Microstructure and Properties of M 3 B 2 -Type Boride-Based Cermet Coatings Prepared by Laser Cladding Synthesis

: Although Q235 steel materials are widely used in o ﬀ shore engineering, the service life is severely shortened by its inferior resistance to wear and corrosion in harsh marine working environments. Boride-based cermet composites could be a good surface-protective coating to enhance surface hardness, wear resistance, and corrosion resistance. M 3 B 2 (M: Mo, Ni, Fe, Cr) boride-based cermet coatings composed of hard ceramics of M 3 B 2 -type complex borides and an {Fe, Ni} metal matrix was fabricated on Q235 steels with mixed Mo, Cr, B, and Ni60 powders using a laser cladding synthesis technique. The inﬂuences of laser cladding parameters on the microstructure, phase composition, microhardness, and corrosion resistance of the coatings were comprehensively investigated. Results showed that the microstructures of the coatings mainly consisted of three layers, which were, from the top to bottom layer, a metal layer with fewer ceramic phases, a ceramic layer with fewer metal phases, and another metal layer with fewer ceramic phases. The ceramic phases were mainly M 3 B 2 -type borides, and the metal phases were mainly {Fe, Ni} alloys. The appearance of Fe-enriching metal phases was due to the supply of Fe elements from Q235 substrates. With squash pretreatment and without a remelting aftertreatment, ceramics uniformly dispersed in the cermet coatings, and their sizes decreased. The results of microhardness showed that the microhardness of the coating ﬁrst increased and then decreased from the top layer to the bottom layer, and maximum microhardness was obtained in the layer of ceramics with less metal phases. An electrochemical corrosion test showed that the cermet coatings ( j corr = 6.35 µ A / cm 2 ) could improve the corrosion resistance of Q235 steels ( j = 43.76 µ A / cm 2 ) by one order of magnitude.


Introduction
Because of its superior properties, such as high strength, plasticity, and weldability, Q235 steel materials are widely used in the fields of offshore engineering and offshore facilities [1]. However, in harsh marine working conditions, such as high humidity, wear, scour, and corrosion, the service life of Q235 steels is severely shortened because of its inferior resistance to wear and corrosion [2]. To maintain the effective service life of Q235 steels, surface-protective coatings need to be prepared on Q235 steels to enhance the surface hardness, wear resistance, and corrosion resistance.
Boride-based cermet composite coatings consist of a transition metal base matrix with dispersed hard phases, such as TiB 2 , Mo 2 FeB 2 , and Mo 2 NiB 2 , and they possess high hardness, high melting point, good wear and corrosion resistance, and good optical and thermal properties [3][4][5][6][7][8]. These cermet coatings have already been applied to wear-and corrosion-resistant applications such as injection molding machine parts, bearings for sea water pumps, and offshore engineering parts [9]. In particular, as emerging and promising materials with superior strength, hardness, and wear and corrosion Before the laser cladding process, Q235 substrates were sandblasted for 10 min to remove rust oil, then they were cleaned with alcohol. Mo, Ni60, Cr, and B powders were mixed and ball milled for 2 h. During the laser cladding process, a Laserline 4000-100 laser was used for the laser source, and the wavelength of the laser was 1080 nm with a laser spot diameter of 5 mm. Two different powder pretreatment methods, including polyvinyl butyral (PVB) binder presetting and squash presetting, and two different aftertreatment methods, including remelting and no remelting, were used to investigate their influences on M 3 B 2 -based cermet coatings in this experiment. Detailed laser cladding parameters are given in Table 2, and the production process of M 3 B 2 -based cermet coatings is exhibited in Figure 1. After preparation, the coated specimens were cut into small specimens of 10 mm × 10 mm for further characterization.  Before the laser cladding process, Q235 substrates were sandblasted for 10 min to remove rust oil, then they were cleaned with alcohol. Mo, Ni60, Cr, and B powders were mixed and ball milled for 2 h. During the laser cladding process, a Laserline 4000-100 laser was used for the laser source, and the wavelength of the laser was 1080 nm with a laser spot diameter of 5 mm. Two different powder pretreatment methods, including polyvinyl butyral (PVB) binder presetting and squash presetting, and two different aftertreatment methods, including remelting and no remelting, were used to investigate their influences on M3B2-based cermet coatings in this experiment. Detailed laser cladding parameters are given in Table 2, and the production process of M3B2-based cermet coatings is exhibited in Figure 1. After preparation, the coated specimens were cut into small specimens of 10 mm × 10 mm for further characterization.

