复旦大学：《材料失效分析 Materials Failure Analysis》课程教学资源（教学案例）22. Microstructure and mechanical properties of dissimilar materials joints between T92 martensitic and S304H austenitic steels

Materials and Design 32(2011)2763-2770 Contents lists available at ScienceDirect Materials and design ELSEVIER journalhomepagewww.elsevier.com/locate/matdes Microstructure and mechanical properties of dissimilar materials joints between t92 martensitic and s304h austenitic steels Jian Cao, Yi Gong, Kai Zhu, Zhen-Guo Yang Xiao-Ming Luo, Fu-Ming Gu of Special Equipment Inspection Technical Research, Shanghai 200062, PR China ARTICLE O A BSTRACT Article history: In this paper, 192 martensitic steel and S304H austenitic steel were welded by gas tungsten arc welding (GTAW)process. Microstructural features and mechanical properties of T92 and S304H dis rials joints were investigated. The results showed that the part of the joints with relativel Available online 8 January 201 strength was T92 coarse-grained heat affected zone(CGHAZ), while the part of the joints which revealed elatively we hess was weld metal. The decrease of t strength in 192 CGHAZ parse ten martensite structure. Weak toughness of the joints was resulted from the coarse of the weld metal. however. the weld in transverse direction of the joints milar materials joints was provided higher tensile strength by the orientation distribution of grains compared with T92 CGHAZ. e 2011 Elsevier Ltd. All rights reserved. 1 Introduction tensitic steels[14-17. Thus, it is widely used for superheaters and reheaters which have the abominable service environment in us Increased heat efficiency and improved environmental protec- boilers. Normally, t92 steel can be used as pipes linking superheat- tion are always the innovative driving forces in the development ers and reheaters. In this case, the welding between T92 and S304H of ultra supercritical(USC) boilers for fossil power plants, whose steels will be necessary. steam temperature is up to 600C and pressure exceeds 27 MPa Until now, many researches have mainly focused on the proper Under this USC condition, the heat efficiency can rise to around ties of T92 and S304H steels. As for T92 and S304H dissimilar 45%, compared with the value as 41% of supercritical (SC)boilers. materials joints, however, there almost hasn't any report about However, with the increase of steam parameters, requirements it. Since T92/S304H dissimilar materials joints is obtained by using for the materials applied in the USc boilers components are melted filler material to join two steels, the melted filler material becoming higher. Thus, many new generation steels have been will re-crystal to form the weld metal part of the joints after weld- developed in recent years, including T92(9Cr05Mo2WVNb) ing. In addition, due to the effect of welding thermal cycles, not martensitic steel and S304H (18Cr9Ni3CuNbN)austenitic steel only the microstructure of T92 adjacent weld metal but also the T92 steel was developed by the Nippon Steel Corporation of microstructure of S304H adjacent weld metal will both change Japan 1 in the late 1990s by modifying chemical compositions during the welding process. Considering that the mechanical prop- pon T91(9Cr1MoVNb) for even more preferable mechanical erties of the joints are closely linked with its microstructure. Thus, properties at high temperatures. This steel has the manufacturers an in-depth insight into the structure-property relationships of designation as NF616(ASTM Stands A213)and contains 0.5% Mo, T92/S304H dissimilar materials joints may have great significan 88 W, as well as small additions of Nb, V and B Creep strength for both the dissimilar steels welding process between new gener f T92 at 600C increases about 10-20% compared with that of ation martensitic and austenitic steels, and the safety of the Usc T91 at 600C[2-13 S304H steel was developed by Sumitomo Me- boilers In our work, on the one hand, mechanical properties of tal Industries Ltd on the base of TP304H(OCr19Ni10. As a new T92/S304H dissimilar materials joints were carried out through type of austenitic steel, S304H possesses not only excellent resis- tensile and impact tests. On the other hand the microstructures tance to high-temperature corrosion and steam oxidation mainly across the entire joints were also investigated. What's more, the due to high Cr content, but also superior creep strength than mar- detailed mechanism governing the microstructural evolution of the joints during welding process was analyzed by means of the electron back-scattered diffraction(EBSD) technique, which was Corresponding author. Tel. +86 21 65642523: fax: +86 21 65103056 firstly used to study the process of grain structure development of dissimilar materials joints 0261-3069/s- see front matter o 2011 Elsevier Ltd. All rights reserved doi:10.1016 mates.201101.008

Microstructure and mechanical properties of dissimilar materials joints between T92 martensitic and S304H austenitic steels Jian Cao a , Yi Gong a , Kai Zhu a , Zhen-Guo Yang a,⇑ , Xiao-Ming Luo b , Fu-Ming Gu b aDepartment of Materials Science, Fudan University, Shanghai 200433, PR China b Shanghai Institute of Special Equipment Inspection & Technical Research, Shanghai 200062, PR China article info Article history: Received 8 September 2010 Accepted 4 January 2011 Available online 8 January 2011 Keywords: Microstructure Mechanical properties Dissimilar materials joints abstract In this paper, T92 martensitic steel and S304H austenitic steel were welded by gas tungsten arc welding (GTAW) process. Microstructural features and mechanical properties of T92 and S304H dissimilar mate￾rials joints were investigated. The results showed that the part of the joints with relatively weak tensile strength was T92 coarse-grained heat affected zone (CGHAZ), while the part of the joints which revealed relatively weak toughness was weld metal. The decrease of tensile strength in T92 CGHAZ was due to its coarse tempered martensite structure. Weak toughness of the joints was resulted from the coarse dendritic austenite of the weld metal. However, the weld metal in transverse direction of the joints was provided higher tensile strength by the orientation distribution of grains compared with T92 CGHAZ. 2011 Elsevier Ltd. All rights reserved. 1. Introduction Increased heat efficiency and improved environmental protec￾tion are always the innovative driving forces in the development of ultra supercritical (USC) boilers for fossil power plants, whose steam temperature is up to 600 C and pressure exceeds 27 MPa. Under this USC condition, the heat efficiency can rise to around 45%, compared with the value as 41% of supercritical (SC) boilers. However, with the increase of steam parameters, requirements for the materials applied in the USC boilers components are becoming higher. Thus, many new generation steels have been developed in recent years, including T92 (9Cr0.5Mo2WVNb) martensitic steel and S304H (18Cr9Ni3CuNbN) austenitic steel. T92 steel was developed by the Nippon Steel Corporation of Japan [1] in the late 1990s by modifying chemical compositions upon T91 (9Cr1MoVNb) for even more preferable mechanical properties at high temperatures. This steel has the manufacturer’s designation as NF616 (ASTM Stands A213) and contains 0.5% Mo, 1.8% W, as well as small additions of Nb, V and B. Creep strength of T92 at 600 C increases about 10–20% compared with that of T91 at 600 C [2–13]. S304H steel was developed by Sumitomo Me￾tal Industries Ltd on the base of TP304H (0Cr19Ni10). As a new type of austenitic steel, S304H possesses not only excellent resis￾tance to high-temperature corrosion and steam oxidation mainly due to high Cr content, but also superior creep strength than mar￾tensitic steels [14–17]. Thus, it is widely used for superheaters and reheaters, which have the abominable service environment in USC boilers. Normally, T92 steel can be used as pipes linking superheat￾ers and reheaters. In this case, the welding between T92 and S304H steels will be necessary. Until now, many researches have mainly focused on the proper￾ties of T92 and S304H steels. As for T92 and S304H dissimilar materials joints, however, there almost hasn’t any report about it. Since T92/S304H dissimilar materials joints is obtained by using melted filler material to join two steels, the melted filler material will re-crystal to form the weld metal part of the joints after weld￾ing. In addition, due to the effect of welding thermal cycles, not only the microstructure of T92 adjacent weld metal but also the microstructure of S304H adjacent weld metal will both change during the welding process. Considering that the mechanical prop￾erties of the joints are closely linked with its microstructure. Thus, an in-depth insight into the structure–property relationships of T92/S304H dissimilar materials joints may have great significances for both the dissimilar steels welding process between new gener￾ation martensitic and austenitic steels, and the safety of the USC boilers. In our work, on the one hand, mechanical properties of T92/S304H dissimilar materials joints were carried out through tensile and impact tests. On the other hand, the microstructures across the entire joints were also investigated. What’s more, the detailed mechanism governing the microstructural evolution of the joints during welding process was analyzed by means of the electron back-scattered diffraction (EBSD) technique, which was firstly used to study the process of grain structure development of dissimilar materials joints. 0261-3069/$ - see front matter 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.matdes.2011.01.008 ⇑ Corresponding author. Tel.: +86 21 65642523; fax: +86 21 65103056. E-mail address: zgyang@fudan.edu.cn (Z.-G. Yang). Materials and Design 32 (2011) 2763–2770 Contents lists available at ScienceDirect Materials and Design journal homepage: www.elsevier.com/locate/matdes

