In order to study the mechanism of the influence of the microstructure evolution of tungsten carbide (WC) in the composite coating on the crack generation, three kinds of Ni50A/ WC composite coatings were prepared by single-layer laser cladding, transition layer gradient cladding and double-layer cladding. The morphology and microstructure of the coatings, the characteristics and causes of crack generation were analyzed, and the influence of the microstructure evolution of WC on the crack generation was explored. The results show that the influence of WC microstructure evolution of different cladding methods on crack generation is mainly composed of the crack source formed by the internal cracking of residual WC particles and the segregation of components caused by hard phase elements. Compared with the single-layer cladding coating, the residual WC particle content of double-layer cladding and gradient cladding coatings is reduced by 32.7% and 37.9% due to the absorption of more energy by the powder, which reduces the internal crack source of the coating. The mass fraction of W element in the eutectic compound also decreases from 0.534 in the single-layer cladding coating to 0.417 in the double-layer cladding coating and 0.386 in the gradient cladding coating, which reduces the concentration of hard phase elements, reduces the segregation of coating components, and reduces the crack sensitivity of the coating. This study has certain guiding significance for improving the cracking problem of laser cladding composite coatings and improving the yield of composite coatings.

Metal-ceramic composite coatings have the advantages of high hardness, wear resistance and heat resistance. The preparation of composite coatings by laser cladding technology is an important research direction in the field of surface modification technology [1]. Tungsten carbide (WC) is a commonly used ceramic phase powder. Due to its excellent properties such as high hardness, high melting point, and small hardness drop at high temperature, it is often used as a coating reinforcement phase in laser cladding. However, the thermal physical properties of WC, such as linear expansion coefficient, elastic modulus, thermal conductivity and melting point, are quite different from those of the base metal. The coating is prone to produce large residual thermal stress, which increases the crack sensitivity of the coating [2]. Ni50A nickel-based self-fluxing alloy powder has the advantages of good wettability, heat and corrosion resistance, and high temperature oxidation resistance. It can be used as a bonding phase material for WC. At the same time, the coating prepared with Ni50A has high density and good bonding performance. It can be used as a transition layer between the ceramic material coating and the substrate to improve the bonding strength of the coating and reduce the crack sensitivity of the coating [3]. It is an important research direction to improve the crack defect problem of laser cladding composite coating by using transition layer gradient cladding and studying the mechanism of crack generation.

SHU et al. [4] synthesized a high volume fraction WC reinforced nickel-based composite coating in situ on a low-carbon steel substrate by gradient laser cladding, providing a new idea for preparing a composite coating without pores or cracks. ZHIKUN et al. [5] prepared Ni/WC composite coatings with different mass fractions (0~0.6) of WC particles on the surface of stainless steel. The coatings have high density and no pores or crack defects inside. LEE et al. [6] studied the cracking trend of Co-based WC+NiCr composite layer during laser cladding. The study showed that the tensile stress in the cladding layer can be used as a potential driving force, combined with crack initiation sites such as cracked WC particles, pores and solidification cracks, to provide an easy crack path for large brittle fracture. SHI et al. [7] prepared gradient composite coating on the surface of 20CrMnTi alloy steel by laser cladding method. Laser power, cladding scanning rate and powder flow rate were selected as influencing factors of orthogonal experiment. Process parameters were optimized by orthogonal experiment and variance analysis to obtain gradient composite coating with excellent performance without crack defects. MA et al. [8] prepared WC particle reinforced nickel-based composite coating on Q550 steel substrate by broadband laser cladding technology. The effect of broadband laser power adjustment on the growth behavior of eutectic structure was studied, which provided a certain theoretical basis for inhibiting crack defects caused by hard particles such as ceramics. WU [9] prepared Fe50A/WC coating by gradient laser cladding technology to repair the blade of garbage crusher. The study showed that the higher the WC mass fraction, the higher the crack sensitivity of the coating; and the use of transition layer can reduce the crack sensitivity and reduce the occurrence of crack defects. SONG et al. [10] used 316 stainless steel laser cladding to repair 304 stainless steel and used WC powder for surface alloying. The results showed that with the increase of WC particles, the hardness of the coating and the undissolved WC particles also increased, and the wear mechanism changed from adhesive wear to abrasive wear. ZHOU et al. [11] used laser cladding and laser induction cladding to prepare nickel-based WC resurrection coatings. The results showed that laser induction cladding can improve the uniform distribution of WC particles. LÜ et al. [12] used laser cladding to prepare titanium-nickel composite coatings and studied the effects of adding different contents of rare earth Ta. The results showed that increasing the content of rare earth Ta can improve the wear resistance and oxidation resistance of the coating. YOU et al. [13] used laser cladding technology to in-situ synthesize Ti-Al-N composite coatings on the surface of TC4 titanium alloy and studied the in-situ synthesis mechanism of the composite coating. The results showed that adding a small amount of Al powder can promote the reaction of TiN and TiAl in the molten pool, thereby significantly increasing the content of Ti2AlN MAX phase in the coating.

