking, Eng. Fract. Mech. 110 (2013) 364-377.
[163] J. Zhang, Z.H. Chen, C.F. Dong, Simulating Intergranular Stress Corrosion Cracking in AZ31
Using Three-Dimensional Cohesive Elements for Grain Structure, J. Mater. Eng. Perform. 24 (2015)
4908-4918.
[164] H. Sheng, Electrochemical Mechanism and Numerical Simulation on Stress Corrosion
Cracking of 2024 Aluminium Alloy in 𝐂𝐥− Containing Solution, Beijing, University of Science and
Technology Beijing, 2012.
[165] D.D. Macdonald, The point defect model for the passive state, J. Electrochem. Soc. 139 (1992)
3434-3449.
[166] Y. Ji, C. Dong, L. Chen, K. Xiao, X. Li, High-throughput computing for screening the
potential alloying elements of a 7xxx aluminum alloy for increasing the alloy resistance to stress
corrosion cracking, Corros. Sci. 183 (2021) 109304.
[167] J.R. Scully, P.V. Balachandran, Future frontiers in corrosion science and engineering, Part III:
the next “leap ahead” in corrosion control may be enabled by data analytics and artificial
intelligence, Corrosion 75 (2019) 1395-1397.
[168] R. Liu, A. Kumar, Z. Chen, A. Agrawal, V. Sundararaghavan, A. Choudhary, A predictive
machine learning approach for microstructure optimization and materials design, Sci. Rep. 5 (2015)
11551.
[169] A. Garcia, O. Anjos, C. Iglesias, H. Pereira, J. Martinez, J. Taboada, Prediction of mechanical
strength of cork under compression using machine learning techniques, Mater. Des. 82 (2015)
304-311.
[170] G. Pilania, C. Wang, X. Jiang, S. Rajasekaran, R. Ramprasad, Accelerating materials property
predictions using machine learning, Sci. Rep. 3 (2013) 2810.
[171] Y. Liu, T. Zhao, W. Ju, S. Shi, Materials discovery and design using machine learning, J.
Materiomics 3 (2017) 159-177.
[172] D.T. Ahneman, J.G. Estrada, S. Lin, S.D. Dreher, A.G. Doyle, Predicting reaction
performance in C-N cross-coupling using machine learning, Science 360 (2018) 186-190.
[173] P. Raccuglia, K.C. Elbert, P.D.F. Adler, C. Falk, M.B. Wenny, A. Mollo, M. Zeller, S.A.
Friedler, J. Schrier, A.J. Norquist, Machine-learning-assisted materials discovery using failed
experiments, Nature 533 (2016) 73-76.
 Figure and table captions
Table 1 Corrosion behavior, input and output parameters, computing methods, typical software and
experiments of corrosion science and engineering [15, 23, 31, 57, 86, 129-148].
Table 2 Geometrical parameters and adsorption energy of H2O and O2 co-adsorbed structures on
Al2CuMg and Al surfaces [45].
Fig. 1. Publications statistics of last three decades until 2020: Various methods used to do
calculating and simulating in corrosion science.
Fig. 2. Knowledge map of mechanisms acting at different simulation methods and hotpot issue of
corrosion science and engineering.
Fig. 3. Integrated computation of corrosion, a construct used to illustrate the multiscale modelling
and experimental verification.
Fig. 4. Flow chart of cross-scale calculation for atmospheric corrosion of Al alloys.
Fig. 5. (a) Top and side view of adsorbed structures of H2O on the oxygen pre-adsorbed Al (111)
surface[33]; (b) the difference charge density of structures of the co-adsorption of H2O and O2 on
Al(111) surface[33]; and the difference charge density of adsorbed structures of O atoms on the (c)
Al (100), Al (111), and Al (101) surfaces[45]; (d) Al2CuMg (101) surfaces[45]; and (e) Al2CuMg
(001) surfaces[45].
Fig. 6. (a) Adsorbed structures of Cl- on hydroxylated γ-Al2O3(010) surface and its corresponding
adsorption energies; (b) penetrated structure of Cl– from surface to interior of γ-Al2O3(010) surface
and its corresponding penetration energies.
Fig. 7. (a) Configuration of ∑5 (100)/36.87° GB and (b) its H-atom solution energies of I1-I4; (c)
the diffusion path of H atoms in Al bulk and (d) its corresponding energy barrier; (e) the diffusion
path of H atoms in Al GB and (f) its corresponding energy barrier.
Fig. 8. (a) Suspected stable sites of H atoms on the Al (111) surface; (b) on the η’ (001) surface for
H coverage of 1/9 ML; (c) most stable sites in Al/MgZn2 grain boundary; (d) hydrogen atom
transfer between precipitate phase and aluminum matrix[16].
Fig. 9. Scheme through a cross-scale calculation from atomic scale to macro-scale for intergranular
cracking [48].
Fig. 10. (a)-(d) Nucleation and propagation process of cracks in oxide film near the notch of WOL
specimen; (e) experimental morphology of pitting induced crack; (f) strain distribution in pitting
which is the crack initiation sit[164].
Fig. 11. Computation flow of the integration process represented by DFT and PD [16]
Fig. 12. (a) Side and top view of the defective Cu2O (111) surface; (b) impedance spectra, (c)
Mott-Schottky, and (d) polarization curve of pure copper in 0.1 mol/L NaOH solution [25].
Fig. 13. (a) Calculated results of diffusivity of cation defect in copper oxide with or without Cl-[27];
(b) calculated and experimental results of diffusivity of point defects in passive film on copper in
0.1 mol/L NaCl + 5 × 10-4 mol/L Na2S·9H2O solution[15]; (c) side and top view of the defective
Cu2S (100) surface[15].
Fig. 14. Band structure and side view of defective Cu2O (111) surface with (a) VCu-1 and (b) VCu-2;
(c) formation enthalpies for point defects in Cu2O; (d) relationship between band gap and vacancy
concentration[21].
Fig. 15. (a) Experimental results of band structures of passive film on copper formed in various Cl -
concentrations; (b) band structures and DOS of Cu2O (111)[28]
Fig. 16. Flowchart of machine learning for electrochemical data [69].
Fig. 17. Machine learning results for electrochemical data, (a)-(e) prediction of potential
polarization curve, (f)(g) EIS and (h) passivation potential with chloride ion concentration and
temperature[69].
Fig. 18. Molecular dynamic and first-principle calculations of 3D printing AlSi10Mg: (a)
agglomeration of silicon atoms with different temperature, (b) the differences in laser melting
temperature causes segregation or to form a silicon-rich phase at the edges, (c) XRD pattern and (d)
the surface crystal structure[23].
 Table list:

