2016 Volume 57 Issue 12 Pages 2097-2103
The creep rupture strength and the degradation mechanism of 23Cr-43Ni-7W alloy weld joints were investigated using creep rupture tests and microstructure observation of the ruptured specimens. The creep rupture tests were conducted at 973, 1023, and 1073 K at stresses from 80 to 180 MPa. The creep strength of the weld joints was higher than that of the base metal. The ruptured area of every specimen was over 10 mm away from the bond line. The microcracks increased gradually up to 10 mm from the bond line, beyond which they suddenly increased. In the grains, traces of M23C6 were observed near the bond line, and the sizes of the M23C6 deposits increased further away from the bond line. But the size of the Laves phase was constant in every observed area. On the grain boundary, the grain boundary shielding ratio by precipitates was constant until 10 mm from the bond line, beyond which it decreased. These observations show that the creep strength near the bond line is higher than that of the base metal because of both precipitation strengthening in the grain and grain boundary.
This Paper was Originally Published in Japanese in J. Japan Inst. Met. Mater. 79 (2015) 348–355.
In recent years, a further increase in steam temperature has been required of coal-fired power plants in order to improve power generation efficiency and to reduce CO2 emissions. To meet this demand, the development of Advanced-Ultra Super Critical (A-USC) thermal power plants, which can withhold steam temperatures of 700℃, 100℃ higher than the latest 600℃ class Ultra Super Critical (USC) thermal power plants,1) is actively being conducted.2–5) Ferritic heat resistant steel and austenitic heat resistant steel are used in current designs. However, they cannot be used in the highest-temperature parts of thick-walled pipes and heat transfer pipes of A-USC plants, as the higher steam temperature leads to problems from the viewpoint of creep strength and high-temperature corrosion. Therefore, Ni-based alloys, which are stronger and have a higher corrosion resistance, are required. In Japan, the application of three alloys, 23Cr-45Ni-7W (HR6W),6) 30Cr-50Ni-4W-Ti (HR35),7) and Ni-22Cr-12.5Co-9Mo (Alloy 617), is being considered for the manufacture of the thick-walled pipes of A-USC plants. In particular, HR6W has received the most attention as a potential material for the main pipes because it has higher ductility at high temperatures and it exhibits excellent resistance to thermal fatigue in comparison with other Ni-base alloys.8,9)
On the other hand, because there have been no cases where a Ni-based alloy was used to build boilers, it is necessary to investigate the long-term creep rupture properties of the welds that are vital and most important for the production of a boiler. This is because the base metal near a weld bond line generally has a thermal history different from that of the original base metal due to the heat generated from forming the weld, and consequently the structure and strength of the base metal near the bond line may be significantly different. In addition, knowledge on the fracture morphology and fracture mechanism of the welds is important from the viewpoint of service and maintenance. However, there are no such reports on Ni-based alloys as far as the authors of this article know. Therefore, the purpose of this study is to conduct creep rupture tests on the Ni-based alloy HR6W, verify the creep rupture strength of the weld joints, and clarify the fracture morphology and structural characteristics of the welds by observing from ruptured materials.
In this study, HR6W plates with a thickness of 25 mm and subjected to a 1,220℃ solution treatment were used as samples. Table 1 shows their chemical compositions. Alloy 617 was used as the welding material, and weld joints were prepared by gas tungsten arc welding (GTAW). Figure 1 shows a cross-section macro photograph of the weld joints. As shown in Fig. 2, creep rupture specimens were taken from the weld joints so that a weld bond line was located at the center of the parallel portion of each specimen. The shape of the specimens had a parallel portion with a diameter of 6 mm and a length of 30 mm. Creep rupture tests were conducted under the temperatures of 700℃–800℃ and stresses of 80–180 MPa. The maximum test time was approximately 10,000 hours at each temperature. After creep rupture, Vickers hardness tests and macro and micro observations were conducted on the surface of each specimen, obtained by cutting the parallel portion of each specimen parallel to the stress axis and the radial direction. The Vickers hardness was measured with a load of 98 N at intervals of 1 mm from the weld metal to the base metal. An optical microscope (OM), a scanning electron microscope (SEM), and a scanning transmission electron microscope (S-TEM) were used for the micro observations. Secondary electrons were used for crack observations by SEM, and backscattered electrons were used for precipitate observations. In addition, S-TEM observations were conducted by using carbon extraction replicas. For precipitates on the grain boundaries, the grain boundary shielding ratio ρ in multiple regions was calculated by formula (1).10)
| \[\rho = 1 - \sum \lambda i/L \] | (1) |
| (mass%) | ||||||||
|---|---|---|---|---|---|---|---|---|
| TP | C | Cr | W | Ti | Nb | B | Fe | Ni |
| HR6W | 0.082 | 23.1 | 7.07 | 0.10 | 0.20 | 0.0042 | 24.1 | Bal. |
Cross section macrostructure of HR6W weld joints.