In this paper, the authors compared the single-layer Ni50A/WC composite coatings, Ni50A/WC composite coatings with Ni50A transition layer, and double-layer Ni50A/WC composite coatings prepared by different cladding methods under the same linear energy density process conditions, studied the cracking characteristics of the coatings, and analyzed the mechanism of WC affecting the crack generation, which provides certain theoretical guidance for better solving the cracking problem of laser cladding metal ceramic composite coatings, and has certain significance for solving the problem of cracks in laser cladding composite coatings.

1 Experimental materials and methods

1. 1 Experimental materials

The laser cladding experimental substrate is H13 steel (hot work die steel 4Cr5MoSiV1), the substrate size is 90 mm×45 mm×11 mm, and the chemical composition (mass fraction) is shown in Table 1. The cladding layer powder material is Ni50A nickel-based self-fluxing alloy powder, and the metal ceramic composite powder with a mass fraction of 0.1 of cast WC powder is mixed on the basis of Ni50A. The main chemical composition of Ni50A material and the main chemical composition of WC material are shown in Table 2 and Table 3. The scanning electron microscope (SEM) of the powder morphology before and after mixing is shown in Figure 1. The powder particle size is 48 μm ~ 106 μm. Before the laser cladding experiment, the samples were coarsely ground by a grinder to remove the surface oxide, and the surface oil was ultrasonically cleaned with anhydrous ethanol. The samples and cladding powder were dried in a vacuum drying oven for 2 h before laser cladding to remove moisture.

1.2 Coating preparation

The laser cladding equipment consists of a 1.4 kW fiber-coupled semiconductor laser, a lateral synchronous powder feeding device, a cooling water circulation system, an argon protective gas circuit and a powder feeding gas circuit, and an xy-axis mobile platform built by the research team. The protective gas pressure during the cladding process is 0.1 MPa. The process parameters of laser cladding are as follows: laser power is 817 W, scanning rate is 1.5 mm/s, and defocus is -10 mm.

Sample 1 uses a single layer of cladding, and the powder feeding rate of Ni50A/WC is 6 g/min; Sample 2 uses Ni50A powder as a transition layer, and the powder feeding rates of the first transition layer and the second layer are both 3 g/min, in order to control the coating thickness of each sample to be close; Sample 3 directly uses Ni50A/WC as an intermediate layer, that is, Ni50A/WC is directly divided into two layers for cladding, and the powder feeding rate of each layer is 3 g/min.

1.3 Material characterization and performance test method

The oxide layer on the surface of the sample coating is cleaned by a wire brush and then cleaned with anhydrous ethanol and other detergents. The crack defects of the sample coating are detected using Lifetime610 X-ray diffraction micro-computed tomography (CT). Select the stable and uniform part in the middle of the cladding layer, cut a 15 mm × 15 mm × 13 mm sample along the cross section of the cladding layer by wire cutting, then polish the cross section step by step with metallographic sandpaper, and finally use polishing liquid to mechanically polish the sample until the cross section reaches a mirror effect. Use a metallographic etchant of aqua regia prepared in a ratio of 3:1 with concentrated hydrochloric acid and concentrated nitric acid, and the etching time at room temperature is 15 s~20 s. After the etching is completed, rinse it clean, blow dry it, and use OLS4000 laser confocal microscope to observe the microstructure. The S3400N tungsten filament SEM and energy dispersive spectroscope (EDS) were used to analyze the cladding layer structure; the D/MAX-RB 12 KW X-ray diffractometer (XRD) was used to analyze the phase of the cladding layer, and the HVS-1000Z microhardness tester was used to analyze the microhardness of the coating-heat affected zone-substrate, with a load of 200 g and a loading time of 15 s.

