The polishing experimental device is depicted in Fig. 4, including polishing tools and sample fixtures. The polishing tool is mounted with an axially magnetized permanent magnet and installed on the spindle of the high-precision controlled three-axis linkage CNC machine tool, which can realize linear motion and rotational motion. The spindle rotation can drive the rotation of the permanent magnet, thereby driving the rotation of the polishing pad. The sample is clamped on the pliers, and the polishing pressure of the MRP pad on the polishing area is changed by adjusting the gap between the polishing tool and the sample.

In this study, 316L metal cube samples formed by L-PBF are used as polishing objects. Spherical 316L stainless steel powder (Beijing Yijia 3D Co., Ltd.) is used as the experimental material (Fig. 5). The composition is depicted in Table 1, and the particle size distribution is depicted in Table 2. The equipment adopts SLM-125 laser selective melting equipment (Fig. 5), the equipment adopts a single laser (1 × 400 W) IPG fiber laser, the maximum molding size is 125 mm × 125 mm × 125 mm, the maximum scanning speed is 10 m s-1, the layer thickness is 0.02–0.075 mm, and the molding room uses argon as protective gas. The forming process parameters are shown in Table 3.

Figure 6a shows the polished sample of 316L stainless steel formed by L-PBF. The sample size is designed to be 30 mm × 30 mm × 35 mm hollow pentahedron, and the original surface roughness Ra is 5–15 µm. Figure 6b shows the surface micro-morphology of L-PBF-formed 316L stainless steel polishing sample. Adhesion and spheroidization of powder are evident on the surface, and argon is used as the protective gas in the molding chamber. Therefore, gas flow in the cabin drives the surrounding powder to cause the adhesion of the powder on the surface. Powder spheroidization occurs due to the molten liquid scattered on both sides of the molten pool after cooling and solidification. Spheroidization will lead to irregular shapes along the scanning trajectory, so the weld gap will occur repeatedly, thereby affecting the surface quality of the sample. Before the experiment, the sample is roughly cast to about 1.7 µm in the grinder.

In the MRP process of the L-PBF-formed 316L stainless steel polishing sample, MRP fluid and process parameters are the main influencing factors. The primary variables affecting polishing quality and efficiency are abrasive particle size and hardness. Diamond, SiC, Al2O3, and other polishing materials are frequently utilized. Table 4 displays the abrasive particle performance. Al2O3 powder is chosen as the abrasive particle of the slurry for 316L stainless steel that SLM has manufactured. Al2O3 powder with a particle size of 3.0–4.0 µm is chosen to complement iron powder because carbonyl iron powder has a particle size of 3.0–4.0 µm (Nagdeve et al., 2018).

Four factors including polishing abrasive particle content, magnetic particle content, working gap, and spindle speed are selected to carry out single-factor experiments. The factor level data are depicted in Table 5. After polishing, the surface roughness of the sample was measured by Mahr-XR20 roughness meter, and the average value of three measurement points was taken as surface roughness Ra to evaluate the polishing effect. The surface morphology of the sample was observed by a white light interferometer.

During MRP, magnetic particles can clamp polishing abrasive particles to remove the material on the workpiece surface, and the content of polishing abrasive particles will affect the polishing effect to a certain extent. MRP fluid containing 3 % vol, 4 % vol, 5 % vol, and 6 % vol of brown corundum abrasive particles (namely, Al2O3) was selected for single-factor polishing experiments under the following conditions: polishing time of 30 min, working gap of 0.4 mm, spindle speed of 600 r min-1, and 45 % vol of iron powder in the polishing fluid. Figure 7 shows the influence of the polishing abrasive on the surface roughness and residual stress of the sample. With increasing content of the polishing abrasive, the surface roughness first decreases and then increases, and the maximum difference is 0.12 µm. When the polishing abrasive is 4 % vol, the surface roughness Ra of the sample reaches a low value of 0.18 µm. The residual stress increases from 41.16 to 71.2 MPa, and the overall broken line shows a slight upward trend, indicating that the increase in abrasive content had little effect on residual stress.

