Effects of rare earth metals on microstructure, mechanical properties, and pitting corrosion of 27% Cr hyper duplex stainless steel

2.1 Material

High-purity metals, including Fe, Cr, Ni, Mo, Cu, Co, and ferrochromium nitride, were used as raw materials. The experimental alloy was melted in a vacuum induction furnace (ZG-100 type; Suzhou Zhenwu Electric Furnace Co., Ltd, Suzhou, China). Different contents of REM (65 ± 5 wt% Ce–35 ± 5 wt% La–0.05 wt% Nd–0.05 wt% Pr) with purity greater than 99% were added in the tapping stage. Table 1 shows the chemical composition of the test alloy. The samples (20 mm × 20 mm × 120 mm) were cut from the alloy by wire electrical discharge and solution treated at 1,100°C for 120 min followed by quenching in water.

2.2 Experimental methods

In order to observe the optical microstructure, the alloy was ground into 1,200 particles with SiC sandpaper, polished with emery gypsum, and etched with a solution composed of 2 g of potassium sulfite, 30 mL of hydrochloric acid, and 100 mL of water for 2–5 min. The fracture morphology and corrosion pit morphology were analyzed by scanning electron microscope (SEM; ZEISS EVO-18, Carl Zeiss AG, Jena, Germany), and the chemical composition of the inclusion was analyzed by energy dispersive spectrometer (EDS) attached to SEM.

Tensile properties and Charpy impact were tested according to ASTM A370 standard test methods. Cylindrical specimens with 6.25 mm in diameter and 25 mm in gage lengths were used to test the tensile properties by 2 mm·min⁻¹ on the testing machines (DDL300; Sinotest Equipment Co., Ltd, Changchun, China). The V-notched specimens (55 mm × 10 mm × 10 mm) were used to measure the Charpy impact on the testing machine (NI150C; NSC Testing Technology Co., Ltd, Beijing, China). Mechanical properties were determined by means of the average values of four specimens of each alloy. The data were reported as mean ± standard deviation values.

In order to study the effect of REM on the pitting resistance of the alloys, the action potential anodic polarization curve and electrochemical impedance were measured by an electrochemical workstation (Autolab PGSTAT128N; MetrohmAG, Herisau, Switzerland) in a deaerated 3.5 wt% NaCl solution at room temperature. The electrochemical experiment adopts a three-electrode system; the auxiliary electrode is a large-area graphite electrode; the reference electrode is a saturated calomel electrode; and the working electrode is a test alloy sample with an exposed area of 1 cm². The scan rate of the potentiodynamic polarization curve was 20 mV·min⁻¹, and the scan range was −600 to 1,200 mV.

3.1 Effect of REM addition on microstructures

Figure 1 shows the optical metallographic structures of the test alloys. The alloys are solidified in the form of ferrite–austenite [28,29,30]. At first, the primary ferrite dendrite begins to be solidified, and the austenite phase is formed with the segregation of Ni and N elements into the melt. In the as-cast microstructure, the lighter phase is austenite, and the darker phases are ferrite and secondary phases. The secondary phases are dominated by the σ phase (Figure 2) from the ferrite phase after the slow cooling process in the ingot mold. It is obvious that the microstructure of DSS 1 alloy without the addition of REM is coarse dendrite, while with the addition of REM, the as-cast structure of the alloys is obviously refined, and the microstructure distribution is more uniform. When the content of REM increases to 0.057 wt%, the as-cast structure did not continue to be more significantly refined. After solution treatment at 1,100°C, the microstructure of the alloy is a typical two-phase structure, that is, the lighter austenite phase is distributed on the darker ferrite matrix. It is basically consistent with the as-cast microstructure. Moreover, with the addition of REM, the size of austenite and ferrite became finer and more uniform.

Figure 1

Microstructural evolution of (a) as-cast and (b) solution-treated alloys.

Figure 2

X-ray diffraction patterns of the as-cast alloys.

