Comparison of Substrate Preheating on Mechanical and Microstructural Properties of Hybrid Specimens Fabricated by Laser Metal Deposition 316 L with Different Wrought Steel Substrate

The combination of additive manufacturing and conventional metal forming processes provides the possibility for improvements of forming efficiency and flexibility. Substrate preheating is an implementable technique to improve the interface adhesion properties of the hybrid forming method. The present experiment investigates the adhesion of additive manufactured 316 L steel on P20 and 1045 steel substrates under two substrate temperatures, and the geometrical characterization, interfacial microstructure and mechanical property of the hybrid specimens were compared. As a result, it was found that the ratio of deposition height to the width was reduced and the width was increased under substrate preheating. Tensile results show that the ultimate strength of 1045 and 316 L hybrid specimens was obviously increased, while the properties of hybrid specimens P20 and 316 L were similar, under different substrate temperature conditions. For the hybrid specimens with the metallurgically bonding characteristic, the tensile properties can reach the level of 316 L depositioned specimens fabricated by laser metal deposition (LMD). Furthermore, substrate preheating had little effect on the microstructure of the laser metal deposition zone, and significant influence on the microstructure of the heat affected zone, which was reflected in the difference of the hardness distribution.


Introduction
In order to improve the building efficiency of current forming technology, integrating additive manufacturing with conventional manufacturing (metal forming) processes is an economical way to fabricate large-scaled structures with local precision parts [1]. Some functional structures can be manufactured by building multilayer sections with laser metal deposition (LMD) on existing substrates using precision casting or wrought. This includes using a mold with conformal cooling channels [2,3], component repairing [4], and surface strengthening [5,6]. This hybrid manufacturing process has been thoroughly discussed for titanium alloy [7] and steel [8].
On the one hand, when steel is used for mold and die (especially high carbon steel), it is difficult to deposit many layers on conventional manufacturing substrate without defects, as it is an alloy with as follows: 0.006 C, 0.17 Cr, 0.007 Si, 0.015 Mn, 0.128 Ni, 0.026 Mo, and bal. Fe. The average particle diameter of the powder was 50~150 µm and in a spherical particle morphology.
The laser metal deposition experiment was done with a self-build device as mentioned in reference [22]. The following process parameters were selected: laser powder 1.6-2 kW, scanning speed 4-6 mm/s, and powder feeding rate 10 g/min, which were the optimized parameters for laser deposition. Based on the formal experiment using the same device, the deposition height was uniform and no interfacial defects were found under these process parameters. First, single deposition layers on the P20 and 1045 steel substrate, with a thickness of 0.5~0.8 mm, were produced under the aforementioned parameters. The same parameters were used for different substrates to compare the microstructure and mechanical property difference of hybrid samples with different substrate steel. Second, thin wall deposition layers samples with the same height as the substrate were produced by zigzag method with laser powder 1.8 kW, laser spot 3 mm and scanning speed 6 mm/s. In order to control defect formation, the previous layers' hatch space was 1.5 mm, while the hatch space of the subsequent layer was 2 mm. Both single deposition layer and hybrid forming multilayer samples were produced with substrate temperatures of 25 • C and 210 ± 10 • C, respectively. Before the hybrid samples with preheated substrate were produced, the substrate was heated to 200 • C to 220 • C by a heater, and thermocouples were fixed on the top surface to measure the temperature variation. Moreover, a temperature control system was connected to the thermocouple and heater to control, as in, facilitate/monitor any increase beyond 200 • C but below 220 • C. The preheating temperature range of 200 • C to 220 • C was selected in consideration of substrate oxidation behavior, and the difficulty of the large-scale part heating process. All deposition experiments were done in room air.
