Figure 3 shows the morphology of the powders used, obtained by scanning electron microscopy. Figure 3a) shows pores and shells for 1559-40 alloy. Figure 3b) shows some f satellite particles for 1560-00 alloy. Figure 3c) shows an irregular morphology for Rockit alloy 401. Figure 3d) shows WC 4590 ceramic reinforcement with regular spherical morphology.
3.2 Surface morphology
3.2.1 Single beads morphology
The parameterization used for the deposition of the single beads (Table 3) resulted in different geometric characteristics, where, as expected: with the increase in the scanning speed, beads of lower height were obtained, and keeping the scanning speed fixed and increasing the power, higher dilution values of the coating on the substrate were obtained.
With a power of 1000 W and a scanning speed of 15 mm/s, the beads presented lower porosity, high microhardness, total adhesion to the substrate, geometry with good wetting and low dilution, as shown in Fig. 4. For this preliminary single bead deposition test, reinforcement containing 20 and 30% Vol. WC were used only for the nickel-based alloys, which resulted that the hard WC particles significantly influenced the generation of porosity in the metallic matrix, especially when the proportion of reinforcing element was 30% by volume.
This behavior was also evidenced by ZHANG et al. [16]. The authors observed that by increasing the WC-12Co content to 30% by weight in the Ni-Cu matrix, the pores in the coating layer increased significantly. ZHOU et al. [17] and TEHRANI et al. [18] explain that during the laser cladding process, excessive heating causes the WC particles to partially dissolve and produce carbon, causing a reaction with a small amount of oxygen to generate CO and CO2. As laser cladding is a fast solidification process, the gas does not have enough time to escape from the molten pool resulting in pores in Ni/WC MMCs.
It was also possible to see that the increase in power (see Table 3) resulted in a marked increase in the internal porosity of single beads containing 30%Vol.WC. This phenomenon can be explained due to the fact that the greater the power/energy delivered by the laser, the greater the absorption of thermal energy by the powder particles. As the WC particles have a much higher melting point than the metallic matrix, they enter the melting pool very hot, taking longer to cool down. Therefore, they act as a source of heat within the bead. In the cooling process, the bead as a whole cool down, either by heat transfer with the substrate or by convection with the external environment, however the inner part of the bead takes longer to cool down, causing the ceramic particles to continue boiling inside of the bead, dissipating heat, and generating vaporization of a portion of the matrix, which generates the pores.
The behavior of increased energy density generating greater internal porosity was also evidenced by WENG et al. [7].
3.2.2 Cracks in coatings
After the sample grinding process, some cracks on the surfaces were observed. To visually identify the cracks on the surfaces, liquid penetrant tests were carried out. Figures 5 and 6 show the grinded samples and the samples with revealed cracks, respectively.
It was possible to identify cracks along the entire surface of the pure coatings (0 Vol.%WC), as well as for the composites with addition of WC (20 Vol.%WC and 30 Vol.%WC), except for the named sample I, Rockit 401 alloy with 0 Vol.%WC. Therefore, for samples J and K, the cause of the cracks were the WC particles. The authors WANG, LI & TAO state that cracks can easily form when the WC volume fraction exceeds 9.1% in iron-based metallic matrices. For low volume fractions of WC, the stress would not be enough to generate cracks in WC/Fe composites [10].
In all cases, the cracks had a growth behavior perpendicular to the laser scanning direction during the deposition of the first two beads, as shown in Fig. 7a). In the following beads, the growth behavior occurred randomly, but tending to the perpendicularity of the laser beam deposition strategy and following the previous cracks, as shown in the illustration in Fig. 7b). At the center of the coatings, it is possible to see that the intersection of larger and smaller cracks occurs, due to the alternation of residual stresses resulting from contraction stresses and thermal stresses in the coating during the cooling process.
Along the deposition, closer to the final deposited region, on the right side, it is possible to observe that the cracks grow back in a perpendicular direction to the laser scan and the smaller cracks that crossed each other decrease. This behavior must have occurred because the substrate was heated by previous depositions, resulting in a higher temperature at the end of the deposition process. This same crack formation behavior was also observed by the authors SHI et al. in a nickel-based alloy [19]. These authors claim that the crack can be prevented by the combination of preheating to 300°C and adding insulating plate. Adding insulating plate can play a similar role to furnace-cooling in terms of crack control.