The solidification of molten steel in non-equilibrium state usually forms grains by non-uniform nucleation. According to the heterogeneous nucleation theory put forward by Tumbull–Vonnegut [31], the interface of liquid metal needs to be provided by solid particles with a melting point higher than the liquid melting point in the solidification process. In addition, these high melting point particles need to have a low degree of mismatch with the metal crystal phase. Combining with oxygen and sulfur elements in steel, the REM of cerium and lanthanum with high chemical activity are easily formed into high-melting-point rare earth oxides, rare earth sulfides, and rare earth oxysulfides. It is generally believed that when the mismatch degree is less than 12%, the high melting point compound can become the core of non-spontaneous nucleation [32]. Gao et al. [33] reported that Ce₂O₂S (0001) has a mismatch of 7.9% with the δ-Fe crystal plane (111), which is more likely to become the core of non-spontaneous nucleation. In addition, REM inclusions as secondary phase particles pinning grain boundaries and phase boundaries, leading to the inhibition of grain growth. Therefore, with the increase of REM content, the number of micro/nanoscale REM inclusions with high melting point increases, resulting in more core of non-spontaneous nucleation. At the same time, the more the REM atoms segregated at the grain boundary, the greater the drag effect on the grain boundary, resulting in the refinement of the as-cast structure of the alloy. Under the same heat treatment condition, the coarsening trend of the microstructure is basically the same, and the grain size after solution treatment was basically the same as that of the as-cast state.

3.2 Effect of the REM addition on the distribution of inclusions

Figure 3 shows the optical metallographic images of the as-cast alloys. Twenty fields were selected for each sample to make statistics on the size and distribution of inclusions, as shown in Figure 4. The alloy with higher purity was obtained by experimental smelting, in which the content of oxygen (<50 ppm) and sulfur (<0.005 wt%) were lower, and the inclusion with a size <3 µm accounted for a large proportion of the alloy. The largest size of the inclusion was <10 µm. In general, when the content of REM increased to 0.031 wt% (DSS 3), the average diameter and area fraction of inclusions in DSS 1 alloy decreased gradually from 3.73 µm and 0.275% to 2.33 µm and 0.154%, respectively. At the same time, the proportion of inclusions <3 µm increases, while the proportion of inclusions >3 µm decreases greatly. However, when the REM addition increase to 0.057 wt% (DSS 4), the size and area of inclusions in the alloy increase sharply, in which the proportion of inclusions <3 µm decreases. However, the number of inclusions >3 µm increases obviously. Therefore, the optimum content range of REM for inhibiting the formation of inclusions is 0.018–0.031 wt%. The statistical trend is consistent with the report of Liu et al. [24], who indicated that the addition of a certain amount of REM (0.012–0.022 wt%) in 718H steel significantly reduced the proportion of large-sized inclusions, increased the proportion of fine inclusions, and reduced the average equivalent diameter of inclusions. However, with increasing content of REM up to 0.07 wt%, the size and number of inclusions increased significantly.

Figure 3

Distribution of inclusions in optical metallographic of the as-cast alloys: (a) DSS 1, (b) DSS 2, (c) DSS 3, and (d) DSS 4.

Figure 4

Effects of REM addition on the size, area, and distribution of inclusions per frame area of the alloys: (a) the average size and fraction of inclusions and (b) size distribution of inclusions.

Figures 5a and b show the SEM-EDS analysis of the typical inclusions in DSS 1 alloy. The energy spectrum composition analysis demonstrates that the inclusions are mainly composed of Al₂O₃, MnS, and their composite inclusions. Without the addition of REM to molten steel, oxygen and sulfur elements mainly exist in the form of dissolved and deoxidized product, such as Al₂O₃. As REMs are added into the alloy as active elements, on the one hand, they react with oxygen and sulfur elements to form spherical or spindle-shaped (REM)₂O₃ and (REM) _x S _y . On the other hand, the inclusions of aluminosilicate, Al₂O₃, and sulfide in the alloy are modified into rare earth oxides, rare earth sulfides, rare earth oxide aluminum compounds, and their composite products with small sizes (Figure 5c and d). The above inclusions have greater negative free energy, which will take precedence over the formation of other oxides and sulfur oxides [18,24,34]. The micro/nanoscale REM inclusions is smaller than that of non-metallic inclusions, such as Al₂O₃ and MnS. Furthermore, a recent study reported the inclusions evolution pattern with the addition of different REMs in 718H steel [24]. The results showed that with the addition of REM, irregular MnS and Al₂O₃ inclusions were modified to elliptical REM inclusions. In addition, with the increasing content of REM, the formation process of REM inclusions was deduced as (RE₂O₂S, RE₂O₃, and REAlO₃) → (RE₂O₂S, RE₂O₃, and RES) → (RE₂O₂S and RES). Through the inclusion of thermodynamic calculation, SEM and EDS observations are consistent with the actual precipitation sequence. However, when the content of REM increases up to 0.057 wt%, excessive REM adsorbs on the formed rare earth oxygen and sulfur inclusions, promoting the absorption of more impurity elements, such as phosphorus and sulfur, as shown in Figure 5e. In addition, excessive REM addition can increase the concentration of inclusions in the alloy, and the collision and agglomeration of inclusions are more likely to form large inclusions or dense inclusions, which is detrimental to the mechanical properties of the alloy.

