Environmentally Assisted Cracking of Duplex and Lean Duplex Stainless Steel Reinforcements in Alkaline Medium Contaminated with Chlorides

Abstract

Herein, the corrosion performance of different stainless steel (SS) reinforcing bar grades in alkaline solution is presented, including UNS S32205 duplex stainless steel (DSS), UNS S32304 and UNS S32001 lean DDS (LDSS). The electrochemical dissolution kinetics were studied by potentiodynamic polarization and the Tafel slope method. The environmentally assisted cracking (EAC) mechanisms of the different SS grades in the presence of Cl⁻ were revealed with the slow strain rate test (SSRT). The higher activation of the anodic branch and the loss of toughness were related to the austenite-to-ferrite phase ratio. UNS S32205 DSS presented the slowest anodic dissolution kinetics, mainly due to the higher austenite content compared to the other LDSS; however, it suffered a more severe EAC than the UNS S32304 LDSS. In the case of UNS S32001 LDSS, even while having the lowest Ni content (i.e., large ferrite α-phase ratio), it experienced the least decrease in elongation as well as low anodic dissolution kinetics for Cl⁻ contents up to 8 wt.%, where the Cl⁻ threshold was reached.

1. Introduction

Duplex stainless steels (DSS) were developed based the idea that the superior properties of the duplex microstructure, consisting of an even ratio of ferrite (α-phase) and austenite (γ-phase), give enhanced strength and corrosion resistance over austenitic stainless steels (ASS), such as UNS S32205. The corrosion resistance of DSS is attributed to its alloy composition, promoting the formation of a stable Cr-rich film, whose thickness is higher than the one formed in ASS, as well as showing a low density of microstructural point defects (lower oxygen vacancies and metal interstitials) [1]. To increase the passive film stability, the γ/α-phase ratio needs to be adjusted, which in most cases, is achieved through thermal treatments [2].

In addition to the DSS, novel low-Ni lean-duplex SS grades have been developed in the recent decades to reduce the cost, due to price fluctuations of Ni in the stock market. In this regard, the manufacturing of lean-duplex SS with lower Ni and/or Mo content led to the new LDSS, including UNS S32304, UNS S32101, UNS S32001, and UNS S32404, among others [3,4]. The use of LDSS types is suggested mostly based on their pitting and environmentally assisted cracking (EAC) resistance in aggressive environments, and they are a cheaper substitute for ASS [5]. UNS S32304 was the first commercialized lean duplex SS, being developed as a σ-phase-free DSS, in addition to its lower cost due to the Mo savings [6]. In recent years, LDSS low-nickel UNS S32001, and UNS S32101 have been used as SS reinforcements, compensating their lower nickel content with a higher Mn content, making them more economically interesting than UNS S32205 and UNS S32304, but with all four SS grades presenting similar resistance to chloride induced corrosion [7].

However, when DSS and LDSS are welded or thermally treated, the γ/α-phase ratio is modified, thus promoting a diffusion of chromium and lowering nickel in the γ-phase region [8]. The lower the γ-phase, the lower the pitting corrosion resistance, which can be determined by the critical pitting temperature and the pitting resistance equivalence number calculations [4,9]. Pit nucleation is favored by the α-phase, as reported throughout the literature, due to the α-phase having lower chloride corrosion threshold values [10]. In addition, due to the presence of the duplex microstructure, the development of the galvanic micro-couple corrosion is triggered, which is one of the key factors for the lowering of the pitting corrosion resistance of DSS [11]. It has been reported in the literature that a difference of 25% in the modulus of impedance between the γ-phase and the α-phase was recorded through localized electrochemical impedance spectroscopy (LEIS), while an over six times increase in current (from 0.01 to 0.06 mA) was observed using the scanning vibrating electrode technique (SVET). Cao et al. studied the galvanic micro-couple of DSS 2205 via single-phase samples prepared by selective dissolution, where immersion experiments were performed in 1 M H₂SO₄, showing that as the immersion time increased, the α-phase dissolved due to galvanic corrosion, increasing the pit depth [12]. On the corrosion behavior of these LDSS, it has been shown that at low pH, the passive film formed on LDSS 2001 is richer in Cr³⁺ than that formed on AISI 316 ASS [13]. While in high-alkaline simulated concrete pore solution (pH 12.6), LDSS 2001 reinforcements have shown similar corrosion resistance to AISI 304 ASS, in both carbonated (pH 9.1) and non-carbonated electrolyte solutions contaminated with chlorides [14]. LDSS 2304 reinforcement in simulated concrete pore solution showed lower corrosion resistance than DSS 2205, but still outperformed the pitting breakdown potential of AISI 204 ASS.

