www.nature.

com/npjmatdeg

ARTICLE OPEN

Galvanic corrosion behavior of welded joint in marine

atmosphere environment based on capillary microcell
1,2 ✉
Xinyu Wang1,2, Feng Huang , Qian Hu1,2, Jiawei Zhang1,2 and Jing Liu1,2

The kinetical behavior and mechanism of micro-galvanic corrosion of Q690 bainitic weathering steel welded joint in marine
atmosphere were investigated. The results showed that a micro-galvanic corrosion cell formed among different regions of the Q690
welded joint. The kinetical behavior of micro-galvanic corrosion could be divided into four stages: the initiation and development of
micro-galvanic corrosion, the decreasing of galvanic corrosion rate, the polarity reversal of galvanic couple, and the disappearance
of galvanic effect. The difference of corrosion rate between Q690 steel and its welded joint is closely related to the micro-galvanic
corrosion kinetic behavior.
npj Materials Degradation (2024)8:1 ; https://doi.org/10.1038/s41529-023-00417-7

INTRODUCTION determined by SKP to investigate the corrosion thermodynamic

1234567890():,;

Bridge steel has become a widely used steel after years of behavior of each micro-region of welded joints. However, conven-
development. With the rapid development of social progress and tional micro-electrochemical methods such as SKP, SVET, etc., could
bridge technology, higher requirements have been put forward only give the surface potential or current distribution through the
for bridge steel, i.e., high strength, good corrosion resistance and electrical signal between the probe and the metal surface, the
welding performance, especially in the harsh marine environment. electrochemical polarization behavior of each micro-region still
However, the microstructure of steel welded joints is often cannot be obtained through the above technologies.
affected by the welding thermal cycling, resulting in a significant The self‐designed three-electrode capillary micro-cell can accurately
change compared with that before welding1, thus becoming the locate the tested micro-region by adjusting the capillary port
weak link of corrosion2, especially the local corrosion dominated diameter, thereby obtaining the potentiodynamic polarization curves,
by galvanic corrosion3–5 and stress corrosion6,7. The galvanic and then study polarization behaviors of different micro-regions of
corrosion is the most common, with a high local corrosion rate the welded joint. Therefore, the technology can overcome the current
and hidden, and a high likelihood of early catastrophic failure. technical deficiencies and provide a great help for deeply studying
There are many reports on galvanic corrosion of welded joints of the electrochemical mechanism of galvanic corrosion in the micro-
steel structures, mainly focusing on the research of influencing regions of welded joint3. Recently, our group improved the structure
factors, including electrochemical factors (potential difference and of capillary microcell, optimized the preparation parameters of
polarization ability8,9), geometric factors (cathode-anode area ratio, Ag / AgCl reference electrode23, and used it to study the mechanism
galvanic distance and spatial position8–12), and environmental of inclusion-induced pitting initiation in stainless steel22,24.
factors (temperature, oxygen content, conductivity Ph, In the paper, the 3.5% NaCl was used to simulate the marine
etc.9,10,13–16). For the above factors, researchers have tried a variety atmospheric environment. The surface electrochemical activity
of quantitative and qualitative testing methods for different and galvanic current of the welded joint of the new generation
macroscopic galvanic corrosion, such as weight loss measurement, Q690 bainitic weathering bridge steel were studied by employing
morphology observation, and traditional electrochemical measure- SKP, Scanning Reference Electrode Technology (SRET), Zero
ment technology17–19. However, with the continuous development Resistance Current Meter and self‐designed capillary microelec-
trode technique. At the same time, combining with Raman
of various new technologies, various micro-zone testing methods
spectroscopy, field emission electron probe microanalysis (FE-
have been employed to study the micro-galvanic corrosion, which
EPMA) and other characterization techniques, and the influence
can be induced by the difference in microstructure between the
law and mechanism of micro-galvanic corrosion on the corrosion
micro-regions, such as base metal (BM), heat-affected zone (HAZ)
kinetics behavior of the whole welded joint would be clarified,
and weld metal (WM). Wu et al.20 detected the high and low
which can provide theoretical basis and data support for the
potential areas on the surface of traditional and new 3% Ni
service safety and protection of 690 MPa high strength bainitic
weathering steel by scanning Kelvin probe (SKP). Li et al.21 simulated
weathering steel in simulated marine atmospheric environment.
the welded joint according to the composition and area ratio by
using the array electrode technology, and studied its initial corrosion
behavior. Wint et al. measured1 the current density distribution in
the laser-welded joints of high-strength low alloy (HSLA) and hot METHODS
stamping ultra-high strength steel (UHSS), and comparatively Material preparation
studied the galvanic corrosion behavior by employing scanning The Q690 high-strength weathering steel used in this research was
vibrating electrode technique (SVET). In our previous work22, the supplied by Wuhan Iron and Steel Corporation, and the welded
Volta potential distribution of A710 steel welded joints was joints was obtained using a gas-shielded arc process. The chemical

1
The State Key Laboratory of Refractories and Metallurgy, Wuhan University of Science and Technology, 430081 Wuhan, Hubei, China. 2Hubei Engineering Technology Research
Centre of Marine Materials and Service Safety, Wuhan University of Science and Technology, 430081 Wuhan, Hubei, China. ✉email: [email protected]

Published in partnership with CSCP and USTB

 X. Wang et al.
2
Table 1. Chemical composition of base metal and welding wire (wt. %).

C Si Mn P S Cr Ni Cu Mo Fe

Q690 0.06 0.21 1.78 0.006 ≤0.005 ≤0.6 ≤0.3 ≤0.4 / Bal.
XY-ER80QNH ≤0.1 ≤0.65 ≤2.00 ≤0.015 ≤0.015 ≤0.4 ≤1.62 ≤0.25 ≤0.60 Bal.