2764 J. Cao et al Materials and Design 32(2011)2763-2770 2.1. Materials and welding procedure Scales of the two base materials t92 and s304h used in the S304H present investigation were 57 150D x 5.08 mm and 570D x 4.5 mm in thickness respectively. Heat treatment conditions of them were described below: 1)T92: austenitization was carried Unit: mm out for 20 min at 1050C and then tempering for 60 min at 60C. 2)S304H: solution treatment at 1100C, followed by water cooling until room temperature. Fig. 1 was schematic diagram of ig2 Schematic diagram of welding process of T92/S304H dissimilar materials corresponding to INCONEL 82/182)was chosen as the filler mate- rial. The chemical col tions of all three materials are given in Table 1. The t92/S304H dissimilar materials joints were weld by means of gas tungsten arc welding (GTAw) with pure argon gas(Ar)as the shielding gas. The arc voltage and the arc current PWH used in welding process were 14 V and 230A(type: balance) spectively. The argon purity used for GTAW in this experiment is 99.99%. Fig 2 was the schematic diagram of the welding process part marked with dotted line in Fig. 2 represented the 730-760℃ egion. After welding, post welding heat treatment was carried out for 1 h at 730-760 C to eliminate the welded residual stress. Fig 3 was the schematic diagram of the Heating rat ≤220℃h PWhT process of T92/S304H dissimilar materials joints and Fig. 4 vas the schem of T92/S304H dissimilar materia 2. 2. Test methods Tensile and impact tests of t92/S304H dissimilar materials Time(min) joints were done respectively at temperature according to Fig 3. Schematic diagram of post welding heat treatment process of T92/S304H the ASTM E8-04 [18 and E23-02a [19]standards. Optical micros- dissimilar materials joints. copy(oM)was used to observe the metallographic microstructures across the joints. Three etching solutions were used: (1)T92: picric acid (2, 4, 6-trinitrophenol)2.5 g, ethanol 20 ml and H20 20 ml; (2) weld metal: FeCl3 5 g, HCI 20 ml and H20 20 ml; (3)S304H: CuSOA T92 ERNICr-3 S304H 4 g, HCI 20 ml and ethanol 20 ml PHILIPS XL30FEG scanning elec B M WM 1100°C Fig. 4. Schematic diagram of T92/S304H dissimilar materials joints. D20min S304H tron microscopy(SEM)was employed to analyze the fractograph 1050°C T92 760°C of impact specimens. The specimens used for EBSD investigation were obtained from the joints in location A and b( Fig. 5),and 60min suitable surface finished for EBSD was obtained by applying the Heating rate mechanical polishing followed by jet electro-polishing in 6% solu- ≤300℃/h tion of perchloric acid in methanol. The schematic diagram of the specimen in EBSD system was displayed in Fig. 6, in which ND RD and tD stood for normal, rolling and transverse direction espectively. EBSD micrographs were taken on the TD-Rd plane Fig. 1 Schematic diagram of heat treatment processes of 192 and S304H steels so that the rd is horizontal and the nd is vertical Table 1 Chemical compositions of the base materials and filler material samples(wt%). 8.76 0.4600160.002 0044001 1.6300033 S304H sample 0.09 18.38 004004 8910.8200330 0.1 ERNiCr-3 flle 0030 20.0 25290 004