Then, the hybrid forming specimens with single deposition layer were sectioned along the longitudinal direction for geometrical characterization observation, as seen in Figure 1a. The deposition width (W1), deposition height (H2), and melt height (H1) under different processing parameters were measured by a Dino-Lite digital microscope (Taiwan Premier, (Suzhou, (JiangSu), China), for comparing the geometrical characterization's difference. The geometric characteristic of the cross-section is an effective parameter for deposition quality, thermal and subsequently mechanical analysis [23] during the laser metal deposition process, which could reflect the quality of the deposition layer. Samples of 2.5 mm in width were cut along the longitudinal direction from the multilayer hybrid forming samples. Microstructure analysis and room temperature test samples were selected, as shown in Figure 1b, and the microstructure and microhardness were done on the transverse direction section. The unit in Figure 1b is mm. Fe. The average particle diameter of the powder was 50~150 µ m and in a spherical particle morphology. The laser metal deposition experiment was done with a self-build device as mentioned in reference [22]. The following process parameters were selected: laser powder 1.6-2 kW, scanning speed 4-6 mm/s, and powder feeding rate 10 g/min, which were the optimized parameters for laser deposition. Based on the formal experiment using the same device, the deposition height was uniform and no interfacial defects were found under these process parameters. First, single deposition layers on the P20 and 1045 steel substrate, with a thickness of 0.5~0.8 mm, were produced under the aforementioned parameters. The same parameters were used for different substrates to compare the microstructure and mechanical property difference of hybrid samples with different substrate steel. Second, thin wall deposition layers samples with the same height as the substrate were produced by zigzag method with laser powder 1.8 kW, laser spot 3 mm and scanning speed 6 mm/s. In order to control defect formation, the previous layers' hatch space was 1.5 mm, while the hatch space of the subsequent layer was 2 mm. Both single deposition layer and hybrid forming multilayer samples were produced with substrate temperatures of 25 °C and 210 ± 10 °C , respectively. Before the hybrid samples with preheated substrate were produced, the substrate was heated to 200 °C to 220 °C by a heater, and thermocouples were fixed on the top surface to measure the temperature variation. Moreover, a temperature control system was connected to the thermocouple and heater to control, as in, facilitate/monitor any increase beyond 200 °C but below 220 °C . The preheating temperature range of 200 °C to 220 °C was selected in consideration of substrate oxidation behavior, and the difficulty of the large-scale part heating process. All deposition experiments were done in room air.
Then, the hybrid forming specimens with single deposition layer were sectioned along the longitudinal direction for geometrical characterization observation, as seen in Figure 1a. The deposition width (W1), deposition height (H2), and melt height (H1) under different processing parameters were measured by a Dino-Lite digital microscope (Taiwan Premier, (Suzhou, (JiangSu), China), for comparing the geometrical characterization's difference. The geometric characteristic of the cross-section is an effective parameter for deposition quality, thermal and subsequently mechanical analysis [23] during the laser metal deposition process, which could reflect the quality of the deposition layer. Samples of 2.5 mm in width were cut along the longitudinal direction from the multilayer hybrid forming samples. Microstructure analysis and room temperature test samples were selected, as shown in Figure 1b, and the microstructure and microhardness were done on the transverse direction section. The unit in Figure 1b is mm.  The specimens for microstructure and geometric characterization observation were mounted, polished, then etched in chloroazotic acid. The substrate steel was etched by a nitric acid-alcohol solution for optical microscopy observation. The microstructure of those specimens was observed by stereology microscopy (Stemi 2000, Carl Zeiss AG, Oberkochen, Germany) and optical microscopy (Axiovert 200 MAT, Carl Zeiss AG, Oberkochen, Germany). Microstructure and fracture surface were observed by means of Hitachi S-3400 N (Hitachi, Tokyo, Japan) and TESCAN MAIA3 (TESCAN China, (Shanghai, China) scanning electron microscope (SEM). Elemental analysis was done by energy dispersive spectrometry (EDS) (TESCAN China, (Shanghai, China)) during the SEM analysis. Based on the guidelines given in GB/T 228.1-2010, the sizes of tensile test specimens were determined, as shown in Figure 1b. The RT (room temperature) tensile testing was conducted on the 5582 Instron testing machine (Instron, Boston, Massachusetts, USA). The hardness was measured by Vickers hardness tester (AMH43) with an indentation load of 200 g for 15 s. The residual stress was measured by a Photo-LXRD X-ray diffraction device (P'Nalytical B.V., Almelo, the Netherlands), and the testing points are marked in Figure 1.