Furthermore, DSS present improved EAC performance compared to ASS, in particular in relation to stress corrosion cracking (SCC), as the γ-phase stops the cracking, stopping the crack branching, and making the crack advance trasngranularly by α-phase or intergranularly through the α/γ-phase interface [15,16,17]. This is because of the high stress intensity factor for SCC (K_ISCC) that DSS have compared to austenitic, in particular, the K_ISCC of the γ-phase [18]. The α-phase has a higher cracking susceptibility because of its body-centered cubic (BCC) crystal structure and corresponding slip system [19]. However, under severe deformation degree, such in the case of cold rolling, the γ-phase gathers higher residual stresses than the α-phase, making it more susceptible to develop SCC [20]. However, the α-phase is still the preferential site for pit nucleation, which increases its corrosion rates as the degree of cold rolling increases. Martin et al. showed that UNS S32205 reinforcements under slow strain rate testing (SSRT) when immersed in high-alkaline solution had a maximum crack velocity of 5.27 × 10⁻⁹ m/s in 8 wt.% Cl⁻, which increased over 100% for the same chloride concentration in carbonated buffered solution, reaching 1.23 × 10⁻⁸ m/s [21]. The solution became further acidified due to the formation and dissociation of carbonic acid, leading to higher cracking susceptibility of the ferrite cleavage facet; however, it remained at a low crack velocity. Ruel et al. found that LDSS 2101 had lower SCC resistance than LDSS 2304 in acid solution (pH 2.8) contaminated with chlorides (50 g/L NaCl) at room temperature (20 °C), suggesting that Mn has a negative effect on SCC resistance, while N has a positive one [22]. Briz et al. tested the SCC behavior of B500SD carbon steel (CS), DSS 2205 and LDSS 2001 reinforcements in concrete pore solution contaminated with chlorides, showing that while both UNS S32001 and UNS S32205 were immune to SCC under the tested EAC conditions, as their elongation-to-fracture ratios were close to unity, CS B500SD showed the greatest reduction in the elongation to fracture (24%), as well as a 35% reduction in σ_UTS [23].

Nevertheless, there is still a lack of knowledge about the EAC mechanisms, and in particular, the electrochemical corrosion kinetics and the mechanical degradation of DSS immersed in alkaline environments contaminated with chlorides. For that reason, this study seeks to unravel the EAC mechanisms of the three main DSS/LDSS reinforcements available in the market, S32205, S32304, and S32001. To measure the change in the corrosion kinetics, potentiodynamic polarizations were monitored as a function of the chloride content, and for the degradation of mechanical properties, SSRT was the preferred method for studying EAC.

2. Materials and Methods

2.1. Materials

The reinforcing bars used for this study were 10 mm diameter (size #3) UNS S32205 and UNS S32304, and 3 mm diameter UNS S32001. The elemental composition of the DSS and LDSS reinforcing steels is shown in

Table 1. Elemental composition of reinforcing bar (wt.%), Fe balance.

Alloy	C	Cr	Mn	Ni	Mo	N	Si
UNS S32205	0.017	22.76	1.57	4.64	3.21	0.171	0.34
UNS S32304	0.019	22.75	0.81	4.32	0.29	0.138	0.35
UNS S32001	0.028	20.07	4.19	1.78	0.22	0.129	0.65

. Three type of specimens were used: (i) for the potentiodynamic polarization testing a 3 cm length rebar was used, where only 1 cm² was exposed (samples were degreased with acetone and ethanol); (ii) SSRT specimens had a circular 60° V-notch in the center of the sample to accelerate the cracking process; and (iii) for microstructure characterization, the samples were epoxy mounted (showing the cross-section of the rolling direction) and polished with different grades of SiC paper up to mirror finishing, finishing with diamond powder (1 μm). The microstructure was revealed with an electrochemical attack with 40 wt.% NaOH at 3 V for 5 s [24].

