Aiming at the problem of easy cracking of nickel-based tungsten carbide metal composite ceramic coating with high tungsten carbide (WC) content prepared by laser cladding on the surface of nickel-based high-temperature alloy, the effects of different laser cladding scanning speeds (300, 500, 700 mm/min) and substrate preheating temperatures (room temperature, 300 ℃, 500 ℃) on the heating and cooling process of the molten pool were explored by using a self-assembled molten pool temperature detection device, and the cracking mechanism inside the coating was analyzed by combining the coating microstructure observation. The results show that both laser cladding parameters and coating materials will affect the cracking behavior of the coating. On the one hand, the heating and cooling rates in the molten pool during laser cladding are positively correlated with the scanning speed, but negatively correlated with the preheating temperature. Lower scanning speeds and higher preheating temperatures can obtain smaller temperature change rates, which are more conducive to inhibiting the overall cracking of the coating; on the other hand, due to the large differences in melting points and thermal expansion coefficients between WC and NiCrSiBC, WC is prone to cracking or dissociation when the molten pool cooling rate is too high or too low. Therefore, it is necessary to optimize the cooling rate of the molten pool in a targeted manner to effectively improve the overall forming quality of the coating.

Compared with traditional wear-resistant protection technologies such as plasma spraying, supersonic flame spraying, and tungsten arc cladding, the preparation of ceramic-metal composite high-wear-resistant coatings based on laser cladding additive manufacturing technology has attracted more and more attention. However, when using laser cladding technology to prepare ceramic-metal composite coatings such as NiCrSiBC-60% WC, it is easy to produce more serious cracking problems and it is difficult to solve. This is mainly because the molten substrate and powder in the laser cladding process will form a molten pool of liquid and solid mixed on the surface of the substrate, and there is a relatively drastic short-term rapid heating and rapid cooling transition inside. Due to the large difference in thermal expansion coefficient and melting point between the ceramic phase and the metal phase, local heating, rapid cooling or the presence of a large thermal gradient may cause drastic changes in residual stress, leading to the formation of cracks.
Some scholars have tried to add rare earth elements or oxides to ceramic-metal composites to improve the wettability of the metal matrix by reducing the surface tension. The results show that this method helps to refine the grains and increase the grain boundary area, thereby hindering the formation of defects such as cracks and pores. However, the addition of rare earth elements cannot solve the problem of residual stress caused by the large difference in thermophysical properties (i.e., the thermal expansion coefficient of the material) between the coating materials. Therefore, many research works are devoted to improving the cladding quality of the coating by optimizing the mass fraction of ceramic and metal matrix in the composite coating. Acker et al. found that a WC ratio of more than 30wt% will lead to the formation of cracks and clusters, causing the coating to become fragile. At the same time, some scholars studied the effect of laser cladding power on the microstructure of Fe-Si-B coatings and found that with the increase of power, the number and size of (Fe, Si) dendrites increased, and the hardness increased by 3.7 times relative to the substrate. Zhou et al. used an induction heater to heat the substrate in the range of 573 to 1173 K, effectively reducing the thermal gradient inside the molten pool, thereby inhibiting the generation and propagation of cracks. However, it is worth noting that when the substrate preheating temperature is too high, the WC particles may dissolve, resulting in a decrease in the hardness and wear performance of the coating. In addition to preheating, some scholars have tried to use nickel-coated WC particles to reduce the crack density in WC-based ceramic-metal composite coatings.

In summary, the generation of cracks in ceramic-metal composite coatings can be inhibited by controlling the heating and cooling processes during laser cladding. However, there are few studies on the specific relationship between the heating and cooling processes during laser cladding and the cracking of the coating. Therefore, this study aims to use an infrared pyrometer to monitor the heating and cooling changes of the molten pool and establish a qualitative relationship between the molten pool cooling rate, residual stress and cracks. On this basis, the thermal damage mechanism of WC particles and its effect on the mechanical properties of the coating were studied.

1 Experimental materials and methods

1. 1 Experimental materials
Figure 1 (a) shows the morphology of NiCrSiBC-60% WC powder (Metco 52052), which is a mixture of cast tungsten carbide and self-fluxing NiCrSiB-WC material. The powder particles are spherical and the size ranges from 45 to 106 μm. Figure 1 (b) shows the XRD phase analysis of the powder. It can be seen that the main components are γ-Ni, Cr2B, CrB and WC. Due to the trace amount of Si, no detection is made. A nickel-based high-temperature alloy (Inconel 718) with a size of 25 mm × 25 mm × 8 mm is used as the substrate. The chemical composition (mass fraction, %) of the Inconel 718 alloy substrate is 17.81Cr, 20.11Fe, 3.12Nb, 2.01Mo, 0.01Mn …. 32Al, 0. 89Ti, 0. 05Mn, 0. 31Si, 0. 005B, Ni balance. The chemical composition of NiCrSiBC-60% WC powder is shown in Table 1.

