2019 Volume 60 Issue 9 Pages 1968-1976
The effect of Ti addition on the oxidation resistance of high-purity 19%Cr ferritic stainless steels has been investigated during isothermal heat treatment at temperatures between 1073 and 1273 K in air. The microstructures of the scale and scale/metal interface were investigated in detail using scanning electron microscopy-electron backscatter diffraction, field emission-transmission electron microscopy, together with micro-energy dispersive X-ray spectroscopy.
Ti addition increased the oxidation mass gains and simultaneously improved the oxidation resistance limit temperature. The formed scale consisted mainly of Cr2O3 regardless of the addition of Ti, but the addition of Ti increased the thickness of the Cr2O3 layer. Moreover, the addition of Ti considerably reduced the grain size of Cr2O3, and this was inferred to increase the oxidation mass gain owing to the easy diffusion of metal ion through grain boundaries. Furthermore, Ti was oxidized in the region underneath the scale/metal interface and formed internal complex oxides, such as Al2TiO5 and Al2Ti7O15 owing to the presence of small amounts of Al in the used steels. θ-Al2O3 also formed in the somewhat deeper region from the interface, where Ti could not be oxidized. The oxygen-getter effect of Ti atoms was postulated to be responsible for improving the oxidation resistance of the alloys.
This Paper was Originally Published in Japanese in J. Japan. Inst. Met. Mater. 82 (2018) 130–139. In order to clearly explain, the captions of figures and tables were changed. The references were also changed.
Fig. 12 TEM images and EDS maps illustrating Ti content at Cr2O3 grain boundaries. Points (a), (d), (e) are located at grain boundaries and points (b) and (c) are located inside grains.
Stainless steels contain large amounts of Cr. Exposing stainless steels to high-temperature environments leads to the formation of Cr-rich protective oxide films or scale on their surfaces. This scale confers stainless steels excellent corrosion resistance, such as oxidation resistance, at elevated temperature. Moreover, as they are high alloyed steels, stainless steels exhibit excellent mechanical properties at elevated temperature, and therefore, are used as heat-resistant materials.1)
Recently, a category of steels called “high-purity ferritic stainless steels” has been developed.2) These are the interstitial-free ferritic stainless steels, which contain ultra-low amounts of C and N, along with added Ti and/or Nb as stabilizing elements. These steels present improved mechanical properties and weld corrosion compared to conventional ferritic stainless steels such as Type 430, and therefore, can be used for a wide range of applications, including automotive exhaust systems. As these steels are used at elevated temperature, oxidation resistance is one of their important required properties.
The excellent oxidation resistance of ferritic stainless steels depends on the formation of Cr2O3 films on their surface. As elements diffuse through it slowly, this film acts as protective film and suppresses the oxidation of Fe.3) When the Cr2O3 film acts as protective film, the oxidation rates of steels are low and steels are not oxidized. This is called normal oxidation. However, when the Cr2O3 films break owing to any changes in circumstances, the steels lose the oxidation protective properties. Thus, Fe can be oxidized resulting in rapid formation of Fe oxides, and therefore, the oxidation mass gain increases. This is called abnormal oxidation. The transition temperature from normal to abnormal oxidation is known as the oxidation resistance limit temperature. This is largely affected by the Cr content of steels, but the effects of other elements, such as Si, cannot be ignored.4)
Ti is an active element which oxidizes easier than Cr, and simultaneously forms compounds with C, N, and S in steels. Therefore, considering the effect of Ti on the oxidation resistance of steels, the effect of the reactions between Ti and these elements cannot be neglected.
Moroishi et al.5) and Fujikawa et al.6) investigated the effect of C, N, Si, Zr, Nb, and Ti on the oxidation resistance of ferritic stainless steels, and concluded that Ti can improve their oxidation resistance. Moreover, they subjected 17.5Cr–0.5Si–0.5Mn steels which contained 0.4–0.6%Ti to oxidation at 1273 K for up to 900 ks, and proposed that C and N were detrimental elements for the oxidation resistance as they promoted the formation of γ phase at high temperatures. Ti addition stabilized C and N as carbides and nitrides, which suppressed the formation of the γ phase, and therefore improved oxidation resistance. However, they did not investigate the oxidation behavior of the steels in normal oxidation state, as their experiments were performed at 1273 K, where the samples presented abnormal oxidation state and Fe oxides formed on the surfaces of the samples.