2. Experimental 2.1. Materials and welding procedure Scales of the two base materials T92 and S304H used in the present investigation were 57.15OD 5.08 mm and 57OD 4.5 mm in thickness respectively. Heat treatment conditions of them were described below: 1) T92: austenitization was carried out for 20 min at 1050 C and then tempering for 60 min at 760 C. 2) S304H: solution treatment at 1100 C, followed by water cooling until room temperature. Fig. 1 was schematic diagram of heat treatment processes of T92 and S304H steels. AWS ERNiCr-3 (corresponding to INCONEL 82/182) was chosen as the filler mate￾rial. The chemical compositions of all three materials are given in Table 1. The T92/S304H dissimilar materials joints were welded by means of gas tungsten arc welding (GTAW) with pure argon gas (Ar) as the shielding gas. The arc voltage and the arc current used in welding process were 14 V and 230A (type: balance) respectively. The argon purity used for GTAW in this experiment is 99.99%. Fig. 2 was the schematic diagram of the welding process and the part marked with dotted line in Fig. 2 represented the welding region. After welding, post welding heat treatment (PWHT) was carried out for 1 h at 730–760 C to eliminate the welded residual stress. Fig. 3 was the schematic diagram of the PWHT process of T92/S304H dissimilar materials joints and Fig. 4 was the schematic diagram of T92/S304H dissimilar materials joints. 2.2. Test methods Tensile and impact tests of T92/S304H dissimilar materials joints were done respectively at room temperature according to the ASTM E8-04 [18] and E23-02a [19] standards. Optical micros￾copy (OM) was used to observe the metallographic microstructures across the joints. Three etching solutions were used: (1) T92: picric acid (2, 4, 6-trinitrophenol) 2.5 g, ethanol 20 ml and H2O 20 ml; (2) weld metal: FeCl3 5 g, HCl 20 ml and H2O 20 ml; (3) S304H: CuSO4 4 g, HCl 20 ml and ethanol 20 ml. PHILIPS XL30FEG scanning elec￾tron microscopy (SEM) was employed to analyze the fractographs of impact specimens. The specimens used for EBSD investigation were obtained from the joints in location A and B (Fig. 5), and a suitable surface finished for EBSD was obtained by applying the mechanical polishing followed by jet electro-polishing in 6% solu￾tion of perchloric acid in methanol. The schematic diagram of the specimen in EBSD system was displayed in Fig. 6, in which ND, RD and TD stood for normal, rolling and transverse direction respectively. EBSD micrographs were taken on the TD-RD plane so that the RD is horizontal and the ND is vertical. Fig. 1. Schematic diagram of heat treatment processes of T92 and S304H steels. Table 1 Chemical compositions of the base materials and filler material samples (wt.%). Material C Cr Mo V Nb Ni Mn P S Si N Al W B Cu T92 sample 0.11 8.76 0.36 0.21 0.059 0.25 0.46 0.016 0.002 0.39 0.044 0.01 1.63 0.0033 / S304H sample 0.09 18.38 / 0.040 0.49 8.91 0.82 0.033 0 0.025 0.11 0.009 / 0.004 2.96 ERNiCr-3 filler 0.030 20.0 / / 2.40 72.5 2.90 / 0.001 0.04 / / / / 0.01 Fig. 2. Schematic diagram of welding process of T92/S304H dissimilar materials joints. Fig. 3. Schematic diagram of post welding heat treatment process of T92/S304H dissimilar materials joints. Fig. 4. Schematic diagram of T92/S304H dissimilar materials joints. 2764 J. Cao et al. / Materials and Design 32 (2011) 2763–2770

J. Cao et aL/ Materials and Design 32 (2011)2763-2770 not present the obvious oriented feature and most of the grains in T92 WM S304H T92 CGHAZ are not conducive to epitaxial growth However while in the weld metal zone. most of coarse colum- nar grains take on green and the direction where these grains grow up is from the interface between T92HAZ and weld metal to the centre part of weld metal zone( Fig. 8b). Furthermore, it can be A 边B seen from Fig 9b that the maximum oriented intensity value of weld metal is about 16, which indicates that the weld metal pre- sents the oriented feature, i. e grains in weld metal zone grow along the transverse direction of the joints after welding. Fig 10 shows the grain oriented distribution map of weld metal near S304H HAZ side. The black line in Fig 10 illustrates the inter Fig. 5. The locations of specimens for EBSD ce between S304H HAZ and weld metal. Similarly, the color of grain in Fig. 10 stands for its orientation. On the left side of the interface, we can estimate that the columnar grains in weld metal zone have similar colors. Furthermore, the trend that grains grow up is from the interface between S304H HAZ and weld metal to the centre part of weld metal zone. In addition, it can be seen from Tested surface RD intensity value of weld metal is about 16 indicating weld metal presents the oriented feature. That is to say, the weld metal near S304H HAZ grow up along the transverse direction of the joint after welding, which is according with the orientation test results of TD weld metal near the t92 HAZ side(fig. 9b). Therefore, weld metal part takes on oriented distribution along transverse direction of the Fig. 6. The location of specimen in EBSD system. On the right side of the interface, S304H HAZ, however, is in contrast with weld metal. The colors of equiaxed grains are ran- 3. Results and discussion domly distributed. In addition, it can be seen from Fig. 11b that the maximum oriented intensity value of S304H HAZ near the weld 3. 1. Microstructure metal is about only 4.5, which shows grains in S304H HAZ do not display oriented feature during the welding process and most of The microstructure of T92 base material is shown in Fig. 7a. It the grains in S304H HAZ are not conducive to epitaxial growth. consists of fully tempered martensite where carbide particles M23C6, MC)precipitated from the lath martensites and the prior 3. 2. Mechanical properties austenite grain boundaries during the tempering process[20] The microstructure of S304H base material shown in Fig. 7b con- 3. 1. Tensile property sts of equiaxed austenitic grains with average grain size of Results of the transverse tensile test are listed in table 2. from 12 Hm What's more, M23C6, NbCrN and NbX precipitates are dis- which the tensile strength values of two group specimens are tributed within the austenite grains, which can assure excellent respectively 689 and 677 MPa, both are higher than that of the strength for S304H steel at room temperature 21. The INCONEL ASME T92 and $304H standards. So the tensile strength of the 82/182 weld metal exhibits a fully austenitic microstructure and joints can ensure the USC boilers safe service. In addition, the frac- the shape of grains is dendritic(Fig. 7c). T92 heat affected zone ture of the joints is located in T92 CGHAZ, which indicates tensile (HAZ) consists of tempered martensite. Furthermore, the region strength of T92 CGHAZ part is relatively weak. in the t92 heat affected zone(HAz) near the weld metal is As can be seen from the Fig. 7, although the microstructures of coarse-grained heat affected zone(cghaz) with an average grain T92 base material and its haz are both tempered martenste, aver size of 25 um(Fig. Sa). Beyond this region, the fine-grained heat age grain sizes of them are different. Average grain size of T92 affected zone (FGHAz) with an average grain size of 9 um CGHAZ is larger than that of t92 base material and any part in ( Fig. 7d)is seen adjacent to the unaffected T92 base metal. In addi- T92 CGHAZ with the same volume as t92 base material has less tion, a lot of carbide particles(M23C6, MC)precipitate and distrib- quantity of interface between grains than that in t92 base mate- ute within the grains after PWHT, which improves the strength and rial. It is well known that interface between grains can effectively toughness of T92 HAz. The S304H HAZ has a large equiaxed obstruct the movement of dislocation and improve the strength of stenitic structure with a grain size of 22 Hm(Fig. 10). Beside material [22-24]. Thus, the strength in T92 CGHAZ is weaker than S304H HAZ/weld metal are both bonded well and the fusing line creased. While in T92 FGHAZ, with the decreasing al ae c haz de- these, the two interfaces including t92 HAZ/weld metal and that in T92 base material, inducing the strength in CGhaz de- is very clear, seen in Fig. 7e and f. Therefore the microstructural size, any part in t92 FGHAZ with the same volume as T92 base characteristics of the whole joints can be obviously reflected by material has more quantity of interface between grains than that in T92 base material. Hence the strength in T92 FGHAZ is stronge Fig 8 shows the grain oriented distribution maps of T92CGHAz than that in t92 base material and, correspondingly, the strength and the weld metal near T92 CGHAZ side. the color of grain inin T92 FGHAZ is improved. Similarly, compared with the $304H Fig 8 stands for its orientation. The observation in Fig 8a demon- base material, average grain size of S304H HAZ(Fig. 10)is larger strates that t92 CGHAZ is made up of much equiaxed grains whose than that of S304H base material and any part in S304H HAZ with colors are randomly distributed. Moreover, combined with the in- the same volume as S304H base material has less quantity of inter verse pole figure map of T92 CGHAZ (Fig. 9a)it is well known that face between grains than that in $304H base material, which re- the maximum oriented intensity value of T92 CGHAZ near the weld sults in a consequent decrease of strength in S304H HAZ. In metal is about only 3. This means that the grains in t92 CGHAZ do addition, although S304H base material has almost the same grain