Table 2. Gas shielded welding process parameters.

Welding wire Tensile strength Rm (MPa) Yield strength Rp0.2 (MPa) Elongation A (%) Protective gas

XY-ER80QNH ≤280 ≤28 ≤30 80%Ar+20%CO2

1234567890():,;

Fig. 1 FE-SEM images of the microstructure of Q690 welded joint. a The junction of WM and HAZ, (b) The junction of BM and HAZ, (c) BM,
(d) FGHAZ, (e) CGHAZ, (f)WM, and (g) Q690 welded joint.

composition of the base metal and the welding wire were shown The total duration of the test was 576 h, and sampling was
in Table 1. And the welding parameters was shown in Table 2. performed at 24 h, 48 h, 72 h, 96 h, 192 h and 576 h. Seven
Fig. 1 shows the microstructure of the welded joints of 690 bainite specimens were used for each type of steels for reproducibility of
weathering steel. It is seen that the base metal (BM) is primarily the results, where five parallel specimens (60 × 40 × 5 mm in
composed of lath bainite (LB) and a small amount of ferrite (F), and dimension) were used for weight-loss testing, one (20 × 15 × 5 mm)
the microstructure is relatively uniform (Fig. 1c). The fine grain heat was used for Raman analysis, one (20 × 15 × 5 mm) was used for
affected zone (FGHAZ) mainly consists of polygonal ferrite (PF) and morphological and elemental analysis by FE-EPMA. The test solution
granular bainite (GB), with different grain sizes (Fig. 1d). The coarse was 3.5 wt.% NaCl solution. The temperature in the test chamber
grain heat affected zone (CGHAZ) (Fig. 1e) has coarse grains, and a during the drying period was (45 ± 2) oC, and the relative humidity
large number of primary austenite boundaries (PABs) are observed was (70 ± 2) %. Each dry/wet cycle was 60 min, with a drying time of
(red dotted line in Fig. 1e), the microstructure is mainly composed of 48 min and an infiltration time of 12 min.
bainite and ferrite. The weld metal (WM) is mainly composed of non- After wet/dry alternating corrosion test, the specimens used for
uniform bainite and ferrite with coarse grains. Fig. 1g is the cross- weight-loss testing were ultrasonically cleaned in a descaling solution
sectional morphologies on welded joints which present various (500 mL hydrochloric acid +500 mL distilled water +3.5 g hexam-
areas. There are significant differences in the grain size and ethylenetetramine), and then washed with absolute ethanol, dried in
microstructure in the different regions of the welded joint, which high-purity nitrogen (99.999%) and weighed by a balance (mt) with
would result in the difference in the electrochemical activity between an accuracy of 0.1 mg. To determine the error caused by dissolution
the different regions, and then initiating the galvanic effect between of the steel specimen in the descaling solution, a blank specimen
different regions25–27. (without corrosion test) of the same size was used for weight-loss
correction in the descaling process. The blank specimen was placed
in the etching solution with the same time, and the weight loss was
Wet/dry alternating corrosion test
recorded as m1. The corrosion rate was calculated by:
The wet/dry alternating corrosion test was conducted to investigate
corrosion of Q690 welded joint steels in simulated marine atmo- m0  mt  m1
Va ¼ (1)
spheric environments according to standard GB/T 19746 − 201828. St

npj Materials Degradation (2024) 1 Published in partnership with CSCP and USTB
 X. Wang et al.
3

Fig. 2 Self‐designed capillary micro-cell test device schematic diagram. 3D schematic and 2D schematic of the device.

