3.1 Coating Microstructure
Figure 3 shows the polished cross-sectional microstructure of the cold-sprayed WC-Ni coating deposited using the optimal spraying parameters. As observed in the image, the interface between the coating and substrate does not contain any porosity and is well-adhered to the substrate material. Good adhesion and dense coatings can be attributed to the nature of the cold spray procedure as it is well documented for producing very dense coatings [26], [41]. There appears to be good retention of WC in the coating; however, the WC phase seems to be distributed non-homogeneously, which can lead to varying hardness at different locations in the coating. Similar morphology has been found by other researchers for cold-sprayed WC-Ni coatings [26, 42].
The porosity of the coating deposited using the OSP was quantified using image processing by analyzing a minimum of 4 different SEM images taken from the coating cross section. It was found that the average cross-sectional porosity of the coating was 0.98 ± 0.14%, which is significantly smaller than the surface porosity of 6.8 ± 0.6% for the OSP coating sample. The porosity measurement for the OSP coating is lower compared to the other coatings as illustrated in Figure 4. During the spray process, as the feed rate of the powder increases, the hard particles often rebound from the coating surface leaving surface defects that contribute to the higher surface porosity. A similar trend was found by Melendez and McDonald, where increased hard particle content lead to decreased deposition rates due to hard particle rebounding during impact [23]. The effect of particle rebound contributes less to the cross-sectional porosity. This is because the successive deposition of particles causes the pores in the coating cross-section to collapse, forming a more consolidated structure.
3.1 Analysis of Parameter Effects
To understand the effect of each spray parameter on the coating, response plots for each parameter and coating property were created as illustrated in Figure 4. The response plots illustrate the system response at each, separate level. It is observed in Figure 4 that the feed rate has the largest effect on the response between the three levels for coating thickness. According to Taguchi and factorial design principles, the larger the difference between the responses, the more influential the spray parameter is. Therefore, it can be concluded that the feed rate has the greatest influence on the coating thickness. It is evident from Figure 4, that increasing the feed rate increases the coating thickness. This can be attributed to the fact that increasing the feed rate in a spray process allows for more powder to be dispensed to the spray nozzle, which in turn leads to greater deposition onto the substrate, resulting in a thicker coating. It is also observed that increasing the gas temperature to 500℃ causes the thickness response to increase. The deposition rate of particles increases with increase in temperature as increasing the temperature allows for enhanced thermal softening which is an essential bonding mechanism in cold spray coatings [30]. Increasing the stand-off distance from 3 mm to 5 mm also causes an increase in the thickness response. This can be explained by a phenomenon explored by Sarikaya where the in-flight particle temperature is described as a function of stand-off distance [37]. Increasing the stand-off distance increases the particle dwell time in the plume of the spray, which causes particle temperatures to rise [37]. Particles that attain higher temperatures will have improved bonding with the substrate thereby improving the deposition efficiency and coating thickness.
Similarly, for the coating hardness, it is seen in Figure 4, that the feed rate is the most influential spray parameter as it exhibits the largest difference in the response. It is observed that the hardness response increases significantly when the feed rate is increased to its maximum [23.7 g/min]. As the feed rate increases, the content of WC in the coating also increases as more particles are dispensed in the spray plume. The increased content of WC promotes load sharing between the coating matrix and the hard particles, thereby yielding higher hardness values [26]. Greater content of WC in the coating increases the interfacial area available to share the load from the coating matrix during application of external loads [26]. Munday, et al. also concluded that increasing hard particle content leads to a decrease in the mean free path between hard particles which inhibits plastic deformation of the coating under loaded conditions by stopping coalescence of nucleated voids, thereby leading to harder coatings [26].