.Characterization of M3B2-Based Cermet Coatings
The microstructure and composition distribution of the M3B2-based cermet coatings were characterized by scanning electron microscopy (SEM, Hitachi TM3030, Tokyo, Japan) and energydispersive X-ray spectroscopy (EDS, Oxford Swift 3000, Oxford, UK). The phase composition of specimens was analyzed by X-ray diffraction (XRD, Rigaku Ultima IV, Tokyo, Japan) with Cu-Kα radiation (λ = 1.54 A) operated at 40 kV and 20 mA. The 2θ angle was detected from 20° to 100°.
The square cross-sections were rubbed by finer and finer sandpaper, then they were polished with a polishing cloth. Microhardness of the boride-based composite coating was measured by a

Characterization of M 3 B 2 -Based Cermet Coatings
The microstructure and composition distribution of the M 3 B 2 -based cermet coatings were characterized by scanning electron microscopy (SEM, Hitachi TM3030, Tokyo, Japan) and energy-dispersive X-ray spectroscopy (EDS, Oxford Swift 3000, Oxford, UK). The phase composition of specimens was analyzed by X-ray diffraction (XRD, Rigaku Ultima IV, Tokyo, Japan) with Cu-Kα radiation (λ = 1.54 A) operated at 40 kV and 20 mA. The 2θ angle was detected from 20 • to 100 • .
The square cross-sections were rubbed by finer and finer sandpaper, then they were polished with a polishing cloth. Microhardness of the boride-based composite coating was measured by a Vickers hardness tester (HXD-1000TMC/LCD, Shanghai TaiMing, Shanghai, China) with a load of 100 gf for a dwell time of 15 s.
The corrosion resistance of the M 3 B 2 -based cermet coatings was evaluated by an electrochemical workstation (Autolab PGSTAT302N) in 3.5 wt % NaCl solution at 25 • C. A standard three-electrode system was used. The reference electrode was a silver chloride electrode (Ag/AgCl), the counter electrode was platinum, and the working electrode coated specimens with a surface area of 1 cm 2 . The open circuit potential (OCP) was measured for 1 h. Potentiodynamic polarization experiments were scanned at 1 mV/s in the range of OCP ± 0.8 V, and corrosion potential (E corr ) and corrosion current (I corr ) were obtained by the Tafel extrapolation method. The electrochemical impedance spectra (EIS) made up of Nyquist and Bode plots were measured at OCP with a potential amplitude of 5 mV and a frequency of 0.01 to 100,000 Hz. Prior to each corrosion experiment, the coatings were immersed in 3.5 wt % NaCl solution for 1.5 h.

Microstructure and Composition of M 3 B 2 -Based Cermet Coatings
The XRD patterns of M 3 B 2 -based cermet coatings are shown in Figure 2. Results indicate that all samples consisted of the (M 2 M) 3  Vickers hardness tester (HXD-1000TMC/LCD, Shanghai TaiMing, Shanghai, China) with a load of 100 gf for a dwell time of 15 s. The corrosion resistance of the M3B2-based cermet coatings was evaluated by an electrochemical workstation (Autolab PGSTAT302N) in 3.5 wt % NaCl solution at 25 °C. A standard three-electrode system was used. The reference electrode was a silver chloride electrode (Ag/AgCl), the counter electrode was platinum, and the working electrode coated specimens with a surface area of 1 cm 2 . The open circuit potential (OCP) was measured for 1 h. Potentiodynamic polarization experiments were scanned at 1 mV/s in the range of OCP ± 0.8 V, and corrosion potential (Ecorr) and corrosion current (Icorr) were obtained by the Tafel extrapolation method. The electrochemical impedance spectra (EIS) made up of Nyquist and Bode plots were measured at OCP with a potential amplitude of 5 mV and a frequency of 0.01 to 100,000 Hz. Prior to each corrosion experiment, the coatings were immersed in 3.5 wt % NaCl solution for 1.5 h.