Geometric Characterization
The comparison of these calculated results is listed in Figure 2. The X-axis in Figure 2 represents laser power (P) and scanning speed (S) under different combination of process parameters. As shown in Figure 2a,b, the melt height (H1) and deposition width (W1) are significantly increased under the effect of substrate preheating. However, the deposition height (H2) of specimens with substrate preheating are slightly lower or equal to the specimens without substrate preheating. Similarly, this result was also found in laser deposition of 316 L on P20 steel. Evidently, the remolten size of the substrate increased with the help of substrate preheating, which would lead to the increase of W1 and H1. The calculated results of W1, H1 and the sum of H1 and H2 are closely related to laser power and scanning speed. The sum of H1 and H2, which indicates the height of the remolten pool section, is increased as a result of the increased powder catchment efficiency [24] and remolten zone volume. However, no clear relationship between H2 and laser powder and scanning speed was found. Similarly, the variation of W1, H1 and H1 + H2 of 316 L with P20 steel substrate shows the same variation tendency with the change of process parameters (laser power and scanning speed).
The values of H2 and W1 were assessment indexes of the deposition layer. The statistical results shown in Figure 2a indicate thatW1 of deposition specimens on 1045 steel substrate ranges from 3.679 to 4.207 mm among the parameters listed in Figure 2; while the range from 3.962 to 4.744 mm exists for substrate preheating, which is greater than the laser beam diameter (3 mm) as mentioned by Liu et al. [25]. The W1 ranges of the different materials deposited under various heat input is lower than the ranges when using substrate preheating. The H2 of deposition specimens on 1045 steel ranges from 0.481 to 0.792 mm and ranges from 0.489 to 0.679 mm with substrate preheating. This means H2 ranges of the different materials deposited under various heat inputs is greater than the ranges when using substrate preheating. For specimens under the same heat input, the value of W1 was increased and H2 was decreased, and the ratio of H2/W1 is decreased under substrate preheating. The H2/W1 ratios of 316 L on 1045 specimens were 0.141 (1600 W, 6 mm/s) and 0.142 (2000 W, 6 mm/s) without preheating, decreased to 0.127 and 0.121, respectively, under substrate preheating as illustrated in Figure 3a. Similarly, the H2/W1 ratios of 316 L on P20 specimens decreased to 0.118 (1600 W, 6 mm/s) and 0.109 (2000 W, 6 mm/s) under substrate preheating, from 0.131 and 0.125, respectively, as illustrated in Figure 3b. This indicates that the contact angle (illustrated in Figure 3) was reduced under the effect of substrate preheating. The contact angle was reduced under substrate preheating, which makes the top surface of the deposition layer smoother than the surface without preheating. The values of H2 and W1 were assessment indexes of the deposition layer. The statistical results shown in Figure 2a indicate thatW1 of deposition specimens on 1045 steel substrate ranges from 3.679 to 4.207 mm among the parameters listed in Figure 2; while the range from 3.962 to 4.744 mm exists for substrate preheating, which is greater than the laser beam diameter (3 mm) as mentioned by Liu et al. [25]. The W1 ranges of the different materials deposited under various heat input is lower than the ranges when using substrate preheating. The H2 of deposition specimens on 1045 steel ranges from 0.481 to 0.792 mm and ranges from 0.489 to 0.679 mm with substrate preheating. This means H2 ranges of the different materials deposited under various heat inputs is greater than the ranges when using substrate preheating. For specimens under the same heat input, the value of W1 was increased and H2 was decreased, and the ratio of H2/W1 is decreased under substrate preheating. The H2/W1 ratios of 316 L on 1045 specimens were 0.141 (1600 W, 6 mm/s) and 0.142 (2000 W, 6 mm/s) without preheating, decreased to 0.127 and 0.121, respectively, under substrate preheating as illustrated in Figure 3a. Similarly, the H2/W1 ratios of 316 L on P20 specimens decreased to 0.118 (1600 W, 6 mm/s) and 0.109 (2000 W, 6 mm/s) under substrate preheating, from 0.131 and 0.125, respectively, as illustrated in Figure 3b. This indicates that the contact angle (illustrated in Figure 3) was reduced under the effect of substrate preheating. The contact angle was reduced under substrate preheating, which makes the top surface of the deposition layer smoother than the surface without preheating.   