Oxidation mechanism of high-volume fraction SiCp/Al composite under laser irradiation and subsequent machining

Combining the advantages of the particles and the metal, particle-reinforced metal matrix composites (MMCs) were developed several decades ago and are now used in various areas.^1–3 Among the most widely used MMCs are those comprising an aluminum matrix reinforced with SiC particles (hereinafter referred to as SiC_p/Al composites). They have excellent properties such as high specific strength and good wear resistance and dimensional stability, which is why SiC_p/Al composites are used widely in aerospace, electronics, and other industries.^4,5 However, their inherent properties of high hardness, high abrasion resistance, and multi-interface characteristics complicate their mechanical machining.

SiC_p/Al composites can be machined mechanically using traditional machining processes such as drilling, turning, milling, and grinding,^6,7 but nontraditional machining processes can also be used, such as electrical discharge machining (EDM), wire cut EDM (WEDM), abrasive water jet machining, and laser-assisted machining (LAM). Xiang et al.⁸ researched the mechanisms for tool wear during the drilling of an SiC_p/Al composite with 65% particles by volume (hereinafter referred to as a 65% SiC_p/Al composite), and their results showed that the best drill was a carbide one coated with diamond via chemical vapor deposition, which has the advantages of stable cutting force, reduced wear, increased number of machined holes, and acceptable machining quality; the main wear mechanisms of the polycrystalline diamond (PCD) tool were abrasive, adhesive, and chemical transformation. Yin et al.⁹ established theoretical and finite-element models for the grinding force in single-grain diamond grinding of SiC_p/Al composites; they investigated the grinding force, stress distribution, and material removal mechanisms, and the results indicated that the material removal can be divided into four stages: (i) plastic removal of the aluminum matrix, (ii) crack initiation in the SiC particles, (iii) crack propagation in the SiC particles, and (iv) brittle fracture of the SiC particles.

Because of the high hardness of SiC particles, the processing quality obtained by traditional processing methods is relatively poor, with inevitable and serious tool wear, particle breakage, and surface defects.^10,11 To solve these problems, several nontraditional processing methods have been applied,^12–14 with one of the most widely used being EDM, which can machine electrically conductive materials regardless of their hardness.¹⁵ Zhang et al.¹⁶ used magnetically assisted EDM on a 65% SiC_p/Al composite, showing that the pulse current is the main factor influencing the surface integrity. Indeed, magnetically assisted technology can improve the surface integrity significantly. Meanwhile, laser machining is a material removal method that involves using high-energy laser beams; it can process materials with high melting point, hardness, and brittleness, and it is characterized by great flexibility.¹⁷ Zhang et al.¹⁸ used a nanosecond fiber laser to irradiate an SiC_p/AA2024 composite and then investigated the laser–material interaction mechanisms; when irradiated by the laser, the SiC particles bulged and the Al alloy matrix was melted and sputtered, and the heat-affected zone (HAZ) increased logarithmically with the pulse energy density.

LAM uses a high-energy laser beam to irradiate the local workpiece surface in front of the cutting edge, thereby reducing both the strength and hardness of the material, which is then removed by the cutting tool; consequently, the cutting force is reduced and the tool life is improved.^19,20 Kong et al.²¹ evaluated the tool life and studied the wear mechanisms of both uncoated and coated carbide tools in the LAM of a 45% SiC_p/Al composite; the tool life was longer with LAM compared to conventional machining, and the most predominant wear mechanism in LAM was abrasive wear. However, to soften a workpiece material such as an SiC_p/Al composite with high hardness and melting point, high laser energy is required, and so the HAZ is relatively large and the machined surface integrity is reduced. Also, the cutting tool must remove the softened material at high temperature, and so its cutting performance is reduced. To solve these problems, Zhao et al.^22,23 proposed the hybrid process of laser-induced oxidation-assisted milling (LOAM). During this process, the target area is irradiated by a pulsed laser beam, and the material reacts with oxygen that is supplied simultaneously; this generates a porous oxide layer that is removed quickly by the cutting tool, thereby improving significantly both the machinability of the workpiece material and the tool life.