Macrostructure of a weld joint showing the sampling position of creep specimen in HR6W.
The grain boundary shielding ratio ρ of each region was calculated from all views, after 10 images were photographed at a magnification of 5,000 times, with the grain boundary triple point located at the center of each view. In addition, an average grain diameter of the Laves phase in grains was determined by averaging the values of 100 equivalent circle diameters observed in backscatter electron images. An average grain diameter of the M23C6 carbide was determined by averaging the values of 100 equivalent circle diameters observed using S-TEM.
Figure 3 shows the Vickers hardness profile of the welded joints. The bond line in the figure is marked as 0 mm, with the positive values indicating the base metal and negative values indicating the weld metal. The Vickers hardness of the weld metal was approximately 250 HV at its hardest, and its hardness was 216 HV when measured at a distance of 1 mm from the bond line. The Vickers hardness decreased as the distance from the bond line on the base metal increased, to a value of 177 HV, which was the hardness of the base metal before welding, at a distance of 10 mm from the bond line. Figures 4 and 5 show cross-sectional OM images and SEM images respectively taken at each distance from the bond line in the prepared weld joint. Significant coarsening of grains or a recrystallized structure was not found at the distances of 1 mm, 5 mm, and 10 mm from the bond line. In addition, precipitates were not found, and the main difference between the weld-heat affected zone (HAZ) and the base metal of the welded material was a change in hardness. Therefore, it can be suggested that the changes in hardness from the bond line were caused by differences in the amount of plastic strain introduced when the weld cooled, i.e. differences in dislocation density.11) Thus, the region up to 10 mm from the bond line was determined as the HAZ.
Vickers hardness distribution across the fusion line for weld joint in HR6W.
Optical micrograph of weld joint in HR6W. (a) 1.0 mm, (b) 5.0 mm, and (c) 10.0 mm from the fusion line.
Secondary electron images of weld joint in HR6W. (a) 1.0 mm, (b) 5.0 mm, and (c) 10.0 mm from the fusion line.
Figure 6 shows the creep rupture test results of the weld joints. The dashed lines indicate average lines of creep rupture strength for the base metal.12) In every test result for this study, lasting approximately 600 h to 10,000 h, the creep rupture strength of the weld joints at 700℃ was higher than the average strength of the base metal. This was also observed at temperatures of 750℃ and 800℃.
Stress-rupture data of weld joints for HR6W.
Figure 7 shows the cross-sectional macro observation results of specimens ruptured at the longest time for each temperature. Each specimen was ruptured at the base metal at a distance of 10 mm or more from the bond line. When the same observation was conducted for specimens of other test conditions, all of them were found to be ruptured at the base metal.
Cross section macrostructures of HR6W weld joints ruptured at the following conditions; (a) 700℃ 140 MPa, (b) 750℃ 100 MPa, (c) 800℃ 80 MPa.
Figure 8 shows SEM observation results at distances of 1 mm (a), 10 mm (b) and 15 mm (c) from the bond line of the creep-ruptured material, tested at 800℃ and 80 MPa. Cracks and creep voids did not occur on grain boundaries at distances of 1 mm (a) and 10 mm (b). Creep cracks with a length of approximately 150 μm were observed at a distance of 15 mm (c), and the length and number of cracks increased as the distance from the ruptured area decreased. The region up to 10 mm from the bond line was defined earlier as the HAZ. Therefore, the cracks did not occur in the HAZ. In order to examine the cause of this, detailed observations were conducted on the precipitates in the HAZ and the base metal.
Secondary electron images of HR6W weld joint ruptured at 800℃, 80 MPa. (a) 1.0 mm, (b) 10.0 mm, and (c) 15.0 mm from the fusion line.
Creep strength of HR6W increases by precipitation strengthening of the Laves phase and the M23C6 carbide.6) Therefore, attention was focused on these two types of precipitates, and differences between the structures of the HAZ and the base metal were considered.
