3.1.1 Ultrasonic amplitude optimization of UV assisted LDED process

Figure 1 shows the microstructure of the single-pass single-layer LDED samples without and with ultrasonic vibration (UV) treatment (ultrasonic amplitude: 18 μm). Without UV treatment, large amounts of obvious columnar crystal structures perpendicular to the substrate existed in the LDED sample, as shown in Figure 1(a). The columnar crystal structure was also observed in the additive manufacturing of 304 stainless steel, Cu alloy and high-entropy alloy [23-25]. The microstructure evolution is affected by two key control parameters, namely temperature gradient (G) and solidification rate (R_s) [26, 27]. Based on solidification theory, the ratio of the G to R_s determines the forms of the microstructure [28, 29]. A complex physical metallurgical process occurred during laser additive manufacturing process, and the material melting, solidification and cooling during the forming process are completed under extremely fast conditions. The solidification process has the characteristics of high temperature gradient and high cooling rate during additive manufacturing, and it is favor to form a coarse columnar crystal structure. During the layer-by-layer deposition process, the columnar crystals in the pre-deposited layer have a “memory function”, and will further epitaxially grow into the new deposited layer, thus forming the coarse columnar crystals that penetrate all the deposited layers. On the contrary, it is obvious that the number of columnar crystals reduces and the overall microstructure is relatively uniform, which is composed of massive equiaxed crystals in the UV assisted LDED sample, as shown in Figure 1(b). With the introduction of ultrasonic vibration, the refined microstructure of LDED A517Q steel can be achieved, and the transformation from columnar crystals to equiaxed crystals occurs in the deposition layer.

Figure 1全屏查看

Microstructure of LDED A517Q steel: (a) Without UV; (b) With UV

Figure 2 shows the microstructure of the YOZ cross-sectional samples under different ultrasonic amplitudes. The microstructure of the deposition layer obviously changes with UV treatment. The large columnar crystals with some equiaxed crystals for the LDED sample are transformed into large numbers of fine equiaxed crystals with some columnar crystals for the UV assisted LDED samples. With increasing ultrasonic amplitudes, the characteristics of columnar crystals in LDED sample gradually weaken and the proportion of equiaxed crystals gradually increase. When the amplitude reaches 14 μm, the coarse columnar crystals penetrating throughout the whole deposition layer significantly reduce, as shown in Figure 2(c). With the amplitude increasing to 18 μm, almost all the columnar crystals structure in the deposition layer are transformed into equiaxial crystals, as shown in Figure 1(b). Based on the above analysis, it can be concluded that ultrasonic vibration has a significant effect on microstructure refinement during LDED process.

Figure 2全屏查看

Microstructure of LDED samples under different ultrasonic amplitudes: (a) 6 μm; (b) 10 μm; (c) 14 μm; (d) AM schematic

Figure 3 shows the microhardness distribution of LDED samples under different ultrasonic amplitudes. Ultrasonic vibration has a significant effect on microhardness value and microhardness distribution of LDED sample, and the microhardness characteristics are listed as follow. Firstly, the microhardness value of the UV assisted LDED samples are significantly higher than that of the LDED sample without UV due to the grain refinement caused by ultrasonic vibration. Secondly, the fluctuation range of the microhardness value of the UV assisted LDED sample is smaller than that of the LDED sample without UV, which indicated that ultrasonic vibration makes the properties of the deposited layer uniform. It can be concluded that the inhomogeneity of the microstructure and properties of LDED sample reduces due to the ultrasonic vibration effect. It is noted that the fine microstructure and uniform microstructure distribution are obtained for the UV assisted LDED sample when 18 μm ultrasonic amplitude is selected.