3. Results and discussion 3.1. Microstructures The microstructure of T92 base material is shown in Fig. 7a. It consists of fully tempered martensite where carbide particles (M23C6, MC) precipitated from the lath martensites and the prior austenite grain boundaries during the tempering process [20]. The microstructure of S304H base material shown in Fig. 7b con￾sists of equiaxed austenitic grains with average grain size of 12 lm. What’s more, M23C6, NbCrN and NbX precipitates are dis￾tributed within the austenite grains, which can assure excellent strength for S304H steel at room temperature [21]. The INCONEL 82/182 weld metal exhibits a fully austenitic microstructure and the shape of grains is dendritic (Fig. 7c). T92 heat affected zone (HAZ) consists of tempered martensite. Furthermore, the region in the T92 heat affected zone (HAZ) near the weld metal is coarse-grained heat affected zone (CGHAZ) with an average grain size of’ 25 lm (Fig. 8a). Beyond this region, the fine-grained heat affected zone (FGHAZ) with an average grain size of 9 lm (Fig. 7d) is seen adjacent to the unaffected T92 base metal. In addi￾tion, a lot of carbide particles (M23C6, MC) precipitate and distrib￾ute within the grains after PWHT, which improves the strength and toughness of T92 HAZ. The S304H HAZ has a large equiaxed austenitic structure with a grain size of 22 lm (Fig. 10). Beside these, the two interfaces including T92 HAZ/weld metal and S304H HAZ/weld metal are both bonded well and the fusing line is very clear, seen in Fig. 7e and f. Therefore, the microstructural characteristics of the whole joints can be obviously reflected by Fig. 7. Fig. 8 shows the grain oriented distribution maps of T92CGHAZ and the weld metal near T92 CGHAZ side. The color of grain in Fig. 8 stands for its orientation. The observation in Fig. 8a demon￾strates that T92 CGHAZ is made up of much equiaxed grains whose colors are randomly distributed. Moreover, combined with the in￾verse pole figure map of T92 CGHAZ (Fig. 9a) it is well known that the maximum oriented intensity value of T92 CGHAZ near the weld metal is about only 3. This means that the grains in T92 CGHAZ do not present the obvious oriented feature and most of the grains in T92 CGHAZ are not conducive to epitaxial growth. However, while in the weld metal zone, most of coarse colum￾nar grains take on green and the direction where these grains grow up is from the interface between T92HAZ and weld metal to the centre part of weld metal zone (Fig. 8b). Furthermore, it can be seen from Fig. 9b that the maximum oriented intensity value of weld metal is about 16, which indicates that the weld metal pre￾sents the oriented feature, i.e. grains in weld metal zone grow up along the transverse direction of the joints after welding. Fig. 10 shows the grain oriented distribution map of weld metal near S304H HAZ side. The black line in Fig. 10 illustrates the inter￾face between S304H HAZ and weld metal. Similarly, the color of grain in Fig. 10 stands for its orientation. On the left side of the interface, we can estimate that the columnar grains in weld metal zone have similar colors. Furthermore, the trend that grains grow up is from the interface between S304H HAZ and weld metal to the centre part of weld metal zone. In addition, it can be seen from the inverse pole figure map in Fig. 11a that the maximum oriented intensity value of weld metal is about 16 indicating weld metal presents the oriented feature. That is to say, the weld metal near S304H HAZ grow up along the transverse direction of the joint after welding, which is according with the orientation test results of weld metal near the T92 HAZ side (Fig. 9b). Therefore, weld metal part takes on oriented distribution along transverse direction of the joints. On the right side of the interface, S304H HAZ, however, is in contrast with weld metal. The colors of equiaxed grains are ran￾domly distributed. In addition, it can be seen from Fig. 11b that the maximum oriented intensity value of S304H HAZ near the weld metal is about only 4.5, which shows grains in S304H HAZ do not display oriented feature during the welding process and most of the grains in S304H HAZ are not conducive to epitaxial growth. 3.2. Mechanical properties 3.2.1. Tensile property Results of the transverse tensile test are listed in Table 2, from which the tensile strength values of two group specimens are respectively 689 and 677 MPa, both are higher than that of the ASME T92 and S304H standards. So the tensile strength of the joints can ensure the USC boilers safe service. In addition, the frac￾ture of the joints is located in T92 CGHAZ, which indicates tensile strength of T92 CGHAZ part is relatively weak. As can be seen from the Fig. 7, although the microstructures of T92 base material and its HAZ are both tempered martenste, aver￾age grain sizes of them are different. Average grain size of T92 CGHAZ is larger than that of T92 base material and any part in T92 CGHAZ with the same volume as T92 base material has less quantity of interface between grains than that in T92 base mate￾rial. It is well known that interface between grains can effectively obstruct the movement of dislocation and improve the strength of material [22–24]. Thus, the strength in T92 CGHAZ is weaker than that in T92 base material, inducing the strength in CGHAZ de￾creased. While in T92 FGHAZ, with the decreasing of average grain size, any part in T92 FGHAZ with the same volume as T92 base material has more quantity of interface between grains than that in T92 base material. Hence the strength in T92 FGHAZ is stronger than that in T92 base material and, correspondingly, the strength in T92 FGHAZ is improved. Similarly, compared with the S304H base material, average grain size of S304H HAZ (Fig. 10) is larger than that of S304H base material and any part in S304H HAZ with the same volume as S304H base material has less quantity of inter￾face between grains than that in S304H base material, which re￾sults in a consequent decrease of strength in S304H HAZ. In addition, although S304H base material has almost the same grain Fig. 6. The location of specimen in EBSD system. Fig. 5. The locations of specimens for EBSD. J. Cao et al. / Materials and Design 32 (2011) 2763–2770 2765

