E ﬀ ect of multiple laser re - melting on microstructure and properties of Fe - based coating

: The aim of this article is to explore the e ﬀ ect of re - melting times on the microstructure and properties of Fe - based coating. In this study, the Fe - based coating is prepared on 316L stainless steel by laser cladding and laser re - melting. Meanwhile, the microstructure and properties of the coating are studied by 3D laser scanner, Vickers microhardness tester, X - ray di ﬀ ractometer, and scanning electron microscope. In addition, the e ﬀ ect of laser re - melting times on microstructure formation that is analyzed by numerical simulation. The results show that re - melting can lead to the decrease in coating height, increase in coating width, and increase in both depth and width of melting pool. The hardness of coatings is enhanced by six times compared with the substrate. However, it was found that the hardness of the coating decreased with the increase in laser re - melting times. The abnormal decrease in hardness was analyzed because of the continued growth of crystals in the coating and an increase in the coating dilution rate. The ﬁ rst laser re - melting results in the obvious change of coating crystal. The crystals of the multiple laser re - melting coating con - tinue to grow. Our research results can provide reference for laser multiple re - melting in industry.


Introduction
FeCrMo alloys are potential candidates due to their superior mechanical properties. Because of its low production cost, 316L is commonly used in industry. Laser melting is just an advanced surface modification technology. By converting light energy into heat energy, the substrate and alloy powder are fused to form molten metal by high energy density. The molten metal is solidified and metallurgically bonded with the substrate to develop a high-performance coating [1][2][3]. FeCrMo can be coated on 316L by laser cladding, which has become a low cost method to prepare highperformance coatings.
In conventional thermal processing, it is difficult to form a coating that meets the requirements (dilution rate, porosity, surface roughness, etc.) [4][5][6]. Laser re-melting is a way of laser melting after the first sample preparation has been completed [7]. In recent years, researchers have used laser re-melting to produce coatings that meet these needs [8][9][10].
Fals et al. [11] sprayed the NbC coating flame onto the AISI 304L substrate and re-melted it using a fiber laser. It is found that the hardness of laser re-melting zone is higher than that of the substrate. Panziera et al. [12] conducted laser re-melting on the surface of the coating. By changing the scanning speed and laser beam power, laser re-melting can obtain a coating with high hardness, no pores, or obvious defects. Zhang et al. [13] modified the surface coating of plasma spraying by laser re-melting and Al deposition to enhance its high-temperature corrosion resistance. It was found that the minimum surface roughness of the dense laser re-melting coating was the main reason for improving the high-temperature corrosion resistance. Godec et al. [14] modified the thermal barrier coatings with different shapes by laser re-melting technology and found that they had dense columnar crystal structure and high microhardness.
Wang et al. [15] prepared Fe-based coating on the surface of H13 steel by laser cladding and re-melting.
It was found that after re-melting, the defects such as cracks and pores on the coating surface were significantly reduced. After re-melting, the overall corrosion resistance of the re-melting layer is greater than that of the coating. Xi et al. [16] found that re-melting not only improved the morphology of the coating but also increased the microhardness and wear resistance of the coating. When the laser power is 650 W, the coating has not only small surface roughness and no defects but also excellent mechanical properties.
As can be viewed above, laser re-melting is an effective method to improve the minimum surface roughness, wear resistance, and corrosion resistance of coatings. This work is to study the effect of laser multiple remelting on the microstructure and properties of coatings. Through the combination of numerical simulation and experiment, the influence of re-melting times on the coating was compared and analyzed. In this article, the effects of multiple laser re-melting on the macroscopic morphology, microstructure forming ability, hardness, and microstructure of Fe-based coatings were studied. The effect of laser re-melting cooling rate on microstructure forming ability was analyzed by finite element numerical simulation.

