Introduction

Steel structures last for decades in dry parts of inland, but fail within months in humid and salty coastal areas (Ref 1). Most of these failures are caused by corrosion under thin electrolyte layer (TEL) (Ref 2, 3). A strong corrosive TEL forms on metal surfaces near sea level due to high humidity, salinity, and long wetting time, especially in subtropical and tropical sea areas (Ref 4, 5). Thus, severe localized corrosions are often observed under TEL at marine atmosphere with high temperature and increasing salinity concentration caused by cyclic wet–dry process (Ref 3, 6-8).

Many studies (Ref 9-13) systematically revealed that hydrogen reductions extensively exist during corrosion under wet–dry cycles in marine atmosphere. For example, Huang et al. (Ref 10) found that hydrogen permeation current is produced during wet stages; this current increases with increased concentration of Cl− and SO4 2+ under the TEL. Generally, hydrogen reduction during wet–dry cycles is as follows (Ref 11, 12): The wet–dry cycles concentrate ions inside the local blocking sites, such as pits under product layer and crevices in junctions. The hydrolyzation of Fe2+ ions inside these localized blocking sites acidifies the local environment. Abundant chlorides in marine atmosphere also promote acidification. The resulting H+ ions weaken the passive film with chloridion joint and promote anodic dissolution. The Fe2+ ions newly generated by anodic dissolution are hydrolyzed, thereby producing many H+ ions. The repeating process continuously acidifies the medium in those local blocking sites. Huang et al. (Ref 9) verified by direct measurement that the pH beneath the water film and rust layer can be approximately 4.5. Consequently, hydrogen evolution reactions are promoted with increasing concentration of H+ ion. Therefore, hydrogen embrittlement (HE) or hydrogen-induced stress corrosion in steels might still occur in marine atmosphere (Ref 14-16). Ma et al. (Ref 17) studied the stress corrosion behaviors of E690 steel under marine atmosphere. They found that cathodic reactions are controlled by both oxygen diffusion and hydrogen reduction. Moreover, the resulting FeSO4 promotes reproduction of hydrogen ions, which induce the stress corrosion cracking (SCC) of E690 under marine TEL. Li et al. (Ref 18) evaluated the susceptibilities of high-strength steels to delay fracture; they found that the fracture stresses of AISI 4135 steels decrease with increased diffusible hydrogen content, which increases with wet–dry cycles. Tsuru et al. (Ref 19) indicated that the hydrogen evolution mechanism on passive films is different from that on a bear iron surface, which enhances hydrogen permeation. Zhang et al. (Ref 20) investigated the HE of galvanized steel in simulated marine atmosphere. Their results showed that galvanized steel exhibits high HE susceptibility in marine atmosphere. This susceptibility is also affected by the relative humidity and the zinc coating coverage. Therefore, the hydrogen behavior and its effects on stress corrosion of steels at marine atmosphere or under TEL are complicated, and they need further investigations. Moreover, reports focusing on the effect of hydrogen on stress corrosion of stainless steel under marine atmosphere or TEL are few given.

Due to its relatively high strength and good resistance to pitting and SCC, duplex stainless steel (DSS) is extensively applied in some severe environments (Ref 21-26). However, DSS is attacked by SCC in marine atmosphere (Ref 4, 27-30). Glowacka et al. (Ref 31) found that α phase in DSS gradually decays, in addition, the stacking faults in γ phase significantly increase with prolonged hydrogen charging time; this increase is probably related to strong grain refinement. However, the relation between stress corrosion and microstructure under hydrogen charging is not considered in their study. Zakroczymski et al. (Ref 32) revealed that the reduced ductility of steel is closely associated with the hydrogen concentration cathodically charged hydrogen. In their other research (Ref 33, 34), they found that hydrogen behaves differently in α and β phases: Diffusible hydrogen in ferrite matrix and reversibly trapped hydrogen in austenite phase induce different concentrations in different phases. Luu et al. (Ref 35) also detected large hydrogen distribution and high permeation and diffusion rates in ferritic phase. Kheradmand et al. (Ref 36) also found that the hydrogen effects on different phases are strongly associated with the existing residual stress, and these effects differ in each phase. Silverstein et al. (Ref 37, 38) applied two advanced techniques, namely thermal desorption spectrometry and secondary ion mass spectrometer, on examining the local hydrogen concentration, trapping, and distribution in 2205 DSS; their study provided further understanding regarding the HE mechanism of DSS. Although considerable efforts have been devoted to determine how hydrogen affects the mechanical properties of DSS, most of the existing reports have ignored the role of environment, such as the scale effect of TEL.