J. Cao et al Materials and Design 32(2011)2763-2770 (a) 20un 120um 20 WM S304H WM T92 20m Fig. 7. Metallographic structures of T92/S304H dissimilar materials joints, exhibiting (a)T92 base material. (b)S304H base material, (c)weld metal. (d)T92 fine grained HAZ (FGHAZ) (e)interface between T92 HAZ and weld metal, and (f) interface between S304H HAZ and weld metal. 100u Fig 8. EBSD grain orientation map of (a)T92 CGHAZ, and (b) weld metal near T92 CGHAZ side. size as t92 base material, the strengths of them are different Due solid solution strengthening and dispersion strengthening effects to high alloy elements content in S304H austenitic steel (Table 1), caused by these alloy elements are stronger in $304H base material

size as T92 base material, the strengths of them are different. Due to high alloy elements content in S304H austenitic steel (Table.1), solid solution strengthening and dispersion strengthening effects caused by these alloy elements are stronger in S304H base material Fig. 8. EBSD grain orientation map of (a) T92 CGHAZ, and (b) weld metal near T92 CGHAZ side. Fig. 7. Metallographic structures of T92/S304H dissimilar materials joints, exhibiting (a) T92 base material, (b) S304H base material, (c) weld metal, (d) T92 fine grained HAZ (FGHAZ), (e) interface between T92 HAZ and weld metal, and (f) interface between S304H HAZ and weld metal. 2766 J. Cao et al. / Materials and Design 32 (2011) 2763–2770

J. Cao et aL/ Materials and Design 32(2011)2763-2770 6 06301 101 p13001 0 Fig 9. Inverse pole figure maps of (a) T92 CGHAZ, and(b) weld metaL Tensile test results of T92/S304H dissimilar materials joints. ple no Tensile strength(os MPa) T92 CGHAZ ASME S304H ing effect caused by grain boundary is weakest for its coarse den- dritic austenitic grain, orientation distribution of grains in weld metal zone, giving rise to higher strength than T92 CGHAZ in trans verse direction of the joints( Figs. 8 and 10). As a result, the fracture location is taken place at the t92CGHAZ finally Fig. 10. EBSD grain orientation map of weld metal near S304H HAZ side. 3. 2. 2. Impact toughness The impact test results of the joints are given in Fig. 12. It is than that in T92 base material, indicating that S304H base material clear that the average impact strength value of $304H base mate has higher strength than T92 base material [21]. Likewise, as the rial is the highest. In contrast, the average impact strength value average grain size of S304H HAz is similar to that of T92 CGHAz of weld metal is the lowest. Furthermore, the average impact and the strengthening effect caused by alloy elements for S304H strength value of weld metal is 57 J which is higher than ASME HAZ is stronger, in this case, S304H HAZ has higher strength than standard value(41 J), so the impact strength of the joints is qual T92 CGHAZ. With regard to weld metal, although the strengthen- ified. Meanwhile, the average impact strength values of T92 CGHAZ pole figure maps of (a)weld metal, and(b)S304H HAZ

than that in T92 base material, indicating that S304H base material has higher strength than T92 base material [21]. Likewise, as the average grain size of S304H HAZ is similar to that of T92 CGHAZ and the strengthening effect caused by alloy elements for S304H HAZ is stronger, in this case, S304H HAZ has higher strength than T92 CGHAZ. With regard to weld metal, although the strengthen￾ing effect caused by grain boundary is weakest for its coarse den￾dritic austenitic grain, orientation distribution of grains in weld metal zone, giving rise to higher strength than T92 CGHAZ in trans￾verse direction of the joints (Figs. 8 and 10). As a result, the fracture location is taken place at the T92CGHAZ finally. 3.2.2. Impact toughness The impact test results of the joints are given in Fig. 12. It is clear that the average impact strength value of S304H base mate￾rial is the highest. In contrast, the average impact strength value of weld metal is the lowest. Furthermore, the average impact strength value of weld metal is 57 J which is higher than ASME standard value (>41 J), so the impact strength of the joints is qual￾ified. Meanwhile, the average impact strength values of T92 CGHAZ Fig. 10. EBSD grain orientation map of weld metal near S304H HAZ side. Fig. 9. Inverse pole figure maps of (a) T92 CGHAZ, and (b) weld metal. Fig. 11. Inverse pole figure maps of (a) weld metal, and (b) S304H HAZ. Table 2 Tensile test results of T92/S304H dissimilar materials joints. Sample no. Tensile strength (rs, MPa) Rupture position 1 689 T92 CGHAZ 2 677 T92 CGHAZ ASME T92 P620 / ASME S304H P655 / J. Cao et al. / Materials and Design 32 (2011) 2763–2770 2767

J. Cao et al Materials and Design 32(2011)2763-2770 and S304H HAZ are lower than that of their base materials respec- ively, while average impact strength value of T92 FGHAZ is higher than that of t9 2 base material Fig. 13 shows the impact fractograph of the joints. The fracture surfaces of t92 base material and t92 FGHaZ. as well as S304H base material exhibit ductile fracture characteristic and consist of fiber zone and radiation zone, especially the fiber zone occupys a big proportion. Beside these, there are a large number of small- sized dimples with different sizes and depths on their fracture su faces, which displays excellent plasticity of them. As for the S304H HAZ and t92 CGhaz, their fracture modes both belong to quasi- cleavage fracture. Compared with the fracture surfaces of two base materials, a larger brittle area is seen on their fracture surfaces. Thus, it can be inferred that the toughness of two base materials is better than that of two hazs. furthermore the amount of tear ridge around dimple in Fig. 13d is much more than that in Fig. 13e, indicating that the toughness of T92 CGHAZ is weaker BM FGHAZ CGHAZ WM S304H S304H than that of s304H haz in addition it can be seen from the frac- Fig. 12. Impact test results of T92/S304H dissimilar materials joints. ture surfaces of two base materials and their hazs that there are some precipitates in the dimples. For T92 base material and it's (b) (d) ( Fig. 13. SEM fractographs of (a)T92 base material, (b)92 FGHAZ (c)S304H base material, (d)S304H HAZ, (e)T92 CGHAZ, and( O weld metal

and S304H HAZ are lower than that of their base materials respec￾tively, while average impact strength value of T92 FGHAZ is higher than that of T92 base material. Fig. 13 shows the impact fractograph of the joints. The fracture surfaces of T92 base material and T92 FGHAZ, as well as S304H base material exhibit ductile fracture characteristic and consist of fiber zone and radiation zone, especially the fiber zone occupys a big proportion. Beside these, there are a large number of small￾sized dimples with different sizes and depths on their fracture sur￾faces, which displays excellent plasticity of them. As for the S304H HAZ and T92 CGHAZ, their fracture modes both belong to quasi￾cleavage fracture. Compared with the fracture surfaces of two base materials, a larger brittle area is seen on their fracture surfaces. Thus, it can be inferred that the toughness of two base materials is better than that of two HAZs. Furthermore, the amount of tear ridge around dimple in Fig. 13d is much more than that in Fig. 13e, indicating that the toughness of T92 CGHAZ is weaker than that of S304H HAZ. In addition, it can be seen from the frac￾ture surfaces of two base materials and their HAZs that there are some precipitates in the dimples. For T92 base material and it’s Fig. 12. Impact test results of T92/S304H dissimilar materials joints. Fig. 13. SEM fractographs of (a) T92 base material, (b) T92 FGHAZ, (c) S304H base material, (d) S304H HAZ, (e) T92 CGHAZ, and (f) weld metal. 2768 J. Cao et al. / Materials and Design 32 (2011) 2763–2770

