Correlation between evolution of inclusions and pitting corrosion in 304 stainless steel with yttrium addition

Effects of the evolution of inclusions on the pitting corrosion resistance of 304 stainless steel with different contents of the rare-earth element yttrium (Y) were studied using thermodynamic calculations, accelerated immersion tests, and electrochemical measurements. The experimental results showed that regular Y2O3 inclusions demonstrated the best pitting resistance, followed in sequence by (Al,Mn)O inclusions, the composite inclusions, and irregular Y2O3 inclusions. The pitting resistance first decreased, then increased, and then decreased again with increasing Y content, because sulfide inclusions were easily generated when the Y content was low and YN inclusions were easily generated at higher Y contents. The best pitting corrosion resistance was obtained for 304 stainless steel with addition of 0.019% Y.

With the progress of smelting technology, sulfide inclusions have been largely removed and gradually replaced by oxide inclusions, which greatly enhances pitting corrosion resistance. Jun et al. argued that conditions where the number of sulfide inclusions was smaller than that of the oxide inclusions and the size and distribution of oxide inclusions were small and dispersed would improve the pitting corrosion resistance of high clean 304 stainless steel 24 . There are, however, few studies focusing on modifying oxide inclusions by adding rare-earth Y to improve the pitting corrosion resistance of clean 304 stainless steel.
In this study, we determined the effects of the evolution of inclusions on the pitting corrosion resistance of 304 stainless steel with Y addition using thermodynamic calculations, potentiodynamic polarization and immersion tests, and scanning electron microscopy with energy-dispersive spectroscopy (SEM-EDS) analysis of inclusions.

Methods
Materials and specimen preparation. The raw material used in this study was 304 stainless steel of the chemical composition given in Table 1. The experimental alloys were prepared using a Si-Mo electrical resistance-heated furnace. The Y contents of the experimental alloys were 0, 0.007%, 0.013%, 0.019%, and 0.049%. Figure 1 shows the Si-Mo furnace with (a): resistance furnace body; (b): body sketch; (c): program console; (d): argon tank. A smelting process was executed by switching on the flows of argon gas and cooling water, covering the furnace mouth with refractory bricks, and then programming an appropriate procedure into the console. After cooling of the furnace, the specimen was not heat treated in any other way. Specimens of the 304 stainless steel with different Y contents were used for counting inclusions and for the immersion and electrochemical tests. To avoid surface defects, the test surfaces were ground with 2000 grit silicon carbide paper and polished with 0.5 μm diamond paste, then rinsed with deionized water, degreased in alcohol, and dried immediately.
Determination of inclusion evolution process. The types and sizes of inclusions in the 304 stainless steel specimens with different Y contents were observed and analysed by SEM-EDS. Several hundred inclusions were randomly selected and classified according to their distribution on a test surface using Factsage 7.0 software to reveal their evolution. After choosing the database, the Equilib software module was selected, the alloy compositions were input (where the Y content was set as a variable), the desired phase (as determined by SEM) was selected, and the compositions of inclusions with increasing Y contents in the matrix were calculated.

Electrochemical measurements.
To reveal the pitting trends of 304 stainless steel with different Y contents, potentiodynamic polarization tests were conducted using a three-electrode configuration in 3.5% NaCl solution at 298 K. A copper wire was attached to the rear side of each specimen and mounted in an epoxy resin and the test surface was ground and polished. A platinum sheet and a saturated calomel electrode (SCE) were used as the counter and reference electrodes, respectively. The working electrode was the specimen, the exposed area (test surface) of which was 1 cm2. Potentiodynamic polarization tests were carried out using a Solartron 1287 power supply. Tests were conducted in the potential range of −0.5 V SCE to + 0 V SCE at a scanning rate of 3.8 × 10 −4 V/s.  The test solution comprised 350 ml deionized water with 69.9 g FeCl 3 •6H 2 O and 20 ml HCl (36-38 mass%). Each specimen was sealed with epoxy resin except for the test surface to avoid formation of porosity and cracks around the surface. After reaching the set immersion time, the specimen was immediately removed from the test solution, rinsed, and dried. SEM-EDS was used for observing the corrosion morphology and analysing the compositions of inclusions after different immersion times.