Figure 5

SEM-EDS analysis of inclusions in the alloys: (a and b) DSS 1, (c and d) DSS 3, and (e) DSS 4.

3.3 Effect of the REM addition on mechanical properties

Figure 6a shows the effect of REM on the mechanical properties of the experimental alloys at room temperature. With the increasing content of REM, the yield strength, tensile strength, and elongation of the alloy reached the maximum value at 0.031 wt%, which are 726 ± 8 MPa, 951 ± 12 MPa, and 44 ± 2%, respectively. Compared with DSS 1 alloy, they are increased by 3.7, 4.2, and 14.1%, respectively. Figure 7 shows that the number of dimples in the fracture of DSS 1 alloy is small, and there are large inclusions at the bottom of the hole. The dimples of the middle axis in the fracture of DSS 3 alloy become more and deeper. However, the number of dimples in the DSS 4 alloy became shallow overall despite its high number. The addition of REM can modify the nonmetallic inclusions into ellipsoidal forms such as rare earth oxysulfide and rare earth sulfides, which have higher hardness and thermal expansion properties compared to Al₂O₃ and MnS. They can improve the coordination of deformation with the surrounding matrix [25,34,35,36], thereby improving the mechanical properties and fatigue life of the material [26]. In addition, the inclusion of REM can also inhibit the segregation of harmful residual elements, such as arsenic and phosphorus, at grain boundaries and enhance the grain boundary cohesion, thereby improving the grain boundary-related properties [25]. Ahn et al. [37] presented that with the addition of 0.015 wt% Gd to 27Cr–7Ni hyper DSSs, the tensile strength, hardness, and impact toughness of the alloy were improved. There are several reasons for the excellent strength and plasticity of DSS 3 alloy. First, a large number of micro/nanoscale and uniformly distributed REM inclusions were formed with the addition of REM, which makes the microstructure refined obviously as shown in Figure 1. According to the Hall–Petch equation [38,39], the yield strength increases with the refinement of grain size, which is consistent with the experimental results. With the addition of REM, there was a small amount of solution REM in the alloy, which can also play the role of solid solution strengthening, so as to improve the strength of the alloy. In addition, the fine and uniform spherical REM inclusions are usually not formed between grains along the grain boundaries, which can bond well to the matrix and are less likely to concentrate stress, thus improving the ductility of the material [17,19]. In contrast, the size and volume of inclusions in DSS 3 were reduced and evenly distributed, which is helpful to improve the tensile properties of the alloy. However, with the increasing content of REM (0.057 wt%), the mechanical properties of DSS 4 alloy decreased. Recent research conducted in situ tensile tests on 718H steel with REM content of 0.022 wt% to elucidate the mechanism of crack initiation, propagation, and inclusions shedding [24]. The results showed that when the matrix undergoes plastic deformation, cracks started to initiate and aggregate at the ends of the REM inclusions. Under continuous loading, internal cracks and boundary cracks were generated in the REM inclusions until they were finally separated from the matrix. Therefore, the larger the average diameter of REM inclusions after plastic deformation, the faster the crack propagation rate, and thus a certain degree of strength reduction occurs. The decrease in the strength of DSS 4 alloy was most likely due to a significant increase in the average diameter and number of inclusions, which increases the crack source and crack propagation rate.

Figure 6

Effects of REM on the mechanical properties of the alloys: (a) tensile stress vs strain curves of solid-solution-treated specimens and (b) Charpy impact energy of solid-solution-treated specimens.

Figure 7

The SEM images of tensile fracture of the alloys: (a) DSS 1, (b) DSS 2, (c) DSS 3, and (d) DSS 4.