By contrast, Okabe et al.7) reported that the addition of Ti decreased oxidation resistance as Ti addition increased the oxidation mass gain during the oxidation tests they performed in pure oxygen atmosphere using 18Cr–0.1Si–0.1Mn steels containing up to 0.9%Ti. However, the oxidation resistance limit temperature of the analyzed steels was 1473 K regardless of the amount of Ti added. They also investigated the microstructure of the samples in normal oxidation state and noted that Cr2O3 scale formed and Ti was present in outermost layer of the scale. Moreover, internal oxides formed. However, they only mentioned that Ti addition likely decreased oxidation resistance by increasing the pores of the scale.
However, many questions on the behavior of Ti during oxidation have not been answered yet. Therefore, the effect of Ti addition on oxidation behavior has not been comprehensively elucidated. As stainless steel consists of high amounts of various elements, which present complex interactions, it is difficult to clearly understand the behavior of each element during oxidation. Moreover, practical stainless steel comprises more elements, and therefore, it is difficult to distinguish the effect of each element on oxidation resistance. In particular, the effects of the added elements on the scale and scale/metal interface have not been understood in detail.
Thus, in the present study, the oxidation behavior of high-purity 19%Cr ferritic stainless steels without and with Ti was investigated to discuss the effect of Ti addition. In particular, we focused on investigating the structures of the scale and scale/metal interface.
Several 50 kg ingots of high-purity 19%Cr ferritic stainless steel which contained approximately 50 mass ppm C (mass ppm is hereafter abbreviated as ppm), 100 ppm N, and 0 to 0.5 mass% (mass% is hereafter abbreviated as %) Ti were melted in a vacuum induction furnace. The chemical compositions of the Ti-free, 0.15%Ti, and 0.30%Ti ingots, are summarized in Table 1. The 19Cr basic composition was adopted to eliminate the structural factors that could affect oxidation behavior, to avoid the formation of the γ phase. The range of Ti addition matched that of the practical Ti-containing high-purity stainless steel.
The obtained hot ingots were used as starting material, and were hot- and cold-rolled to a thickness of 2 mm followed by annealing in air at 1153 K for 60 s. The produced steel sheets were used as sample materials.
Oxidation test specimens which were 2 mm thick, 20 mm wide, and 20 mm long were machined from the steel sheets, were subsequently finished using emery paper up to #400, and were lastly degreased using acetone.
2.2 Oxidation test in airOxidation tests were conducted in still air in a tube furnace. The test temperature ranged from 1073 to 1273 K and the test time was 720 ks. Each oxidation test specimen was obliquely placed on a quartz rod holder and inserted into the furnace, which was maintained at the test temperature. Scale spallation from specimens on cooling after removal from the furnace was considered to accurately measure the oxidation mass gain. The specimens were held in the furnace for the predetermined test time, following which they were removed from the furnace, transferred one by one into a lidded container, and cooled. All spalled scale was thus recovered. One or two specimens were tested at each temperature and time.
2.3 Scale evaluation testsThe appearance of the samples was analyzed and the oxidation mass gain was measured using a balance. As the spalled scale was also recovered, the oxidation mass gain in this report included the spalled scale.
All specimens were analyzed using X-ray diffraction (XRD) utilizing Cu Kα radiation to identify the phases of the surface layer.
The surfaces and nearby cross sections of some specimens were observed using field emission-scanning electron microscopy (FE-SEM). Cross-section specimens were prepared using a cross-section polisher (CP). Elemental analysis and elemental mapping were also performed using energy dispersive X-ray spectroscopy (EDS).
Cross-section specimens embedded in resin were additionally prepared and their scale was structurally analyzed using scanning electron microscopy-electron backscatter diffraction (SEM-EBSD). The OIM-5 (TSL) software was used for EBSD measurement control and analysis.
