Arc-based directed energy deposited Inconel 718: role of heat treatments on high-temperature tensile behavior

Abstract

This study evaluated the effect of dedicated heat treatments (1050°C, 1100°C, 1142°C, and 1185°C/2 h + double-aging) on the uniaxial tensile properties at elevated temperature (650°C) of Inconel® 718 fabricated via arc plasma directed energy deposition. They enabled to meet, for the first time, the AMS 5662 requirements at elevated temperature. Tensile tests exhibited ductile strain–stress curves. The 1100°C/2 h + double-aging showed the best performance (YS_0.2%, UTS, and elongation of 967 MPa, 1126 MPa, and 18.7%, respectively). Additionally, vertical specimens evidenced dynamic strain aging, although no brittle-like features were observed.

GRAPHICAL ABSTRACT

IMPACT STATEMENT

Dedicated heat treatments developed for Inconel® 718 fabricated via arc plasma directed energy deposition enabled compliance with AMS 5662 requirements at elevated temperature (650°C).

KEYWORDS:

• Inconel 718
• heat treatment
• tensile behavior at elevated temperature
• crystallographic texture
• arc plasma directed energy deposition

1. Introduction

The Nickel-based superalloy 718 (Inconel® 718—IN-718) is a precipitation-strengthened material typically applied in hot sections of jet engines and power generation gas turbines due to its outstanding oxidation resistance and mechanical strength at elevated temperatures [1–3]. This alloy has moderated machinability and superior weldability compared to Ni-based superalloys typically applied at elevated temperatures (e.g. Hayness 282 and Waspaloy) [4–8], making it appropriate for fusion-based additive manufacturing (AM). Among the AM processes, arc plasma directed energy deposition (AP-DED) stands out due to its higher deposition rate and lower feedstock material (solid wire) and 3D-printer cost [9,10], being suitable for fabricating and/or repairing medium/large parts (e.g. flanges and open ports) [11]. However, a coarse and oriented microstructure with significant interdendritic segregation is often observed in Ni-based superalloys fabricated via AP-DED due to its lower cooling rate compared to laser-DED and powder bed fusion (PBF) [12–17], which may limit the response of IN-718 fabricated via AP-DED (IN-718 AP-DED) to the mandatory post-deposition heat treatment (PDHT) [18,19].

Seow et al. [20] and Kindermann et al. [21,22], IN-718 AP-DED, reported that low-temperature PDHTs, e.g. direct-aging and solution + aging (AMS 5662), have a low dissolution effect on eutectics (Laves and MC-type carbides). These eutectics are rich in Nb, leading to a heterogeneous distribution and lower content of γ^″ compared to the homogenous material (wrought) [23,24]. Xu et al. [25] highlighted that the AMS 5662 PDHT promoted the δ phase precipitation in Nb-rich zones. However, high-temperature PDHTs (AMS 5383 and API 6ACRA) [20–22] showed a remarkable Laves dissolution and low δ content despite the AMS 5662 requirements were not met. Furthermore, Xi et al. [26,27] (IN-718 AP-DED) demonstrated that the highest PDHT temperature (1185°C—eutectic incipient melting temperature for IN-718 welded [28]) could almost completely dissolve the Laves and drive a high hardening phases (γ′ and γ^″) content, meeting the yield strength and elongation at room temperature requirements (AMS 5662).

These findings [20–22,25–27] indicate that the typical PDHTs applied to wrought (AMS 5662 and API 6ACRA) and cast (AMS 5383) material did not perform equivalently for IN-718 AP-DED, diverging from IN-718 fabricated via PBF and laser-DED, which met the wrought material requirements using typical PDHTs [29,30]. Moreover, for applications at elevated temperatures, IN-718 AP-DED must also have an equivalent behavior to wrought material, which has not yet been achieved. Seow et al. [20], Zhang et al. [31], and James et al. [32] reported that heat-treated IN-718 AP-DED (modified AMS 5383 and AMS 5662) exhibited inferior performance compared to other AM processes [33–35] and non-compliance with AMS 5662, which emphasizes the importance of developing PDHTs specifically tailored for IN-718 AP-DED to allow its use in critical engineering applications.