Figure 4a,c shows micrographs of hybrid specimens made by LMD 316 L with 1045 and P20 steel, respectively, LMD 316 L with 1045 and P20 steel specimens with substrate preheating are also respectively shown in Figure 4b,d for comparison. The deposition boundaries are marked by red and yellow lines in Figure 4. The deposition boundaries of zigzag structure became the adjacent layer, scanning along the opposite direction. As shown in Figure 4b, LMD 316 L formed metallurgically sound and dense bonding with 1045 steel under substrate preheating, while defects formed near the bonding interface and terminated on the bottom melting line of the third deposition layer without substrate preheating. Unmolten powder was found in the defects zone as shown in Figure 5b, which verifies the insufficient bonding behavior. Meanwhile, internal stress caused cracking near the interface, as marked by the yellow dash line Figure 5c. The difference in interface morphology was mainly due to the morphology of the deposition layer. Assuming the wider deposition width of the single scanning path and the smoother surface morphology of one deposition layer under the substrate preheating, the insufficient bonding behavior between adjacent layers of multiple layers was eliminated. However, both LMD 316 L with P20 steel specimens have metallurgically bonding characteristics without defects under different substrate temperatures. Defects were generated by insufficient bonding between the adjacent deposition tracks. Comparing the difference of deposition width between 1045 and P20 steel as shown in Figure 2b,d, it is evident  Figure 4a,c shows micrographs of hybrid specimens made by LMD 316 L with 1045 and P20 steel, respectively, LMD 316 L with 1045 and P20 steel specimens with substrate preheating are also respectively shown in Figure 4b,d for comparison. The deposition boundaries are marked by red and yellow lines in Figure 4. The deposition boundaries of zigzag structure became the adjacent layer, scanning along the opposite direction. As shown in Figure 4b, LMD 316 L formed metallurgically sound and dense bonding with 1045 steel under substrate preheating, while defects formed near the bonding interface and terminated on the bottom melting line of the third deposition layer without substrate preheating. Unmolten powder was found in the defects zone as shown in Figure 5b, which verifies the insufficient bonding behavior. Meanwhile, internal stress caused cracking near the interface, as marked by the yellow dash line Figure 5c. The difference in interface morphology was mainly due to the Crystals 2020, 10, 891 6 of 15 morphology of the deposition layer. Assuming the wider deposition width of the single scanning path and the smoother surface morphology of one deposition layer under the substrate preheating, the insufficient bonding behavior between adjacent layers of multiple layers was eliminated. However, both LMD 316 L with P20 steel specimens have metallurgically bonding characteristics without defects under different substrate temperatures. Defects were generated by insufficient bonding between the adjacent deposition tracks. Comparing the difference of deposition width between 1045 and P20 steel as shown in Figure 2b,d, it is evident that the deposition width on 1045 steel is slightly smaller than P20 steel under the same parameter. The boundaries of different deposition tracks are clearly distinguished in Figure 4a,c. The height of the deposition track with substrate preheating is larger than other tracks. This phenomenon is more clearly evident in LMD 316 L with P20 steel.   Figure 6 presents the microstructural characteristics of the interface between 316 L deposition layer with 1045 and P20 substrate. As shown in Figure 6, the flat structure is presented at the interface under the effect of high temperature gradient and low solidification velocity. Fine cellular and columnar dendritic have grown perpendicular to the interface, which was caused by temperature gradient perpendicular to the bottom of the molten pool. The microstructure of laser metal deposition 316 L is mainly composed of δ, γ phases, as reported in the literature [26,27]. The matrix phase (dark region in the dendritic region) is the γ austenite phase, and the interdendritic phase is δ-ferrite or γ-austenite in reticular distribution morphology. However, it is difficult to distinguish the δ-ferrite from the SEM images. In comparison to the deposition microstructure, it is apparent that the microstructure morphology of the deposition layer is similar. The average dendrite arm space of the microstructure with different substrate temperatures is below 10 μm, while the dendrite arm space only minorly increased with the increasing substrate temperature. The impacts of substrate preheating on the microstructure of the deposition layer is mild.  Figure 6 presents the microstructural characteristics of the interface between 316 L deposition layer with 1045 and P20 substrate. As shown in Figure 6, the flat structure is presented at the interface under the effect of high temperature gradient and low solidification velocity. Fine cellular and columnar dendritic have grown perpendicular to the interface, which was caused by temperature gradient perpendicular to the bottom of the molten pool. The microstructure of laser metal deposition 316 L is mainly composed of δ, γ phases, as reported in the literature [26,27]. The matrix phase (dark region in the dendritic region) is the γ austenite phase, and the interdendritic phase is δ-ferrite or γ-austenite in reticular distribution morphology. However, it is difficult to distinguish the δ-ferrite from the SEM images. In comparison to the deposition microstructure, it is apparent that the microstructure morphology of the deposition layer is similar. The average dendrite arm space of the microstructure with different substrate temperatures is below 10 µm, while the dendrite arm space only minorly increased with the increasing substrate temperature. The impacts of substrate preheating on the microstructure of the deposition layer is mild.

