Preparation of high wear resistance nickel based WC coating by carefully adjusting interface structure

In recent years, many scholars have paid attention to wear-resistant coatings for shield machine cutterheads due to their very high consumption rates. Among these coatings, nickel-based tungsten carbide (Ni-based WC) is one of the best, showing both corrosion resistance and wear resistance. However, to further improve the wear resistance of such coatings, there are still numerous issues that need to be resolved. Herein, a new method, distinct from conventional methods, is presented. Specifically, the brittle phase W2C is not widely regarded as the main wear-resistant phase, but we were surprised to find that careful adjustment of its rigid structure can yield satisfactory results. Experimental results and first-principles simulations have indicated that the friction coefficient and weight loss of a coating with a suitable distribution of W2C are only half of those of a traditional Ni-based WC coating (about five times higher than those of the substrate), which can mainly be attributed to the excellent thermal expansion coefficient and hardness of the W2C phase. As we expected, the surface morphology of the material after wear revealed that the suitable W2C layer has a well-defined friction morphology. We hope to provide new ideas for the study of Ni-based WC coatings in shield machine cutterheads.


Introduction
In recent years with the rapid development of urban rail transit, people have put forward an urgent demand for wear-resistant coating of shield machine cutterhead [1][2][3]. Among these coatings, nickel-based tungsten carbide coating (Ni-based WC) has attracted much interest for its excellent wear resistance, service life, and other properties [4]. However, the performance of Ni-based WC coating can not fully meet the current actual demand, due to the interface between WC particles and Ni-base alloy being prone to stress concentration, which would lead to coating crack extension, peeling, and a series of problems [5,6]. Therefore, it is crucial to develop a high wear resistance and high stability of Ni-based WC coating.
In order to solve those dilemmas, some researchers have done some work on preparation processes. For instance, by laser hot-wire deposition, high volume fraction of the reinforcements with ex situ eutectoidstructured WC/W 2 C particles can be obtained, thus the coating hardness and wear resistance are improved to about 3.5 times and 4.5 times more than that of substrate [7]. Instead, using cold spraying technology to prepare Ni-WC composite coating, can effectively avoid the WC decarbonization phase (W 2 C phase), which contributes to the formation of WC coating with higher filling rate [8]. According to the above explanations, the true modification mechanism of Ni-based WC coating is not fully understood.
Recent mechanistic studies of wear resistance of intermediate transition phase W 2 C have shown that W 2 C seems to be a potential for interface adjustment. Classically, Tillmann W et al carried out high-speed arc spraying Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.
of WC-W2C iron-based coatings results in WC-W2C particle-enhanced coating properties through a dense microstructure, good bond strength to the substrate, and the smooth surface [9][10][11][12]. Meanwhile, Iizuka et al [13] analyzed the data from the Ni-based WC coating composite interface and concluded that the reaction between tungsten and carbon sources resulted in the growth of nano-sized W 2 C particles in situ, thus reinforcing the composite interface. Hence, the ultimate wear resistance of wear-resistant coatings is enhanced by the introduction of W 2 C particles resulting in the improvement of coating properties, which has also been confirmed in coating applications. However, although these endeavors, these studies are limited to the interrelationship between interfacial tuning and wear resistance, while theoretical predictions and computational simulations are still not clearly available for mechanical understanding [14][15][16].
Therefore, Therefore, this paper discusses the thermodynamic mechanism of the reaction between tungsten and carbon in tungsten carbide and calculates the microstructure of the primary phase interface by advanced first principles. It was demonstrated that the main phase with soft secondary phase reinforcement was obtained through thermomechanical control, thereby upgrading the wear resistance and stability of the nickel-based tungsten carbide coating and addressing the problem of spalling and cracking of the shield blade coating to extend its service life and reduce costs. More importantly, the slow release mechanism of intermediate transition relative coating interface was found out which can put forward new theoretical guidance for the research of wear-resistant materials.

