3.1 Surface Roughness characterization after EP, CP, and as-built AM samples
Surface characterization of the samples was conducted utilizing the KEYENCE Digital Microscope VHX-7000. Figure 5 summarizes different roughness parameters for the AM surface after three types of cleaning. Figure 4a and 4b, present the outcomes of the electropolishing surface finishing technique, revealing a substantially flatter topography with diminished hills and peaks. The associated surface roughness parameters are: Ra = 4.85 µm, Rz = 22.89 µm, RzJIS = 13.02 µm, Rp = 11.76 µm and Rv = 11.13 µm (Fig. 5). Figure 4c and 4d, illustrate the surface finish post chem-polishing. The corresponding surface roughness measurements are as follows: Ra = 11.65, Rz = 56.84 µm, RzJIS = 17.79 µm, Rp = 33.33 µm and Rv = 23.50 µm (Fig. 5). Figure 4e-f, depict the surface topography of the as-built samples. The surface roughness parameters for the as-built specimens are as follows: Ra (arithmetic mean roughness) = 15.95 µm, Rz (the average maximum peak) = 86.99 µm, RzJIS (Ten-point mean roughness) = 38.54 µm, Rp (Maximum profile peak) = 40.73 µm,and Rv (Maximum profile valley depth) = 46.26 µm (Fig. 5). Figures 4 and 5 suggests that surface topography and roughness changed significantly after chempolishing and electropolishing. Figures 4 and 5 suggests that surface topography and roughness changed significantly after chempolishing and electropolishing.
3.2 Surface Roughness characterization after Nickel Plating
To gain the insight about the role of surface treatment method and different coating parameters we investigated the 9 samples prepared as per the Taguchi Design of experiment scheme (Table 1). Figure 6 shows the topography of different DOE samples, whereas Fig. 7 summarizes the roughness data on each sample quantitatively. Figure 6a in our study portrays the outcomes of Design of Experiment (DOE) #1, where a specific set of experimental parameters was employed (Table 1). These parameters encompassed the utilization of a high phosphorus nickel solution, an electropolished surface finish, orientation within the XY plane, and an elevated temperature exceeding the optimum by 5°C. (Fig. 7)Fig. 6b, denoting DOE#2, features a distinct combination of experimental parameters involving a high phosphorus nickel solution, chem-polished surface finish, alignment along the YZ plane, and an optimal temperature setting (Fig. 7). The resultant surface roughness measurements are as follows: Ra = 10.90 µm, Rz = 49.993 µm, RzJIS = 34.61 µm, Rp = 23.1 µm, Rv = 26.89 µm, Rc = 37.03 µm, Rt = 49.99 µm, and Rq = 12.90 µm (Fig. 7). In Fig. 6c, we present the outcomes of DOE#3, executed with the utilization of a high phosphorus nickel solution, an as-built surface finish, alignment along the XZ plane, and a temperature setting lower than the optimum by 5°C. The measurements of surface roughness revealed the following values: Ra = 15.7639 µm, Rz = 70.21 µm, RzJIS = 19.22 µm, Rp = 35.01 µm, Rv = 35.20 µm, Rc = 35.63 µm, Rt = 70.23 µm, and Rq = 18.77µm.
Figure 6d, representing DOE#4, comprises parameters involving a mid-phosphorus nickel solution, an electropolished surface finish, alignment along the YZ plane, and a temperature set lower than the optimum by 5°C. The surface roughness characteristics of this configuration resulted in: Ra = 15.15 µm, Rz = 67.91 µm, RzJIS = 36.42 µm, Rp = 38.38 µm, Rv = 29.5267 µm, Rc = 43.58 µm, Rt = 67.94 µm, and Rq = 18.21 µm. Figure 6e, representing DOE#5, encapsulates an experimental framework characterized by the use of a mid-phosphorus nickel solution, chemo-polished surface finish, alignment along the XZ plane, and an elevated temperature surpassing the optimum by 5°C. Surface roughness measurements for this setup yielded the following results: Ra = 14.59 µm, Rz = 73.4 µm, RzJIS = 36.18 µm, Rp = 32.95 µm, Rv = 40.45 µm, Rc = 44.43 µm, Rt = 73.4 µm, and Rq = 17.475 µm. Figure 6f, representative of DOE#6, outlines a specific combination of parameters that incorporates a mid-phosphorus nickel solution, as-built surface finish, alignment along the XY plane, and the optimal temperature. The corresponding surface roughness measurements are as follows: Ra = 17.56 µm, Rz = 84.77 µm, RzJIS = 32.82 µm, Rp = 46.77 µm, Rv = 38.03 µm, Rc = 50.1 µm, Rt = 84.81 µm, and Rq = 21.0 µm (Fig. 7).