Interfacial Morphology
Nevertheless, the microstructure of the HAZ (heat affected zone) in the substrate is different between the microstructure of different substrate temperatures. The difference is also found in specimens of both of 316 L with 1045 and P20 substrate. The unaffected microstructure of 1045 steel consisting of ferrite and pearlite is shown in Figure 7a; the ferrite is the dark phase etched by a nitric acid-alcohol solution; and the pearlite phase consists of alternating layers of ferrite (dark region) and cementite (white phase in long strips). This is consistent with the optical result as shown in Figure 4g, the white zone is the ferrite, and the dark zone is pearlite. During laser deposition, the ferrite and pearlite austenitized, cooled rapidly, and then formed a nonequilibrium microstructure during the cooling process. Rashid et al. [28] reported the HAZ microstructure of mild steel during laser deposition consisted of martensitic lath plus ferrite and bainite plus ferrite. The resulting HAZ microstructure is affected by peak temperature and cooling rate [29], which is closely related to the processing parameters and substrate properties. The resultant microstructure within the scope of this research mainly consists of pearlite in different morphologies, and the microstructure is refined. The composition of the ferrite phase in the HAZ is obviously decreased as shown in Figure 4e Nevertheless, the microstructure of the HAZ (heat affected zone) in the substrate is different between the microstructure of different substrate temperatures. The difference is also found in specimens of both of 316 L with 1045 and P20 substrate. The unaffected microstructure of 1045 steel consisting of ferrite and pearlite is shown in Figure 7a; the ferrite is the dark phase etched by a nitric acid-alcohol solution; and the pearlite phase consists of alternating layers of ferrite (dark region) and cementite (white phase in long strips). This is consistent with the optical result as shown in Figure 4g, the white zone is the ferrite, and the dark zone is pearlite. During laser deposition, the ferrite and pearlite austenitized, cooled rapidly, and then formed a nonequilibrium microstructure during the cooling process. Rashid et al. [28] reported the HAZ microstructure of mild steel during laser deposition consisted of martensitic lath plus ferrite and bainite plus ferrite. The resulting HAZ microstructure is affected by peak temperature and cooling rate [29], which is closely related to the processing parameters and substrate properties. The resultant microstructure within the scope of this research mainly consists of pearlite in different morphologies, and the microstructure is refined. The composition of the ferrite phase in the HAZ is obviously decreased as shown in Figure 4e,f. The HAZ microstructure of 1045 steel under substrate preheating is coarsened with preheating. The unaffected substrate microstructure of P20 steel is composed of tempered martensite, as shown in Figure 7b. The resultant microstructure at the HAZ experienced different degrees of tempering during laser deposition, which was also verified by Chen et al. [12]. Carbide precipitated and grew between martensitic lathes, and the carbide is marked by a purple circle in the highly magnified image shown in Figure 6d. The HAZ microstructure had more carbide phases under substrate preheating, as shown in Figure 6c,d. The compositional profiles of the interface between 316 L with 1045 and P20 steel are presented in Figure 8a,b, respectively. The compositional profiles of the interface under substrate preheating are shown in Figure 8c,d for comparison. The EDS testing lines are marked in Figure 6a-d. Because of composition difference between the lower alloyed substrate and the 316 L deposition layer, the contents of Cr, Ni and Fe elements tend to be found towards sharp decreases at the interface. The width of the composition variation zone is nearly equal to the width of the flat crystals near the fusion line. An EDS mapping performed at the fusion zone are shown in Figure 9, which demonstrates that the clear composition variation happened in the fusion line. The EDS mapping image was from specimens of laser deposition 316 L on 1045 steel without preheating (the testing The unaffected substrate microstructure of P20 steel is composed of tempered martensite, as shown in Figure 7b. The resultant