Microstructure and mechanical properties of UNS N10003 alloy welded joints

Materials Science & Engineering A 682 (2017) 168–177

Contents lists available at ScienceDirect

Materials Science & Engineering A journal homepage: www.elsevier.com/locate/msea

Microstructure and mechanical properties of UNS N10003 alloy welded joints

crossmark

⁎

Shuangjian Chena,b, Xiang-Xi Yea, Kun Yua,b, Chaowen Lia, , Zhijun Lia, Zhong Lia, Xingtai Zhoua a b

Center for Thorium Molten Salt Reactor System, Shanghai Institute of Applied Physics, Chinese Academy of Sciences, Shanghai 201800, PR China University of Chinese Academy of Sciences, Beijing 100049, PR China

A R T I C L E I N F O

A BS T RAC T

Keywords: Nickel based alloy Welding M6C Mechanical performance Strengthening mechanism

Microstructure and mechanical performance of the welded joints of UNS N10003 alloy have been investigated in this work. Primary precipitates in base metal and eutectic precipitates in Heat Affected Zone (HAZ) and weld metal have been characterized and identified as M6C type. The hardness value of HAZ (Eutectic zone) is significantly higher than the rest of joint, including weld and base metal. Tensile tests suggested the welded joints possess a very stable mechanical performance at elevated temperature from 650 °C to 725 °C. Moreover all the tensile samples were fractured in base metal, indicating that the eutectic carbides have no adverse effects on the short-time mechanical performance of joint. The fine carbides, acting as dispersion strengthening in weld metal, are main contribution to enhance the hardness and strength of weld. The good mechanical performance of HAZ is ascribed to the presence of eutectic carbides and twins.

1. Introduction Molten salt reactor (MSR), which uses molten salt as coolant and nickel base alloy as metallic structural material, is one of the generation IV nuclear reactors [1–3]. Hastelloy N (ASME designated as UNS N100003) alloy, as a representative solid solution strengthened Ni-MoCr alloy, was developed by Oak Ridge National Laboratory (ORNL) in 1950s specially as a structural material for Molten Salt Reactor Experiment (MSRE) due to its high molten salt corrosion resistance and excellent high temperature strength [4–8]. However, some challenges still exist in nuclear application of Hastelloy N, one of which is the welding issues of this alloy. Since the weldment is commonly considered as the weakest area of structural components, it is essential to study the weldability and evaluate the microstructure and mechanical performance of welded joints of Hastelloy N. The weldability of Hastelloy N alloy was first studied by using Gas Tungsten Arc Welding (GTAW) in 1960s, and a good weldability and mechanical performance were obtained to this alloy [9–11]. McCoy [12] studied hot cracking in the welded joint with manual-TIG in a heavily restrained condition, and found the cracking was associated with the segregation of chemical elements in the HAZ. Additionally, some non-equilibrium intermetallic phases were formed in the heataffected zone (HAZ) when thermal cycle peak temperatures > 1300°C and microprobe analyses showed the brittle eutectic-type structure had a different composition from that of the matrix [13,14]. To the authors'

⁎

knowledge, the types of precipitates in the welded joint are still subjected to debate. D.Bhattacharyya et al [15] characterized the precipitates in different regions of the fillet weld of Ni-Mo-Cr-Si alloy, and found that the primary carbides in base metal and eutectic carbides in HAZ possess similar chemical composition closer to Ni2(Mo,Cr)4(Si,C) and Ni3(Mo,Cr)3(Si,C) by transmission electron microscopy equipped with energy dispersive spectrometer (TEM-EDS). The results could be better convinced if another identification method is applied since content of C cannot be measured accurately by TEM-EDS. He [16] studied the microstructure of eutectic carbides of Ni-Mo-Cr alloy through simulated heat-affected zone (HAZ) thermal cycle treatment and considered the precipitates in HAZ as M3C2 but without taking account of the type of precipitate in weld and base metal. In addition, the effects of precipitate carbides on the mechanical performance of welded joint and strengthening mechanism of welded joint are lacking in-deep understanding. Yang [17] investigated the influence of eutectic carbides on the mechanical performance of Ni-17Mo-6Cr alloy by using Gleeble simulator and found no adverse effect brought by eutectic carbides. On contrary, hot cracks caused by eutectic carbides were detected by Jiang [13] when Ni-Mo-Cr alloy with high silicon was heated to more than 1335°C. These contradictory conclusions cannot provide direct application guidance to a real welded joint of UNS N10003 alloy, further related experiments are needed to evaluate the effect of eutectic carbides on the mechanical performance of joints.

Corresponding author. E-mail address: [email protected] (C. Li).

http://dx.doi.org/10.1016/j.msea.2016.10.122 Received 18 August 2016; Received in revised form 29 October 2016; Accepted 31 October 2016 Available online 09 November 2016 0921-5093/ © 2016 Elsevier B.V. All rights reserved.

Materials Science & Engineering A 682 (2017) 168–177

S. Chen et al.

With the above problems in mind, to further accurately determine the types of precipitates in different regions of the welded joint, several analytical methods were used in this work, such as electro-probe micro analyzer (EPMA), electron backscattered diffraction (EBSD), X-ray diffraction (XRD). These methods were employed in a combined way to characterize and analyze the precipitates of as-welded joint of GH3535 alloy which belongs to the group of UNS N10003 alloy. In view of the narrow region of eutectic precipitates in the welded joint, Gleeble simulator was applied to simulate heat-affected zone (HAZ) thermal cycle treatment to produce a large sample with the same microstructure as the eutectic region near fusion line. Besides, we attempted to clarify the dependence of the eutectic precipitates on the hardness and tensile properties of the real welded joint and the strengthening mechanism of the weld metal and HAZ in the welded joint. The tensile tests were carried out at three typical service temperatures for TMSR, namely, 650°C, 700°C and 725°C. The results can substantially contribute to the understanding of microstructure and mechanical properties of welded joint involved in application of this alloy.