These results and preliminary conclusions can be justified by the various scientific works on preheating of Ni-based coatings, such as the work of the authors SADHU et al. [20]. The authors conclude that the use of substrate preheating effectively controls thermal gradients, resulting in crack-free multilayer deposition. Furthermore, with decreasing scanning speed and increasing substrate preheating temperature, the cooling rate decreased, thus, residual stresses decreased with decreasing cooling rate [21].
ZHOU et al. state that the crack sensitivity of nickel-based coatings is related to the scanning speed and the laser power used, where the number of cracks increases with the increase in laser power and scanning speed. The authors also claim that after absorbing the laser energy, the metal-ceramic composite powders and the substrate surface are fused together, and then solidified into the composite coating. It is this solidification that leads to shrinkage stress. Thermal stress is induced by the high temperature gradient during laser cladding [17].
Similarly, SHI et al. state that due to out-of-equilibrium solidification in laser plating, the high temperature gradient leads to great thermal stress. Meanwhile, there is a large amount of hard Cr-rich precipitates and eutectic structures around the γ(Ni) grain boundaries, which increases the hardness of the coating but weakens the strength of the grain boundaries. These microstructures provide easy routes for crack creation and propagation [19]. Therefore, the Ni60A alloy coating prepared by laser cladding has a high susceptibility to cracking, so it can be said that the 1560-00 coating has a similar behavior because it has a chemical composition very similar to the coating reported by the authors.
WANG, LI & TAO use iron-based 316 stainless steel as a matrix and state that cracks can easily form when the WC volumetric fraction exceeds 9.1%. For low volume fractions of WC, the stress would not be enough to generate cracks in WC/Fe composites [10]. This behavior is like the MMC composed of the Rockit 401 alloy, also based on iron in this study.
Finally, BANAIT et al. state that the coefficient of thermal expansion of stainless steel 316L is almost twice that of Ni-Cr-B-Si and the thermal conductivity of Ni-Cr-B-Si is considerably higher than 316L, so these thermophysical properties also influence the generation of flaws in coatings deposited by laser cladding [22]. As an alternative to avoid cracks, the authors preheated the substrate to 400°C to obtain a slow cooling during the process.
Figure 8 shows the morphology of the cross-section obtained by MO of the MMCs as named in Table 2. For the MMCs an excellent distribution uniformity of the WC particles spread over the two layers of the deposited coating was obtained. In general, all the coatings showed some vertical cracks except sample I, whose matrix was the pure Rockit 401 alloy. WC volumetric fraction develops pores of larger diameter when using energy density of 55.6 J/mm². This behavior was also evidenced in the depositions of single beads in this study, where the increase in the volumetric fraction is directly related to the increase in porosity, mainly when power densities from 133.33 to 233.33 J/mm² were used.
WANG, LI & TAO report that cracks of this nature appear after the deposition of 5 layers due to cyclic thermal stresses when using an energy density of 27.27 J/mm² (P = 300 W, V = 0.3 m/min D = 2.2 mm) and 72.72 J/mm² (P = 800 W, V = 0.3 m/min D = 2.2 mm), when using an MMC containing 28.6%Vol.WC and 16.7%Vol.WC respectively [10]. In this study, the authors used a 316L stainless steel metallic matrix on an 8 mm thick 304 stainless steel substrate. These results, compared to the present study, show that the higher the energy density used, even decreasing the volumetric fraction of WC, the greater the probability of occurrence of cracks on the surface of the MMC coating. However, it is important to point out that a minimum energy density must be known for powder melting, used to obtain good coatings.
Figure 9 shows cracks in the cross section of sample H, which cross the hard WC particles, where this behavior was seen in all MMCs studied. Other authors report a similar behavior [10, 17].
LIU et al. report that there are two types of cracks, those that cross the WC particles and those that occur between the interface of the matrix with the WC particles [23]. This second case was not observed in the present study, demonstrating a strong connection between the metallic matrix interface and the reinforcing particles during the laser cladding process.