Figure 6g corresponds to DOE#7, characterized by the use of a low phosphorus nickel solution, an electropolished surface finish, alignment along the XZ plane, and the optimal temperature setting. Surface roughness measurements for this particular configuration revealed the following values: Ra = 18.54 µm, Rz = 87.49 µm, RzJIS = 17.44 µm, Rp = 46.06 µm, Rv = 41.42 µm, Rc = 25.57 µm, Rt = 87.54 µm, and Rq = 22.22 µm. Figure 6h, emblematic of DOE#8, incorporates a set of parameters that entail a low phosphorus nickel solution, chemo-polished surface finish, alignment along the XY plane, and the temperature setting at the optimum level. The associated surface roughness characteristics are as follows: Ra = 17.44 µm, Rz = 71.88 µm, RzJIS = 20.65 µm, Rp = 32.87 µm, Rv = 39.01 µm, Rc = 55.84 µm, Rt = 71.9 µm, and Rq = 20.44 µm. Finally, Fig. 6i represents DOE#9, wherein the experimental parameters encompass the utilization of a low phosphorus nickel solution, an as-built surface finish, alignment along the YZ plane, and a temperature exceeding the optimum by 5°C. The surface roughness measurements for this configuration resulted in the following values: Ra = 11.30 µm, Rz = 51.26 µm, RzJIS = 19.56 µm, Rp = 26.08 µm, Rv = 25.17 µm, Rc = 25.49 µm, Rt = 51.26 µm, and Rq = 13.66 µm (Fig. 7).
3.4 Scratch testing
The utilization of scratch testing serves as a fundamental technique for the comprehensive analysis and characterization of mechanical wear behaviors. The meticulous application of precisely defined scratches, performed in a consistent and reproducible fashion, becomes of paramount importance in the pursuit of surface wear resistance characterization. In the scope of our study, we opted for the standard 10 N scratch test methodology, conducted on samples coated with nickel (Fig. 9). Each of the nine samples (Table 1) was subjected to scratch testing (Fig. 9a-f). We were successful in obtaining clear scratch on nickel coated sample with high phosphorous content (Fig. 9a-c), medium phosphorous content (Fig. 9d-f), and low phosphorous content (Fig. 9g-i). However, a major challenge was in analyzing different scratches. The acquisition of precise measurements of the scratch width from the microscopic imagery was difficult due to the scratch's inconsistent width. Consequently, the methodology has been adapted to calculate the projected area of the scratch, which is then divided by its length to achieve a standardized measure (Fig. 10a).
In the realm of image segmentation, the Segment Anything Model (SAM), developed by Meta, previously known as Facebook, has been employed. As a foundational model for segmentation, SAM has undergone training on a dataset encompassing 11 million images and in excess of one billion masks. The architecture of SAM is tripartite, consisting of an image encoder, a prompt encoder, and a mask decoder. The strength of SAM lies in its dual utility, offering both a no-code and a code-based solution. For the purposes of our experiment, the no-code, fully online option was selected. Subsequent to this, we applied denoising and thresholding processes using ImageJ to refine the results (example shown in, Fig. 10b). Once we have performed instance segmentation and thresholding on the scratch, we proceed to determine the material's level of hardness using a formula.
