Influence of Annealing Treatment on Microstructure and Properties of Ni-Rich NiTi Alloy Coating Prepared by Laser Cladding

NiTi alloys are widely known for their shape memory effect and super-elasticity. In this study, the laser cladding method was applied to prepare Ni-rich NiTi alloy coatings on 316L stainless steel substrate. The microstructure, phase composition, element distribution and phase transformation behavior of the coatings were investigated in as-fabricated and annealing-treated states. The results indicated that the recrystallized microstructure obtained and the content of Ni3Ti and Ti2Ni phases increased significantly with a rising annealing temperature. Annealing treatment also induced a decrease in the phase-transition enthalpy and a rise in the transformation temperature, even though no obvious martensite transformation was observed. This was suppressed due to the Fe element diffused from the substrate and was probably retarded by the mounting metallic compounds formed during annealing as well. The mechanical properties have also improved obviously; coatings annealed under 850 °C exhibited the highest microhardness of 839 HV, and the wear resistance of the coatings after annealing was enhanced with an 11% average wear mass loss reduction.


Introduction
NiTi shape memory alloy (SMA) was discovered by William J. Buehler et al. at the Naval Ordnance Laboratory (NOL) in the late 1950s, so it was also named as Nitinol [1]. In addition to the unique shape memory effects, NiTi alloys are also well known due to their super-elasticity and biocompatibility [2,3]. Since then, NiTi alloys have been widely used in aerospace, marine, military, automobile, biomedicine and other fields for applications such as antennas, connectors, actuators, cryogenic fittings, stents, tissue implants and medical devices [4][5][6]. In addition to the near equiatomic NiTi alloy (also called 55NiTi with 55 wt.% Ni), 60NiTi (with 60 wt.% Ni) is a Ni-rich version of the equiatomic NiTi alloy, which exhibits not only super-elasticity and shape memory effects, but also high tensile strength as well as excellent abrasion and corrosion resistance; therefore it is considered an ideal material for aerospace bearing applications [7,8].
However, the outstanding properties of NiTi alloys also make them difficult to work with; the manufacturing and processing of these alloys are very challenging, which limits their further extensive utilization [9][10][11]. In recent years, with the development of additive manufacturing (AM), which produces parts or components with certain shapes by adding layers of powders progressively [12], the application of AM methods has gained a lot of attention, such as laser metal deposition (LMD), selective laser melting (SLM) and laser cladding, for the manufacturing of NiTi alloy parts [13,14]. Laser cladding with NiTi alloy powders, applied as a new coating technology, has become one of the research hotspots for improving substrate surface performance. For example, Norhafzan et al. prepared NiTi cladding layers on tool steel and found that the surface hardness was almost three times that of its substrate [15]. Mokgalaka  A high-power fiber laser (YLS-10000, IPG Photonics, Oxford, MA, USA), as shown in Figure 1, was employed for laser cladding.