Furthermore, the structures of the scale and regions near the scale/metal interface were observed in detail using field emission-transmission electron microscopy (FE-TEM). Thin-film specimens for TEM were prepared using a focused ion beam (FIB) miller. At that time, C was evaporated from the thin-film specimens in the FIB miller to protect the scale. Then, the prepared thin-film specimens were structurally observed using FE-TEM. Structural analysis using micro electron beam diffraction (µ-Diff) and element analysis using µ-EDS were performed on the observed phases to identify the substances contained therein. Semi-quantitative analysis was conducted on the constituent elements of the phases using their µ-EDS intensity profiles.
Figure 1 presents the appearances of the 19Cr, 19Cr–0.15Ti and 19Cr–0.3Ti steel specimens after 720 ks of oxidation. The Ti-free steel exhibited abnormal oxidation at 1173 K and above. Thus, the oxidation resistance limit temperature of the 19Cr steel was 1123 K. On the other hand, Ti-containing steels presented abnormal oxidation at 1223 K and above. Therefore, the oxidation resistance limit temperature was 1173 K, which was 50 K higher than that of the Ti-free steel. This increase in limit temperature indicated that Ti addition improved the oxidation resistance of the samples.
Appearance of 19Cr, 19Cr–0.15Ti, and 19Cr–0.3Ti steels oxidized in air at 1073–1273 K for 720 ks.
Figure 2 illustrates the oxidation mass gain of the specimens after 720 ks of oxidation relative to the test temperature. The oxidation mass gain of the 19Cr steel was small until it reached its oxidation resistance limit temperature of 1123 K. However, at a temperature higher than 1123 K the oxidation mass gain increased sharply during abnormal oxidation. Figure 2 also clearly indicates that the Ti-containing steels behaved similarly, but their limit temperature was 50 K higher than that of the Ti-free steel. In addition, at the normal oxidation state, the oxidation mass gains of the Ti-added steels were higher than those of the Ti-free steel.
Temperature dependence of mass gain of 19Cr, 19Cr–0.15Ti, and 19Cr–0.3Ti steels oxidized in air at 1073–1273 K for 720 ks.
Table 2 summarizes the results of the XRD analysis of the specimens oxidized at 1123 K. Two phases were identified for the 19Cr steel: α-Fe(Cr) as metal and Cr2O3 as scale. In addition to α-Fe(Cr) and Cr2O3, the presence of rutile TiO2 was confirmed in the XRD spectra of the Ti-containing steels, such as the 19–0.15Ti and 19Cr–0.3Ti steels.
To observe the microstructure of the scale and scale/metal interface in detail, the specimens that underwent normal oxidation at 1123 K, were selected for FE-SEM observations. Figure 3 depicts the surface observation results. Grains in the µm size range were observed for the 19Cr steel. The XRD results suggested that these grains consisted of Cr2O3. Most of these grains detected in the structure of the 19Cr–0.15 Ti steel were smaller than those of the 19Cr steel, but some hexagonal prismatic grains were large, and reached 10 µm in size. Most of the grains identified in the structure of the 19Cr–0.3Ti steel were much smaller, while some hexagonal columnar grains were larger and reached 10–20 µm in size.
Surface FE-SEM images of 19Cr, 19Cr–0.15Ti, and 19Cr–0.3Ti steels oxidized in air at 1123 K for 720 ks.
Figure 4 illustrates the results of the EDS analysis for the 19Cr–0.15Ti steel. The small and large grains (Nos. 041 and 039, respectively) were identified to be Cr2O3. However, the presence of Ti was only detected in the small grains, not the large grains. The results for grain No. 040 were omitted because they were identical to those of grain No. 39.
Surface FE-SEM image and EDS results of 19Cr–0.15Ti steel oxidized in air at 1123 K for 720 ks.