The response reaches a maximum when the stand-off distance goes from 3 mm to 5 mm (Figure 4); however, further increasing the stand-off to 10 mm decreases the response. The variation in the response as the stand-off distance is increased can be attributed to the powder particle impact velocity, which is influenced by the stand-off distance. Goyal, et al. found that having smaller stand-off distances increases the in-flight velocity of the impacting particles; therefore, for smaller stand-off distances, the impacting velocity of the particles is higher [34]. This higher net impact velocity creates a peening effect in the coating, consolidating it and causing it to become more dense. This consequently increases the coating hardness [34]. The phenomenon found by Sarikaya also explains the varying response of the coating hardness [37]. As established earlier, increasing the stand-off distance increases the particle dwell time in the spray plume causing the particle temperatures to increase. However, increasing the stand-off distance past an optimal point causes the particles to decrease in temperatures as the isotherms in the plume begin to decay [37]. Therefore, increasing the stand-off distance past a certain point causes the deposition rate to decrease and, consequently, causing the coating hardness to decrease. Coating hardness is also greatly affected by the amount of reinforcing particles like WC in the coating. Although the content of WC particles in the powder blend is held constant for this study, the amount of WC particles in the coating varies with changing operation parameters [26]. Therefore, it can also be concluded that the WC content in the coating is maximized when a stand-off distance of 5 mm is used, thus yielding higher hardness. Increasing the temperature to 500℃ causes the coating hardness to increase. This response is in agreement with the coating thickness response to temperature increase.
Next, the surface porosity appears to be most influenced by the feed rate and the stand-off distance, as illustrated in Figure 4. As the stand-off distance increases from 3 mm to 10 mm, the coating porosity decreases significantly. A similar trend was found by Qiao, et al. where increasing the stand-off distance led to increased particle dwell time in the plume leading to higher particle temperatures [40]. The increased particle temperatures enhance thermal softening leads to flatter particles upon impact that are easier to tile with the previously deposited particles, consequently collapsing the pores in the coating [40]. Increasing the feed rate causes the porosity response to increase (Figure 4), as hard particle rebounds increase with increasing feed rates leaving pores in the rebound locations. Though increasing the feed rate does yield a thicker coating, the coating contains more pores due to particle rebounds, thereby causing the porosity response to increase. The porosity response to temperature variation does not exhibit a clear trend (Figure 4) as the feed rate and stand-off distance exhibit antagonistic responses and their coupled response may overwhelm the response due to temperature variation alone. Additional fractional factorial studies can be conducted to get information about the coupled relationships of varying the feed rate, stand-off distance and temperature to make conclusions about the coupled responses.
3.2 Analysis of Variance (ANOVA)
The analysis of variance (ANOVA) was conducted using the statistical software Minitab to determine the significance of each spray parameter on the coating thickness, hardness and surface porosity, which is crucial to understanding and predicting the response of the system. ANOVA is performed by comparing the variability in the response contributed by each spray parameter with the total variability of the response. The total variability is the sum of squared deviations (SS) from the total mean of the response (MS) for a total of N samples. The F ratios are calculated for each spray parameter and compared with the critical F ratio which is found using the selected confidence interval α and the degrees of freedom for the selected parameter k. For this study, if the F ratio for a specific spray parameter is larger than 4.20, that spray parameter can be concluded to have significant influence on the coating quality.

Using equations 1 and 2, ANOVA was performed for the coatings generated. The ANOVA results for coating thickness, hardness and porosity are summarized in Table 2, Table3 and Table 4, respectively. It was found that the feed rate and stand-off distance both had significant influence on the coating thickness as their F ratios were larger than 4.20. It was also found that the interaction between temperature and stand-off distance also had significant influence on the coating thickness, as seen in Table 2. The feed rate and stand-off distance were also the largest contributors in the coating response, contributing 31% and 28% of the total response respectively, as tabulated in Table 2. All the other parameters and interactions were found to have insignificant influence on the coating thickness as they have F ratios of less than 4.20.