Material and Sample Preparation
Pure Ni powder and 55NiTi alloy powder (99.9% purity, Jiangsu Willari New Material Technology Co., Ltd., Xuzhou, China) with spherical morphology and a particle size distribution of d10 = 55.25 μm, d50 = 85.73 μm and d90 = 125.90 μm were used as feedstock materials for laser cladding. The chemical composition of the applied powders and substrate (AISI 316L stainless steel, with dimension size of 100 mm × 100 mm × 10 mm) are shown in Tables 1 and 2, respectively, in which Bal. means balance element. NiTi alloy powder and pure Ni powder are mixed according to the weight proportion of 8:1 at 250 rpm by a planetary ball mill (QM-3SP4, Nanda Instrument, Nanjing, China) for 8 hours and dried in an oven under 110 °C, while research shows that pre-alloyed powders are beneficial for a more homogeneous microstructure and stable mechanical properties [33].  Figure 1, was employed for laser cladding. The diameter of the laser beam spot was fixed by 7.2 mm with a wavelength of 1080 nm. Based on previous studies [34,35], the parameters for coating fabrication were under: 2.0 kW power, 2 mm/s laser scanning speed, 50 rpm powder feeding rate and 55% overlap ratio. Argon (99.9 vol.%) with a flow rate of 5 L/min was applied as shielding gas to prevent oxidation of the molten pool and powders. Apart from that, it was used as a powder feeding gas with a flow rate of 8 L/min as well. The substrate was preheated to 250 °C to release internal stress before laser cladding. After laser cladding, the substrate with coating was cut into numerous cuboid samples with dimensions of 5 mm × 5 mm × 10 mm for further investigation. Samples for the subsequent microstructure observations and hardness test were cut perpendicular to the laser scanning direction. These cross-section samples were grinded, polished and etched by an HF: HNO3: H2O solution with a dilution of The diameter of the laser beam spot was fixed by 7.2 mm with a wavelength of 1080 nm. Based on previous studies [34,35], the parameters for coating fabrication were under: 2.0 kW power, 2 mm/s laser scanning speed, 50 rpm powder feeding rate and 55% overlap ratio. Argon (99.9 vol.%) with a flow rate of 5 L/min was applied as shielding gas to prevent oxidation of the molten pool and powders. Apart from that, it was used as a powder feeding gas with a flow rate of 8 L/min as well. The substrate was preheated to 250 • C to release internal stress before laser cladding. After laser cladding, the substrate with coating was cut into numerous cuboid samples with dimensions of 5 mm × 5 mm × 10 mm for further investigation. Samples for the subsequent microstructure observations and hardness test were cut perpendicular to the laser scanning direction. These cross-section samples were grinded, polished and etched by an HF:HNO 3 :H 2 O solution with a dilution of 1:4:5. Samples for differential scanning calorimetry (DSC) tests were cut in the form of a cylinder with a height of 2 mm, a diameter of 3 mm and an average weight about 50 mg. These samples were polished to remove surface impurities.

Experimental Procedure
The heat treatment of the prepared cuboid samples was conducted with the use of an electrical resistance furnace (YFX16, Shanghai Y-FENG Electrical Furnace Co., Ltd, Shanghai, China), as shown in Figure 2. With a heating rate of 10 • C/min, the furnace was always heated 10 • C higher than the desired temperature, so that its temperature was allowed to drop to the designated temperature when the furnace chamber was opened later on to put in the aforementioned samples. A timer would be activated once the furnace temperature or hold time reached the set value, and subsequently, the related samples were taken out of the furnace and air-cooled under room temperature. As shown in Table 3, there were 4 groups of samples, which were annealed separately according to the different temperatures and holding time, and each group had 4 pieces to check their repeatability. In addition, HT is defined as the abbreviation of heat treatment in this article.
Materials 2021, 14, x FOR PEER REVIEW 4 of 16 1:4:5. Samples for differential scanning calorimetry (DSC) tests were cut in the form of a cylinder with a height of 2 mm, a diameter of 3 mm and an average weight about 50 mg. These samples were polished to remove surface impurities.

Experimental Procedure
The heat treatment of the prepared cuboid samples was conducted with the use of an electrical resistance furnace (YFX16, Shanghai Y-FENG Electrical Furnace Co., Ltd, Shanghai, China), as shown in Figure 2. With a heating rate of 10 °C/min, the furnace was always heated 10 °C higher than the desired temperature, so that its temperature was allowed to drop to the designated temperature when the furnace chamber was opened later on to put in the aforementioned samples. A timer would be activated once the furnace temperature or hold time reached the set value, and subsequently, the related samples were taken out of the furnace and air-cooled under room temperature. As shown in Table 3, there were 4 groups of samples, which were annealed separately according to the different temperatures and holding time, and each group had 4 pieces to check their repeatability. In addition, HT is defined as the abbreviation of heat treatment in this article.

