3.1.

Surface Morphology and Roughness

There are many factors affecting the surface morphology and roughness of laser polishing Ti6Al4V. The main factors depend on the choice of process parameters. To obtain a high precision machined surface, it is necessary to optimize the process parameters. Liang et al.¹² and Zhou et al.⁴⁸ experimentally studied the relationship between the laser polishing process parameters of Ti6Al4V alloy and the surface quality and finally obtained the optimal combination of process parameters. Table 1 summarizes the surface roughness of laser polished Ti6Al4V alloy and lists the optimum process parameters corresponding to the surface roughness. It can be seen that the ultimate surface roughness after laser polishing can reach $0.3\mu \mathrm{m}$, which is a 95% improvement compared with the roughness of as-received component.

Table 1

Surface roughness and the corresponding process parameters of laser polishing Ti6Al4V.

In the laser polishing process, the process parameters directly affect the temperature gradient, which has a great influence on the roughness of laser polished parts. Therefore, the surface with the highest precision can be machined by the optimized process parameters. Liang et al.¹² studied the influence of three different laser power densities ($0.853\times {10}^5$, $1.703\times {10}^5$, and $2.503\times {10}^5\mathrm{w}/{\mathrm{cm}}^2$) on the surface roughness of polished workpiece and found that the Ra value at $1.703\times {10}^5\mathrm{w}/{\mathrm{cm}}^2$ was the minimum, reaching $2.1\mu \mathrm{m}$. With the increase of power density ($2.503\times {10}^5\mathrm{w}/{\mathrm{cm}}^2$), surface roughness increased. Further analysis showed that it was the SSM regime when the power density was $0.853\times {10}^5$ and $1.703\times {10}^5\mathrm{w}/{\mathrm{cm}}^2$. With the increase of power density ($2.503\times {10}^5\mathrm{w}/{\mathrm{cm}}^2$), the temperature of the workpiece increases, and the depth of the molten pool exceeds the peak-valley value, which is the SOM regime. At the same time, Ti6Al4V has a negative temperature coefficient of surface tension, that is, the surface tension of molten metal decreases with increasing temperature. The temperature in the center of the molten pool is the highest, and the temperature at the edge of the molten pool is lower. Therefore, the distribution of surface tension gradually increases from the center of molten pool to the edge of molten pool, forming the surface tension gradient. The molten liquid is pulled to the solidification front (the edge of the molten pool) under the action of the surface tension gradient,⁵³ causing a height difference in the liquid level. Under gravity, the molten liquid flows back again, forming a ripple, as shown in Fig. 2, and resulting in an increase in surface roughness.

The effect of laser power density on the surface roughness of components has been generally accepted. Similarly, the combination of scanning speed and laser power is also very crucial to the roughness of laser polished titanium alloy. Zhou et al.⁴⁸ investigated the effect of scanning speed and laser power on the surface roughness of Ti6Al4V alloy after laser polishing by the orthogonal test. It was found that the minimal roughness ($0.56\mu m$) was acquired at higher power (150 w) and high scanning speed ($30\mathrm{mm}/\mathrm{s}$). This finding is very similar to that of Jaritngam et al.⁴⁹ The surface roughness decreases slightly with the increase of scanning speed and power. Generally speaking, the increase or decrease of the surface roughness is related to the heat input and laser–material interaction time. Higher laser power results in sufficient heat input, and the high scanning speed avoids overmelting caused by long interaction time, thus ensuring the SSM regime. Then, the surface roughness of laser polished Ti6Al4V alloy is reduced. If low power and low scanning speed are used, the long-time laser–material interaction can also ensure sufficient heat input, but the processing efficiency is greatly reduced.

In addition to the selection of process parameters, the influence of other factors on the surface roughness of laser-polished Ti6Al4V alloy should also be taken into account. Tian et al.⁴² measured the surface roughness of laser-polished Ti6Al4V using a Nanofocus μScan laser profilometer and a white light interferometer with different resolutions and found that the Sa parameter was reduced from as-received 21.46 to $5.5\mu \mathrm{m}$ and $0.51\mu \mathrm{m}$, respectively. The difference in the measurement results was mainly due to the size of the sampling area, namely, $2\mathrm{mm}\times 3\mathrm{mm}$ and $0.13\mathrm{mm}\times 0.17\mathrm{mm}$. The small sampling area was similar to the $\beta$ grain size, and it revealed that the roughness after laser polishing was related to surface relief between individual grains and grain boundary grooving generated during the solidification process. The surface morphology is shown in Fig. 3. Zhang et al.⁵⁰ studied the in-situ polishing of additive manufactured Ti6Al4V with a fiber laser. There was no need to clamp the components twice after 3D printing. It was the same as that of the printing processing environment and almost no oxidation occurred. The results showed that surface roughness Sa was $3.32\mu \mathrm{m}$ after polishing when the laser power was 400 W and the feed rate was $0.5\mathrm{m}/\mathrm{s}$.

