SCC behavior and susceptibility prediction mode of SMA490BW weathering steel under cathodic potential

SMA490BW weathering steel has been used to manufacture high speed train bogie structures. However, its stress corrosion cracking (SCC) behavior was not very clearly, let alone the prediction mode. In this work, combined with fracture surface analysis by SEM, electrochemical impedance spectroscopy (EIS) and potentiodynamic polarization analysis, the SCC behavior of SMA490BW steel in 3.5 wt.% NaCl solution under cathodic potentials from open circuit potential (OCP) to −1200 mVSCE (SCE = saturated calomel electrode) was studied using slow strain rate tensile (SSRT) method at 25 °C. The results indicated that the SCC susceptibility index (ISSRT) decreased with the applied potential negative shift from OCP (ISSRT = 0.31) to −800 mVSCE (ISSRT = 0.11) by anodic dissolution (AD) mechanism, and then sharply increased with the potential negatively increasing (ISSRT = 0.41 at −1200 mVSCE) by hydrogen embrittlement (HE) mechanism. The SCC susceptibility could be preliminarily predicted using the mode founded as a function of charge transfer resistance (Rt): I S S R T = 0.329 × exp − R t 2840.965 + 0.086 . This study will be helpful for the service assessment of high speed train bogie structure and other structures manufactured by SMA490BW steel.


Introduction
SMA490BW steel is a type of low alloy high strength weather steel manufactured according to JIS-G3114-2008 standard. It is widely used on bridges, constructions and other structure components due to the perfect strength and weldability. SMA490BW steel is also the optimizing material for the high speed train bogie structures [1][2][3], and its structural integrity directly concerns the safely service of high speed train. In December 2017, a Shinkansen train was found serious cracking on bogies. If the crack furtherly propagates 30 mm, the bogie structure will completely failure and lead to serious derailment accident. This crack is not only caused by manufacture problems, stress corrosion cracking (SCC) and/or hydrogen induced stress corrosion cracking (HISCC) are also main factors.
SCC is a localized corrosion failure form subject to tensile load and specific corrosive environment for a SCC susceptive metal [4,5]. In SMA490BW weathering steel, Cu, P, Cr and Ni is added as main alloying element to effectively resist the corrosive medium contacting with the steel by forming dense and well adhesive rust layer. But in marine atmosphere, the protective rust layer is hard to form due to the effects of Cl − [6]. Hydrogen evolution may appear during the corrosion reaction on the steel surface was controlled by cathodic polarization, and resulted in hydrogens adsorb on the steel surface and diffuse into the steel. The hydrogens aggregate in the defects traps such as inclusions, grain boundaries, and dislocations, induced decohesion and/or local high hydrogen pressure, facilitate the hydrogen-induced micro-cracks initiation and propagating and result in hydrogen embrittlement (HE), which is regarded as a main mechanism of SCC [7][8][9][10]. The SCC behavior is also Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI. strongly dependent on the microstructures [11,12]. For example, pearlite was considered to be susceptible to HE due to its layered structure [13][14][15].
Electrochemical methods have been used for the SCC mechanism and protective potential prediction. The electrochemical properties of the crack tip and crack wall could be respectively represented using polarization curves with fast and slow sweep rate. Thus, the SCC mechanism of a material can be primarily determined by the difference of the polarization curves [16,17]. Charge transfer resistance (R t ) measured by electrochemical impedance spectroscopy (EIS) was performed for the protective potential prediction of metals [14,18]. However, there is no mode was found to predict the SCC susceptibility using the corrosive parameters measured by electrochemical analysis.
Many studies have been applied on the general corrosion properties of weather steels, mainly about the effect of alloying elements and/or corrosive environment [19][20][21][22]. However, the SCC behavior of this type of steels is still not clear, let along the prediction mode.
In the present work, SCC behavior of SMA490BW steel in 3.5 wt.% NaCl solution under various applied cathodic potentials from open circuit potential (OCP) to −1200 mV SCE (SCE=saturated calomel electrode) was studied by slow strain rate tensile (SSRT) method at 25°C. Fracture surface analysis by SEM, potentiodynamic polarization and EIS analysis were used to study the SCC mechanism. And a SCC susceptibility prediction mode for SMA490BW steel was also founded as a function of R t .

