Regular Article
Effect of Al Addition on Thermal Fatigue Deformation Morphology in High Heat-resistant Ferritic Stainless Steel SUS444
2024 Volume 64 Issue 12 Pages 1838-1846
Details
2024 Volume 64 Issue 12 Pages 1838-1846
Ferritic stainless steels are used for automobile exhaust parts because of their high heat and corrosion resistance. Among them, parts located upstream near the engine, so-called hot-end parts, for example, exhaust manifolds, are required to show excellent heat resistance. Since thermal fatigue is induced by repeating heating and cooling with mechanical strain restriction, thermal fatigue resistance is one of the most important properties of upstream exhaust-parts materials.
In this study, the effect of Al addition on thermal fatigue deformation morphology was investigated for high heat-resistant ferritic stainless steel SUS444 which has been used for automobile exhaust parts. In contrast with the steel without Al addition, which fracture morphology in thermal fatigue under the maximum temperature of 1173 K was necking, cracking was predominant in the steel with Al addition without necking. Al addition has the effect to prevent necking in thermal fatigue under the maximum temperature of 1173 K due to solid solution strengthening by Al.
Ferritic stainless steel has excellent corrosion and heat resistance, and it is a resource-saving material compared to austenitic stainless steel because it does not contain the rare metal Ni. While austenitic stainless steel has an advantage in terms of high-temperature strength (yield strength), ferritic stainless steel with the BCC (body-centered cubic) crystal structure has an advantage in the coefficient of thermal expansion being approximately 30% lower than that in austenitic stainless steel with the FCC (face-centered cubic) structure.1,2,3,4,5) As to an automobile exhaust pipe, thermal fatigue is a type of fracture that occurs when a material connected at both ends is repeatedly heated and cooled by exhaust gas, and then thermal expansion and contraction continue to be restricted. Since thermal fatigue is caused by strain accumulation due to the restricted thermal expansion of the material, a low coefficient of thermal expansion, in addition to the high yield strength of the material (especially the high-temperature yield strength), is effective in improving the thermal fatigue resistance. Therefore, this is an advantageous property for ferritic stainless steel compared to austenitic stainless steel.5,6,7,8) Until now, Type 429 stainless steel (14%Cr-0.4%Nb), which has high-temperature resistance by adding Nb with a solid solution strengthening effect, has been used as a type of ferritic stainless steel with high heat resistance.9,10) Similarly, SUS444 (18%Cr-0.5%Nb-2%Mo) is further strengthened with Mo, a solid solution strengthening element, to further increase its high-temperature yield strength.11,12,13) Recently, stainless steel using precipitation strengthening of Cu instead of Mo, a rare metal, has been developed and put into practical use.14,15,16)
The authors found that for SUS444, the most heat-resistant of the ferritic stainless steels, the failure mode in thermal fatigue tests differed between maximum temperatures below 973 K and above 1073 K and that necking deformation was more pronounced at higher temperatures.17) Takushima et al. reported that the addition of 2% Mo suppressed dynamic recovery during the thermal fatigue process and improved the thermal fatigue resistance by about 100 K.18) Hamada et al. investigated the effect of grain size on thermal fatigue resistance. They reported that the finer the grain, the smaller the amount of plastic strain and the longer the fatigue lifetime.19) Furthermore, Ota et al. reported that when the maximum temperature was 1073 K, the necking resistance was improved by adding a small amount of Al.20) However, they pointed out that although the necking was suppressed by adding a small amount of Al, necking eventually occurred and led to failure, so there is room for further improvement. Therefore, in this study, to clarify the effect of Al addition on improving necking resistance, we increased the maximum temperature to 1173 K and investigated the effect of Al addition on thermal fatigue resistance under test conditions where necking is likely to occur.
