key: cord-0321832-boakvomb authors: Patel, Vivek; Sali, Akash; Hyder, James; Corliss, Mike; Hyder, David; Hung, Wayne title: Electron Beam Welding of Inconel 718 date: 2020-12-31 journal: Procedia Manufacturing DOI: 10.1016/j.promfg.2020.05.065 sha: 80c9324d8f3f58388a5c0453f0e7cb7119b27dfc doc_id: 321832 cord_uid: boakvomb Abstract This paper presents a study of Electron Beam Welding (EBW) of rolled Inconel 718 (IN718) plates. The experiment utilized nine different heat inputs at three different levels of welding speed and beam current. The weld quality was characterized and ranked based on weld geometry, microhardness and mechanical properties. Fusion area and weld penetration depth were found to be proportional to the heat input. Microhardness measurement showed a wider and softer weld zone at the beam entrance. Voids, found at weld bottoms when welding at the highest beam power of 3250 watt and scanning speed exceeding 0.78 m/min, can be removed by a suitable machining process. All samples but one had tensile strength exceeding the specification for rolled IN718 (837 MPa); ductility of welded specimens also met the specification (57%) when welding with heat input less than 225 J/mm. At low heat inputs, the welded samples fractured outside of the weld zone and indicated their exceptional weld strength. The samples welded at higher heat inputs cracked and fractured at the weld. These results provide a benchmark for comparison with future study on EBW of selective laser melted IN718. Inconel 718 (IN718) is a nickel-based austenitic super alloy which can be hardened by precipitation. Due to its superior high-temperature mechanical properties and corrosion resistance, it has wide range of applications in aerospace, nuclear and petrochemical industry [1, 2] . Inconel also has good weldability as compared to other super alloys of similar chemical characteristics [3] . Advanced products operating at demanding conditions normally require components with complex geometries and tight tolerances; therefore, IN718 parts are generally difficult and costly to be manufactured by conventional manufacturing processes. Selective laser melting (SLM), a powder bed fusion technique, can produce near net shaped metals parts; however, one of the constraints on current SLM equipment is its limited working volume. Inconel 718 (IN718) is a nickel-based austenitic super alloy which can be hardened by precipitation. Due to its superior high-temperature mechanical properties and corrosion resistance, it has wide range of applications in aerospace, nuclear and petrochemical industry [1, 2] . Inconel also has good weldability as compared to other super alloys of similar chemical characteristics [3] . Advanced products operating at demanding conditions normally require components with complex geometries and tight tolerances; therefore, IN718 parts are generally difficult and costly to be manufactured by conventional manufacturing processes. Selective laser melting (SLM), a powder bed fusion technique, can produce near net shaped metals parts; however, one of the constraints on current SLM equipment is its limited working volume. There is a need to produce large components by joining smaller SLM parts. Electron beam welding (EBW) is among the promising joining processes for high quality and deep penetration [4] . Published literatures include welding of rolled or extruded IN718 by different methods; joining of additively manufactured IN718 by Tungsten Inert Gas (TIG) welding and brazing were performed by few researchers but none is yet to be found for EBW of 3D printed IN718. The objective of this research is to optimize parameters for EBW of rolled IN718 for best mechanical properties. The results will be used at a later stage as references for comparison against SLM IN718 components after joining by EBW. Various methods to join IN718 have been previously investigated. Laser beam welding and EBW are most widely researched joining methods for conventionally manufactured IN718. Whereas, Tungsten Inert Gas (TIG) welding and brazing have been explored for joining of 3D printed IN718. Laser welding of IN718 was investigated [5] . Several 1.3mm-thick plates were welded using four different heat inputs ranging from 74.5 J/mm to 126.6 J/mm. The heat input was defined as: Defect-free and full penetration welds were achieved and had a wider wine-glass shape as the heat input was increased. The CO2 laser welding was performed on pre-treated 5mm-thick plates, pre-and post-treated at three different conditions (as received, 955°C solution treatment, 955°C solution treatment + aging) [6] . A minimum power of 6 kW was required for full penetration with optimal focal point on the surface of the weld. Tensile strength of the as-received welded specimen was 1410 MPa as compared to 1380 MPa of the base metal. The least strength of 980 MPa was obtained for samples pretreated by solution treatment following by aging and no post-weld heat treatment. Many researchers have studied joining of additively manufactured IN718. Efforts have been made to join