First, the HAZ and the base metal were observed for precipitates on grain boundaries by using backscatter electrons. Figure 9 shows the results. Precipitates with white color are the Laves phase, and precipitates with gray color are the M23C6 carbide. The size of the Laves phase on grain boundaries was nearly constant regardless of the distance from the bond line. However, the size and the amount of M23C6 carbide precipitates were different depending on the distance from the bond line. Therefore, the fine M23C6 carbide, with a diameter of approximately 0.5 μm, was precipitated so as to closely cover the grain boundaries at a distance of 1 mm (a) from the bond line. The diameter of the M23C6 carbide precipitates at a distance of 5 mm (b) from the bond line was approximately 1 μm, which was slightly larger than the ones at a distance of 1 mm, while the ratio of shielding of grain boundaries was nearly equal to the ones at a distance of 1 mm. However, the size of the M23C6 carbide precipitates near the ruptured area at a distance of 15 mm (c) from the bond line was approximately 4 μm, which was larger than the ones at distances of 1–5 mm, and the ratio of shielding of grain boundaries also decreased. Takeyama et al. reported that the creep strength increased as the ratio of shielding of grain boundaries (grain boundary shielding ratio) increased in any Ni-based alloy.13–15) Therefore, the relationship between the distance from the bond line of the creep-ruptured material and the grain boundary shielding ratio was quantified. The results are shown in Fig. 10. The grain boundary shielding ratio at distances of 1–10 mm from the bond line was approximately 80% and nearly constant while the grain boundary shielding ratio near the ruptured area at a distance of 10 mm or greater from the bond line decreased to approximately 65%. As described in Section 3.3, the number of cracks increased in the region at a distance of 10 mm or greater from the bond line, and this trend was in agreement with the changes on the grain boundary shielding ratio.
Backscattered electron images of grain boundaries for HR6W weld joint ruptured at 800℃ and 80 MPa. (a) 1.0 mm, (b) 5.0 mm, and (c) 15.0 mm from the fusion line.
The difference of grain boundary shielding ratio by precipitates for HR6W weld joint ruptured at 800℃ and 80 MPa.
Next, the comparison of precipitates in grains was conducted. Because the Laves phase and the M23C6 carbide within the interior of grains were different in size, the Laves phase and the M23C6 carbide were observed in backscatter electron images and S-TEM bright field images respectively. Figure 11 shows backscatter electron images in grains. White needle-like precipitates observed in all regions are the Laves phase. There was little change in the size and the amount of precipitation of the Laves phase in the HAZ at distances of 1 mm (a) and 5 mm (b) from the bond line and in the base metal at a distance of 15 mm (c). Figure 12 shows the differences in the average grain size of the Laves phase in grains depending on the distance from the bond line. The average grain diameter of the Laves phase was approximately 0.8 μm and nearly constant regardless of their distance from the bond line. In addition, their size distribution was nearly constant. Thus, precipitation strengthening by the Laves phase in grains was considered to be almost the same in the HAZ and the base metal.
Backscattered electron images of intragranular grain for HR6W weld joint ruptured at 800℃ and 80 MPa. (a) 1.0 mm, (b) 5.0 mm, and (c) 15.0 mm from the fusion line.
The average diameters of Laves phase for HAZ and base metal ruptured at 800℃ and 80 MPa.
Next, Fig. 13 shows observation results of M23C6. Spherical M23C6 with sizes of approximately 10–50 nm was observed at a distance of 1 mm (a) from the bond line. There was little change in the sizes at a distance of 4 mm (b) from the bond line. However, the sizes at a distance of 10 mm (c) from the bond line were slightly larger than those at distances of 1 mm (a) and 4 mm (b). Therefore, the average grain diameter of the M23C6 carbide was calculated and plotted as a function of the distance from the bond line. The results are in Fig. 14. The average grain diameter of the M23C6 carbide was 38.1 nm at a distance of 1 mm from the bond line, 38.3 nm at a distance of 4 mm from the bond line, and nearly constant. On the other hand, the average grain diameter of the M23C6 carbide at a distance of 10 mm from the bond line was 52.9 nm, which was approximately 1.4 times larger than that in the other regions. In addition, it was found that the maximum grain diameter at a distance of 1 mm from the bond line was approximately two or more times smaller than the ones at a distance of 10 mm.
