materials Article Full Optimization of an Electroless Nickel Solution: Boosting the Performance of Low-Phosphorous Coatings Asier Salicio-Paz 1 , Ixone Ugarte 1 , Jordi Sort 2,3 , Eva Pellicer 2 and Eva García-Lecina 1, * 1 CIDETEC, Basque Research and Technology Alliance (BRTA), E-20014 Donostia-San Sebastián, Spain; asalicio@cidetec.es (A.S.-P.); iugarte@cidetec.es (I.U.) 2 Departament de Física, Universitat Autònoma de Barcelona, E-08193 Bellaterra, Spain; jordi.sort@uab.cat (J.S.); eva.pellicer@uab.cat (E.P.) 3 Institució Catalana de Recerca i Estudis Avançats (ICREA), E-08180 Barcelona, Spain * Correspondence: egarcia@cidetec.es Abstract: Univariate and multivariate optimizations of a novel electroless nickel formulation have been carried out by means of the Taguchi method. From the compositional point of view, adjustment of the complexing agent concentration in solution is crucial for fine-tuning free Ni2+ ions concentra- tion and, in turn, the mechanical properties of the resulting coatings. The Ni (II) concentration and the pH are the main parameters which help restrict the incorporation of phosphorous into the Ni layers. On the other hand, the stirring rate, the pH and the reducing agent concentration are the most influential parameters for the corrosion resistance of the coatings. Multivariate optimization of the electrolyte leads to a set of optimized parameters in which the mechanical properties (hardness and   worn volume) of the layers are similar to the optimal values achieved in the univariate optimization, but the corrosion rate is decreased by one order of magnitude. Citation: Salicio-Paz, A.; Ugarte, I.; Sort, J.; Pellicer, E.; García-Lecina, E. Full Optimization of an Electroless Keywords: electroless nickel; Taguchi; optimization; hardness; wear; corrosion Nickel Solution: Boosting the Performance of Low-Phosphorous Coatings. Materials 2021, 14, 1501. https://doi.org/10.3390/ma14061501 1. Introduction Since the discovery of electroless nickel plating (ENP) by Brenner and Riddell in Academic Editors: Eugenijus Norkus 1946 [1], this formulation has become one of the preferred solutions for engineering applica- and Loreta Tamasauskaite- tions in which high corrosion resistance and superior mechanical properties are required [2]. Tamasiunaite In essence, and leaving aside the mechanistic phenomenon behind the electroless formation of NiP, which still remains unsolved [3,4], the electroless plating method is an autocatalytic Received: 21 February 2021 metal deposition in which the reduction in metallic cations is driven by a reducer, present Accepted: 15 March 2021 in the same solution, without the use of external power supply [5]. Published: 18 March 2021 ENP electrolytes are multicomponent electrolytes typically containing a source of nickel cations, a reducer and several complexing agents, in addition to stabilizers, buffers, Publisher’s Note: MDPI stays neutral accelerators and wetting agents [6]. All these components fulfil essential functions during with regard to jurisdictional claims in published maps and institutional affil- the electroless process and their careful selection, along with the fine tuning of the experi- iations. mental conditions, have a pronounced effect on the properties of the resulting coatings [7]. The properties of electroless nickel coatings are mainly governed by the phosphorous (P) content in the alloy [8]. Low phosphorous deposits (3–5 wt.% P) exhibit superior mechanical properties but poor corrosion resistance in acid media; medium phosphorous coatings (6–9 wt.% P) show moderate corrosion resistance and good mechanical properties Copyright: © 2021 by the authors. and, finally, high phosphorous layers (10–12 wt.