where Va is the average corrosion rate (g m−2 h−1). S is the sample scanning rate was 0.25 mV s−1. The frequency range of electro-
area (m2). t is time(h). m0 is the initial weight of the specimen. mt is chemical impedance was 10−2 ~ 105 Hz, and the amplitude of AC
the weight loss of the specimen. ml is the weight loss of blank excitation signal was 10 mV, and the temperature was 25 ± 2 °C.
specimen.
Galvanic current measurement in different regions of
SKP test welded joints
The Volta potential distributions for different regions of the Q690 Fig. 3 shows the schematic diagram of the galvanic couple current
weathering steel welded joints were measured using a VersaSCAN measurement between different regions of Q690 steel welded joint.
electrochemical workstation with the SKP technique with a size of In the dry / wet cycle accelerated corrosion testing, the CST500(ZRA)
50 × 10 mm scanned area, a 100 μm s−1 scanning step at room was used to continuously measure the galvanic current in different
temperature (~25 °C) and a relative humidity of 40%. A tungsten regions of the Q690 steel welded joint. The different parts of the
wire probe with a diameter of 100 μm was invoked as the needle, Q690 welded joint were divided, and the BM was used as the
whose amplitude and frequency were 30 μm and 1 kHz, working electrode 1. In the galvanic current measurement device,
respectively. And the average distance between the sample surface the HAZ + WM used he working electrode II. If the current measured
and the Kelvin probe was approximately 100 μm during the testing. by CST500 is positive, the working electrode I is the anode; on the
The welded joint specimens were wet ground with silicon carbide contrary, the working electrode II is the anode. The size of electrode
papers up to 2000 grit, and then polished with diamond 1 (BM) was 28 × 20 mm and the size of electrode 2 (HAZ + WM) was
suspensions (size of 3 μm). After that, the specimen was etched 26.7 × 20 mm. In addition, it is worth emphasizing that the area ratio
slightly with 4% nital (4 mL nitric acid + 96 mL ethyl alcohol) to of BM, HAZ and WM was the same as that in the dry/wet accelerated
distinguish the BM, HAZ and WM regions of the welded joint. corrosion cycle testing, and the measurement would be stopped
until the galvanic current approached zero.
SRET test
The surface current distribution of the welded joint during corrosion Morphological and element composition characterization
was measured by XMU-BY-LG scanning electrochemical workstation The morphologies of the steel specimens after dry/wet alternating
with a scanning size of 20 mm × 10 mm. The sample was first etched corrosion testing for various times were observed using a digital
to show the three regions according to the experimental steps camera (LEICAVARIO-SUMMILUX-H). After the test was completed,
described in 2.3, and the boundary of each region was marked and the specimens were sealed in epoxy resin. The cross-sectional
then the sample was re-polished. According to the previous marker, morphology of the corroded specimen was characterized by
the scanning area was selected to include BM, HAZ and WM. FE-EPMA-8050G with an acceleration voltage of 15 kV, irradiation
current of 100 nA, and beam diameter of 1 μm.
Micro-electrochemical measurements Raman spectroscopy (Renishaw, inVia Qontor, England) was
The electrochemical polarization behavior of different regions of used to determine the composition of the corrosion products,
Q690 steel welded joint was studied using a self‐designed capillary with the laser wavelength of 532 nm, the power of 50 mW and the
microcell, as shown in Fig. 2. The steel sample is used as the exposure time of 10 s. The scanning range was within
working electrode (WE), and the Ag / AgCl wire in the glass 0∼1800 cm−1, with a linear scan range of 200 μm and a sampling
capillary as the reference electrode (RE), the platinum wire dimension of 1 point per 90 μm. The results were analyzed using
wrapped outside the glass capillary as the counter electrode (CE), NanoScope Analysis 2.0 software.
in other words, RE and CE is isolated and insulated by the glass
capillary in the glass tube filled with corrosive media. The circular
ring with a diameter of Φ200 is a glass tube mouth, and the surface RESULTS
is coated with a silicone ring. Moreover, each micro-area Volta potential distribution of Q690 steel welded joint
electrolytic cell is isolated and insulated from other micro-areas The Volta potential distribution of the Q690 welded joint
due to corrosive media, which is just in the area to be tested. The measured via Kelvin probe is shown in Fig. 4. From Fig. 4b, it is
linear polarization scanning range was ± 10 mV (vs OCP) and the seen that the potential for different regions of the Q690 welded

Published in partnership with CSCP and USTB npj Materials Degradation (2024) 1
 X. Wang et al.
4

Fig. 3 The schematic diagram of galvanic couple current measurement in different regions of welded joint under wet/dry alternating
environment. These are the schematic diagrams of the detection device, periodic infiltration device, and sample for detection.

Fig. 4 Volta potential distribution in different regions of Q690 welded joint. a SKP macro testing area; (b) SKP 3D view; (c) SKP line
distribution map.

npj Materials Degradation (2024) 1 Published in partnership with CSCP and USTB
 X. Wang et al.
5

Fig. 5 Ampere SRET diagram of Q690 steel welded joint in different micro regions. a 3D figure; (b) Line distribution map; (c) SRET macro
testing area.

joint progressively decreases in the order of BM, FGHAZ, WM and CGHAZ had higher electrochemical activity, which acts as the
CGHAZ, and the maximum potential difference between these anode of the micro-region galvanic corrosion cell. This is
regions exceeds 100 mV, reaching the electrochemical conditions consistent with the testing results of SKP.
for galvanic corrosion29.
M · Stratmann30 proved that the self-corrosion potential Ecorr of Corrosion kinetics of Q690 steel and its welded joint under
the metal had the following relationship with the volta potential Δ cyclic immersion test
Φ measured by the scanning Kelvin probe test :
Figure 6 shows the average corrosion weight loss rate of
E corr ¼ ΔΦ þ Const (2) Q690 steel base metal and welded joint after the wet/dry
alternating test. It is seen that both the corrosion rates of Q690
where Const is the constant determined by the experiment. base metal and welded joint increase firstly (24h–72h), and then
Therefore, according to the measured volta potentials, the order
reaches the maximum at 72 h, and then gradually decreases.
of the Volta potentials of the different regions in the welded joint
Comparing these two corrosion rates, it can be found that the
could be judged as follows: BM (482 mV) > FGHAZ (435 mV) > WM
(417 mV) > CGHAZ (379 mV). The large potential difference in the corrosion rate of welded joint is larger than that of base metal,
different regions indicates that there may be galvanic corrosion indicating that the corrosion resistance of welded joint is worse
for the Q690 steel welded joint in the harsh marine atmospheric than that of base metal. The dotted line in Fig. 6 indicates the
corrosion environment. corrosion rate difference between the base metal and welded
joint. It could be seen that with the increase of immersion time,
the corrosion rate difference increases first and then decreases,
Surface current distribution of Q690 steel welded joint
reaching the maximum on the 6th day (0.2874 g m−2 h−1).
Figure 5 gives the surface current distribution of Q690 steel The double logarithmic model is usually used to study the
welded joint in 0.01 wt % NaCl solution. An anodic current of atmospheric corrosion kinetics of metal materials31:
about 2.944 μA is observed on CGHAZ, which is higher than those
of other regions, while a cathodic current of about −5.296 μA is
observed on BM. The higher anodic current means that the lgðCÞ ¼ nlgðtÞ þ lgðAÞ (3)