SLM IN718 by various methods. Tungsten Inert Gas welding (TIG) was performed on 3.4-mm thick specimens that were hot isostatic pressed at 1160˚C and 105 MPa for three hours. Although small grains were achieved, large sized cracks were evident in the welded microstructure [7] . Brazing of 20 x 20 x 4 mm IN718 specimens was investigated using brazing filler metal BrazeLet (BNi-2) held between the specimens at 50 kPa pressure [8] . Brazed parts were then cut to 40 x 3 x 4 mm for testing. Brazing led to formation of isothermally solidified zone at the center of joint which had 460 HV hardness and surrounded by diffusion affected zone with hardness of 490 HV, as compared to 475 HV hardness of the base metal. The shear strength of brazed samples, averaged at 802 MPa, only reached about 77% of that from the base metal. Many attempts have been made to join IN718 using electron beam in a vacuum. Microhardness measurement was used to characterize the weld and surrounding heat affected zones (HAZ) [9] . Two types of specimens were tested: the asreceived and the solution heat treated and aged. Although the HAZ in solution pre-treated samples could be identified with microhardness measurement, the surrounding HAZ in asreceived and precipitation pre-treated samples could not be distinguished from the central weld zone. After metallographic etching, scanning electron microscopy (SEM) examination could identify a HAZ that was within 100 microns for the solution-and-aged pre-treated samples. The anodic potentiostatic etching technique was recommended for identifying HAZ in IN718 welds. Some researchers depicted that introduction of electron beam oscillations reduced niobium segregation, therefore, improving the weld quality [10] . Their experimental IN718 specimens (205 x 105 x 3.1 mm, solution pre-treated) were first welded at heat inputs of 50 J/mm in oscillated conditions, and then subjected to three different post-weld heat treatments. The best results obtained in this research are summarized below. Other researchers showed that increasing heat input would increase the width of the welds along with reduction in microcracks in the weld and HAZ [11, 12] . Variation in lower values of heat inputs had minimal effect on the mechanical properties of the weld. The 2-mm thick rolled sheets were EBW'ed without pre-or post-weld heat treatment. The best mechanical properties of welded specimens (with 884 MPa tensile strength and 489 MPa yield strength) were obtained after EBW at the lowest heat input of 36 J/mm. Other researchers investigated the weld geometry due to different base metal conditions [12] . The weld geometry transitioned from stemless wine glass shape to nail head shape, when sample was pre-heated at 1100°C for one hour followed by air cooling. An alteration was reported in chemical composition of the weld region because of different pre-weld treatments. There was also an increase in total crack length from 40 µm to 675 µm due to pre-heating of the base metal at the same temperature and for the same time duration. Effect of processing parameters was also reported [13] . The fast heating and cooling rates at lower heat input would (i) reduce the niobium segregation, and (ii) decrease the fractional areas of precipitation of low melting eutectic inter-dendritic brittle phase of carbide and Laves phase. A wider HAZ was observed when EBW'ed at 120 J/mm heat input level, 70 kV voltage, and 2 m/min welding speed. A lower heat input also restricted the maximum penetration depth that can be achieved for the given conditions. Two weld passes were attempted for EBW of IN718 plates [14] . The plates (100 x 120 x 13.35 mm thick) were subjected to the first EBW pass (60 kV, 120 mA, and 11 mm/s) followed by the second pass (60 kV, 30 mA, and 11 mm/s). The authors then tested their samples at 650˚C for mechanical properties at the top, middle and bottom of the thick weld. The resulting mechanical properties varied from better to worse from bottom (with smaller grain size) to top (with large grain size) as shown in the following table. An attempt to improve mechanical properties of weldment by specific post-weld heat treatments was made [15] . However, the results were still inferior compared those form the base metal as shown below. Friction Stirring Process (FSP) as a post-welding treatment process was investigated [16] . The FSP was conducted using a WC-Co tool (Ø6 mm, 1.8 mm long). To obtain a homogenous material surface, the tool was tilted 3° forward from the vertical and argon gas was utilized to prevent surface oxidation. The tool was rotated at 200 rpm, applied a downforce of 39.2 kN, and travelled at 100 mm/min speed. The researchers compared the effectiveness of FSP as a post-weld treatment on IN718 overlap welds, with and without heat treatment (720°C for 8 hours, 620°C for 6 hours, 1 x 10 -5 Torr vacuum pressure, then air-cooling). Hardness of the base material increased by >20% after FSP and >70% after both mechanical and thermal treatments. The results indicated refinement of grain structure with reduced