$$Hs = 8P/\varPi {w}^{2}$$
Where,
Hs = Scratch hardness number (MPa)
P = Normal force (N)
w = scratch width
Upon a meticulous analysis of the results derived from the 10 N scratch test, a conspicuous enhancement in scratch resistance was observed both prior to and subsequent to the test. As depicted in Fig. 9, it is discernible that Design of Experiment (DOE) #2 and #7 exhibited a substantial augmentation in scratch resistance, reaching an impressive increase of up to 50%. Furthermore, DOE #4, #5, #8, and #9 displayed a marked improvement in surface hardness, manifesting enhancements that ranged from 51–86%. Of notable significance, DOE #6 exhibited the most substantial advancement in surface hardness, registering an impressive escalation of 128%. However, it is worth mentioning that the remaining experiments failed to yield any substantial improvements in surface hardness. Hardness data before and after nickel coating has been summarized in Fig. 11.
Table 2
Surface Hardness before and after Ni coating
| DOE | Hardness before Ni deposition | Hardness after Ni deposition |
| 1 | 125.13 | 332.64 |
| 2 | 119.67 | 338.10 |
| 3 | 207.47 | 587.53 |
| 4 | 218.19 | 610.02 |
| 5 | 194.94 | 538.22 |
| 6 | 297.36 | 890.48 |
| 7 | 219.64 | 643.89 |
| 8 | 233.46 | 660.62 |
| 9 | 269.18 | 815.58 |
3.5 Taguchi Analisis
Multi-plots serve as a visualization tool for elucidating the impacts of numerous factors on a response variable (Fig. 12). Within these plots, the presentation encompasses both main effects and interaction effects of the factors on the response variable (Fig. 12). The main effects delineate the individual influence of each factor on the response variable, while the interaction effects depict the collective impact of two or more factors on the response variable. Phosphorous level impacted significantly (Fig. 12a); high Phosphorous conetent appear to yield low hardness as compared to the medium and low phosphorous content. Interestingly, electropolish and chempolish surface treatment yielded similar effect (Fig. 12b). The unpolished surface behaved very differently (Fig. 12a). The surface orientation and temperature factor’s level impact was relatively weaker (Fig. 1c,d). Different sample orientation produced different impact (Fig. 12c); XY orientation behaved significantly different than that of XZ and YZ orientation (Fig. 12c). Temperature effect was similar for level T and T + 5, and it was different as compared to T-5 (Fig. 12d). We also studied the strength of interaction between two parameters. All potential interactions between pairs of two factors are computed (Fig. 13.a). The interaction pairs are presented in a descending order based on their Severity Index (SI), which is expressed on a scale from 0 to 100%. In cases involving interactions among pairs of factors with 3 levels, the Severity Index (SI) is indicative of the highest angle within the array of feasible combinations of line segments (Fig. 13.a). Interaction data shows relative independence of a factor in relation to other factor. The highest interaction strength of 92.78 was observered between the orientation and temperature; it means changing orientation will necessiate adjustment in the plating temperature for the desired results (Fig. 13a). On the other hand, low severity index of 6.92 phosphorous and surface finish interaction shows their independence from each other. We also investigated the impact of individual factor on film hardness (Fig. 13b). Phosphorus content is main factor in deciding the nickel coating hardness (Fig. 13b). This result is in line with the prior literature relating the phosphorous content to the hardness of the nickel coating (Fig. 13b). Interestingly, surface finishing method was the second most important factor (Fig. 13b). This result is of critical importance in the light of the degree of complexity involved in AM geometry. Hardness is highest for the As built surface. The reason is that high roughness enable creating better grip of the coating material on the surface (Fig. 13b), this result is consistent with the prior literature defining the impact of roughness on the film adhesion.
We also employed Taguchi Design of experiment analysis to investigate the combination of parameters that will lead to the highest hardness. The optimal table represent the predictive equation delineating the anticipated performance under optimal conditions as well as any conceivable alternative conditions. The numerical values presented in the table are derived from computations conducted under the optimal condition, a state determined by the chosen quality characteristic for analysis. Conventional practice dictates the inclusion of only statistically significant factors (without pooling) in the computation of anticipated performance, aligning with established analytical methodologies. In the context of this experiment, The optimal condition for achieving the highest hardness is identified as a low-phosphorus nickel solution, coupled with a as-built surface finish, XY orientation, and a solution temperature of 90 degrees Celsius.