The typical components of Al-free steel (Steel A) and 3 mass% Al-added steel (Steel B) are shown in Table 1. Here, Si was added in small amounts to increase oxidation resistance and reduce the effects of oxidation. Ti was added in small amounts to form preferentially carbonitrides and effectively utilize the solid solution strengthening of Nb. In addition, to clarify the effect, the amount of Al added was set at 3 mass%, which is higher than the 0.5 mass% used by Ota et al. Figure 1 shows the manufacturing method for both test materials. The test material was a 50 kg research steel ingot made by small-scale vacuum melting, heated at 1443 K for 3.6 ks, and then hot-rolled into a sheet bar with a thickness of 35 mm (assuming rough hot rolling). The sheet bar was then reheated at 1323 K for 1.8 ks and then hot-rolled to produce a 5 mm thick hot-rolled sheet (assuming finish hot rolling). The hot-rolled sheets were annealed at 1313 K for 60 s, then cold-rolled to a thickness of 2.0 mm and annealed at 1313 K for 60 s. At this time, the crystal structure of both Steel A and Steel B was a single ferrite phase with a grain size of approximately 50 μm. High-temperature tensile test specimens with a parallel section width of 10 mm and gauge distance of 50 mm were prepared from the obtained cold-rolled and annealed sheets by cutting and grinding. The surface finish of the end surface was set to Ra, approximately 0.8. In addition, a small piece of 10 mm × 50 mm was cut from the same cold-rolled and annealed sheet and subjected to aging treatment at 1173 K in the atmosphere for 30 ks (maximum holding temperature in a thermal fatigue test, equivalent to 1000 cycles). The oxidation scale on all six surfaces was removed by polishing with emery paper and used to analyze the precipitate residue.
| Steel No. | C | Si | Mn | P | S | Cr | Al | Mo | Nb | Ti | N |
|---|---|---|---|---|---|---|---|---|---|---|---|
| A | 0.0058 | 0.82 | 0.16 | 0.022 | 0.0007 | 18.5 | 0.02 | 1.82 | 0.52 | 0.19 | 0.0065 |
| B | 0.0052 | 0.86 | 0.15 | 0.023 | 0.0011 | 18.4 | 2.91 | 1.79 | 0.52 | 0.21 | 0.0068 |
In addition, the sheet bar produced above was reheated at 1373 K for 3.6 ks and forged into a 30 mm square bar (assuming finish hot rolling). After cutting this square bar to a length of 150 mm, it was annealed at 1313 K for 300 s, followed by accelerated air cooling (cooling rate: approximately 50 K/s). At this time, the crystal structure of both Steel A and Steel B was a single-phase ferrite with a grain size of about 100 μm. Round bar specimens of the shape shown in Fig. 2 were fabricated by cutting from the square bars and subjected to thermal fatigue tests. The surface finish was Ra, about 0.8.
Automobile exhaust pipes have a pipe shape or a shell shape made by stacking press-formed steel plates, but in thermal fatigue tests using pipe-shaped test pieces, the test piece shape (roundness, welded part shape, etc.) makes it difficult to evaluate the thermal fatigue properties of the material itself. Oku et al. performed thermal fatigue tests using both solid round bar specimens and pipe-shaped specimens and found that most of the test fracture surfaces using solid round bar specimens exhibited thermal fatigue-like failure.21) In addition, in a previous report by the authors,17) beach marks due to fatigue were observed on the fracture surface of the crack after thermal fatigue tests on solid round bar specimens, which means thermal fatigue failure in a broad sense occurred in solid round bar specimens. For these reasons, thermal fatigue tests were conducted using solid round bar test pieces in this study as well to exclude the influence of the test piece shape.
2.2. High-temperature Tensile TestHigh-temperature tensile tests were conducted at 973 K, 1073 K, and 1173 K. The temperature was increased by approximately 0.5 K/s, and after reaching the test temperature, the test was started after holding for 900 s. The crosshead speed was 0.2 mm/min up to 0.2% proof stress and 5 mm/min after 0.2% proof stress.