microstructure at the HAZ experienced different degrees of tempering during laser deposition, which was also verified by Chen et al. [12]. Carbide precipitated and grew between martensitic lathes, and the carbide is marked by a purple circle in the highly magnified image shown in Figure 6d. The HAZ microstructure had more carbide phases under substrate preheating, as shown in Figure 6c,d. The compositional profiles of the interface between 316 L with 1045 and P20 steel are presented in Figure 8a,b, respectively. The compositional profiles of the interface under substrate preheating are shown in Figure 8c,d for comparison. The EDS testing lines are marked in Figure 6a-d. Because of composition difference between the lower alloyed substrate and the 316 L deposition layer, the contents of Cr, Ni and Fe elements tend to be found towards sharp decreases at the interface. The width of the composition variation zone is nearly equal to the width of the flat crystals near the fusion line. An EDS mapping performed at the fusion zone are shown in Figure 9, which demonstrates that the clear composition variation happened in the fusion line. The EDS mapping image was from specimens of laser deposition 316 L on 1045 steel without preheating (the testing region has been marked in Figure 6a), and the results of the other three specimens were similar in comparison (not shown to avoid repetition). The composition of the first deposition layer is homogeneous, no macrosegregation is found. This can be largely attributed to the sufficient mixing in the molten pool and the compositional difference between the substrate material and the deposition material. Macrosegregation might happen at the first deposition layer as mentioned by Liu [30]. The composition variation of different wrought materials under substrate preheating are the same as that without preheating, as shown in Figure 8c,d.     Figure 10 shows the stress-strain curves of specimens selected from LMD built from 316 L with wrought 1045 and P20 steel. The hybrid samples of 316 L/1045 steel cracked under a lower stress level. At the same time, there are large differences in mechanical properties among samples 1, 2, 3. Under substrate preheating, the average ultimate strength of hybrid forming specimens increases from 390 MPa to 557 MPa, and the average elongation increases from 3.8% to 37.3% as shown in Figure 11a. The weakness of tensile properties is correlated with the interface cracks. However, the stress-strain curve of 316 L/P20 steel hybrid specimens is similar, whether or not the substrate preheating was done, as shown in Figure 10b. The average ultimate tensile strength and elongation  Figure 10 shows the stress-strain curves of specimens selected from LMD built from 316 L with wrought 1045 and P20 steel. The hybrid samples of 316 L/1045 steel cracked under a lower stress level. At the same time, there are large differences in mechanical properties among samples 1, 2, 3. Under substrate preheating, the average ultimate strength of hybrid forming specimens increases from 390 MPa to 557 MPa, and the average elongation increases from 3.8% to 37.3% as shown in Figure 11a. The weakness of tensile properties is correlated with the interface cracks. However, the stress-strain curve of 316 L/P20 steel hybrid specimens is similar, whether or not the substrate preheating was done, as shown in Figure 10b. The average ultimate tensile strength and elongation of hybrid specimens of 316 L with P20 steel are 565.3 MPa and 35.9% respectively, while this changes to 550 MPa and 32% under substrate preheating, as shown in Figure 11b.    Figure 12 shows the hybrid formed plate tensile samples. The fracture surfaces of the tensile test specimens are presented in Figure 13. The white line in Figure 12 represents the interface of 316 L and substrate, and the shrinkage of the neck is presented in a large magnified image. Fractures of    Figure 12 shows the hybrid formed plate tensile samples. The fracture surfaces of the tensile test specimens are presented in Figure 13. The white line in Figure 12 represents the interface of 316 L and substrate, and the shrinkage of the neck is presented in a large magnified image. Fractures of 316 L with 1045 steel are located in the interface without substrate preheating, whereas the location  Figure 12 shows the hybrid formed plate tensile samples. The fracture surfaces of the tensile test