LOAM is an effective means of reducing the cutting force and prolonging the tool life in the machining of SiC_p/Al composites, but its mechanism for material modification induced by laser irradiation requires further study. An SiC_p/Al composite comprises SiC particles and an aluminum metal matrix, and various laser interactions may occur because of their different optical and thermal properties. The fundamental issue in LOAM is the oxidation mechanism of the SiC_p/Al composite under laser irradiation, and understanding that issue would offer better understanding of the oxidation behavior and a method for controlling the laser-induced oxidation process. Herein, we investigate the oxidation mechanism with a 65% SiC_p/Al composite as the workpiece material and a nanosecond pulsed laser with a series of pulse energy densities. Laser irradiation experiments were carried out using oxygen, nitrogen, or argon as the auxiliary gas, and the mechanism for forming an oxide layer was analyzed by conducting multiple scans of the composite for each combination of laser parameters. Furthermore, the cutting force and surface quality are compared between LOAM and conventional milling (CONM).

In this study, the workpiece material was a 65% SiC_p/Al composite, and its mechanical properties as provided by the supplier are listed in Table I. The average particle size of the SiC particles was 10 μm, and the morphology of the workpiece is shown in Fig. 1. The samples were cut into 30 × 5 × 5 mm³ pieces using WEDM and were then ground using a boron carbide abrasive. Moreover, to remove any impurities, the samples were cleaned ultrasonically in acetone for 15 min.

To analyze the reaction mechanism between the laser and the material in different gas environments, we used oxygen, nitrogen, or argon as the auxiliary gas with the same laser parameters. To analyze the formation process of an oxide layer under laser irradiation, we performed once, twice, and five and 50 times for each auxiliary gas and set of laser parameters. The scanning speed v was 1 mm/s, and the track displacement between adjacent scanning lines was 10 μm. The laser scanning paths are shown schematically in Fig. 3.

A vertical milling center was used in the milling experiment, and the milling tool was a straight-tooth PCD one with a cemented carbide shank brazed with PCD plates. The morphology of the PCD milling tool is shown in Fig. 4, and its parameters are listed in Table II. The LOAM parameters were a spindle speed of n = 10 000 rpm, a feed rate per tooth of f_z = 7.5 μm/z, and a depth of cut of a_p = 0.1 mm. For comparison, CONM was also carried out using the same parameters and material removal.

A scanning electron microscope (SEM) (S8220; Hitachi) was used to observe the working material, the laser-treated surface, and the HAZ. The elemental compositions were measured using energy-dispersive spectroscopy (EDS), and the products were detected using x-ray diffraction (XRD). The machined surface morphology was observed using another SEM (S4800; Hitachi), and the surface roughness was obtained using an M1 perthometer; the workpiece was measured three times in different areas of the machined surface to obtain the arithmetic mean roughness. The measurement process complies with the national standard GB1031-83. The milling forces were measured using a dynamometer (9256C1; Kistler) with a maximum sampling frequency of 30 kHz and a threshold force of 0.002 N.

The pulse energy density F is the main factor affecting the interaction between the pulsed laser and the material. To analyze the effects of the laser energy density on the 65% SiC_p/Al composite, we used F = 7.84, 11.75, 15.67, 19.59, and 23.51 J/cm². The surface morphologies of the material obtained after a single oxygen-assisted scan are shown Fig. 5; Figs. 5(a)–5(e) show SEM images of the irradiated surface under different pulse energy densities, and Figs. 5(f)–5(j) show the corresponding cross-sectional morphologies.

From Figs. 5(a)–5(e), both the processing area and the HAZ on the surface increase with increasing laser energy density, and from Figs. 5(f)–5(j), so does the laser processing depth. For F = 7.84 J/cm² [Fig. 5(a)], a small groove appears on the surface of the material and follows the laser scanning path, with white fine granular oxide distributed on both sides of the groove; the groove is very shallow and there is almost no HAZ [Fig. 5(f)], and tiny oxide deposits on the scanning area protrude from the surface. For F = 11.75 J/cm² [Fig. 5(b)], the width of the HAZ on the machined surface increases to 220.4 μm, and the oxide becomes denser and accumulates on both sides of the groove; the groove depth on the corresponding cross section increases to 60.1 μm, and the HAZ is more obvious [Fig. 5(g)]. For F = 19.59 J/cm² [Fig. 5(d)], the HAZ width is 282.1 μm, and the oxide on both sides of the groove increases in volume and becomes denser; the groove depth increases to 271.3 μm [Fig. 5(i)]. For F = 23.51 J/cm², the HAZ width is 320.2 μm [Fig. 5(e)] and the groove depth is 308.8 μm [Fig. 5(j)]. Therefore, with increasing laser energy density, both the HAZ width and groove depth increase, as does the amount of oxide deposited on both sides of the groove.