Table 2 Welding parameters.

2. Experimental procedure

Layer (Each side)

Current (A)

Welding speed (mm/ min)

Pulse frequency (Hz)

Peak pulse duration (%)

Base pulse value/ peak pulse value (%)

Gas flow rate (L/ min)

1 2–6 7

230 240 230

90 110 100

2.5 2.5 2.5

50 50 50

50 50 50

15 15 15

Fig. 1. Macroscopic morphology of the GH3535 welded joint.

Automatic GTAW with pulsed current was employed as the welding method in this work. ERNiMo-2 filler wire with diameter of 1.2 mm was used as welding consumables and the GH3535 alloy plates with thickness of 16 mm were applied. The nominal chemical compositions for the parent alloy and filler wire are listed in Table 1. The GH3535 alloy plates in solution annealed state were procured from Special Steel Shares Co., LTD (China). Before welding, the plate was firstly cut into two test pieces with a sizes of 300 mm×100 mm×16 mm, then in each of which double V grooves with angle of 60° were prepared. The grooves and adjacent plates in the range of 20 mm were cleaned thoroughly. High purity Argon (99.99%) was used as welding shielded gas during the whole welding process. The optimized welding parameters used in this experiment have been qualified by the American Society of Mechanical Engineers (ASME) section IX [18] as presented in Table 2. The root pass and the final pass were inspected by visual inspection, and the surface morphology of the welded joint is shown in Fig. 1. It can be observed that a good appearance of weld was achieved without visual imperfection. Then the weld was inspected by penetration testing and x-ray, no weld defects were detected. To characterize the microstructure, the cross-sections of welded joints were polished with 0.5 µm diamond paste after series of grinding procedures, and then etched with solutions including 70cc H2O, 10ccHCl mixed with 10gCuCl2 for 30 seconds at room temperature (RT). Microstructure characterization and analysis were carried out by ZEISS Axio Cam optical microscope (OM) and ZEISS LEO 1530VP SEM equipped with an Oxford EBSD system. The measurements of chemical compositions were conducted by EPMA and TEM-EDS. The polished specimens of welded joint were conducted on ZHV 30 micro Vickers under load of 500gf to measure the hardness values which were calculated automatically after measuring diagonal length of indentation on the machine software. Tensile tests with a set of 3 specimens were carried out in the temperature range of 25–725°C on a Zwick Z100 universal testing machine. The strain rates were respectively 0.005/ min and 0.05/min before and after being yielded according to ASTM E21. Yield strength and ultimate tensile strength were defined at 0.2% offset and the maximum stress respectively. The elongation percentage was determined as the maximum elongated gauge length divided by the

original gauge length. The gauge zone with a width of 50 mm was marked before tensile tests and then the elongation of gauge zone was measured manually after test. The geometry of tensile specimens is presented in Fig. 2. Fracture morphology after tensile tests was observed by SEM. The precipitates in different areas of the welded joint were analyzed by a Bruker D8 Advance XRD with a CuKα1 radiation source (λ=1.5406 Å) and an analyzed area size of 0.4×12 mm at step size of 0.02° in 0.15 s. Measurements were conducted in θ-2θ mode geometry at a tube power of 40 kV/40 mA to maintain linearity in the detector response. To analyze the narrow eutectic region in the welded joint, Gleeble3500 was applied to simulate heat-affected zone (HAZ) thermal cycle treatment. The detailed principle and application of Gleeble3500 have been described [19]. Fig. 3 displays the thermal cycle curve, in which the peak temperature is the key parameter to govern the final microstructure in the alloy. In this work, it was set to 1340 °C which is higher than the melting point of eutectic precipitate according to our previous work [16].

3. Results 3.1. Microstructure of the welded joint Optical micrographs of the cross-section of as-welded GH3535 alloy joint are shown in Fig. 4. It is evident that three different zones can be observed, including base metal, heat affected zone (HAZ) and weld metal. The base metal is a single-phase austenite with chains of precipitates in the γ matrix. As for HAZ, it is divided into three parts, including eutectic zone, coarse grain zone and heat affected base metal zone with no significant change in microstructure. For the “eutectic zone” with width of about 300 µm near fusion line, it is noticeable to observe some much darker chains of carbides. HAZ (E) stands for the eutectic zone in HAZ. The range of “coarse grain zone” in HAZ is approximately 600 µm and a slight increase in grain sizes is observed compared with the adjacent out part of HAZ. The left part of HAZ is the area with no observed microstructure change and its width cannot be determined accurately from OM photos. HAZ(R) stands for the rest part of the HAZ except eutectic zone. In the weld, there exist precipitates in the matrix and columnar crystals almost perpendicular to the fusion line. Besides, migrated grain boundaries (MGBs) and solidification sub-grain boundaries (SSGBs) are observed. The presence of MGBs in the fusion zone may cause failure during tensile loading since some element segregation (S, P, B, O, etc.) can occur along MGBs

Table 1 Chemical compositions of GH3535 alloy and filler wire (wt%). Alloy

Mo

Cr

Fe

C

Si

Mn

Ni

GH3535 ERNiMo-2

16.5 16.4

7.0 8.0

4.0 5.0

0.06 0.05

0.27 1

0.5 1

Bal. Bal.