2.3. Thermal Fatigue TestThe thermal fatigue test was conducted with an Instron model 8861 thermal fatigue testing machine under the conditions shown in Fig. 3. The minimum test temperature Tmin was 473 K, and the maximum Tmax was 1173 K. The heating rate was 4 K/s, the cooling rate was 2 K/s, and the holding time at the maximum and minimum temperatures was 30 s each. First, in order to measure the amount of free thermal expansion, the temperature was increased from 473 K to 1173 K at 4 K/s without stress, and immediately after reaching 1173 K, it decreased to 473 K at 2 K/s. This is repeated for 3 cycles, and the coefficient of thermal expansion is calculated based on the amount of thermal expansion in the 3rd cycle when the behavior becomes stable and the amount of strain (modal strain) to be applied during the thermal fatigue test is calculated by multiplying the restraint factor η was calculated. In this study, the constraint ratio η was set to 0.5. The starting temperature for the test was intermediate (=823 K) between the maximum temperature Tmax (=1173 K) and the minimum temperature Tmin (=473 K). During the test, the maximum tensile stress at 473 K was recorded for each cycle, and the test was continued until the number of cycles at which the maximum tensile stress decreased to 50% with respect to the initial (5th cycle) value. The repeated number of cycles at which it decreased to 75% was evaluated as the number to failure (fatigue lifetime).22) In order to observe the deformation behavior of the test piece during the test, fixed-point photography was taken at the soaked parallel part of the test piece, and the area of the most necked part was measured as the cross-sectional area.
The precipitates were extracted by constant current electrolysis in 10 vol% acetylacetone and 1 mass% tetramethylammonium chloride-methanol using an electrolyzer SPEED ANALYZER FV-128 manufactured by Fujiwara Seisakusho. A cellulose acetate membrane filter (pore size 0.2 μm, 47 mm diameter) was used to filter the extraction residue. The collected residue was placed in a platinum crucible with the filter and incinerated at 853 K. 0.75 g Na2O2 + 0.75 g LiBO4 was added and melted with a gas burner. The melt was dissolved by adding 25 mL of 0.8 mass% tartaric acid and 10 vol% sulfuric acid; the volume was brought to 100 mL with pure water. Subsequently, the precipitated Nb, Mo, and Al amounts were determined using an ICP emission spectrometer.23)
Figure 4 shows the stress-strain curve obtained from the high-temperature tensile test. At this time, the strain values in Figs. 4(a), 4(c), and 4(e) showing the entire curve were obtained from the amount of displacement of the crosshead. The initial strain value in the enlarged curve described in Figs. 4(b), 4(d), and 4(f) was determined from the displacement of the extensometer. The entire curve was calculated from the displacement of the crosshead because the displacement of the extensometer often became discontinuous due to the extensometer slipping during the test. In the 973 K test, work hardening was observed in the early stage of the test, and stress discontinuity (strain rate dependence) was not observed when the strain rate was changed around 0.6% elongation in both steels, indicating intermediate temperature deformation. On the other hand, in the tests at 1073 K and 1173 K, work hardening was not almost observed, and stress discontinuities were clearly seen at the strain rate change points, indicating high-temperature deformation exhibiting strain rate dependence. Table 2 shows the yield strength (0.2%PS), ultimate tensile strength (UTS), and elongation (El) obtained from the high-temperature tensile test. At 973 K and 1073 K, Steel B with Al had higher 0.2%PS and UTS and lower elongation than those of Steel A without Al. On the other hand, at 1173 K, the 0.2%PS and UTS of Steel B were higher than those of Steel A, and the elongation of Steel B was more than three times greater.
| 973 K | 1073 K | 1173 K | |||||||
|---|---|---|---|---|---|---|---|---|---|
| 0.2%PS (MPa) | UTS (MPa) | EL (%) | 0.2%PS (MPa) | UTS (MPa) | EL (%) | 0.2%PS (MPa) | UTS (MPa) | EL (%) | |
| Steel A | 136 | 210 | 12 | 35 | 67 | 32 | 21 | 39 | 45 |
| Steel B | 243 | 332 | 6 | 115 | 162 | 15 | 25 | 57 | 148 |
Figure 5 shows the maximum tensile stress change during the thermal fatigue test for both steels. Steel B with Al had higher stress and a significantly longer lifetime (2296 cycles) than Steel A without Al (847 cycles).