J. Cao et aL/ Materials and Design 32(2011)2763-2770 IAZ, EDX results show the precipitates on the fracture surfaces of process, the grain size in the HAzs of 192 and S304H have both them are composed of Cr, Fe, W and c elements( Fig. 14a). So it can changed, but orientation factor has not increased since no orienta- inferred that the type of precipitates is a M23C6 particle[1]. For tion was formed. Thus, the toughness of them is decided by grain S304H base material and it's HAZ, EDX results show the precipi- size. For T92 CGHAZ and S304H Haz, due to larger grain size than tates on the fracture surfaces of them consist of two types particles. their respective base materials, correspondingly, the toughness of One is composed of Cr, Fe, Ni and C elements(Fig. 14b), and the them decreases. while in t92 FGHAZ, the decrease of grain size other is composed of Nb, Cr and n elements( Fig. 14c). Thus, it leads to improvement of its toughness. As a result, the toughness can be inferred that the two types precipitates are M23 C6 and of T92 FGHAZ is better than that of T92 base material. However. NbCrN respectively [21]. Since the precipitates on the fracture sur- as for weld metal, during the welding process, not only are the faces of two base materials are the same with that of their own coarse austenitic grains produced, but also the orientation is HAZs, it also implies that there is no any new type precipita formed in the transverse direction of the joints. In this case, under ppearing in the two HAZs after welding. The fracture surface con- the impact force which is perpendicular to the transverse direction sists of many dissociation surfaces at weld metal part, has been of the joints, crack is formed and easily expands along the grain ggested to be the typical brittle fracture feature( Fig. 13f). Hence, boundaries of coarse austenitic grains causing brittle fracture. Con the toughness of weld metal part is the weakest, which is accor- sequently, weld metal part of the joints has the lowest toughness dance with impact test results(Fig. 12). The changes of toughness at different parts of the joints can be ell explained by means of Cottrell-Petch theory [25 According 3.3. Microstructural evolution mechanism to the classic Cottrell-Petch theory, when the temperature keeps constant, two factors, grain size and average orientation factor, Based on former analyses, it is clear that GTAW has taken effect are responsible for the toughness of metals and alloys. Moreover, on the microstructures of HAZs and weld metal during the welding with the increase of grain size or average orientation factor, the process. At the initial stage of welding, the filler material under toughness of metals and alloys will be weakened, and vice versa. goes melting. When the welding process is finished, melted filler Based on previously microstructural analyses, during the welding material begins to crystallize to form weld metal zone. According Mo 0.0 2000.002.00400 00100012.001400 2000.002.004.006.0080010001200140 Energy-Kev Energy-Kev 2000.002004006.00800100012.00140016001800 Kev Fig. 14. EDX spectrums of (a) precipitates on the fracture surfaces of 192 base material and it's HAZ; ( b)and (c) precipitates on the fracture surfaces of S304H base material

HAZ, EDX results show the precipitates on the fracture surfaces of them are composed of Cr, Fe, W and C elements (Fig. 14a). So it can be inferred that the type of precipitates is a M23C6 particle [1]. For S304H base material and it’s HAZ, EDX results show the precipi￾tates on the fracture surfaces of them consist of two types particles. One is composed of Cr, Fe, Ni and C elements (Fig. 14b), and the other is composed of Nb, Cr and N elements (Fig. 14c). Thus, it can be inferred that the two types precipitates are M23C6 and NbCrN respectively [21]. Since the precipitates on the fracture sur￾faces of two base materials are the same with that of their own HAZs, it also implies that there is no any new type precipitate appearing in the two HAZs after welding. The fracture surface con￾sists of many dissociation surfaces at weld metal part, has been suggested to be the typical brittle fracture feature (Fig. 13f). Hence, the toughness of weld metal part is the weakest, which is accor￾dance with impact test results (Fig. 12). The changes of toughness at different parts of the joints can be well explained by means of Cottrell–Petch theory [25]. According to the classic Cottrell–Petch theory, when the temperature keeps constant, two factors, grain size and average orientation factor, are responsible for the toughness of metals and alloys. Moreover, with the increase of grain size or average orientation factor, the toughness of metals and alloys will be weakened, and vice versa. Based on previously microstructural analyses, during the welding process, the grain size in the HAZs of T92 and S304H have both changed, but orientation factor has not increased since no orienta￾tion was formed. Thus, the toughness of them is decided by grain size. For T92 CGHAZ and S304H HAZ, due to larger grain size than their respective base materials, correspondingly, the toughness of them decreases. While in T92 FGHAZ, the decrease of grain size leads to improvement of its toughness. As a result, the toughness of T92 FGHAZ is better than that of T92 base material. However, as for weld metal, during the welding process, not only are the coarse austenitic grains produced, but also the orientation is formed in the transverse direction of the joints. In this case, under the impact force which is perpendicular to the transverse direction of the joints, crack is formed and easily expands along the grain boundaries of coarse austenitic grains causing brittle fracture. Con￾sequently, weld metal part of the joints has the lowest toughness value. 3.3. Microstructural evolution mechanism Based on former analyses, it is clear that GTAW has taken effect on the microstructures of HAZs and weld metal during the welding process. At the initial stage of welding, the filler material under￾goes melting. When the welding process is finished, melted filler material begins to crystallize to form weld metal zone. According Fig. 14. EDX spectrums of (a) precipitates on the fracture surfaces of T92 base material and it’s HAZ; (b) and (c) precipitates on the fracture surfaces of S304H base material and it’s HAZ. J. Cao et al. / Materials and Design 32 (2011) 2763–2770 2769