number of defects after FSP. Table 4 summarizes the mechanical properties of the test specimens. Majority of published literature focuses on understanding of weld geometry and microstructure, effects of welding parameters on niobium segregation, Lave formation, and ensuing defects in weld and base metal. Considering favorable consequences of different variables, it may be inferred that there exists a compromise in the selection of the process parameters to mitigate possible defects and achieve quality welds. Although many researchers have studied and suggested optimized welding conditions for rolled /extruded IN718, no published data is yet found for EBW of additively manufactured IN718. This paper establishes a reference by presenting EBW results of rolled IN718; the results will then be compared with those from EBW of SLM IN718 in a different paper. The IN718 plates (305 × 152 × 13 mm thick), provided by Huntington Alloys Corporation, were used in this study [18] . The as-received plates conforming to the SAE AMS 5596M / ASTM B670-07 standards had rolling direction (RD) along the plate length (305 mm). Its chemical composition (wt%) was 53.48Ni 18.15Cr 17.88Fe 5.17Nb 3.00Mo 0.98Ti 0.53Al 0.18Co 0.14Cu 0.09Si 0.07Mn and 0.03C. Table 5 lists the material mechanical properties. Table 5 . Mechanical properties of rolled IN718 [18] The as-received plate was sectioned using wire-type Electrical Discharge Machining (wire EDM) across the length into nine equal welding coupons (152.4 × 33.7 × 13 mm thick). Each weld coupon was engraved at both ends for ease of identification before welding. The coupons were assembled for welding with uniform spacers to mitigate heat propagation amongst the coupons (Fig. 1) . Electron beam weld beads were produced without beam oscillations on the coupons along the rolling direction using the Sciaky EBW system. The beam was focused on the surface for all coupons. Three levels of welding parameters were selected for welding speeds and beam currents. Constant beam voltage of 50 kV and low vacuum pressure <1 µTorr were used for all experiments. These parameters were selected to produce heat input within the range from 180 to 300 J/mm for deep weld penetration. A total of nine samples, with two replicates for each welding condition, were produced at different welding parameters (Table 6) . After welding, the hardness coupons (20 × 6 × 13 mm thick) were sectioned from each welded specimen for fractography and microhardness study. The rest were stress relieved (970˚C, 1 hour), milled equally on both sides to 6 mm thick, then EDM'ed to tensile specimens according to ASTM E8/E8M standard (Fig. 2) . For fractography and microhardness study, the surface of coupons, perpendicular to the RD and welded section, was hand ground with abrasive papers of 240, 320, 400 and 600 grits on the Buehler Handimet grinder. The coupons were then polished in two steps, first with 5-7 µm and then with 1 µm diamond paste. Ultrasonic cleaning in isopropyl alcohol was used to clean the samples and to remove remnant abrasives after each step. The hardness samples were etched using the Aqua Regia etchant (33% HCl + 67% HNO3) by swapping with a cotton tip for 25 s to reveal the weld regions. The 0.1-µm resolution Olympus STM6 optical microscope was used to observe weld microstructure and measure the weld profile, penetration depth, fusion zone and defects. The base metalfusion zone interface was traced, and resultant zone was scaled using image processing to obtain fusion zone area. Microhardness test on the polished surface was carried out using the Vickers hardness tester (Wilson VH1102) at 100 gf load and 15 s dwell time. The indentations were made from the weld center through HAZ to the base metal in both directions at two different weld depths of each sample (x = 0.3 and 4.8 mm). At each depth, the distance between two successive indentations was kept at least 0.18 mm to avoid possible work hardening effect from adjacent indentations. Between the two depths, four equally spaced indentations (at weld depths, x = 1.2, 2.1, 3.0 and 3.9) were made along the penetration depth on the weld centerline. Tensile tests were carried out on the MTS 810 material tester. Displacement-control approach was adopted with crosshead speed of 0.5 mm/min and load-cell of 100 kN. The data acquisition frequency was set at 5 Hz. Typical "nail shape" weld was seen for all samples (Fig. 3) . The large "nail head" shape was seen at the beam entrance (top region of the weld) and the slender and sharp "nail point" shape was seen at the deepest region of the weld. Different penetration depths, fusion areas, and defects were found depending on the (linear) heat input. Simulation of pore formation was reported for selective laser melting process when a laser beam reacts with metal powder [17] . The scanning laser formed a trench filled with molten metal below the beam. The dynamic molten metal flowed rapidly upward on both sides of the trench, then collapsed and trapped pores after the laser beam passed by. Assuming similar pore forming