Damage evaluation of the austenitic heat-resistance steel subjected to creep by using Kikuchi pattern parameters

1 Introduction

To improve the thermal efficiency and reduce the CO₂ emission, many countries are adopting the high-capacity ultra-supercritical (USC) power plants. By increasing the steam temperature and pressure entering into the turbines, the efficiency can be significantly enhanced, which reduces the fuel consumption and harmful gas emission [1,2,3]. Compared with traditional power plants, the heat efficiency of USC power plants had increased to about 47%, which could also lead to a 20–25% reduction in the carbon dioxide emission. In the meantime, new-grade heat-resistant steels with higher creep strength and oxidation resistance are required to satisfy such severe service conditions [4,5].

Super304H steel is a novel type of austenitic heat resistance steel, which is being widely employed in USC power plants as the super-heater and re-heater tubes, owing to its superior high-temperature strength, good oxidation resistance, and steam corrosion resistance [6,7]. However, these steel tube subjected to the creep stress at a high temperature, during their long-term service process, creep damage will inevitably happen over these tube components, which impair the microstructure stability and mechanical properties of Super304H steel [8,9]. To avoid the premature failure of these critical tube components, more research work should be devoted to the creep damage evaluation of Super304H steel.

The electron backscatter diffraction (EBSD) technique is a novel method emerged in recent years, which has been recognized as a promising approach to characterize the micro-structural damage in metallic materials [10,11,12]. Physically, by using the electron beam scanning, the Kikuchi pattern and lattice orientation data of sample can be acquired, and the deformation/damage of materials can be evaluated at the sub-micro level [13,14,15]. Recently, Guo et al. [16] used the band contrast (BC) maps to investigate the twin transmission process in Mg alloys. The results indicated that the evolution of twin can be observed in the corresponding BC images, so that the development or obstruction of twin chain accompanied by the crystal lattice distortion can be revealed; Gupta et al. [17] used the BC value to measure the effect of environment on fatigue behavior in Al–Cu–Mg alloy 2024. They suggested that the BC could be directly related to the perfection of the crystal lattice. The BC map results obtained by them revealed that there is a plastic-deformed region in the vicinity of crack-wake, which demonstrates that the BC value could be used to measure the plastic zone size. From above, the capacity and effectiveness of Kikuchi pattern parameters can be fully recognized. However, the research about the utilization of these parameters upon creep damage evaluation of heat-resistance steel is still rare. In this study, the Super304H steel was selected as the research object, and the creep damage process of this material was tried to investigate by using the EBSD technique with Kikuchi pattern parameters. The purpose is to provide the valuable information for the evaluation of creep damage of this austenitic heat-resistance steel.

2 Experiments

The typical Super304H austenitic heat-resistance steel was used in this study, with the chemical composition of C: 0.10%, Mn: 0.90%, Si: 0.31%, Cr: 18.78%, Ni: 9.74%, Cu: 3.25%, N: 0.052%, Mo: 0.36%, Nb: 0.49%, and Fe: Bal. A set of creep specimens were cut from the Super304H tube (see Figure 1(a) and (b)) and machined into the rod-shaped specimen with the dimension that are shown in Figure 1(c). The creep rupture experiment was first performed on the type-N92 creep test machine, which was carried out at 650°C with applied stress of 210 MPa. Then, the interrupted creep experiments were performed at 510, 960, 1,400, 1,465, and 1,530 h to study the damage evaluation of Super304H steel (see Figure 1(e)). The load error of this equipment is not more than ±1%, the temperature control error is not more than ±3°C, and the temperature gradient is less than 3°C. For EBSD measurements, a Symmetry^® EBSD detector with 50 nm spatial resolution was used to measure the external surface (S) of EBSD sample (see Figure 1(d)) by combing a Zeiss-Sigma 300 SEM. The scanning step was set as 1 μm, and the direction of loading (LD), transverse (TD), and normal (ND) are given in Figure 1(d). Ultimately, the EBSD testing dataset was analyzed by using the ATEX-4.03 software [18].

Figure 1

The creep experiments of Super304H austenitic heat-resistance steel. (a) The Super304H tube; (b) the sampling method; (c) the dimension of specimen; (d) the diagram of creep specimen; and (e) the interrupted creep experiments program.

