Introduction

Additive manufacturing (AM) has attracted significant attention for metallic component manufacturing in that it offers a wide variety of advantages compared with traditional manufacturing methods such as reductions of material waste, processing complexity, lead time, and transportation cost, and also it provides much more freedom concerning the geometry and alloy constituent design.[1,2] Powder-based laser metal deposition (LMD) is a directed energy deposition (DED) type of metal AM techniques. In this process, a melting pool on the build surface is generated by the laser beam first. A nozzle simultaneously feeds metal powders into the melt pool. This creates beads that are welded to one another. The process is repeated layer by layer until the entire structure is formed on the substrate. In general, LMD is still regarded as less competitive in manufacturing structural components as compared to the traditional manufacturing methods, because of several major limitations, including porosity,[3] low surface precision,[4] detrimental residual stress distribution,[5] and hard-to-control microstructure.[6] To eliminate material defects and thus improve the mechanical properties of additively manufactured metal parts, various post-treatment methods have been adopted, such as post-heat treatment,[7] isostatic pressing,[8] and shot peening.[9] Although the post treatments could enhance the mechanical properties to a certain extent, they have various disadvantages, such as high energy consumption, long lead time, and adverse effect on part precision. Therefore, in-situ material modification during metal AM processes has become an important research topic thanks to the simplified processing procedure, high integrability, and reduced cost. More importantly, in-situ material modifications may improve the mechanical properties and microstructure of the entire component, instead of being limited to the surface region.

In the related literature on metal AM research, various techniques traditionally developed for surface modification have been explored for AM applications. For instance, Aimangour demonstrated the improved surface quality and hardness of additively manufactured 17-4 stainless steel by post-shot peening.[10] Post-heat treatment of shot peened parts further promoted the surface hardening effect.[11] Similarly, Uzan et al. applied post-shot peening to improve the fatigue properties of SLM-produced AlSi10Mg.[12] Laser shock peening was also applied as an effective post-treatment measure to refine surface microstructure and obtain multi-directional mechanical twins.[13] Post-laser shock peening was reported as helpful to obtain homogenous microstructure and better shape memory effect for additively manufactured NiTi alloy.[14] Donoghue et al.[15] applied in-situ rolling deformation for wire arc additive manufacturing (WAAM) process. A 100-mm-diameter roller ran across the top of the deposited walls after each layer of deposition. The results demonstrated that normally coarse centimeter-scale columnar grain structure could be refined down to 100 µm. Zhao et al.[16] conducted similar research by combining synchronous rolling and laser multilayer cladding. It was found that columnar dendrite crystals broke into uniform equiaxed crystals, which enhanced the anti-wear performance of the deposited material. Kalenticsa et al.[17] applied in-situ layer-wise laser shocking peening (LSP) during SLM of 316 stainless steel. A significant increase in the magnitude and depth of compressive residual stress was obtained with in-situ LSP, as compared to the SLM parts, or those parts traditionally treated on the surface by LSP as a post process.

A few studies have reported using ultrasonic impact peening (UIP) as a material modification technique by taking advantage of its low cost, easy operation, and excellent strengthening effect. Zhang et al.[18] applied UIP every two layers during SLM of Ti6Al4V alloy. The results indicated that through in-situ UIP, the defects could be hammered flat or even eliminated. Meanwhile, the epitaxial growth of columnar grains was prevented, and fine equiaxed grains were formed due to plastic deformation and recrystallization. Gale et al.[19] investigated the effect of in-situ UIP on material behavior during SLM of 316L stainless steel. Their results indicated that the mechanical properties including hardness and yield strength were significantly improved. However, both Zhang’s and Gale’s works were based on powder bed fusion type of AM process, in which the powder processing, e.g., filling and removal of powder from the bed, was complicated during the intermittent cycles of powder fusion and peening. The challenge can be greatly alleviated by using a DED type of AM process, and thus in this study the in-situ UIP was attempted in combination with the powder-based LMD process. More importantly, the underlying mechanism of microstructure refinement during UIP-assisted AM processes has been lacking. In addition, due to severe plastic deformation and stress generation resulted from the in-situ UIP, it is reasonable to hypothesize that a post-heat treatment would influence the final microstructure and mechanical properties of AM materials. However, such investigation is still uncharted. Therefore, this study aims to fill these gaps by combining laser metal deposition and in-situ UIP to improve the microstructure and mechanical properties of Inconel 718 (IN718), and to develop a good understanding on the strengthening mechanism. Also, a standard post-solution treatment is applied to study the combined effect of in-situ peening and post-heat treatment on the material behavior.