Microstructure and Composition of M3B2-Based Cermet Coatings
The XRD patterns of M3B2-based cermet coatings are shown in Figure 2.  The diffraction peak intensity of the {Fe, Ni} phase was very strong, as shown in Figure 2d, and this gradually decreased in Samples 1 and 3. The diffraction peak intensity of the M3B2 phase showed no obvious differences between Samples 1 and 2, but it was stronger in Samples 3 and 4, as shown in Figure 2c,d compared to Figure 2a,b. The distribution of the phase in the four samples was different according to the SEM image in Figure 3, which was the major cause of the difference in intensity in the XRD patterns of the samples. The 10 mm × 10 mm surface of the coating was scanned by a Rigaku The diffraction peak intensity of the {Fe, Ni} phase was very strong, as shown in Figure 2d, and this gradually decreased in Samples 1 and 3. The diffraction peak intensity of the M 3 B 2 phase showed no obvious differences between Samples 1 and 2, but it was stronger in Samples 3 and 4, as shown in Figure 2c,d compared to Figure 2a,b. The distribution of the phase in the four samples was different according to the SEM image in Figure 3, which was the major cause of the difference in intensity in the XRD patterns of the samples. The 10 mm × 10 mm surface of the coating was scanned by a Rigaku diffractometer. Two phases were found in the top layer of the three-layer coating, and the proportion of the phases were different, as shown in Figure 4. The {Fe, Ni} phase was richer in Sample 4, and the M 3 B 2 phase was richer in Sample 3. diffractometer. Two phases were found in the top layer of the three-layer coating, and the proportion of the phases were different, as shown in Figure 4. The {Fe, Ni} phase was richer in Sample 4, and the M3B2 phase was richer in Sample 3.  Microstructures of the samples are shown in Figure 3. White, light gray, and dark gray colored phases were found in all samples. The dark gray phase is the substrate Q235 steel. The other two phases were found in the coating, which was consistent with the XRD results. The cross-section in  diffractometer. Two phases were found in the top layer of the three-layer coating, and the proportion of the phases were different, as shown in Figure 4. The {Fe, Ni} phase was richer in Sample 4, and the M3B2 phase was richer in Sample 3.  Microstructures of the samples are shown in Figure 3. White, light gray, and dark gray colored phases were found in all samples. The dark gray phase is the substrate Q235 steel. The other two phases were found in the coating, which was consistent with the XRD results. The cross-section in Microstructures of the samples are shown in Figure 3. White, light gray, and dark gray colored phases were found in all samples. The dark gray phase is the substrate Q235 steel. The other two phases were found in the coating, which was consistent with the XRD results. The cross-section in the coating, shown in Figure 3, took on a visible structure of stratification, with a large amount of the light-gray phase but less of the white phase in the top layer of the coating, more white phase and less light-gray phase in the middle layer of the coating, and more light-gray phase in the bottom layer of the coating. The white phase was dendritic in the top layer, and it was flocculent in the middle layer, shown in Figure 3a, which used a PVB binder presetting method. The flocculent white phase in the middle of the coating diffused to the surface and the binder interface, and the thickness became thicker after being remelted, as shown in Figure 3b. The thickness of the coating was much thicker when the squash presetting method was used, according to previous experiments which used the PVB binder presetting method with a maximum coating thickness of 0.8 mm. The stratification shown in Figure 3c,d is not obvious, as it is also shown in Figure 3a,b. The top and bottom layers were smaller than those present in coatings using the PVB binder presetting method. The microstructure of the coating using the squash presetting method was very uniform and fine. The white structure of the coating changed from square-to needle-shaped after it was remelted, the thickness increased, and the structure became very uniform.
The elemental map in Figure 5 shows the spatial distribution of elements in the coating by SEM with EDS. The elemental distribution of the white phase was composed of a large amount of Mo and less Fe, Ni, and B, whereas the light-gray phase contained large amounts of Fe and Ni. It is considered that the white phase was M 3 B 2 and the light-gray phase was the {Fe, Ni} phase, which agrees with the XRD analysis. the coating, shown in Figure 3, took on a visible structure of stratification, with a large amount of the light-gray phase but less of the white phase in the top layer of the coating, more white phase and less light-gray phase in the middle layer of the coating, and more light-gray phase in the bottom layer of the coating. The white phase was dendritic in the top layer, and it was flocculent in the middle layer, shown in Figure 3a, which used a PVB binder presetting method. The flocculent white phase in the middle of the coating diffused to the surface and the binder interface, and the thickness became thicker after being remelted, as shown in Figure 3b. The thickness of the coating was much thicker when the squash presetting method was used, according to previous experiments which used the PVB binder presetting method with a maximum coating thickness of 0.8 mm. The stratification shown in Figure 3c,d is not obvious, as it is also shown in Figure 3a,b. The top and bottom layers were smaller than those present in coatings using the PVB binder presetting method. The microstructure of the coating using the squash presetting method was very uniform and fine. The white structure of the coating changed from square-to needle-shaped after it was remelted, the thickness increased, and the structure became very uniform. The elemental map in Figure 5 shows the spatial distribution of elements in the coating by SEM with EDS. The elemental distribution of the white phase was composed of a large amount of Mo and less Fe, Ni, and B, whereas the light-gray phase contained large amounts of Fe and Ni. It is considered that the white phase was M3B2 and the light-gray phase was the {Fe, Ni} phase, which agrees with the XRD analysis.