2 Results and analysis

2.1 Analysis of the influence of laser cladding methods on coating morphology and structure

The thickness of the coatings prepared by the three laser cladding methods is close to 2.1 mm, and the surface morphology is shown in Figure 2. Comparing the surface morphology of the coatings of sample 1, sample 2 and sample 3, it can be seen that the surface of the coatings prepared by gradient cladding and double-layer cladding is very flat and smooth, while the surface of the coating prepared by single-layer cladding is uneven, and there are particles that are not fully melted and diffused into the molten pool. Since argon gas is used as the protective gas in the cladding process, and the powder contains Si and B elements with good deoxidation and slag-making properties, none of the three coatings is oxidized. Analyzing the surface morphology of the three coatings, the single-layer cladding coating has a short existence time in the molten pool during the cladding process, resulting in insufficient expansion, and even some powders are condensed into larger particles after being irradiated by laser and then do not expand and diffuse into the molten pool. This phenomenon is a direct reflection of the insufficient molten pool temperature, indicating that the powder fails to absorb enough energy, resulting in an uneven coating surface and particles that are not melted and diffused into the molten pool. However, the powder feeding rate of the gradient cladding and double-layer cladding coatings is half of that of the single-layer cladding. Therefore, when the laser power, scanning rate, and defocus amount are the same, that is, the laser specific energy is the same, the powder absorbs more energy and the molten pool is fully expanded, so the coating surface is smooth and flat.

Figures 3 and 4 are enlarged views of the interfaces of the three Ni50A/WC composite coatings. As shown in the figure, the three composite coatings first form a plane crystal with a thickness of 10 μm at the lowest point. This is because the temperature gradient G at the solid-liquid interface is large when the molten pool is formed, the crystallization rate R is small, and the G/R ratio is large, so the plane crystal is formed [14]. Along the interface toward the coating surface, the temperature gradient G gradually decreases, the crystallization rate R gradually increases, and the organizational growth mode on the plane crystal changes from cellular crystal to dendrite and equiaxed crystal. WC particles can be seen at the bottom of the single-layer cladding and double-layer cladding coatings. This phenomenon shows that during the laser cladding process, the WC particles in the mixed powder are not completely dissolved in the molten pool, but are distributed in the coating in the form of residual particles. However, this situation does not appear at the bottom of the gradient cladding coating, indicating that the undissolved residual WC particles in the second layer (Ni50A/WC layer) have not entered the first layer (Ni50A layer). Analyzing the bonding between the coating and the substrate, both interface magnifications show that the interface between the coating and the substrate is dense, indicating that a good metallurgical bonding is formed between the coating and the substrate.

2.2 Characteristics and causes of coating cracks

The crack detection of X-ray diffraction micro-CT is shown in Figure 5. Figure 6 is a SEM cross-sectional view of the coating. Combined with the observation of the gap width of the coating cracks, the characteristics of the cracks in the three coatings can be summarized. Cracks appeared in all three coatings, but judging from the number of cracks, total length and crack gap width, the crack length of the coating prepared by single-layer cladding is the longest among the three, and there are two coarse cracks spanning multiple cladding layers, so the cracking is the most serious. The coatings prepared by gradient cladding and double-layer cladding have only one crack, and the crack gap width is similar and smaller than the crack of the single-layer cladding coating, so the degree of cracking is relatively light. In addition, multiple cracks occurred in the lower part of the single-layer cladding coating, and the most serious crack extended from the interface between the coating and the substrate to the coating surface. There were also some cracks at the interface between the coating and the substrate. Although the coatings with transition layer and double-layer cladding also had cracks, there was no cross-over of multiple cracks in the single-layer cladding coating. In addition, cracks appeared in the substrate and the interface between the substrate and the coating of the double-layer cladding. The comprehensive level of cracking of the double-layer cladding coating on the surface was more serious than that of the gradient cladding.