Experimental
FeCrMo powder was employed as the cladding material. Table 1 lists the contents of FeCrMo powder and 316L stainless steel. The substrate was polished with sandpaper before laser treatment, cleaned by anhydrous ethanol, and dried in the drying cabinet.
The laser source is provided by 2,000 W fiber laser (XL-F2000T). Prefabrication method was used to coat powder on the substrate. The processing schematic diagram is shown in Figure 1. The laser cladding process is divided into two stages: laser cladding and laser remelting. According to the relevant literature [15,17], low power and low scanning speed are selected by melting, and high power and high scanning speed are selected by laser re-melting. Laser cladding power is 300 W, scanning speed is 300 mm·min −1 , laser re-melting power is 1,600 W, scanning speed is 2,400 mm·min −1 , defocusing is 0 mm, and the experimental scheme is shown in Table 2. Re-melting is carried out in inert argon at 5 L·min −1 . Use corrosion solution (aqua regia) for 10 s after processing. The macro morphology of the coating was measured by Scan Tech 3D laser scanner (PRINCE335). The microhardness of the coating was measured by Wilson Vickers microhardness tester (BUEHLER VH1202). The crystalline phase composition of the coating was analyzed by X-ray diffraction (XRD). The microstructure of the coating was observed by JSM 6460 scanning electron microscope (SEM). Figure 2 shows the coating schematic.

Finite element analysis
For crystalline alloys, cooling rate is the most important factor to restrain nucleation and growth of competitive crystalline phase. The cooling rate of the coating under multiple laser re-melting also affects its microstructure forming ability. Therefore, the finite element method is   Re-melting passes 0 1 2 used to solve the heat conduction equation, and the temperature change process of the coating during laser re-melting is obtained, which can provide prediction for the thermal cycle of microstructure formation and change.
In the machining process, with the effect of laser heat source, the temperature of the workpiece changes, and the thermal physical parameters of the material also change. Therefore, the analysis of temperature field is an obvious nonlinear transient heat conduction problem. The three-dimensional temperature field equation is shown as where ρ and c are material density and specific heat capacity, respectively; k x , k y , and k z are the thermal conductivity in three directions in turn; T is the temperature field distribution function; and Q is the thermal power per unit volume. Convection and radiation heat dissipation Q1 exists on the coating surface (interface between substrate material and air) and around the substrate surface, which satisfies [19,20] ( where T 0 is the ambient temperature; h c is the convection coefficient, W·m −2 ·K −1 ; surface emissivity of ε matrix; and σ is the Stefan-Boltzmann constant of 5.67 × 10 −8 W·m −2 ·K −4 . At present, the moving heat source models for numerical simulation of laser machining mainly include surface heat source model, body heat source model, and combined heat source model [21][22][23]. The heat source expression of conical heat source model is as follows [24]: where E is the heat generation rate, α is the substrate absorption efficiency, β is the powder loss efficiency, η is the laser power efficiency, P is the total laser power, r is the laser spot radius, d is the total heat source depth of the cone, and a is the XOY plane offset coefficient. The heat source model of the conical heat source model is shown in Figure 3. By Ansys Parametric Design Language (APDL), some heat generating units (changing d 1 ) can be selected to adjust the heat source model of the conical heat source model.
According to the relevant literature [25][26][27], the nonlinear physical parameters of FeCrMo and 316L were added in Ansys material library. Table 3 shows the physical parameters of FeCrMo and 316L: T-temperature, ρmaterial density, c-specific heat capacity, and k-thermal conductivity.
To simplify the model, only half of the workpiece model is simulated by symmetrical modeling. In the finite element numerical simulation, the coating size is 1.8 mm × 0.5 mm × 20 mm, and the 316L stainless steel substrate size is 50 mm × 50 mm × 2 mm. In numerical simulation, the software environment is set as follows:   (4) In order to improve the convergence of the calculation results, the empirical formula for heat transfer coefficient is used as follows: , and surface emissivity of ε matrix. (5) Adiabatic boundary condition is applied on the symmetric surface of the symmetric model, and heat transfer coefficient (4) is applied on the other outer surfaces. Considering the latent heat of phase transformation of materials, it can be expressed as where f 1 is the liquid mass fraction and L is the solidification latent heat value. According to ref. [27], the solid phase line of 316L was 1520.5 K, the liquid phase line was 1,609 K, and the latent heat value was 2.8 × 10 5 K·kg −1 . The gasification temperature is 3,133 K and the latent heat is 6.1 × 10 6 K·kg −1 . Due to the lack of relevant heat values of FeCrMo, 316L latent heat data were used.
where T l is the liquid line, T s is the solid line, and T is the unit temperature.   (Figure 5c) has a small fluctuation, and the maximum height difference is 0.09 mm. The morphology of the first re-melting coating changed obviously. The coating height (CH) decreases along the X axis. In other words, the CH decreases slightly along the scanning path. The surface of sample 1 coating appeared bump phenomenon. The high-power heat input   is performed again by the re-melting process, and the unevenness of the sample 2 is alleviated, while the CH is reduced and the coating width (CW) is increased.
Increasing the re-melting times, the above phenomenon of sample 3 is further obvious. The increase in re-melting times reduces the fluctuation of coating surface. Table 4 shows the cross sections of coating and molten pool of each sample. The dilution rate formula is calculated according to equation (7) [17,28]. Figure 2b shows the schematic of coating geometry. The CH of sample 1 is 0.76 mm. The CH of sample 2 is 0.60 mm. The CH of sample 3 is 0.54 mm. The CH of sample 2 was significantly lower than that of sample 1, and the CH of sample 3 was relatively lower than that of sample 2. The CW of each sample and the depth of molten pool (DMP) and width of molten pool (WMP) have similar CH. The above phenomena indicate that re-melting will lead to the decrease in CH and increase in CW, and the size of the first re-melting coating changes significantly. It is found that re-melting will increase the DMP and WMP, and the size of the first re-melting molten pool changes significantly.
where S 1 is the area of cladding layer and S 2 represents the area of molten substrate on the cross-section of a sample.