In the present work, a typical DSS, namely 2205 DSS, was used as the experimental material. The electrochemical and stress corrosion behaviors of 2205 DSS charged with different densities of cathodic current were investigated under 3.5 wt.% NaCl TEL by electrochemical measurements, slow strain rate tensile (SSRT) tests, and fracture surface characterizations. The effect of hydrogen on the SCC behavior of 2205 DSS under TEL was discussed. Conclusions were drawn on the basis of the results.

Materials and Experiments

Materials and Sample Preparation

The 2205 DSS was used as-received. The chemical composition is listed in Table 1. The metallographic microstructure is shown in Fig. 1. Carbon and sulfur contents were measured by infrared absorption method. Other contents were measured by inductively coupled plasma atomic emission spectrometry. The metallographic microstructure was obtained by etching the polished matrix using aqua regia; the volume ratio of HNO3 (16 mol/L) to HCl (12 mol/L) is 1:3. Austenite (white areas) and ferrite (gray areas) phases are distributed alternatively and uniformly with approximately equal volume shared. The mechanical properties of the studied 2205 DSS were evaluated according to the standard GB/T 228-09. The test was performed on a MTS Landmark 370 material test system. Size of the specimens is shown in Fig. 2. The proof stress (σ p0.2), tensile strength (σ b), elongation (δ), and reduction-in-area (ϕ) of the 2205 DSS are 583 MPa, 777 MPa, 19.6%, and 58.6%, respectively.

Table 1 Chemical composition of the 2205 duplex stainless steel (wt.%)
Full size table
Fig. 1
figure 1

Optical microstructure of 2205 DSS

Full size image
Fig. 2
figure 2

Size of specimens used for mechanical testing

Full size image

Specimens for hydrogen charging and determination and electrochemical measurements were machined plates with dimensions of ϕ5.64 mm × 3 mm. The specimens were welded to a copper wire for electric connection and mounted in epoxy resin with an exposed area of 10 mm2. The exposed surface was progressively ground with emery to 1500 grit, cleaned with deionized water and alcohol, and dried with a blower.

Specimens for SSRT tests were machined according to the standard GB/T15970.7-2000. The specimen size is shown in Fig. 2. The working section was ground with emery paper to 1500 grit, then degreased with alcohol, and rinsed with deionized water. The specimens were covered with silica gel, and the working section was left exposed.

Hydrogen Charging and Determination

Hydrogen was introduced in advance in the matrix by electrochemical method. The charging solution was aqueous with NaOH (0.25 M) and thiourea (0.25 g/L). Pure nitrogen was piped in the solution for 2 h to deaerate before charging. A typical three-electrode system was applied for hydrogen charging; this system included a 2205 DSS specimen as the working electrode (WE), a saturated calomel electrode (SCE) as the reference electrode (RE), and a platinum plate as the counter electrode (CE). Constant cathodic currents of 10, 20, 50, and 100 mA cm−2 were applied on specimens. In this study, the hydrogen charging current density (HCCD) magnitude may be considerably higher than the magnitude in actual service. The hydrogen content of substrate in actual service resulted from years of accumulation. HCCD in laboratory should be tens or hundreds of times that of hydrogen reduction current density in free and actual corrosion so that it can simulate the actual hydrogen content of the substrate after years of service. This HCCD magnitude was also adopted in many previous studies (Ref 39, 40). Hydrogen charging continued for 24 h for each experiment. According to the method above, specimens for hydrogen determination, electrochemical measurements, and SSRT tests were precharged cathodically.