Figure 6 shows the change in the cross-sectional area of the most necked part of the specimen during the thermal fatigue test. Necking occurred, and the cross-sectional area decreased after 100 cycles in Steel A, whereas there was almost no decrease in the cross-sectional area in Steel B.
Figure 7 shows the optical microscope photograph of the longitudinal section of the parallel section in the specimen after the thermal fatigue test. In Steel A, the necking of the specimen was clear, whereas in Steel B, necking was not observed, and cracks occurred near the R portion at both ends of the parallel section.
Figure 8 shows the SEM photograph of the fracture surface generated by a crack (the area indicated by the arrow in Fig. 7(b)) after the thermal fatigue test of Al-added Steel B. Nevertheless, the unevenness was somewhat unclear by oxides formed on the fracture surface since cracks occurred during the test, beach marks could be seen all over the place, for example at the locations indicated by arrows in the photograph, meaning that they were caused by fatigue. In the thermal fatigue test of Steel A, cracks were not observed, so the fracture surface could not be observed.
Figure 9 shows the hysteresis curves obtained in the 5th cycle of the thermal fatigue test. During the temperature increase, the compressive stress reached its maximum at around 800 K for Steel A and around 900 K for Steel B, and then the compressive stress of both steels decreased and reached the same value of around 1100 K. The tensile stress of Steel B was higher than that of Steel A in the entire temperature range after being held at 1173 K. The total strain Δεt applied during a cycle expressed as the difference between the maximum strain and minimum hysteresis strain was 0.494% for Steel A and 0.551% for Steel B. This is because the thermal expansion coefficient measured before the thermal fatigue test for each steel was 13.99×10−6/K for Steel A and 15.47×10−6/K for Steel B. This means that Steel B has a longer lifetime than Steel A, even though the total strain Δεt is larger. In addition, in the hysteresis curve shown in Fig. 9(b), which shows the relationship between strain and stress, when the Young’s modulus of these steels was calculated in the region where the stress decreased linearly from the maximum tensile stress, it was 224 GPa for Steel A. In contrast, Steel B was slightly larger at 244 GPa. When the Young’s modulus became larger, the amount of plastic strain indicated by the difference in strain at two points where the stress value was 0 in the hysteresis curve shown in Fig. 9(b) becomes larger. Therefore, this is a disadvantageous factor from the viewpoint of thermal fatigue resistance in Steel B.
Table 3 shows the amounts of precipitated Nb, Mo, and Al before and after aging. The amounts of precipitated Nb and Mo in Steel A and Steel B were almost equal before and after aging. After aging, more than 80% of the added Nb, and about 40% of the added Mo were precipitated. Furthermore, in Steel B, the amount of Al precipitated after aging were very small at 0.007%, and it was thought that the contribution of Al addition to precipitation strengthening is negligible.
| Precipitates of Nb | Precipitates of Mo | Precipitates of Al | ||||
|---|---|---|---|---|---|---|
| Steel A | Steel B | Steel A | Steel B | Steel A | Steel B | |
| Before aging | 0.035 | 0.038 | 0.009 | 0.013 | 0.001 | 0.001 |
| After aging for 500 min | 0.438 | 0.446 | 0.757 | 0.767 | 0.002 | 0.007 |
The authors have already reported that the fracture morphology due to thermal fatigue in high heat-resistant ferritic stainless steel SUS444 can be broadly divided into two types, necking and crack propagation, and it changes mainly depending on temperature.17) Under the thermal fatigue test conditions conducted in this study, that is, the maximum test temperature was 1173 K with the restraint ratio of 0.5, necking was dominant with no cracks occurring in the test for Steel A without Al addition, and crack propagation was dominant without necking in Steel B with Al addition. The composition of Steel A is almost the same as the SUS444 used in the previous report,17) and the fatigue lifetime and necking behavior were also similar in thermal fatigue tests at a maximum test temperature of 1173 K with a restraint ratio of 0.5. On the other hand, in Steel B with Al addition, necking was not almost observed. Crack propagation was dominant in the test with a maximum temperature of 1173 K, similar to the previously reported case of SUS444 tested with a maximum test temperature of 973 K. Then, the fracture morphology in thermal fatigue tests is discussed below.