specimens are presented in Figure 13. The white line in Figure 12 represents the interface of 316 L and substrate, and the shrinkage of the neck is presented in a large magnified image. Fractures of 316 L with 1045 steel are located in the interface without substrate preheating, whereas the location of the fracture changes to the LMD zone under substrate preheating. All fractures of 316 L with P20 steel specimens are in the LMD zone, despite the fractures near the interface. Neck contraction is found on samples which did not crack in the interface, as shown in Figure 12b-d. The resulting mechanical properties verify that the tensile strength of hybrid formed specimens with different substrates are very different. During the laser metal deposition process, some defects might form as a result of insufficient bonding, which deteriorates the adhesion properties. In this study, the 316 L and 1045 steel hybrid samples keep the interface bond characteristics safe under substrate preheating, which increases the bonding properties more so than without preheating. Under substrate preheating the deposition layer is smooth and wide, which will promote adhesion properties under the scanning method in the same hatch space. However, the mechanical properties of the hybrid specimens of 316 L with P20 steel were similar between different substrate temperatures, because no defects appeared on the interface zone without substrate preheating. The deposition width on 1045 steel was slightly smaller than P20 steel on the same parameter, as verified in Figure 2. As a result, the hybrid samples of 316 L and P20 can realize safe interface bonding even without the help of substrate preheating.

Tensile Properties
The tensile strength properties of hybrid forming 316 L with conventional manufacturing steel and without interface defects are found to be comparable to that of the laser metal deposition 316 L in the literature [29], which has ultimate tensile strength (540-625 MPa) and elongation (35-85%). The tensile strength of wrought 1045 steel and P20 steel is above 600 MPa and 1250 MPa, respectively. Additionally, the fracture of hybrid specimens occurs on the weakest position on the 316 L side, away from the interface.
Representative fractographs of the tensile specimens are shown in Figure 13. Flat fracture surface is shown in Figure 13a, which was caused by a crack formed during the laser metal deposition. The brittle fracture with dimple gliding fracture is shown in Figure 13b. Some micro cracks are discovered on the fracture surface. Conversely, the fracture surface under substrate preheating shows a dimpled ductile mode of tensile failure. The voids of unmelted powders in the laser metal deposition specimens could be avoided by using more laser power, however excessive The resulting mechanical properties verify that the tensile strength of hybrid formed specimens with different substrates are very different. During the laser metal deposition process, some defects might form as a result of insufficient bonding, which deteriorates the adhesion properties. In this study, the 316 L and 1045 steel hybrid samples keep the interface bond characteristics safe under substrate preheating, which increases the bonding properties more so than without preheating. Under substrate preheating the deposition layer is smooth and wide, which will promote adhesion properties under the scanning method in the same hatch space. However, the mechanical properties of the hybrid specimens of 316 L with P20 steel were similar between different substrate temperatures, because no defects appeared on the interface zone without substrate preheating. The deposition width on 1045 steel was slightly smaller than P20 steel on the same parameter, as verified in Figure 2. As a result, the hybrid samples of 316 L and P20 can realize safe interface bonding even without the help of substrate preheating.
The tensile strength properties of hybrid forming 316 L with conventional manufacturing steel and without interface defects are found to be comparable to that of the laser metal deposition 316 L in the literature [29], which has ultimate tensile strength (540-625 MPa) and elongation (35-85%). The tensile strength of wrought 1045 steel and P20 steel is above 600 MPa and 1250 MPa, respectively. Additionally, the fracture of hybrid specimens occurs on the weakest position on the 316 L side, away from the interface.