Materials
In the present study, 304 L stainless steel (100 mm × 100 mm × 5 mm, Alibaba, Shanghai) was used as the substrate, and its chemical composition is shown in table 1. The coating material is a Ni-WC powder (40 wt% Ni, 60 wt% WC, Alibaba, Shanghai) with a particle size of 100-150 mesh.

Plasma surfacing process
The substrate (304 L stainless steel) was first subjected to high-energy shot peening derusting, ultrasonic cleaning with anhydrous ethanol, and air-drying treatment [17]. Before coating, the substrate was annealed at 350°C for 2 h to eliminate any residual stress therein [17]. Plasma coating was carried out using DML-V03BD and BM09DF three-way machine tools, the operational parameters of which are listed in table 2. Considering the melting points of WC and Ni powder, repeated experimental welding gave coatings with a good metallurgical bond above welding currents of 140 A; thus, three welding currents of 140 A, 160 A, and 180 A were selected for experiments under otherwise identical conditions. After coating, the coated material was cut into equal 10 mm × 10 mm squares using a wire-cutting machine, which were easy to observe utilizing a scanning electron microscope (SEM). Abrasive papers of 400#, 600#, 800#, 1000#, 1500#, and 2000# were then sequentially applied for preliminary grinding and fine grinding [13].

Characterization
Surface microstructures of coating layers were observed with a Hitachi 4800 SEM, which was equipped with energy-dispersive spectroscopy (EDS) attachment to analyze the elemental compositions and distributions. A Tecnai G2 20S-TWIN transmission electron microscope (TEM), operated at a working voltage of 160 kV, was used to observe the microstructures of the samples. A D/max-2200 x-ray diffractometer (XRD) was used to determine the phase composition of the coating layer, employing a Cu-K α x-ray source (1486.6 eV, λ = 0.15406 nm), a working potential of 30 kV, a current of 30 mA, a step size of 0.5°, and a scanning speed of 4°·min −1 over the range 10°-90° [14,15]. To further investigate elemental valence states on the surface of the coating layer, x-ray photoelectron spectroscopy (XPS) on a PH I5000 instrument was used (Al x-ray source). The effect of interfacial reaction on friction resistance was investigated using HSC Chemistry software. The friction and wear properties of the coating were tested with an MFW-02 reciprocating friction and wear testing machine, employing a load of 60 N, a test time of 20 min, and a running speed of 200 rad min −1 . Three-dimensional surface profiles of the coating layer were obtained using an atomic force microscope (AFM, Nanoscope V, MultiMode ® 8, Bruker).

First-principle calculation
All calculations were performed using the first-principles total energy program CASTEP [16]. Here, the plane wave pseudopotential describes the electron wave function, and the generalized gradient approximation (GGA-PBE) is used to deal with the exchange-correlation function. The cut-off energy is 400 eV. The convergence precision of energy is 10 −4 eV cell, and the convergence precision of force is -0.02 eV A −1 , so as to build the interface structure for calculation. A vacuum model with periodicity should be built when building the crystal plane model. Brillouin zone sampling was carried out using Monkhorst-Pack k-point meshes [17]. The structures were optimized by the conjugate-gradient algorithm method. To further verify the mechanical stability of these polymorphs, the elastic constants are calculated with the strain-stress method [18]. Besides, the bulk modulus B and the shear modulus G are estimated via the Voigt-Reuss-Hill approximation by using the obtained elastic constants C ij . Young modulus and Poisson ratio are calculated in accordance with the formulas Y and v [19]. The theoretical Vicker's hardness is estimated by the model. The details of convergence tests have been described elsewhere [20].