Material and Methods

A laser metal deposition system was employed to additively manufacture the IN718 samples. The system consists of a 2000 W fiber laser unit with a beam size of about 2 mm, a multi-axis CNC platform, an argon protective environment, a multi-hopper powder feeder, and a coaxial conical focused powder feed nozzle. The commercially available IN718 powders were used as the starting material. The powder diameter ranged from 20 to 180 μm and has the average value of 110 µm. The particle morphology and size distribution are shown in Figure 1. Table I summarizes the chemical composition of IN718 powders. In the LMD experiment, a 20-mm-thick IN718 plate was used as the substrate. IN718 was chosen as the target material because it is a widely used superalloy in various industries.[20] It is a nickel-based alloy with a substantial amount of iron. The addition of chromium contributes to the excellent anti-corrosion performance, and the fine precipitates resulted from solution and aging treatment lead to a significant strengthening effect.[21,22,23] The main LMD parameters took the following settings: laser power was 1500 W, powder feed rate was 30 g/min, laser scan speed was 20 mm/s, and hatch space was 0.8 mm. After each layer of deposition, the material was cooled to below 100 °C. Then a UIP treatment was applied uniformly on the material surface. The UIP treatment adopted a flat peening head with 5 mm in diameter, and a peening frequency of about 28 kHz. The output power of the UIP unit was 1 kW. The peening head assembly was held vertical above the LMDed material and no additional force was applied along vertical direction, meaning that the contact force between the peening head and the material surface was the gravity force of the peening head assembly (around 25 N). The lateral moving speed of the peening head was 3 mm/s. The snapshots of the proposed hybrid manufacturing process were shown in Figure 2. Both UIP and LMD adopted the same unidirectional scan strategy along the X direction, as shown in Figure 3. It should be noted that the LMD-only region and LMD + UIP region were built simultaneously. However, due to the drastic material compression resulted from the in-situ peening, LMD + UIP region have slightly smaller layer thickness as compared to LMD-only region. Thus, one more layer was deposited on the LMD + UIP region after the LMD-only region reached the desired height of 5 mm, so that eventually both regions exhibit the same height. Specifically, the LMD-only region consists of 13 layers and the LMD + UIP region consists of 14 layers. The average layer thicknesses are about 385 and 357 μm, respectively, for the LMD-only and LMD + UIP regions.

Fig. 1
figure 1

Inconel 718 powder information (a) powder morphology and (b) particle distribution

Full size image
Table I Chemical Compositions of IN718 (in Weight Percent)
Full size table
Fig. 2
figure 2

Snapshots of LMD-UIP hybrid manufacturing process (a) LMD process and (b) UIP process

Full size image
Fig. 3
figure 3

Manufacturing strategy of in-situ UIP-assisted LMD process

Full size image

The deposited material was then detached from the substrate by wire EDM cutting. A post-solution treatment at 980 °C in a vacuum furnace with 1-hour holding, followed by air quench,[7] was carried out for selected samples.