The three-electrode system was also used for hydrogen determination. Specimens were rinsed with NaOH (0.2 M) solution immediately after charging and placed in NaOH (0.2 M) solution for hydrogen determination. Pure nitrogen was piped in the solution for 2 h before the determination and continuously piped during determination. Specimens were positively polarized to 168 mV (versus SCE). Subsequently, the polarization current was recorded as a function of time. A specimen without hydrogen charging was also prepared using the same process as the blank control. The hydrogen oxidation current was varied with time by subtracting the polarization current of the blank specimen from that of the charged specimen. Hence, the hydrogen content introduced in the charged specimen was quantified by integrating oxidation current on time.

Electrochemical Measurement

The electrochemical measurements conducted on the precharged specimens under TEL included potentiodynamic polarization and electrochemical impedance spectroscopy (EIS) measurements. The setup for electrochemical measurements under TEL was in accordance with that in the literature (Ref 3). The WE, RE, and CE were encapsulated together with epoxy resin (Fig. 3) and gradually ground with emery paper to 1500 grits. Prior to measurement, specimen was hydrogen-charged according to the method mentioned in section 2.2. Afterward, the capsule was placed on a horizontal platform within 3.5 wt.% NaCl atmosphere for 2 h to form a steady TEL on its surface. During the measurement, temperature was kept at 30 ± 2 °C. The TEL thickness was measured by a device consisting of ohmmeter, screw micrometer, and bronze needle and kept at around 100 μm by controlling the salt mist flow rate. The thickness (100 μm) utilized in this study was within the range of that adopted by most previous studies (Ref 41, 42). A Princeton VMPS electrochemical workstation was used to acquire data. All potential data were obtained versus SCE. For potentiodynamic polarization measurements, the potential scanning rate was 0.5 mV/s, and the potential polarization ranged from −750 to 2000 mV versus open circuit potential. For EIS measurements, the test frequency ranged from 10 kHz to 0.1 Hz, and the AC amplitude was 10 mV. A software ZsimpWin V3.20 was used to fit the EIS data.

Fig. 3
figure 3

Sketch of electrode capsule for electrochemical measurement under thin electrolyte layer (TEL)

Full size image

SSRT

SSRT tests were carried out on the charged specimens with a material test system (Letry Co., Ltd., Model: WDML-3). The setup used for SSRT test under TEL is depicted in Fig. 4. Mist nebulized from 3.5 wt.% NaCl solution was continuously injected in the box to form a TEL on the working section. To keep the TEL thickness at approximately 100 μm (same as that of electrochemical measurement), the flow rate was consistent with that of the electrochemical measurement. The temperature was also kept at 30 ± 2 °C. The tensile test began after 2 h of mist injection. The tensile rate was 0.0012 mm/min.

Fig. 4
figure 4

Setup used for slow strain rate tensile (SSRT) test under TEL environment

Full size image

After tensile test, the specimen was descaled with nitric acid (100 mL of HNO3 with distilled water added to 1000 mL), cleaned, and dried. Reduction-in-area and elongation were calculated according to the measurement of cross-sectional area and the length of the working segment before and after fracture. Finally, the fracture surfaces were observed using scanning electron microscope.

Results

Electrochemical Results

The potentiodynamic polarization curves of 2205 DSS with different HCCDs under 3.5 wt.% NaCl TEL are shown in Fig. 5. Hydrogen charging mainly promotes the anodic part of polarization curves. With increased HCCD, the passivation current increases, and the pitting potential decreases, thereby decreasing the passive zone. Information obtained from Fig. 5, including pitting potential, corrosion potential, polarization current density at 0.1 V, and corrosion current density, was replotted against HCCD in Fig. 6. Corrosion potential and corrosion current density were obtained from the intersection of linear fitting lines of anodic and cathodic strong polarization zones in activation phase. With increased HCCD, the pitting potential and corrosion potential decrease at current density of 0.1 V, and the corrosion current density increases. Moreover, all the variation tendencies change slowly when HCCD exceeds 50 mA cm−2. Hence, the uptake amount or the electrochemical effect of hydrogen may approach a limit when HCCD exceeds 50 mA cm−2.