169

Materials Science & Engineering A 682 (2017) 168–177

S. Chen et al.

Fig. 2. The geometry of the tensile tests specimen.

columnar dendrite both are detected in the weld metal. Large number of precipitates mainly along the boundaries of dendrites were observed in the weld metal. It is evident that these precipitates have two sorts of appearance (Fig. 6(c)), one is a particle with the size less than 1 µm, and the other is larger with a similar morphology as eutectic phase in the HAZ but of different sizes and structures. 3.2. Precipitates in the welded joint It can be deduced at least two sorts of carbides existed in the aswelded joint from the microstructure in Fig. 6. To further understand the relation and difference between primary and eutectic precipitates, the area near fusion line in HAZ was analyzed by EPMA. Fig. 7 shows that the primary precipitates in the base metal and the eutectic precipitates in HAZ are the carbides which are rich in Mo, Si and C, but lack of Ni, Fe and Cr. It is of note that, between the two kinds of carbides, concentration of Mo and Si of the eutectic carbides are significantly lower than those of the primary carbides. In contrast, the concentration of Fe in the eutectic carbides is apparently higher and the concentrations of Cr, Ni are slightly higher than the primary carbides. In view of the lamellar structure of eutectic carbide, the concentration of the eutectic carbide is certain to be affected by the matrix since the spot diameter of electron beam is similar to the thickness of the lamella in the eutectic carbide, the slight difference of element content between the primary and eutectic carbides may be due to the measuring error of EPMA. TEM samples of carbides were sampled from three regions of the welded joint, namely, base metal, HAZ and weld zone. TEM images of those samples are shown in Fig. 8. It can be clearly seen that three images are all consisted of γ matrix and precipitates. The selected area electron diffraction (SAED) pattern of the precipitate from base metal is shown in Fig. 8(a) where a block of carbide in the image is visible, it clearly matches the pattern of an FCC/Fd3m structure taken on a [110] zone axis. The lattice parameter was determined to be~11.1 Å. As for the samples of precipitates in HAZ and weld metal, approximately parallel and long stripes of precipitates can be observed in Fig. 8(b) and Fig. 8(c), SAED patterns of the [1–21] and [−110] zone axis for the precipitates were obtained, they both match the pattern of an FCC/ Fd3m structure as well. In addition, their lattice parameters were determined to be ~11.03 Å and ~11.01 Å respectively, which indicates

Fig. 3. The simulated HAZ thermal cycle curve.

in the solid state to lead to ductility - dip cracking [20]. High magnification OM and SEM images of base metal, HAZ and weld metal are shown in Fig. 5(a)–(c) and Fig. 6(a)–(c) as follow, respectively. (1) Base metal. The diameters of planar grains are in the range of 50– 100 µm and ASTM grain size is about 4.5 as shown in Fig. 5(a) and Fig. 5(a). The primary precipitates, in chains with the sizes in the scope of 1–10 µm, are more or less equiaxed in general within the grain and grain boundary. In fact, the chain of precipitates aligning with rolling direction of the plate is a common phenomenon in the wrought condition [10]. (2) HAZ. It is obvious that welding brings a significant change to the microstructures of HAZ. Precipitates in the eutectic zone of HAZ possess a eutectic-like morphology, which is shown more clearly at the high magnification SEM image in Fig. 6(b). The structure of eutectic precipitates is in the form of lamellae and skeleton. Meanwhile, several partially melted primary precipitates are observed to be surrounded by eutectic precipitates in Fig. 6(b). Some twins are also observed in the matrix of eutectic zone (Fig. 5(b)). (3) Weld metal. The microstructure of the weld metal is typical cast structure in Fig. 5(c) and Fig. 6(c). The equiaxed grains and

Fig. 4. Optical metallographic photos of welded joint (The width of HAZ is about 5 mm).

170

Materials Science & Engineering A 682 (2017) 168–177

S. Chen et al.

Fig. 5. High magnification OM images of the welded joint, (a) base metal; (b) HAZ; (c)weld metal.

3.3. Mechanical performance of the welded joint

the carbides with different morphologies in different regions of welded joints have the same crystal structure. Energy dispersive X-ray spectroscopy (TEM-EDS) was conducted on the samples to determine the chemical composition of the precipitates and the results are shown in Table 3. It should be noted that the C concentration in the samples of this work are all about zero due to the difficulty to detect the light elements by EDS, thus all the measurements of chemical elements are not very accurate, and these quantities are only given as estimation for guidance. As shown in Table 3, the precipitates in different regions have similar concentration of elements, which are all rich in Mo, Si, and depletion of Ni, Cr and Fe compared with the matrix. The results are consistent with the qualitative elemental analysis of carbides in HAZ by EPMA (Fig. 7). The Ni: (Mo+Cr) ratios of precipitates in base metal and HAZ(E) are both close to 0.75. As for the precipitates in weld, the Ni: (Mo+Cr) ratio is close to 0.61. The difference between the Ni: (Mo+Cr) ratios of carbides in different regions of welded joint may due to metal atom substitution in M6C structure, non-stoichiometric measurment or slightly different variation in the overall composition of local areas [15]. Fig. 9 shows a SEM image and corresponding EBSD phase map of γ-matrix and eutectic in the HAZ. In the EBSD analyses, the Ni matrix and Ni3Mo3C were presupposed as the main two phases which exist in the GH3535 alloy. As shown in Fig. 9(a), the eutectic carbides are in the upper part of the image, while in the lower part, several partially melted primary carbides can be found. Besides, the structure of eutectic carbide is distinctly different from the primary carbide and significant skeleton carbides are detected to extend from the particle primary carbides. Phases map in Fig. 9(b) is composed of γ matrix and Ni3Mo3C, indicating the primary carbides and the eutectic carbides have the same phase. XRD patterns of the base metal, eutectic zone and weld metal are presented in Fig. 10. The XRD pattern of the base metal reveals the presence of γ matrix and M6C carbides. For the weld metal, compared with base metal, no new diffraction peak appeared in the XRD pattern except the diffraction peaks of matrix phase and M6C carbide phase. As for the HAZ (eutectic zone), a large sample stimulated by Gleeble 3500 was characterized to determine the phase, whereas the result does not bring any new different diffraction peak, just like those of the weld metal.