Fig. 3

Roughness reduction by laser polishing measured by surface profiling at different scales: (a) an SEM image showing the contrast between the original and polished surfaces; (b) the equivalent transition using the laser profilometer, with an area of $2\mathrm{mm}\times 6\mathrm{mm}$; (c) the LP surface measured over a $0.6\mathrm{mm}\times 0.86\mathrm{mm}$ area; and (d) measured over a $0.13\mathrm{mm}\times 0.17\mathrm{mm}$ area by white light interferometery.⁴²

In a word, the surface morphology and roughness are related to numerous factors. Besides the process parameters such as laser power density and scanning speed, there are the initial surface roughness of metal, homogeneity of the material, thermos-physical properties, and so on. These factors affect the polishing results to varying degrees. There are still many challenges in the surface finish of laser polishing, so tremendous efforts are needed to satisfy the requirements of titanium alloy in the aerospace, marine, and biomedical fields.

3.2.

Microstructure

The microstructure changes of Ti6Al4V alloy after laser polishing are usually shown in five aspects: surface oxidation, microcracks, recast thickness, phase change, and heat-affected zone depth. The above changes are closely related to heat input. During the process of laser irradiation, the temperature of the workpiece ascends gradually, and the heat accumulation is sufficient with the increase of heat input, leading to the thickening of the recast layer and the deepening of the heat-affected area. Jaritngam et al.⁴⁹ studied the effects of different laser powers and scanning speeds on the thickness of the recast layer and the depth of the heat-affected zone. It was found that when the power was 30 W and the scanning speed was $50\mathrm{mm}/\mathrm{s}$, the thickest recast layer was $70\mu \mathrm{m}$, and the deepest heat affected zone was $262\mu \mathrm{m}$. When the original surface roughness of workpiece was larger, the lower the scanning speed was, the thicker the recast layer would be under the same laser power. This is because the rougher the surface is, the higher the absorption of light is. However, the depth of the heat-affected area was independent of the original roughness. When the laser power was the same, the lower the scanning speed was, the deeper the heat-affected area was, as shown in Fig. 4. Furthermore, the thick and brittle recast layer was prone to crack under the action of high thermal stress. On the contrary, the crack is not obvious when using low power (10 and 20 w), as shown in Fig. 5. In short, the thickness of the recast layer and the depth of the heat-affected zone are related to the heat input. The more heat input, the more heat accumulation, the thicker the recast layer, and the deeper the heat-affected zone. This causes microcracks on the surface and subsurface. Based on the premise of ensuring the surface roughness, we can appropriately increase the scanning speed, reduce the laser power, or increase the repetition frequency to reduce the heat input to alleviate the thermal damage.

Fig. 4

(a) Thickness of recast layer and (b) depth of heat-affected zone induced by the different laser polishing conditions and initial surface roughness.⁴⁹

Fig. 5

Subsurface of laser-polished region obtained under the different processing conditions.⁴⁹