mechanism for energy beams such as laser or electron beam, a pore could be formed in EBW fusion zone due to either solidification shrinkage or trapping of pore from rapidly flowing and collapsing of molten metal. A low viscous molten metal would fill all pores if time permits, but a molten metal with high viscosity would flow slowly and might solidify too quickly thereby preventing the pores in the weld from being filled. Pores were seen only at the weld bottoms of the samples A4 and A7, which were welded at relatively high scanning speeds (787.4 and 914.4 mm/min) and the highest beam current (65 mA). Such conditions would promote fast cooling rates and formation Author name / Procedia Manufacturing 00 (2019) 000-000 of high viscous molten metal that fails to fill the large pores at the weld bottom (Fig. 4) . Except for samples A4 and A7 where voids existed at the weld bottom, all other samples were free of defect when viewed at the highest magnification on the Olympus microscope. As expected, the heat input directly affected both weld penetration depth and fusion area. In fact, linear relationships were found for the ranges of experimental parameters in this study . The measured penetration depths and computed fusion zone areas are tabulated in Table 6 with respect to corresponding heat input. A coordinate system is shown in Fig. 3 for each weld. The coordinate origin is at the beam entrance point, the x-direction indicates the weld depth while the y-direction is along the weld width. Fig. 7 shows microhardness of the typical sample A9 across the weld at different weld depths (x = 0.3 and 4.8 mm). The fusion zone widths are schematically represented at different weld depths by solid and dashed lines at x = 0.3 mm and x = 4.8 mm respectively. The weld zone is wider at the "nail head" location compared to slender width at the "nail stem." Table 7 lists microhardness of the same sample measured inside and outside the weld at two different depths, 0.3 mm and 4.8 mm. The bottom half of Table 7 lists the microhardness measured at the weld center along the weld depth. Typical hardness values for sample A9 (with low heat input of 180.4 J/mm) and those for sample A3 (with high heat input of 249.8 J/mm) are included for comparison. It was found that: -No obvious hardness variation was observed in the weld, HAZ, and base metal at depth greater than 1.2 mm. -Microhardness at weld center near the beam entrance (x = 0.3 mm) is lower than that at the deeper zone (x = 4.8 mm). The material in the nail head region was (i) exposed to the beam for a longer time, and (ii) surrounded with a larger pool of molten metal in the weld. Both factors would promote grain growth and a softer nail head region. -Similar hardness trends were observed regardless of the heat input values. Micro-hardness (HV) Distance from weld center (mm) at x=0.3 mm level at x=4.8 mm level at x=1.2 mm level (on the weld centeline) at x=2.1 mm level (on the weld centerline) at x=3 mm level (on the weld centerline) at x=3.9 mm (on the weld centerline) weld zone x=0.3mm weld zone x=4.8mm x y The slight difference in hardness near the beam entrance and voids at the bottom of weld zone would not affect the tensile test results since both the nail head region and voids were removed during machining to fabricate the tensile specimens. Although the experimental setup is different from true welding of two separated parts, each tensile specimen is with a full penetration weld that separates the two halves of a specimen. The tensile test in this would represent an ideal case in welding with complete penetration, precise machining and no misalignment of welded parts. Pores were found at the weld bottom of samples A4 and A7; however, these defects were machined away when fabricating the tensile specimens. Figure 8 shows consistent stress-strain curves for sample A9 and its replicate. Similar consistency was found for other identical pairs. Figures 9,10 and Table 8 show the mechanical properties of all tested EBW samples. With an exception of one sample (A6, 209.6 J/mm), all sample tensile strengths exceeded the material specification of 837 MPa. The elongation of EBW specimens, however, decreased below the specification of 57% when the heat input values were higher than 225 J/mm. Perhaps a higher temperature and fast cooling rate associated with higher heat input promotes forming of brittle Laves particles in the microstructure, therefore, degrading the ductility of welded samples. Material toughness, represented by area under the stress-strain curve, also reduced with increasing of heat input (Table 8 ). This supported the hypothesis that brittle particles might be the cause when welding at high heat input conditions. Some samples fractured within the weld zone, while other samples fractured randomly and away from the weld. Selected fractured samples were polished along the gage length and chemically etched to confirm that the fractures were actually confined within the weld zone. Study of fracture surfaces helps to explain the measured ductility results. The