Hardness
The micro Vickers hardness test point positions of the boride-based composite coatings are shown in Figure 6. Seven points were tested on the cross-section, and three group points of each sample were tested. The average hardness data of three groups are shown in Figure 7. The results show that the trend of cross-section hardness in the four samples was identical. The hardness increased from the top layer of the coating to the middle layer of the coating, then it decreased towards the substrate. The presetting method and process parameters influenced hardness, although the same variation trend regarding hardness was observed in all samples. The mean hardness of Sample 1 was 1343 HV, and the maximum value reached was 1647 HV (in the middle layer). The mean hardness of Sample 3 was 1421 HV, and the maximum value reached 1647 HV in the middle layer. The squash method resulted

Hardness
The micro Vickers hardness test point positions of the boride-based composite coatings are shown in Figure 6. Seven points were tested on the cross-section, and three group points of each sample were tested.

Hardness
The micro Vickers hardness test point positions of the boride-based composite coatings are shown in Figure 6. Seven points were tested on the cross-section, and three group points of each sample were tested. The average hardness data of three groups are shown in Figure 7. The results show that the trend of cross-section hardness in the four samples was identical. The hardness increased from the top layer of the coating to the middle layer of the coating, then it decreased towards the substrate. The presetting method and process parameters influenced hardness, although the same variation trend regarding hardness was observed in all samples. The mean hardness of Sample 1 was 1343 HV, and the maximum value reached was 1647 HV (in the middle layer). The mean hardness of Sample 3 was 1421 HV, and the maximum value reached 1647 HV in the middle layer. The squash method resulted The average hardness data of three groups are shown in Figure 7. The results show that the trend of cross-section hardness in the four samples was identical. The hardness increased from the top layer of the coating to the middle layer of the coating, then it decreased towards the substrate. The presetting method and process parameters influenced hardness, although the same variation trend regarding hardness was observed in all samples. The mean hardness of Sample 1 was 1343 HV, and the maximum value reached was 1647 HV (in the middle layer). The mean hardness of Sample 3 was 1421 HV, and the maximum value reached 1647 HV in the middle layer. The squash method resulted in a harder coating compared to the PVB binder method. Remelting process technology was used in Samples 2 and 4. The hardness curves show that the hardness of Sample 2 was higher than that of Sample 1, whereas the hardness of Sample 4 was lower than that of Sample 3, indicating that the remelting process had a large effect on the hardness. in a harder coating compared to the PVB binder method. Remelting process technology was used in Samples 2 and 4. The hardness curves show that the hardness of Sample 2 was higher than that of Sample 1, whereas the hardness of Sample 4 was lower than that of Sample 3, indicating that the remelting process had a large effect on the hardness.

Electrochemical Impedance Spectroscopy
The EIS analyses of the substrate and the coatings of all samples are shown in Figures 8 and 9. It can be seen that the Nyquist plots of samples formed a single capacitive reactance arc. The diameter of the capacitive reactance arc reflects the rate of the electrochemical reaction [27]. As the Nyquist plots in Figure 8 show, the diameter of the arc increased from substrate to the coating of Sample 1. This indicates that the corrosion resistance was improved by the coating, and the diameter arc was at its maximum when the PVB binder presetting was used without remelting. The frequency-phase Bode plots of all samples are shown in Figure 9. They exhibited the same features as the Nyquist plots. The Z value was lowest for the substrate and highest in Sample 1 at the lowest frequency, which indicates that Sample 1 had the best corrosion resistance.