Figure 6 shows more intuitively how the HAZ width and groove depth change with the laser energy density, and also included are the data obtained with nitrogen and argon for the same energy densities. From Fig. 6(a), with increasing F, the HAZ width increases with the assistance of oxygen, nitrogen, or argon, and the observed growth trends are similar. The HAZ width increases rapidly at lower F but slower at higher F, suggesting that it approaches a fixed value for sufficiently high F; we attribute this to the limitation of the spot diameter. Furthermore, under the same energy density, the HAZ widths obtained with the different auxiliary gases are similar, albeit that that obtained in an oxygen environment is slightly higher because oxygen is more active than nitrogen or argon and so reacts more easily with the high-temperature material after laser irradiation. From Fig. 6(b), the depth of the groove obtained by laser irradiation also increases with increasing F. The rate of increase of the groove depth is larger than that of the HAZ width, and the groove depths obtained with the same laser energy density but different auxiliary gases are very similar.

Therefore, the laser energy density influences significantly both the width of the HAZ on the material surface and the depth of the groove. With increasing laser energy density, the energy absorbed by the material increases, and this intensifies the reaction of the workpiece material, thereby increasing both the HAZ width and the groove depth. Using oxygen as the auxiliary gas promotes the HAZ on the material surface, but for the same laser energy density, the choice of auxiliary gas has little effect on the machining groove depth.

The quality of the oxide layer generated in an oxygen environment is better than that in an atmospheric environment, and the oxide layer is denser and easier to remove under the same laser parameters. To investigate further how the gas environment influences the formation of an oxide layer in the laser-induced oxidation process, we selected oxygen, nitrogen, or argon as the auxiliary gas, and we subjected the machined surface to EDS and XRD to analyze the products in the different gas environments.

Figure 7 shows SEM images of the machined surface for a scanning speed of 1 mm/s, a pulse energy density of 23.51 J/cm², a track displacement of 10 μm, and laser scanning performed 50 times. Figure 7(a) shows the morphology of an oxide layer obtained in oxygen. A white porous oxide layer forms in the laser scanning area, being very loose and easily removed by ultrasonic cleaning. The surface of the oxide layer is wavy and undulating and contains some white flocs; these are oxides generated from the oxide layer under the blowing of the oxygen flow, which adhere to the oxide layer surface. According to the cross section in Fig. 8(a), the oxide layer has a relatively large volume, and it protrudes to a certain height from the surface of the material at this pulse energy density. On the other hand, there are many holes in the lower part of the oxide layer, and the part near the material is irregularly shaped; this part comprises the bottom of the groove and the resolidified material. Figure 7(b) shows the morphology of a recast layer obtained in nitrogen. Its surface is rough, irregular, and composed of many fine particles and clusters with dense gaps between them. The recast layer is black in color and difficult to remove. Figure 8(b) shows the recast layer in cross section, again protruding from the material surface. However, the volume of the recast layer is less than that of the oxide layer, even though the former contains many folds and crevices that increase its volume. The bottom morphology of the recast layer is similar to that of the oxidation layer, and this can be attributed to the gas being unable to penetrate the material far from the workpiece surface. Figures 7(c) and 8(c) show the morphology and cross section, respectively, of the recast layer obtained in argon. The surface morphology of this recast layer is similar to that obtained in nitrogen and is a typical morphology for laser reaction with a material in an inert gas.

From comparing the morphologies obtained in the three auxiliary gases, oxygen is best for promoting the formation of the oxide layer. In an inert gas, there is almost no reaction between the gas and the material, and under laser irradiation, the material removed by sputtering solidifies again to form a recast layer with high hardness and irregular morphology and that is very rough and difficult to remove; consequently, this has adverse effects on cutting. However, in oxygen, the molten material that would otherwise have formed a recast layer instead reacts with the oxygen to form a loose oxide layer. This happens because oxygen reacts relatively easily with the material, and compared with the recast layer, the oxide layer is larger and easier to remove, which is conducive to further machining.

