Microstructure characteristics in weld zone of the novel thick 08Cr9W3Co3VNbCuBN heat-resistant steel welded joint by fusion welding

ABSTRACT One of the primary candidate materials used in 650°C ultra-supercritical (USC) thermal power plant units is a novel thick grade of heat-resistant steel, 08Cr9W3Co3VNbCuBN (G115), for which China has complete intellectual property rights. In this study, G115 steel with a 115 mm thick wall was successfully welded using gas tungsten arc welding (GTAW) + shielded metal arc welding (SWAM), which was a significant step toward achieving its application in large-diameter boiler tubes. The influence of the welding process on the microstructure and mechanical properties of the G115 welded joint after post-weld heat treatment (PWHT) at 770°C for 11 h was investigated. The findings demonstrate that the weld zone (WZ) after PWHT was composed of martensite and ferrite, and M23C6 was the precipitate. The average hardness of WZ was 273.1 HV, and the average impact toughness was above 39J, which was sufficient for actual production applications. It offered the necessary theoretical foundation for the practical development of thick-walled G115 heat-resistant alloy steel during welding. GRAPHICAL ABSTRACT


Introduction
Manufacturing and processing of heat-resistant steel tubes in high-temperature sections is the most important factor that restricts the development of complete sets of technology of USC thermal power units [1].However, the working temperature of traditional martensitic/ferritic heat-resistant steels containing 9∼12 wt% Cr (such as T/P91 steel, T/P92 steel, SAVEI2AD steel, etc.) is limited to 600∼620°C [2].A brand-new martensitic heat-resistant steel with full intellectual property rights, 08Cr9W3Co3VNbCuBN steel (known as G115 steel in China), was independently researched and developed in that country.It can be used for an extended period of time at temperatures ranging from 620 to 650°C and is primarily used in large diameter boiler pipes of USC thermal power plant units [3,4].The carbonisation of M 23 C 6 was effectively controlled by reasonably controlling the ratio of B and N.In addition to Nb and V, some Cu elements are added to increase the precipitation, further strengthening the effect [5][6][7].The G115 steel has excellent microstructure stability in the temperature range of 620°C to 650°C [8], and its target allowable stress at 625°C and 650°C is more than 1.5 times that of T/P92 steel [5,9].
T/P91 and T/P92 heat-resistant steels welded by the same or different kinds, typically adopting electric arc welding processes including SMAW, GTAW, electron beam welding (EBW), etc., has become the main focus of martensitic heat-resistant steel connection work in recent years [10][11][12][13].The weldability of P91 steel, however, is a significant issue [12].With an increase in working temperature and applied stress, the creep fracture life of P91 steel will decrease, and the creep at 650°C will increase the coarsening rate and deteriorate the properties [14].Additionally, regardless of the arc welding procedure and heat treatment employed, the creep fracture life of P91 base metal (BM) is greater than that of weld metal (WM) [15].Dissimilar welding process between martensitic heat-resistant steels or between martensitic heat-resistant steels and austenitic stainless steels is also one of the research focuses.Pandey et al. [16] compared autogenous tungsten inert gas (A-TIG) welding and GTAW welded joints of P91 and P92 dissimilar steels after PWHT.Compared with GTAW weld, δ-ferrite was observed in A-TIG weld, resulting in poor peak hardness and impact toughness of the weld.According to Dak et al. [17], hot cracking and carbon migration were caused by differences in the chemical composition, mechanical properties, and metallurgical characteristics of austenitic stainless steel and P91/P92 steel.The key challenges in dissimilar metals welding also included the generation of brittle intermetallic compounds, the development of a soft heat-affected zone in martensitic steel, and the formation of δ-ferrite in the fusion zone.Using the cold metal transfer plus pulse (CMT + P) method, Cai et al. [18] successfully welded G115 steel with a thickness of 15 mm while researching the impact of heat input on the microstructure and mechanical properties of G115 welded joints.Compared to Cai, this study used the more popular fusion welding (GTAM + SMAW) approach, which was simpler to implement in real-world production applications.In addition, the thickness was also thicker in this study (about 7.7 times that of the former).
The thick-walled G115 steel welded has not yet been the subject of any research reports.The GTAW backing, SMAW filling, and covering processes were used in this article to effectively weld the G115 welding joint, which had a thickness of 115 mm.By using an optical microscope (OM), a thermal field emission scanning electron microscope (SEM), an energy dispersive spectrometer (EDS), and X-ray diffraction (XRD), the microstructure characteristics of the G115 welding zone were carefully examined.A crucial theoretical foundation for the practical welding application of thick-walled G115 heat-resistant alloy steel was provided by the analysis of the impact of the welding process on the microstructure and mechanical characteristics in weld zone of joints.