Published in partnership with CSCP and USTB npj Materials Degradation (2024) 1
 X. Wang et al.
6

Fig. 6 The average corrosion weight loss rate of Q690 steel base metal and welded joint after the wet/dry alternating test. Corrosion
kinetics curves (a), corrosion depth-time double logarithmic curve of Q690 steel (b), and it’s welded joint (c) under wet/dry cycle accelerated
corrosion test.

where C is the reduction in thickness (μm) of metal material; t is metal and welded joint are composed of two layers of corrosion
the corrosion time (h). A and n are constants. products: reddish brown outer layer and black inner layer, and
The double logarithmic model was employed to explore the gradually becomes dense (Fig. 7a). After removing the rust layer,
corrosion kinetics of Q690 steel base metal and its welded joint the surface morphology of base metal after different corrosion
with the measured weight loss corrosion rate, as shown in cycles is almost the same, showing uniform corrosion character-
Fig. 6(b, c). The n value in Eq. (1) could reflect the characteristic of istics (Fig. 7b). While for the welded joint, the samples emerge
the corrosion kinetics. n < 1 means a declined corrosion process, different degrees of groove-like corrosion morphology in the weld
and n > 1 means an accelerated corrosion process, while n = 1 area after removing the rust layer, and the corrosion groove
means a constant corrosion process. The n value transfers from gradually becomes obvious (Fig. 7b).
n > 1 to n < 1 between 72h–96h, which is consistent with the
change of corrosion rate with time in Fig. 6a. In general, the Cross-section morphology and element distribution of the
greater the n value, the greater the corrosion trend32. In the initial rust layer
stage of corrosion, the n value of the base metal and the welded The cross-sectional morphology and the element distribution of
joint is quite different (Δn = 0.12), but Δn decreases to 0.01 after the rust layers of Q690 base metal and welded joint after corrosion
72 h. Therefore, the difference in corrosion rate between for 24 days are shown in Fig. 8. It could be seen that both the rust
Q690 steel base metal and its welded joint decreases after layers of BM and WM are composed of inner and outer layers. The
forming a stable rust layer in the later stage of corrosion. overall thicknesses of the rust layers of BM and WM are about
530 μm and 580 μm, respectively. There is not much different in
The surface macro morphology of Q690 base metal and the thickness of the inner rust layer, but the cavities and cracks in
welded joint before and after removal of corrosion products the BM rust layer are less than those in the WM rust layer, and the
Fig. 7 shows the macroscopic morphology of Q690 base metal and rust layer of BM is denser. In addition, Cr, Si, and Cu elements are
welded joint before and after rust removal in different corrosion enriched in the inner rust layer, and Cu and Cr are enriched in the
cycles. With the increase of corrosion time, the rust layers on base inner rust layer near the substrate side. Furthermore, the

npj Materials Degradation (2024) 1 Published in partnership with CSCP and USTB
 X. Wang et al.
7

Fig. 7 The macroscopic morphology of Q690 steel and its welded joint. The macroscopic surface morphology of Q690 steel and its welded
joint samples before (a) and after (b) removal of corrosion products as a function of test time.

enrichment degree of Si, Cr, and Cu elements in BM rust layer is galvanic corrosion mechanism, the electrochemical corrosion
higher than that in WM rust layer, which should be relationship to behavior of different regions of Q690 steel welded joint was
the higher content of these elements in the base metal33. And our measured by self‐designed capillary microcell.
previous research also found that the Si content in steel around
0.2 wt.% had a better protective effect on the rust layer34. Micro-regions corrosion potential measurement. Figure 10 shows
the variation curve of open circuit potential with time in three
Raman spectrometer analysis of the rust layers of Q690 steel different regions (base metal, heat affected zone, and weld metal)
base metal and welded joint of Q690 steel welded joint. It could be seen from Fig. 10 that the
The Raman spectra of the rust layers formed on the Q690 steel open circuit potentials of the three different regions of the
and welded joint are shown in Fig. 9. It can be seen from Fig. 9 Q690 steel welded joint gradually decline until stable. The order of
that there is no difference in the chemical composition of the stable open circuit potentials in the three different regions is: BM
inner and outer rust layers between BM and WM, which are (−0.474 V) > FGHAZ (−0.515 V) > WM (−0.526) > CGHAZ (−0.538 V).
composed of α-FeOOH, γ-FeOOH and Fe3O4. The position of the The large potential difference between BM and CGHAZ indicates
main peak corresponding to α-FeOOH is at around 387 cm−1 20,35. that the CGHAZ has a greater corrosion tendency than the other
regions36, and the overall welded joint is more susceptible to
It is worth noting that the intensity of α-FeOOH peak in BM inner
galvanic corrosion.
rust layer is higher, which indicates that the content of α-FeOOH in
BM inner rust layer is higher. Especially at point 1, the enrichment
of Cr element can promote the transformation of α-FeOOH in the Micro-regions linear polarization (LP) measurement. The polariza-
rust layer33. In addition, the enrichment of Si result in the tion resistance Rp could reflect the corrosion rate of metal to a
formation of a stable amorphous SiO2 with a three-dimensional certain extent37. A large Rp value indicates the high corrosion
network structure in the rust layers promoting the formation of α- resistance of the steel. The linear polarization curves of the
FeOOH34. different regions in Q690 steel welded joint are shown in Fig. 11.
The fitted polarization resistance Rp values are: BM (2163.55Ω
cm−2) > FGHAZ (1705.51Ω cm−2) > WM (1284.9Ω cm−2) > CGHAZ
Electrochemical corrosion behavior of different micro-regions (960.38Ω cm−2), which indicates that BM has the best corrosion
in Q690 steel welded joint resistance, while CGHAZ has the worst corrosion resistance.
From the above experimental results, it is seen that macroscopic
galvanic corrosion could occur on the Q690 steel welded joint in Micro-regions electrochemical impedance spectroscopy measure-
the marine atmospheric environment. In order to further clarify its ment. Electrochemical impedance spectroscopy (EIS) is one of