The following are the reasons why the fracture morphology of Steel A and Steel B differs.
4.1.1. Amount of Plastic StrainBecause necking is caused by plastic instability, the amount of plastic strain that occurs during the thermal cycle affects whether necking occurs. Even if the restraint ratio is the same at 0.5, the total amount of strain generated is different because the coefficient of thermal expansion of the steel is different. Figure 10 shows the total strain, elastic strain, and plastic strain of each steel during the thermal fatigue test. The slight decrease in the total strain and plastic strain of Steel B at the beginning of the test is thought to be due to instability at the beginning. For Steel A, the amount of plastic strain increased towards the end of the test due to the softening of the material itself as well as a decrease in cross-sectional area due to necking. On the other hand, in Steel B, plastic strain suddenly decreased towards the end of the test without increasing as in Steel A due to the constraint at the openings caused by microcracks on the surface of the test piece. This is thought to be due to relaxation, suggesting that the amount of plastic strain changes differently depending on the type of fracture (necking or cracking). As mentioned above, since the total strain was larger for Steel B, which has a higher coefficient of thermal expansion, and the elastic strain was also larger for Steel B, the plastic strain is larger for Steel B than Steel A despite a higher high-temperature yield strength. Therefore, whether necking occurs cannot be explained solely by the amount of plastic strain, and other factors must be considered.
In general, pure metals have different deformation mechanisms in the temperature range above and below half the melting point (0.5 × Tm (K)).24) Slip deformation becomes dominant at the lower temperature, and creep deformation due to atom diffusion becomes dominant at the higher temperature. In tensile tests at lower temperatures, after yielding beyond the elastic range, work hardening occurs as the strain increases. After reaching the maximum tensile strength, the deformation stress drops sharply, leading to fracture. On the other hand, at higher temperatures, no work hardening is observed because the dislocation substructure recovers, and the maximum tensile strength is reached with a relatively small strain after yielding. Then the deformation stress gradually decreases until fracture. Another difference is that in higher-temperature deformation, the dependence of stress on strain rate is remarkable. In deformation in the intermediate temperature range between the lower and higher temperatures (medium-temperature deformation), the strain rate dependence shows weak.25)
In high-temperature tensile tests at 973 K shown in Fig. 4, stress discontinuities due to changes in strain rate were not observed in both steels, and work hardening was noticeable up to about 5% strain. This is considered a so-called medium-temperature deformation. On the other hand, in tensile tests at 1073 K and 1173 K, discontinuities in deformation stress were observed in both steels due to changes in strain rate, and the deformation stress gradually decreased after reaching the maximum tensile strength without showing any work hardening and the elongation became relatively large (over 20%). This suggests that higher-temperature deformation occurs above 1073 K. Regarding elongation, the elongation was significantly larger at the higher temperature of 1173 K than at 1073 K, and in particular, Steel B was considerably larger by about 150%. The melting point Tm of ferritic stainless steel is approximately 1773 K, and 0.5×Tm becomes approximately 887 K. However, the transition temperature of the deformation mechanism varies depending on the material’s alloy composition, purity, strain rate, etc. In the case of this study, the experimental results show that the deformation mechanism transitions around 975 K (0.55 × Tm) for both steels. It is thought that there is no significant difference in the transition temperature between Steel A and Steel B. Okada et al. reported that this transition point ranges from 0.35 × Tm to 0.68 × Tm for various pure metals,24) and both steels were within this temperature range.