According to literature survey [20–22,25–27,31,36–41], dedicated PDHTs for IN-718 AP-DED should consider higher homogenization temperatures (≥1050°C) and longer soaking times (1–2 h) to effectively dissolve the Nb-rich eutectics. Additionally, PDHTs should not induce secondary grain growth [42], explaining the homogenization soaking time range, which was selected following the AMS 5383. To prevent δ phase precipitation, the solution step was avoided (per API 6ACRA and in contrast to AMS 5383) [43]. Therefore, given the industrial demand of PDHTs for IN-718 AP-DED and the limited studies [20,31,32] addressing its behavior at elevated temperatures, this work analyzed the effects of dedicated PDHTs on the uniaxial tensile behavior at 650°C of IN-718 AP-DED.

2. Materials and methods

IN-718 single-bead multi-layer parts were built in an in-house developed DED 3D-printer—welding machine (CITOWAVE III 520) mounted in a 3-axis CNC [12,13]. The feedstock material (solid wire; 1.2 mm) was deposited on a hot-rolled steel plate. The printed parts underwent proposed PDHTs, which consist of two steps: (i) homogenization (1050°C, 1100°C, 1142°C, and 1185°C/2 h) and (ii) double-aging (AMS 5662).

Synchrotron X-ray diffraction (SXRD) in transmission mode (0.14235 Å) was used to characterize the heat-treated IN-718 AP-DED. 2D Debye–Scherrer diffraction rings were acquired with a Perkin Elmer detector. The data post-processing was carried out in Fit2D [44]. The pole figures estimation from SXRD data followed the Wenk and Grigull procedure [45].

Thermodynamic simulations were performed to support microstructure analysis. The Scheil–Gulliver model (Thermo-Calc®, TCNI11 and MOBNI5 database) considers back diffusion, carbon fast diffusion, and a cooling rate of 10²°C/s [16,46]. The estimated chemical compositions (dendritic core and interdendritic region) were utilized to simulate the IN-718 AP-DED time-transformation-temperature (TTT) diagram (JMatPro®).

The response of IN-718 AP-DED to PDHTs was evaluated following the AMS 5662 procedure. Uniaxial tensile tests (650°C) were conducted on specimens extracted along the horizontal (deposition) and vertical (building) directions.

For feedstock material composition, deposition parameters, PDHTs, and tensile test specimen details, see Supplemental material.

3. Results and discussion

3.1. Microstructure characterization

The 2D Debye–Scherrer diffraction rings (Figure 1) exhibited discontinuous characteristic features with intense dotted patterns, indicating a coarse and oriented microstructure. This observation was supported by their respective pole figures and orientation image map (Supplemental material), which showed coarse columnar grains and cube texture (100<100>) aspect, as commonly reported for fusion-based additively manufactured Ni-based superalloys [12,13,47]. During the melting pool solidification, dendrites growth direction followed the maximum thermal gradient [48–50]. Additionally, crystallographic planes and directions with low planar and linear densities (100 and <100>) possess a higher growth rate [51]. Thus, considering the multi-bead welds solidification mechanism (epitaxy and competitive growth [52–54]), the grains with higher growth rates (<100> almost parallel to building direction) [53] are favored layer-by-layer, promoting the observed coarse- and cube-texturized microstructure [20–22,25–27].

Figure 1. 2D Debye–Scherrer diffractions rings and (200) pole figures of the IN-718 AP-DED.

Figure 1 also showed that even the highest homogenization temperature was unable to alter the coarse grain- and cube-texturized microstructure (orientation image map and grain size data, Supplemental material), i.e. induce nucleation and significant grain boundary migration. Furthermore, it is worth mentioning that the difference in texture index between the heat-treated conditions can be considered negligible. Although the SXRD analyses had been conducted at the same position, it is difficult to control the position within a specific layer (close to the fusion line or in the middle of the layer), which justified the observed variation in texture index. Thus, given the volumetric characteristic of SXRD, all tested conditions showed a similar moderate cube texture aspect. Furthermore, the proposed PDHTs were unable to alter the typical primary microstructure of the as-built IN-718 AP-DED [20–22,25–27].

The corresponding diffractograms (Figure 2) showed characteristic peaks of the secondary phases (Laves and MC-type carbide), which are related to the IN-718 solidification sequence, as predicted in Scheil–Gulliver model (Figure 3(a)). These phases persisted after PDHTs, i.e. they were not completely dissolved, aligning with the literature reports [20–22,25–27,31,36,37,39–41]. Moreover, the Scheil–Gulliver model is per the experimental solidification models [55–57], which indicated that the interdendritic segregation, particularly of Nb and C (Figure 3(b)), promotes the first eutectic reaction (L → γ + MC-type carbides). With the continuous increase of Nb and Mo in the liquid and a reduction in C content (MC-type carbide formation), the second eutectic reaction (L → γ + Laves) occurred. Furthermore, despite the Scheil–Gulliver prediction of the δ phase, its formation is typically associated with the solid-state transformation from the Nb-rich zones (e.g. Laves phase) due to the multiple arc plasma DED reheating thermal cycles [16,58–61].