The EDS results for the surface of the workpiece irradiated in oxygen, nitrogen, and argon are shown in Figs. 7(d)–7(e), respectively, for the sampling areas marked by red boxes A, B, and C in Figs. 7(a)–7(c), respectively. In Fig. 7(d), the oxygen content is the highest (51.1%) and the aluminum content is the lowest (9.3%). The oxide layer consists mainly of SiO₂ and Al₂O₃. The vaporization point of aluminum (2740 K) is lower than that of silicon carbide (3103 K), so when the temperature rises under laser irradiation, aluminum is more likely to vaporize and escape into the air, even if it reacts with oxygen to form Al₂O₃. It is also blown away by the airflow, making it difficult for it to accumulate on the workpiece surface. Silicon carbide has a higher vaporization temperature and mainly forms molten matter at high temperatures. Moreover, it forms SiO₂ with oxygen, which accumulates easily in the oxide layer. Therefore, the Si content is increased and the Al content is decreased within the oxide layer. As shown in Figs. 7(e) and 7(f), the results of EDS analysis under nitrogen and argon are similar, i.e., lower oxygen content and higher carbon and aluminum contents. The main components of the recast layer are resolidified silicon carbide and aluminum; the carbon in silicon carbide does not react with nitrogen or argon, which is why the carbon content is increased, but because aluminum is active, that on the surface of the recast layer reacts with oxygen in the atmosphere after laser irradiation to form Al₂O₃, which is why less oxygen is detected.

To analyze further the reaction mechanism and the formation process of an oxide layer under laser irradiation, oxygen was used as the auxiliary gas during the irradiation experiment with a laser energy density of 23.51 J/cm². The scans were performed once, twice, and five and 50 times, and the scanning interval was 10 μm. SEM images of the irradiated surface with different scanning times are shown in Fig. 10, and SEM images of the cross section are shown in Fig. 11.

As shown in Fig. 10(a), under laser irradiation, a groove is formed in the central area of the scanning path, with convex oxide deposits distributed on both sides of the groove. In Fig. 10(b), when scanning the second path, the laser beam moves 10 μm to the right with respect to the position of the first scan. There is still a groove on the surface, which is located in the center of the second scan path, but the groove formed in the first scan is filled and covered by the oxide. The HAZ width is 174 μm on the left side and 150 μm on the right side. As shown in Fig. 10(c), after scanning five times, there is only one groove on the surface, which is distributed in the center of the fifth scanning path; the grooves generated in the first four scans are covered by the oxide. The HAZ width on the left side of the groove is enlarged to 213 μm, and that on the right side is 140 μm. The oxide accumulation area on the left side also widens gradually. After 50 scans, the groove is located in the center of the 50th scan path, as shown in Fig. 10(d), and the oxide layer accumulated on the left side forms a porous oxide layer.

Figure 11(a) shows a SEM image of the cross section obtained after one scan; the groove generated by laser irradiation penetrates the material, with little protruding oxide on either side. Figure 11(b) shows the cross section after two scans; compared with Fig. 11(a), the groove is now slightly wider with more oxide on its left side. Figure 11(c) shows the cross section after five scans; the groove is now even wider with even more oxide on its left side. Figure 11(d) shows the cross section after 50 scans; a loose oxide layer was produced in the scanning area, and the groove formed during the final scanning remains on the far right. In summary, the oxidation products generated by each scan fill the groove generated by the previous one, and the oxides accumulate gradually to form the oxide layer; this is shown also by the texture trend of the end face of the oxide layer, which is growing to the left.

As shown in Fig. 12(d), the generated oxidation products (SiO₂ and Al₂O₃) are deposited and stacked on both sides of the grooves, with CO₂ released into the air. Figures 12(e)–12(h) show the second laser scanning process, whose reaction process is the same as that of the first one. The difference now is that the oxide generated by the second laser irradiation accumulates in the groove generated by the first one. Figure 12(i) shows a schematic of the situation after multiple scanning, where the oxide accumulates gradually in the groove until a loose oxide layer is formed.

To verify the feasibility of LOAM, experiments were performed in which the 65% SiC_p/Al composite was machined using both LOAM and CONM. The milling parameters used in both processes were a spindle speed of n = 10 000 rpm, a feed rate per tooth of f_z = 7.5 μm/z, and a cut depth of a_p = 0.1 mm.