3.3.1. Hardness of the welded joint Micro Vickers technique was employed to evaluate the hardness change of welded joint as shown in Fig. 11. It is of note that HAZ was divided in two parts according to the hardness values. HAZ (E) stands for the eutectic zone in HAZ and HAZ(R) stands for the rest part of the HAZ except eutectic zone. The base metal has the lowest hardness of the welded joint. Close to the weld, the hardness is highest in the range from base metal to HAZ (E), in some cases exceeding 300 HV. As for the weld metal, hardness slightly falls compared with HAZ (E) and fluctuates between 250 HV and 290 HV. The mean hardness values of different regions are shown in Fig. 11(b), in which the hardness change follows a similar tendency as that in Fig. 11(a). The mean hardness value of HAZ (E), 281 HV is largest in the welded joint and peaks at 281 HV, while that of weld metal and HAZ(R) are 271 HV and 260 HV, respectively. Besides, base metal possesses the minimum hardness value of 250 HV in the welded joint. According to the hardness test, it is worth noting that welding bring a significant change to the HAZ (E) where hardness value is the highest. 3.3.2. Tensile properties of the welded joint at different temperatures Fig. 12 shows the tensile curves and transverse tensile properties of welded GH3535 alloy joints and the base metal at different temperatures in the range of 650 °C to 725 °C. The data of base metal was obtained from the relevant acceptance report of GH3535 alloy. As for the base metal, the previous studies [21,22] indicated Hasterlloy N and GH3535 alloy have stable tensile properties before 550 °C and a minor decline in the range of 550 °C to 800 °C. In addition, the types of precipitates in Hastelloy N alloy are unchangeable to 1180°C, and the microstructures are relatively stable at elevated temperatures below 700°C in a short time [23,24]. Fig. 12(a) reveals tensile curves of specimens of welded joints. The curves at different elevated temperature overlap, indicating those specimens possess similar strengths and elongations at the temperatures in the range from 650 °C to 725 °C. Fig. 12(b) shows the tensile properties of welded joints and base metal. It is interesting that the yield strength (σ0.2) of joints are always 20– 30 MPa higher than those of the base metal at all test temperatures. As for RT, σ0.2 of base metal and welded joint are 350 MPa and 373 MPa respectively. With increasing temperature from 650 °C to 725 °C, σ0.2

Fig. 6. SEM images, (a) base metal; (b) HAZ; (c) weld metal.

171

Materials Science & Engineering A 682 (2017) 168–177

S. Chen et al.

Fig. 7. EPMA images of carbides in HAZ of welded joint.

declines very slowly for the base metal and welded joint, their values are in the vicinity of 210 MPa and 240 MPa respectively, which indicates the σ0.2 of GH3535 alloy is stable at elevated temperature in the range from 650 °C to 725 °C. As for the ultimate tensile strength (UTS) of welded joint, it falls slowly from 512 MPa at 650 °C to 484 MPa at 725 °C. Base metal has similar UTS tendency and values at the elevated temperatures. However, the trend mentioned above is not applicable for the tensile ductility. The elongations (δ) of base metal are almost twice as that of the welded joints at variously elevated temperatures. The minimum δ with values of 33% and 17% for the base metal and welded joint are both obtained at 650 °C, as temperature rises, δ has a slight increase at the given temperature range.

Table 3 Quantitative elemental analysis of carbides in different regions by TEM. Element (at%)

Mo

Ni

Cr

Fe

C

Si

Matrix Carbides in Base metal Eutectic carbides in HAZ Eutectic carbides in weld metal

11.7 47.1 45.5 48.8

72.4. 38.3 40.9 35.3

9.1 5.7 7.9 8.9

5.0 1.0 1.7 1.5

0.0 0.0 0.0 0.1

0.4 7.7 4.1 5.5

temperatures. The typical tensile fracture locations for the as-welded joint at RT and different elevated temperatures all occurred in the base metal as shown in Fig. 13(a)–(d), respectively. A typical ductile fracture of the sample stretched at RT is shown in Fig. 13(a). Dimples and microscopic cracks are distinctly seen on the fracture surfaces, and carbides are also detected at the bottom of dimples. Careful observation of top right magnification image reveals

3.3.3. Fractures morphologies of the tensile specimens of welded joint Fig. 13 shows the fractures morphologies of welded joints at various

Fig. 8. TEM SAED patterns, (a) Base metal;(b) HAZ;(c) Weld metal.