Figure 3全屏查看

Microhardness distribution of LDED samples under different ultrasonic amplitudes

3.1.2 Microstructure and mechanical properties of UV assisted LDED sample

Based on the above analysis, ultrasonic amplitude of 18 μm is selected to treat single-pass multilayer LDED sample. Figure 4(a) shows the microstructure of the YOZ cross-sectional single-pass multilayer UV assisted LDED sample. From Figures 4(b) and (c), the bottom and upper part microstructure of the deposition layer is relatively uniform, no coarse columnar crystals with obvious orientation are found, which is significantly different from the microstructure of LDED sample without ultrasonic treatment. As shown in Figure 4(b), a few fine rod-like dendrite structure exists at the bottom part near the substrate. During the UV assisted LDED process, ultrasonic vibration and laser additive manufacturing are carried out simultaneously, and ultrasonic vibration acts on each layer of single-pass multilayer UV assisted LDED sample. The previous deposition layer will be remelted under the laser action when forming the next layer, the remelting will break the unique microstructure characteristics formed under the ultrasonic vibration. The microstructure of the deposition layer is generated by the joint action of ultrasonic vibration treatment and subsequent laser remelting effect, and only the last layer is spared from the laser remelting effect. From Figure 4(d), large amounts of granular bainite (GB) and polygonal ferrite (PF) with a small amount of martensite (M) are found at the bottom of UV assisted LDED sample. When a few layers are deposited firstly, the heat from the deposition layer can be dissipated through the substrate, and a huge temperature gradient will result in the formation of martensitic structure. However, when the subsequent layer is deposited, the laser remelting will heat the previous layer, and the temperature is over austenitizing temperature. The temperature gradient is less than that of the rapid cooling of the liquid, the microstructure of remelted layer is transformed into GB, PF and some residual M under the condition of lower cooling rate with large heat accumulation. From Figure 4(e), the upper microstructure of UV assisted LDED sample is mainly composed of M. No subsequent continuous heat input accumulates at the top of deposition layer, and the heat exchange with air occurs and large temperature gradient forms, the transformation of austenite into martensite can be obtained.

Figure 4全屏查看

Microstructure of single-pass multilayer LDED sample with ultrasonic vibration: (a) Overall microstructure; (b, d) Bottom part microstructure; (c, e) Upper part microstructure

Figure 5 shows the XRD patterns of LDED samples under different ultrasonic amplitudes. With increasing ultrasonic amplitude, the peak intensity for LDED sample significantly reduces. The peak intensity relates to the diffraction intensity of the crystal plane, indicating the formation of a relatively random crystallographic orientation distribution [30]. It is indicated that the grains of LDED samples gradually refine with increasing ultrasonic amplitude, which is consistent with the analysis above. When under ultrasonic vibration, even if the low amplitude (6 μm), is introduced, the (110) peak of XRD is shifted to the left compared to the sample without UV, which shows that lattice expansion and residual tensile stress is produced in the sample under the introduction of ultrasound vibration [28, 30, 31].

Figure 5全屏查看

XRD patterns of LDED samples under different ultrasonic amplitudes

Figures 6(a) and (b) show the inverse pole figures (IPF) of the single-pass multilayer LDED samples without and with ultrasonic vibration (UV), and the reference direction are perpendicular to the substrate. For the LDED sample without UV (Figure 6(a)), most microstructures with obvious orientation are continuously growing along a specific direction, and the continuous block distributed microstructures is found in the deposition layer. For the UV treated sample (Figure 6(b)), the microstructure is relative uniform, and no obvious continuous orientation forms. For the overall morphology, the microstructure of the UV treated sample is composed of equiaxed crystals with a few fine columnar crystals, which also proves that obvious grain refinement is obtained in the deposition layer by introduction of UV. Figure 6(c) shows the percentage of grain size distribution interval and the results are listed as follow. Without UV treatment: grain size ranges 10-30 μm² area: ~51%, grain size >100 μm² area: ~22%. With UV treatment: grain size ranges 10- 30 μm² area: ~60%, grain size >100 μm² area: ~12%. It indicates the grains are refined by the introducing of synchronous ultrasonic vibration. Meanwhile, the proportion difference of the aspect ratio, the ratio between the length of the long side and the short side of the grains, is large for the LDED sample without UV treatment, while the proportion difference of the aspect ratio reduces for the UV treated sample (Figure 6(d)). In general, the introduction of ultrasonic vibration can effectively promote the transformation of coarse columnar crystals to fine equiaxed crystals, refine the grains size and improve the grains uniformity of the deposition layer obviously. Figures 6(e) and (f) show the grain boundary maps of the single-pass multilayer LDED A517Q alloy steel without and with UV. Large numbers of coarse columnar crystals consisted of many sub-grains exist in LDED sample without UV treatment. The grain boundary distribution is mainly low angle grain boundary (LAGB, 2°-15°) and the percentage is 70.9%. For the UV treated sample, the microstructure is mainly composed of fine equiaxed crystals, and its grain boundary distribution is mainly large angle grain boundary (HAGB, >15°) and the percentage is 57.2% (Figure 6(f)). Large numbers of sub-grains are generated in the deposition layer under ultrasonic vibration effect during UV assisted LWAM process, and these regions possessed lattice distortion and plastic deformation due to ultrasonic vibration, which is beneficial for the recrystallization of sub-grains under subsequent continuous heat input [32, 33]. Therefore, the transformation of LAGB to HAGB occurs.