Hardness
The microhardness of the samples along the cross section was measured under 0.2 kg loading for 10 s. Each sample has 30 indentation points, 3 along the Y axis direction and 10 along the Z axis direction. As shown in Figure 6, the hardness of sample 1 reaches 1,400 HV 0.2 , which is six times higher than that of matrix (about 200 HV 0.2 ). The hardness of samples 1, 2, and 3 coatings decreased in turn. It shows that re-melting will lead to a decrease in hardness.

XRD and microstructure
The XRD patterns of the three sample coatings are shown in Figure 7. It can be seen that the diffraction peaks of the three samples have no obvious change. It can be said that the increase in the number of re-melted elements will not continue to generate in the form of compounds. Figure 8 shows SEM photographs of different parts of three samples. Figure 2c shows the position distribution   Figure 6: Distribution of microhardness of coatings.
of SEM observation. Some crystal structures can be seen, and the microstructure of the coating changes obviously after re-melting. No crystals appear in Figure 8a. A large number of fine flower crystals (I) appear in Figure 8b. Figure 8c shows coarse snowflake crystals (II). It was found that at the top of the coating, the crystal was formed in sample 2 compared with sample 1, while the crystal continued to grow in sample 3. In the middle of the coating, there was no crystal in Figure 8d, and a small amount of coarse snowflake crystals were formed after laser re-melting (Figure 8e III). A large number of coarse snowflake crystals were formed after laser re-melting (Figure 8f IV). At the bottom of the coating, the matrix below the interface appears martensite structure. However, different crystals appear above the interface. Figure 8g shows that no crystal appears in the fused layer. In Figure 8h, some vertical columnar dendrites (V) form arc bands. A large number of columnar dendrites (VI) parallel to the heat flow direction form larger arc bands in Figure 8i. It shows that the first laser re-melting leads to the crystallization of the coating, forming an arc crystal band as shown in the figure, and the laser re-melting continues to form a larger arc crystal band. According to the rapid solidification theory [29], higher cooling rate and undercooling rate can be obtained by laser scanning speed, and the grains grow in columnar dendrites parallel to the direction of heat flow (V, VI). Figure 9 shows grain size of different parts of three samples. At the top, the grain sizes of samples 1, 2, and 3 are 0.339, 0.806, and 3.04 μm, respectively. In the middle, the grain sizes of samples 1, 2, and 3 are 0.534, 1.44, and 6.24 μm, respectively. At the bottom, the grain sizes of samples 1, 2, and 3 are 0.311, 1.66, and 6.52 μm, respectively. The grain size of samples 1, 2, and 3 in each part increases in turn.   Figure 10a shows the temperature distribution of sample 1 laser melting process. It is clear that the moving heat source is comet-like. Figure 10b-d shows the comparison between the numerical simulation of each sample and the actual molten pool. The red region is higher than the melting point of 1,609 K (as the coating is assumed rectangular size, it is not shown in the picture). It indicates that the red region is melted during laser melting. The comparison between the experimental results and the temperature field shows that the temperature predicted by the established model is in good agreement with the simulation results, which can be used for the thermal cycle prediction of laser re-melting process. Figure 11 shows the temperature change calculation results of each sample during laser re-melting. It can be  seen that the highest temperature of sample 1 is 1,946 K, sample 2 is 2,616 K, and sample 3 is 2,656 K at the top of the surface. We calculated the cooling rate from the highest temperature to 700 K. The cooling rates of samples 1-3 are 1523.4, 6922.2, and 5609.7 K·s −1 , respectively. The interfacial cooling rates of samples 1-3 were 1332.3, 3207.4, and 2547.1 K·s −1 , respectively. It was found that the cooling rates of the surface vertices and interfaces of the three samples had similar trends. Sample 2 has the highest cooling rate and sample 1 has the lowest cooling rate.