The material microstructure was characterized using an optical microscope system (model: Amscope X900) and an FEI Quanta SEM. The elemental composition of interested areas was acquired by an Oxford silicon EDS (energy dispersive spectroscopy) detector. For EBSD analysis, samples were sectioned perpendicular and parallel to the build direction. To avoid residual deformation on the sample surface and obtain high-quality diffraction images, the EBSD samples were mechanically polished and then electro-polished for 10 seconds in a 5 pct perchloric solution at 20 V and – 10 °C. A JOEL100CXII TEM system was used to characterize the material microstructure at nano scale. Thin foils for TEM observations were prepared using a twin jet electropolish thinning system under 20 V and − 20 °C, and the composition of the electrolyte was 90 pct C2H5OH + 10 pct HClOH. The residual stress was measured by a side inclination method using a Proto LXRD system. It was measured along the build direction (Z) and horizontal direction (X), and the profiles of in-depth stress were recorded. In the measurement, a Cr tube operated at 30 kV and 10 mA was used for producing Kβ X-rays. The monochromatic beam was about 50 µm size. To obtain the residual stress distribution along the vertical direction, σz, 10 points along the depth direction (Z) with a separation distance of 300 µm were analyzed. Moreover, to measure the horizontal residual stress σx without the errors induced by stress release, electro-polishing was performed to remove the material between two adjacent measurements. For each residual stress measurement, diffraction was at 5 angles, namely, ψ = 0, 18, 27, 33, and 45 deg. The adopted diffractive face is the (311) crystal plane of nickel, corresponding to a Bragg angle (2θ) of 157.73 deg. The residual stress was calculated by the sin2ψ method (i.e., obtaining the gradient of the line from five points on the 2θ − sin2ψ diagram using the least-square method). A micro Vickers hardness tester (Model: MC010) was used to examine the microhardness of prepared samples. All hardness and residual stress measurements were performed along the center axis of each sample.

Results and Discussion

Microstructure Observation

Figure 4 shows the side view of an as-built sample on the substrate. The left half of the sample is processed by LMD-only, and the right half is processed by LMD + UIP hybrid process. It is observed that whether or not the in-situ UIP is performed, the material can be deposited successfully on the substrate without apparent macroscopic defects, such as cracks, debonding, deformation, etc. The LMD-only portion of the sample exhibits an obvious rough surface with loose particles attached to it, while the LMD + UIP portion of the sample shows a finer surface finish. The top surface is smoothed and shows metallic luster thanks to the peening process, and loose particles are hardly observed.

Fig. 4
figure 4

The as-deposit IN718 workpiece made by LMD-only (left) and LMD + UIP hybrid process (right)

Full size image

Figures 5(a) and (b) show the microstructure of the as-built samples, in which the view planes are parallel to the build direction. The horizontal melt pool tracks can be clearly seen in both cases. Directional columnar dendrites are found to dominate the microstructure. Because heat rapidly dissipates to the solidified material from the melt pool, highly directional heat flux accompanied with very high cooling rate is present during the solidification process. Thus, the columnar grains tend to grow along the build direction with abundant sub-grain dendritic structure. The repetitive melting and cooling during the additive process also favor the columnar grains to grow across a number of layers. Interestingly, the layer thickness of the sample with in-situ peening is found to be slightly smaller than that without peening. For instance, the average layer thicknesses are about 460 and 430 µm for the cases without and with in-situ peening, respectively. Since the powder feed rate is constant for both LMD-only and LMD + UIP regions, the reduction of layer thickness in the LMD + UIP region is apparently caused by the plastic deformation resulted from in-situ peening. Moreover, a zoomed SEM image shown in Figure 5(d) shows clear columnar-to-equiaxed transition on the interlayer regions. This type of transition is less often observed in the samples without peening, where most grains continuously grow cross layer boundaries, as shown in Figure 5(c). In order to explain the columnar-to-equiaxed transition phenomenon, the relation between mechanical deformation and material microstructure evolution must be taken into consideration. During the manufacturing stage of LMD + UIP sample, there are two major mechanisms that should be accounted for the formation of the fine and equiaxed crystals. The first one is dynamic recrystallization upon peening, when the material undergoes heavy deformation. The material deformation due to ultrasonic peening is significant. It results in the high compressive stress in the interlayer region, providing a driving force for the occurrence of mechanical twining and dynamic recrystallization. The second mechanism is annealing recrystallization during laser scan of the subsequent layer. An existing layer undergoes significant compressive stress after UIP and compressive plastic deformation can induce high dislocation density and stacking fault energy. The deposition of the subsequent layer introduces excessive heat to the existing layer. As a result, the rapid heating in combination with the stored deformation energy triggers effective annealing recrystallization in the interlayer regions. However, due to prominent gradient strain distribution (e.g., higher strain close to surface) and the rapid cooling of laser scan, both dynamic recrystallization and annealing recrystallization only occur close to the top surface of every new layer, also represented as the interlayer region. As a result, the equiaxed recrystallization region lies in between the two adjacent layers at the as-built condition, as shown in Figure 5(d).