In a previous report,17) in thermal fatigue tests of SUS444 at 1073 K and 1173 K, in which necking was dominant, the compressive stress during heating decreased rapidly around 1073 K due to the formation of the dislocation substructure by recovery with the temperature rise. On the other hand, in the test at 973 K, since no decrease in compressive stress was observed when the temperature was increased, and an increase in the maximum tensile stress was observed at the beginning of the test, it was considered that the dominant factor was the initiation and propagation of cracks since the dislocation substructure was not recovered and work hardening occurred during the cycle. However, in this study, for Steel B, which does not cause necking and is dominated by crack initiation and propagation, a decrease in compressive stress at around 1100 K was observed, and no work hardening was observed in the initial stage of the test. Therefore, it is thought that the recovery of the dislocation substructure has occurred, and this hypothesis does not provide a sufficient explanation.
4.1.3. Relationship between Tensile Stress and Proof StressIn a previous report,17) necking in thermal fatigue tests did not occur when the maximum test temperature was 973 K but occurred only when the maximum test temperature was 1073 K or higher. Considering that necking is caused by tensile stress rather than compressive stress, looking at the temperature range where tensile stress is applied in the hysteresis curve, the entire temperature range below 1173 K in the case when the maximum test temperature is 1173 K and approximately 823 K or lower in the case when the maximum temperature was 973 K respectably. Based on this and the fact that the deformation mechanism transitions from 973 K to 1073 K and the yield strength rapidly decreases in the tensile test, it is inferred that necking occurs in the tensile stress loading range at temperatures higher than 973 K. In other words, in the previous report, the maximum test temperature was 973 K, so no necking occurred because no tensile stress was applied in the higher temperature range above 823 K.
Figure 11 shows a plot of the 0.2%PS value obtained from the high-temperature tensile test shown in Table 2 on the high-temperature part of the hysteresis curve described in Fig. 9. Considering that the strain rate during cooling in a thermal fatigue test was approximately 1.5 × 10−5/s, which is slightly slower than that during yield strength measurement in a high-temperature tensile test at approximately 5.6 × 10−5/s, tensile stress was applied at 1073 K almost equal to the proof stress in Steel A. It is thought that plastic strain occurred at this time, leading to necking progress. At 1073 K, no work hardening occurred, and necking progressed along with creep deformation accompanied by atomic diffusion. Therefore, once necking occurs, the true stress increases in the area where the cross-sectional area becomes smaller, and then the deformation progresses significantly. On the other hand, the stress generated by the thermal cycle was almost equal to that of 0.2%PS only at a maximum temperature of 1173 K in the test of Steel B, but 0.2%PS was higher than the tensile stress applied in the temperature range below 1173 K, therefore necking rarely occurred. Furthermore, as stated by Ota et al.,20) the addition of Al suppressed micro-necking, which prevented deformation from progressing due to local stress concentration, and then contributed to a longer lifetime.
Looking at the stress-strain curve in Fig. 4, Steel B, which has a high maximum tensile strength at 1173 K, has a slower stress drop after reaching the maximum tensile strength than Steel A. This means the overall uniform deformation occurs rather than the local plastic deformation (necking). This is also thought to be the reason for the increased elongation of Steel B.