Figure 2. Synchrotron X-ray diffractograms of the heat-treated IN-718 AP-DED. The asterisk (*) denotes a second harmonic reflection.

Figure 3. Thermodynamic simulations of the IN-718 feedstock material. (a) Scheil–Gulliver solidification, (b) chemical composition during solidification, (c) equilibrium diagram, and (d) transformation-time-temperature diagram. The continuous, dotted, and double-dotted lines represent feedstock material, interdendritic, and dendritic core, respectively.

The equilibrium diagram (Figure 3(c)) indicates that the MC-type carbide are stable within the temperature range of PDHTs, which reinforces its presence in diffractograms (Figure 2). In contrast, even for the 1185°C/2 h PDHT (Laves phase solvus temperature for the cast IN-718—1162.8°C) [62], characteristic peaks of Laves phase can still be observed, confirming its presence after the tested PDHTs. Further, all PDHTs induced a significant eutectics dissolution in relation to as-built condition (Supplemental material). Additionally, due to the aging step (Figure 3(d)), the hardening phases (γ′ and γ^″ phases) were also identified. Figure 2 shows that all PDHTs can promote significant hardening phase content, which was emphasized by the broadening of γ peaks (‘shoulder’ aspect) and statistically equal microhardness (Supplemental material). In general, the diffractograms and pole figures had similar features, indicating that the proposed heat treatments did not alter the columnar grain microstructure and induced comparable eutectics dissolution and hardening phase content.

3.2. Tensile test at elevated temperature (650°C)

The IN-718 AP-DED exhibited an almost similar response to the proposed PDHTs (Figure 4), which can be correlated with similarities in microstructure and texture aspects and the final effect on the precipitation strengthening mechanism (Figures 1 and 2). Additionally, for the first time, it has been demonstrated that IN-718 AP-DED can meet the AMS 5662 requirements at elevated temperatures (Figure 4(b)) using the typical process route (3D-printing + PDHT). All tested conditions met the ultimate tensile strength (UTS) and elongation requirements; however, the yield strength (YS_0.2%) requirement was met only for vertical specimens. Compared to Zhang et al. [31] (AMS 5662) and Seow et al. [20] (1185°C/40 min + aging), the proposed PDHTs (Figure 4(b)) showed a significant improvement in elongation, YS_0.2%, and UTS, which can be related to the insufficient eutectic dissolution (low-temperature PDHT [31]) and undesirable secondary grain growth [20] observed by these authors. Relative to laser-DED [34] and PBF [35] (high-temperature PDHTs), the present results exhibited almost similar strength performance in the vertical direction and lower in the horizontal direction, which can be associated with the inefficiency of the heat treatments in alter the texture and grain morphology aspects of IN-718 fabricated via AP-DED (gas metal arc-based—GMA-DED) [16,24,26,27,46]. However, for low-temperature heat-treated PBF (AMS 5662) [30,33], the present results were even superior. This finding suggests that the homogenization PDHT soaking time and peak temperatures were correctly designed for IN-718 fabricated via GMA-DED. According to the literature survey [21,22,25–27,31,36,37,39–41], longer soaking time can be applied to IN-718 fabricated via GMA-DED. This material did not show undesirable secondary grain growth, unlike what was observed for IN-718 fabricated via AP-DED (plasma transferred arc) [20], laser-DED [29], and PBF [63,64]. Thus, conditions recommended for hot isostatic pressing (1100–1185°C/6 h) and ingots homogenization (1150–1190°C/20 h) [43,65] may also be adopted for IN-718 fabricated via GMA-DED.

Figure 4. IN-718 AP-DED response to post-deposition heat treatments: (a) tensile test at elevated temperature and (b–c) comparison with the literature. The references are Seow et al. [20], Zhang et al. [31], Sui et al. [34], Sun et al. [33], Teng et al. [35], and Trosch et al. [30]. ①, ②, and ③ refer to AP-DED, laser-DED, and PBF, respectively.