Published in partnership with CSCP and USTB npj Materials Degradation (2024) 1
 X. Wang et al.
8

Fig. 8 Section and element distribution of rust layer of Q690 steel base metal and welded joint. a BM rust layer, (b)WM rust layer.

the effective methods to study the corrosion resistance of highest (worst corrosion resistance), which is consistent with
steel38,39. Figure 12 shows the Nyquist and Bode diagrams of the linear polarization results.
the four regions (BM, FGHAZ, CGHAZ, and WM). It is seen that the
impedance spectra of the four regions emerge as an extended Galvanic corrosion between different regions in Q690
semicircle (Fig. 14a). welded joint
In order to quantitatively analyze the corrosion resistance of
From the electrochemical measurement of each region in the
the four regions of Q690 steel welded joint, the EIS were fitted
Q690 welded joint (Figs. 10–12), it can be seen that the surface
with the equivalent circuit in Fig. 14a. The fitting results are
shown in Table 3. The CPE-Y and CPE-n are a parameter related electrochemical activity of BM is quite different from the other
to capacitance and exponential term related to roughness of three regions. Combined with the surface volta potential (Fig. 4)
electrode surface, respectively. Notably, the Y values for and current distribution (Fig. 5), it could be preliminarily inferred
CGHAZ, WM, FGHAZ and BM exhibit an upward trend, that in the early stage of corrosion, BM is inclined to be protected
indicating a gradual loosening of the corrosion product films40. in galvanic corrosion, while the corrosion of HAZ (FGHAZ and
The electrode/solution transfer resistance Rct in Nyquist CGHAZ) is inclined to be accelerated, especially for CGHAZ.
diagram, which reflects the difficulty of charge transfer at the Therefore, there may be galvanic corrosion between the local
electrode / solution interface during corrosion, is in the order: regions of Q690 steel welded joint.
Rct (BM) > Rct (FGHAZ)> Rct (WM)> Rct (CGHAZ). Meanwhile, the To intuitively and quantitatively analyze the galvanic corrosion
evolution of |Z | 0.01Hz used for qualitatively characterize the behavior of Q690 steel welded joint, the galvanic current
protectiveness of the rust film exhibits a consistent result in the between the BM and HAZ + WM couples was measured, as
corresponding Bode diagram. Therefore, the corrosion resis- shown in Fig. 13. In each dry/wet cycle, there are current
tance of BM is the best, and the surface activity of the CGHAZ is fluctuations during alternate dry/wet cycles, and no current is

npj Materials Degradation (2024) 1 Published in partnership with CSCP and USTB
 X. Wang et al.
9

Fig. 9 Raman spectrometer of the rust layers. Roman spectrometer line distribution of rust layer cross section of Q690 steel (a) BM and (b)
WM samples.

Fig. 10 Open circuit potential of different micro-regions of Q690 Fig. 11 Linear polarization curves of different micro-regions of
welded joint. The open circuit potentials of different regions in the Q690 welded joint. The linear polarization curves of different
welded joint of Q690 steel are represented using different colors. regions in the welded joint of Q690 steel, along with their respective
fitting curves, are represented using different colors.

displayed during the 48 min of drying period of each cycle. In

addition, if the measured current is positive, the working DISCUSSION
electrode I (WM + HAZ) is the anode. It can be seen from Fig.
Generally speaking, the microstructure has a great influence on
13 that a positive galvanic current is present in the initial stage of
the corrosion resistance of weathering steel before the formation
corrosion, which indicates that the WM + HAZ acts as anode and
of rust layer41. The Q690 bainitic weathering steel welded joint in
BM acts as cathode. By calculating the average value of the
this study has different electrochemical activities in different
galvanic current in the immersion stage of each cycle, the regions (BM, HAZ, WM) due to the difference in the microstructure
variation of the galvanic current with time can be quantitatively in these regions (Fig. 1) (Figs 4–5, Fig. 10). Then the corrosion
analyzed. As shown in in Fig. 14, the change of galvanic current of behavior in different regions is significantly different (Figs 10–12),
Q690 welded joint under dry/wet cycle conditions can be which leads to the occurrence of galvanic corrosion between
obviously divided into four stages: (I) initiation stage; (II) different regions (Fig. 13).
development stage; (III) polarity reversal inversion stage; (IV) Fig. 15 shows the four-stage diagram of the evolution
galvanic effect disappearance stage. Correspondingly, the galva- mechanism of the galvanic corrosion of the Q690 steel welded
nic current first rises, and reaches the highest value of 0.00256 A / joint in the marine atmosphere. In the early stage of galvanic
cm2, and then gradually decreases to zero, and finally the polarity corrosion (before 72 h) (Fig. 15a), the HAZ + WM region with high
is reversed. The negative galvanic current also increases first and electrochemical activity acts as the anode of the galvanic
then gradually decreases to zero. corrosion cell to dissolve preferentially, while the BM acts as the

Published in partnership with CSCP and USTB npj Materials Degradation (2024) 1
 X. Wang et al.
10

Fig. 13 The change curve of actual galvanic current of Q690

welded joint with time under wet/dry cycle alternation. The real-
time monitored galvanic current graph, along with its local
magnified view. The galvanic current during the drying stage is
around 0.