Sample Characterization
In order to investigate the microstructure of the samples before and after heat treatment, an optical microscope (OM, GX51, Olympus, Tokyo, Japan) and a scanning electron microscope (SEM, Zeiss Gemini 300, Oberkochen, Germany) with a configured energydispersive spectroscopy (EDS, OXFORD Xplore, Oxford, England) were used for observation. In addition, X-ray diffraction (XRD, Bruker D8 ADVANCE, Karlsruhe, Germany) was performed to study the phase composition of the coating with Cu Kα radiation at 40 kV and 100 mA and a scanning speed of 2°/min, ranging from 10° to 80°. The differential scanning calorimetry (DSC250, TA Instruments, New Castle, DE, USA) with a measuring

Sample Characterization
In order to investigate the microstructure of the samples before and after heat treatment, an optical microscope (OM, GX51, Olympus, Tokyo, Japan) and a scanning electron microscope (SEM, Zeiss Gemini 300, Oberkochen, Germany) with a configured energydispersive spectroscopy (EDS, OXFORD Xplore, Oxford, England) were used for observation. In addition, X-ray diffraction (XRD, Bruker D8 ADVANCE, Karlsruhe, Germany) was performed to study the phase composition of the coating with Cu Kα radiation at 40 kV and 100 mA and a scanning speed of 2 • /min, ranging from 10 • to 80 • . The differential scanning calorimetry (DSC250, TA Instruments, New Castle, DE, USA) with a measuring range from −180 • C to 725 • C, temperature accuracy of 0.05 • C and enthalpy precision of 0.08%, was used to study the coating phase transformation behavior. The DSC test refers to ASTM Standard F2004-17 and was under a nitrogen atmosphere with temperature scanning ranges from −80 • C to 180 • C and a heating and cooling speed of 10 • C/min. For the friction and wear test, the experiment was conducted in a ball-on-plate configuration (Ht-1000, ZKKH Science and Technology Development Co., Ltd., Lanzhou, China) using 5 mm diameter Si 3 N 4 balls at room temperature under a load of 15 N and a rotation speed of 224 r/min. Furthermore, a microhardness tester (HVS1000A, HuaYin Testing Machine Technology Co., Ltd., Laizhou, China) was employed in the experiment with a load of 1.961 N and a holding time of 15 s. Hardness was measured from the coating surface to the substrate at an interval of 0.2 mm and repeated 3 times at the same depth to calculate the average value.

Coating Phase Composition Analysis
The analysis of coating phase composition is based on the XRD results. Figure 3 shows the XRD patterns of the as-fabricated and heat-treated 55NiTi+5Ni laser cladding samples, which has been normalized according to a standard intensity scale.
range from −180 °C to 725 °C, temperature accuracy of 0.05 °C and enthalpy precision of 0.08%, was used to study the coating phase transformation behavior. The DSC test refers to ASTM Standard F2004-17 and was under a nitrogen atmosphere with temperature scanning ranges from −80 °C to 180 °C and a heating and cooling speed of 10 °C/min. For the friction and wear test, the experiment was conducted in a ball-on-plate configuration (Ht-1000, ZKKH Science and Technology Development Co., Ltd., Lanzhou, China) using 5 mm diameter Si3N4 balls at room temperature under a load of 15 N and a rotation speed of 224 r/min. Furthermore, a microhardness tester (HVS1000A, HuaYin Testing Machine Technology Co., Ltd., Laizhou, China) was employed in the experiment with a load of 1.961 N and a holding time of 15 s. Hardness was measured from the coating surface to the substrate at an interval of 0.2 mm and repeated 3 times at the same depth to calculate the average value.