Fig. 5
figure 5

Microstructure of the as-built samples processed by (a) LMD-only optical image, (b) LMD + UIP optical image, (c) SEM image of an interlayer region in (a), and (d) SEM image of an interlayer region in (b)

Full size image

Figures 6(a) and (b) present the sub-grain microstructures of LMD-only and LMD + UIP materials, respectively. Dendritic cast structure due to fast cooling and microsegregation is clear, and the interdendritic region is decorated with a large amount of precipitated particles. The hard-to-dissolve elements (e.g., Nb, Mo) are pushed outside of the solidification front during fast cooling, and become enriched at the interdendritic regions. In the investigated range, the application of in-situ UIP hardly affects the morphology of dendritic pattern. The interdendritic spacing is measured, and the average values turn out to be about 10 µm for both cases. The EDS line scan result is shown in Figure 6(c). It can be clearly seen that the interdendritic regions are rich in Nb and Mo and lack of Cr. The high concentration of Nb and the chain-like geometry indicate those interdendritic particles are most possibly Laves phase.[24] TEM observations further explain how nanoscale defects behave upon going through the UIP process and consequently influence the material microstructure. The samples are taken from around 200 µm below the top surface of samples. Figures 7(a) and (b) show the bright-field TEM micrographs of a LMD-only sample and a LMD + UIP sample, respectively. The high density of dislocations is visible for both cases. As shown in Figure 7(a), due to the blocking effect of grain boundaries, dislocations resulted from high thermal strain are found in the vicinity of grain boundaries. The hindering effect on dislocation propagation by grain boundaries enhances the material hardness and strength as more energy is required to generate new dislocations in more grains. This is consistent with literature findings that metal parts produced by laser additive processes are hard and brittle.[25] Due to the severe plastic deformation induced by UIP, mechanical twinning is clearly visible as shown in Figure 7(b). The mechanical twinning happens at the nanoscale and the width of the twin grain is measured to be about 250 nm. It is known that mechanical twinning can occur in lattice structures when it is highly strained in that they do not have sufficient slip systems to account for the arbitrary shape changes.[26] Literature studies have also demonstrated that deformation-induced microstructural refinement features are identified as mechanical twinning.[27,28] Meanwhile, mechanical twinning is an important indicator of dynamic recrystallization.[29] Wang et al.[30] observe fractions of 20 pct of twins at the interface between initial grains and dynamically recrystallized grains when austenitic steel alloy 800H is compressed to a strain of 0.49 at 1100 °C. It can be concluded here that the severe plastic deformation induced by UIP is responsible for producing defects in the material such as dislocations and stacking faults, while the resulted deformation twinning and recrystallization, in turn, lead to a refinement of the microstructure.