3 Results and discussion

The Kikuchi pattern can be recognized as a reflection of the inner microscopic crystal structure of the material. Much information about crystal structure, orientation, and symmetry can be extracted from it. The quality of Kikuchi patterns could reflect the degradation degree of a material by handling the sharpness of Kikuchi bands at any given point [10], since the Kikuchi line broadening is related to the number of lattice defects in the measured sample. In this research, several parameters that reflecting the quality of the Kikuchi pattern were used to characterize the creep damage of Super304H steel.

BC value is a representative Pattern Quality parameter, which is derived from the Hough transform, a ratio of the average intensity of Kikuchi bands relative to the overall intensity within the electron backscatter patterns (EBSP). Consequently, the ratio is scaled to the numeric value from 0 to 255, the higher the value the better the pattern quality. The BC maps of Super304H austenitic heat-resistance steel are shown in Figure 2, and the change of BC value with different creep times cannot be intuitively seen through Figure 2(a)–(f). Thus, the data average processing was conducted and the mean BC value (BC_ave) of each map was obtained. From Figure 2(g), it can be recognized that the BC_ave did not show a monotonic changing trend, which was different from the situation of plastic deformation of ordinary steels [11]. In our previous research [11], we try to use the three parameters to study the degradation behavior in 316 steel during plastic deformation. The in situ EBSD results revealed that with the increase of plastic strain, the BC value monotonously decreased in both grain and grain boundaries of 316 steel. Here, the results indicate that the overall statistics of the observation area of the specimens by BC parameter may not be effective to assess the creep damage of austenitic steel.

Figure 2

BC of Super304H steel. (a–f) Creep for original state, 510, 960, 1,400, 1,465, and 1,530 h, respectively. The higher the value (i.e., the lighter the color), the better the pattern quality. (g) The mean value of BC of specimens at different creep stages, and the data are obtained by averaging BC values of all pixels in the field of view of each specimen.

Nevertheless, in-depth research that distinguished grain boundary and grain interior regions provides meaningful results. A detailed view of BC maps is displayed in Figure 3(a), which corresponds to the yellow frames in Figure 2. In each specimen, nine pixels were arbitrarily selected along the grain boundaries (GBs, red spots) and inside the grain (yellow spots), respectively. The average value of these two kinds of pixels is shown in Figure 3(b) and (c). In the original specimen, the BC_ave along GBs was 197.97, which was lower than that BC_ave inside grain (i.e., 206.98), the result also verified that the GB region has a poor quality of Kikuchi pattern, which could be due to the irregular array of atoms at this location. With the increase of creep time, the BC_ave in both cases showed a monotonic decreasing trend, in spite of the difference of BC value among these 9 pixels may large (see error bar). Besides, the falling rate of BC_ave along GBs or inside grain in the acceleration creep stage was much faster than that in the initiation and steady stage, which manifests that the deterioration of microstructure of Super304H austenitic heat-resistance steel was more severe at the acceleration creep stage. In addition, comparing the error bars in Figure 3(b) and (c), it can be clearly seen that the error of BC_ave within the grain is significantly less than that along the GBs. Therefore, the BC_ave was more suitable for the damage assessment of grain instead of grain boundary (GB).

Figure 3

A detailed view of BC maps and the statistical results. (a) The selected pixels along the GBs and inside the grain of each specimen. The selected grain of each specimen is shown in the yellow frame in Figure 2(a)–(f), respectively. (b and c) The statistical results of specimens of the BC along the GBs and inside the grain at different creep stages.

The mean angular deviation (MAD) is another Quality parameter, which represents how well the simulated EBSP overlays the actual EBSP. It was given in degree and specifies the averaged angular misfit between detected bands and simulated Kikuchi bands. In contrast to BC, the smaller the MAD value, the better the pattern quality and the lower the degradation degree. The MAD maps of Super304H austenitic heat-resistance steel are shown in Figure 4(a)–(f). It can be seen that the MAD value presented an overall upward trend with increasing creep time. However, after we extracted the data of pixels and conducted the average processing, we found that the change of MAD_ave was not in a strict monotonic increase (see Figure 4(g)), especially at the early creep stage (0–510 h). Therefore, just immediately using the MAD_ave to assess the creep damage of Super304H may not be possible.

Figure 4

MAD of Super304H steel. (a–f) Creep for original state, 510, 960, 1,400, 1,465, and 1,530 h, respectively. The smaller the value (i.e., the darker the color), the better the pattern quality. (g) The mean value of MAD of specimens at different creep stages, and the data are obtained by averaging MAD values of all pixels in the field of view of each specimen.