Figure 6b shows the effect of REM addition on the impact properties of the alloys at room temperature. DSS 3 has the highest impact toughness. Compared with DSS 1 alloy, the average impact toughness increases from 99.5 ± 5 to 138.1 ± 4 J. Figure 8 shows the impact fracture and energy spectrum analysis of the samples, all of which show a ductile fracture. Figure 8a shows the fracture of DSS 1 with large and shallow dimples and large inclusion particles at the bottom of the dimples. Energy spectrum analysis (Figure 8e) shows brittle inclusions such as MnS, which are easy to become crack initiation points and reduce the toughness of the material. The number of dimples on the fracture surface of DSS 3 alloy obviously increased, and the dimples became deeper. The inclusions with large size at the bottom of the dimples decreased obviously. Although there is a large number of dimples in the fracture surface of DSS 4 alloy, there are dense aggregations of REM inclusions exist (Figure 8f). It is widely accepted that the addition of appropriate amounts of REM can purify the steel, modify irregular inclusions, and refine the grain, so as to improve the toughness of the alloy. The form of impact fracture is the interaction of crack initiation and propagation. The purification of the steel by REM addition leads to a reduction in the number of inclusions, which can significantly reduce the source of cracks. Grain refinement can effectively inhibit the expansion of cracks under impact loading and reduce grain boundary stress concentration. In addition, the micro/nanoscale ellipsoidal REM inclusions can effectively reduce the stress concentration around them. It has been reported that tiny REM inclusions have a lower activity than that of most other coarse inclusions during crack initiation and propagation [37]. The REM inclusions can produce large plastic deformation and consume more energy in the process of impact fracture, therefore improving the impact toughness. Diffusely distributed fine spherical REM inclusions can be less prone to breakage during crack extension and avoid stress concentration effectively [40]. Meanwhile, another research reported that with the increasing content of REM, REM inclusions trapped the harmful elements, such as sulfur, arsenic, and phosphorus, in high-carbon chromium-bearing steel, which gradually reduced the segregation at grain boundaries and purified the grain boundaries. The addition of REM can enhance the grain boundary strength and modify the crack extension path, and thus the transverse and longitudinal impact toughness can be improved [25]. In addition, several studies have shown that REM solute segregation at grain boundaries can increase grain boundary cohesion and improve grain boundary strength [41,42]. However, when excessive REMs were added to steel, the growth mechanism of inclusions changed from precipitation to aggregate growth [25], with a continuous increase in size and a gradual deterioration in distribution and morphology. The larger inclusions formed due to excessive REM addition have different coefficients of thermal expansion from the matrix, as shown in Figure 8f. The type of inclusion tends to cause stress concentration, which significantly increases the number of crack sources and accelerates crack expansion, leading to a sharp decrease in the impact toughness of DSS 4 alloy.

Figure 8

The SEM images of impact fracture of the alloys: (a) DSS 1, (b) DSS 2, (c) DSS 3, and (d) DSS 4. SEM images and its EDS peaks of (e) DSS 1 and (f) DSS 4.

3.4 Effect of the REM addition on the resistance to pitting corrosion

Figure 9 shows the potentiodynamic polarization curves of the alloys in deaerated 3.5 wt% NaCl solution. The pitting potential (E _p) of DSS 3 alloy (E _p: 1.13 V_SCE) is higher than that of DSS 1 alloy (E _p: 1.02 V_SCE). The corrosion current density (I _corr) of DSS 3 alloy (I _corr: 9.81 × 10⁻⁸ A·cm⁻²) is obviously lower than that of DSS 1 alloy (I _corr: 5.88 × 10⁻⁷ A·cm⁻²). Compared with other alloys, DSS 3 alloy has a lower corrosion current density.

Figure 9

Potentiodynamic anodic polarization curves of the alloys after solution treated in deaerated 3.5 wt% NaCl solution.

Figure 10 shows the SEM diagrams of the alloys after overpotential polarization test in deaerated 3.5 wt% NaCl solution. The pitting corrosion occurs in the bright area in the picture, and it can be seen that with the addition of REM, the number of pitting corrosion in the experimental alloy is less than that in DSS 1 alloy, in which DSS 3 alloy showed better results.

Figure 10

SEM images of the alloys after potentiostatic polarization test in deaerated 3.5 wt% NaCl solutions: (a) DSS 1, (b) DSS 2, (c) DSS 3, and (d) DSS 4.