Coating Phase Composition Analysis
The analysis of coating phase composition is based on the XRD results. Figure 3 shows the XRD patterns of the as-fabricated and heat-treated 55NiTi+5Ni laser cladding samples, which has been normalized according to a standard intensity scale. It can be seen that the coating was mainly composed of NiTi and Ni3Ti phase, which show the strongest diffraction peaks. In addition to that, secondary phases such as TiO2, Fe2Ti and Ni4Ti3 were also detected with a relative weak diffraction intensity after the annealing treatment. Of these, Ni4Ti3 is a metastable phase with nanoscale structure, which has similar diffraction peaks to NiTi phase. The formation of Fe2Ti was due to diffusion of the Fe atom from the substrate, and TiO2 was attributed to the oxidation of the coating surface, which has not been observed before annealing. It is worth noting that, compared with the patterns before annealing, the intensities of the characteristic diffraction peak of Ni3Ti (2θ = 46.53°) and Ti2Ni (2θ = 41.56°) are getting obviously stronger. This was probably caused by the decomposition of NiTi and Ni4Ti3 phase (NiTi / Ni4Ti3 → Ni3Ti + Ti2Ni), It can be seen that the coating was mainly composed of NiTi and Ni 3 Ti phase, which show the strongest diffraction peaks. In addition to that, secondary phases such as TiO 2 , Fe 2 Ti and Ni 4 Ti 3 were also detected with a relative weak diffraction intensity after the annealing treatment. Of these, Ni 4 Ti 3 is a metastable phase with nanoscale structure, which has similar diffraction peaks to NiTi phase. The formation of Fe 2 Ti was due to diffusion of the Fe atom from the substrate, and TiO 2 was attributed to the oxidation of the coating surface, which has not been observed before annealing. It is worth noting that, compared with the patterns before annealing, the intensities of the characteristic diffraction peak of Ni 3 Ti (2θ = 46.53 • ) and Ti 2 Ni (2θ = 41.56 • ) are getting obviously stronger. This was probably caused by the decomposition of NiTi and Ni 4 Ti 3 phase (NiTi / Ni 4 Ti 3 → Ni 3 Ti + Ti 2 Ni), which were unstable when heated to a higher temperature [36,37]. In addition, the content of Fe 2 Ti increased significantly when annealing at a higher temperature, which was due to more active diffusion activity.

Coating Microstructural Analysis
The surface morphology of the coatings before and after the annealing treatment is shown in Figure 4. A few unfused powders are found on the surface of the coating; while laser cladding is a quick heating and cooling process, the temperature gradient becomes larger from the molten pool to the upper surface, and therefore, these powders could not absorb the energy sufficient to melt. However, with an increasing annealing temperature, the surface flatness shows the trend improving. It is worth noting that the grain boundaries are more obviously seen in the samples annealing at 450 • C and 650 • C; however, they are no more visible when annealing at 850 • C, which is due to the occurrence of recrystallization. which were unstable when heated to a higher temperature [36,37]. In addition, the content of Fe2Ti increased significantly when annealing at a higher temperature, which was due to more active diffusion activity.

Coating Microstructural Analysis