The vertical specimens outperformed the horizontal ones for all tested conditions. At elevated temperatures, the grain boundaries become weaker in relation to the grain core [66]. Thus, considering that horizontal specimens were positioned transversely to columnar grains, they would have a higher grain boundary area compared to vertical specimens [13,47], resulting in lower material strength. Solid solution [67], dislocation density [68], and precipitation [69] strengthening mechanisms are linearly correlated to the Taylor factor ($M$), i.e. IN-718 AP-DED strengthening mechanisms are direction-dependent [70,71]. Consequently, they can be linked with the crystallographic texture (Figure 1), where cube-texturized grains induce a lower $M$ (∼2.4; Supplemental material) in relation to a non-oriented microstructure (3.05–3.1) [71–74]. Furthermore, based on previous works [20,69,75–79], the cross-section transverse to the building direction (vertical specimens) exhibit a slightly inferior cube texture index (slightly higher $M$) compared to cross-sections parallel to the building direction (horizontal specimens), which also promoted an anisotropic behavior. Additionally, Gokcekaya et al. [80] estimated, for a cube-texturized material (IN-718 fabricated via PBF), that the majority of the grains had a Schmid factor (0.408) similar in both horizontal and vertical directions, i.e. macroscopically there is not a preferential direction that can easily active some slip system [72]. However, when the columnar grains were loaded along their principal axes (vertical direction), they experienced an almost isodeformation state from grain to grain. Consequently, due to the volumetric cube texture aspect (Figure 1), slip can occur simultaneously in most of the <001> oriented grains. Otherwise, when columnar grains were loaded transversely to their main axes (horizontal direction), they underwent an isostress state, and the deformation could concentrate in a small number of grains with a Schmid factor higher than that observed for cube-texturized columnar grains, resulting in lower stress for yielding [81,82]. These findings support the correlation between the observed anisotropy behavior of IN-718 AP-DED and its dependence on $M$ (texture aspects) and grain morphology (columnar) [83–86], explaining the observed anisotropy at elevated temperature and the challenge to meet the wrought material (non-oriented) properties (Figure 4(b)).

Figure 4(a) also shows the dynamic strain aging (DSA) phenomenon, which is characterized by the serrated aspect in the strain–stress curve [87]. According to Nalawade et al. [88], at 650°C both interstitial and substitutional elements diffuse to generate Cottrell’s atmospheres around dislocations in motion [89–91], which can cause consecutive dislocation locking and unlocking cycles, driving the serrated aspect. Beese et al. [92] did not observe DSA for Inconel® 625 fabricated via laser-DED. These authors described that the cube texture and coarse microstructure resulted in positive strain-rate sensitivity ($\dot{\gamma }$) [93] and lower work-hardening rate ($\dot{n}$) so that the critical strain necessaries to promote the DSA was not achieved. Otherwise, Banait et al. [94,95] (IN-718 lattices fabricated via PBF) observed a negative $\dot{\gamma }$ (occurrence of DSA) for all tested strain rates and temperatures, which was associated with the fine and almost non-oriented microstructure, reinforcing the correlation between the DSA and the crystallographic texture aspects in additively manufactured Ni-based superalloys. Additionally, Hayes [96] highlighted the interaction between the dislocation and interstitial elements (fast diffusion—C). If MC-type carbides remove C (promoting a sink effect) from the dislocation in motion faster than the Cottrell’s atmosphere formation, DSA will not occur (positive $\dot{\gamma }$). Thus, given that the MC-type carbides are stable (Figure 3) and disperse intergranularly (Supplemental material), the sink effect will be similar among the heat-treat conditions. Also, despite heavy elements (e.g. Nb and Mo) can also forming Cottrell’s atmosphere, its efficiency is secondary in relation to C due to the lower diffusion coefficient [88]. Therefore, the lower work-hardening (texture and coarse grain) and presence of high content of MC-type carbides (solidification sequence and inefficient of heat treatment to dissolve it) can explain the difficult to observe DSA for IN-718 AP-DED (GMA). Regarding the present work, DSA is observed (Figure 4(a)) only in the stronger and less oriented direction (vertical), particularly for higher strength conditions (1142°C/2 h + aging). For the softer direction (horizontal), it is believed that the critical strain for a negative $\dot{\gamma }$ was not achieved due to the lower $\dot{n}$ typically observed in coarse and oriented microstructures [21,22,92,97,98].

The DSA occurrence in the present work differs from Seow et al. [20] (IN-718 fabricated via plasma transferred arc DED). These authors observed expressive serration (DSA type C) in both directions, which can be related to the almost absence of cube texture (higher $\dot{n}$) after the heat treatment (recrystallization). Otherwise, Zhang et al. [31] did not verify DSA for IN-718 AP-DED, which was associated with the insignificant effect of the AMS 5662 PDHT on primary microstructure (coarse and oriented microstructure) and limited PDHT response (low material strength). Therefore, the occurrence of DSA in IN-718 AP-DED shows ambiguous results, which can be related to its process and PDHT dependence. Furthermore, the DSA was correlated with crystallographic texture and microstructure aspects, where a coarser and cube-textured microstructure showed less susceptibility to DSA [20,31,94,95], as stated by Beese’s et al. model.