Fig. 12 Electrochemical impedance spectroscopy of different

regions of Q690 steel welded joints. a Nyquist diagram and
equivalent circuit diagram, (b) Bode diagram and phase angle
diagram. Rs is solution resistance, Rct is charge transfer resistance,
CPE is double charge layer constant phase element.

Table 3. Q690 welded joint fitting results.

Rs/ (Ω·cm2) Rct/ (Ω·cm2) CPE-Y (10−4Ω−1cm−2sCPE-n) CPE-n

BM 45.02 2316.49 3.436 0.6298

FGHAZ 44.41 1257.39 5.078 0.6924
WM 44.89 1231.87 5.689 0.6912
CGHAZ 44.78 882.95 6.682 0.6844 Fig. 14 The variation curve of the average value of the galvanic
current of the Q690 welded joint with time under wet/dry cycle
alternation. The entire galvanic corrosion process is divided into
cathode to be protected. This forms a small anode/large cathode four stages, designated as I, II, III, and IV in the graph, with the
galvanic corrosion cell. (the area of BM in the welded joint is larger dashed line indicating the zero current level.
than that of WM + HAZ). With the formation of rust layer in the
area of WM + HAZ, the area of small anode decreases and the
corrosion rate increases. The corrosion of welded joint enters the the third stage — polarity reversal stage of galvanic corrosion (Fig.
first stage of galvanic corrosion evolution (Figs 13–14), and 15c).
corrosion grooves begin to appear in the weld zone accordingly, In the later stage of corrosion (288 h – 576 h), with the extension
and gradually deepens (Fig. 7b). of dry / wet cycle accelerated corrosion time, all the regions of the
In the middle stage of corrosion (72 h – 144 h), most of galvanic corrosion cell of welded joint are covered by a stable rust
WM + HAZ is covered by the rust layer (Fig. 15b). Although the layer. The corrosion enters the fourth stage of galvanic corrosion
area of the small anode is further reduced, some areas in the BM (Fig. 15d). Then the galvanic effect is weakened until it disappears,
are also covered by the rust layer, resulting in a decrease in the and the galvanic current approaches to zero (Fig. 14). At this time,
overall galvanic corrosion rate (Fig. 6. 72 h – 144 h). At this time, the difference in the corrosion rate between the base metal and
the corrosion of welded joint enters the second stage with a the welded joint is reduced, the difference in corrosion depth
declined galvanic current (Fig. 14). Due to the existence of between the different regions is also gradually decreasing, and the
galvanic corrosion, the corrosion of the welded joint is corrosion grooves disappear with the time (Fig. 7. 288 h – 576 h).
decelerated. The galvanic effect43, and the corrosion resistance of the welded
With the further corrosion, the surface of WM + HAZ is covered
joint is mainly determined by the protective performance of the
by corrosion products in the middle and late stages of corrosion
(144 h – 288 h), the self-corrosion potential of WM + HAZ shifts rust layer (Fig. 8). In addition, due to the enrichment of Cr, Cu, Si
positively, which is higher than that of BM, resulting in the polarity elements and α-FeOOH which endues the rust layer with good
reversing of the galvanic corrosion cell, the BM acts the anode and protection, the synergistic effect of these factors could greatly
the WM + HAZ acts as the cathode in corrosion couple42. improve the protection ability of the rust layer, so that the
Therefore, the corrosion galvanic current also became negative corrosion rate difference between the base metal and the welded
value from the positive value in the previous stage, and entering joint tends to be stable (Fig. 6).

npj Materials Degradation (2024) 1 Published in partnership with CSCP and USTB
 X. Wang et al.
11

Fig. 15 Mechanism of galvanic corrosion kinetics of Q690 welded joint in marine atmosphere. The a, b, c, and d correspond to the four
stages (I, II, III, IV) in Figure 14, where the gray-white color represents the steel substrate and the orange-yellow color represents the rust.