A detailed view of MAD maps is also given in Figure 5(a), similarly, nine pixels were selected along the GBs and inside the grain, which are the same position as selected in BC maps. It can be seen from the error bar directly that the error of MAD_ave in each creep stage is relatively large, both at the GBs and inside the grain, indicating that this evaluation parameter has a certain instability. Figure 5(b) and (c) shows the average value of these two kinds of pixels, unintelligibly, the MAD_ave does not present a monotonic increasing but appears a decreasing tendency in the initiation creep stage and most of the time in the steady creep stage. Logically, the formation of defects during creep will result in the increase in deviation between the fitting Kikuchi bands and the detected bands and lead to the increase in MAD value. Here, the decrease in MAD_ave has not yet been fully understood. Combined with the above results, it is found that the MAD may not be an effective parameter to assess the creep damage of austenitic heat-resistance steel.

Figure 5

A detailed view of MAD maps and the statistical results. (a) The selected pixels along the GBs and inside the grain of each specimen. The selected grain of each specimen is shown in the yellow frame in Figure 4(a)–(f), respectively. (b and c) The statistical results of specimens of the MAD along the GBs and inside the grain at different creep stages.

Band slope (BS) is an image quality factor derived from the Hough transform in an EBSP. This parameter reflects the maximum intensity gradient at the margins of the Kikuchi bands. The result values are also scaled from 0 to 255, the higher the value, the sharper the band, oppositely, the lower the value, the blurrier the Kikuchi band, and the larger the deformation degree of the material. Figure 6(a)–(f) shows the BS maps of Super304H steel with different creep times, and the change in BS can be observed here, especially at the acceleration creep stage, since it manifested as a significant decrease in BS value (see Figure 6(g)), which indicates that this parameter is effective upon the creep damage characterization of Super304H.

Figure 6

BS of Super304H steel. (a–f) Creep for original state, 510, 960, 1,400, 1,465, and 1,530 h, respectively. The larger the value (i.e., the darker the red), the better the pattern quality. (g) The mean value of BS of specimens at different creep stages, and the data are obtained by averaging BS values of all pixels in the field of view of each specimen.

A detailed view of BS maps is displayed in Figure 7(a), the mark points in white color are those pixels along the GBs, while those points in black are the pixels that are distributed inside the grain. Figure 7(b) and (c) shows the change in BS_ave of these two kinds of pixels. Different from MAD_ave, the BS_ave displayed a monotonic decreasing trend, in both the steady creep stage and the acceleration creep stage. In addition, the decline in BS_ave is more obvious in those pixels along the GBs (42.84%), especially at the later steady creep stage. This result illustrates that the GB regions are more vulnerable to suffer the creep damage. However, it is worth noting that the BS_ave error increases largely after entering the acceleration stage, indicating that the reliability of this parameter decreases with the increase in creep time. For BS_ave inside the grain, although the parameter changes little in the first two stages, the data change dramatically after entering the acceleration stage while keeping the error low. Therefore, the results confirm that the two BS_ave parameters all can be used to evaluate the creep damage of austenitic heat-resistance steel, but both have their own applicable ranges.

Figure 7

A detailed view of BS maps and the statistical results. (a) The selected pixels along the GBs and inside the grain of each specimen. The selected grain of each specimen is shown in the yellow frame in Figure 6(a)–(f), respectively. (b–c) The statistical results of specimens of the BS along the GBs and inside the grain at different creep stages.

4 Conclusions

The damage evaluation of austenitic heat-resistance steel during creep was investigated in this article by using the EBSD technique. The parameters based on the image quality of the Kikuchi bands, e.g., BC, MAD, and BS, were used to characterize the damage degree of Super304H steel. The results manifested that, after the average processing of data in the observation area, the BC_ave and MAD_ave showed a non-monotonic change with increasing creep time; however, the BS_ave displayed an approximate monotonic decreasing trend, and the falling range is distinct at the acceleration creep stage. An in-depth research that distinguished grain boundary and grain interior regions provided more results. With increasing creep time, the MAD_ave still did not present a regular change, while the BC_ave and BS_ave displayed a monotonic decreasing trend, especially at the GBs. The results indicated that the BC_ave especially inside the grain one could be an effective diagnostic indicator that helps estimate the damage degree of austenitic heat-resistance steel, the combination of the two BS_ave takes second place, while the MAD may not be suitable for the damage evaluation of austenitic heat-resistance steel.