3.1. Phase Composition
Figure 3 provides the phase compositions of composite coatings with and without ultrasonic vibrations. Based on the figure, the coating phase consists of FeNi
3, NbC, B
4C and CrB
2, indicating a similar phase composition whether ultrasonic vibrations are applied or not. Ultrasonic vibrations do not affect the phase compositions of in situ synthesized NbC and cladding layers.
Coatings were analyzed by an EDS test to identify the phase morphology (see
Figure 4 and
Table 4).
Figure 4a indicates three kinds of color particles—silver flower-like, black flocculent, and gray cubic particles. The content of Nb atom in the silvery white material at point A is 52.03%, and the content of C atom is 36.85%. Combining with the XRD results, it is speculated that it is an in situ synthesized reinforcing phase NbC. The content of Cr element in the gray massive material at point B is up to 40.92%, containing relatively high C and Fe, which is presumed to be the strengthening phase of Cr compound. The black area of point C is rich in C, and the content reaches 54.06%. The C in the coating mainly comes from the decomposition of B
4C, which is presumed to be the residual C atom of the reaction.
Due to the big difference in the grain morphology in coatings, the investigation of the grain-forming factors can reveal the solidification mechanism of the molten pool.
Figure 5 demonstrates that during the molten pool solidification, grain morphologies depend on the molten pool temperature gradient (
G) divided by grain growth rate (
R). The grain size is determined by the product of
G and
R. The morphology of grains changes with the decrease in
G/R value, and the grain size is reduced as the cooling rate increases.
Figure 6 shows the grain growth at the coating bottom. The heat diffuses rapidly to the substrate with a high
G/R value at the coating bottom (namely the dilution rate area). The high
G/R ratio leads to the plane growth of grains at the bottom of coatings, creating plane crystals. The planar crystals nucleate at the fusion line of the substrate and grow toward the internal coating along the opposite direction of the heat diffusion [
23]. The growth rate of grains in the middle of the dilution rate region increases with decreasing
G/R value, leading to grain transformation from planar to cellular crystals. Grains gradually grow into columnar dendrites with a further decreased
G/R value. The
G/R value determines the grain size, which gradually decreases as the cooling rate accelerates [
24]. It is known that cavitation bubbles generated by ultrasonic vibration will produce instantaneous high temperature and high pressure on the surrounding solution at the moment of collapse, increasing the local temperature gradient (
G). This gives rise to an increase in
GR and a decrease in grain size. So, ultrasonic vibration can refine grains [
25].
where Δ
T is the subcooling degree; Δ
Gv is the free-energy difference;
Tm is the melting point; and
Lm is the latent heat of crystallization. Based on Equation (2), the free-energy difference increases with the increased subcooling degree. The higher free-energy difference increases crystallization driving forces and accelerates crystallization rates. Therefore, the equation suggests a relationship between nucleation rate and subcooling degree.
3.2. Influences of Ultrasonic Vibrations on the Pore Area
The experimental results were analyzed by a fitting regression model.
Figure 7 provides the normal probability of the pore area. The residuals have an S-shaped distribution around the predicted line, indicating a normal distribution of the data. Equation (3) shows the mathematical model of the pore area according to the regression analysis.
Table 4 outlines the results of the variance analysis.
According to
Table 5, DF is the amount of information in the data, which is used by analysis to estimate the value of unknown population parameters. The adjusted sum of squares (Adj SS) is a measure of variation in different components of the model. The adjusted mean square (Adj MS) measures the variance of an item or model interpretation. The F value is the test statistic used to determine whether the item is associated with the response.
p value is a probability used to measure the evidence that negates the original hypothesis.
The
p value is less than 0.001, indicating a high precision of the model. As the indicators of the model precision, R
2, R
2adj and R
2pre are closer to 100%, suggesting higher fitting accuracy of the model with smaller errors. The high fitting accuracy is also exhibited by the slight difference between R
2adj and R
2pre with less than 20%. Based on the significance level analysis, the process parameters are significant when the
p value is lower than 0.05.
Table 5 displays that UP, F, and the interaction items between UP and F have significant effects on the model.
Figure 8a,b are the surface and contour plots among the pore area, ultrasonic power, and frequency. The pore area decreases and then increases with increased ultrasonic power and frequencies. The pore area is minimized at approximately 800 W ultrasonic power and approximately 32.5 kHz amplitude.