Results
Counting inclusions and thermodynamic calculations. Five compositions of 304 stainless steel with different Y contents were obtained by tube furnace smelting. The principle of counting related to the type and size of inclusions after cooling in the furnace. The results of the thermodynamic calculations explained the evolution of inclusions with increasing Y content. The statistical results for inclusions in 304 stainless steel with different Y contents are shown in Fig. 2. When the Y content was zero, the proportion of (Al,Mn)O inclusions was largest, followed by those of (Al,Mn,Si)O wrapped in (Al,Mn)O. When the Y content was 0.007%, the inclusions were mainly MnS, followed by (Al,Y) O wrapped in (Al,Y) x (SO) y and (Y,Mn) x (SO) y wrapped in MnS. Only two types of inclusions could be found in the steel matrix when the Y content increased to 0.013%: a larger proportion of irregular Y 2 O 3 inclusions and a smaller proportion of regular Y 2 O 3 inclusions. The proportion of irregular Y 2 O 3 inclusions was significantly smaller than that of regular Y 2 O 3 inclusions when the Y content reached 0.019%. YN inclusions mainly presented in 304 stainless steel containing 0.049% Y, followed by regular Y 2 O 3 inclusions and then irregular Y 2 O 3 inclusions. The average size of the inclusions decreased and then increased with increasing Y content. Inclusions with the smallest average size were found in 304 stainless steel containing 0.013% Y.
The morphologies of the main inclusions in 304 stainless steel with different Y contents are shown in Fig. 3, where a, b, and c denote the main inclusions in steel matrices containing 0%, 0.007%, and above 0.013% Y, respectively.
The results of the thermodynamic calculations (calculated by Factsage 7.0) are shown in Fig. 4. Figure 4a shows that the proportion of Y 2 O 3 inclusions gradually increased and reached a peak at ~0.012% Y, while YN inclusions gradually increased. Figure 4b shows that the proportions of Al 2 O 3 and MnO inclusions decreased while those of MnS increased and then decreased. YN inclusions were surprisingly generated at ~0.012% Y. The  Electrochemical results. Potentiodynamic polarization tests were performed on the 304 stainless steels with different Y contents to determine their pitting corrosion resistance. The results are shown in Fig. 5; the corresponding corrosion potentials and pitting potentials are shown in Table 2. When the Y content was 0.007%   or 0.049%, the corrosion potential (−0.35 V) was lower than for other Y contents, which indicated that severe pitting corrosion susceptibility occurred at these Y contents. The highest pitting potential (−0.11 V) was achieved by 0.019% Y, indicating that the 304 stainless steel with 0.019% Y possessed the best pitting corrosion resistance.    Table 2. The corrosion potential and pitting potential in potentiodynamic polarization curves of 304 stainless steel with different Y contents.  Figure 5 shows that the pitting potential disappeared in 304 stainless steel containing 0.007% Y because the corrosion potential was greater than the pitting potential, so no passive region was attained: pitting corrosion occurred, indicating the weakest resistance to pitting corrosion by this steel composition. From these data, it was evident that the pitting potential first decreased, then increased, and then decreased again with increasing Y content.       Miro-crevices formed in the inclusion/matrix boundary at 5 min ( Fig. 6a and b), while Fig. 6c shows the formation of a micro-crevice after 5 s. The morphologies of inclusions in 304 stainless steel containing 0.007% Y after immersion for 0 s, 5 s, 2 min, and 5 min are shown in Fig. 7, where a shows (Al,Y) Figure 7a and b show serious corrosion, resulting in large pits after 5 s, while Fig. 7c shows the formation of a micro-crevice after 5 s and the development of corrosion along the inclusion/ matrix boundary. The inclusion in Fig. 7d exhibited the best pitting corrosion resistance: serious corrosion only occurred after 5 min.
The morphologies of inclusions in 304 stainless steel containing 0.013% Y after immersion for 0 s, 5 s, 15 min, and 18 min are similarly shown in Fig. 8, where a and b represent irregular and regular Y 2 O 3 inclusions, respectively. Figure 8a shows pit formation after 5 s, the depth of which increased with time, while b showed no dissolution until 18 min. This trend is similar to that shown in Fig. 9, which represents 304 stainless steel containing 0.019% Y, where Fig. 9a and b show regular and irregular Y 2 O 3 inclusions, respectively.
The inclusion morphologies in 304 stainless steel containing 0.049% Y after immersion for 0 s, 5 s, and 18 min are shown in Fig. 10, where a represents YN inclusions and b shows regular Y 2 O 3 inclusions. Figure 10a shows that large pits formed after 5 s, while b exhibited little or no corrosion after 18 min. This indicated that lower pitting corrosion resistance was achieved by 304 stainless steel containing 0.049% Y because of the generation of YN inclusions.
Pitting corrosion is randomly induced in stainless steel 26,27 . We therefore calculated the proportions of pits induced by inclusion after immersion for 5 s, 5 min, and 18 min in 304 stainless steel with different Y contents, as shown in Table 3. The results showed that different types of inclusions induced different trends of pitting initiation at these times. Pitting initiation induced by regular Y 2 O 3 inclusions did not all occur after 18 min, which indicated that these inclusions had the best pitting corrosion resistance.

Discussion
The evolution of inclusions in 304 stainless steel with different Y contents directly affected their pitting resistance. The inclusion size first decreased and then increased with increasing Y content (Fig. 2). When the 304 stainless steel contained 0.013% Y, the average inclusion size in the matrix was lowest, but its pitting corrosion resistance was weaker than that of 304 stainless steel containing 0.019% Y (Fig. 5). Kim et al. argued that the addition of rare-earth elements changed the composition and shape of inclusions and decreased the inclusion size in duplex stainless steel 4 . Ha et al. illustrated that rare-earth elements reduced the size and surface density of (Mn,Cr,RE)-oxysulfide inclusions 20 (Figs 2-4). This evolution behaviour ensured that 304 stainless steel containing 0.007% Y gave the weakest resistance to pitting corrosion, because most inclusions (the composite inclusions) were etched and almost half were severely etched (Fig. 7). A large number of YN inclusions were severely etched after 5 s in 304 stainless steel containing 0.049% Y, but its pitting corrosion resistance was greater than the steel containing 0.007% Y because of the generation of regular Y 2 O 3 inclusions (Fig. 10). The best pitting corrosion resistance was exhibited by 304 stainless steel containing 0.019% Y, which contained the largest proportion of regular Y 2 O 3 inclusions (Figs 5, 9).

Conclusions
In 304 stainless steel with Y addition, pitting corrosion resistance induced by inclusion was not predominantly related to the inclusion size, but also to the inclusion type. The composite inclusions and irregular Y 2 O 3 inclusions showed the weakest resistance to pitting corrosion.
The pitting corrosion resistance first decreased, then increased, and finally decreased again with increasing Y content (within the range of 0%-0.049% Y) in 304 stainless steel. MnS inclusions and the composite inclusions were produced when Y contents were relatively low and YN inclusions formed at relatively high Y contents, both of which deteriorated the pitting corrosion resistance. Regular Y 2 O 3 inclusions gave the best pitting corrosion resistance: the higher the number of inclusions, the better was the pitting corrosion resistance. The best pitting corrosion resistance was exhibited by 304 stainless steel containing 0.019% Y.