The distribution and state of ceramics in the composite coating have an important influence on the cracking of the coating. In order to study the cracking causes of the three composite coatings, the distribution characteristics of residual WC particles in the three coatings were analyzed. As shown in Figure 6, there are incompletely dissolved WC ceramic particles, that is, residual WC particles, in all three coatings. The distribution of the residual WC particles in the coating has related characteristics and similarities and differences. In Figure 6a, the residual WC particles are mainly distributed in the middle and lower part of the single-layer cladding coating, and the closer to the bottom, the more WC particles there are; in Figure 6b, the residual WC particles are mainly distributed in the middle of the coating, that is, the upper part of the Ni50A transition layer and the bottom of the Ni50A/WC layer, and the closer to the bottom of the Ni50A/WC coating, the more WC particles there are; in Figure 6c, the residual WC particles are mainly concentrated in the middle and bottom of the coating, that is, the bottom of the second layer Ni50A/WC and the bottom of the first layer Ni50A/WC, among which the second layer has more WC particles and the first layer has fewer. The distribution forms of the residual WC particles of the three coatings are different, but they all have the same characteristic, that is, the residual WC particles will gather at the bottom of the Ni50A/WC layer. Therefore, the distribution of residual WC particles in the three coatings is due to the coagulation effect of WC [4], which is manifested in single-layer cladding, gradient cladding, and double-layer cladding. Combining the cracking characteristics of the three coatings with the distribution characteristics of residual WC particles, the results show that the area where residual WC particles gather coincides with the area where the coating cracks severely. The dissolution and precipitation state of ceramic particles in the molten pool affects the number of residual WC particles, and the number of residual WC particles affects the cracking of the coating. Through image processing, the residual WC particles in the coating area of ​​the SEM cross-sectional images of the three coatings with a resolution of 1180×700 were image recognized. The number of residual WC particles in the three coatings was reflected according to the area ratio of the residual WC particles in the resolution area. The results showed that the residual WC particles in the single-layer cladding coating had the highest area ratio, reaching 2.11%, and the double-layer cladding and gradient cladding coatings were 1.42% and 1.31%, respectively. Compared with the single-layer cladding coating, the residual WC particle content was reduced by 32.7% and 37.9%. It can be seen that double-layer cladding and gradient cladding can effectively reduce the residual WC particle content and promote the dissolution of ceramic particles in the molten pool.

The causes of cracks in the coatings of the three samples were analyzed. Laser cladding is a process of rapid heating and sudden cooling, so the coating will have a large residual thermal stress due to shrinkage, which is manifested as tensile stress. Adding WC ceramic particles will increase the difference in physical properties such as linear expansion coefficient and melting point between the substrate and the coating, which will further increase the residual thermal stress, so the coating has a higher cracking sensitivity. On the one hand, the more residual WC particles there are, the more they gather in the same area, and the ceramic hard particles will cause the internal stress concentration problem of the coating in this area to be more serious, so the crack defects of the coating are more serious. On the other hand, the closer the residual WC particles are to the H13 steel substrate, the greater the difference in thermophysical properties between the two than the nickel-based components of the coating, the greater the residual thermal stress, the more serious the cracking of the coating, and the interface cracks shown in Figure 6a appear. Since double-layer cladding and gradient cladding powder absorb more energy than single-layer cladding, the molten pool temperature is higher and the existence time is longer, and the dissolution of WC ceramic particles in the molten pool is more complete, so the residual WC particle content of the coating is effectively reduced. In addition, the transition effect of the Ni50A layer makes the cracking degree lighter, and the cracks are mainly in the Ni50A/WC layer. Although double-layer cladding also absorbs more energy than single-layer cladding, due to the lack of transition layer, the WC in the first Ni50A/WC layer precipitates at the bottom of the coating, which also causes residual WC particles to approach the H13 steel substrate. Higher energy will also cause more W elements to diffuse into the heat-affected zone of the substrate, increasing the hardness and brittleness, and eventually causing interface cracks and heat-affected zone cracks.