Figure 9a,b present the coating morphologies with and without ultrasonic vibrations under the same process parameters (LP = 1700 W, SS = 4 mm/s).
Figure 9a indicates that coatings without ultrasonic vibrations have cracks and pores with an 11.75% porosity rate. B
4C is decomposed in the formation of the molten pool to produce C atoms. C is oxidized to generate gas during laser cladding. As the molten pool is solidified, part of the bubbles cannot escape in time due to the rapid solidification speed, leaving pores in the cladding layer.
Figure 9b demonstrates that the coatings under ultrasonic vibrations do not have large pores with a 4% porosity rate. Lower porosity suggests that the cavitation effect caused by ultrasonic vibrations can facilitate gas escape in the molten pool, thus eliminating coating defects.
Based on
Figure 10, the solution is torn by the local transient negative pressure during the negative pressure phase of ultrasonic vibration. A low-pressure cavitation bubble is formed in the molten pool. When the ultrasonic positive pressure phase is applied, the pressure around the cavitation bubble reaches the threshold of the pool tension. Consequently, the cavitation bubble rapidly shrinks and collapses. At the moment of cavitation bubble collapse, instantaneous high temperature and high pressure will be generated in the surrounding local area. The huge pressure gradient due to high temperature will create micro-jet around the bubble to form local excitation wave and shatter the surrounding bubbles. Therefore, ultrasonic vibration can effectively reduce the pore area of the coating [
26].
With the further increase in ultrasonic power, the enhanced cavitation effect will produce large volume of cavitation bubbles in the coating, according to Stokes’ Law (see Equation (4)).
where
F is the friction between fluid and particle;
R is the radius of the sphere;
v refers to the velocity of the sphere relative to the liquid;
η is the viscosity coefficient of the liquid. The increase in pore radius leads to a higher resistance for pores to escape from the coating. Laser cladding has a high cooling rate, resulting in a large number of pores remaining in the coating and an increase in pore area. Therefore, a proper increase in ultrasonic vibration power and frequency can reduce coating defects, while an excessive increment will generate more coating defects [
23].
3.3. Influences of Ultrasonic Vibrations on Hardness
The fitting regression model is adopted in this research.
Figure 11 presents the normal distribution of the experimental results. The residual value is fitted around the prediction straight line with an S-shape distribution, suggesting the normal distribution of the data and the reliability of the model [
24]. Equation (5) provides the hardness mathematical model obtained by regression analysis.
Table 4 lists the results of the variance analysis.
The
p-value is less than 0.01, indicating that the model has a high precision (see
Table 6). R
2, R
2adj, and R
2pre are more approximate to 100%, suggesting higher fitting accuracy of the model with smaller errors. Less than a 20% difference between R
2adj, and R
2pre implies remarkable fitting accuracy of the model.
Table 6 demonstrates that UP, F, and the interaction items of UP and F significantly affect the model.
Figure 12a,b provide the surface and contour plots of coating hardness affected by ultrasonic power and frequencies. Hardness increases and then decreases with increased ultrasonic power and frequencies.
Figure 13a,b present the microstructure of the junction between the cladding layer bottom and the substrate with and without ultrasonic vibrations.
Figure 13a indicates that the grains at the bottom of the coating and the substrate are mainly planar and cellular crystals. As the temperature gradient decreases, most grains are transformed into dendrite crystals with unmelted B
4C particles remaining in the coating. Based on
Figure 13b, the cellular crystals in the coatings significantly reduce in size with a finer distribution under ultrasonic vibrations. The cavitation bubble generated by ultrasonic vibration will form a high-speed shock wave at the moment of collapse. The bottom dendrites are ruptured by the instantaneous spatiotemporal bubbles to produce the impact force to destroy the primary dendrites and form more nucleation points. Further, bubbles created by the cavitation effect absorb a large amount of heat from the surrounding molten pool in the expansion process, increasing the undercooling of local areas. A large surface tension gradient strengthens solution flow in the molten pool, causing the dendrites to break. The dendrite fragments act as the nuclei of new grains [
27,
28,
29] to grow more fine grains.