1. 2 Experiment and analysis
A 2 kW YB fiber laser (IPG photonics, YLR 2000) with a wavelength of 1. 07 μm was used as the energy source. The laser can operate in continuous wave and modulation modes. The cladding process uses a coaxial laser cladding head, which is integrated into the laser module through an optical fiber. The laser cladding head uses a convex lens with a focal length of 200 mm, and the laser system uses a laser beam with a wavelength of 16 mm. Table 2 shows the process parameters. The experiment is divided into two modes: no preheating and preheating. The substrate preheating temperature is kept below 500 °C and 3 layers of cladding are performed. However, during the actual laser cladding operation, the temperature of the substrate may rise to 650 °C or above due to laser scanning heating.
The substrate preheating uses a self-assembled substrate preheating temperature control device as shown in Figure 2. The specific operation process is: the substrate is placed on a stainless steel plate above the resistance heater, and its temperature is measured and calibrated according to the voltage change using a K-type thermocouple. An infrared pyrometer (Micro Epsilon, model: CTLM-2HCF3-C3H) is used to monitor the thermal change process of the molten pool during the deposition process. A notch filter with a spectral range of (1064 ± 25) nm is introduced into the optical system of the infrared pyrometer to block the laser radiation. The infrared pyrometer is calibrated before use to ensure the accuracy of the temperature measurement results. The pyrometer and the substrate remain stationary, and the laser cladding head with a coaxial powder feeding device is scanned on the substrate.

After laser cladding, the sample is cut using wire cutting to prepare the metallographic specimen. SEM (Zeiss, EVO 18 Research) and X-ray analyzer were used to analyze the microstructure, crack morphology and phase composition of the coating cross section. X-ray diffraction (PANalytical-Empyrean) was used to measure the residual stress. Before the test, the equipment was calibrated on a stress-free austenite standard sample with a Mn target. The diffraction angle 2θ was measured to be 152°, and the maximum error was no more than ± 14 MPa. Under normal circumstances, the error was within ± 6. 9 MPa, indicating that the instrument calibration was qualified. A 2 mm X-ray detection probe was selected, and the multi-exposure mode was set. The single exposure time was 10 s, the incident tilt angles were 0°, ± 24. 1°, ± 35. 3°, ± 45°, the target voltage was 30 kV, and the target current was 6. 7 mA.

2 Experimental results and discussion

2. 1 Effect of process parameters on the thermal change process of the molten pool
Figure 3 shows the temperature change curve of the molten pool at different substrate preheating temperatures when the scanning speed is 700 mm/min. It can be seen that the peak temperature of the molten pool decreases with the increase of the substrate preheating temperature, and the time to reach the peak temperature is delayed, and the cooling rate slows down. The cross-sectional morphology of the coating at different preheating temperatures is shown in Figure 4. With the increase of the preheating temperature, the overall cross-sectional area of ​​the laser cladding coating tends to increase.

The size of the typical area of ​​the cross section of the cladding layer changes with the scanning speed and preheating temperature as shown in Figure 5. As shown in Figure 5, the size of each typical area of ​​the cross section of the cladding layer increases with the increase of the substrate preheating temperature and decreases with the increase of the scanning speed. This is because at room temperature, the melt width will be smaller than the laser spot diameter due to the influence of the laser absorptivity on the substrate surface. With the increase of the preheating temperature, the laser absorptivity on the substrate surface tends to increase, thereby enhancing the absorbed laser power, thereby improving the melting effect of the substrate and the powder, resulting in a larger melt width. As the radial range increases, the melt height also increases, thereby improving the powder collection efficiency and increasing the overall volume of powder deposition.

Figure 6 shows the temperature change rate of the molten pool at different scanning speeds and preheating temperatures. As the scanning speed increases, the heating rate and cooling rate of the molten pool both tend to increase. This is mainly because during the interaction between the laser beam and the material, the rate of increase of the material temperature will gradually slow down with the increase of the action time, showing a change trend similar to a logarithmic function. Conversely, the cooling process is also the same. As the scanning speed increases, the interaction time between the material and the laser beam becomes shorter. Therefore, the heating rate and cooling rate of the material will also increase in a short time, thereby increasing the temperature gradient in the matrix, resulting in a higher cooling rate as shown in Figure 6 (b). However, as the preheating temperature increases, the temperature gradient in the matrix decreases, resulting in a decrease in the cooling rate.

2.2 Effect of process parameters and preheating temperature on crack mitigation
Figure 7 shows the cross-sectional morphology of the cladding samples at different scanning speeds and preheating temperatures. As shown in Figure 7 (a, d, g), at room temperature, as the scanning speed decreases from 700 mm/min to 300 mm/min, the cooling rate decreases from 8.3 × 103 °C/s to 3.8 × 103 °C/s. Despite the reduction in cooling rate, multiple cracks are still formed. This shows that controlling the thermal gradient and cooling rate in the molten pool alone may not be sufficient to prevent the formation of cracks. Similarly, as shown in Figure 7 (b), when the cladding is preheated to 300 °C and at a higher scanning speed (700 mm/min), more obvious cracks will also be generated. However, as shown in Figure 7 (e, h), when the preheating temperature is 300 °C, the coating does not produce cracks after reducing the scanning speed to 500 mm/min and 300 mm/min. As shown in Figure 7 (c, f, i), by increasing the preheating temperature to 500 °C, the temperature gradient between the substrate and the molten pool and within the molten pool is further reduced, so that no cracks are generated when preparing the coating at different scanning speeds.