% P) exhibit superior corrosion resistance Licensee MDPI, Basel, Switzerland. at the expense of limited mechanical performance [9–13]. In addition to the phosphorous This article is an open access article distributed under the terms and content, careful choice of bath composition, operating pH and temperature of the system conditions of the Creative Commons have an effect on the properties of ENP layers [14]. Attribution (CC BY) license (https:// In this work, the best mix of bath composition (i.e., components and their concen- creativecommons.org/licenses/by/ tration) and operational parameters were targeted in order to optimize a proprietary 4.0/). electroless nickel formulation. As a result of the large number of factors to be considered, Materials 2021, 14, 1501. https://doi.org/10.3390/ma14061501 https://www.mdpi.com/journal/materials Materials 2021, 14, 1501 2 of 18 the Taguchi method was used to elucidate the effect each factor had on the performance of the bath and the properties of the resulting coatings. The Taguchi method consists of three phases; namely system design, parameter design, and tolerance design [15]. System design involves the selection of the optimal factor levels for developing the quality characteristics. In this research, system design was utilized in combination with multivariable analysis to determine the optimum set of bath components, concentration and working parameters which yielded the best corrosion resistance, mechanical and tribological properties of the coatings. In contrast to full factorial analysis, the Taguchi method reduces the number of experimental runs by using orthogonal arrays (OAs) in which all the factors and their levels are included, defining the experimental setup during the research [16]. In our study, an L8 OA was utilized to explore the effect of the most influential parameters of the new ENP bath, where the letter L in this notation stands for Latin square. The concentrations of the nickel salt (A), the reducing agent (B), the complexing agents 1 (C) and 2 (D), the stirring rate (E), the temperature (F) and the pH (G) were selected as the most influential parameters. Additionally, two levels were set for each factor, depicted by “−” for the lowest and “+” for the highest levels, respectively, completing the L8 OA as shown in Table 1. The effect of process parameters and bath composition on the thickness, hardness, corrosion resistance and wear of the coatings was evaluated separately based on individual optimization stages for each feature. Nonetheless, coatings working under real- life conditions are usually subject to more than one degradation mechanism (i.e., corrosion, wear, fretting, etc.) simultaneously. For this reason, a multivariate analysis was carried out aimed at gathering the conditions leading to both best wear and corrosion resistance for the here considered low phosphorous electroless nickel coatings. Table 1. Experimental matrix derived from L8 orthogonal array (OA). C E F Factors A B Interaction D Interaction Interaction G A-B A-D B-D Ni (II) Reducer Complexing 1 Complexing 2 Stirring T Experiment ◦C pH g/L g/L g/L g/L rpm 1 4.8 20.0 1.0 7.0 100 75 5.2 2 4.8 20.0 1.0 12.0 250 90 6.5 3 4.8 30.0 3.0 7.0 100 90 6.5 4 4.8 30.0 3.0 12.0 250 75 5.2 5 6.5 20.0 3.0 7.0 250 75 6.5 6 6.5 20.0 3.0 12.0 100 90 5.2 7 6.5 30.0 1.0 7.0 250 90 5.2 8 6.5 30.0 1.0 12.0 100 75 6.5 2. Materials and Methods 2.1. Electroless Nickel Bath and Operating Conditions A proprietary low P electrolyte has been used in this study. Main components of the solution include NiSO4 as Ni (II) cations source, sodium hypophosphite as reducing agent, short chained organic acids as complexing agents (i.e., malic acid, glycine, citric acid, propionic acid, etc.) and ammonia and diluted sulfuric acid as pH modifiers. Apart from bath compositional factors, some working parameters, namely solution pH, temperature and stirring rate were included as parameters to be optimized. All the reagents were of analytical grade and were dissolved in distilled water. Note that the quality of reagents is important in electroless nickel formulations as these elec- trolytes are very sensitive to metallic impurities, usually present in low purity chemicals. Importantly, the ENP formulation used here meets the requirements of the RoHS and WEE standards [17]. Coatings were plated in 1L cylindrical glass cells for a total plating time of 1 h. Four NiP coatings were manufactured per condition during the optimization stage. Materials 2021, 14, 1501 3 of 18 Temperature of the electrolyte was kept fixed by means of a PT-1000 probe attached to a hot plate magnetic stirrer (RTC basic, IKA, Staufen im Breisgau, Germany). The probe ensured a maximum temperature deviation of ± 1 ◦ C. The electrolyte was stirred with a Teflon® -coated magnetic bar (Wilmington, NC, USA) placed at the bottom of the cell. Meanwhile, a polyethylene cover was placed on top of it to