Discussion
In Figure 5, the increase in laser re-melting times leads to a decrease in the surface undulation of the coating. With the increase in the number of laser re-melting, the accumulated heat input makes the temperature of the coating exceed the melting point (as can be found in Figure 11), and the coating is repeatedly re-melted and solidified, resulting in the decrease in the fluctuation of the upper surface of the coating. The results are consistent with previous studies by Zhang et al. [13]. Laser re-melting can reduce the hardness of the coating (as can be seen in Figure 6). The heat input of the laser re-melting will cause the coating to continue to be heated at high temperatures, causing the amorphous structure to crystallize and the grain to continue to grow (as can be seen in Figure 8). We have measured the grain size in Figure 8, and the results are shown in Figure 9. The grain size is increased by the re-melting process. At the same time, laser re-melting increases the dilution rate (Table 4), which can result in a decrease in hardness [12,30]. Sousa et al. [31] also concluded in his study. In XRD analysis (Figure 7), it can be found that no new compound was formed in the coating after laser re-melting.
Combined with the analysis of Figure 8, the tissue of sample 2 changed significantly compared with sample 1. According to the rapid solidification theory [29], from the substrate to the coating, the microstructure will show a typical structural transformation order of planar crystal, cellular crystal, dendritic crystal, and equiaxed crystal. But only dendritic and snowflake crystals were found in the three samples. In the cold zone rate, sample 1 is the smallest, and there is no obvious crystal formation. The cooling rates of sample 2 and sample 3 are similar. After the first laser re-melting, sample 2 crystal was formed. After laser re-melting again with the same process parameters, sample 3 crystal grew. However, the tissue changes of sample 3 were not obvious compared with sample 2. This is because the behavior of laser melting before laser re-melting has preheating effect on the substrate. The substrate temperature increases during laser re-melting, resulting in a decrease in temperature gradient. But the crystal continues to grow as the laser re-melting continues to heat input again.

Conclusions
Fe-based composite coating was synthesized by laser cladding and laser re-melting. The microstructure and mechanical properties of the samples were studied. Laser remelting was simulated by finite element method. According to the experimental and numerical simulation results, the effects of multiple laser re-melting on microstructure forming ability and coating performance were analyzed.
(1) The morphology and size of the first re-melting coating changed obviously. The fluctuation of the coating surface decreases after multiple re-melting. (2) Multiple laser re-melting results in the decrease of coating hardness. The reason is that the crystal phase continues to grow and the coating dilution rate increases. (3) After the first laser re-melting, the multiple re-melting had no effect on the coating compounds.
Finally, considering the influence of multiple laser re-melting on the macroscopic morphology, microstructure forming ability, hardness and microstructure of the coating, the appropriate number of laser re-melting can be selected. Author contributions: Linyi Xie: conceptualization, methodology, formal analysis, writingoriginal draft, and writingreview and editing. Wenqing Shi: conceptualization, formal analysis, super-vision, and project administration. Teng Wu, Meimei Gong Jiang Huang, and Yuping Xie: discussions. Kuanfang He: fund. All authors have read and agreed to the published version of the manuscript.

Conflict of interest:
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this article.