In the simulated marine atmospheric environment, the whole 11. Jia, J. X., Song, G. & Atrens, A. Experimental measurement and computer simu-
corrosion process of Q690 weathering steel welded joint could be lation of galvanic corrosion of magnesium coupled to steel. Adv. Eng. Mater. 9,
divided into two stages: accelerated corrosion and decelerated 65–74 (2007).
corrosion. In the same corrosion cycle, the corrosion rate of Q690 12. Yaro, A. S., Khadom, A. A. & Idan, M. A. Galvanic corrosion inhibition behavior of
coupled copper—Steel alloys in cooling water system. J. Environ. Chem. Eng. 2,
welded joint is larger than that of its base metal, and their
2120–2128 (2014).
corrosion rate difference increased first and then decreased until 13. Zhang, G. A., Yu, N., Yang, L. Y. & Guo, X. P. Galvanic corrosion behavior of
stable with the corrosion time goes by. The corrosion rates of the deposit-covered and uncovered carbon steel. Corros. Sci. 86, 202–212 (2014).
base metal and welded joint increase first with the galvanic effect, 14. Cui, T. F. et al. Effect of NaCl concentration, pH value and tensile stress on the
and tend to be stable after the reversal. The more significant the galvanic corrosion behavior of 5050 aluminum alloy. Mater. Corros. 67, 72–83
galvanic effect, the greater the difference in the corrosion rate (2016).
between the base metal and welded joint. 15. Xing, S. H. et al. Effect of dissolved oxygen and coupled resistance on the gal-
vanic corrosion of Cr-Ni low-alloy steel/90-10 cupronickel under simulated deep
sea condition. Mater. Corros. 68, 1123–1128 (2017).
DATA AVAILABILITY 16. Tian, H. Y. et al. Combined effect of cathodic potential and sulfur species on
calcareous deposition, hydrogen permeation, and hydrogen embrittlement
The raw/processed data required to reproduce these findings cannot be shared at
of a low carbon bainite steel in artificial seawater. Corros. Sci. 158, 108089
this time due to technical or time limitations.
(2019).
17. Nicolas, A., Mello, A. W. & Sangid, M. D. Relationships between microstructure and
Received: 30 July 2023; Accepted: 12 December 2023; micromechanical stresses on local pitting during galvanic corrosion in AA7050.
Published online: 04 January 2024 Corros. Sci. 154, 208–225 (2019).
18. Deng, P. et al. Vertical galvanic corrosion of pipeline steel in simulated marine
thermocline. Ocean Eng. 217, 107584 (2020).
19. Varma, D. et al. Comparison of galvanic corrosion potential of metal injection
REFERENCES molded brackets to that of conventional metal brackets with nickel-titanium and
1. N. Wint. et al. The galvanic corrosion of welded ultra-high strength steels used for copper nickel-titanium archwire combinations. J. Contemp. Dent. Pract. 14,
automotive applications. Corros. Sci. 136, 366–373 (2018). 488–495 (2013).
2. Wang, Y. X., Tsutsumi, S., Kawakubo, T. & Fujii, H. Microstructure and mechanical 20. Wu, W., Cheng, X. Q., Zhao, J. B. & Li, X. G. Benefit of the corrosion product film
properties of weathering mild steel joined by friction stir welding. Mater. Sci. Eng. formed on a new weathering steel containing 3% nickel under marine atmo-
A 823, 141715 (2021). sphere in Maldives. Corros. Sci. 165, 108416 (2020).
3. Zhu, J. Y. et al. Galvanic corrosion of a welded joint in 3Cr low alloy pipeline steel. 21. Yadong, L. I., Qiang, L. I., Tang, X. & Yan, L. I. Reconstruction and characterization
Corros. Sci. 111, 391–403 (2016). of galvanic corrosion behavior of X80 pipeline steel welded joints. Acta Metallurg.
4. Lin, Y. J. & Lin, C. S. Galvanic corrosion behavior of friction stir welded AZ31B Sin. (Eng. Lett.) 55, 801–810 (2019).
magnesium alloy and 6N01 aluminum alloy dissimilar joints. Corros. Sci. 180, 22. Huang, C. et al. The galvanic effect of high-strength weathering steel welded
109203 (2021). joints and its influence on corrosion resistance. Corros. Eng. Sci. Technol. 54,
5. Zhu, G. Y., Li, Y. Y. & Zhang, G. A. Interaction between crevice and galvanic 556–566 (2019).
corrosion of X65 carbon steel in the CO2-saturated NaCl solution under the 23. Lu, S. W., Hu, Q., Liu, J. & Huang, F. Parameter optimization for capillary micro-
coupling of crevice and galvanic effects. J. Electroanal. Chem. 918, 116482 (2022). electeode preparation. J. Wuhan Univ. Sci. Technol. 42, 6 (2019).
6. Gao, F. Y. et al. Stress corrosion characteristics of electron beam welded titanium 24. Zhang, Y., Liu, J., Huang, F., Hu, Q. & Lu, S. Effect of composition and size of oxide
alloys joints in NaCl solution. Mater. Charact. 192, 112126 (2022). inclusions on pitting initiation of 2205 duplex stainless steel. Corros. Sci. Protect.
7. Ma, H. C. et al. Stress corrosion cracking of E690 steel as a welded joint in a Technol. 30, 8 (2018).
simulated marine atmosphere containing sulphur dioxide. Corros. Sci. 100, 25. Liu, P., Han, S. G., Yi, Y. Y. & Yan, C. X. Corrosion behavior of welded joint of Q690
627–641 (2015). with CMT twin. Int. J. Corros. 2018, 1–9 (2018).
8. Zhang, D.-L., Wang, W. & Li, Y. An electrode array study of electrochemical 26. Ahmadi, M. & Mirsalehi, S. E. Investigation on microstructure, mechanical prop-
inhomogeneity of zinc in zinc/steel couple during galvanic corrosion. Corros. Sci. erties and corrosion behavior of AISI 316L stainless steel to ASTM A335-P11 low
52, 1277–1284 (2010). alloy steel dissimilar welding joints. Mater. High Temp. 32, 627–635 (2015).
9. Shi, L. J. et al. Effect of corrosive media on galvanic corrosion of complicated tri- 27. Tian, H. Y. et al. Corrosion evolution and stress corrosion cracking behavior of a
metallic couples of 2024 Al alloy/Q235 mild steel/304 stainless steel. J. Mater. Sci. low carbon bainite steel in the marine environments: effect of the marine zones.
Technol. 35, 1886–1893 (2019). Corros. Sci. 206, 110490 (2022).
10. Fushimi, K., Naganuma, A., Azumi, K. & Kawahara, Y. Current distribution during 28. Professional Standards Compilation Group of People’s Republic of China, Corro-
galvanic corrosion of carbon steel welded with type-309 stainless steel in NaCl sion of Metals and Alloys—Alternate Immersion Test in Salt Solution[S]. China
solution. Corros. Sci. 50, 903–911 (2008). National Standardization Administration (2018).