Figure 8 shows the polishing patterns of the workpiece observed by the white light interferometer under different polishing abrasive contents. When the content is 3 %, the number of abrasive particles involved in polishing is small, the magnetic particles can better constrain the polishing abrasive particles, and the surface roughness of the sample can reach 0.23 µm. However, the sample plot has many weld defects and obvious plastic removal traces because the polishing abrasive particles are less and the removal amount of the abrasive particles on the overall material is also less; as such, surface defects cannot be completely removed. When the contents are 5 % and 6 %, the surface processing quality is poor, and weld defects and obvious plastic removal traces are found on the surface of the sample. With increasing content of abrasive particles, the number of abrasive particles between magnetic particles increases, thereby affecting the formation of flux linkage and weakening the magneto-rheological effect. The abrasive particles involved in polishing are easy to escape the constraint, lacking the effect of only magnetic particles on the extrusion of the workpiece, and cannot effectively achieve the removal effect. The surface of the sample has many weld defects. When the abrasive content is 4 %, the surface quality is the best and the sample surface is flat and smooth with fewer defects.

The magnetic particles in MRP fluid are arranged into chains along the magnetic field line. The abrasive particles originally dissociated in the polishing fluid are clamped between magnetic particles and between flux linkage. The content of magnetic particles directly affects the formation of flux linkage, thereby affecting the polishing effect. MRP fluid containing 35 % vol, 40 % vol, 45 % vol, and 50 % vol hydroxyl iron powder was selected for single-factor polishing experiment under the following conditions: polishing time of 30 min, working gap of 0.4 mm, spindle speed of 600 r min-1, and 5 % vol polishing fluid. Figure 9 shows the changes in the surface roughness and residual stress of the sample with different hydroxyl iron powder contents. The surface roughness decreases first and then increases with increasing magnetic particle content, and the maximum difference is 0.12 µm. When the content is 40 %, the surface quality is good and the surface roughness reaches 0.13 µm. The residual stress increases from 68.72 to 91.71 MPa, showing a slight upward trend as a whole, indicating that the increase in the magnetic particle content had slight effect on residual stress.

Figure 10 shows the polishing patterns of workpieces observed by a white light interferometer with different magnetic particle contents. When the content is 35 %, the sample has a small number of pits and plastic removal traces. The iron powder is less than the abrasive, and the clamping force provided by the iron powder is limited; as such, the abrasive cannot completely remove the surface defects. The surface quality is low, and the corresponding Ra value is 0.15 µm. When the content of iron powder is 40 %, the pits and removal traces are significantly reduced compared with that at 35 %. With increasing number of iron powder particles and the same content of abrasive particles, the number of iron powder and abrasive particles is close to the equilibrium. The clamping force provided by iron powder particles cannot only remove the surface defects of abrasive particles but also enable them to be timely separated to update the abrasive particles. Finally, the overall material removal is increased and the surface quality is improved. The optimal Ra value is 0.13 µm. When the contents are 45 % and 50 %, the surface of the sample has pits and removal marks. The amount of iron powder exceeds the number of abrasive particles due to the increase in the iron powder content, and the clamping force provided by iron powder particles is large. As such, the abrasive particles involved in polishing have difficulty to escape constraints, resulting in the difficulty of updating the passivated abrasive particles and affecting the quality of the final surface.

In different sizes of the working gap, the different thicknesses of the polishing pad formed by polishing liquid may affect the polishing force of the sample surface and the material removal rate and surface quality. The control experimental conditions are as follows: polishing time of 30 min, spindle speed of 600 r min-1, polishing liquid composition with iron powder concentration of 45 % vol, brown corundum concentration of 5 % vol under 0.2, 0.4, 0.6, and 0.8 mm processing conditions. Figure 11 shows the influence of working gap on the surface roughness and residual stress. With decreasing working gap, the surface roughness Ra first decreases and then increases, the maximum difference is 0.151 µm and reaches a small value at 0.6 mm ,and the surface roughness Ra is 0.099 µm. The residual stress decreases from 236.74 to 43.32 MPa with increasing working gap.