Experimental
In this experiment, the test material was 08Cr9W3Co3VNbCuBN (G115) large thick-walled heat-resistant steel tube with Φ530 × 115 mm.The chemical composition (mass fraction, %) of the base metal was Cr 8.82, Co 2.99, W 2.66, Cu 0.94, Mn 0.58, Si 0.34, V 0.19, Nb 0.08, C 0.07, Ni 0.03, B 0.02, N 0.01, and the rest was Fe.Vickers hardness (HV) range of G115 steel BM is HV 196-265, the yield strength (YS) range is 674-706 MPa, the ultimate tensile stress (UTS) range is 792-836 MPa, the elongation (E) range is 23-25%, and the impact energy range is 48-76 J.The welding processing was performed using gas tungsten arc welding (GTAW, also TIG welding) backing, shielded metal arc welding (SWAM) filling and covering.The following welding consumables were used: GTR-W93 solid core wires and GER-93 electrodes (special for G115 steel).The chemical composition (mass fraction, %) of GTR-W93 and GER-93 were Cr 8.82, Co 2.82, W 2.31, Cu 0.01, Mn 0.60, Si 0.36, V 0.19, C 0.10 and Cr 9.00, Co 2.72, W 2.75, Cu 0.83, Mn 0.72, Si 0.13, V 0.17, C 0.07 respectively.It is worth mentioning that the chemical composition of filling materials is similar to that of G115 steel base material, which is specially designed for G115 steel arc welding and has not been circulated in the market at present.
A nearly V-shaped joint groove was designed due to the large wall thickness of the weldment, and the groove and edge were polished after processing.The physical drawing of the weldment and specific details of the groove are shown in Figure 1.Before welding, the joint shall be preheated to 200°C to prevent cold cracks.The welding process was multi-layer, multi-pass welding, and the interlayer temperature was maintained at 200-250°C.The specific process parameters were shown in Table 1.For 11 h, the PWHT was tempering at 770°C with air cooling to room temperature.The tempering temperature for the heat treatment of large diameter pipe fittings made of G115 steel was 770-790°C, and the holding duration was 3.5-5 min/mm and not less than 4 h, according to the research findings of Cong et al. [19].In this study, welding was done on G115 steel that was 115 mm thick.The heat treatment process parameters of 770°C/ 11 h were chosen in consideration of the substantial wall thickness.
A series of specimens were cut from the WZ after PWHT, and then these specimens were made into metallographic samples.The samples were etched  using mixed solution 5 g FeCl 3 + 50 ml HCL + 100 ml H 2 O.The G115 steel welded by GTAW + SMAW was cut by a wire cutting machine to prepare samples having a size of 10 × 10 × 10 mm, and the sample interception positions were A (welded cover), B (welded centre) and C (welded root), respectively (as shown in Figure 2a).First, XRD analysis samples were cut from the WZ.Then, the samples were also cut to 10 mm thick along the cross-section and polished to a bright surface.And then, the samples were cut mechanically to make a square with a thickness of 5 mm.At last, the cut samples etched were used for OM and SEM studies.At room temperature, the polished component of the joint WZ was tested for hardness using a Vickers hardness tester (HV1000), and the hardness distribution was measured using a 1.961 N load and a 10 s residence period.The WZ was tested from the top to the root, and every 4 mm, the hardness value was recorded.The hardness distribution curve for WZ was developed to fit the average value and standard deviation of the results, which were used to construct the microhardness distribution map.The impact toughness analysis samples were prepared by a lining cutting equipment, and the size of the sample was 55 × 10 × 10 mm.Three samples were taken at the root, centre and top of the WZ, respectively.The Charpy V-notch impact toughness test was performed in accordance with the ASTM E23 standard (USA) at room temperature.A pendulum impact tester was used to conduct the impact test (JBS-300H).The impact absorbed energy of the test sample under a 2 mm hammer blade was tested.The schematic diagram of the impact sample is shown in Figure 2b.