Electrochemical Impedance Spectroscopy
The EIS analyses of the substrate and the coatings of all samples are shown in Figures 8 and 9. It can be seen that the Nyquist plots of samples formed a single capacitive reactance arc. The diameter of the capacitive reactance arc reflects the rate of the electrochemical reaction [27]. As the Nyquist plots in Figure 8 show, the diameter of the arc increased from substrate to the coating of Sample 1. This indicates that the corrosion resistance was improved by the coating, and the diameter arc was at its maximum when the PVB binder presetting was used without remelting. in a harder coating compared to the PVB binder method. Remelting process technology was used in Samples 2 and 4. The hardness curves show that the hardness of Sample 2 was higher than that of Sample 1, whereas the hardness of Sample 4 was lower than that of Sample 3, indicating that the remelting process had a large effect on the hardness.

Electrochemical Impedance Spectroscopy
The EIS analyses of the substrate and the coatings of all samples are shown in Figures 8 and 9. It can be seen that the Nyquist plots of samples formed a single capacitive reactance arc. The diameter of the capacitive reactance arc reflects the rate of the electrochemical reaction [27]. As the Nyquist plots in Figure 8 show, the diameter of the arc increased from substrate to the coating of Sample 1. This indicates that the corrosion resistance was improved by the coating, and the diameter arc was at its maximum when the PVB binder presetting was used without remelting. The frequency-phase Bode plots of all samples are shown in Figure 9. They exhibited the same features as the Nyquist plots. The Z value was lowest for the substrate and highest in Sample 1 at the lowest frequency, which indicates that Sample 1 had the best corrosion resistance. The frequency-phase Bode plots of all samples are shown in Figure 9. They exhibited the same features as the Nyquist plots. The Z value was lowest for the substrate and highest in Sample 1 at the lowest frequency, which indicates that Sample 1 had the best corrosion resistance. The equivalent circuit model shown in Figure 10 was used to fit the EIS data [28,29]; Rs is the solution resistance from the reference electrode to the working electrode, Rf is the surface layer resistance, and Rct is the constant phase elements of the charge transfer resistance. The impedance of the regular phase-angle element Q is related to the angular frequency (w) by the following relation: ZCPE = Y0 -1 (jw) −n , where Y0 is a proportionality factor and n is the deviation parameter, which reflects the roughness of the surface; Qf is the constant phase angle element (CPE) between the solution and the surface, and Qdl is the CPE of the interface double layer between the solution and the matrix. The fitting results of the EIS data of samples immersed in 3.5% NaCl solution with different processing parameters are listed in Table 3. It shows that Rs changed from 1.17 to 4.51 Ω·cm 2 , indicating that the environmental conductivity was relatively stable. Rf changed from 2.32 to 710 Ω·cm 2 , and Rct changed from 980 to 2460 Ω·cm 2 . The polarization resistance Rp (Rp = Rf + Rct) is used to evaluate the corrosion resistance of materials [30]: the greater the Rp value, the better the corrosion resistance. The Rp value of Sample 1 indicated that it had the best corrosion resistance of all the samples. The results are consistent with the previous analysis of the Nyquist and Bode plots. Table 3. Fitted values of the equivalent circuit of the electrochemical impedance spectra (EIS) diagram.