Figure 10 show that the WC particles obtained an excellent distribution in the metallic matrices of the coatings, because of a good methodology for mixing the powders and the use of an adequate process parameter. In addition, it can be seen in this figure that the WC particles were well adhered to the matrix, showing up as points of greater resistance to abrasive wear. WU et al. state that the distribution of ceramic particles in laser ceramic-metal composite coatings is a very important factor, which affects not only the wear resistance properties, but also the quality and processing of the coatings [2]. Unlike this study, another reinforcement particle distribution behavior can occur. Other research shows that because the WC particles have a high density in relation to the matrix and a high melting point, they tend to sink towards the substrate [2, 5, 6, 9], this also happens due to the Marangoni convection currents unable to act efficiently so that this unwanted behavior is avoided.
3.3 Microhardness
3.3.1 Microhardness of single beads
The average microhardness values of the single beads, referring to the metallic matrices, are shown in Fig. 11. It is possible to notice that despite using different process parameters, the microhardness values were close, taking into account their respective standard deviations, except for the parameter of 800 W and 5 mm/s for the Rockit 401 alloy, which showed a high value.
It is worth noting that the dilution factor did not alter the microhardness values, as this only occurs in a more representative way for the power of 1400 W. Therefore, for this power used in the MMCs, a decrease in values due to the dilution generated between coating and substrate.
SHI et al. report that the large line energy (ratio between laser power and scanning speed (J/mm)) and low powder feed rate are helpful to prevent crack but lead to low microhardness [19]. The authors also state that the increase in line energy would produce a decrease in the solidification rate of the molten pool. Therefore, the higher energy of the line helps to increase the bubble rise time and the formation of slag that are generated by the Si and B elements in the molten pool, which can reduce coating defects. In addition, the higher line energy would extend the high temperature dwell time, which would increase the high temperature plastic flow and effectively reduce the residual stress of the coating. Also, the dilution rate increases due to more energy obtained by the substrate, and the increase in the Fe element inhibits the hard precipitates of chromium borides and chromium carbides, which improves the crack resistance of the Ni60A coating.
3.3.2 Microhardness of coatings
The average microhardness values of the metal matrix composite coatings along with the proportions of WC as a reinforcing element are illustrated in Fig. 12. The coatings named by 0%Vol.WC*, were deposited on a 1020 steel substrate, to compare with 304 stainless steel substrates.
It is possible to notice that there was a change in the microhardness of the individual beads compared to the coatings. For the Rockit 401 alloy there was an increase of 10% (485.6 HV and σ = 23.0 to 531.3 HV and σ = 19.6). Conversely, for alloy 1559-40 there was a 10% reduction in values (667.8 HV and σ = 15.2 to 614.7 HV and σ = 29.6). Alloy 1560-00 showed similar results in the comparison (857.1 HV and σ = 13.9 to 852.7 HV and σ = 36.2).
It was observed that the average values of microhardness increase with the addition of WC inserted in the metallic matrix, as well as its standard deviations. This fact occurs due to the heterogeneity of the coating layer. During indentation, some of these were made on hard tungsten carbide particles that can reach values greater than 2500 HV1. ZHANG et al. recommend that the indentations should be made deviating from the WC reinforcement particles, with the test being performed only on the nickel-based matrix [16]. They noticed that when 30% by weight of WC-12Co was used on the Ni-Cu matrix, there was a 20% increase in the microhardness value when compared to the matrix in its pure state.
GUO et al. state that the NiCrBSi/WC-Ni composite coating has a higher microhardness than the coating with NiCrBSi as the metallic matrix, due to the formation of hard ceramic phases including WC and W2C in the composite coating [6].
BANAIT et al. [22] and SOUSA et al. [24] also found an average value of 645.2 HV1.96 and 580 HV1 for the Ni-Cr-B-Si coating deposition respectively, in comparison to alloy 1560-00 in this study. This discrepancy is mainly due to the higher carbon content present in the 1560-00 alloy.