Results and discussion
Figure 3 depicts the microstructure of BM of G115 steel after PWHT, which was mostly composed of lath martensite.Previous austenite grain boundaries, lath pockets, lath blocks, and lath were used to make the martensite.The previous austenite grain had 3-5 randomly oriented martensite laths, multiple parallel lath bundles near the lath bundle boundary, and several dislocations in the lath.SEM image and EDS scanning results show that the BM of G115 steel was mainly composed of Fe, Cr, Co and W (see Figure 3b).This result was consistent with that reported in Section 2.
Figure 4 shows the microstructure on the welded surface of the G115 steel connected by GTAW + SMAW under different weld locations.The microstructure of the weld metal was composed of martensite and ferrite.The microstructure of the welded cover was mainly composed of ferrite and tempered lath martensite (see Figure 4a).It can be seen from Figure 4a and b that the welded cover was formed well, and there was no lack of fusion.The grains near the weld were mainly equiaxed.The tempered lath martensite transformed from lath martensite after PWHT was discontinuous, and the boundary of the lath bundle became unclear.A number of fine precipitates were found at the boundaries of martensite laths and lath bundles, where distribution was dispersed (see Figure 4b).
It can be seen from Figure 4c that the microstructure of the welded centre was mainly ferrite and lath martensite.It can be seen from Figure 4d that the tempered martensitic grain size in the centre of the weld was smaller than the cover.It was because, in the process of multi-layer and multi-pass welding, the filler layer reheated the previous layer of weld, causing the weld structure to undergo phase transformation and recrystallization, forming fine equiaxed crystals and refining grains.The martensite at the root of WM was consisted of lath martensite and block martensite, showing tempered martensite structure, and the grain boundary was clearer (see Figure 4e).In the process of root GTAW, part of the heat between the WP and the BM on both sides was transmitted, resulting in a relatively fast cooling rate [20], which increased the undercooling of the weld metal, improving the nucleation rate and promoting the formation of equiaxed crystals.Therefore, the grain at the root of the weld was fine.None of the obvious black holes in the root of the weld were found (see Figure 4f).
Figure 5 shows the SEM morphology with EDS results of the welded surface of the steel.The red boxes in Figure 5a, d   f, i correspond to the EDS energy spectrum and element composition content of test points, respectively.According to Figure 4a and b, it can be seen that the microstructure of the welded cover was mainly composed of tempered lath martensite.There were several black holes and precipitated second-phase particles in and inside the grain boundary in the welded cover.The holes may be defects such as pores, enclosed slag or incomplete fusion.And part of the second phase was distributed on the grain boundary.The EDS analysis was performed on the surface of spectrum 1 and spectrum 2, respectively.Based on the element composition of literature [21], EDS results and XRD analysis (see Figure 6), it is indicated that the matrix of the original alloy α-Fe solid solution.A large number of Si-rich phases were detected in the chemical composition (see Figure 5c1) of the black hole at spectrum 1 in Figure 5b1, and it was speculated that the black hole may be a Si-rich inclusion.This was mainly due to the fact that Si was added to the electrode as a deoxidising element to prevent the oxidation of iron, and the products were difficult to float out of the molten pool, resulting in slag inclusion in the weld metal.According to the above analysis of process and microstructures, the grains at the weld cover were coarser, which often led to high brittleness and low toughness of the joint.In addition, the spectrum 2 pointed out in Figure 5b2 represented the chemical composition of the martensite matrix area (see Figure 5c2), which had 81.51 Fe, 10.95 Cr, 3.62 Co and 3.91 W (at. %).This result showed that the weld metal had a chemical composition similar to BM.
It can be seen from Figure 5d that after the weld underwent multi-layer and multi-pass welding, the filler layer reheated the previous weld, resulting in phase transformation and recrystallization of the weld structure.A large number of fine equiaxed crystals appeared, and granular second phase particles appeared inside the crystal grains.A higher-magnification SEM image was selected in the weld centre, and EDS results were analysed to characterise the precipitate (see Figure 5e).EDS and XRD analysis of Figure 5e1 and e2 showed that the chemical composition (at.%) of the two points was Fe: 54.71, Cr: 20.93, W: 7.93, Mn 1.41, C 15.02 (see Figure 5f1) and Fe: 81.74, Cr: 13.41, W: 3.74, Mn 1.10 (see Figure 5f2), respectively.According to the EDS spectra (Figure 5f1), XRD analysis (Figure 6) and the literatures [18,22], it could be concluded that M 23 C 6 phase was rich in Cr and Fe and contained a small amount of W precipitated on the grain boundary.Since M 23 C 6 phase precipitated from the grain boundary, the strength and hardness of WZ were improved [18,23].The chemical composition of the grey dot in Figure 5f2 was similar to that of the martensite matrix, but no Co element was detected, so it was considered that there may be no cobalt-rich phase.It is worth mentioning that the weld centre of the EDS surface scanning analysis results and weld cover were not much different.As a result, the fine second phase M 23 C 6 precipitated in the weld centre.
Figure 5g shows the SEM morphology with EDS results of the welded surface of the root under the GTAW process.According to the EDS and XRD results, the element content of point spectrum 1 (Figure 5h1) was similar to that of Figure 5e1, which indicated that the second phase precipitated in this region was mainly M 23 C 6 (see Figure 5i1).Due to the narrow root, the cooling rate of molten pool metal was faster, and the grain size was finer than in other areas.It can be seen that the sediment at the root (Figure 5h) was smaller than that at the centre (Figure 5e) of the same scale.Spectrum 2 in Figure 5h2 indicated that the chemical composition (at.%) was Fe 86.08, Cr 10.70, W 1.89 and Mn 1.33 (Figure 5i2), which was not different from the weld cover (see Figure 5f2).It was worth mentioning that the content of the W element in the root was twice that of other areas, which may be caused by the difference in chemical composition of filling materials.As a result, the granular second phase precipitates at the grain boundary, which will hinder the dislocation movement.In addition, the intergranular strengthening phase by dispersion distribution will improve the strength of the WM [21,24].The EDS results also showed that Co 3 Fe 7 phase did not appear in the grey particles near the black hole.Therefore, it can be inferred that the precipitation phase in the defects of different welding methods will be slightly different but will not affect the stability of the overall structure.Combined with XRD analysis results (see Figure 6), the microstructure of the WM should be a structure composed of α-Fe matrix, Co 3 Fe 7 phase, WC phase and M 23 C 6 phase (mainly Cr 23 C 6 and Fe 23 C 6 ).
Figure 7 shows the distribution of microhardness of WZ.The test sequence was from the top to the root of WZ along the weld centre line.The average value and standard deviation of HV were listed at the top of the figure: HV = 273.1 ± 11.7 (the average value was before pluses or minuses, and the standard deviation were after pluses or minuses).The curve in the figure was the fitting result of the WM hardness value.Among them, the average hardness value of the weld root was relatively high, followed by the centre, and the average hardness value of the cover was the lowest.The point with the highest hardness value appeared near the centre and was 297.7 HV.The main factors affecting the hardness were the microstructure and alloy element composition, which had a specific relationship with the welding method used.First of all, the welded root was affected by alloy element segregation, and the content of main elements such as Cr, W, Mn, etc. increased, which had a specific strengthening effect on the weld root metal and improved the hardness value [25].Secondly, the welded cover was less affected by multilayer welding tempering heating, so the grains were relatively rough, which was reflected in the low hardness value.In addition, the welded root was formed using GTAW process, the welding heat input was relatively low, and the grains after welding were relatively small.The filler surface and covering surface welded by SMAW were gradually widened due to groove design, and the welding heat input was gradually increased from bottom to top, leading to the gradual increase of grain size and decrease of hardness value.In addition, different filler materials in the weld zone may also lead to different hardness values.
The impact test results are shown in Table 2.The mean values of impact energy in WZ were 45.33, 44.79 and 39.22 J, respectively.The Charpy impact toughness for WZ of G115 steel produced by SMAW + GTAW process was  lower than that of the BM [18].The impact absorption energy at the root of the weld was low, and the area was almost all equiaxed grains with anisotropic grains.The grain size and orientation tended to be the same, which was prone to crack [26].In addition, the difference in composition of the filling materials may also be one of the reasons for the poor root impact toughness.According to the NB/ T47014 standard (China) for pressure equipment welding process assessment, at least three samples should be taken from each region, and the average impact absorption energy should not be less than 31J.Each test area conformed to the standard and met the practical requirements.
The link between the microstructure and the fracture characteristics of each section of the welded joint may be seen in the study of impact toughness fracture.As shown in Figure 8, the impact toughness fracture morphology was illustrated.Overall, the fracture morphology revealed a mixed fracture mechanism of brittle zone and shear dimple, indicating a similar toughness.However, the fracture morphology of the low-toughness sample (root) showed a large brittle area, and the dimple was large and shallow, indicating low impact toughness.The brittle zone was made up of smaller separation surfaces connected by the tear ridge, which had a distinct river pattern and flowed in the same direction as the crack propagation.Owing to the presence of brittle regions, the WZ retained some brittleness.As a result of the fracture morphology inspection, the fracture mode of WZ was a mixed fracture of toughness (dimple) and brittleness (cleavage plane), with a tendency to be brittle.