Figure 6全屏查看

EBSD results of the single-pass multilayer LDED samples: (a) IPF without UV; (b) IPF with UV; (c, d) Grain size distribution interval; (e, f) Grain boundary maps

In fact, the effect of ultrasonic vibration on microstructure is classified into two types of mechanisms according to our previous work [16, 17, 34]. Ultrasonic vibration is applied to both the molten pool and the solid deposition layer. The effects of ultrasonic vibration on molten pool mainly include ultrasonic cavitation, acoustic fluidization and dendrite breaking, and effects of ultrasonic vibration on solid deposition layer is mainly consisted of plastic deformation and recrystallization. In this work, ultrasonic vibration was applied simultaneously to press onto the deposition layer at a constant offset. The ultrasonic vibrations were followed by a certain spacing between the two effects. In such way, the regulation effects of ultrasonic energy field in the liquid melt pool as well as the dynamic recrystallization effects of high-temperature deposited layer induced by the ultrasonic impact can be achieved. Firstly, ultrasonic cavitation, acoustic fluidization and dendrite breaking exist in the liquid molten pool, secondly, the solid deposition layer is still at high temperature when ultrasonic vibration is conducted at solid layer, thus, dynamic recrystallization of solid deposition layer takes place and the grains refinement occurs.

Figure 7(a) shows the microhardness distribution of LDED samples without and with ultrasonic vibration (UV). With ultrasonic vibration, the mean microhardness obviously improves from 280HV0.2 for LDED sample to 447HV0.2 for UV assisted LDED sample, the increasing rate is about 59.6%. The microstructure determines the mechanical properties of the deposition layer [26, 35-37]. Due to the introducing of ultrasonic vibration, the microstructure and grains of deposition layer refine, increasing the numbers of dislocations and grain boundaries, which is beneficial for the improvement of mechanical properties of the deposition layer [26, 38]. However, it should be noted that the huge fluctuation range of microhardness values at different measurement positions of UV assisted LDED sample is remained. It still needs post-treatment to eliminate the inhomogeneity of microstructure and performance of the UV assisted LDED sample. Figure 7(b) shows the stress-strain curves of the LDED samples without and with UV. The ultimate tensile strength (UTS) and yield strength (YS) of the sample with UV treatment are higher than that of the sample without UV. When the ultrasonic amplitudes of 18 μm is selected, the UTS and YS of the sample with UV are 947 MPa and 734 MPa, which are improved by 43.7% and 49.4% compared to the original LDED sample, respectively. Under this condition, the elongation after breaking (EL) of the sample with UV have some slight decrease and the value is ~8%. Under the ultrasonic vibration effect, the grains and microstructure of UV assisted LDED samples refine, resulting in the increase of grain boundaries, hindering the movement and expansion of dislocation, which makes the plastic deformation of the deposition layer difficult to occur and the strength increase [38, 39]. Meanwhile, residual tensile stress is produced in the UV assisted LDED sample by the introduction of ultrasound vibration, which may have negative effect on the EL of deposition layer [40].