J. Cao et al Materials and Design 32(2011)2763-2770 to non-spontaneous nucleation theory of welding crystallography Acknowledgements [26-28]. firstly, melted filler material crystallizes on the base of djacent base material, and then produces epitaxial growth to form The work was supported by both National Natural Science coarse columnar crystals under the effect of temperature field. Foundation of China(Grant 50871076)and Shanghai Leading Aca- Thus, the direction where the grains in weld metal zone grow up demic Discipline Project(Project Number: B113) decided by variation of temperature field. Taking into account lat the direction where temperature decreases is from the inter- References face between base material and weld metal towards the e middle part of the weld metal zone, the melted weld metal crystallizes [1 Ennis P). zielinska-lipiec A, Wachter o, Czyrska-filemonowicz A. tion. In other words, the grains in weld metal zone grow up from (2 Kodak &. Hemas aw Kielbus at s uastrcture stability of h the two interface sides towards the middle part of weld metal zon respectively and result in orientation distribution along the trans- [3I Kimura K, Sawada K Effect of stress on the creep deformation of ASME Grade transverse tensile strength of weld metal is improved for its orien- 1492 anew. ande nberghe B. Hahn B, Heuser H. Jochum C T/P23.24,911 and erse direction of the joint, seen in Figs. 8 and 10. Therefore, the tation in this direction. While in the two HAZs, because of narrow I5)Haarmann K vaillant JC, Vandenberghe B, Bendick W, Arbab A. The T91/p91 perature gradient change. This suggested that orientation distribu- [6] Richardot D, Vaillant JC, Arbab A, Bendick w. The T9 tions of grains in the two HAZs are random. On the other hand, for T92 HAZ, since the temperature is higher than its initial austenitic transition temperature(Ac, during the welding process, tempered martensite begins transforming into [81 Abe F, Horiuchi T, Taneike M, Sawada K. Stabilization of martensitic ustenite. Furthermore, in t92 CGHaZ, the temperature even ex- are in advanced 9Cr steel during creep at high temperature. ceeds its fully austenitic transition temperature (Ac3), which [91 makes austenitic grains grow up and results in coarsening. while in T92 FGHAZ, as the temperature is between Aci and Ac3. the fine 101 Hu ZF, Yang zG. An investigation of the embrittlement inX20CrMov121power austenitic grains can be obtained. After welding, with the continu exposure at elevated te ous cooling, austenitic grains transform into lath-shaped martens- 11 Hu ZF, Yang zG. Identification of the precipitates by TEM and EDs in ite. Finally, t92 CGHAZ has martensitic structure with coarse grain ong-term service at elevated temperature. J. Mater En 2003:12(1):106-11 size and t92 FGHAZ has martensitic structure with fine grain size. [121 Serrea l, Vogt JB. Mechanic rties of a 316L/191 weld joint tested in 30(9):3776-83. austenite structure. Thus, microstructural evolution leads to the [14] Sawaragi Y, Otsuka N. of a new18-8 austenitic steel tube (Super304H) for fossil fired boilers after service exposure with high elevated variation of mechanical properties of the joints. [15 Dischi A, Kenny JM, Mecozzi MG. Development of high nitrogen low nickel aragi Y. Properties and experiences of a new 18-8 austenitic 4. Conclusions stainless steel tube(0. 1C-18Cr-9Ni-3Cu-Nb, N) for boiler tube application. The investigations performed on the T92/S304H dissimilar [171 Yasutaka N, Mitsuo M, Hirokazu O, Masaaki L Effect of grain size on creep fatigue properties of 18Cr-9Ni-3Cu-Nb-N steel under uniaxial and torsional materials joints have led to the following conclusions [18] ASTM E8-04. Standard test methods for tension testing of metallic materials. materials joints obtained by GTAW process can meet the USC 19 ASTM E23-02a. Standard test methods for notched bar impact testing of Tensile strength and impact toughness of T92 /S304H dissimilar boiler's requirements. Furthermore the part of the joints with [201 Maruyama k, Sawada k Koile J Strengthening mechanisms of creep resistant relatively weak tensile strength was T92 CGHAZ, while the part (21) Huang YZ, Titchmarsh JM. TEM investiga intergranular stress corrosion cracking for 316 stainless steel in PwR environment. Acta Mater weld metal 2. The coarse tempered martensitic structure of T92 CGHAZ makes [22] Kashyap KT, Chandrashekar T. Effects and mechanisms of grain refinement in the strengthening effect caused by grain boundary decreased [23] Sleboda T. Muszka K, Majta I. Hale P. Wright RN. The possibilities of and results in relatively weak tensile strength in T92 CGHAZ echanical property control in fine grained structures. J Mater Process Tech 3. The weak toughness of weld metal is attributed to its 2006;177(1-3)461-4. dendritic austenitic structure. Furthermore, the impact [24] Herring DH. Grain size and its influence on materials properties. Heat Treat graph of weld metal takes on typical brittle re [25] He Z]. Mechanical properties of metal materials. 1st ed. Beijing: Metallurgical characteristics 4. Grains in weld metal zone grow up respectively from the two 126 Ne son W, Ltppoid j c.mis My.nature and evol ution ot the tusion bounday interface sides towards the middle part of weld metal resulting in its orientation distribution in transverse direction of the [27] Savage wF, Lundin CD, Aronson AH. Weld metal solidification med joints.Weld metal. due to its orientation distribution, has v)Yyhermoch Acta 1996: 280: 303-17 ystallization kinetics of amorphous alloys. higher tensile strength than T92 CGHAZ