Figure 5(a) and (b) presents the cross-section SEM images of the 19Cr and 19Cr–0.15Ti steel specimens, which were oxidized at 1123 K for 720 ks, respectively. The scale of the 19Cr steel was uneven and its thickness was approximately 2 µm. In addition, an internal oxidation zone was uniformly formed from beneath the scale to approximately 10 µm in depth. On the other hand, 3–4 µm thick uniform scale was formed for the 19Cr–0.15Ti steel, and some parts of the scale were projected. The FE-SEM surface observations suggested that the hexagonal columnar Cr2O3 grains represented the projected part. An internal oxidation zone, which was approximately 10 µm thick, was observed underneath the scale. The internal oxidation zone was slightly shallower than that of the 19Cr steel. Furthermore, the particles on the left side of the figure were suggested to be Ti(C, N). Figures 6(a) and (b) illustrate the back scattered electron (BSE) images and the EDS characteristic images of the 19Cr and 19Cr–0.15Ti steels, respectively. The scale of the 19 Cr steel is known to consist of Cr2O3, owing to the accumulation of Cr and O on the surface of the steel. Figure 6 also clearly indicates that the internal oxidation phase consists of Al oxides. On the other hand, the scale of the 19Cr–0.15Ti steel consisted of a Cr2O3 layer (similar to that of the 19Cr steel), which was attributed to the accumulation of Cr and O. In addition, the accumulation of Ti on top of the Cr2O3 scale was observed, which suggested that Ti oxides, such as TiO2 were formed. Also, the accumulation of Al and Ti suggested that their oxides were present in the internal oxidation zone.
Cross-sectional FE-SEM images of (a) 19Cr and (b) 19Cr–0.15Ti steels oxidized in air at 1123 K for 720 ks.
Cross-sectional FE-SEM and EDS images of (a) 19Cr and (b) 19Cr–0.15Ti steels oxidized in air at 1123 K for 720 ks.
Figure 7 illustrates the cross-sectional EBSD maps of the scale of the 19Cr and 19Cr–0.15Ti steel specimens after oxidation at 1123 K for 720 ks. The SEM images and inverse pole figures (IPFs) of various phases in the thickness direction are also depicted. The scales of both steels consisted of Cr2O3 layers. The EBSD maps enabled us to elucidate the shapes and sizes of the grains in the IPF images. Comparing Figs. 7(a) and (b), it was determined that the grain sizes of the Cr2O3 layers of the two steels were markedly different. The Cr2O3 grains of the 19Cr steel were approximately 1–2 µm in size, which was equivalent to the thickness of the scale, while the Cr2O3 grains of the 19Cr–0.15Ti steel were much smaller. It is likely that Ti addition refined the grain size of the Cr2O3 layer.
Cross-sectional IPF maps of Cr2O3 and αFe(Cr) of (a) 19Cr and (b) 19Cr–0.15Ti steel samples oxidized in air at 1123 K for 720 ks.
To examine the scale and scale/metal interface in greater detail, the cross sections of the specimens subjected to oxidation at 1123 K and 720 ks were analyzed. Figures 8(a) and (b) illustrate the bright field FE-SEM images and elemental EDS mappings of the 19Cr and 19Cr–0.15Ti steels, respectively.
Cross-sectional FE-TEM and EDS images of (a) 19Cr and (b) 19Cr–0.15Ti steels, oxidized in air at 1123 K for 720 ks.
The scale of the 19Cr steel mainly consisted of columnar grains 2 µm in depth and 1 µm in width. The results of the EDS mapping, XRD analysis and so on suggested that the scale only consisted of Cr2O3. Thin Fe–Cr oxide films, which were described by Mishima8) were not observed. Thus, the oxidation of Fe was considered to be suppressed for these specimens. The internal oxidized particles consisted of Al oxides.
The Cr2O3 grains of the 19Cr–0.15Ti steel were finer than those of the 19Cr steel. Also, the scale of the 19Cr–0.15Ti steel was slightly thicker than that of the 19Cr steel. In addition, Fig. 8 indicates that Ti accumulated at the surface of the scale, and the internal oxidized particles of the 19Cr–0.15Ti steel were finer and denser than those of the 19Cr steel. Moreover, Al oxides formed in regions deeper than those where Ti oxides formed.