Scanning transmission electron images of intragranular grain for HR6W weld joints ruptured at 800℃ and 80 MPa. (a) 1.0 mm, (b) 4.0 mm, and (c) 10.0 mm from the fusion line.
Change in the size of M23C6 carbide in HAZ with the distance from the fusion line of weld joint in HR6W ruptured at 800℃ under 80 MPa.
It was suggested that the reason why the M23C6 carbide in the HAZ was smaller than that in the base metal was differences in dislocation density due to initial differences in Vickers hardness. Many results were reported showing that when a Ni-based alloys and austenitic steels were subjected to cold working, they were more stably refined for a long time by precipitation on dislocations of the M23C6 carbide.16–19) In addition, it was reported that the M23C6 carbide in the grains improved creep strength.16) Thus, it was suggested that the reason why the M23C6 carbide in the HAZ was smaller than the ones outside the HAZ was that the M23C6 carbide precipitated on dislocations introduced at the time of welding and was stable for a long time. From the above, it can be said that one of the reasons why the creep strength of the HAZ improved was that the M23C6 carbide in the grains of the HAZ acted on creep strengthening more significantly than the M23C6 carbide with in in the grains of the base metal.
3.5 Effect of precipitates on creep rupture strength of weld jointsFigure 15 shows a schematic diagram of microstructural changes in the HAZ and the base metal before and after creep tests. Before the creep tests, they were in the solution-treated state, and precipitates were not observed in the HAZ near the bond line and in the base metal at a distance of 10 mm from the bond line. In addition, significant coarsening of grains or recrystallization did not occur. After creep rupture, the size of the Laves phase in grains and on grain boundaries was nearly constant regardless of their distance from the bond line. On the other hand, the M23C6 carbide in grains and on grain boundaries of the HAZ was precipitated more finely than that in the base metal. Thus, the difference in microstructure between the base metal and the HAZ was fine precipitation of the M23C6 carbide in grains and on grain boundaries. In other words, as the distance from the bond line increased, so did the size of the M23C6 carbide increase, and became nearly equal to that of the base metal at a distance of approximately 10 mm. On grain boundaries, the grain boundary shielding ratio at distances of up to 10 mm from the bond line was nearly constant while the grain boundary shielding ratio at more distant positions decreased. These effects of precipitation strengthening in grains and on grain boundaries appeared in the creep strength of the HAZ and the base metal. On the other hand, cracks did not occur in the HAZ at distances of up to 10 mm from the bond line as described by the SEM observations in Section 3.3, but increased at a distance of 10 mm or more. From the above, the region where cracks occurred and increased was in close agreement with the region where precipitation strengthening decreased. Thus, it became clear that the creep rupture of HR6W weld joints was caused by differences in the forms of precipitates in grains and on grain boundaries of the HAZ and the base metal.
Schematic illustrations showing the precipitates of weld joint in HR6W before and after creep at 800℃ under 80 MPa.
The creep rupture strength of 23Cr-45Ni-7W alloy (HR6W) weld joints was clarified, and the relationship between their fracture morphology and structure was considered. The findings are described below.
(1) The creep rupture strength of HR6W weld joints was equal to or higher than the average rupture strength of the base metal. From the macro observations, the rupture position of every specimen was in the base metal at a distance of 10 mm or more from the bond line.
(2) From the micro observations of ruptured materials, cracks and creep voids were not observed in the HAZ at distances of 1–10 mm from the bond line. However, cracks were observed in the base metal at a distance of 10 mm or more, and the length and number of cracks increased as the distance from the ruptured area decreased.
(3) The grain boundary shielding ratio by the Laves phase and the M23C6 carbide was constant in the HAZ. However, it decreased in the base metal at a distance of 10 mm or more. This is because the size of the M23C6 carbide increased as the distance from the base metal decreased.
(4) The average grain diameter of the Laves phase in grains was constant regardless of their distance from the bond line. On the other hand, the average grain diameter of the M23C6 carbide increased as the distance from the bond line increased. It was suggested that the precipitation of the M23C6 carbide in the HAZ was caused by precipitation on dislocations and was distributed more finely than in the base metal.
From the above, it was concluded that the reason why cracks were not observed in the HAZ of HR6W weld joints was that M23C6 in grains was precipitated more finely than in the other regions, and that the high grain boundary shielding ratio contributed largely to precipitation strengthening in grains and on grain boundaries.