Published in partnership with CSCP and USTB npj Materials Degradation (2024) 1
 X. Wang et al.
12
29. Zhang, T. Y. et al. Determining on welding methods and parameters for the 3%Ni ACKNOWLEDGEMENTS
steel welded joints-insights into the corrosion resistance of the fusion zone. This work was supported by the National Natural Science Foundation of China (No.
Mater. Chem. Phys. 276, 125365 (2022). U21A20113) and Major Program (JD) of Hubei Province (2023BAA003).
30. Stratmann, M. The investigation of the corrosion properties of metals covered
with adsorbed electrolyte layers-a new experimental technique. Corros. Sci. 27,
869 (1987). AUTHOR CONTRIBUTIONS
31. Ma, Y. T., Li, Y. & Wang, F. H. The atmospheric corrosion kinetics of low carbon
X.W.: data curation, formal analysis, investigation, methodology, and writing—
steel in a tropical marine environment. Corros. Sci. 52, 1796–1800 (2010).
original draft. F.H.: data curation, investigation, methodology, and writing—review &
32. Hao, L., Zhang, S. X., Dong, J. H. & Ke, W. Evolution of corrosion of MnCuP
editing. Q.H.: conceptualization and data curation. J.Z.: sample preparation. J.L.:
weathering steel submitted to wet/dry cyclic tests in a simulated coastal atmo-
investigation and supervision.
sphere. Corros. Sci. 58, 175–180 (2012).
33. Sun, M. H. et al. Fundamental understanding on the effect of Cr on corrosion
resistance of weathering steel in simulated tropical marine atmosphere. Corros.
Sci. 186, 109427 (2021). COMPETING INTERESTS
34. Liu, H. X. et al. Essential role of element Si in corrosion resistance of a bridge steel The authors declare no competing interests.
in chloride atmosphere. Corros. Sci. 173, 108758 (2020).
35. Demoulin, A. et al. The evolution of the corrosion of iron in hydraulic binders
analysed from 46- and 260-year-old buildings. Corros. Sci. 52, 3168–3179 ADDITIONAL INFORMATION
(2010). Correspondence and requests for materials should be addressed to Feng Huang.
36. Zhang, T. Y. et al. On how the corrosion behavior and the functions of Cu, Ni and
Mo of the weathering steel in environments with different NaCl concentrations. Reprints and permission information is available at http://www.nature.com/
Corros. Sci. 192, 109851 (2021). reprints
37. Nishikata, A., Zhu, Q. J. & Tada, E. Long-term monitoring of atmospheric corrosion
at weathering steel bridges by an electrochemical impedance method. Corros. Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims
Sci. 87, 80–88 (2014). in published maps and institutional affiliations.
38. Bongiorno, V., Gibbon, S., Michailidou, E. & Curioni, M. Exploring the use of
machine learning for interpreting electrochemical impedance spectroscopy
data: evaluation of the training dataset size. Corros. Sci. 198, 110119 (2022).
39. Wang, Y. et al. Understanding the role of alloyed Cu and P in the initial rust Open Access This article is licensed under a Creative Commons
composition of weathering steel formed in a simulated coastal-industrial atmo- Attribution 4.0 International License, which permits use, sharing,
sphere. Corros. Sci. 193, 109912 (2021). adaptation, distribution and reproduction in any medium or format, as long as you give
40. Sun, Y. P. et al. The dual nature of Cu content on the corrosion resistance of low appropriate credit to the original author(s) and the source, provide a link to the Creative
alloy steels in sulfuric acid: A time-dependent reversal of corrosion resistance. Commons license, and indicate if changes were made. The images or other third party
Corros. Sci. 224, 111486 (2023). material in this article are included in the article’s Creative Commons license, unless
41. Zou, Y., Wang, J. & Zheng, Y. Y. Electrochemical techniques for determining indicated otherwise in a credit line to the material. If material is not included in the
corrosion rate of rusted steel in seawater. Corros. Sci. 53, 208–216 (2011). article’s Creative Commons license and your intended use is not permitted by statutory
42. Zhang, W. H. et al. Corrosion behavior of the low alloy weathering steels coupled regulation or exceeds the permitted use, you will need to obtain permission directly
with stainless steel in simulated open atmosphere. Mater. Chem. Phys. 288, from the copyright holder. To view a copy of this license, visit http://
126409 (2022). creativecommons.org/licenses/by/4.0/.
43. Yang, Y. et al. A study of rust layer of low alloy structural steel containing 0.1% Sb
in atmospheric environment of the Yellow Sea in China. Corros. Sci. 188, 109549
(2021). © The Author(s) 2024, corrected publication 2024

npj Materials Degradation (2024) 1 Published in partnership with CSCP and USTB