3.5 Mechanism of the effect of REM addition on the resistance to pitting corrosion

Figure 11 shows the SEM images of the inclusions in the alloys after the overpotential anodic polarization test in deaerated 3.5 wt% NaCl solution. Non-metallic inclusions, such as Al₂O₃ and MnS, in DSS 1 alloy were easy to be formed at the interface between ferrite and austenite, in which the inclusions have been basically dissolved to form corrosion pits (Figure 11a). Most of the REM inclusions pitting does not happen, but very few REM inclusions cause pitting corrosion due to the gap between them and the matrix [34], which can be seen in the DSS 3 alloy (Figure 11b). However, none of the REM inclusions dissolves, and the corrosion resistance of REM inclusions seems to be higher than that of the metal matrix.

Figure 11

SEM-EDS analysis of inclusions in the alloys after potentiostatic polarization test in deaerated 3.5 wt% NaCl solutions: (a) DSS 1 and (b) DSS 3.

As for the cause of pitting formation, the theory of passivation film failure has been recognized by most scholars. Suter and Böhni [43] found that the region with the highest electrochemical activity is not the inclusions or metal matrix, but the interface between the inclusions and metal matrix, which is more likely to become a sensitive site for pitting corrosion. Related researches have verified that (Cr, Mn, and Al) oxidation inclusions and MnS form Cr-depleted zones with low polarization potential and strong activation [34,44]. Therefore, the areas around such inclusions are most susceptible to corrosion. Most of the non-metallic inclusions are in sharp angle shape, and the coefficient of thermal expansion is quite different from that of the metal matrix, which is easy to form cracks at the interface [34,45]. It can be directly exposed to the Cl⁻ corrosion environment, become the origin of pitting corrosion, and then expand to the inclusions until they are dissolved. Generally, the addition of REM into an alloy can form spherical rare earth oxysulfide and rare earth oxyaluminum compounds with more negative free energy, which inhibit the formation of other coarse oxides and sulfur oxides [18,24,34]. Lu and Ives [36] have found that the thermodynamically highly stable ceria blocks the cathodic and anodic reactivity, which weakens the electrochemical corrosion of the micro-region. The small and dispersed spherical inclusions are difficult to dissolve, which can effectively prevent the corrosion of Cl⁻ to the matrix. Since the thermal expansion coefficient of rare earth oxides is close to that of the iron matrix, it is not easy to form cracks during the hot working process [34,40]. The REM inclusions reduce the micro-gap between the matrix and the inclusions, thereby reducing the possible position of pitting corrosion. In addition, a Cr-rich area is formed around the REM inclusions, which reduces the Cr-depleted area formed around non-metallic inclusions such as (Al₂O₃ and MnS) and improves the pitting corrosion resistance of the alloy [34]. Moreover, it has been discovered that the size of inclusions affects the formation of sub-stable pits, in which ultra-fine size inclusions (less than 1 µm) are difficult to form steadily growing and expanding pitting pits [46,47]. The REM inclusions marked B in Figure 11b did not form pitting pits, which is consistent with their findings. The addition of appropriate amounts of REM can purify the alloy and refine and modify the coarse and irregular inclusions. The ellipsoidal REM inclusions were likely to cause weakened electrochemical corrosion in nearby micro-regions and reduced pitting susceptibility. However, once an excessive amount of REM was added, the number of REM inclusions formed increased and the size became larger (Figures 3d and 8f), which can increase the pitting susceptibility to a certain extent and decrease the pitting resistance of the alloy.

Figure 12 shows the optical metallographic photograph of the experimental alloy after the overpotential anodic polarization test in deaerated 3.5 wt% NaCl solution. As for the DSS 1 alloy without REM addition, pitting preferentially and randomly occurs with non-metallic inclusions in the α phase (darker color) and γ phase (lighter color) (Figure 12a). However, in DSS 3 alloys with REM, pitting occurs more randomly in the α phase and extended to the γ phase (Figure 12b). Due to the higher content of the N element in DSSs, the solubility of the N element in the γ phase is much higher than that of the α phase, which significantly increases the PREN value of the γ phase. Therefore, some scholars [48,49] have proposed to increase the N content coefficient from 16 to 30 in the DSSs calculation formula. The value of PREN for γ phase in DSS 1 (PREN_γ = 60.3) is much higher than that of the α phase (PREN_α = 49.9). However, since the stable REM inclusions seem to be less susceptible to corrosion dissolution, pitting occurs more randomly in the α phase first.

Figure 12

Optical microstructures of the alloys after potentiostatic polarization test in deaerated 3.5wt% NaCl solutions: (a) DSS 1 and (b) DSS 3.