Representative fractographs of the tensile specimens are shown in Figure 13. Flat fracture surface is shown in Figure 13a, which was caused by a crack formed during the laser metal deposition. The brittle fracture with dimple gliding fracture is shown in Figure 13b. Some micro cracks are discovered on the fracture surface. Conversely, the fracture surface under substrate preheating shows a dimpled ductile mode of tensile failure. The voids of unmelted powders in the laser metal deposition specimens could be avoided by using more laser power, however excessive heat input would also result in severe deformation of the wrought substrate. The fracture surfaces of 316 L with P20 steel, conversely, are ductile fracture mode free from the influence of substrate preheating. The dimple size observed on the fracture surface is similar. Some voids were also found, which were caused by incomplete melted powder, as mentioned by Sun et al. [14] and Yadollahi et al. [31]. The fracture surfaces are consistent with tensile properties as shown in Figures 10 and 11.
Crystals 2020, 10, x FOR PEER REVIEW 14 of 17 which were caused by incomplete melted powder, as mentioned by Sun et al. [14] and Yadollahi et al. [31]. The fracture surfaces are consistent with tensile properties as shown in Figures 10 and 11.

Microhardness Distribution
The microhardness distributions along the building direction of the laser metal deposition 316 L on P20 and 1045 steel are given in Figure 14a,b, respectively. A schematic about the test point of hardness distribution is also shown. The mean hardness values of 316 L deposition zones for all specimens are around 200 HV0.2. In addition, hardness distribution of different deposition layers is similar. However, sharp changes in hardness are found in the interface between different materials. The substrate HAZ hardness shows a different transformation phenomenon. The hardness of 1045 steel increases, while the HAZ hardness of P20 steel decreases, when comparing the affected substrate.

Microhardness Distribution
The microhardness distributions along the building direction of the laser metal deposition 316 L on P20 and 1045 steel are given in Figure 14a The influence of preheating on the substrate hardness is different between 1045 and P20 steel, which is closely related to the microstructure transformation. The hardness in 1045 steel HAZ region is increased, which is because the microstructure was refined and the proeutectoid ferrite disappeared during austenitization. The variation in characterization of substrate hardness is similar with different substrate temperatures. Conversely, the HAZ microstructure of P20 substrate experienced severe tempering after substrate preheating, which makes the HAZ hardness (300-330 HV0.2) lower than without preheating. The decreasing width of the hardness region is correlated with the HAZ width under substrate preheating. The hardness of P20 substrate without preheating is slightly decreased compared to the unaffected substrate hardness (higher than 400 HV0.2).
The stress between first deposition layer and substance was usually the largest one among all the layers, and the residual stress at the border position of hybrid formed specimens was measured. The average value of residual stress in 316 L/1045 steel was 280 MPa, while after substrate preheating the stress decreased to 262 MPa. The average value of stress in 316 L/P20 steel specimens decreased from 303 MPa to 273 MPa under substrate preheating. Under substrate preheating, the stress of deposition and substrate was reduced, while in this research, the influence of preheating on the residual stress was small. The influence of substrate preheating on the thermal and residual stress will be discussed further by high-scale laser cladding in hybrid forming process. The preheating temperature higher than 200 °C might be more suitable to guarantee safe bonding and a lower stress level, but it is hard to achieve this in a large scale part.

Conclusions
The followings are significant conclusions which can be drawn from this study: 1. Under substrate preheating, the deposition width (W1) was increased and deposition height (H2) showed little difference. At the same time, the variation amplitude of deposition height (H2) as heat input increased was weakened, but the variation amplitude of deposition width W1 was broadened. In addition, the H2/W1 ratio was decreased, making the deposition layer surface smoother. Similar morphology variation was found with both 1045 and P20 substrates. 2. LMD 316 L formed metallurgically sound and dense bonding with 1045 steel substrate under substrate preheating, while defects were formed near the bonding interface without substrate preheating. LMD 316 L with P20 steel specimens have metallurgically bonding characteristic without defects under different substrate temperature. The microstructure of deposition layers was similar between different substrate materials and temperatures. 3. The adhesion properties of hybrid formed 316 L with 1045 were obviously decreased without substrate preheating, and brittle fractures of 316 L with 1045 steel were located in the interface without substrate preheating. The adhesion properties of hybrid formed 316 L with P20 had no relationship with substrate preheating because of the metallurgic bonding characteristic in the interface. The difference between P20 and 1045 steel substrate was mainly because of