The ANOVA results for coating hardness and surface porosity concluded that none of the spray parameters had significant influence on the coating hardness and porosity as the F ratios were all below 4.20. It was found for the surface porosity, the greatest contribution to the response was 45% from the temperature. For the coating hardness, as indicated by the response plots (Fig. 4), the interaction between the parameters had a greater effect than the individual parameters themselves. As seen in Table 3, the largest contributor for the coating hardness is the interaction between feed rate and stand-off distance, contributing 37% of the overall response. The interaction between feed rate and temperature, and temperature and stand-off distance come in next contributing 27% and 23% of the effect, respectively. This further concludes that, unlike coating thickness and surface porosity, the coating hardness response is governed by the interaction effects between the spray parameters. Since the parameters all had F values lower than 4.20 for the coating hardness and porosity (Table 4), their effects are concluded to be insignificant; however, it is advantageous to know the largest contributors that affect the coating hardness and surface porosity to allow for further investigations and optimization studies. The insignificance of these parameters for the coating hardness and surface porosity can be due to the large size of the experimental design or inherent inconsistencies in the system which overwhelm the statistical studies.
Table 2: ANOVA results for coating thickness.
Symbol
|
Parameter
|
Degrees of Freedom
|
Mean Sum of Squares
|
F value
|
Contribution [%]
|
A
|
Feed Rate
|
2
|
0.176
|
9.26
|
31.22
|
B
|
Temperature
|
2
|
0.072
|
3.81
|
12.85
|
C
|
SOD
|
2
|
0.160
|
8.41
|
28.35
|
A*B
|
Feed Rate * Temperature
|
4
|
0.016
|
0.83
|
2.80
|
A*C
|
Feed Rate * SOD
|
4
|
0.058
|
3.04
|
10.25
|
B*C
|
Temperature * SOD
|
4
|
0.082
|
4.31
|
14.53
|
Error
|
|
8
|
0.019
|
|
|
Total
|
|
26
|
|
|
|
Table 3: ANOVA results for coating hardness.
Symbol
|
Parameter
|
Degrees of Freedom
|
Mean Sum of Squares
|
F value
|
Contribution [%]
|
A
|
Feed Rate
|
2
|
241.15
|
0.42
|
8.27
|
B
|
Temperature
|
2
|
91.01
|
0.16
|
3.15
|
C
|
SOD
|
2
|
26.13
|
0.05
|
0.98
|
A*B
|
Feed Rate * Temperature
|
4
|
799.07
|
1.38
|
27.17
|
A*C
|
Feed Rate * SOD
|
4
|
1097.16
|
1.90
|
37.40
|
B*C
|
Temperature * SOD
|
4
|
676.25
|
1.17
|
23.03
|
Error
|
|
8
|
578.55
|
|
|
Total
|
|
26
|
|
|
|
Table 4: ANOVA results for surface porosity.
Symbol
|
Parameter
|
Degrees of Freedom
|
Mean Sum of Squares
|
F value
|
Contribution [%]
|
A
|
Feed Rate
|
2
|
4.028
|
0.1
|
4.74
|
B
|
Temperature
|
2
|
37.918
|
0.95
|
45.02
|
C
|
SOD
|
2
|
19.037
|
0.48
|
22.75
|
A*B
|
Feed Rate * Temperature
|
4
|
10.643
|
0.27
|
12.80
|
A*C
|
Feed Rate * SOD
|
4
|
10.028
|
0.25
|
11.85
|
B*C
|
Temperature * SOD
|
4
|
2.438
|
0.06
|
2.84
|
Error
|
|
8
|
39.891
|
|
|
Total
|
|
26
|
|
|
|
3.3 Optimizing Operation Parameters
Using the results from the effects plots and ANOVA, an optimization study was performed using Minitab to determine the optimal spray parameters (OSP) that would produce the best coating quality possible for this system. The goals for this optimization study were to maximize the coating thickness and hardness, while minimizing the surface porosity. The goals were selected in order to produce the thickest and hardest coating possible to achieve improved wear and erosion performance [42], and to achieve thicker coatings to increase the coating life span [43]. The porosity was selected to be minimized to achieve a more uniform, less porous coating to avoid localized areas of increased stress in loaded conditions [44]. The optimal spray parameters (OSP) were found to be 23.7 g/min for the feed rate, 500℃ for the temperature and 10 mm for the stand-off distance. The OSP would yield a coating that is 1.22 ± 0.06 mm thick, with a hardness of 364.5 ± 8.5 HV, and a porosity of 6.8 ± 0.6%.