There are discrepant cases, such as found by DESCHUYTENEER et al. [11]. In this study, the authors obtained lower hardness values with the use of 10 and 20%Vol.WC when compared to the pure matrix of NiCrBSi, and only with 30%Vol.WC the hardness value surpasses the pure matrix. Throughout the work, the authors describe that the microstructure of the coating and thermal input may have led to this result.
3.4 Volumetric loss
The volumetric loss in the time intervals of 10, 20 and 30 minutes of the samples coated with alloys 1559-40, 1560-00 and Rockit 401, as well as only the materials used as substrates, respectively, are presented in Figs. 13, 14, 15 and 16.
The volumetric loss during the first 10, 20 and 30 minutes occurred linearly, showing a linear correlation trend between 0.98 and 1. With these results, it is possible to see that the coating showed the same resistance to abrasive wear at 3 bodies, considering differences in coating density, porosity and, in some cases, the appearance of cracks.
Figures 13 and 14 show that, despite changing the substrate material, such as 1020 steel and 304 stainless steel, for 1559-40 and 1560-00 alloys respectively, using the matrix in its pure state (0%Vol.WC*), there was no significant change in volumetric loss. It was also observed that the variability of the standard deviation values of each material were similar.
Figure 17 shows a comparison of the volumetric loss for the three different MMCs, as well as for the substrate. As mentioned in the methodology of this work, for the MMCs containing 1559-40 and 1560-00 alloys, stainless steel 304 was used as substrate and for the alloy Rockit, carbon steel 1020.
The higher volumetric loss of Rockit 401 and 1559-40 alloys in its pure state is evident, that is, without the addition of WC, obtaining approximate values of 515.3 and 488.5 mm³ respectively, when compared to the 1560-00 alloy with 79.60 mm³. The amount for 1020 steel and 304 stainless steel substrate materials was 169.61 mm³ and 150.58 mm³, respectively.
Abrasive wear resistance was directly proportional to the microhardness values of the MMCs, that is, with higher microhardness values presented by alloy 1560-00, less material was removed. The same was noticed for the other alloys. This behavior would be related because this alloy has elements such as chromium and a high carbon content in its chemical composition, which results in the formation of chromium carbides, as well as a matrix of high hardness and wear resistance.
Figure 18 shows in more detail the volumetric loss of the MMCs, for 20 and 30%Vol.WC, for the three alloys used. The addition of WC as a reinforcing element in metallic matrices provides a reduction in the volumetric loss on the surface. In the case of Rockit 401 alloy, there was a reduction of approximately 89,2% and 90,7%, compared to the matrix in its pure state, containing 20 and 30%Vol.WC, respectively. Using 1559-40 alloy, there was a reduction of 92.5% and 95% while for 1560-00 alloy, there was a reduction of 79.1% and 85.8%, compared to the matrix in its pure state, containing 20 and 30%Vol.WC, respectively.
In the literature, there is not enough information about the tribological behavior of nickel-based alloys, especially when subjected to abrasive tests. SOUSA et al. deposited alloy 1545-00 based on Ni-Cr-B-Si by laser cladding and obtained a volumetric loss at the end of the test of 0.32 mm³ and 0.20 mm³ when the deposition trajectory was parallel and perpendicular to the ASTM G65 test, respectively [25]. The sweep trajectory has a significant influence on the volumetric loss.
SILVA et al. [26] used Inconel alloy 625 with a nickel-based composition, but containing molybdenum and a higher chromium content, when compared to alloy 1560-00 in this work. In the ASTM G65 test, the authors obtained a volumetric loss of 41.5 mm³, practically half when compared to alloy 1560-00 in its pure state. This behavior confirms that the presence of different chemical elements, such as Mo and Cr, directly interfere with wear resistance. It is also noteworthy in this work that the microhardness values found by the authors were only 247 HV.