2.3 Analysis of the organizational evolution of WC and its influence on the mechanism of coating crack generation

During the laser cladding process, the cast WC particles will dissolve in the high-temperature molten pool, gradually become free atoms of tungsten (W) and carbon (C) diffuse in the high-temperature metal liquid, and finally precipitate and form complex compounds. Due to the short existence time of the molten pool, the dissolution process of WC is not completely sufficient, so residual WC particles still remain in the coating. The dissolution of residual WC particles and the formation of compound precipitation are shown in Figure 7. The edge of the particle is left with a dissolved phenomenon, and more compound precipitation is generated at the boundary and its vicinity.

Previous studies have shown that the dissolution of WC particles and the precipitation of complex compounds determine the microstructure evolution of WC in the molten pool [8]. Figure 8 shows the compound precipitation in the area 0.7 mm away from the substrate inside the single-layer cladding coating, indicating that there are also more compound precipitation in the area where the residual WC particles are gathered. The elemental composition was analyzed by EDS, and Figure 9 is the EDS surface scan of the element distribution of the compound precipitation. The scanning results show that the compound precipitation is mainly enriched in Cr and W, among which Cr is more obvious, while the compound precipitation is mainly enriched in Ni and Fe elements. In order to analyze the element distribution of the compound more clearly, EDS point scanning is performed on it. Figure 10 is the EDS point scan of the compound precipitation area. In the figure, the first sampling point is selected on the compound precipitation, and the second sampling point is selected in the area outside the precipitation. The results show that a large amount of Cr and W elements are detected in the compound precipitation, and a large amount of Ni elements are detected outside the compound precipitation. The distribution of the elements shows that the compound precipitation concentrates the main hard elements Cr and W of the coating, while the basic component elements Ni and Fe of the coating are enriched outside the compound. Therefore, the complex compounds precipitated in the coating exist as hard compound precipitation.

The coating is analyzed by XRD phase analysis, and Figure 11 is the XRD peak diagram of the coating. In the figure, three very obvious characteristic peaks appear at 44.28°, 51.54° and 75.64°, respectively. The results show that the Ni50A/WC metal ceramic composite coating is mainly composed of γ-Ni(Fe) solid solution, Cr hard phase, and W hard phase. The γ-Ni(Fe) solid solution accounts for the vast majority of the coating, mainly as the toughness phase of the coating, followed by the Cr hard phase, and finally the W hard phase. Cr mainly exists in the form of Cr23C6 and Cr7C3 carbides, with a small amount of CrB boride. The Cr element promotes the solid solution strengthening of the γ-Ni(Fe) phase, and other solid solutions are generated at the junction of the cladding layer and the substrate, which improves the high temperature resistance of the repaired sample[15]. Previous studies have shown that M23C6 carbides are complex cubic crystals, in which M can be Cr, Fe or W[16-18]. At the same time, the research of Wang et al.[19] shows that M23C6 carbides are mainly eutectic structures. During the evolution of the structure of WC, WC and Ni50A powder reacted in a complex manner, in which W mainly existed in the form of carbides of WC and W2C, and a small amount of boride W2B5, and a eutectic mixture could be formed between WC and W2C, indicating that after the evolution of the structure, WC finally existed in the form of a eutectic mixture of W hard phase in the precipitated compounds. Therefore, the compound precipitated in the coating is actually a eutectic compound composed of Cr and W carbide hard phases. The eutectic compound has a high hardness, but because the hard phase is concentrated, it will cause the coating to segregate, thereby increasing the cracking sensitivity. Table 4 is a comparison of the hard phase element content of the eutectic compounds in the three coatings by EDS. Compared with the double-layer cladding and gradient cladding and single-layer cladding, the W and Cr content of the eutectic compounds are reduced, and the mass fraction of the W element decreases from 0.534 to 0.417 and 0.386, respectively, which reduces the concentration of hard phase elements and reduces the composition segregation of the coating. In addition, this composition segregation generally occurs in the area where the residual WC particles are gathered, because there are more eutectic compounds generated in the residual WC particle area, resulting in the most serious cracking in the area where the residual WC particles are gathered in all three coatings.