relative location of a fractured surface from the weld zone is found after polishing and etching along the gage length of a tensile specimen after testing. The fractured surfaces on two samples A3 (249.8 J/mm heat input) and A9 (180 J/mm heat input) are shown in Figs. 11 and 12 . The brittle sample A3 (50% ductility) fractured at the weld while the more ductile sample A9 (65% ductility) fractured away from the weld. Shown in Fig. 11a , the weld bead contains no visible crack even when viewing at a high magnification. In Fig. 11b , a series of shear bands near the fractured surface, located away from the weld zone and oriented about 45° to the tensile loading direction, suggest ductile fracture and are confirmed with the high ductility of 65% in tensile testing ( Table 8 ). The opposites are found on sample A3 where cracks are initiated and propagated within the weld zone as shown in Fig. 12a . Other secondary cracks, oriented about 90° to the tensile loading direction, are visible in the brittle weld zone as shown at the lower left of Fig. 12b . These facts explain the low ductility of samples when EBW'ed at high heat input greater than 225 J/mm. In EBW and selective laser melting, the melting and cooling of molten IN718 forms the dendrite solid solution γ phase, NbC niobium carbide, and the Laves phase [19] [20] [21] [22] [23] : Liquid Inconel 718 → γ + NbC + Laves (2) The Laves phase, forming due to fast cooling of molten IN718, perhaps are present in the weld zone when cooled from a higher weld temperature due to high heat input. Such brittle phases would be responsible for the brittle nature of the specimen. Electron beam welding of rolled Inconel 718 was performed. This study showed: 1) Voids can be formed at the weld bottom when welded at high beam power at high welding speed. The only two welded samples containing voids, out of nine samples in this study, were welded at beam power of 3250 W (65 mA, 50 kV) and at welding speed faster than 0.78 m/min. 2) All samples but one sample exhibit higher tensile strength than the specification for rolled IN718 (837 MPa). The ductility also exceeded the specification except when joining at heat input above 225 J/mm. 3) The heat inputs dictate fracture behavior of EBW'ed IN718. A low heat input resulted in strong welds since the ductile fracture surface occurred in the base material and outside of the weld zone. In contrast, a high heat input resulted in brittle fracture inside the weld zone. The brittleness perhaps caused by the presence of Laves phase formed by fast cooling in the weld a) Studying the microstructure of the weld and confirming the presence of brittle Laves particles in a weld zone when joining at high heat input. b) Comparing EBW results of rolled samples versus those from samples produced by selective laser melting. c) Improving the mechanical properties of EBW'ed IN718 with different post processing techniques before tensile testing. High temperature deformation of Inconel 718 A Review on Superalloys and IN718 Nickel-Based Inconel Superalloy Weldability of Inconel 718 -A Review Advanced Welding Processes. Cambridge: Woodland Publishing Limited A Study on Laser Beam Welding (LBW) Technique: Effect of Heat Input on the Microstructural Evolution of Superalloy Inconel 718 Microstructures and mechanical properties of Inconel 718 welds by CO2 laser welding Crack Varestraint weldability testing of additive manufactured alloy 718 Microstructure and Properties of 3D Printed Inconel 718 Joint Brazed with BNi-2 Amorphous Filler Metal A study of the heat affected zone (HAZ) of an Inconel 718 sheet welded with electronbeam welding (EBW) Improvement of mechanical properties of Inconel 718 electron beam weldsinfluence of welding techniques and postweld heat treatment Effect of Heat Input on Microstructure and Mechanical Properties of inconel-718 EB Welds Effect of base metal and welding speed on fusion zone microstructure and HAZ hotcracking of electron-beam welded Inconel 718 Studies on Electron Beam Welded Inconel 718 Similar Joints. Procedia Manufacturing Microstructures and high temperature mechanical properties of electron beam welded Inconel 718 superalloy thick plate Microstructure and mechanical properties of Inconel 718 electron beam welds Investigation of Microstructure and Mechanical Properties on Surface-Modified Inconel 718 Alloy Laser powder-bed fusion additive manufacturing: Physics of complex melt flow and formation mechanisms of pores, spatter, and denudation zones Microstructure and mechanical properties of selective laser melted Inconel 718 compared to forging and casting Electron metallography of alloy 718. Superalloys 718, 625, 706 and Various Derivatives The microstructure and mechanical properties of deposited-IN718 by selective laser melting Microstructure and hardness studies of Inconel 718 manufactured by selective laser melting before and after solution heat treatment The authors thank Mr. Rodney Inmon of Materials and Testing Lab, Department of Aerospace Engineering at Texas A&M University for helping with the tensile testing.