Figure 7全屏查看

Mechanical properties and tensile fracture morphology of LDED samples: (a) Microhardness; (b) Stress-strain curves (c) Fracture morphology of LDED sample; (d) Fracture morphology of UV assisted LDED sample

Figure 7(c) shows the tensile fracture morphology of LDED sample. Typical ductile fracture characteristics with large numbers of dimples [41] can be found. The equiaxial dimple characteristics with fine size, shallow depth are shown in Figure 7(b). The size, quantity and depth of the dimples are usually related to the size and quantity of inclusions of the material during fracture process. Large amounts of fine particles can be seen in the fracture surface, and the existence of these particles can serve as nucleating points, which is helpful for the formation of the dimples [42]. Meanwhile, many defects, such as insufficient melting wires and pores, are found. These defects are often the crack source in the service process, it is needed to treat the deposition layer properly to eliminate these defects. Figure 7(d) shows the tensile fracture morphology of UV assisted LDED sample. Similar fracture characteristics can be found, however, the inclusions size, dimple size and depth significantly reduce compared to the LDED sample, which indicates that the plasticity of the UV treated sample is lower than that of the sample without UV.

3.2.1 Microstructure and microhardness of LDED sample treated by tempering

Figures 8 shows the microstructure of LDED sample tempering at 650 ℃ for 120 min. After tempering treatment, columnar crystals at the bottom part reduces and selectively grain growth weakens compared to the original LDED sample. The bottom part microstructure of the deposition layer is composed of GB, PF and tempered sorbite (TS). The boundary between GB and PF gradually weakens in the tempering state, as shown in Figure 8(c). The upper part is mainly composed of TS, as shown in Figure 8(d). During LDED process, the solidification of the molten pool forms in a short time due to the characteristics of fast heating and fast cooling, which is equivalent to the quenching process. Under this condition, high-temperature tempering at 650 ℃ on the LDED sample is equivalent to thermal refining treatment conducted on the deposition layer. Therefore, M changes into TS during high-temperature tempering treatment [43].

Figure 8全屏查看

Microstructure and microhardness of the single-pass multilayer LDED sample tempering at 650 ℃ for 120 min

Figure 9 shows the EBSD results of LDED sample with tempering treatment. After tempering, the grain orientation weakens to some extent and the grain distribution is uniform, no obvious boundaries between different crystal are found, which may be caused by the precipitation carbide during tempering. It should be noted that columnar crystals and fine dispersion distributed precipitates exist (Figure 9(a)). Figure 9(b) reveals that the tempering treatment has limited effect on the grain boundary of the deposition layer. Figures 9(c) and (d) show the grains quantitative statistics result of LDED sample with tempering treatment. After tempering, the percentage of grain size >100 μm² area has a slight increase to 10%, while the percentage of the grain size ranges 10-30 μm² area is the same as the original LDED sample. During tempering treatment, sub-grains merge with each other and grow into grains with HAGB. The aspect ratio of grains has little change for the two states sample.

Figure 9全屏查看

EBSD result of the single-pass multilayer LDED sample tempering at 650 ℃ for 120 min: (a) IPF; (b) Grain boundary; (c) Grain size; (d) Aspect ratio

Figure 10 shows the microhardness distribution of single-pass multilayer LDED sample without and with tempering treatment. It is seen that the microhardness distribution becomes uniform, while the overall microhardness value obviously decreases, and the mean value is about 243HV0.2 after tempering treatment. During the tempering process at 650 ℃, the martensite structure in the deposition layer decomposes to form stable TS, which reduces the microhardness of the sample. In general, it can be concluded that the tempering treatment can modify the microstructure and improve the inhomogeneity of mechanical properties of the single-pass multilayer LDED sample, however, the mechanical properties of the sample decreased sharply.

Figure 10全屏查看

Microhardness distribution of the single-pass multilayer LDED sample tempering at 650 ℃ for 120 min

3.2.2 Effect of tempering treatment on UV assisted LDED-ed sample

Figure 11 shows the microstructure of the single-pass multilayer UV assisted LDED sample tempering at 650 ℃ for 120 min. The bottom part microstructure is mainly composed of tempered sorbite (TS) with some GB and PF, as shown in Figure 11(c), while the upper part is mainly composed of TS (Figure 11(d)). The microstructure orientation of the deposition layer significantly reduces compared to the three states samples mentioned above.