Conclusions
In this paper, 115 mm-thick wall G115 heat-resistant steel was successfully welded by the SMAW + GTAW welding method.The influence of the welding process on the microstructure, microhardness, and impact toughness of post-weld heat treatment (PWHT, 770°C + 11 h) welded joints was studied.This study supported the use of G115 steel by confirming the viability of GTAW + SMAW process to joint G115 steel with a thickness of 115 mm.The present conclusions can be drawn as follows: (1) The weldability issue in weld metal of G115 steel welded joint was resolved, which had satisfactory weld formation and left no evident defects (undercut, hole, crack, and insufficient penetration) in the macroscopic weld zone by GTAW + SMAW process.(2) After post-weld heat treatment, the weld metal was composed of martensite and ferrite, and the grain size shrank from the weld cover to the root.Following root welding, the rapid cooling caused the formation of finer grains, and M 23 C 6 precipitates at or inside the grains, which may impede dislocation movement.Additionally, the strength of the weld metal will be improved through the dispersed intergranular strengthening phase.(3) Average hardness value in weld zone is 273.1 HV, while the hardness trended upward from the cover to the welded root with the impact toughness trending downward.This result was mainly caused by the difference in chemical composition of the filling materials.The impact fracture mode was a combination of ductile and brittle fracture with a propensity towards brittle fracture.