After laser polishing, the chemical composition of the Ti6Al4V alloy surface also changed. Vaithilingam et al.⁵⁴ analyzed the effect of laser remelting on surface chemistry of AM Ti6Al4V alloy and found the oxide film on the remelted surface was rich in aluminum, while the nonremelted surface was rich in titanium. The oxide layer of the remelting surface was thicker than that of the nonremelting surface. Generally speaking, the corrosion resistance of parts is directly proportional to the thickness of the oxide layer. However, the concentration of Al increases during rapid solidification because the solubility of Al in Ti is very low when the temperature reaches 500°C to 600°C and ${\mathrm{Ti}}_3\mathrm{Al}$ phase precipitates. When the heat input is higher, such as through decreasing the scanning speed, the material maintains a high temperature for a relatively long time and promote further precipitation. The high concentration of aluminum would affect the corrosion resistance of the workpiece. Therefore, the scanning speed and laser power should be properly controlled to avoid excessive heat input. On the other hand, the formation of the acicular ${\alpha }^{\prime }$ martensite phase after laser remelting also resulted in poor corrosion resistance of the component. Zhou et al.⁴⁸ optimized the process parameters through 16 groups of orthogonal experiments and obtained the lowest average roughness of workpiece of $0.56\mu \mathrm{m}$ when the power was 150 W, the scanning speed was $30\mathrm{mm}/\mathrm{s}$, and the spot overlap rate was 92%. Through electrochemical experiments, it is found that the corrosion resistance of parts after laser polishing is improved. The main reason is that the large peak valley value accelerates the corrosion when the surface of Ti6Al4V alloy is rough, while the surface roughness is small after laser polishing. Hence, the corrosion resistance is enhanced. In addition, the grain sizes also affect the corrosion resistance. The corrosion rate of fine grains is faster than that of coarse grains. Accordingly, the grain size in the polishing zone of Ti6Al4V alloy becomes larger after polishing, as shown in Fig. 6. In short, the change of microstructure such as chemical composition and grain size will affect the properties of materials, so it is necessary to optimize the process parameters to process the parts that meet the requirements.

Fig. 6

Electron backscatter diffraction (EBSD) results and grain size.⁴⁸

Another significant change of microstructure after laser polishing is phase transformation. Ti6Al4V is an $\alpha +\beta$ dual-phase Ti alloy, where Al is an $\alpha$ phase stabilizer, which can increase the phase transition temperature, and V is a $\beta$ phase stabilizer, which can reduce the $\beta$ phase transition temperature. During laser polishing, as the temperature rises to the beginning of the transition, the phase transition $\alpha \to \beta$ occurs. When it exceeds the transition temperature (about 1273 K), the $\alpha +\beta$ phase is almost completely transformed into the $\beta$ phase. As the laser beam leaves, the temperature decreases, and the surface of workpiece cools rapidly. Its cooling rate ($1.209\times {10}^4\mathrm{k}/\mathrm{s}$) far exceeds the critical cooling rate ($410\mathrm{k}/\mathrm{s}$), and the $\beta$ phase finally transforms into a needle-like α’ Martensite phase. Ma et al.³⁸ verified this phase transition process by SEM observation and XRD analysis before and after laser polishing, and the results are shown in Fig. 7. However, Marimuthu et al.⁵¹ had different opinions. He found through experiments that the columnar microstructure before and after laser polishing did not change much, mainly due to the similar cooling rate of powder particles during AM and laser polishing. Furthermore, he pointed out that the formation of the α phase could be explained by the oxygen concentration and material hardness before and after laser polishing. But, experimental results showed that maximum oxygen concentration was $<10\%$, which was similar to that before polishing, and the microhardness of the substrate and polished zone were 380HV and 420HV, respectively, which was not much different, so it revealed that there was no α phase and no change to the subsurface characteristic.

Fig. 7

Microstructure analysis of laser-polished TC4 surface: (a) overview, (b) microstructure of polished zone, (c) microstructure of substrate, and (d) XRD profiles.³⁸

In general, the different research results are due to the different lasers selected, the processing parameters being distinct, the heat absorbed by the workpiece being also not alike, and the changes in the microstructure being diverse. Oxidation during the polishing process is inevitable, but further study on the thickness and chemical composition of the oxide layer will help to improve the understanding. Phase transition can easily cause changes in mechanical properties and should be slowed down as much as possible. However, since the melting temperature (1923 k) exceeds the transformation temperature (1273 k), the first stage of phase transition($\alpha +\beta \to \beta$) is inevitable. The formation of secondary acicular Martensite(${\alpha }^{\prime }$) occurs in the cooling period. The cooling rate can be reduced by adding thermal insulation materials to slow down the phase transformation in the second stage. In a word, surface oxidation, microcracks, recast layer thickness, and heat-affected zone depth are all related to each other, and heat input is a key factor affecting surface quality and subsurface damage. Therefore, laser polishing of Ti6Al4V should be further optimized to improve surface quality and mechanical property, while minimizing thermal damage and other defects. The combination of multiobjective optimization and the Taguchi experiment can determine the best processing conditions based on the relevant response.

4.1.