Fig. 6
figure 6

SEM observation of dendritic microstructure of (a) LMD-only, (b) LMD + UIP, and (c) EDS line scan of elemental distribution for LMD-only

Full size image
Fig. 7
figure 7

TEM micrographs from 200 µm below the top surface under the as-built conditions (a) LMD-only sample and (b) LMD + UIP sample

Full size image

Crystallographic Texture

The grain morphology and crystallographic texture under the as-built and solution-treated conditions are investigated using EBSD technique. The inverse pole figure mapping and pole figures are shown in Figures 8 and 9, respectively. EBSD data are collected from the millimeter-scale regions of sectioned surfaces to ensure that the results are statistically representative of the bulk material as much as possible. The view plane is parallel to the build direction (Z). The diffraction results are indexed according to pure Ni. Overall, the effect of UIP on microstructure is significant for both the as-built and solution-treated conditions. Under the as-built condition, the grain morphology is coarse with an average grain size of about 420 µm. The material also exhibits a strong crystallographic texture close to the 〈100〉 direction, as shown in Figure 9(a). The multiple of uniform density (MUD), an useful indicator of texture strength, has a maximum value of 3.07. When the in-situ UIP treatment is applied, a large amount of fine recrystallized crystals can be observed with more equiaxed morphology. In the investigated area, the average grain size is significantly reduced to 280 µm. Moreover, the material texture is weakened. The preferred 〈100〉 texture as seen in the LMD-only case obviously diminishes, and it tends to move toward 〈111〉 direction, as shown in Figure 9(b). The MUD value shows an overall decreasing trend with a maximum value of 1.96, indicating that the material becomes less anisotropic. The influence of post-solution treatment is also found to be significant. For the LMD-only case, as shown in Figures 8(c) and 9(c), the grain morphology and grain size do not change much after solution treatment, and the overall texture distribution is hardly affected, while the MUD value shows a slight drop from 3.06 to 2.65. However, a further microstructure refinement can be seen for the LMD + UIP case after solution treatment, as shown in Figure 8(d). A large number of equiaxed grains emerge in the entire investigated area due to the annealing recrystallization. This can be attributed to the combined effect of a proper post-solution temperature and the UIP-induced drastic plastic strain. The average grain size reduces to about 80 µm. Furthermore, the maximum MUD value reduces to 1.7, indicating a significantly homogenized microstructure with minimal anisotropy. In the post heat-treatment stage, annealing recrystallization plays the major role in refining microstructure because the treatment temperature is well above the recrystallization temperature of Inconel 718.[31] The long heating time as well as the stored deformation energy induced by peening is able to complete full recrystallization of materials, and this leads to significantly refined and equiaxed grain morphology in the bulk material. The abundant annealing recrystallization can be further validated from the grain boundary misorientation angle distribution, as shown in Figure 10(d). For all cases except the UIP + heat treatment, the highest fraction of grain boundary misorientation angle lies in the < 5 deg region, which belongs to the sub-grain boundary. For the solution-treated LMD + UIP sample, the fraction of 60 deg misorientation angle is significantly promoted as those boundaries correspond to the primary recrystallization twin boundary of Inconel 718 matrix with misorientation angle of 60 deg about the 〈111〉 axis.[32]

Fig. 8
figure 8

EBSD inverse pole figures maps for materials under various conditions (a) LMD-only as-built, (b) LMD + UIP as-built, (c) LMD-only solution treated, and (d) LMD + UIP solution treated

Full size image
Fig. 9
figure 9

Pole figures obtained from EBSD for materials under various conditions (a) LMD-only as-built, (b) LMD + UIP as-built, (c) LMD-only solution treated, and (d) LMD + UIP solution treated

Full size image
Fig. 10
figure 10

Distribution of grain boundary misorientation angles for materials under various conditions (a) LMD-only as-built, (b) LMD + UIP as-built, (c) LMD-only solution treated, and (d) LMD + UIP solution treated

Full size image

Figure 11 shows the microstructure of LMD + UIP sample after one-hour solution treatment at 980 °C. It can be observed the heat-treated material is effectively homogenized as most dislocation tangles and interdendritic microsegregation particles in the as-built samples are now absent. Needle-like δ precipitates with length from 50 to 100 μm and width about 50 nm are uniformly distributed over the entire matrix. They have orthorhombic structure and Ni3Nb stoichiometric composition. The solution treatment temperature of 980 °C is in the precipitation range of δ phase, which grows along a certain direction in the matrix because the preferential growth direction of δ phase is along (100) plane as reported previously.[33] For both LMD-only and LMD + UIP cases, the amount and morphology of δ precipitates are similar.