Materials and experiment methods
Hot rolled SMA490BW weathering steel was used in this study. The chemical composition (wt.%) is listed in table 1. The addition of Ni, Cr and Cu improved its corrosion resistance. Optical metallography on the cross section etched by nital (figure 1(a)) shows banded shape pearlite and ferrite.
According to Liu [16] and Zhang [17], in corrosive environment, the crack wall could be fully polarization and in a quasi-stable state, the fresh crack tip was in a nonequilibrium state. Thus, the SCC mechanism of a material can be primarily determined by the difference of slow and fast sweep rate polarization curves. Figure 1(b) shows the potentiodynamic polarization curves of SMA490BW steel with a slow sweep rate of 0.1 mV s −1 and fast sweep rate of 10 mV s −1 in 3.5 wt.% NaCl solution at 25°C. The whole potential range could be divided into three zones at −530 mV SCE and −1040 mV SCE . At the potential above −530 mV SCE (Zone I), the SCC cracks initiation and propagation promoted by anodic dissolution (AD) mechanism due to both the crack wall and crack tip were in the anodic polarization state. At the potential between −530 mV SCE and −1040 mV SCE (Zone II), the hydrogen evolution reaction occurs but not very tempestuous while the AD reaction turns slower. So the susceptibility is not high in this zone. When the potential is below −1040 mV SCE (Zone III), tempestuous hydrogen evolution reaction occurs at both crack tip and wall, resulting in high SCC susceptibility by HE mechanism.
Based on the difference between polarization curves with fast and slow sweep rate, SSRT tests at a strain rate of 1×10 −6 s −1 controlled by a SSRT test machine were conducted in 3.5 wt.% NaCl solution at 25°C. Typical circular cross section specimens (figure 1(c)) with a gauge length of 22 mm were used according to ISO 7539-7-2005 standard. The referred to SCE potentials of OCP (−553 mV SCE ), −600 mV SCE , −800 mV SCE , −1000 mV SCE and −1200 mV SCE were applied on the specimen by three-electrodes system using CHI 660C Electrochemical Workshop as shown in figure 1(d), where a platinum sheet worked as auxiliary electrode, saturated calomel electrode worked as reference electrode, the specimen itself acted as work electrode. As shown in figure 1(e), cathode reaction occurred and resulted in hydrogen atoms adsorbed and diffused into the steel when applied cathodic potential on the specimens during SSRT tests.
The SCC susceptibility index (I SSRT ) was used to compare the SCC susceptibility of SAM490BW steel under different applied cathodic potentials, which could be calculated using equation (1): where s fw and s fA respectively are the fracture strength in corrosive and inert environment. d fw and d fA are the elongation in corrosive and inert environment, respectively. SCC susceptibility increases with I SSRT increasing from 0 to 1. The fracture side and surface morphologies were also examined after SSRT tests using optical microscope (Zeiss Stemi 2000-C) and SEM (QUANTA FEC 250). EIS test was performed to measure out the electrode reaction process. It was carried out under the potentials as same as that applied on SSRT tests with frequencies from 100 kHz to 0.01 Hz. Ten points per tenfold frequency was chosen, and 5 mV perturbation amplitude was set. The EIS test specimens were covered by epoxy resin and a 1 cm 2 area was exposed as the test surface. Before EIS test, the exposed surface was grinded and followed by polishing using silicon carbide polishing paste with a particle size of 2.5 μm. The corrosion parameters were fitted using Zview3.0a software. The test under various potentials was repeated tress times.