Figure 13 shows how the milling force measured for each machining process varied, where the machining length of each tool feed was 50 mm. For LOAM, in the first eight tool feeds, the milling tool mainly cut off the loose oxide layer, and the ninth and tenth tool feeds mainly removed the residual oxide layer, transition layer, and some matrix materials to avoid the influence of the HAZ. For CONM, the milling force components F_x and F_y were 15.1 and 14.2 N, respectively, for the first tool feed, and with more tool feeds, the tool wear increased gradually, as did F_x and F_y; at the tenth tool feed, F_x and F_y reached 24.8 and 19.4 N, respectively. For LOAM, F_x and F_y were 3.2 and 2.7 N, respectively, in the first tool feed. Because the loose oxide layer was always removed during the first eight tool feeds and the tool was almost free of wear, there was no significant change in milling force. In the eighth tool feed, F_x and F_y were 3.9 and 2.8 N, respectively. Because of the loose combination between oxides, it was easy to remove the oxide layer and produce powdery chips. Thus, the milling force was very small. The ninth tool feed removed the transition layer and some base materials, and F_x and F_y increased to 8.2 and 7.4 N, respectively. The transition layer contains residual oxide, the recast layer, and heat-affected material, which is easier to remove than workpiece material. Thus, the cutting force when removing the transition layer was less than that of CONM using the same cutting parameters. The tenth tool feed mainly removed the matrix material to avoid the influence of the HAZ. At this time, F_x and F_y were 16.2 and 15.3 N, respectively. Only slight tool wear occurred when removing the transition layer, and therefore the milling force was similar to that in the initial processing stage in CONM. By using LOAM, the cutting force can be reduced greatly and the tool life can be improved greatly.

Figures 14(a) and 14(b) show SEM images of the machined surface obtained by LOAM and CONM, respectively, with the same selective magnification. The surface roughness R_a was measured to be 0.131 and 0.265 μm, respectively. According to Fig. 14(a), the surface obtained by LOAM was very flat with few machined surface defects. As can be seen, there were only a few pits with small size and depth, which can meet the application requirements of high-volume-fraction SiC_p/Al composites. According to Fig. 14(b), the surface obtained by CONM had more defects. Compared with LOAM, CONM produced significantly more pits in the same area, and their average size was much larger than that under LOAM. In addition, the trace of the cutting path is more obvious. In the LOAM experiment, the oxide layer was loose and porous, so the milling force when removing it was very small. Consequently, the tool was hardly worn and remained sharp, so the compression was small when removing silicon carbide particles. Thus, the number and average size of pits generated on the machined surface were reduced. Because of the sharp tool nose and cutting edge, the cutting load was small and the cutting process was stable, so the cutting path was not obvious. Therefore, the surface quality obtained by LOAM is better than that by CONM.

Based on the above analysis, LOAM is effective in reducing machined surface defects and improving the quality of the machined surface. This is because the tool in LOAM is removing the loose oxide layer most of the time, so the tool wear is very small, and the tool remains sharp in the final finishing stage.

In this paper, the oxide-layer formation laws and mechanism for a 65% SiC_p/Al composite under nanosecond pulsed laser irradiation were studied in detail, and LOAM and CONM were compared using the same milling parameters. We draw the following main conclusions.

With increasing laser energy density, both the ablated groove depth and HAZ width increase, but there is no influence on the groove depth with different gas assistance. Every time the laser scans the material area, the sputtered molten material reacts with oxygen to form oxides, and the accumulated oxides gradually form an oxide layer with more scans.

An oxygen environment promotes the formation of an oxide layer. In an oxygen environment, a loose oxide layer forms that is relatively easy to remove.

Compared with CONM, LOAM is better for reduced milling force and tool wear. Furthermore, machining damage is contained and higher surface quality is achieved with the LOAM.

This work was supported by the Fundamental Research Funds for the Central Universities (Grant No. NT2021020).

The authors have no conflicts to disclose.

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Hanliang Liu is a Senior Engineer at Beijing Spacecrafts, China Academy of Space Technology, Beijing, China. His research interest include advanced manufacturing technology for spacecraft, and digital and intelligent manufacturing technology.

Guolong Zhao is a professor at Nanjing University of Aeronautics and Astronautics. His research interest are hybrid machining, micro-machining, and cutting tool technology.

Zhiwen Nian is a Ph.D. student at Nanjing University of Aeronautics and Astronautics. His research interest are hybrid machining and machining technology for ceramic matrix composites.

Zhipeng Huang is a Senior Engineer at Beijing Spacecrafts, China Academy of Space Technology, Beijing, China. His research interest include digital workshop management, construction, and applications.

Kai Yang is an Engineer at Beijing Spacecrafts, China Academy of Space Technology, Beijing, China. His research interest include CAD/CAM technology and CNC machining technology.

Conghua Liu is an Engineer at Beijing Spacecrafts, China Academy of Space Technology, Beijing, China. His research interest include CAD/CAM technology and CNC machining technology.

Peng Wang is an Engineer at Beijing Spacecrafts, China Academy of Space Technology, Beijing, China. His research interest include CAD/CAM technology and CNC machining technology.

Zhenkuan Diao is an Engineer at Beijing Spacecrafts, China Academy of Space Technology, Beijing, China. His research interest include CAD/CAM technology and CNC machining technology.