After the WC particles undergo the organizational evolution of dissolution, a large number of residual WC particles are still left. Figure 12a is a residual WC particle diagram of the single-layer cladding coating. It can be seen that cracks appear inside the residual WC particles in the coating. After multiple sampling observations, the cracking of the internal residual WC particles is a common phenomenon. This phenomenon shows that the residual WC particles that are not fully dissolved have large residual thermal stresses inside the particles themselves due to the large difference in thermophysical properties from the nickel-based self-fluxing alloy coating, and the rapid heating and cooling process of laser cladding. In addition, the residual WC hard particles are also stress concentration points in the coating, so cracks are easy to appear inside the residual WC particles. As shown in Figure 12b and Figure 12c, hard particles with crack defects inside are also easy to become crack sources, providing paths for cracks in the coating, causing cracks in the coating to spread and aggravate the severity of cracks in the coating.

The evolution of WC organization is also reflected in the microhardness of the coating. Figure 13 shows the microhardness distribution of the three coatings. It can be seen that the single-layer cladding coating has microhardness peaks at a depth of about 1 mm to 1.5 mm from the coating surface (the middle and lower layer area of ​​the Ni50A/WC single layer), about 1 mm (the second layer, i.e. the bottom of the Ni50A/WC layer) and about 2 mm (the area at the substrate interface) of the sample 2/3. The area where the peak is located, the area where the residual WC particles gather, and the area where the coating cracks are serious are all consistent. This phenomenon shows that the deposited WC and its evolution products enhance the coating hardness in this area [4]. Therefore, the coagulation effect of WC causes the residual WC particles to gather in certain areas according to the coagulation effect, and more eutectic compounds composed of carbides of W and Cr hard phases are precipitated in this area. The eutectic compounds and residual WC particles increase the crack sensitivity of the coating and also increase the hardness of the coating in this area.

3 Conclusions

Ni50A/WC composite coatings of the same thickness were prepared by single-layer laser cladding, Ni50A transition layer gradient cladding, and double-layer cladding under the same laser process parameters. The morphology, organization, and crack characteristics of the three coatings were analyzed, and the organizational evolution of WC and its influence on the crack generation mechanism were explored.

(1) Due to insufficient energy density, the surface morphology of the coating was uneven and the wettability was relatively poor in single-layer cladding. However, due to the reduction of single-layer powder amount, gradient cladding and double-layer cladding obtained more energy under the same laser process parameters, so the coating surface was smoother and flatter, and the wettability was better.

(2) The composite coating prepared by single-layer cladding had the most serious cracking among the three. The crack generation characteristics of the coating were mainly affected by the number and distribution of residual WC particles. The more residual WC particles there are, the more crack sources there are; the more concentrated the distribution is, the more serious the stress concentration is; the distribution of residual WC particles in the three coatings is mainly the manifestation of the WC coagulation effect in different cladding methods. Double-layer cladding and gradient cladding powder absorb more energy, WC dissolves more fully, and the residual WC particle area ratio is reduced by 32.7% and 37.9% respectively compared with single-layer cladding, and the disadvantage of concentrated distribution of single-layer cladding is also improved.

(3) The main components of the three coatings are γ-Ni (Fe) solid solution, Cr and W hard phases. After WC particles are dissolved in the molten pool during the cladding process, eutectic compounds mainly composed of carbides WC and W2C are precipitated. These eutectic compounds have high hardness and are easy to aggregate with Cr phase to form hard phase precipitation, which significantly improves the hardness of the coating. The maximum hardness of the coating reaches 850 HV, but it also causes component segregation, improves the hardness and brittleness of the coating, and increases the cracking sensitivity of the coating. The W element content and Cr element content of the eutectic compounds in the double-layer cladding and gradient cladding coatings are reduced, and the mass fraction of the W element decreases from 0.534 to 0.417 and 0.386, respectively, which reduces the concentration of hard phase elements and reduces the component segregation of the coating.