Results and discussion
3.1. Thermodynamic analysis of interfacial reactions As shown in figure 1, WC and W 2 C belong to the cubic and orthogonal crystal system respectively with space group of Pm3m and Pnnm [21]. In this work, Ni-based WC interface structure and crystal structure are established and can be observed in figures 1(a) and (b). The coefficients of thermal expansion of Ni, WC, and W 2 C crystals calculated by first principles are given in figure 1(c) [22]. When the temperature exceeds 200 K, the average thermal expansion coefficient of the W 2 C phase, Ni and WC particles is 22.312 × K 10 , Ni particles, the coating is easy to peel off, W2C has better high-temperature heat resistance and is suitable as an interfacial transition phase, where stress-relieving effects exist to increase the bonding between Ni and WC particles. The coefficient of linear expansion of W 2 C is relatively flat and the value is not large. For WC/Ni composites, the generation of interfacial phase W 2 C helps to regulate the difference of thermophysical properties between Ni and reinforcing particles WC, thus effectively mitigating the stress concentration at the junction of Ni and reinforcing particles, which leads to particle detachment [23,24]. It is apparent that W 2 C is dispersed in the matrix as an intermediate phase.
In general, to clearly understand the mechanism of the formation of the second phase [13], at the high temperature of a plasma arc, the main possible metallurgical reactions in the W-C system to be considered are as follows [25]: HSC chemistry analysis software was used to calculate the Gibbs free energies associated with reactions according to equations (1)-(4) at different temperatures, and the results are shown in figure 1(d). Apparently, in the plasma arc coating process of W-C systems, the reactions according to equations (1) and (4) are negative in a certain temperature range. Consequently, the formation and decomposition of W2C is thermodynamically possible. We know that the ease with which a chemical reaction can occur depends mainly on the reaction Gibbs energy ΔG r the degree of negative. Therefore, it can be seen from the thermodynamic analysis that equations (2) and (3) would not proceed in the molten liquid pool [26]. As can be seen from figure 1(d), when the temperature exceeds 2615 K, the free energy associated with 2WC=W 2 C+C is lower, so WC particles in the coating will preferentially melt and decompose into the hard W 2 C phase and free C at high temperatures. It can also be seen that when the temperature is lower than 1690 K, the stability of W 2 C in the coating decreases, and W 2 C begins to transform into WC and W phases. This means that at high temperature, the tungsten carbide particles melt at high temperature and undergo metallurgical reaction, and decompose into W 2 C and C [27]. In summary, WC particles melt at high temperature and undergo metallurgical reaction to decompose into W 2 C and C.
Using the Voigt-Reuss-Hill approximation, the three-dimensional distribution of WC and W 2 C elasticity performance was evaluated by ElasticPOST code [28]. Figure 2 shows the mechanical properties derived from the elastic constants in different directions, the 3D modulus surfaces of WC and W 2 C deviate from the sphere, indicating that the Young's modulus is anisotropic, further predicting the microcracking [29]. Figures 2(a) and (c) shows that the Young's modulus of WC and W 2 C are 570.56 GPa and 430.71 GPa, respectively, and the theoretical predicted hardness values of 34.82 GPa and 16.77 GPa in figures 2(b) and (d). It is noteworthy that the [0 0 1] crystal hardness of WC is significantly higher than that of W 2 C, while the opposite is true for W 2 C, which is more favorable for interfacial bonding and stable interfacial phase formation due to the cluster dragging effect and the ability of W 2 C to maintain good wettability with WC.