Figure 9 illustrates the phases of the 19Cr steel, identified using µ-Diff and µ-EDS. As presented in Fig. 9(I), the scale consisted of Cr2O3 and its impurities were really small. As illustrated in Fig. 9(II), the internal oxidation phase consisted of θ-Al2O3. In our previously published papers,9,10) we mentioned that when Nb-containing ferritic stainless steels which also contained small amounts of Al were oxidized in air, the internal Al-containing oxidized phase consisted of θ-Al2O3, not α-Al2O3. Those results were in agreement with the findings of this study where we used Nb-free steels.
Cross-sectional FE-TEM images of 19Cr steel oxidized in air at 1123 K for 720 ks: (I) scale and (II) internal oxide region. Sets of µ-diffraction patterns and EDS spectra of the spots indicated in the figures are also illustrated.
Figure 10 depicts the results of identified phases of the 19Cr–0.15Ti steel. The scale mainly consisted of Cr2O3 along with very small amounts of impurities (Fig. 10(I)). However, Cr2O3 grains which contained Ti, and TiO2 grains were observed in the surface region of the scale (Fig. 10(II)). These results suggested the outward diffusion of Ti. As illustrated in Fig. 10(III), Al2TiO5 and Al2Ti7O15, which are Ti–Al complex oxides, were detected in the metal beneath the scale. Furthermore, as presented in Fig. 10(IV), θ-Al2O3 was identified in the deeper region of the metal, similarly to the 19Cr steel.
Cross-sectional FE-TEM images of 19Cr–0.15Ti steels oxidized in air at 1123 K for 720 ks: (I) and (II) scale, and (III) and (IV) internal oxide regions. Sets of µ-diffraction patterns and EDS spectra of the spots indicated in the figures are also included.
The results of the SEM, EBSD, and TEM tests indicated that Ti addition caused the Cr2O3 scale to grow thicker and the grains of the scale to become refined. The representative TEM images presented in Fig. 11 indicate that Ti addition refined the grains of the Cr2O3 scale.
TEM images illustrating Cr2O3 grains in scale of (a) 19Cr and (b) 19Cr–0.15Ti steels oxidized in air at 1123 K for 720 ks.
The µ-EDS analysis results depicted in Figs. 9 and 10 confirmed that Ti accumulated in the outermost region in the scale where it dissolved into Cr2O3 grains and/or formed TiO2 grains. However, the status of Ti in the rest of the scale has not been elucidated yet. Therefore, we examined the status of Ti in the scale except for its outermost region. Figure 12 presents the results of µ-EDS analysis of grains and scale grain boundaries near the scale/metal interface. The presence of Ti could not be confirmed at the analyzed points of the grains (Figs. 12(b) and (c)). However, Ti was identified at grain boundaries where the average amount of Ti was approximately 2.4%. Therefore, Ti existed at the grain boundary in dense and it was suggested to diffuse outward through Cr2O3 grain boundaries.
TEM images and EDS maps illustrating Ti content at Cr2O3 grain boundaries. Points (a), (d), (e) are located at grain boundaries and points (b) and (c) are located inside grains.
Figure 13 depicts the schematic scale structure changes caused by the addition of Ti. For the 19Cr steel (Fig. 13(a)), the scale consisted of Cr2O3 and the size of grains of the scale was similar to the scale thickness. Moreover, the internal oxidation phase consisted of θ-Al2O3. However, for the 19Cr–0.15Ti steel (Fig. 13(b)), the scale also consisted of Cr2O3, but small amounts of TiO2 formed above it. Ti addition refined the grains of the Cr2O3 scale. Furthermore, Al2TiO5 and Al2Ti7O15, which are Ti–Al complex oxides, formed in the region near the scale/metal interface, while θ-Al2O3 only formed in the deeper region of the internal oxidation phase.
Schematic illustration of scale structure of (a) 19Cr and (b) 19Cr–0.15Ti steels oxidized in air at 1123 K for 720 ks.
The scale of Ti-containing steel grew thicker than that of Ti-free steel for the normal oxidation state.