Figure 12 shows the surface topography of the workpiece observed by a white light interferometer under different working gaps. When the working gap is 0.2 mm, the surface of the sample has fewer residual pits and plastic removal traces during polishing. The polishing pad is thick, the magnetic field intensity is high, the binding force between the magnetic particles is strong, the position between the particles is relatively fixed, and the polishing force is large due to the small working gap. With larger binding force of the magnetic flux on the abrasive particles, the larger the polishing force is, the larger the residual stress is, and the surface roughness can reach 0.18 µm. However, due to the small working gap and the short magnetic flux, the abrasive particles are less, thereby affecting the removal effect. When the working gap is 0.6 mm, the working effect is better, the sample surface is smooth, and the pits are fewer; when the working gap is 0.8 mm, many unfinished meteoric plastic removal traces are found on the figure. The gap is too large and the magnetic field intensity is low; as such, the magnetic particle chain does not have sufficient clamping force on the abrasive particles. When the abrasive particles remove the material, the cutting force is greater than the clamping force, so the abrasive particles are separated from the clamping of the magnetic particles, thereby forming meteoric plastic removal traces.

Spindle speed affects the relative linear velocity between the polished particle and the surface, thereby affecting the material removal of the sample surface by the MRP pad. The control experimental conditions are as follows: the working gap is 0.4 mm, the polishing time is 30 min, and the polishing liquid composition is 45 % vol of iron powder and 5 % vol of brown corundum. Figure 13 shows the influence of spindle speed on surface roughness and residual stress at different rotating speeds. With increasing spindle speed, the surface roughness Ra first decreases and then increases, and the maximum difference is 0.11 µm. When the spindle speed reaches 600 r min-1, the surface roughness is the lowest, reaching 0.10 µm. The residual stress decreases with increasing working gap from 119.34 to 39.89 MPa.

Figure 14 shows the surface topography of the workpiece observed by the white light interferometer at different spindle speeds. When the spindle speed is 200 r min-1, the micro-pattern has many defects and plastic removal traces and a visible channel direction. When the spindle speed is low, the polishing times of the workpiece are limited and the material removal amount is less in unit time; that is, the surface roughness is high. With increasing spindle speed, the material removal linear velocity and particle trajectory density of abrasive particles on the surface of the sample increase, thereby increasing the material removal rate and the surface quality of the sample increases and decreasing the surface roughness. When the speed reaches 600 r min-1, the surface roughness value is the lowest, reaching 0.10 µm. When the rotational speed exceeds 600 r min-1, the surface roughness increases. With the increase in the rotational speed of the spindle, the centrifugal force of the particles in the MRP pad increases, resulting in the deformation of the flux linkage, the decrease in the clamping force on the particles, and the escape of the particles from the magnetic chain. The particles are marginalized in the polishing area, the update speed of the particles in the flux linkage decreases and the material removal amount per unit time is less, thereby increasing the surface roughness.

Based on the above experimental results of each process parameter, the optimized process parameters of L-PBF formation of 316L stainless steel are as follows: abrasive content of 4 %, magnetic particle content of 40 %, working gap of 0.6 mm, and spindle speed of 600 r min-1. The sample was polished for 30 min under the optimized parameters. The surface morphology of the workpiece observed by the white light interferometer of the 316L stainless steel sample formed by L-PBF is depicted in Fig. 15. The surface roughness Ra of the workpiece decreased from the initial 1.59 µm to 61.43 nm. The comparison before and after polishing is shown in Fig. 16, and the surface became smooth.