172

Materials Science & Engineering A 682 (2017) 168–177

S. Chen et al.

Fig. 9. Photos of eutectic carbides and primary carbides. (a) SEM image of carbides in HAZ, (b) Corresponding EBSD phase map.

process, some carbides may break into two pieces where micro cracks start to generate after local plasticity is exhausted. Fig. 13(b) shows fracture morphology after tensile tests at 650 °C. In this case, the cleavage surface and tearing ridge could be observed on fracture surface, which are taken as typical brittle fracture features. It can be deduced that the inter-granular fracture becomes dominating fracture mode although the ductile fracture is marked by the existence of dimples. This brittle fracture morphology reflects the intermediate temperature embrittlement of GH3535 alloy at elevated temperatures, and it is a common phenomenon for many types of nickel based alloys at elevated temperatures in the range from 600°C to 800°C [25,26]. In terms of the temperature at 700 °C and 725 °C, the inter-granular fracture still dominates the fracture surfaces. Similar fracture morphologies as in Fig. 13(b) are observed in Fig. 13(c) and (d) as well. Moreover, micro cracks are detected along the grain boundary from the high-magnification images at the top right corner in Fig. 13(b), (c) and (d). It is evident that the grain boundary cohesion becomes weaker and inter-granular fracture becomes dominant with the increase of testing temperature in the range from 650 °C to 725 °C. Fig. 14 shows the appearance of fractured tensile sample tested at 650°C. The weld bead is located in the middle of sample as shown in Fig. 14(a) and (b) which is the contour profile of the fractured sample.

Fig. 10. XRD patterns of the base metal, eutectic zone in HAZ and weld metal.

the morphology of cracks where granular carbides with smooth crosssection surfaces are embedded in the matrix. It should be mentioned that these fracture features indicate the micro cracks initiate at the carbides. Since carbides are so hard and brittle compared with matrix that the uncoordinated deformation is inevitable during the tensile

Fig. 11. Hardness values of the welded joint under load of 500gf. (a) Hardness value of the welded joint (b) Mean hardness value of different micro areas.

173

Materials Science & Engineering A 682 (2017) 168–177

S. Chen et al.

Fig. 12. (a)Tensile curves of welded joints at different temperatures. (b) Comparisons between base metal and welded joint at different temperatures.

A slight necking can be observed in the weld except for a significant one near fracture in the base metal as shown in Fig. 14(c). Fig. 15 shows twins boundaries of HAZ and base metal by EBSD after tensile tests. The length fractions of twins boundaries (Σ3 type) in HAZ near fusion line and base metal of as-welded joints are 45.2% and 49.9% respectively in Fig. 15(a) and (b). Fig. 16 shows the distribution of twins in different regions of welded joint by EBSD after tensile tests. The fractions of twins in HAZ near fusion line of as-welded joints are 42.3% (Fig. 16(a)). As for the regions HAZ(R) near HAZ(E) (Fig. 16(b)) and HAZ(R) near base metal (Fig. 16(c)), they possess similar fractions of twins with 33.4% and 35.6% respectively. However, the fraction of twins in base metal near fracture position is only 27.9% (Fig. 16(d)). Moreover, it is interesting to note the grain size in HAZ almost remains unchanged before and after tensile tests as shown in Fig. 15(a) and Fig. 16(a), while for the regions in Fig. 16(b) and (c), it is found that the grains become slight larger and elongated. A significant change is detected in Fig. 16(d) that the grains of base metal are elongated along the load direction.

Fig. 14. Fractured samples after tensile tests at 650°C (a) fractured sample of welded joint, (b) the contour profile of fractured sample, (c) different parts of the welded joint near the fracture.

4. Discussion 4.1. Microstructure evolution in weld and HAZ Solid-solution strengthened nickel-base alloys are primarily strengthened by the addition of substitutional alloying elements including Cr, Fe, Mo, W. The additions not only result in a net strengthening (or hardening) of the austenite phase but also contribute

Fig. 13. Fracture morphology of tensile specimens at (a) RT, (b) 650 °C, (c) 700 °C, (d) 725 °C.

174

Materials Science & Engineering A 682 (2017) 168–177

S. Chen et al.

Fig. 15. EBSD images of twins boundaries of as-welded joint. (a) HAZ (E) near fusion line, (b) Base metal.

γ + M6 C(p) → L → L + γ + M6 C(e) → γ + M6 C(e)

to form MxC-type carbides [27]. Under most processing conditions, the carbides, such as M23C6, M7C3, M6C, are commonly found in these alloys. Although the types of precipitates in the welded joint of UNS N10003 alloy are still subjected to debate, data obtained in this work suggests that the carbides in different regions possess the same chemical composition and crystal structure by EPMA, EBSD, TEM and TEM-EDS. With further analysis on the phase constitution by XRD, it can be determined that the carbides in different regions of the welded joint are all M6C type. In addition, an empirical equation is also helpful to identify the type of the carbide as follows [28].

R = XCr /(XCr + XMo + 0.7XW ) at. %

(2)