Figure 8 (a-d) shows the high-magnification cross-sectional morphology of the coating at different scanning speeds and preheating temperatures. It can be clearly seen that most of the microcracks inside the coating penetrate the WC ceramic particles. This is because the expansion coefficient of WC is small ((6.5~7.4)×10-6K-1), which is almost half of that of NiCrSiBC coating ((13.3~16.8)×10-6K-1) and Inconel 718 alloy matrix ((12.8~14.2)×10-6K-1), while the melting point of WC (3100℃) is almost twice that of NiCrSiBC coating (1300℃) and Inconel718 alloy matrix (1600℃). Therefore, during the solidification process, the liquid metal around the WC particles begins to shrink, exerting high tensile force on the particles at high temperature, causing cracks to nucleate along the direction of high stress concentration points, that is, WC particles. This is very obvious in Figure 8(d). In addition, the temperature gradient between the substrate and the deposited material with a relatively low thermal expansion coefficient will generate internal stress and residual stress.

In addition to the microstructural analysis, the residual stress in the deposited layer was analyzed by X-ray diffraction. As can be seen from Figure 8(e), the internal residual stress of the coating changes with the substrate preheating temperature and the laser cladding scanning speed. At the same scanning speed, the residual stress is the largest when preheating at 300 °C, and the cladding residual stress is the smallest at room temperature. At the same preheating temperature, the residual stress value increases with the increase of the scanning speed. The stress value is the largest under the scanning speed of 700 mm/min and the preheating condition of 300 °C.

2.3 Effect of cooling rate on microstructure evolution
Figures 9 and 10 show the morphology of WC particles inside the coating at different scanning speeds and preheating temperatures and the XRD spectra of the coating, respectively. As shown in Figure 9 (a, b), without preheating, carbides produced by dissolution and interdiffusion will precipitate around the WC particles at scanning speeds of 700 and 300 mm/min. In addition, as the cooling rate decreases from 8.3 × 103 ℃/s to 3.8 × 103 ℃/s, these precipitates around the WC particles transform into coarse dendritic carbides.
Under preheating conditions, as shown in Figure 9 (c, d), at the same scanning rate (500 mm/min), as the preheating temperature decreases (500 ℃ to 300 ℃), it is monitored that the cooling rate is 1. 8 × 103 ℃ /s, at a scanning rate of 500 mm /min, the change of preheating temperature has no effect on the cooling rate of the coating. From the microstructure, it can be seen that the larger dendrite structure disappears, and the tungsten carbide particles gradually show grain refinement. The cooling rate is further reduced to study the transformation of the coating structure under different preheating temperatures at lower scanning rates. As shown in Figure 9 (e, f), at the same scanning speed (300 mm /min), as the preheating temperature decreases (500 ℃ to 300 ℃), the larger dendrite structure in the coating disappears. This is mainly because maintaining the same scanning speed and reducing the preheating temperature accelerates the solidification rate of the molten pool, which in turn leads to the refinement of the grains in the coating. As shown in Figure 10, as the cooling rate decreases, the content of these precipitates (MC) increases.

Figure 11 is a schematic diagram of the dissolution mechanism of WC particles in Ni-based WC coatings at different cooling rates. As shown in Figure 11(a), the shell-core structure is formed by the interdiffusion of carbide precipitates around WC. As mentioned in Section 3.1, the cooling rate decreases with the heating of the substrate. It has been shown that when WC particles are subjected to thermal shock or heating, stresses form nucleation cracks, leading to their fracture. As mentioned earlier, due to the high temperature gradient, the cooling rate is too fast, and these cracks tend to coalesce, resulting in through cracks from the interface to the surface.

On the other hand, if the cooling rate is slower, a longer melt pool life with good wettability can be provided, and the liquid matrix can fill these cracks and separate the broken WC particles. In the case of Figure 11(b), it can be clearly observed that the spherical WC particles dissociate into blocky angular WC particles together with the surrounding carbide precipitates. Further reduction in the cooling rate leads to the complete dissolution of the WC particles into the matrix, making the coating brittle, as shown in Figure 11(c). In addition, it can be observed that as the WC dissolves, the coating becomes brittle, residual stress is generated and microcracks are induced. In all coatings, a considerable number of WC particles always remain due to the high concentration or mass fraction of WC.

Conclusion
1) By varying the scanning speed, the cooling rate can be controlled. At room temperature, stresses due to the thermal gradient between the substrate and the deposited layer can lead to cracks in the cladding layer.
2) In the case of preheating the substrate, the thermal gradient can be controlled, resulting in crack-free multilayer cladding. The cooling rate decreases with decreasing scanning speed and increasing substrate preheating temperature.
3) Residual stress decreases with decreasing cooling rate. There is an optimal cooling rate. Above this rate, cracks dominate, and below this rate, WC particles dissociate, resulting in embrittlement of the coating.