Fig. 11
figure 11

Precipitation of δ precipitates in solution-treated LMD + UIP samples (a) SEM and (b) TEM observations

Full size image

Residual Stress

Residual stress distribution is also an important indicator of material homogeneity, and it significantly affects material performance, such as crack resistance, corrosion resistance, and fatigue strength. Residual stress distributions along the depth direction are measured, and two stress components, e.g., along the horizontal direction (σy) and along the vertical direction (σz), are depicted in Figure 12. The profiles of σy along depth direction at the as-built and at 980 °C solution heat-treated conditions are shown in Figure 12(a). For the as-built condition, it can be observed that the horizontal residual stress (σy) is overall tensile for the LMD-only condition. The maximum value occurs at the very top layer with a value of 158 MPa, and the stress value gradually decreases as the depth increases and then becomes stabilized around 125 MPa with some oscillation. The prominent tensile residual stress along the horizontal direction is caused by the uneven heating and cooling during LMD. The schematic of residual stress generation and evolution during the LMD + UIP process is shown in Figure 13. During LMD, the fast-moving laser beam interacts with the metal powders on the top surface, resulting in a longitudinal heat-affected zone and molten pool in the new layer along the laser scan direction (y-direction). Upon cooling/solidification, shrinkage of newly added layer occurs above the solidified layers. Due to the tight metallic bonds between two layers, the shrinkage is constrained and thus a significant tensile stress is generated on the newly deposited material. In the meantime, a bending torque is generated on the underlying layers and causes the deposited material to bend upward. The effect of in-situ UIP on residual stress distribution is found to be significant. With UIP, the profile of σy shows more drastic oscillation—either compressive or tensile σy can be observed from the stress profile. However, its average magnitude is significantly reduced to around 15 MPa. The in-situ UIP is not most effective to introduce compressive σy because the peening is applied in the vertical direction, and therefore plastic deformation mainly takes place along the vertical direction. Meanwhile, the effect of post-solution treatment on residual stress component σy is noticeable. For all experiment conditions, σy significantly reduces to a more constant value, around 20 MPa after the solution treatment. The increasing or decreasing trend along the depth direction as seen for the as-built condition also becomes less prominent. This illustrates that the post-heat treatment is effective in homogenizing the material microstructure and reducing the detrimental residual stress. The average values of σy are 21 and – 14 MPa under LMD + solution-treated and LM + UIP + solution-treated conditions, respectively.

Fig. 12
figure 12

Residual stress profile along the depth direction from the top surface under various processing conditions showing (a) horizontal stress component σy, and (b) vertical stress component σz

Full size image
Fig. 13
figure 13

Schematic of residual stress generation during LMD + UIP hybrid manufacturing process