The surface morphology of the coatings before and after the annealing treatment is shown in Figure 4. A few unfused powders are found on the surface of the coating; while laser cladding is a quick heating and cooling process, the temperature gradient becomes larger from the molten pool to the upper surface, and therefore, these powders could not absorb the energy sufficient to melt. However, with an increasing annealing temperature, the surface flatness shows the trend improving. It is worth noting that the grain boundaries are more obviously seen in the samples annealing at 450 °C and 650 °C; however, they are no more visible when annealing at 850 °C, which is due to the occurrence of recrystallization.  Studies reported that the recrystallization temperature of 55NiTi alloy is between 550 °C and 650 °C and would further increase with the rising content of Ni in the alloy [22,27]. As the typical dendrite structure still remains in the coating after annealing under 650 °C, which is shown in Figure 5c, it can be shown that the recrystallization temperature of the 55NiTi+5Ni coating is at least higher than 650 °C. Figure 5d and Figure 5e show the microstructure of coating after annealing under 850 °C, which displays  Figure 5 displays the SEM morphology of the coating cross-sections before and after the annealing treatment. Studies reported that the recrystallization temperature of 55NiTi alloy is between 550 • C and 650 • C and would further increase with the rising content of Ni in the alloy [22,27]. As the typical dendrite structure still remains in the coating after annealing under 650 • C, which is shown in Figure 5c, it can be shown that the recrystallization temperature of the 55NiTi+5Ni coating is at least higher than 650 • C. Figure 5d,e show the microstructure of coating after annealing under 850 • C, which displays only large single grains with black inclusions and indicates that the recrystallization has been completed. In addition, the average grain size of the coating displays an enlargement trend with an increasing annealing temperature. According to ASTM E112, the linear intercept method was applied to measure the grain size. The grain intercept count (GIC) stands for the number of times a test line cut through the grains, in which tangent hits and cutting lines that end within a grain are considered to be a 0.5 interception. As remarked in Figure 5a, L 1 and L 2 are the intercept lines, and the grain size is calculated through the following formula: The measuring results are listed in Table 4, and Figure 5f is the bar chart of the results. It can be seen that the coating average grain size became larger with an increasing annealing temperature and a prolonged holding time as well.
stands for the number of times a test line cut through the grains, in which tangent hits and cutting lines that end within a grain are considered to be a 0.5 interception. As remarked in Figure 5a, L1 and L2 are the intercept lines, and the grain size is calculated through the following formula: The measuring results are listed in Table 4, and Figure 5f is the bar chart of the results. It can be seen that the coating average grain size became larger with an increasing annealing temperature and a prolonged holding time as well. Table 4. Measuring results of average grain size with linear intercept method. In comparison to the coating without heat treatment, there are more black inclusions found in the heat-treated coatings, which occur among the dendrites and inter-dendrites area. According to the EDS regional scanning results of these inclusions, shown in Figure  6, it can be concluded that the black inclusions are titanium oxide.    In comparison to the coating without heat treatment, there are more black inclusions found in the heat-treated coatings, which occur among the dendrites and inter-dendrites area. According to the EDS regional scanning results of these inclusions, shown in Figure 6, it can be concluded that the black inclusions are titanium oxide.