Rct (Ω·cm 2 )
Qdl, Y0  The equivalent circuit model shown in Figure 10 was used to fit the EIS data [28,29]; R s is the solution resistance from the reference electrode to the working electrode, R f is the surface layer resistance, and R ct is the constant phase elements of the charge transfer resistance. The impedance of the regular phase-angle element Q is related to the angular frequency (w) by the following relation: where Y 0 is a proportionality factor and n is the deviation parameter, which reflects the roughness of the surface; Q f is the constant phase angle element (CPE) between the solution and the surface, and Q dl is the CPE of the interface double layer between the solution and the matrix. The equivalent circuit model shown in Figure 10 was used to fit the EIS data [28,29]; Rs is the solution resistance from the reference electrode to the working electrode, Rf is the surface layer resistance, and Rct is the constant phase elements of the charge transfer resistance. The impedance of the regular phase-angle element Q is related to the angular frequency (w) by the following relation: ZCPE = Y0 -1 (jw) −n , where Y0 is a proportionality factor and n is the deviation parameter, which reflects the roughness of the surface; Qf is the constant phase angle element (CPE) between the solution and the surface, and Qdl is the CPE of the interface double layer between the solution and the matrix. The fitting results of the EIS data of samples immersed in 3.5% NaCl solution with different processing parameters are listed in Table 3. It shows that Rs changed from 1.17 to 4.51 Ω·cm 2 , indicating that the environmental conductivity was relatively stable. Rf changed from 2.32 to 710 Ω·cm 2 , and Rct changed from 980 to 2460 Ω·cm 2 . The polarization resistance Rp (Rp = Rf + Rct) is used to evaluate the corrosion resistance of materials [30]: the greater the Rp value, the better the corrosion resistance. The Rp value of Sample 1 indicated that it had the best corrosion resistance of all the samples. The results are consistent with the previous analysis of the Nyquist and Bode plots. Table 3. Fitted values of the equivalent circuit of the electrochemical impedance spectra (EIS) diagram.

Sample No. Rs (Ω·cm 2 ) Rf (Ω·cm 2 )
Qf, Y0 (m·Ω −1 ·cm −2 ·s n )  The fitting results of the EIS data of samples immersed in 3.5% NaCl solution with different processing parameters are listed in Table 3. It shows that R s changed from 1.17 to 4.51 Ω·cm 2 , indicating that the environmental conductivity was relatively stable. R f changed from 2.32 to 710 Ω·cm 2 , and R ct changed from 980 to 2460 Ω·cm 2 . The polarization resistance R p (R p = R f + R ct ) is used to evaluate the corrosion resistance of materials [30]: the greater the R p value, the better the corrosion resistance. The R p value of Sample 1 indicated that it had the best corrosion resistance of all the samples. The results are consistent with the previous analysis of the Nyquist and Bode plots.

Polarization Curves
The polarization curves of the substrate and coating samples immersed in experimental solutions are shown in Figure 11. It clearly shows that the activated state shifted to a passivated state in the substrate and all samples. A stable passivation film formed on the surfaces of the substrate and coatings. The transition-passivated state of the substrate and Sample 4 are more clearly shown in Figure 11. It shifted from a passivated state to an over-passivated state quickly, and the passivation film located on the substrate surface dissolved. The coatings' passivated curve interval increased similarly to that of the substrate. The passivation film was quickly formed on Sample 1-3, and the time it took for the film to dissolve was much longer in all samples compared to the substrate. The polarization curve shifted to the right, which indicates that the corrosion rate decreased. The corrosion potential of the substrate was the lowest, and that of Sample 1 was the highest in Figure 11. The curves show that the coatings had good corrosion resistance, and Sample 1 had the best corrosion resistance.

Polarization Curves
The polarization curves of the substrate and coating samples immersed in experimental solutions are shown in Figure 11. It clearly shows that the activated state shifted to a passivated state in the substrate and all samples. A stable passivation film formed on the surfaces of the substrate and coatings. The transition-passivated state of the substrate and Sample 4 are more clearly shown in Figure 11. It shifted from a passivated state to an over-passivated state quickly, and the passivation film located on the substrate surface dissolved. The coatings' passivated curve interval increased similarly to that of the substrate. The passivation film was quickly formed on Sample 1-3, and the time it took for the film to dissolve was much longer in all samples compared to the substrate. The polarization curve shifted to the right, which indicates that the corrosion rate decreased. The corrosion potential of the substrate was the lowest, and that of Sample 1 was the highest in Figure 11. The curves show that the coatings had good corrosion resistance, and Sample 1 had the best corrosion resistance. The electrochemical parameters related to polarization curves are shown in Table 4. OCP is open circuit potential, Ecorr is the corrosion potential, icorr is the corrosion current, and jcorr is the corrosion current density. The corrosion potential Ecorr of coatings increased compared to the substrate, and the corrosion current icorr decreased. The corrosion current of Sample 1 was the lowest, and the corrosion rate was the slowest, indicating that it had the best corrosion resistance. Sample 2 also had good corrosion resistance, which was superior to that of Sample 3. The corrosion of Sample 4 was the highest of all coating samples. The analysis is consistent with the EIS results.