Disclosure statement
No potential conflict of interest was reported by the author(s).
Peng Liu, Professor, is mainly engaged in the welding research of high-performance material and additive manufacturing technology for Ni-based alloys and non-ferrous alloys.
Yong-bin Wang, PhD candidate, is mainly engaged in the welding of heat-resistant steel.
Xin-fang Guo, a welding engineer and manager, is mainly engaged in the research of welding process of nuclear power and thermal power heat-resistant steel.

Figure 1 .
Figure 1.Image of welded plate and schematic diagram showing the SMAW + GTAW process.
For the XRD analysis, the D/MAX-rc RU-200B equipment using copper target (Kα) radiation was used.The test conditions are 40 kV voltage and 200 mA current.The microstructure studies were carried out on Nikon Epiphot 300U/200 horizontal metallographic microscope and Zeiss SupraTM 55 scanning electron microscope instruments.The main parameters of the latter are extra high tension (EHT) = 15 kV, working distance (WD) = 9∼11 mm, magnification (Mag) = 5000 X and detector type (signal A) = SE 2.

Figure 2 .
Figure 2. (a) The structure image of G115 heat-resistant steel by arc welding; (b) Schematic diagram of the impact sample.

Figure 3 .
Figure 3. (a) OM image of the G115 BM; (b) SEM image of the G115 BM with EDS surface scanning results after PWHT.

Figure 4 .
Figure 4. Microstructures on WZ of the G115 steel under different locations.(a), (b) weld cover; (c), (d) weld centre; (e), (f) weld root.The former were OM images and the latter SEM images.
, g represent the SEM magnification area, and the high magnification images are shown in Figure 5b, e, h.Figures 5c,

Figure 6 .
Figure 6.XRD analysis on WZ of the steel under different weld locations.

Figure 7 .
Figure 7. Microhardness distribution in weld zone with fitting result.

Figure 8 .
Figure 8. Impact fracture morphology of the WZ under different sampling position: (a) cover; (b) centre and (c) root.

Table 1 .
The welding process parameter of G115 heat-resistant steel.

Table 2 .
Charpy V-notch impact properties of the G115 welded joints.