Fig. 5
figure 5

Effect of hydrogen charging current density (HCCD) on the potentiodynamic polarization curves of 2205 DSS under 3.5 wt.% NaCl TEL

Full size image
Fig. 6
figure 6

Variation in electrochemical indexes obtained from the polarization curves (Fig. 5) with HCCD

Full size image

Figure 7 shows the EIS of 2205 DSS with different prior HCCDs under 3.5 wt.% NaCl TEL and the corresponding simulation circuit. The Nyquist curves exhibit a decreasing radius with increasing HCCD. This result indicates that the corrosion resistance of 2205 DSS decreases with increased HCCD. As shown in Fig. 7(b), R s is the resistance of solution between reference electrode and working electrode, Q f is the capacitance of passive film, R f is the resistance of passive film, Q dl is the capacitance of double layer, R ct is the resistance of charge transfer, and W is the diffusion impedance. R f and R ct variations with HCCD are plotted in Fig. 8. R f decreases remarkably with increased HCCD, whereas R ct increases relatively slightly. Therefore, hydrogen charging suppresses the electrochemical reactivity, but weakens the passive film.

Fig. 7
figure 7

Electrochemical impedance spectroscopy of 2205 DSS with different HCCDs under 3.5 wt.% NaCl TEL: (a) Nyquist curves, (b) bode absolute value against frequency, (c) bode phase value against frequency, and (d) the corresponding simulation circuit

Full size image
Fig. 8
figure 8

Variation in passive film resistance and charge transfer resistance with HCCD

Full size image

Hydrogen Determination

Hydrogen exists in metal matrix in the form of an atom. Driven by the anodic discharging electric field, the diffusible hydrogen in subsurface diffuses to the electric double layer; those hydrogen atoms, as well as some metal atoms, are oxidized to hydrogen anions. Consequently, the anodic oxidation current is generated. Hydrogen-oxidized current, I H, can be obtained by subtracting the anodic oxidation current of the uncharged sample from that of the charged sample. Figure 9 shows the I H variations at different prior HCCDs on discharging time. Evidently, I H decreases markedly before 250 s and slowly after 250 s. This result indicates that hydrogen atoms are mainly absorbed on or near the metal surface. Thus, hydrogen charging may considerably influence the metal surface.

Fig. 9
figure 9

Effect of HCCD on the hydrogen-oxidized current of 2205 DSS under 3.5 wt.% NaCl TEL

Full size image

The damage caused by charging hydrogen can be quantified with the average concentration of the absorbed hydrogen in matrix (C H). C H is related to the corresponding quantity of electricity (Q):

$$C_{\text{H}} = \frac{Q}{zFv},$$
(1)

where z is the electrons whose value for hydrogen atom is 1, F is the Faraday constant (96487 C·mol−1), and v is the effective volume of the specimen (0.3 cm3) in this study. Q in Coulomb is obtained by integrating I H with respect to time, namely by calculating the area under the I H curve (Ref 43).

$$Q = \int {I_{\text{H}} } {\text{d}}t.$$
(2)

The C H values under different HCCDs, integrated in the range of 0–3580 s, are listed in Table 2. Notably, C H increases with increased HCCD. C H values at different HCCDs are replotted in Fig. 11. C H immediately increases with increasing HCCD before 50 mA cm−2 and slowly from 50 to 100 mA cm−2, which is consistent with the contents depicted in Fig. 6 and 8. Thus, the influence of HCCD on the electrochemical behavior of 2205 DSS under the TEL is directly related to the amount of hydrogen uptake. When HCCD exceeds 50 mA cm−2, the amount of hydrogen uptake in 2205 DSS is close to saturation. Given the influence of hydrogen charging, this amount approaches a limit.

Table 2 Hydrogen concentration absorbed in the substrate (C H) under different HCCDs
Full size table

SSRT Tests After Various HCCDs

Figure 10 shows the stress–strain curves of 2205 DSS under different prior HCCDs at 3.5 wt.% NaCl TEL. The 2205 DSS elongation decreases with increased HCCD. The susceptibility of 2205 DSS to SCC is defined as the loss of reduction-in-area (I ψ) or loss of elongation (I δ):

$$I_{\psi } = (\psi_{0} - \psi_{m} )/\psi_{0} \times 100\%$$
(3)
$$I_{\delta } = (\delta_{0} - \delta_{m} )/\delta_{0} \times 100\%$$
(4)

where ψ 0 and ψ m are reduction-in-area measured in air and under the TEL, and δ 0 and δ m are elongation measured in air and under the TEL, respectively. Figure 11 depicts I ψ, I δ, and C H together as a function of HCCD. The variation in I ψ and I δ with HCCD is consistent with that in C H. Apparently, the SCC susceptibility of 2205 DSS directly correlates with hydrogen concentration in substrate.