to non-spontaneous nucleation theory of welding crystallography [26–28], firstly, melted filler material crystallizes on the base of adjacent base material, and then produces epitaxial growth to form coarse columnar crystals under the effect of temperature field. Thus, the direction where the grains in weld metal zone grow up is decided by variation of temperature field. Taking into account that the direction where temperature decreases is from the inter￾face between base material and weld metal towards the e middle part of the weld metal zone, the melted weld metal crystallizes along this direction, i.e. the maximum temperature gradient direc￾tion. In other words, the grains in weld metal zone grow up from the two interface sides towards the middle part of weld metal zone respectively and result in orientation distribution along the trans￾verse direction of the joint, seen in Figs. 8 and 10. Therefore, the transverse tensile strength of weld metal is improved for its orien￾tation in this direction. While in the two HAZs, because of narrow width in transverse direction of the joints, there is no obvious tem￾perature gradient change. This suggested that orientation distribu￾tions of grains in the two HAZs are random. On the other hand, for T92 HAZ, since the temperature is higher than its initial austenitic transition temperature (Ac1) during the welding process, tempered martensite begins transforming into austenite. Furthermore, in T92 CGHAZ, the temperature even ex￾ceeds its fully austenitic transition temperature (Ac3), which makes austenitic grains grow up and results in coarsening. While in T92 FGHAZ, as the temperature is between Ac1 and Ac3, the fine austenitic grains can be obtained. After welding, with the continu￾ous cooling, austenitic grains transform into lath-shaped martens￾ite. Finally, T92 CGHAZ has martensitic structure with coarse grain size and T92 FGHAZ has martensitic structure with fine grain size. However, For S304H austenitic steel, there is no martensite– austenite (M–A) transition during welding process. Only the region adjacent to the weld metal, due to welding heat effect, has coarse austenite structure. Thus, microstructural evolution leads to the variation of mechanical properties of the joints. 4. Conclusions The investigations performed on the T92/S304H dissimilar materials joints have led to the following conclusions. 1. Tensile strength and impact toughness of T92/S304H dissimilar materials joints obtained by GTAW process can meet the USC boiler’s requirements. Furthermore, the part of the joints with relatively weak tensile strength was T92 CGHAZ, while the part of the joints which revealed relatively weak toughness was weld metal. 2. The coarse tempered martensitic structure of T92 CGHAZ makes the strengthening effect caused by grain boundary decreased and results in relatively weak tensile strength in T92 CGHAZ. 3. The weak toughness of weld metal is attributed to its coarse dendritic austenitic structure. Furthermore, the impact fracto￾graph of weld metal takes on typical brittle fracture characteristics. 4. Grains in weld metal zone grow up respectively from the two interface sides towards the middle part of weld metal resulting in its orientation distribution in transverse direction of the joints. Weld metal, due to its orientation distribution, has higher tensile strength than T92 CGHAZ. Acknowledgements The work was supported by both National Natural Science Foundation of China (Grant 50871076) and Shanghai Leading Aca￾demic Discipline Project (Project Number: B113). References [1] Ennis PJ, Zielinska-lipiec A, Wachter O, Czyrska-filemonowicz A. Microstructural stability and creep rupture strength of the martensitic steel P92 for advanced power plant. Acta Mater 1997;45(12):4901–7. [2] Rodak K, Hernas A, Kielbus A. Substructure stability of highly alloyed martensitic steels for power industry. Mater Chem Phys 2003;81(2–3):483–5. [3] Kimura K, Sawada K. Effect of stress on the creep deformation of ASME Grade P92/T92 steels. Int J Mat Res 2008;99(4):395–401. [4] Vaillant JC, Vandenberghe B, Hahn B, Heuser H, Jochum C. T/P23, 24, 911 and 92: new grades for advanced coal-fired power plants – properties and experience. Int J Press Ves Pip 2008;85(1–2):38–46. [5] Haarmann K, Vaillant JC, Vandenberghe B, Bendick W, Arbab A. The T91/P91 book. 1st ed. Boulogne: Vallourec-Mannesmann tubes; 1999. [6] Richardot D, Vaillant JC, Arbab A, Bendick W. The T92/P92 book. 1st ed. Boulogne: Vallourec-Mannesmann tubes; 2000. [7] Sawada K, Kubo K, Abe F. Creep behavior and stability of MX precipitates at high temperature in 9Cr–0.5Mo–1.8W–VNb steel. Mater Sci Eng 2001;A319– 321:784–7. [8] Abe F, Horiuchi T, Taneike M, Sawada K. Stabilization of martensitic microstructure in advanced 9Cr steel during creep at high temperature. Mater Sci Eng 2004;A378:299–303. [9] Korcakova L, Hald J, Somers MA J. Quantification of Laves phase particle size in 9CrW steel. Mater Charact 2001;47(8):111–7. [10] Hu ZF, Yang ZG. An investigation of the embrittlement inX20CrMoV12.1 power plant steel after long-term service exposure at elevated temperature. Mater Sci Eng 2004;A383:224–8. [11] Hu ZF, Yang ZG. Identification of the precipitates by TEM and EDS in X20CrMoV12.1 for long-term service at elevated temperature. J. Mater Eng Perform 2003;12(1):106–11. [12] Serrea I, Vogt JB. Mechanical properties of a 316L/T91 weld joint tested in lead–bismuth liquid. Mater Des 2009;30(9):3776–83. [13] Spigarelli S, Quadrini E. Analysis of the creep behaviour of modified P91 (9Cr– 1Mo–NbV) welds. Mater Des 2002;23(6):547–52. [14] Sawaragi Y, Otsuka N. Properties of a new18–8 austenitic steel tube (Super304H) for fossil fired boilers after service exposure with high elevated temperature strength. Sumitomo Search 1994;10:56. [15] Dischi A, Kenny JM, Mecozzi MG. Development of high nitrogen low nickel 18%Cr austenitic stainless steels. J Mater Sci 2000;35:4803–8. [16] Takao K, Sawaragi Y. Properties and experiences of a new 18–8 austenitic stainless steel tube (0.1C–18Cr–9Ni–3Cu–Nb, N) for boiler tube application. Sumitomo Search 1993;10:45. [17] Yasutaka N, Mitsuo M, Hirokazu O, Masaaki I. Effect of grain size on creep￾fatigue properties of 18Cr–9Ni–3Cu–Nb–N steel under uniaxial and torsional loading. J Soc Mater Sci 2007;56(2):136–41. [18] ASTM E8-04. Standard test methods for tension testing of metallic materials. ASTM International; 2003. [19] ASTM E23-02a. Standard test methods for notched bar impact testing of metallic materials. ASTM International; 2002. [20] Maruyama K, Sawada K, Koile J. Strengthening mechanisms of creep resistant tempered martensitic steel. ISIJ Int 2001;41(6):641–53. [21] Huang YZ, Titchmarsh JM. TEM investigation of intergranular stress corrosion cracking for 316 stainless steel in PWR environment. Acta Mater 2006;54:635–41. [22] Kashyap KT, Chandrashekar T. Effects and mechanisms of grain refinement in aluminum alloys. Bull Mater Sci 2001;24(4):345–53. [23] Sleboda T, Muszka K, Majta J, Hale P, Wright RN. The possibilities of mechanical property control in fine grained structures. J Mater Process Tech 2006;177(1–3):461–4. [24] Herring DH. Grain size and its influence on materials properties. Heat Treat Doct 2005:20–2. [25] He ZJ. Mechanical properties of metal materials. 1st ed. Beijing: Metallurgical Industry Press; 1982. [26] Nelson TW, Lippold JC, Mills MJ. Nature and evolution of the fusion boundary in ferriticaustenitic dissimilar weld metals, part 1 – nucleation and growth. Weld J 1999;78(10):329–37. [27] Savage WF, Lundin CD, Aronson AH. Weld metal solidification mechanisms. Weld J 1965;44(4):175–81. [28] Kaloshkin SD, Tomilin IA. The crystallization kinetics of amorphous alloys. Thermoch Acta 1996;280:303–17. 2770 J. Cao et al. / Materials and Design 32 (2011) 2763–2770