Microhardness

The microhardness of additive manufactured Ti6Al4V alloy after laser polishing is one of the important mechanical properties. The cross-section of the titanium alloy after laser polishing is divided into three areas: the remelting area, the heat-affected zone, and the substrate layer. Compared with the base layer, the microhardness of the remelting area is significantly increased, while the hardness of the heat-affected zone is increased less. There are three main reasons for the increase in hardness of the polishing layer. First, ${\alpha }^{\prime }$ Martensite phase, which has a hexagonal close-packed structure, is formed and the bulk modulus is higher than for the $\alpha$ and $\beta$ phases in the original material, so the hardness increases. Second, the martensite phase has a high dislocation density and a large number of phase boundaries in a double-needle structure. It has an effect of fine grain strengthening and increases the surface hardness. Finally, the reduced porosity in the laser-polished layer increases the hardness and density of the material. Many articles^12,48,52 have similar reports. Li et al.⁵² found that the microhardness was 426 HV, which was about 25% higher than the as-received sample (340 HV), as shown in Fig. 8. The increase of the hardness also improves the wear resistance of the Ti6Al4V alloy. The wear scar of the laser polished surface is shallower than the as-received sample, and the wear rate is 39% lower than the original parts after calculation.

Fig. 8

Microhardness distributions in the laser-polished layer.⁵²

4.2.

Residual Stress and Fatigue Life

In the laser polishing of additive manufactured Ti6Al4V alloy, rapid melting and solidification make laser polishing an uneven thermodynamic process, which is prone to residual stress on the polished surface. Residual stress affects the propagation of the fatigue cracks and has an important impact on the fatigue life of parts. There are many ways to measure residual stress, for example, groove cutting, x-ray diffraction, ion beam milling, and laser line scanning profiling. These methods can obtain the residual stress distribution of the polished area. Tian et al.⁴² used the $\sin (2\theta )$ method by x-ray diffraction to measure the magnitude of residual stress generated during laser polishing. The results showed that there was a residual tensile stress of 580 MPa parallel to the scanning direction after laser polishing, while the residual stress in the vertical direction was 325 MPa. However, the residual stress disappeared at a depth of 1 mm from the surface. This was because the surface of the cold substrate was heated by a small moving heat source during the polishing process. The thermal stress relaxation caused by the local volume expansion under the heat source caused a compressive plastic mismatch. After the laser beam left, the material temperature decreased and the workpiece hardened, so the tensile stress was difficult to release. There was a thermal field elongation characteristic in the direction of heat source travel, so the residual stress parallel to the scanning direction was greater.

Under the action of alternating loads, fatigue failure of AM parts often occurs at cracks and warpage. After laser polishing, the defects are greatly improved, but the fatigue characteristics of the workpiece are still crucial. The fatigue performance of laser polished parts is mainly affected by microstructures, residual stress, and internal defects such as holes and microcracks. Li et al.⁵² studied the fatigue life of Ti6Al4V alloy before and after laser polishing and found that, compared with the as-received parts, the fatigue life of laser polished specimens decreased to ${10}^4$ to ${10}^5$ from ${10}^7$ cycles, as shown in Fig. 9(b). One of the reasons is that the formation of the ${\alpha }^{\prime }$ martensite phase makes the surface of the workpiece hard and brittle. Such a surface has a high notch sensitivity, which is prone to cracking and weakens the fatigue characteristics of the parts. After laser polishing, the elongation decreased by 5%, but the tensile strength and yield strength did not change much, as shown in Fig. 9(a). The residual tensile stress that occurs during laser polishing also accelerates fatigue fracture. Crack initiation occurs at the pores of the polished area and then propagates under the effect of residual stress, as shown in Fig. 9(e), finally resulting in a transient rupture zone with a pit shape, as shown in Fig. 9(c). This is a typical ductile fracture characteristic, indicating that the substrate has good ductility. In addition, the internal defects of the parts will also shorten the fatigue life. Therefore, the top priority for the next step is to reduce the residual stress and defects of additive manufacture, as well as suppressing the transformation of phase.

Fig. 9

(a) Tensile properties, (b) high cycle fatigue life, (c) dimples due to overload in (d), (d) overview of fracture surface, and (e) crack initiation area at the laser polished surface in (d).⁵²