Where M6C(p) and M6C(e) represent primary and eutectic M6C, respectively. As in the HAZ, this region undergoes thermal cycles in the welding process. The only difference between the HAZ and the weld metal is that the peak temperature in HAZ does not reach to the melting point. However, some carbides in the HAZ near fusion line still experience the process of melting and re-solidifying [15,17] due to its lower melting point (1300–1335°C) as compared with the matrix [13]. Fig. 6(b) clearly reveals the transformation process of eutectic carbides during welding process. As shown in Fig. 6(a) in the out part of HAZ, the base metal in this region experiences a moderate temperature during the welding thermal cycle, the peak temperatures of thermal cycles are getting higher as the distance is getting closer to the weld bead as shown in Fig. 6(c). In certain regions of the HAZ near the fusion line where the peak temperature exceeds the eutectic point, eutectic reaction would firstly start in the interface between the carbides and matrix since some element segregation (S, P, etc.) segregated in the interface will cause a relatively lower melting point compared with the matrix and the interface firstly melts in the thermal cycles, after a diffusion in short distance to some elements such Mo, Cr, Ni, etc. in the transitory temperature-fall period of thermal cycle, the liquids will reach to approximate eutectic composition and cause eutectic reaction. This reaction brings a significant change to the primary carbides that the lines of eutectic carbides grow from the primary carbides and alternately parallel arrange with γ matrix. Some carbides cannot melt thoroughly in the critical point of eutectic reaction as shown in the left side of Fig. 6(b), with the temperature increasing as moving to fusion line, the primary carbides would undergo a complete transformation to eutectic carbides. Thus the phase transformation sequences in HAZ are as follows.

(1)

where XCr, XMo and XW are the molar fraction of Cr, Mo, and W, respectively. For R < 0.72, M6C is expected to form. For R > 0.82, M23C6 is expected to form. For 0.72 < R < 0.82, MxC-type is dependent on the heating treatment. As shown in Table 3, the chemical compositions of GH3535 alloy are as follows, Mo, 11.7%; Cr 9.1%; W, 0%. According to Eq. (1), the R value of GH3535 alloy is, R = XCr /(XCr + XMo +0.7 XW) =9.1% /(9.1% +11.75%)=0.436. It is less than 0.72. Therefore, the carbide formed in the GH3535 alloy can be determined positively as M6C, which is consistent with the results obtained above and literature regarding to the carbides in Ni-Mo-Cr alloy [15]. Eutectic structure is a common phenomenon and has been reported to be observed in a variety of nickel-base materials, such as Hastelloy 242 and B-2 [29,30]. As mentioned before, the eutectic carbides are also detected in the welded metal and the HAZ of GH3535 alloy as shown in Fig. 5(b) and (c). According to the results by using several methods, both of the eutectic and primary carbides are actually the same type of carbide, the only difference between them is the morphology. Namely, welding heat cycle only leads to the morphological change but not the type of phase. As for the weld, it is evident that the carbides and matrix both undergo melting and then re-solidifying when the temperature increases from RT to the melting point and falls back to RT during the process of welding. In view of alloy consisting of theγmatrix and the M6C carbide, the phase transformation sequence observed in the weld metal of welded joint of GH3535 alloy can be written as:

Forpartiallymeltedprimarycarbides γ + M6 C(p) → γ + M6 C(p) + L → γ + M6 C(p)+ M6 C(e)

Formoltenprimarycarbides γ + M6 C(p) → γ + L → γ + M6 C(e)

(3) (4)

Fig. 16. EBSD images of twins boundaries of HAZ and base metal in the welded joint after tensile tests. (a) HAZ(E) near fusion line, (b) HAZ(R) near HAZ(E), (c) HAZ(R) near base metal , (d) Base metal near failed position.

175

Materials Science & Engineering A 682 (2017) 168–177

S. Chen et al.

have been determined to lead a large increase on the hardness of HAZ(E) though the detail mechanism are still unclear, needing a further investigation in future. High hardness of HAZ(E) means a high yield strength which results in a non-uniform plastic deformation in HAZ during the tensile tests. On the other hand, another factor contributed to improvement of the strength of HAZ has been detected from the EBSD result of HAZ and base metal after tensile tests as shown in Fig. 15 and Fig. 16. The similar data of twins boundaries (Σ3 type) in HAZ before and after tensile tests and equaxied grains in Fig. 15(a) and Fig. 16(a) indicates a small plastic deformation. As in the regions of HAZ(R) near HAZ(E) (Fig. 16(b)) and HAZ(R) near base metal (Fig. 16(c)) after tensile tests, their length fractions of Σ3 type are lower than in HAZ(E) but higher than base metal near fracture position(Fig. 16(d)). The large number of twins existing in the whole deformation process were reported to be significant to the improvement of tensile strength and ductility since they are believed to be tough and crack-resistant boundaries [40,41]. The obvious decrease in Σ3 type of base metal as shown in Fig. 15(b) and Fig. 16(d) would reduce the ability to block the movement of dislocations and lead to a weaker local strength compared with that of HAZ. In general, the good mechanical performance of weld metal mainly benefits from the dispersively distributed fine carbides which play a role of improving the resistance of dislocation motion. As for HAZ, eutectic M6C carbides and twins are the two factors that result in higher hardness, especially in HAZ(E). These strengthening factors integratedly increase deformation resistance and then improve the yield strength of local matrix. On contrary, base metal with a lower hardness firstly starts to yield in the tensile process and occurs necking and fracture. Besides, we come to an additional conclusion—the eutectic carbides in the real welded joint of UNS N10003 alloy do not bring any negative effects on the short-time mechanical performance. However the results do not imply a same conclusion on the long-term mechanical performance, such as high-cycle fatigue and creep rupture, which needs a further research in future. The results obtained here can provide valuable guide in design and application of GH3535 alloy.