2.2. Testing Method and Environment

Electrochemical testing via potentiodynamic polarization was performed with 3 cm length rebar samples using a three-electrode configuration cell setup in a Gamry 600 potentiostat. The potentiodynamic polarization was selected to show the changes in the electrochemical corrosion kinetics as well as the pitting susceptibility of the different alloys as a function of the chloride content. The reference electrode (RE) used in this test was a saturated calomel electrode (SCE), the counter electrode (CE) used was a graphite rod, and the working electrode (WE) was the reinforcing bar. Three different chloride concentrations were tested (0, 4 and 8 wt.% of Cl⁻) by means of CaCl₂ additions (to avoid any additional cations in solution other than Ca²⁺) to the simulated concrete pore solution (SCPS, pH 12.6) made out of a saturated Ca(OH)₂ aqueous solution. The polarization potential scan ranged from −0.3 V_OCP to +1.0 V_OCP at a potential scan rate of 0.1667 mV s⁻¹, following ASTM G61-86 [25].

SSRT testing was performed using a fresh SCPS electrolyte solution contaminated with 0, 4 and 8 wt.% of Cl⁻ to investigate the influence on the mechanical property degradation. The SSRT were performed using a uniaxial tensile test media following ASTM-G129 [26]. The SSRT were performed at a strain rate of 1 × 10⁻⁶ s⁻¹ to enhance the number of environmental events.

2.3. Characterization Techniques

The characterization of the samples was performed with a scanning electron microscopy (SEM) (Tescan Lyra 3 XMU). For the phase identification, X-ray diffraction (XRD) analysis was conducted using a Rigaku SmartLab-3kW X-ray diffractometer, with a Cu target (K_α = 1.5406 Å), and a scan step of 2°/min over the 2θ range of 35°–100°. Optical imaging was performed after potentiodynamic polarization testing with a metallographic microscope Nikon eclipse MA 100.

3. Results

3.1. Microstructure Characterization

The microstructure in the rolling direction of each SS reinforcement (UNS S32205, UNS S32304, and UNS S32001) can be seen in Figure 1, where the austenite γ-phase grains (light grey) are embedded in the ferrite α-phase matrix (dark grey) [27]. All three SS reinforcing grades with varying γ/α-phase ratios showed similar microstructures. No sign of inclusions, segregations, precipitates or carbides were seen in the as-received samples.

3.2. Slow Strain Rate Testing (SSRT)

The stress–strain curves of the tested SS reinforcements (UNS S32205, UNS S32304 and UNS S32001) for each chloride concentration can be seen in Figure 2 [28]. The EAC behavior for the different chloride additions revealed a severe impact on the elongation to failure (ε_f), followed by the ultimate tensile stress (σ_UTS), and elongation at the σ_UTS (ε_UTS), while the change in yield stress (σ_y) was minor (see

Table 2. Mechanical properties of the reinforcements after SSRT in SCPS (pH 12.6) at different chloride concentrations.

[Cl−] wt.%	σy MPa	σUTS MPa	εUTS %	εf %
UNS S32205
0	510	715	15.91	21.07
4	463	560	13.05	17.37
8	407	479	9.95	12.02
UNS S32304
0	565	759	23.92	27.77
4	631	652	20.65	24.94
8	483	575	17.52	21.11
UNS S32001
0	466	609	29.33	51.66
4	428	532	25.09	42.65
8	388	507	23.88	38.21

) [29].

UNS S32304 showed increased mechanical properties during EAC testing, having the highest σ_UTS and σ_y. However, the alloy with the highest elongation was UNS S32001, despite having low σ_UTS and σ_y values. UNS S32205 had σ_UTS and σ_y values similar to those of UNS S32304 in the absence of chlorides; nevertheless, EAC produced a decrease in the σ_UTS and σ_y values with increased chloride addition.

The UNS S32001 LDSS showed the highest ductility (ε_f) compared to the other two SS reinforcing grades, even after EAC test exposure at high chloride concentrations, which were 3.17 and 1.81 times larger than the values for UNS S32205 and UNS S32403, respectively.

3.3. Potentiodynamic Polarization

The potentiodynamic polarization curves of the SS reinforcements (UNS S32205, UNS S31653, and UNS S32304) immersed in SCPS at the different chloride concentrations are presented in Figure 3. The analysis of the electrochemical curves is summarized in

Table 3. Electrochemical properties of the reinforcements after SSRT in SCPS (pH 12.6) at different chloride concentrations.