Full size image

The effect of in-situ peening and post-solution treatment on the vertical component of residual stress, σz, is shown in Figure 12(b). It is obvious that the in-situ peening is more effective in controlling the residual stress along the vertical direction than that in the horizontal direction. At the as-built condition, σz of the LMD-only sample is overall tensile with an average value of 53 MPa, However, it becomes compressive after the in-situ UIP is applied. The compressive stress reaches the maximum magnitude of – 190 MPa at the top layer, and it gradually decreases as the depth increases to 2.5 mm and finally becomes stabilized around −110 MPa. It should be noted that the last layer of deposit (i.e., the top layer) is subject to the final UIP treatment without further re-melting/re-heating. The stress profile close to the surface thus somehow reveals the depth of UIP-affected zone. From Figure 12(b), it can be seen that the depth of the UIP-affected zone is more than 2.5 mm because the magnitude of σz tends to be stable beyond this depth. In this UIP-affected region, a large amount of stored energy induced by material deformation is released during the subsequent laser re-heating process and provides sufficient driving force for recrystallization and grain refinement. The proposed LMD + UIP hybrid process is somewhat similar to a traditional forging process. Therefore, it is reasonable to observe significant compressive residual stress inside the bulk material. Although all the residual stress measurements give compressive results in the LMD + UIP case, tensile residual stress should be expected in non-measurement regions to balance the compressive stress as a free body. For instance, literature shows that tensile residual stress still presents close to the edge and cavity region of forged Inconel part.[34] After the solution treatment, σz drops in both cases, but the UIP-treated sample still retains significant compressive σz, with an average value of − 35 MPa. The relations of material hardness with respect to depth are also depicted in Figure 14. For the LMD-only sample, the hardness value is hardly affected by the depth from the surface, and it is stable with an average value of about 345 HV0.2. Due to the severe plastic deformation, the application of in-situ UIP effectively promotes the surface hardness increase by 30 pct to 450 HV0.2. For the bulk material, the overall increase in hardness is around 20 pct. After the heat treatment, the hardness of the UIP-treated sample is still significantly higher than that of untreated samples. The average hardness values are 355 and 265 HV0.2, with and without the UIP treatment, respectively. It is believed that the efficient grain refinement makes the classical Hall–Petch theory to become relevant.

Fig. 14
figure 14

Hardness vs depth profile under various processing conditions

Full size image

Conclusions

In this study, a novel hybrid manufacturing process combining laser metal deposition and in-situ ultrasonic impact peening is used to produce Inconel 718 parts. The effects of in-situ layer-wise UIP and post-solution treatment on material microstructure, residual stress distribution, as well as hardness are investigated. The underlying strengthening mechanism is revealed by various material characterization techniques. The results show that this process is capable of producing high-quality metal parts with significantly refined microstructure and beneficial residual stress distribution. The main findings are summarized as follows:

  1. 1.

    The severe plastic strain introduced by in-situ UIP, and the resulted mechanical twinning and dynamic recrystallization play an important role in refining the microstructure of IN 718 in laser metal deposition. The in-situ UIP also leads to more uniform and less texturized microstructure.

  2. 2.

    The material can be effectively homogenized after 1-hour 980 °C solution treatment, as most microsegregation patterns and dislocation tangles observed under the as-built condition disappear. The tensile residual stresses are also significantly alleviated.

  3. 3.

    The post-solution treatment further refines the microstructure due to annealing recrystallization. A large number of equiaxed grains emerge from the deformed microstructure, and the average grain size is reduced from 280 to 80 µm.

  4. 4.

    The in-situ UIP effectively improves the residual stress distribution along the vertical direction, and its effect on the horizontal residual stress distribution is less significant. The UIP treatment alters the residual stress component σz from tensile to compressive, with a maximum magnitude of – 190 MPa on the top surface.

  5. 5.

    At the as-built condition, the microhardness is increased by 30 pct in the surface region and by 20 pct in the interior when the in-situ UIP treatment is employed, as compared with the LMD-only samples. At the solution-treated condition, the microhardness is increased by about 40 pct when the in-situ UIP treatment is employed, as compared with the LMD-only samples.

This study represents the first step to demonstrate the feasibility of the proposed novel technique, and several extension directions should be considered. We plan to conduct material property tests to measure the tensile properties of the materials prepared by this technique. Other mechanical properties such as fatigue and creep will also be evaluated. The results will provide insightful information for many potential applications of this technique. Meanwhile, the effects of LMD and UIP parameters are worth investigation, and the effort can lead to the optimization of process parameters for various materials and applications.