Coatings
As shown in Figure 5b, it is noticeable that there is a lot of titanium oxide dispersed within the dendrites when samples annealed at 450 • C. However, when samples are annealed at higher temperatures, there is less dispersion of titanium oxide observed. On the one hand, due to the increasing temperature, the gap among the grains becomes smaller, which leads to the escape of oxides originally stored in the gaps. On the other hand, due to the more active diffusion activities at higher temperatures, more Fe atoms from the substrate diffused to the coating, which caused the formation of Fe 2 Ti and partly consumed the titanium atoms. Figure 7 shows the SEM morphology of the coating after annealing under 850 • C. 2021, 14, x FOR PEER REVIEW 8 of 16 As shown in Figure 5b, it is noticeable that there is a lot of titanium oxide dispersed within the dendrites when samples annealed at 450 °C. However, when samples are annealed at higher temperatures, there is less dispersion of titanium oxide observed. On the one hand, due to the increasing temperature, the gap among the grains becomes smaller, which leads to the escape of oxides originally stored in the gaps. On the other hand, due to the more active diffusion activities at higher temperatures, more Fe atoms from the substrate diffused to the coating, which caused the formation of Fe2Ti and partly consumed the titanium atoms. Figure 7 shows the SEM morphology of the coating after annealing under 850 °C.   As shown in Figure 5b, it is noticeable that there is a lot of titanium oxide dispersed within the dendrites when samples annealed at 450 °C. However, when samples are annealed at higher temperatures, there is less dispersion of titanium oxide observed. On the one hand, due to the increasing temperature, the gap among the grains becomes smaller, which leads to the escape of oxides originally stored in the gaps. On the other hand, due to the more active diffusion activities at higher temperatures, more Fe atoms from the substrate diffused to the coating, which caused the formation of Fe2Ti and partly consumed the titanium atoms. Figure 7 shows the SEM morphology of the coating after annealing under 850 °C.  It is worth noting that the average grain size of the coating with a holding time of 2 h is larger than that with a holding time of 0.5 h, which explains that, with the extension of the holding time, the grains would further grow after recrystallization [38]. In order to identify the relationship between the coating microstructure and phase composition, EDS was employed to analyze the element distribution within the phases, and the scanning results are shown in Table 5. Based on the EDS results of Position 2 and Position 6, marked in Figure 7, their atomic ratio of Ni and Ti is close to 1:1 so that the separate large grain can be assigned as NiTi phase. In contrast, Position 3 is rich in Ni with a Ni-Ti atomic ratio of about 3:1, which can be identified as Ni 3 Ti. With the diffusion of Fe atoms, which participated in the nucleation process during recrystallization of the coating, Fe 2 Ti phase was formed and precipitated from the Ni 3 Ti phase, where Position 4 and Position 5 were occupied, and the fraction of regions with an eutectic microstructure of Fe 2 Ti and Ni 3 Ti increased due to an accelerated Fe diffusion into the coating, especially under higher temperatures and extended holding time. In addition, the formation of Fe 2 Ti led to an increase in the atomic ratio of Ni to Ti in the coating, further promoting the precipitation of Ni 3 Ti intermetallic compound, which is more stable and has a higher melting point. Based on the EDS result of Position 8, it seemed to be Ni 4 Ti 3 . However, Ni 4 Ti 3 phase is a metastable phase, which would decompose at higher temperatures and with prolonged exposure. In addition to that, Ni 4 Ti 3 has a typical lenticular nanoscale shape [39,40]. Therefore, it was assumed to be the start phase of the decomposition inside the NiTi phase, which needs further TEM observation. Meanwhile, a small amount of Cr atoms, which diffused from the substrate, were detected within the NiTi and Ni 3 Ti phase. According to the site preference rule of ternary alloys to NiTi, Cr atoms are more likely to replace the Ti atoms in the crystal structure, since Cr has similar atom radius and electronegativity to a Ti atom [41].  Figure 8 shows the DSC curves of the 55NiTi+5Ni coating before and after annealing treatment. However, it looks different from the main publication: there is no obvious absorption peak detected during the whole heating and cooling cycle, but this does not mean that the martensite transformation has not taken place in the measuring temperature range. Table 5 shows that a few Fe atoms are detected in the NiTi phase, which substituted the Ni atoms in the B2 structure [41]. Previous studies reported that the addition of Fe has influence on the phase transformation behavior of NiTi alloys, which could suppress the first-order martensite transformation and reserve transformation. In addition, the second-order-like phase transformation from an incommensurate state to a commensurate state might take place [42,43]. In addition to that, according to the XRD results, there were more secondary phases, such as Ni 3 Ti, Ti 2 Ni and Fe 2 Ti, precipitated after annealing. The second-order-like martensite transformation might be retarded by the formation of these metallic compounds as well. Meanwhile, previous research reported that the martensite transformation temperature of the NiTi alloy decreased with the increasing content of Ni [44]. With the decomposition of NiTi into the more stable Ni 3 Ti and Ti 2 Ni by higher annealing temperatures, the content of Ni in the NiTi phase decreased. Therefore, it is supposed that the martensite transformation temperature of the coating would increase with an increase in the annealing temperature. Apart from that, the compensated heat flow for the heat-treated samples was about 0.007 W/g higher than the samples without annealing during both the heating and cooling process, which indicates that the heat-treated coating samples consumed more energy.