4.2. Strengthening mechanism of welded joint Obviously, failure of all the tensile samples occurred in the base metal as shown in Fig. 13 and Fig. 14(a), which means the weakest part of the welded joint is not the weld metal and HAZ but the base metal. Evidently deformation of sample is uneven due to the inhomogeneity of material of the welded joint in the tensile tests as shown in Fig. 14(b). The results of the hardness tests may be used to understand the tensile tests results since hardness is in direct proportion to the yield strength of the metallic material. In the early stage of tensile tests, base metal starts to yield firstly and then plastic deformation occurs since its lower hardness and strength. In the meantime, the areas with higher hardness are still in the elastic stage. As the tensile tests continues, a further deformation happens to the base metal, while HAZ(R) and weld just start to yield and undergo the plastic stage. Finally, necking and fracture successively take place in base metal when its plasticity-reserve runs out, but the plastic deformation of HAZ(R) and weld are still not exhausted. Moreover, it is possible that deformation has not yet occurred in HAZ (E) due to its high hardness. Therefore, a slight necking can be observed in the weld in contrast to a significant one near fracture in the base metal as shown in Fig. 14(c) after the tensile tests. It is generally believed that the weld and HAZ are usually the highrisk failure areas due to the existence of welding defects, element segregation, microstructure variations and performance non-uniformity [20]. In this work, the welded metal and HAZ of the joints possess a better strength than the base metal and reveal a different fracture tendency from the common cases. It is of worth to discuss the strengthening mechanism of the welded metal and HAZ respectively. From the microstructure analyses, the main differences between the three different regions in the welded joint are the grain and carbide morphology. The grain in base metal and HAZ are both equiaxed austenite, but the ones in the weld metal is typically cast microstructure which consists of majority columnar crystal and minority fine equiaxed austenite as shown in Figs. 4–6. Columnar crystal in weld metal of the nickel base alloy normally possesses coarse grain size and therefore less grain boundary compared with those of base metal and HAZ [31], which is not conductive to the hardness and strength of the alloy at room temperature but helpful to improve tensile strength at elevated temperature [32,33]. To some extent, this can explain why all the failure occurred in the base metal at 650-725 °C although it is not consistent with the mechanical results at RT, suggesting varieties of grain morphology is not the crucial factor to affect the mechanical strength of different regions in the welded joint. In term of carbide, a significant difference can be detected in the joint. The primary carbides existed in the base metal are in chains with the sizes in the scope of 1– 10 µm. After the welding process, carbide morphology in weld metal is observed in two forms, minority is bulk with the size up to 15 µm and majority is small particle with the size approximately in the range of 0– 0.5 µm. As mentioned in Fig. 6(c), most of eutectic carbides distribute at the grain boundary and in the interdendritic regions of the weld. The sizes of carbides in the weld are very fine, performing vital function in obstructing the movement of the dislocations to enhance the hardness of alloy [34,35] and are conducive to enhance the tensile strength [36,37]. For HAZ, as mentioned above, it is also a high risk failure region in the service of welded structure. As revealed in Figs. 4–6, the sizes of eutectic carbides, whose length can be up to more than 30 µm in HAZ, are much larger than carbides in other regions of joint. This is mainly caused by the agglomeration of the molten adjacent carbides which fused and formed into whole eutectic carbide during the welding heating and cooling process. Large eutectic carbides may act as crack initiators to be detrimental to the ductility of the alloy to some extent [38,39]. However, in our present work, this deduction also does not match the test results. According to the results of hardness test, eutectic carbides in HAZ, which transformed from primary carbides,

5. Conclusions In this work, GH3535 was welded to evaluate the mechanical and microstructure properties of welded joint. Several conclusions have been achieved as follow. (1) Good welded joints of the GH3535 alloy were obtained by using GTAW. The eutectic carbides are detected in HAZ and weld metal, possessing similar elements concentration to primary carbide, which all are rich in Mo, Si, and lack of Ni, Cr, Fe compared with the concentration of matrix. (2) The eutectic phases in the HAZ and weld metal of welded joint were both identified as M6C type, which are the same as the primary carbides in base metal. The phase transformation of eutectic carbides in weld metal is γ+M6C(p) → L→ L+γ+M6C(e) → γ+M6C(e). As for HAZ, the phase transformation of partially melted primary carbides and molten primary carbides are as follows, γ+M6C(p) → L+γ+M6C(p) → γ+M6C(p)+M6C(e) and γ+M6C(p) → L+γ → γ+M6C(e), respectively. (3) HAZ (E) possesses the maximum value beyond 300HV, which is much higher than weld metal, HAZ(R) and base metal. The significant change of hardness in HAZ(E) is mainly caused by the eutectic carbides transformed from primary carbides. (4) Tensile tests show that the welded joint samples have a very stable mechanical performance at elevated temperature from 650 °C to 725 °C, consequently all tensile samples failed in base metal of welded joint. The tensile fracture at RT was a typical ductile fracture, while the inter-granular fracture dominates the fracture mechanism at the temperature from 650 to 725 °C. Fine carbides acting as dispersoids in the columnar grain boundary and inter176