[Cl−] wt.%	Ecorr mVSCE	icorr A/cm2	βa mV/dec	βc mV/dec	B
UNS S32205
0	−58	5.14 × 10−8	287	249	57
4	−189	3.58 × 10−8	285	225	54
8	−255	5.44 × 10−7	197	176	40
UNS S32304
0	−264	2.38 × 10−7	186	216	43
4	−329	2.14 × 10−6	172	205	40
8	−455	1.61 × 10−6	166	197	39
UNS S32001
0	−248	5.13 × 10−8	277	285	60
4	−254	6.34 × 10−8	202	265	49
8	−268	1.82 × 10−7	154	245	41

where the corrosion potential (E_corr), corrosion current density (i_corr) and Tafel slopes can be found. As a common trend, the higher the chloride content, the lower the E_corr and the higher the i_corr, where UNS S32001 showed the smallest changes, followed by UNS S32205 and UNS S32304. While UNS S32001 and UNS S32205 did not reveal a pitting potential in the selected potential scan range, UNS S32304 did after the 4 wt.% Cl⁻, thus indicating it is the most active reinforcing SS grade from the three in this study [30].

Similarly, the i_corr of UNS S32304 was the greatest among the three of them, regardless of the chloride content. Both parameters are consistent in demonstrating that UNS S32304 is the most active alloy. The only case in which the i_corr was not constantly increasing with the chloride addition was the change from 0 to 4 wt.% Cl⁻ for UNS S32205, where the 0 wt.% Cl⁻ had a slightly higher value. This can be attributed to the heterogeneity of the corrugated section of the rebar; nevertheless, the difference between the conditions is relatively small.

4. Discussion

4.1. Phase Quantification by XRD

As seen in Figure 1, all the duplex SS alloys had a similar microstructure, with only changes in the ratio of the γ/α-phase. Figure 4 portrays the X-ray diffraction pattern of the as-received samples, where both the face-centered cubic (FCC) γ-phase (JCPDS No. 33-0397) and body-centered cubic (BCC) α-phase (JCPDS No. 06-0694) were found [31]. The ratio of the γ/α-phase was quantified by integrating the respective intensity peaks, giving γ/α ratios of 62/38, 51/49 and 45/55, for UNS S32205, UNS S32304, and UNS S32001, respectively [27].

The trend of decreasing γ/α-phase ratio coincides with a decrease in corrosion performance, as seen from the potentiodynamic polarization curves, as well as a loss of mechanical properties, as seen in the stress–strain curves. This is due to the γ-phase having a higher corrosion potential between 40 and 70 mV greater than the α-phase, due to the higher Ni and N constituents, which result in a lower oxidation rate, thus making the α-phase the anodic region [32]. Hence, the lower the γ/α-phase ratio, the more anodic areas there are and the higher the corrosion susceptibility.

4.2. Mechanical Properties Degradation Analysis

To further assess the loss of mechanical properties due to the SCC effect, several ratios based on ASTM G129-21 were calculated, where the values used were taken from Table 2 [26,33]. First, the plasticity lost (I_δ) (see Equation (1)) is as follows [26]:

$$I_{\delta }=\left(\frac{{\epsilon }_{f,b}-{\epsilon }_{f,SCC}}{{\epsilon }_{f,b}}\right)\times 100$$

(1)

where ε_f,SCC is the elongation to failure of the sample tested under SCC and ε_f,b is the elongation to failure of the sample tested in 0 wt.% chlorides.

Additionally, the ratio of elasticity loss (REL) is considered (see Equation (2)) [26,33]:

$$REL=\left(\frac{{\sigma }_{y,b}-{\sigma }_{y,SCC}}{{\sigma }_{y,b}}\right)\times 100$$

(2)

where σ_y,SCC is the yield strength of the sample tested during SCC and σ_y,b is the yield strength of the 0 wt.% chlorides.