Figure 12 schematically shows the relationship between lifetime governing factors and lifetime in thermal fatigue tests. As indicated by the arrow in the figure, when the test cycle progresses (progresses in the horizontal axis direction) at the maximum test temperature on the vertical axis (1173 K in the figure) and reaches the curve shown by the solid line (crack rupture limit line), lifetime is decided due to cracks (lifetime), and when the test reaches the curve shown by the dotted line (necking limit line), necking becomes dominant and the lifetime is decided. In general, necking occurs in a relatively short time at higher temperatures and leads to fracture, whereas necking is less likely to occur as the temperature drops. On the other hand, crack initiation and propagation have less temperature dependence than necking. The crack rupture limit line and the necking limit line intersect, and necking becomes dominant in the higher temperature range, while crack propagation dominates in the lower temperature range. That is, in Steel A, as shown in the previous report and in Fig. 12(a) of this study, necking was the dominant factor at 1173 K and 1073 K, and cracking was the dominant factor at 973 K. On the other hand, in the Al-added Steel B in this study (Fig. 12(b)), necking was suppressed by the addition of Al, leading the necking limit line to shift to the higher cycle side and making it less dominant even in the high-temperature range. The factor becomes crack, and it is thought that the lifetime was longer than that of Steel A, which factor is necking. The reason why necking is suppressed by the addition of Al is that solid solution strengthening increases the deformation stress at local necking (micro-necking) areas, preventing stress concentration and causing uniform deformation. This effect can also be obtained with Nb and Mo, which are well known as solid solution strengthening elements, but as seen in Table 3, Nb and Mo precipitate as intermetallic compounds (Laves phase) during thermal cycles, and the amount of solubility decreases, makes the effect becomes smaller compared to the initial stage of the test. On the other hand, Al hardly precipitates during the thermal cycle and remains a solid solution in the steel, so its effects can be obtained even after a long-term thermal cycle. Addition of Al is thought to be effective against crack initiation, but its effect needs to be clarified and a subject for future study.
As shown in Table 2, Steel B with Al significantly increased 0.2%PS in the temperature range from 873 K to 1173 K compared to Steel A without Al added. In addition, as shown in Table 3, there was no difference in the amount of Nb precipitation and Mo precipitation due to the addition of Al. There was almost no precipitation of Al itself, so even when exposed to high temperatures, Al remains solid solution in the steel, leading to contribution as a solid solution strengthening element. Until now, few reports have reported that Al functions as a solid solution strengthening element for ferritic heat-resistant stainless steels, although Kimura et al. have investigated it at room temperature.26) Most respondents recognized Al as an element that improves their oxidation resistance at high temperatures. However, the atomic size of Al is 0.143 nm which is similar to Nb9) (0.146 nm) and Mo27) (0.136 nm) which are well-known as solid solution strengthening elements having a significant difference from Fe (0.125 nm).20,26,28,29) Also, the diffusion rates are similar (Nb: 3.0×10−14 m2/s, Al: 1.3×10−14 m2/s at 1173 K),30) so it is recommended as a solid solution strengthening element. However, on the other hand, Al has a smaller atomic weight than Nb and Mo, and its atomic vibrations become more significant at high temperatures, so it may behave differently in inhibiting dislocation movement. In particular, the strain rate of thermal fatigue is relatively low, on the order of 10−5/s, and it is necessary to consider the possibility that the effect may be stronger or weaker depending on the operating temperature and strain rate. This is an issue for future study.
To elucidate the effect of Al addition on the thermal fatigue deformation morphology of ferritic stainless steel SUS444, thermal fatigue tests were conducted at a maximum temperature of 1173 K with a restraint ratio of 0.5 using SUS444 and a steel to which 3 mass% of Al was added. As a result, the following findings were obtained.
(1) In the case of Al-free steel, the thermal fatigue deformation was of the high-temperature type (creep deformation). In contrast, it was closer to the intermediate-temperature type (slip deformation), and the lifetime was significantly increased in the Al-added steel.
(2) Al-added steel has higher yield strength at high temperatures than Al-free steel, and the occurrence of plastic deformation (necking) in thermal fatigue tests was suppressed.
(3) Even with the addition of Al, there is no apparent difference in the Nb, Mo, and Al contents in the precipitates, indicating that Al is basically solid solution in the steel and has the effect of strengthening the steel by solid solution strengthening, suppressing thermal fatigue deformation, necking.