Table 1

Corrosion Characterization
Computing Tool Experiment
Behavior Parameter

VASP,
Pourbaix Free Energy,
Pwmat [82, 83], Potentiostatic
Diagram Chemical Potential
Materials Project

Work Function,
Anodic VASP [84-86],
Adsorption, SKPFM, Polarization
Dissolution PWmat, CASTEP
Formation Energy

Semiconductor Photoelectrochemical
VASP, Pwmat [87-89],
Passivity Properties, methods,
CASTEP
Point Defects Mott-Schottky

Transient Current, Metastable Pitting

Cellular Automata [90-92],
Pitting Pit Radius， Curves, Surface
Monte Carlo[93-95]
Pits Interaction Analysis

Cohesive Energy,
Hydrogen LAMMPS [96-98], VASP
Cracking, GB Binding
Induced [99-101], Phase Field Hydrogen Charging
Energy, Deformation,
Cracking Method [102, 103]
Diffusion

Dissolution Processes, Phase Field Method

Stress In situ SCC
Diffusion, Crack, [104-106], VASP [107, 108],
Corrosion Experiments,
Adsorption, Monte Carlo [109, 110],
Cracking Electrochemical Noise
Atom Bonds LAMMPS [111-113]
 Adsorption, Monte Carlo [114, 115],
Inhibitor Immersion, EIS
Chemical Bonds LAMMPS , Gaussian

Binding Energy, LAMMPS[116, 117], VASP Pull-off Adhesion

Coating Chemical Bond, [118], Test, Salt Spray Test,
Adhesion, Monte Carlo [119, 120] EIS

BEASY[121-123], COMSOL
Cathodic Anode Output Current，
[124-126], ANSYS [127, Feed Experiment
Protection Potential Distribution
128], MATLAB [129]
 Table 2

Distance / nm
Adsorbed Al(Cu/Mg)-O
Adsorption Charge
structure (H2O/OH) Al-O Cu-O Mg-O H-O
energy transfer
(O2) (O2) (O2) (H2O/OH)
Top Bridge (eV) (e)

Al (100) 0.175 0.189 0.188 — — 0.098 10.87 3.39

Al (111) 0.182 0.187 0.184 — — 0.101 10.76 3.42

Al (101) 0.174 0.189 0.175 — — 0.97 10.69 3.32

S-1 0.210 — 0.178 0.200 0.209 0.98 9.32 3.3

S-2 0.176 0.192 0.179 0.194 0.197 0.99 9.37 2.92

S-3 0.208 — 0.179 0.194 0.197 0.98 9.44 3.14

J-1 0.184 0.189 0.176 — — 0.10 10.33 3.3

J-2 — 0.199 0.185 — 0.195 0.97 9.63 2.92

 1100
First principle
1000 Molecular dynamics 182

Monte carlo
900 Cellular automata
169
Finite element method
800 Phase field method 158

Others*
Simulation method

700 134

114

600 117

109

500 85
110

274
96

400 105
81
228

85
77
70 64
300 143
70 56
53
69
123
41
200 41
57 58 41 100
40 311
28 48 29 86 287
30 48 65

100 26 26
25
34
39

23
23
29
22 33
60
160
209

24 143
24 34
26 101 105
24 59 66 73
37 45
23
0
1990 1995 2000 2005 2010 2015 2020
* including Numerical simulation,
Years Peridynamic and Thermo-Calc.

Fig. 1.
Fig. 2.
Fig. 3.
Fig. 4.
Fig. 5.
Fig. 6.
Fig. 7.
Fig. 8.
Fig. 9.
Fig. 10.
Fig. 11.
Fig. 12.
Fig. 13.
Fig. 14.
Fig. 15.
Fig. 16.
Fig. 17.
Fig. 18.