In addition, the SCC susceptibility index (I_SCC) was also used as this ratio is widely used to assess how harsh the environment is (see Equation (3)) [34]:

$$I_{SCC}=\left[\left(\frac{1+{\epsilon }_{UTS,b}}{1+{\epsilon }_{UTS,SCC}}\right)\times \left(\frac{{\sigma }_{UTS,b}}{{\sigma }_{UTS,SCC}}-T\right)\right]\times 100$$

(3)

where ε_UTS,SCC is the elongation to σ_UTS of the samples tested under SCC, ε_UTS,b is the elongation to σ_UTS of the 0 wt.% chlorides, σ_UTS,SCC is the σ_UTS of the samples tested under SCC, and σ_UTS,b is the ultimate tensile strength of the sample tested in 0 wt.% chlorides.

Table 4. Mechanical properties of the reinforcements after SSRT in SCPS (pH 12.6) at different chloride concentrations.

[Cl−] wt.%	ISCC %	Iδ %	REL %
UNS S32205
4	33.31	17.56	21.68
8	76.09	42.95	33.01
UNS S32304
4	7.55	10.19	14.10
8	22.23	23.98	24.24
UNS S32001
4	22.30	17.44	12.64
8	27.88	26.04	16.75

combines all the calculated parameters for each DSS reinforcement and environment. It can be seen that, despite the outstanding mechanical properties of UNS S32205, it had the highest I_SCC and REL. Regardless of the better corrosion properties that UNS S32205 showed compared to the other LDSSs, the severe reduction in ε_f was the main cause for the degradation of mechanical properties. UNS S32304 had similar initial mechanical properties as UNS S32205, with the difference that it did not degrade as fast, resulting is a lower percentage for all three calculated parameters.

However, UNS S32001 had the most similar I_SCC values and the lowest REL as the yield did not vary as much with the chloride addition. Furthermore, although the I_δ value was greater than that of UNS S32304, even at the highest chloride concentrations, its ε_f value was almost double that of the other samples. This shows that UNS S32001 could be a great candidate where the structures would require good long-lasting mechanical properties.

4.3. Electrochemical Corrosion Kinetics Analysis

To further compare the corrosion behavior of the alloys, an electrochemical corrosion kinetic analysis was performed by the Tafel method using both the cathodic and anodic branches. The results are presented in Table 3. The Tafel slopes in the cathodic (β_c) and the anodic (β_a) branches represent the reactivity of the processes, where the lower the values the faster the kinetics are. This means that for a small change in potential, there is a decade increase in current. Conversely, the higher the values, the less relevant the process is. When both the anodic and cathodic Tafel slopes are combined, the Stern–Geary constant (B) is obtained (see Equation (4)). As a common rule, for carbon steel reinforcements, a B value of 52 is considered passive, while a B value of 26 is considered active [35].

$$B=\frac{{\beta }_a{\beta }_c}{2.303\left({\beta }_a+{\beta }_c\right)}$$

(4)

The B value can be interpreted as the overall reactivity of the alloy in the immersed environment, where the higher the value the more passive it is [36]. UNS S32205 had B values above 52, even at 4 wt.% chlorides, supporting the low i_corr values (less active); it is only at 8 wt.% chlorides that the B value lowers to 40. UNS S2304 had a starting value of 43, much lower than UNS S32205. However, it did not vary with the chloride increase, remaining at 40. This is also seen in the little change in its β_a value, which did not severely decrease. Finally, UNS S32001 had a B value greater than UNS S32205; nevertheless, for chloride concentrations greater than 0 wt.%, the B values remain similar. It can be seen that the B values never reached the limit of 26, which shows that regardless of the high chloride content, the samples remained fairly passive.

For the SS reinforcements immersed in the high alkaline solution with chlorides, the anodic reactions of the β_a were dominated by the dissolution of the passive film, first the iron oxides (Fe₂O₃ + H₂O + 2e⁻ ⇆ 2Fe₃O₄ + 2OH⁻) and then the chromium oxides (Cr₂O₃ + 5H₂O ⇆ 2CrO₄²⁻ + 10H⁺ + 6e⁻). Here, the γ/α-phase ratio is one of the determining parameters since the austenite is richer in chromium and more stable than the ferrite. Once the passivity breakdown occurs, the dissolution of the metal initiates. Conversely, the cathodic reactions of the β_c are dominated by the oxygen reduction reaction (ORR) (O₂ + 2H⁺ + 4e⁻⇆ 4OH⁻) and the formation of the Fe₂O₃ oxide (as per the Pourbaix diagram). The more relevant the β_c are, the less the passivity is compromised as the SS reinforcements are able to form a more stable passive film.