Fig. 10
figure 10

Effect of HCCD on SSRT stress–strain curve of 2205 DSS under 3.5 wt.% NaCl TEL

Full size image
Fig. 11
figure 11

Variation in stress corrosion cracking susceptibility and concentration of absorbed hydrogen in matrix (CH) with HCCD

Full size image

Figure 12 shows the effect of prior HCCD on the fracture morphology of 2205 DSS in 3.5 wt.% NaCl TEL. The left and right sides are the macroscopic and microscopic views, respectively. Macroscopic images show that the necking degree of the fracture decreases with increasing HCCD. Microscopic images show an increasing brittle rupture with increasing HCCD. The fracture surface characteristic is a combination of dimples and cleavages. With increasing HCCD, the dimple area ratio increases. These results indicate that hydrogen charging remarkably increases the SCC susceptibility of 2205 DSS under the TEL.

Fig. 12
figure 12

Fracture morphologies of SSRT specimens under 3.5 wt.% NaCl TEL with different HCCDs: (a) 0, (b) 10, (c) 20, (d) 50, and (e) 100 mA cm−2

Full size image

The side surfaces of the SSRT specimens were also investigated and are shown in Fig. 12. Microcracks are observed on all side surfaces. Their numbers and lengths both increase with increased HCCD. Furthermore, few pits are observed on the side surface, and not all microcracks initiate from pits or any other form of localized corrosion. These results indicate that hydrogen charging promotes SCC susceptibility of 2205 DSS, mainly through HE rather than localized anodic dissolution. Moreover, the secondary cracks propagate straightly, not zigzag, which means that the crack propagation of 2205 DSS under 3.5 wt.% NaCl TEL is intergranular.

Discussion

Effect of Prior Hydrogen Charging on the Electrochemical Processes

The anodic reaction of 2205 DSS under 3.5 wt.% NaCl TEL is the following iron oxidation (Ref 11, 44):

$${\text{Fe}} \to {\text{Fe}}^{2 + } + 2{\text{e}}^{ - } .$$
(5)

The cathodic reaction is primarily the O2 reduction:

$${\text{O}}_{2} {\text{ + 2H}}_{ 2} {\text{O + 4e}}^{ - } \to 4 {\text{OH}}^{ - } .$$
(6)

Furthermore, oxygen is exhausted in some localized areas, such as the interior of pit holes, crevices, and crack tips, where oxygen diffusion is impeded. These places are commonly called the blocking area. The anodic products (Fe2+ ions) inside these areas are hydrolyzed (Ref 11, 44):

$${\text{Fe}}^{2 + } {\text{ + 2H}}_{ 2} {\text{O = Fe}}\left( {\text{OH}} \right)_{2} {\text{ + 2H}}^{ + } .$$
(7)

In addition to the continual hydrolyzation of Fe2+ ions, the hydrogen ion in the blocking area is remarkably sufficient for hydrogen reduction:

$${\text{H}}^{ + } {\text{ + e}}^{ - } \to {\text{H}}_{\text{ads}} .$$
(8)