Another way to evaluate wear results is with weight loss. Figure 19 shows the weight loss at the end of the ASTM G65 abrasive test, in order to compare the results with other authors. For example, NURMINEN, NÄKKI & VUORISTO [8] used the modified ASTM G65 standard in their study with test times of 1 hour among other modifications. The authors used different matrices and reinforcement elements for the MMCs. For the alloy based on NiCrBSi40 and NiCrBSi60 containing 30 and 50% by volume of WC particles of spherical and angular morphology, the authors obtained weight losses of approximately 100 to 130 mg. The best results found for this MMC were approximately 100 mg for NiCrBSi60 + 30%Vol.WSC and NiCrBSi40 + 50%Vol.WSC MMCs (WC with spherical geometry in this case). It is then noticed that spherical reinforcement particles present better abrasive results and that for the same proportion of 30%Vol. of spherical particles, the NiCrBSi60 alloy (higher percentage of carbon and chromium) presents less weight loss when compared to the NiCrBSi40 alloy.
In the present work, the abrasive test lasted 30 minutes, as recommended by the ASTM G65 standard. Even considering twice the weight loss value obtained for alloy 1560-00, which has a similar chemical composition to the alloy in the study by NURMINEN, NÄKKI & VUORISTO [8], the 1560-00 alloys containing 20% and 30%Vol.WC gave better results.
3.5 Wear surfaces
Figure 20 show the worn surfaces (to the same scale) for alloys 1559-40, 1560-00 and Rockit 401 in their pure state, 20%Vol.WC and 30%Vol.WC. In general, the highest volumetric losses occurred in coatings whose metallic matrix in its pure state presented lower microhardness values, directly influencing the wear mechanisms generated on the surfaces of coatings. Figure 20 also shows that the WC particles obtained an excellent distribution in the metallic matrices of the coatings and it can be seen that the WC particles were well adhered to the matrix, showing up as points of greater resistance to abrasive wear.
Figures 20a), 20 d) and 20 g) show the action of abrasive wear mechanisms on the surfaces of coatings in their pure state 0%Vol.WC. In these tests, it was observed that the action of deep parallel grooves predominated in macroscale, extending throughout the entire wear range. This behavior was very pronounced for Rockit 401 and 1559-40 alloys, which can be explained by the similarity of volume loss values and microhardness values of both alloys in the pure state (0%Vol.WC). Only at 1500x magnifications is it possible to see that the action of the wear mechanisms changes to microplowing and macrocutting, which is detailed in the zoom on Figs. 20a), 20 d) and 20 g).
In a way, all abrasive wear mechanisms cited so far were found in pure coatings, what changes is the intensity of action on the surface. The coating 1560-00 0%Vol.WC presented a less aggressive and more homogeneous surface compared to the other coatings under analysis. This result is related to its greater microhardness and wear resistance, providing less abrasive penetration and material removal during the test.
WC particles maintain their spherical morphology without major damage caused by the abrasive sand particles in the test. An important fact that was observed is that the WC particles interrupt/inhibit the wear mechanisms, and consequently inhibit the removal side effects that were mentioned in the pure matrix because they are so resistant, working as material removal barriers. WU et al. report that the hard phases of WC in the matrix 10.0 Cr, 2.1 B, 2.8 Si, 0.1 C, 14.0 Fe and 71.0 Ni. formed a load-bearing system that greatly reduces the load applied to the abrasives and limits the micro-cutting process [2].
It is noteworthy that in the present study, the cracks pass through the WC particles, as already mentioned, but they do not interfere with a greater volumetric loss, as they do not loosen the metallic matrix as in other studies. SOUSA et al. report that for coatings based on NiCrBSi, the greater the power used, consequently the greater is the dilution [24]. In this study, coatings showed lower microhardness and greater volumetric loss, so that the SEM analyzes revealed detachments of the coating material in the cracks and regions close to them, increasing the volumetric loss.
Figure 21 shows, in greater detail, the wear behavior of MMC composed of alloy 1560-00 with 20% WC by volume. It is possible to observe that there is a periodic wear profile on the worn surface with a characteristic parallel to the direction of the ASTM G65 test. SOUSA et al. studied this aspect, concluding that when the direction of the deposition of the beads is parallel to the direction of the tribological test, the removal of the coating is facilitated, since the sand tends to accumulate in regions of lower resistance, removing a greater amount of material [27]. These regions of lower resistance can be originated due to variations in the thermal input that the laser generates when depositing one bead over the other, generating a gradient of microstructure and microhardness in the matrix.