The acoustic flow effect generated by ultrasonic vibrations promotes thermal convection and accelerates the cooling of the molten pool, causing a decrease in grain size [
30]. According to the Hall–Petch relationship, the finer grains create a larger total area of grain boundaries. More dislocations are accumulated at the grain boundaries during deformation to produce higher dislocation resistance, thus increasing the coating hardness. Equation (6) shows the grain-boundary strengthening mechanism (Δ
σgb) [
31].
where
k is the material constant;
d is the average particle size. The grain-boundary strengthening ability negatively correlates with particle sizes. Grain refinement enhances the coating strength and hardness, as fine grains have large boundary areas to hinder grain dislocation.
According to
Figure 13, the number and size of black unmelted B
4C particles significantly decrease due to the ultrasonic cavitation effect that can accelerate the melting of unmelted particles [
18]. Unmelted B
4C particles decrease, followed by morphology from irregular polygons to spherical, eliminating the agglomeration of nano B
4C powders in the molten pool to a certain extent. Due to the diffusion attenuation of ultrasonic waves, this research applies ultrasonic waves at the substrate bottom. As ultrasonic waves propagate in the molten pool, resistance in the pool causes energy loss during ultrasonic wave propagation. Therefore, the grain morphology does not change evidently at the coating top [
32,
33] (see
Figure 14).
Excessive ultrasonic vibrations create the cavitation effect to produce more energy for coatings, slowing the cooling rate of the molten pool for sufficient grain growth in the molten pool. The strengthening of fine grains disappears under the Ostwald Ripening effect [
34]. Coarse grains develop to reduce coating hardness. Therefore, the cavitation effect refines grains to increase hardness. However, when the cavitation effect reaches saturation, high temperatures and high pressures are generated to grow grains. The coarse dendrites reduce the total surface area of the grain and subgrain boundaries, weakening the ability of the coating microstructure to resist dislocations and decreasing the coating hardness.
3.4. Influences of Ultrasonic Vibrations on Wear Resistance
Figure 15 is the normal distribution of wear resistance analysis. The data conform to the normal distribution as the residual value is fitted around the prediction straight line with an S shape. The reliability of the model is established. Equation (7) provides the mathematical model of wear volume by regression analysis, and
Table 6 lists the results of the variance analysis.
Based on
Table 7, the
p-value of less than 0.01 indicates a high model precision. R
2, R
2adj, and R
2pre are closer to 100%, suggesting that the model has higher fitting accuracy with smaller errors. R
2adj, and R
2pre have less than a 20% difference, implying the remarkable fitting accuracy of the model.
Table 7 shows that the model is influenced significantly by UP, F, and the interaction items of UP and F.
Figure 16a,b provide the surface and contour plots of the interaction between the coating wear volume and ultrasonic power and frequency. The wear volume increases with increasing ultrasonic power and frequencies.
Figure 16 explores the impact of the interaction between ultrasonic power and ultrasonic frequency on the coating wear resistance. However, when ultrasonic power is less than 800 W, the interaction does not affect the wear resistance. To better understand the effects of process parameters, a single-factor method is adopted to explore the individual influence of ultrasonic power and frequency on wear resistance. As shown in
Figure 17a, COF (Coefficient of friction) value is an important index to evaluate the wear resistance of the cladding layer. The increase in COF diminishes the wear resistance of the cladding layer. The red, green, and blue areas in the figure indicate the COF values without ultrasonic vibration, at 700 W ultrasonic power with varying frequencies, and at 28 kHz frequencies with varying ultrasonic power, respectively. The COF decreases and then increases as ultrasonic frequencies and powers increase.
Figure 17b provides the wear volume of coatings, similar to the COF value. The optimal wear resistance of coatings is obtained either at 28 kHz frequencies or at 700 W ultrasonic power. The cavitation effect of ultrasonic vibration refines grains to enhance the coating strength and improve the wear resistance of the coatings. So, the wear volume is minimum.
The cavitation effect of ultrasonic vibration plays the role of grain refinement, which improves the strength of the cladding layer, while the fine and dense grains resist the pressure of the friction pair and increase the coating wear resistance.
Figure 18 shows 3D wear morphologies of coatings at ultrasonic powers of 600, 700, 800, and 900 W, respectively.
Figure 18a indicates that the coating has a small wear volume but with pores. The coating in
Figure 18b has minimum wear volume and depth. The wear width and depth increase significantly with further increased ultrasonic power, as shown in
Figure 18c,d. The maximum wear width and depth are provided in
Figure 18d.