Comparing the Tafel slopes from the three different alloys, it can be seen that UNS S32205 had the highest β_a and that it barely changed, even at the highest chloride concentrations. In this case, the β_c had lower values, being the fastest process, thus showing how passive UNS S32205 was. In contrast, UNS S32304 had the fastest β_a among all three alloys and chloride contents, except for UNS S32001 at 8 wt.% Cl⁻, where it had the lowest value with 154 mV/dec. The drop in β_a at the highest chloride concentration shows that the chloride threshold of UNS S32001 was reached, consistent with previous studies found in the literature [37]. This is also reflected in the highest β_c values which show that their relevance in the kinetics is lower.

4.4. Pitting Assessment after Potentiodynamic Polarization

After the potentiodynamic polarization, the samples were cleaned and studied by microscopy to reveal the severity of the environment on pitting corrosion. For this reason, only the highest chloride concentration (8 wt.% Cl⁻) was used (see Figure 5). It can be seen that UNS S32205 preferentially formed pit clusters; however, their size was relatively smaller than those of the other two DSS grades. Low pit density was found, proving the higher corrosion protection, which is in agreement with the higher β_a values obtained by Tafel analysis. Similarly, UNS S32304 had great corrosion protection properties (see Table 3); nevertheless, more pits nucleated, leading to pit clustering. Finally, UNS S32001 showed the largest pit density surface, having not only greater pit sizes, but also more nucleating spots. The greater number of pits can be related to the large α-phase region, as seen from the XRD. The decrease in the γ/α-phase ratio and the higher activity of the α-phase lead to faster anodic electrochemical kinetics, which increases passivity breakdown potential and, therefore, results in a higher chance of developing pits.

4.5. Fracture Analysis after SSRT

After the SSRT, the samples immersed in 8 wt.% Cl⁻ were selected for a fractographic analysis with the SEM. Figure 6 depicts all three SS reinforcements and their main features. UNS S32205 showed a mainly ductile fracture with some minor ductile overload areas (see Figure 6a). Where most of the ductile overload and ferrite cleavage facets were found, the presence of broken carbides was seen (see Figure 6b). These carbides are the reason for the more brittle failure, developing the cleavage facets. UNS S32304 suffered more from the chloride attack, developing more brittle features such as the greater ductile overload areas with cleavage facets (see Figure 6c), and even some cracks (see Figure 6d). While UNS S32304 had great corrosion properties, the lower γ/α-phase ratio, in conjunction with the applied stress, led to the degradation of the mechanical properties, promoting SCC failure. Finally, UNS S32001 was the sample which developed the most SCC failure, with well-defined cleavage planes (see Figure 6e) and severe cracks along the cleavage planes (see Figure 6f). As was seen with the electrochemistry and kinetics analysis, UNS S32001 reached its chloride concentration, promoting the higher anodic reactions, which now in combination with the stress, resulted in severe SCC susceptibility.

5. Conclusions

In this work, the EAC and corrosion behavior of duplex and lean duplex stainless steel reinforcements (UNS S32205, UNS S32304, and UNS S32001) were studied as a function of the chloride concentration. The main conclusions can be drawn as follows.

UNS S32205 had the highest β_a slope values (slowest anodic electrochemical kinetics) and it barely changed even at the highest chloride concentrations. UNS S32304 had the highest β_a among all three alloys and chloride contents tested, except for UNS S32001 at 8 wt.% Cl⁻, where it had the lowest β_a value with 154 mV/dec. The drop in β_a at the highest chloride concentration showed that the chloride threshold of UNS S32001 was reached.

After the potentiodynamic polarization test, the samples showed increased pit density with the reduction in the γ/α-phase ratio, thus promoting more sites for pit nucleation and clustering.

UNS S32205 had a severe decrease in mechanical properties, with a SCC susceptibility of 76% at 8 wt.% Cl⁻, while UNS S32304 and UNS S32001 only developed 22 and 27% susceptibility, respectively. UNS S32001 had a ε_f value, even at the highest chloride concentrations, that was 80% greater compared to the other SS reinforcements. This shows that UNS S32001 could be a great candidate where structures require good long-lasting mechanical properties.