EIS analysis
The Nyquist plots versus applied cathodic potential resulted from EIS tests are shown in figure 2(a). All Nyquist plots exhibit a single depressed semi-circle. The semi-circles diameters increase gradually with the potential negatively increasing from OCP to −800 mV SCE , and sharp decrease observed when the applied potential negatively increasing from −800 mV SCE to −1400 mV SCE . Figure 2(b) shows the electrical equivalent circuit, where R s is the solution resistance, Q is capacitance and R t is the charge transfer resistance. The R t was obtained by conducting spectra fitting with the equivalent circuit depended on the potentials. Similar to the diameters of the Nyquist semi-circle plots, R t increased with the potential negatively increasing from OCP to −800 mV SCE , and decreased with the potential negatively increasing from −800 mV SCE to −1400 mV SCE ( figure 2(b)).
In this study, the electrode reactions only have one status variable, applied potential (E). In this status, the EIS only exhibited one time constant during each electrode reaction was performed separately. Thus, the Faraday admittance (Y Fi ) expressed as: Where I Ft is Faraday current density, R ti is charge transfer resistance. During each electrode reaction was performed separately, there is also only one capacitive reactance arc. When the two reactions occurred simultaneously, the total Faraday admittance (Y F ) could be expressed as: Therefore, the electrochemical impedance spectrum of the whole electrode also has only one time constant and capacitive reactance arc.
In the mixed potential zone (−1040 mV SCE to −530 mV SCE , figure 1(b)), anodic and cathodic reactions occur simultaneously [14,15]. Thus, R t could be divided into two parts: anodic (R ta ) and cathodic (R tc ) charge transfer resistance ( figure 3(a)). With the potential negatively increasing from OCP to −800 mV SCE , R ta increased rapidly due to the anodic overpotential reduction. However, R tc decreased slightly due to the cathodic overpotential enlarging. Thus the R t increase gradually as a combined result and got the peak value at −800 mV SCE ( figure 2(b)), where the cathodic reaction was controlled by oxygen concentration. Subsequently, the hydrogen evolution reaction was strongly assisted with the applied potential moved to more negative from −800 mV SCE to −1400 mV SCE . Here, the R t could be divided into another two parts ( figure 3(b)): oxygen reduction reaction charge transfer resistance (R tO ) and hydrogen evolution reaction charge transfer resistance (R tH ). The charge transfer resistance of the cathodic reaction notably increased owning to the occurring of hydrogen evolution reaction. As a result, R t declined rapidly with the potential shifting from −800 mV SCE to −1400 mV SCE .  As a result, the R t got the peak value at −800 mV SCE , at where the anodic dissolution was inappreciable, the cathodic reaction was controlled by oxygen concentration with weak hydrogen evolution reaction. Therefore, the SMA490BW steel could both be prevented from AD and avoid HE. As the potentials turned more negative than −800 mV SCE , the R t was sharply reduced, suggesting hydrogen evolution reaction enhancement. These results were consistent with that of potentiodynamic polarization test ( figure 1(b)).

SSRT tests
The SSRT stress-strain curves of SMA490BW steel measured in air and in 3.5 wt.% NaCl solution under various cathodic potential are shown in figure 4(a). Obviously, the corrosive environment and cathodic potential have prominent effects on the SCC behavior of SMA490BW steel. To be more specific, as shown in figure 4(b), the fracture strain values at OCP and various cathodic potential were lower than that in air. It is worth noting that the fracture strain increased slowly from OCP to −800 mV SCE and reached the peak value at −800 mV SCE . Whereafter, the fracture strain decreased sharply with the potential negative shift from −800 mV SCE to −1200 mV SCE . Except at OCP, the fracture strength shows almost no obvious change in air and other various cathodic potential. And the fracture strength shows obvious reduce at OCP is due to the pits initiation and propagation induced by AD. However, as many results from other studies [7][8][9][10], the fracture strength always exhibited little change but heavily decreasing of fracture strain during HE.
The fast and slow sweep rate polarization curves ( figure 1(b))shows that the SCC mechanism at OCP is controlled by AD, which may decreases the strength sharply [17,23]. With the potential turn to negative, the SCC behavior is mainly controlled by HE mechanism and decreases the elongation [14,16,24].
Overall considering the effects of corrosive environment and cathodic potentials on the fracture stain and strength, SCC susceptibility index I SSRT as the function of applied potential calculated by equation (1) is shown in figure 5. The I SSRT versus potential curve exhibited as a 'V' shape and has the lowest value of 0.11 at −800 mV SCE , which is about 3 times lower than that of OCP (I SSRT =0.31), and 4 times lower than that of −1200 mV SCE (I SSRT =0.41). Figure 6 shows the typical fracture surface SEM images after SSRT tests in air, OCP, −800 mV SCE and −1200 mV SCE . The micro-morphologies exhibit ductile failure features of dimples, whereas the morphologies at OCP and −1200 mV SCE show brittle rupture of quasi-cleavage features. Thus indicating that the SMA490BW steel has a certain degree of SCC susceptibility at OCP and −1200 mV SCE in 3.5 wt.% NaCl solution, and that the susceptibility was greatly lowered at −800 mV SCE . These results are consistent with the results of SCC susceptibility in figure 5. Although both of the fracture surface ruptured at OCP and −1200 mV SCE exhibited quasi-cleavage features. It should be pointed out that many corrosion products could be observed on the fracture surface ruptured at OCP, indicating that the SCC mechanism is anodic dissolution. However, there is almost no corrosion products could be observed on the fracture surface ruptured at −1200 mV SCE , indicating that the SCC mechanism is HE.
The OM images on the side face near the fracture surface are shown in figure 6. Corrosion pits could be observed on the specimen ruptured at OCP, and cracks initialed from these corrosion pits ( figure 7(a)). While on the specimen ruptured at −800 mV SCE , the side face exhibited some deformation steps near the fracture surface instead of corrosion pits and cracks ( figure 7(b)). When the potential was negative to −1200 mV SCE , a large amount of cracks could be observed on the side face near the fracture surface without any pits ( figure 7(c)).