Coating Martensite Transformation Behavior
With the decomposition of NiTi into the more stable Ni3Ti and Ti2Ni by higher annealing temperatures, the content of Ni in the NiTi phase decreased. Therefore, it is supposed that the martensite transformation temperature of the coating would increase with an increase in the annealing temperature. Apart from that, the compensated heat flow for the heattreated samples was about 0.007 W/g higher than the samples without annealing during both the heating and cooling process, which indicates that the heat-treated coating samples consumed more energy. As there was no obvious phase transformation observed during the DSC test, the transformation temperatures and phase-transition enthalpy were assessed with TA Instruments software TRIOS. Figure 9 shows the integral calculation process based on the coating DSC curves; the transformation temperatures and heat flows were determined by the intersections of a baseline and the tangents to a thermal peak in software, and the results are listed in Table 6. The phase-transition enthalpy is closely related to the progress of the phase transformation, which is accompanied by the release or absorption of heat. The phase-transition enthalpy of the coatings after annealing is lower than that without heat treatment, which indicates that the martensite transformation is more likely to take place in the coating without annealing. In addition, the calculation result proves that the phase transformation temperature increased after the annealing treatment. As there was no obvious phase transformation observed during the DSC test, the transformation temperatures and phase-transition enthalpy were assessed with TA Instruments software TRIOS. Figure 9 shows the integral calculation process based on the coating DSC curves; the transformation temperatures and heat flows were determined by the intersections of a baseline and the tangents to a thermal peak in software, and the results are listed in Table 6. The phase-transition enthalpy is closely related to the progress of the phase transformation, which is accompanied by the release or absorption of heat. The phase-transition enthalpy of the coatings after annealing is lower than that without heat treatment, which indicates that the martensite transformation is more likely to take place in the coating without annealing. In addition, the calculation result proves that the phase transformation temperature increased after the annealing treatment.
With the decomposition of NiTi into the more stable Ni3Ti and Ti2Ni by higher annealing temperatures, the content of Ni in the NiTi phase decreased. Therefore, it is supposed that the martensite transformation temperature of the coating would increase with an increase in the annealing temperature. Apart from that, the compensated heat flow for the heattreated samples was about 0.007 W/g higher than the samples without annealing during both the heating and cooling process, which indicates that the heat-treated coating samples consumed more energy. As there was no obvious phase transformation observed during the DSC test, the transformation temperatures and phase-transition enthalpy were assessed with TA Instruments software TRIOS. Figure 9 shows the integral calculation process based on the coating DSC curves; the transformation temperatures and heat flows were determined by the intersections of a baseline and the tangents to a thermal peak in software, and the results are listed in Table 6. The phase-transition enthalpy is closely related to the progress of the phase transformation, which is accompanied by the release or absorption of heat. The phase-transition enthalpy of the coatings after annealing is lower than that without heat treatment, which indicates that the martensite transformation is more likely to take place in the coating without annealing. In addition, the calculation result proves that the phase transformation temperature increased after the annealing treatment.   Figure 10a displays the microhardness test results for the coatings along the depth, which could be clearly divided in three areas: coating cladding area, coating-substrate transition area and substrate, respectively. The microhardness distribution was generally consistent from sample to sample, indicating that the samples were fairly homogeneous over the areas measured. Figure 10b shows a hardness boxplot of the coating cladding area, the region that bears load and friction in the coating. The hardness of the coating decreased when annealed at lower temperatures (e.g., 450 • C and 650 • C), which is due to a decrease in the defect density in the coating with the release of internal stress [45]. However, coatings annealed under 850 • C exhibit a recovery of the hardness, which is even a little bit higher than samples without heat treatment. The phase segregation of Ni 3 Ti and Ti 2 Ni causes precipitation hardening, which can be attributed to the combined effect of crystallization, nucleation and grain growth [46]. This conclusion is also supported by the XRD, SEM and EDS results mentioned above.   Figure 10a displays the microhardness test results for the coatings along the depth, which could be clearly divided in three areas: coating cladding area, coating-substrate transition area and substrate, respectively. The microhardness distribution was generally consistent from sample to sample, indicating that the samples were fairly homogeneous over the areas measured. Figure 10b shows a hardness boxplot of the coating cladding area, the region that bears load and friction in the coating. The hardness of the coating decreased when annealed at lower temperatures (e.g. 450 °C and 650 °C), which is due to a decrease in the defect density in the coating with the release of internal stress [45]. However, coatings annealed under 850 °C exhibit a recovery of the hardness, which is even a little bit higher than samples without heat treatment. The phase segregation of Ni3Ti and Ti2Ni causes precipitation hardening, which can be attributed to the combined effect of crystallization, nucleation and grain growth [46]. This conclusion is also supported by the XRD, SEM and EDS results mentioned above.