Discussion
The microstructural XRD and EDS analyses showed that the M3B2-type boride-based cermet coating was successfully fabricated by laser cladding synthesis. The property tests show that the hardness and corrosion resistance greatly improved. The influences of the laser cladding process, including pretreatments and aftertreatments on the microstructure and phase composition of M3B2 cermet coatings and properties, were investigated. The diffraction peak intensity of the M3B2 phase was stronger in coatings using the squash presetting method than the PVB binder presetting method. The intensity of the M3B2 phase of Sample 1 increased after remelting treatment, whereas it decreased when Sample 3 was remelted. The intensity of the {Fe, Ni} phase of Sample 1 decreased after The electrochemical parameters related to polarization curves are shown in Table 4. OCP is open circuit potential, E corr is the corrosion potential, i corr is the corrosion current, and j corr is the corrosion current density. The corrosion potential E corr of coatings increased compared to the substrate, and the corrosion current i corr decreased. The corrosion current of Sample 1 was the lowest, and the corrosion rate was the slowest, indicating that it had the best corrosion resistance. Sample 2 also had good corrosion resistance, which was superior to that of Sample 3. The corrosion of Sample 4 was the highest of all coating samples. The analysis is consistent with the EIS results.

Discussion
The microstructural XRD and EDS analyses showed that the M 3 B 2 -type boride-based cermet coating was successfully fabricated by laser cladding synthesis. The property tests show that the hardness and corrosion resistance greatly improved. The influences of the laser cladding process, including pretreatments and aftertreatments on the microstructure and phase composition of M 3 B 2 cermet coatings and properties, were investigated. The diffraction peak intensity of the M 3 B 2 phase was stronger in coatings using the squash presetting method than the PVB binder presetting method. The intensity of the M 3 B 2 phase of Sample 1 increased after remelting treatment, whereas it decreased when Sample 3 was remelted. The intensity of the {Fe, Ni} phase of Sample 1 decreased after remelting, whereas it increased when Sample 3 was remelted. The same phenomenon was found in the microstructure analysis.
Polyvinyl butyral powder was dissolved in alcohol, then the saturated solution was mixed with raw powder after being milled. The pasty mass was brushed onto the substrate, but uniformity of the application could not be ensured. The lack of uniformity and poor fluidity caused a difference in the quantity of powder that could react. The non-uniform distribution of the M 3 B 2 phase in the coating is shown in the SEM (Figure 3), which caused a three-layer arrangement to form in the coating. The powder was pressed onto the substrate after being milled, the thickness of the layer was even, and the spherical powder automatically filled the gaps under the press. The microstructure was distributed uniformly in the coating, but the three-layer arrangement was not clear in SEM. Because of the high density of the presetting layer, a thicker coating was formed by the squash presetting method. Partially unreacted elements would react after remelting, which caused the M 3 B 2 phase intensity to increase, and a large amount of the M 3 B 2 phase formed in Sample 2. The Fe element of the substrate would dissolve in the coating while all the elements reacted, which caused the {Fe, Ni} phase intensity and dilution rate to increase in Sample 4. The powders melted rapidly after laser scanning, and the powders reacted and released lots of heat; a molten pool was formed by the heat of the reaction and the laser. Parts of the substrate were melted, generating large amounts of heat; the molten pool cooled rapidly after laser scanning, and the bottom layer consisted of the gray phase, formed because of substrate dilution. The thickness of the coating using the squash presetting method resulted in less and less energy being conducted to the substrate, and the bottom layer was thinner, which meant low dilution of the coating.
A harder coating was obtained by the squash presetting method, according to hardness analysis. A large amount of the M 3 B 2 phase formed uniformly in Sample 3, according to XRD and microstructure analyses. Reacting M 3 B 2 with elements accelerated the nucleation, and the hardness increased under dispersion strengthening. On one hand, remelting will reduce the defects of coating, such as holes and cracks, promoting the reaction; on the other hand, remelting will increase the dilution rate for the substrate element dissolved in the coating. This is why the hardness of Sample 2 increased and the hardness of Sample 4 decreased.