Microstructure and phase analysis
In previous work by Weng et al it has been shown that the coatings prepared by Ni-60% WC are dense structures and that they are continuous and uniformly distributed [30]. Hence, the WC particle distribution was controlled by adjusting the welding process parameters. The SEM morphology of the surfacing coatings with different welding currents is shown in figure 3, which presents the grayish-white central area of WC particles, the silverwhite area of WC melting and diffusion, and the outermost dark gray area of Ni-based molten pool. [14] In addition, it can be observed that the W2C particles of the coating as needle-like or dendritic crystals have edges and corners that are present in the nickel-based molten pool as secondary phases. There are different degrees of melting and diffusion phenomena, and the diffusion direction is from the white central area to the surrounding silver-gray area. The sizes of the melting and diffusion areas increase with the increasing welding current. Around the WC particles, there is a small amount of an acicular phase [31], the needle length and content of which depend on the welding current [32]. More significantly, the needle-like phase as stable is a symbol of favorable chemical and metallurgical bonding between the substrate and the coating [33]. The unmelted WC particles inhibit grain growth, thereby achieving grain refinement to form the needle-like phase, and the hard phase of the interlaced needle-like organization is distributed in the coating, which can greatly enhance the hardness and wear resistance of the coating [34][35][36].
As illustrated in figure 3(a), when the welding current is 140 A, it can be seen some WC particles melted and decomposed into whisker-like dendrites at high temperature. The inner layer structure of this area is uniform and dense, while the outer layer is a large number of Ni-based molten pool areas, which scattered with a small number of broken carbide particles and a large number of pores and cracks. When the current is increased to 160 A, the microstructure of the coating in figure 3(b) changes greatly. Not only does the density and homogenization of the Ni-based pool area increase to a great extent, but the length and number of the needlelike microstructure also increase to a great extent. And it is distributed around the WC particles in a crosssuperimposed way, forming a good dispersion and strengthening phase. When the current goes up to 180 A, it can be observed that the length and number of acicular structures are greatly reduced, and short fibrous  Figure 3(d) illustrates the surface of the surfacing coating was analyzed by x-ray diffractometer. As can be seen, the surfacing layer is mainly composed of Fe 0.64 Ni 0.36 phase, WC, W 2 C, and free state C. With increasing welding current, the content of WC in the surface layer varied accordingly, suggesting an effect on the relative stabilities of WC and W 2 C. The results may be interpreted in terms of WC particles melting and decomposing to form W 2 C at high temperature, forming a thick molten and diffusion zone around them, which may suggest the formation of a new phase and dissolution of W in the Ni matrix, in line with the results of Myalska et al [37].
At a welding current of 140 A, the negative temperature gradient is larger and the cooling rate of the molten pool is faster. Only a small amount of W 2 C is transformed into WC in the cooling process. In addition, the interface binding force between the matrix and WC particles will decrease with increasing WC particle content [38,39]. At a welding current of 160 A, more WC decomposes into W 2 C at high temperature. As the negative temperature gradient is smaller and the cooling rate is slower, more W 2 C is transformed into WC upon cooling, and the WC transition layer is the result of the selective dissolution of W 2 C. At a high welding current of 180 A, more WC particles decompose to W 2 C. Upon cooling, only part of the W 2 C phase reverted to WC, so overall W 2 C remained dominant, consistent with the findings in a previous study [40].
To further determine more clearly the diffusion and microstructure of WC particles and Ni-based melt pool elements in the overlay coating, figure 4(a) shows an EDS line scan after the WC particles. The coating is mainly composed of W, Fe, Ni, and C elements, with the W 2 C phase generated 10-40 nm away from the Ni particles. Figure 4(b) shows the different degrees of dissolution diffusion in the crystalline surfaces of WC crystals (1 0 0) in clusters. Combined with figures 1, 3, and 4 the existence of W2C particles can be successfully speculated. Since the melting point of WC is 2870°C and the melting point of Ni is 1453°C, during the overlaying process, the alloy powder melts on the surface of the tungsten carbide particles when it enters the molten pool, and the interior is not melted. Accordingly, as the temperature drops with the solidification process, respectively, around the carbide after diffusion, a thicker layer of melt diffusion zone is produced around the tungsten carbide particles [41]. It was also noted that tungsten and carbon in the acicular carbide diffused into the melt pool, and nickel and iron in the melt pool diffused into the acicular carbide The amount of tungsten and carbon in the base metal decreases gradually with increasing distance from the carbide, with localized high carbon peaks, while the amount of nickel and iron decreases with decreasing distance from the needle carbide [42].   Figure 5(a) confirms that the main surface elements were Ni, W, and C. The W4f pattern in figure 5(c) consists of two pairs of overlapping peaks, from which binding energies of BE Wf5/2 =33.997 eV and BE Wf7/2 =32.597 eV can be derived. These peaks show incomplete symmetrical fittings, thus demonstrating the presence of the W 4+ valence state in the coating. When the welding current was 160 A, the carbide on the surface of the coating was mainly WC, consistent with the results of thermodynamic analysis. Figure 6(a) shows the friction and wear coefficient curves of coatings obtained at different welding currents. It can be seen that, at the beginning of the friction and wear test, the wear coefficient of the coating increased rapidly and reached a small step at 200 s. Thereafter, the wear coefficient continued to increase, corresponding to the running-in stage. Compared with the friction coefficients of coatings produced at welding currents of 140 A and 180 A, that of the coating produced at 160 A is lower, reaching about 0.25 in the stable wear stage. The wear rate (Ws) is calculated according to the following formula:

Ws
Wlost PL a 1 ( ) = Therefore, Ws represents wear rate, W lost , P and L represent wear weight loss, load, and wear length respectively, and the frictional wear of 5 mm travel distance. The wear rate and wear weight loss are positively dependent, whereas the lower the wear coefficient of the material, the smaller the wear weight loss, which means superior friction and wear resistance. As shown in figure 6(b), the wear rate of the coating is the lowest when compared to 140 A and 180 A at the welding current of 160 A for the same time. The results seem to indicate that WC was the main phase in the coating obtained at 160 A and that W 2 C existed in the coating as a secondary phase, enhancing the binding energy between the WC particles and the Ni-based, providing strength, and  contributing to the wear resistance of the coating substrate. Thus, the matrix was further enhanced by secondary carbides, which also reduced deformation [43][44][45][46].
To further verify the above conclusions, the surface morphology of the material after wear was observed by SEM. Additionally, it can be seen in figures 7(a) and (c) that the W 2 C particles constituted a strengthening phase in the material, having a hardness much greater than that of the Ni-base. Therefore, when an indenter cuts through W 2 C particles, small particles are broken, and large particles are directly separated from the matrix in which they are embedded to form hard abrasive particles. As a result, there are a lot of holes and cracks on the surface of the coating, accelerating the wear of the material surface. As shown in figures 7(b) and (d), dispersion of the W 2 C reinforcing phase, with small particle size and uniform distribution, makes the wear trace shallow and leads to a smoother spread zone. The results show that when the welding current is 160 A and the particles are worn, W 2 C enhances the bonding strength between the Ni matrix and WC. With W 2 C embedded in the matrix, the WC particles are less likely to fall away, thus improving the wear resistance of the coating [47][48][49][50][51] .

Conclusions
Based on experimental results and discussion, the following conclusions were summarized as follows: (1)W 2 C particles exist in the nickel-based molten pool as the second phase. The length and content of WC particles distributed in the acicular structure around WC particles vary with the current.
(2)The surfacing layer is mainly composed of Fe0.64Ni0.36 phase, WC, W 2 C, and free C, and each element has different degree of diffusion phenomenon. At a temperature above 2615 K, the tungsten carbide particles melt at high temperature and undergo metallurgical reaction, that is, 2WC=W 2 C+C, as the temperature drops. When the temperature is lower than 1690 K, W 2 C begins to change into the WC phase again. The hardness and wear resistance of the coatings are significantly improved under the combined action of the heterotopic and in situ reinforcers.
(3)W2C needle-like particles as an intermediate transition phase reduces the stress between WC and nickel substrate, and the good bond strength and smooth surface of the substrate reduces the porosity and makes the abrasion marks shallow, which greatly improves the friction and wear performance of the material. (4)The thermal expansion coefficient of each phase at the interface shows that the interfacial transition phase W 2 C helps to regulate the difference in thermophysical properties between the body Ni and the reinforcing particles WC, thus slowing down the generation of thermal cracks at the junction of the body Ni and the reinforcing particles.

Data availability statement
All data that support the findings of this study are included within the article (and any supplementary files).