According to the above-mentioned results, Ti reached the surface of the alloy and was either dissolved into Cr2O3 and/or formed TiO2. Naudomidis et al.11) reported the formation of a solid solution of Ti in Cr2O3, and mentioned that approximately 18% of the cation sites of Cr2O3 could be replaced by Ti at 1273 K. Although no data existed on the diffusion rate of Ti in Cr2O3, Negelsberg12) reported that Ti could pass through Cr2O3 into the oxidizing atmosphere at high temperature. Therefore, it was considered that Ti could diffuse into Cr2O3 and its diffusing path followed the grain boundaries.
Also, it was suggested that grain-refining owing to Ti addition was due to the presence of Ti at Cr2O3 grain boundaries interrupting the transportation of Cr2O3 grain boundaries in lateral direction. The same mechanism was reported in one of our previously published papers9) for Mn addition. The thickening of the Cr2O3 scale by the addition of Ti was attributed to the following mechanism. The grain refinement of the Cr2O3 scale by Ti addition increased grain boundary diffusion, and therefore, the growth rate of the Cr2O3 scale or the oxidation rate increased. For the Ti-containing steels, Ti was oxidized and diffused through the grain boundaries of the Cr2O3 scale, which caused the grain refinement of the Cr2O3 scale. Consequently, the diffusion path increased, which increased the oxidation rate.
4.2 Effect of Ti addition on internal oxidationAs illustrated in Fig. 13, the internal oxidation phase of the 19Cr steel consisted only of θ-Al203. On the other hand, for the Ti-containing steels, Al2TiO5 and/or Al2Ti7O15 were formed in the region near the scale/metal interface and θ-Al2O3 was formed in the deeper regions. In addition, TiO2 was not detected. Ellingham diagrams indicate that it is difficult to oxidize Cr, Ti, and Al in order, and thus Cr2O3, TiO2 and Al2O3 formed in order from the surface toward the bulk of the sample. TiO2 and Al2O3 are soluble in each other, and therefore, Ti–Al complex oxides such as Al2TiO5 and Al2Ti7O15 formed beneath the scale/metal interface. In addition, only θ-Al2O3 formed in the deeper region, where TiO2 was unable to form.
4.3 Effect of Ti addition on oxidation behaviorTi addition increased the oxidation mass gain but increased the oxidation resistance limit temperature (Figs. 1 and 2). Thus, Ti addition improved the oxidation resistance of the steels. One of the reasons for Ti addition increasing the oxidation mass gain was the oxidation of Ti itself, as Ti can oxidize easier than Cr. Ti was dissolved into Cr2O3 and/or formed TiO2 at the surface region of the scale. In addition, Ti oxides formed as internal oxidation phase in the region near the scale/metal interface. Thus, Ti addition caused oxidation mass gain. Moreover, Ti addition caused Cr2O3 grain-refinement and the thickening of the Cr2O3 scale. Similar findings have been reported by Okabe et al.,7) except for the thickening of the Cr2O3 scale. However, these results suggested that Ti itself did not cause the oxidation mass gain to increase significantly. Therefore, the thickening of the Cr2O3 layer was considered to be the main reason for the increase in mass gain.
The increase in oxidation mass gain is usually the reason for the decrease in oxidation resistance. However, the addition of Ti caused the oxidation resistance limit temperature to increase in spite of the increase in oxidation mass gain. As Moroishi et al.5) and Fujikawa et al.6) reported, the suppression of the formation of the γ phase was likely to make a roll. However, the 19Cr steel used in this study contained ultra-low amounts of C and N, and thus, the effect of the γ phase was considered to be limited.
Ike et al.13) suggested that the small amounts of Al in steel could exhibit oxygen-getter effect. This involved the consumption of O to internally oxidize Al, which suppressed the oxidation of Fe, when the inward diffusion of O was small. Ti also exhibited oxygen-getter effect, similarly to Al. Ti was oxidized and consumed O, which suppressed the increase in oxygen partial pressure ($P_{O_{2}}$) at the scale/metal interface and the oxidation of Fe.
We investigated the effect of Ti addition on the air oxidation of high-purity ferritic stainless steels by focusing on the microstructure changes of the scale and scale/metal interface. The following conclusions were drawn.