Materials Science & Engineering A 682 (2017) 168–177

S. Chen et al.

[14] Y. He, J. Yang, C. Qin, S. Chen, Z. Gao, J. Alloy. Compd. 643 (2015) 7–16. [15] D. Bhattacharyya, J. Davis, M. Drew, R.P. Harrison, L. Edwards, Mater. Charact. 105 (2015) 118–128. [16] Y. He, J. Yang, S. Chen, Z. Li, Z. Gao, J. Alloy. Compd. 660 (2016) 266–275. [17] J. Yang, Y. He, C. Qin, W. Zhao, S. Chen, Z. Gao, Mater. Des. 86 (2015) 230–236. [18] ASME, 2010. [19] F.-R. Xiao, B. Liao, Y.-Y. Shan, G.-Y. Qiao, Y. Zhong, C. Zhang, K. Yang, Mater. Sci. Eng.: A 431 (2006) 41–52. [20] J.C. Lippold, S.D. Kiser, J.N. DuPont, Welding Metallurgy and Weldability of Nickel-base Alloys, John Wiley & Sons, Inc., Hoboken, New Jersey, 2011. [21] F.F. Han, B.M. Zhou, H.F. Huang, B. Leng, Y.L. Lu, J.S. Dong, Z.J. Li, X.T. Zhou, Mater. Chem. Phys., 2016. [22] R.W. Swindeman, Oak Ridge National Lab., Tenn., , No. ORNL-2780., 1961. [23] R.E. Gehlbach, H.E. Mccoy, R.E. Gehlbach, H.E. Mccoy, In: Proceedings of International Symposium on Structural Stability in Superalloys, Seven Springs, PA 1968. p. 346-66.. [24] F.F. Han, B.M. Zhou, H.F. Huang, B. Leng, Y.L. Lu, Z.J. Li, X.T. Zhou, Mater. Sci. Eng. A 673 (2016) 299–306. [25] L. Zheng, R. Schlesiger, R. Chellali, D. Baither, G. Schmitz, Mater. Des. 34 (2012) 155–158. [26] R. Chellali, Crit. Rev. Solid State Mater. Sci. 37 (2012) 181–214. [27] J. Guo, Science Press, Beijing, 109, 2008, 292. [28] G. Jianting, Science Press, Beijing, 2008, pp. 373–374. [29] T. Lienert, C. Robino, C. Hills, M. Cieslak, Sandia National Labs., Albuquerque, NM (USA), 1990. [30] M. Maguire, T. Headley, Weldability of Materials: Proceedings of the Materials Weldability Symposium, 1990, pp. 167–173. [31] M. Qian, J. Lippold, Acta Mater. 51 (2003) 3351–3361. [32] P. Sun, C. Yu, P. Kao, C. Chang, Scr. Mater. 52 (2005) 265–269. [33] A. Chokshi, A. Rosen, J. Karch, H. Gleiter, Scr. Metall. 23 (1989) 1679–1683. [34] N. Verdhan, R. Kapoor, Comput. Mater. Sci. 98 (2015) 149–157. [35] A. Thompson, J. Brooks, Acta Metall. 30 (1982) 2197–2203. [36] I. Gutierrez-Urrutia, M. Munoz-Morris, D.G. Morris, Mater. Sci. Eng.: A 394 (2005) 399–410. [37] M. Ahmad, J. Akhter, M. Akhtar, M. Iqbal, E. Ahmed, M. Choudhry, J. Alloy. Compd. 390 (2005) 88–93. [38] L. Lu, H. Soda, A. McLean, Mater. Sci. Eng.: A 347 (2003) 214–222. [39] R. Briggs, Molten-salt Reactor Program Semiannual Progress Report, ORNL-3812. 1965, pp. 71–72 [40] C.Y. Cui, Y.F. Gu, Y. Yuan, T. Osada, H. Harada, Mater. Sci. Eng. A 528 (2011) 5465–5469. [41] T. Watanabe, S. Tsurekawa, Acta Mater. 47 (1999) 4171–4185.

dendritic region of weld metal is the main factor to increase the dislocation motion resistance and enhance the hardness and strength of the welded metal. As for the HAZ, eutectic M6C carbides and twins are the two main factors that result in an ascent of hardness especially in HAZ(E) and a good strength. In addition, the eutectic carbides in the welded joint do not reveal any adverse effects on the short-time mechanical performance. Acknowledgements This work was supported by the National Key Research and Development Program of China (Grant No. 2016YFB0700404), National Natural Science Foundation of China (Grant Nos. 51371188, 51671122, 51671154, 51601213, 51501216), Strategic Priority Research Program of the Chinese Academy of Sciences (Grant No. XDA02004210), and Talent Development Fund of Shanghai (Grant No. 201650). References [1] R. Gehlbach, H. McCoy Jr, International Symposium on Structural Stability in Superalloys, Seven Springs, Pa. Sept. 1, 1968, pp. 346–366. [2] P.N. Haubenreich, J. Engel, Nucl. Technol. 8 (1970) 118–136. [3] M. Übeyli, T. Demir, Mater. Des. 29 (2008) 852–859. [4] L. Mathieu, D. Heuer, R. Brissot, C. Garzenne, C.L. Brun, D. Lecarpentier, E. Liatard, J.M. Loiseaux, O. Méplan, E. Merle-Lucotte, Prog. Nucl. Energy 48 (2006) 664–679. [5] M. Brovchenko, D. Heuer, E. Merle-Lucotte, M. Allibert, V. Ghetta, A. Laureau, P. Rubiolo, Nucl. Sci. Eng. 175 (2013) 329–339. [6] L. McNeese, Oak Ridge National Lab., Tenn., 1971 [7] M. Rosenthal, P. Kasten, R. Briggs, Nucl. Technol. 8 (1970) 107–117. [8] Y. He, J. Yang, H. Shen, L. Wang, Z. Gao, Mater. Des. 104 (2016) 1–9. [9] W. Manley, Fluid-Fueled Reactors, 1958, pp. 595–603 [10] J. Venard, ORNL-TM-1017 (February 1965), 1965. [11] H. HARLAN, V. GRITZNER, Met. Eng. Q. 3 (1963) 196343. [12] H. McCoy Jr., J. Weir Jr., Oak Ridge National Lab., Tenn., 1967. [13] J. Li, S.L. Shrestha, Y. Long, L. Zhijun, Z. Xintai, Mater. Des. 93 (2016) 324–333.

177