Fracture mode
The reaction process on the steel surface at various potentials analyzed by EIS indicated that the steel will be well protected at −800 mV SCE because both of the anodic dissolution and cathodic hydrogen evolution reaction are very weak. As a result, the steel showed almost no SCC susceptibility (I SSRT =0.11) and still kept excellent plasticity with ductile fracture characterize of dimples as same as that in air.
At OCP condition, the steel exhibited very strong SCC susceptibility due to the AD mechanism [14,17]. Figure 8 gives out the schematic of SCC failure controlled by AD mechanism. During AD process, corrosion pits will form on the specimen surface and induce the stress concentration at the bottom of the pits [25]. As a result,    the micro-cracks will initiate and propagate under the external load. Therefore, the fracture surface presented as a quasi-cleavage feature [26].
With the potential negative to −1200 mV SCE , as shown in figure 9(a), drastic hydrogen evolution reaction observed on the specimen surface. The hydrogen evolution reaction in applied cathodic potential is as follow: Where H ads is the hydrogen atoms adhere on the specimen surface. Subsequently, the H ads penetrated into the steel and diffused due to its thermal motion. Hydrogen atoms intend to aggregate in the traps such as impurities, grain boundaries and dislocations, inducing dechesion and/or produced high hydrogen partial pressure [27][28][29]. This may initiate and propagate the hydrogen-induced micro-cracks in these traps, resulting in HE failure. When micro-cracks initiated, hydrogen atoms will diffuse to the crack tip and reach great high concentration ( figure 9(b)). Thus the cracking rate was ulteriorly enhanced. As a result, the steel showed skyscraping SCC susceptibility at −1200 mV SCE (I SSRT =0.41) and exhibited quasi-cleavage feature and secondary cracks on the fracture surface.

SCC susceptibility prediction mode
The charge transfer resistance R t and SCC susceptibility index I SSRT as a function of applied potentials is shown in figure 10(a). R t value increased gradually with the potential negatively increasing from OCP to −800 mV SCE , and sharply decreased with the potential negatively increasing from −800 mV SCE to −1400 mV SCE . On the  contrary, the I SSRT value decreased gradually as the potential negatively increasing from OCP to −800 mV SCE , and increased sharply when the potential negatively increasing from −800 mV SCE to −1200 mV SCE . Therefore, the relationships between R t and applied potentials, I SSRT and applied potentials could be fitted using Gauss function: Here, we suppose all the parameters as dimensionless. The fitting R 2 for R t and I SSRT are 0.94 and 0.96 respectively, which indicating that the function is suitable for fitting the experiment results.
Calculating x in equation (3): Here, ordering y as R t and I SSRT respectively. When at the same applied potential, equation (4) could be rewrite as:  (6), we can found a function between I SSRT and R t : Using equation (7) to fit the experimental data I SSRT and R t for calculating the constants A, B and t, as shown in figure 10(b) and equation (8). Thus, the SCC susceptibility prediction mode for SMA490BW steel is founded. And the R 2 value is 0.89, which indicated that this mode is suitable for the prediction. The SCC susceptibility data of E690 steel in [14] was also fitted using equation (9) and shown in figure 11 and equation (11). The fitting variance R 2 value is 0.84, indicating that this mode is also suitable for the SCC susceptibility prediction of other steels.  Figure 11. Fitting resultes of E690 steel SCC susceptibility data in [14] using equation (9).

Conclusions
In summarizing, the SCC behavior of SMA490BW steel under various applied potentials was investigated by SSRT method and electrochemical analysis.
(1) The SCC susceptibility decreased from with the applied potential negative shift from OCP to −800 mV SCE by AD mechanism, and then increased sharply with the potential turn to more negative by HE mechanism.
(2) A SCC susceptibility prediction mode was found as a function of R t : =´+ -I 0.329 exp 0.086.