The coating resulted in a good protective substrate under all process parameters. The corrosion current of cermet coating using the PVB binder presetting method was 6.35 µA, and it was 10.05 µA when the squash presetting method was used. The diameter of the capacitive reactance arc of Sample 1 was larger, and the polarization resistance R p was also larger in the fitting results (R p Sample 1 = 2534 Ω·cm 2 , R p Sample 3 = 1551 Ω·cm 2 ). All these showed that the corrosion resistance of the coating using the PVB binder method was better. The corrosion current increased, and arc diameter and polarization resistance decreased, when the samples were remelted. The boride cermet phase was surrounded by a metal phase in the coating; the cermet is a cathode, whereas the metal phase is an anode in the corrosion system. Because massive corrosion cells formed, the M 3 B 2 phase promoted the corrosion of metal matrix anodes. Thus, the presence of a large amount of M 3 B 2 phase in the top layer of Sample 3 could accelerate the surface corrosion and worsen the corrosion resistance of the coatings formed using the squash presetting method. In addition, a large amount of the cermet phase formed in Sample 2 after remelting, and the cermet phase diffused to the top layer of the coating, which led to an inferior corrosion resistance of Sample 2 compared to Sample 1. In a similar way, the elements in the substrate dissolved to the top coating in Sample 4, and consequently, the metal matrix corroded in an accelerated manner.
Formation of a passivation film on the surface of cermet coatings could improve the corrosion resistance of the coatings [31]. In the corrosion cells, anodic dissolution of the metal and the cathodic consumption of the soluble O 2 could favor formation of the passivation film. The procedure of the electrochemical reaction and the formation of passivation films during the experiments are as follows: Anodic reaction: Cathodic reaction: The corrosion reaction is as follows: where M is the metal elements in a layer (according to the XRD and SEM analyses, M is Fe and Ni), and n + is the valence of elements. Formation of the passivation film can reduce the surface portion of the anodic dissolution process, consequently reducing the average velocity of the anodic dissolution process and improving the corrosion resistance of the cermet coatings. Passive films formed anodically on carbon steel have been characterized by electrochemical impedance spectroscopy in the work of Hamadou et al. [32]. This indicates that a stable passive state formed, and the number of defects decreased with increasing film thickness. The corrosion behavior of austenitic stainless steels was investigated in NaCl solution at room temperature by Toor [33]. This and similar studies showed that stabilizing the passive film increased the corrosion resistance. The three-layer-type coating has good properties, and it has an especially dense M 3 B 2 phase in the middle layer and a smaller amount of M 3 B 2 phase in the top layer. The combination property of the substrate can be greatly improved, whereas the coating is fabricated on the substrate in the three-layer-type above. The control of the rate of formation of M 3 B 2 in coating using the squash presetting method and the improvement of the uniformity using the PVB binder presetting method will be investigated in the future. Friction and wear properties will also be researched.

Conclusions
M 3 B 2 boride complex coatings were developed by laser cladding synthesis. Four processing parameters were used in the experiments, which influenced the microstructure and properties of the coatings. X-ray diffraction, SEM, and EDS analyses showed a white M 3 B 2 phase and a light-gray {Fe, Ni} phase distributed in the coating. The thickness of the coatings increased after remelting. A thicker coating could be made when the squash presetting method was used. The structure in the coating using the PVB binder presetting method had more obvious stratification than the structure formed using the squash presetting method. The microstructure in the coating using the squash presetting method was highly uniform and fine. The hardness increased, and the maximum value achieved in the middle layer of the coating then decreased from the surface of the coating to the substrate in all samples. The average hardness was higher using the squash presetting method, and remelting technology had a big effect on the hardness. The corrosion resistance of all coatings improved compared to the substrate.
The complex coatings possess good properties, and they can be used for remanufacturing key aspects of moving parts and in additive manufacturing of small complex structure parts applied in marine environments. This new, laser-induced method helps to synthesize a ternary boride complex coating directly on the substrate. Acknowledgments: The authors acknowledge Shanghai Lin Shi Laser Technology Co., Ltd., for experiment support. Professor Chunwang Liu from Shanghai Maritime University is acknowledged for help.

Conflicts of Interest:
The authors declare no conflict of interest regarding the publication of this article.