Coating Wear Resistance
The surface morphologies of coatings after the friction and wear test are shown in Figure 11, and the sliding directions, marked with arrows, are clear to recognize. It is obvious to see that, coating without annealing shows more wear debris on the surface, which is due to the poor surface condition with some unfused powders, as shown in Figure 4a. Figure 11c-e are SEM surface morphologies of the coatings after annealing, which exhibit less wear debris instead. In addition, the coating surface without heat treatment is supposed more brittle, while the surface cooled in the air within a very short time after the high-power laser beam moved forwards. Nevertheless, all surface morphologies share a similar wear appearance but with different levels of severity, indicating the same wear

Coating Wear Resistance
The surface morphologies of coatings after the friction and wear test are shown in Figure 11, and the sliding directions, marked with arrows, are clear to recognize. It is obvious to see that, coating without annealing shows more wear debris on the surface, which is due to the poor surface condition with some unfused powders, as shown in Figure 4a. Figure 11c-e are SEM surface morphologies of the coatings after annealing, which exhibit less wear debris instead. In addition, the coating surface without heat treatment is supposed more brittle, while the surface cooled in the air within a very short time after the highpower laser beam moved forwards. Nevertheless, all surface morphologies share a similar wear appearance but with different levels of severity, indicating the same wear mechanism during the friction and wear test. This is considered delamination wear. Figure 12 displays the results of wear mass loss of the coatings during the friction and wear test. The coatings after annealing under 650 • C and 850 • C show obviously less wear mass loss, which might be attributed to an improvement in the ductility after recrystallization. Generally, with annealing treatment, the wear resistance of the coatings has been significantly improved with an 11% average mass loss reduction. mechanism during the friction and wear test. This is considered delamination wear. Figure 12 displays the results of wear mass loss of the coatings during the friction and wear test. The coatings after annealing under 650 °C and 850 °C show obviously less wear mass loss, which might be attributed to an improvement in the ductility after recrystallization. Generally, with annealing treatment, the wear resistance of the coatings has been significantly improved with an 11% average mass loss reduction.

Conclusions
This paper investigates the influence of annealing treatment on the microstructure, phase composition, transformation behavior and mechanical properties of 55NiTi+5Ni coating prepared via laser cladding. Based on the obtained results and analysis, the following main conclusions can be made: 1. The untreated coating mainly consists of NiTi and Ni3Ti phases with a typical dendrite structure. With the annealing temperature rising, the content of Ni3Ti and Ti2Ni visibly increased.  mechanism during the friction and wear test. This is considered delamination wear. Figure 12 displays the results of wear mass loss of the coatings during the friction and wear test. The coatings after annealing under 650 °C and 850 °C show obviously less wear mass loss, which might be attributed to an improvement in the ductility after recrystallization. Generally, with annealing treatment, the wear resistance of the coatings has been significantly improved with an 11% average mass loss reduction.

Conclusions
This paper investigates the influence of annealing treatment on the microstructure, phase composition, transformation behavior and mechanical properties of 55NiTi+5Ni coating prepared via laser cladding. Based on the obtained results and analysis, the following main conclusions can be made: 1. The untreated coating mainly consists of NiTi and Ni3Ti phases with a typical dendrite structure. With the annealing temperature rising, the content of Ni3Ti and Ti2Ni visibly increased.

Conclusions
This paper investigates the influence of annealing treatment on the microstructure, phase composition, transformation behavior and mechanical properties of 55NiTi+5Ni coating prepared via laser cladding. Based on the obtained results and analysis, the following main conclusions can be made:

1.
The untreated coating mainly consists of NiTi and Ni 3 Ti phases with a typical dendrite structure. With the annealing temperature rising, the content of Ni 3 Ti and Ti 2 Ni visibly increased.

2.
The coating microstructure transformed from dendrite structure to single grains because of recrystallization, and the recrystallization temperature of the 55NiTi+5Ni coating is assumed to be between 650 • C and 850 • C.

3.
There is no martensite transformation observed in the DSC test. The martensite transformation could be suppressed by Fe atoms diffused from the substrate and probably retarded by the numerous metallic compounds after annealing as well.
However, the software calculation results show that the phase-transition enthalpy decreased and the transformation temperature rose after annealing. 4.
The coating annealed under 850 • C shows the highest microhardness due to the precipitation hardening, and the wear resistance was also improved after an annealing treatment with an 11% average wear mass loss.
The research provides a reference for using blend powder to replace the rare 60NiTi alloy powder in the preparation of Ni-rich NiTi alloy coating and proved the possibility of further improving the coating performance via annealing treatment. However, since the phase transformation temperature of the Ni-rich NiTi alloy is very sensitive to the content of Ni and the addition of Fe in the alloy, additional DSC tests are needed for further investigation.