Figure 11全屏查看

Microstructure of UV assisted LDED sample tempering at 650 ℃ for 120 min

Figure 12 shows the EBSD results of the single-pass multilayer UV assisted LDED sample with tempering treatment. Compared to the UV assisted LDED sample (Figure 5(b)), the grain orientation weakens, more dispersed precipitates distribute. Overall microstructure is more uniform after tempering treatment (Figure 12(a)). Compared to the LDED sample with tempering (Figure 9), the grain orientation significantly weakens and the grains refine obviously, which proves that the ultrasonic vibration treatment has a significant effect on grain refinement for the LDED sample during tempering treatment. Figure 12(b) shows the grain boundary maps of the single-pass multilayer UV assisted LDED sample without and with tempering. After tempering, the proportion of LAGB is 55.3%, which decreases by 22% and 23.9% compared with LDED sample (70.9%) and tempered LDED sample (72.7%), respectively, while increases by 29.2% compared with UV assisted LDED sample (42.8%). This may be caused by the precipitation of large quantities of carbides during the tempering process. Figures 12(c) and (d) show the grains quantitative statistics result of the single-pass multilayer UV assisted LDED A517Q alloy steel without and with tempering treatment. After tempering, the percentage of grain size >100 μm² area has a slight increase from 60% to 63% and the aspect ratio of grains ranged 1-2 has an obvious increase from 38% to 46% compared to the original UV assisted LDED sample. The results indicated that the grains were refined and equiaxial of the UV assisted LDED sample after tempering.

Figure 12全屏查看

EBSD results of UV assisted LDED sample with tempering: (a) IPF; (b) Grain boundary; (c) Grain size; (d) Aspect ratio

Figures 13 and Figure 7(b) show the mechanical properties of the single-pass multilayer UV assisted LDED A517Q alloy steel without and with tempering treatment. The microhardness of UV assisted LDED sample after tempering significantly reduces compared to the UV assisted LDED sample, however, the limited fluctuation range of microhardness values at different measurement positions is achieved (Figure 13). The mean microhardness value of 244HV0.2 is obtained for the UV+tempering state and the reduction rate is about 45.4% compared to UV assisted LDED sample (447HV0.2). Due to the complexity of thermal cycle and refinement effect of ultrasonic vibration, martensite with different saturation degrees and inhomogeneous microstructure forms, which result in the large fluctuation of microhardness at different positions of UV assisted LDED sample. After tempering, the transition state martensite and residual austenite structures in the deposition layer are transformed into stable TS and carbide structures under the tempering treatment at 650 ℃, thus, the microhardness difference between each layer or positions are limited. Figure 7(b) shows the stress-strain curves of the single-pass multilayer LDED sample, UV assisted LDED sample, LDED sample with tempering and UV assisted LDED sample with tempering. It is seen that the mechanical properties of the single-pass multilayer LDED sample are improved to a certain extent after ultrasonic vibration treatment or tempering treatment. The maximum ultimate tensile strength (UTS) and yield strength (YS) are obtained for the UV assisted LDED sample, however, the elongation after breaking (EL) has a sharp decrease to 12% and large inhomogeneity of microstructure and performance exists. For the LDED sample with tempering, the UTS, YS and EL are 566 MPa, 481 MPa and 22.5%, respectively. The plasticity has an obvious increase while the UTS and YS have a significant decrease compared to the LDED sample (659 MPa, 491 MPa and 19.5%). For the LDED sample treated by UV with tempering, the UTS, YS and EL are 765 MPa, 657 MPa and 19.5%. The increase rates of UTS and YS are 14.5% and 33.8% compared to LDED sample, respectively. In other words, the ultimate tensile strength and yield strength of the deposition layer can be greatly improved while maintaining plasticity for the LDED sample when it is treated by synchronous ultrasonic vibration combined with subsequent tempering treatment.

Figure 13全屏查看

Microhardness distribution of LDED A517Q alloy steel with tempering

Figure 14 shows the tensile fracture morphology of the single-pass multilayer UV assisted LDED sample with tempering treatment. The uniform distribution of dimples reveals that the ductile fracture occurs for the UV assisted LDED sample with tempering. No obvious defects, such as pores, inclusions, and un-melted material, are found in the fracture surface. All these indicate that the sample has superior comprehensive mechanical properties. For the analysis above, the microstructure orientation of the deposition layer significantly reduces, and the grains are refined and equiaxial for the UV assisted LDED sample after tempering, resulting in the improvement of mechanical properties of samples.

Figure 14全屏查看

Tensile fracture morphology of UV assisted LDED sample with tempering