In this reaction, hydrogen atoms are introduced electrochemically as a simulation of the produced hydrogen above. Figure 5 shows that hydrogen charging decreases the pitting resistance of 2205 DSS under the TEL. The uptake hydrogen increases the corrosion current density and passivation current density and decreases the passivation zone by decreasing the pitting potential. This phenomenon is also observed and explained by previous report about austenitic stainless steel (Ref 45). The charged hydrogen atoms hinder chromium enrichment in surface films. EIS results show that the effect of HCCD on electrochemical process is mainly weakening the R f without increasing the R ct (Fig. 7). This result indicates that the uptake hydrogen mainly blocks the formation of passive film rather than promotes the electrochemical reaction. This observation is verified on the side surfaces of SSRT specimens, where no obvious trace of anodic dissolution can be observed (Fig. 12). Generally, microcracks can be found inside the corrosion pits when the SCC is induced by anodic dissolution; this observation is also demonstrated in many previous reports (Ref 46-48). Therefore, hydrogen charging influences the SCC behavior of 2205 DSS under the TEL by weakening the passive film without promoting localized anodic dissolution. Consequently, the weakened passive film allows the hydrogen atoms (produced in Eq 7) to permeate the matrix easily.

Effect of Prior Hydrogen Charging on the Stress Corrosion Behavior

Figures 10, 11, and 12 indicate that the SCC susceptibility of 2205 DSS increases with increased hydrogen concentration in the matrix. This result confirms that the SCC susceptibility of 2205 DSS under the TEL is closely related to hydrogen. Combined with the discussion in the last paragraph, hydrogen charging can be concluded to influence the SCC of 2205 DSS under TEL mainly through increasing its embrittlement. This result partly verifies that the SCC mechanism of 2205 DSS under the TEL is HE, when the conditions of localized hydrogen charging are available (Fig. 13).

Fig. 13
figure 13

Side morphologies of SSRT specimens under 3.5 wt.% NaCl TEL with different HCCDs: (a) 0, (b) 10, (c) 20, (d) 50, and (e) 100 mA cm−2

Full size image

Furthermore, it is noted that the susceptibility to HE of 2205 DSS under the TEL is highly consistent with hydrogen concentration in matrix. The SCC susceptibility increases more sharply when HCCD ranges from 20 to 50 mA cm−2 than when HCCD is below 20 mA cm−2 or above 50 mA cm−2. It means that the uptake efficiency of hydrogen in matrix is enhanced when HCCD exceeds 20 mA cm−2. There are two sources for this enhancement: ① One is that transformation of Hads (Eq 8) to H2 is depressed when HCCD exceeds 20 mA cm−2 and ② the other one is that the passive film is weakened greatly so that hydrogen atoms can diffuse into the matrix more easily. Generally, transformation of Hads to H2 will be not depressed, but promoted with increasing HCCD. So the first source cannot be the main reason. For the second source, Figure 8 shows that resistance of the passive film decreases greatly with increasing HCCD. It verifies the second source at a certain extent. Thus, it can be speculated that the weakened passive film results in the sharply enhancement of hydrogen uptake efficiency when HCCD ranges from 20 to 50 mA cm−2.

It is also noted that the SCC susceptibility and hydrogen concentration increase slowly when the HCCD exceeds 50 mA cm−2. This trend is assumingly attributed to the saturated hydrogen concentration in the substrate. Farrell (Ref 49) indicated that hydrogen concentration in ferrous alloys reaches a saturation level after prolonged charging. The saturation time and hydrogen concentration depend on charging conditions (electrolyte type, charging current, and charging temperature), type of steel, and sample geometry. However, other researchers believe that a considerably large number of hydrogen bubbles forming on the working electrode prevent the further introduction of hydrogen atoms (Ref 39). Thus, the reason of hydrogen saturation requires further investigation.

Conclusion

  1. 1.

    In this paper, the effects of hydrogen charging the stress corrosion behavior of 2205 duplex stainless steel have been investigated under 3.5 wt.% NaCl thin electrolyte layer. Based on the above results and discussions, the following conclusions can be drawn. 2205 DSS shows a nice resistance to pitting corrosion under hydrogen charging, although hydrogen weakens the passive film.

  2. 2.

    The stress corrosion mechanism of 2205 DSS under 3.5 wt.% NaCl TEL is mainly hydrogen embrittlement rather than localized anodic dissolution.

  3. 3.

    2205 DSS shows a more susceptibility to hydrogen under the TEL when hydrogen charging current density is between 20 and 50 mA · cm-2. Its SCC susceptibility increases remarkably when HCCD increases from 20 to 50 mA cm−2 and slowly when HCCD exceeds 50 mA cm−2.