Figure 5 exemplifies the fractography of the heat-treated IN-718 AP-DED (1100°C/2 h + aging), which had a transgranular ductile aspect with a fracture surface majority composed of shallow and almost equiaxed dimples [99]. The horizontal specimen showed a dimple ‘dendrite’ pattern, attributed to the remaining interdendritic incoherent secondary phases that drove microvoid nucleation (Figure 5(a.1)). These microvoids subsequently grew and coalesced during loading, resulting in a dimple ‘dendrite’ pattern on fracture surface [20,31,33,34,100]. Furthermore, it is worth mentioning that, although DSA was associated with embrittlement phenomena [90,101–104], the present work did not evidence brittle fracture aspects. In summary, all conditions had a transgranular ductile fracture and met the elongation requirement (Figure 4(b)).

Figure 5. Fractography analysis of IN-718 AP-DED.

4. Conclusion

The proposed PDHTs enable the IN-718 AP-DED to meet, for the first time, the AMS 5662 requirements at an elevated temperature (650°C). PDHT not alter the typical cube texture aspect of the IN-718 AP-DED, which showed an anisotropic behavior at elevated temperature. The vertical direction showed a higher strength than the horizonal one, which is associated with strengthening mechanism dependence on texture aspect and grain morphology (columnar). Additionally, heat-treated IN-718 AP-DED showed a typical strain–stress curve with a ductile fracture aspect. The dynamic strain aging phenomena was observed, however without significant effect on the heat treatment response and fracture mode. The present work demonstrated that the heat treatments designed for IN-718 AP-DED may differ from those commonly employed for cast (AMS 5383), wrought (AMS 5662), and PBF (ASTM F3055) materials due to its unique microstructure.

Acknowledgements

CRediT authorship contribution statement: Francisco Werley Cipriano Farias—Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Validation, Writing—original draft. Valdemar R. Duarte—Conceptualization, Formal analysis, Investigation, Methodology, Supervision, Writing—review and editing. João da Cruz Payão Filho—Software, Validation. Norbert Schell—Investigation, Resources. Emad Maawad—Investigation. M. Bordas-Czaplicki—Investigation, Methodology, Data Curation. F. Machado Alves da Fonseca—Investigation, Methodology, Data Curation. J. Cormier—Investigation, Methodology, Resources, Funding Acquisition, review and editing. T.G. Santos—Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Resources, Supervision, Visualization, Writing—review and editing. J. P. Oliveira—Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Resources, Supervision, Visualization, Writing—review and editing.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

Authors acknowledge the Portuguese Fundação para a Ciência e a Tecnologia (FCT—MCTES) for its financial support via the project UID/EMS/00667/2019 (UNIDEMI). JPO acknowledges funding by national funds from FCT—Fundação para a Ciência e a Tecnologia, I.P., in the scope of the projects LA/P/0037/2020, UIDP/50025/2020 and UIDB/50025/2020 of the Associate Laboratory Institute of Nanostructures, Nanomodelling and Nanofabrication—i3N. Funding of CENIMAT/i3N by national funds through the FCT-Fundação para a Ciência e a Tecnologia, I.P., within the scope of Multiannual Financing of R&D Units, reference UIDB/50025/2020-2023 is also acknowledged. FWCF acknowledges Fundação para a Ciência e a Tecnologia (FCT-MCTES) for funding the Ph.D. Grant 2022.13870.BD. The authors acknowledge DESY (Hamburg, Germany), a member of the Helmholtz Association HGF, for the provision of experimental facilities. Beamtime was allocated for proposal I-20210986 EC. The research leading to this result has been supported by the project CALIPSOplus under Grant Agreement 730872 from the EU Framework Programme for Research and Innovation HORIZON 2020. This activity has received funding from the European Institute of Innovation and Technology (EIT) Raw Materials through the project Smart WAAM: Microstructural Engineering and Integrated Non-Destructive Testing. Institut Pprime gratefully acknowledges ‘Contrat de Plan Etat—Région Nouvelle-Aquitaine (CPER)’ as well as the ‘Fonds Européen de Développement Régional (FEDER)’ for their financial support to part of the reported work.