SOUSA et al. report that when the direction of deposition is perpendicular to the direction of the test, the beads form an intermittency in the indentation of the sand, preventing the sand from accumulating, making the test less severe. The authors obtained a volumetric loss of approximately 37% less when using the perpendicular direction of deposition in relation to the ASTM G65 test [27].
Furthermore, in the Fig. 21, in the small space where there are no WC particles, plastic deformation of the matrix occurs, causing greater loss of volume of the material due to the action of the test abrasive. The same behavior is observed for the MMC containing 30%Vol.WC (Fig. 22), but to a lesser extent, due to the greater concentration and distribution of reinforcement particles, since these work as inhibitors of the severe abrasive action on the metallic matrix of lower hardness. The cracks formed (Fig. 22a)) in the coating do not interfere in such a way as to accentuate the volumetric loss during the test.
A very similar metal matrix removal behavior as described above for the 1560-00 alloy occurs for the 1559-40 alloy. Figure 23 shows the MMC for matrix 1559-40 containing 30%Vol.WC, where the greatest material removal occurs in small regions, where the metallic matrix is exposed to abrasive material, as highlighted in red in Fig. 23a. In this way, the increase in the content of reinforcing particles improves resistance to abrasive wear and keeps the matrix protected. However, the hardness and tenacity of the metallic matrix influence the way in which the matrix material is removed, as mentioned by WU et al. [2], in which in this case it is characterized by adhesive wear on the metallic matrix (Fig. 23a). This behavior was also observed in the pin on disc test by other authors [9].
The WC particles attached to the metallic matrix are so well anchored and fixed, as a result of a good chemical bond and mechanical interlocking between matrix and reinforcement [1], in which their wear occurs slowly and gradually, presenting a surface of morphology smooth and very resistant (Fig. 23b)).
This described behavior was also perceived by PENG et al. in the coating based on FeCoCrNi with 20%wt. of WC [28].
In the MMC, whose metallic matrix was the Rockit 401 alloy, it is possible to perceive by the Fig. 24, that it was shown to be more ductile by “wetting” the WC particles during the abrasive test, mischaracterizing the circular shape that was very remarkable for the other metallic matrices the nickel base of this study. Likewise, the WC particles also served as points of greater resistance to wear. However, as the matrix has less hardness and greater ductility, wear occurs by removing the matrix together with the inserted hard WC particles. Thus, the wear mechanism is characterized as adhesive wear, forming large areas of severe material removal. Given this fact, the lower resistance and greater volumetric loss during the ASTM G65 test are justified.
RAAHGINI & VERDI state that abrasive wear mechanisms may involve plastic flow and brittle fracture. Under some conditions, only plastic flow can occur. However, both often occur together, even in regularly thought or ideally fragile materials [1]. In this study, the authors use the In625 alloy as a matrix, where they state that the splintering of the matrix may be the result of the plastic deformation of the matrix caused by the abrasive particles of the ASTM G65 test.
It is noteworthy that the Rockit alloy 401 and 1559-40 with 0% reinforcing element, presented very similar volumetric loss of 515.3 mm³ and 488.5 mm³ respectively. However, when 30%Vol.WC was used, the 1559-40 alloy showed almost 2 times less volumetric loss in the ASTM G65 test, due to the anchoring performance of the hard WC particles in the nickel matrix.
Another factor that influenced the reduction of wear was the presence of dark regions containing oxides in all MMCs, which acted as a protective film. Figure 25 shows the EDS analysis of some regions of the worn surface of samples D, K and H where there is the presence of Si and O, elements from the sand used in the abrasive test, which were impregnated on the surface of the coating. Other authors also found similar results indicating layers of oxides with high Si and O contents in dark regions in coatings [9, 24].
RAAHGINI & VERDI also observed the presence of these Si rich particles in MMC [1]. The authors claim that this type of occurrence, where contamination and debris or inclusions of interacting bodies can be transferred and incorporated into the analyzed surfaces, result in the formation of a “mechanically mixed layer” where this phenomenon may have contributed to the wear behavior of the material. material of interest.