Heat treatment is one of the main post-treatment methods for changing the residual stress. Leuders et al.⁵⁵ used three methods to heat treat SLM Ti6Al4V, namely 800°C (below β-transus) under Argon atmosphere, 1050°C (above $\beta$-transus) in evacuated quartz glass tubes, and hot isostatically pressed (HIP) at 920°C at a pressure of 1000 bar for 2 h. The results showed that the residual stress on the surface of the workpiece was greatly reduced after 800°C heat treatment, especially in the $y$ direction, from the original 235 to 10 MPa. The reduction of residual stress helped to slow down the growth of fatigue cracks. After heat treatment at 1050°C, the size of the $\alpha$ phase and the volume fraction of the $\beta$ phase increased, which enhanced the toughness of Ti6Al4V. The porosity of the workpiece decreased to the lowest after HIP, and the porosity was one of the main reasons for the fatigue crack initiation and fatigue strength reduction. Therefore, the heat treatment not only reduced the residual stress but also increased the fatigue life of the parts. On the other hand, when the stress of the part is large enough to cause plastic deformation or fracture of the material, the post-treatment hardly works. Therefore, the real-time stress control is also very important. The on-line thermal state detection of molten pool or temperature field can be realized by optical detection such as infrared imaging and high-speed x-ray imaging with an appropriate image algorithm. Through the closed-loop feedback control of the input laser energy, the surface temperature field is homogenized, the thermal gradient is reduced, and the residual stress is adjusted.^56,57

4.3.

Simulation Analysis

Numerical simulation provides a reference for experiments and is an indispensable part of much research. In the process of laser polishing, the finite element simulation method is often used to analyze the temperature field distribution, the depth and width of the molten pool, and the fluid flow speed. In general, the experimental outcomes are compared with the simulation results to verify the accuracy of the mold and experiment. Li et al.⁵² considered that the phase transformation process in the laser polishing area mainly depended on the temperature change and the cooling rate. Using the ABAQUS to simulate the temperature field distribution of the Ti6Al4V alloy during laser polishing according to the transient heat balance equation in the model, it was found that the temperature decreased from 2800°C in the melting zone to 1600°C in the heat-affected zone. Then, the temperature field was compared with the cross-sectional view after laser polishing, as shown in Fig. 10, and the formation process of the needle-shaped ${\alpha }^{\prime }$ martensite phase was explained. The heat transfer mechanisms and the change of surface morphology (surface ripple) during laser polishing of Ti6Al4V alloy were studies by Marimuthu et al.⁵¹ It had been found that the width of the molten pool magnified with the increase of the laser power. When the power was 160 W, the width reached the maximum (0.41 mm). As the laser power was further increased, the rate of increase in the width of molten pool decreased exponentially, but the thermal energy increased, which increased the velocity of the molten pool. Under the effect of negative surface tension coefficient, the melted material flowed outward from the center of molten pool, causing the center of the molten pool to swell and the boundary of the molten pool to sag, as shown in Fig. 11. Therefore, the high melt pool velocity made the surface morphology worse. On the other hand, Fig. 12 shows the change of the melt pool speed with the laser power. The surface tension gradient (Marangoni convection) was the main factor affecting the melt pool dynamics and convective fluid flow. At the same time, surface tension was a function of the temperature gradient in the melt pool, and it increases with decreasing temperature. The temperature of the center of the molten pool was the highest and the edge was lower, so the fluid flowed outward under the effect of tension, providing effective heat conduction from the center to the boundary. Ma et al.⁵⁸ studied the relationship between the shape of molten pool and the profile of the solidified surface and the laser pulse duration during the laser polishing of Ti6Al4V alloy through numerical simulation. The molten pool deepened with the prolongation of pulse duration. The peak-valley value (PVH) of the resolidified surface was normalized by the radius of the molten pool, and the normalized PVH changed when the pulse width was obtained, as shown in Fig. 13. This shows that the optimal pulse duration was $0.66\mu \mathrm{s}$ when the workpiece surface reached the smoothest ($\mathrm{PVH}=0$). There are also some references^59–61 that have solved some problems in the experiment by simulating the laser polishing process. However, there are still few reports on numerical models, and that will be the next focus here.

Fig. 10

(a) Temperature distribution of cross-section view, (b) microstructure, (c) polished zone, (d) heat-affected zone, and (e) as-received zone.⁵²

Fig. 11

Comparison of liquid fraction (%) along the midpoint cross-section transverse to polishing direction, showing the change in surface topology of (a) 120 W, (b) 160 W, and (c) 200 W, with a constant speed of $600\mathrm{mm}/\min$.⁵¹

Fig. 12

Comparison of top surface velocity vector (m/s) for laser power of (a) 120 W, (b) 160 W, and (c) 200 W, with a constant speed of $600\mathrm{mm}/\min$.⁵¹

Fig. 13

The relationship between normalized PVH and pulse duration.⁵⁸