prevent from water evapora- tion and consequent changes in the concentration of the species in the electrolyte, and to maintain a stable temperature across the cell. 2.2. Substrate Conditioning The substrates were low carbon steel (AISI 1010) [18] flat panels (100 mm × 75 mm × 1 mm). In order to assure good adhesion between the substrate and the electroless nickel coatings, the former was subject to surface conditioning. Firstly, after soak cleaning, the substrates underwent chemical degreasing (UniClean 251, Atotech Deutsland GmBH, Berlin, Germany) for 5 min at 60 ◦ C followed by electrolytic degreasing at 4 V for 1 min in the same solution. A stainless-steel anode was employed during the electrolytic degreasing. The substrates were then subject to acid pickling for removing the outermost oxide layer from their surface. The acid solution was composed of HCl and H2 SO4 at 40% and 5% (v/v), respectively. Once gently rinsed, the substrates were transferred to the plating cell. 2.3. Selected Orthogonal Array The parameters subject to optimization were the concentrations of nickel cations (A), reducing agent (B) and complexing agents 1 (C) and 2 (D), the stirring rate (E), the temperature (F) and the solution pH (G). In the Taguchi approach, the degrees of freedom (DoF) of a system are defined as DoF = number of factors − 1. In our case, as there were seven parameters to explore, the system had six DoF. Thus, the smallest OA able to allocate six DoFs is an L8 (27 ) matrix, which enables the study of seven factors at two levels, where “−” and “+” in the OA stand for lower and upper levels, respectively, for each factor. The different factors and their levels are distributed in the resulting matrix, giving rise to the sequence of experiments to be done. Each row defines the experimental setup for every single test. Careful positioning of the factors in the L8 OA needs to be done beforehand in order to anticipate possible interactions between factors [19]. An interaction between two factors occurs when the effect of a factor on the considered feature or property depends on the level at which the other factor is operating, thereby hiding or polluting the real effect of the former. As the OA was populated with seven factors, these were placed in matrix positions for which, according to common knowledge in electroless plating, interactions between parameters are not expected. The final experimental matrix derived from the OA is shown in Table 1. The Taguchi methodology involves a graphical representation of the parameters, which further facilitates detection of the most influential ones on the properties to boost. The results are converted into signal-to-noise (S/N) ratio according to three main cate- gories, namely lower-the-better, higher-the-better, and nominal-the-best. The objective of the optimization stage is to maximize the S/N ratio in order to achieve the best quality char- acteristics. According to the methodology, it is possible to predict the response of each feature according to the optimal set of parameters derived from the optimization stage. The prediction is then validated in a confirmation test or experiment [20]. Thus, in order to unveil the effect that the different bath components and the working parameters (pH, stirring rate, and temperature) have on the properties of the resulting coatings, thickness, hardness, wear and corrosion resistance were selected as the outputs of individual univariate optimizations. On a further step, a multivariate optimization was carried out for all the features mentioned above, by prioritizing corrosion and wear resistance as the properties to be improved. Materials 2021, 14, 1501 4 of 18 2.4. Characterization of the Coatings Surface topography and thickness analyses were carried out on a Jeol JSM 5500LV scanning electron microscope (SEM, Tokyo, Japan). Compositional analyses of the samples were performed in the SEM by means of an Oxford instruments INCA X-sight X-ray spectroscopy (EDX, Abingdon, UK) detector. In some cases, coatings were analyzed on a Zeiss Gemini field emission scanning electron microscope (FESEM, Oberkochen, Germany) equipped with an Ametex EDAX Apollo X detector. Thickness and composition of the coatings were double checked by means of Fischer Instruments XDV-SDD X-ray fluorescence instrument (XRF). Cross-sections of the coatings were prepared by cutting and embedding in hot mounting epoxy resin, grinded with SiC paper (1400–4000 grit) and polished down to a mirror-like finishing using diamond paste (1 µm). Crystalline structure of the coatings was analyzed on a Bruker D8 X-ray diffractometer(Billerica, MA, USA) using CuKα radiation in Bragg Brentano geometry. Phase analysis was done using Bruker EVA® software package [21]. Crystallite size was evaluated using the Scherrer’s formula. Hardness of the coatings was assessed with the help of a Fischer Instruments Fisch- erscope HM 2000 (Fischer Instrumentation(GB)Ltd., Lymington, UK), using a pyramidal- shaped diamond Vickers indenter, by applying a normal load of 100 mN. The obtained values correspond to the average of 10 indentations performed on the cross-section of the coatings. Wear performance of the coatings was evaluated using a CSM THT tribometer in ball-on-disc configuration (CSM Instruments, Needham, MA, USA). The tribo-pair con- sisted of an Al2 O3 ball (Ø = 6 mm) in contact with the NiP coatings. The applied normal load was 10 N during a total sliding distance of 3000 m at a rotating speed of 10 mm/s. The wear tracks were evaluated by a Leica Microsystems DCM3D (Leica Microsystems, Wetzlar, Germany) confocal microscope, operated with a blue light (λ = 460 nm) at a 10× magnification by means of a 10× EPI objective. The total worn volume was calculated using LeicaMaps (Leica Microsystems) dedicated software. Corrosion performance of the coatings was studied by electrochemical techniques. Experiments were conducted on a Princeton Applied Research cylindrical cell with a total volume of 250 cm3 and an exposed testing area of 1 cm2 . The electrolyte was NaCl 3.5 wt.% (Scharlau, ACS), a Crison 52–41 Ag/AgCl/KCl (3M, St. Paul, MN, USA) was employed as a reference electrode whereas a platinized titanium mesh was used as a counter-electrode. All measurements were conducted under stagnant conditions at room temperature. During the experiments, open-circuit potential (OCP) values were recorded for 1 h in order to ensure a stable value before running the polarization curves. Separate cathodic and anodic potentiodynamic curves were recorded in the range of (−0.050/+1.0) V vs. OCP and (+0.050/−1.0) V vs. OCP, respectively. Corrosion rate was extrapolated from the analysis of the corresponding Tafel slopes. 3. Results and Discussion 3.1. Univariate Optimization 3.1.1. Composition and Crystallographic Structure In the pursuit of achieving the best mechanical properties, the lowest possible P content was targeted. Therefore, the lower-the-better approach was utilized to find the optimal combination of factors leading to the lowest P content. Note that the S/N ratio is determined differently depending on the type of characteristic being analyzed [22]. As can be seen in Table 2, the phosphorous content varies in the range from 2.2 to 8.5 wt.% P as a function of both bath components concentrations and working parameters. Table 2. Phosphorous content in the coatings obtained from L8 experiments. “S.D.” stands for standard deviation. Experiment 1 2 3 4 5 6 7 8 P content (wt.%) 7.0 3.2 7.8 6.4 2.2 6.7 8.5 4.0 S.D. 0.1 0.2 0.1 0.1 0.3 0.1 0.3 0.1 Table 2. Phosphorous content in the coatings obtained from L8 experiments. “S.D.” stands for standard deviation. Materials 2021, 14, 1501 Experiment 1 2 3 4 5 6 7 58of 18 P content 7.0 3.2 7.8 6.4 2.2 6.7 8.5 4.0 (wt.%) S.D. 0.1 0.2 0.1 0.1 0.3 0.1 0.3 0.1 Figure 1a shows shows that thatthe thepH pHvalue value(G) (G)and thethe and concentration concentration of reducing agent of reducing (B) are agent (B) decisive are in achieving decisive the lowest in achieving P content. the lowest Meanwhile, P content. Meanwhile,stirring and temperature stirring also have and temperature also an effect, have although an effect, lesser. lesser. although This can alsocan This be also observed in the mean be observed response in the values based mean response on values S/N calculations (Table 3). based on S/N calculations (Table 3). Figure 1. Signal-to-noise (S/N) (S/N) ratios ratiosfor for (a) (a) PP content, content, (b) (b) thickness, thickness, (c) (c) hardness, hardness, (d) wear resistance (worn volume), and (e) corrosion resistance (corrosion current) of electroless nickel plating (e) corrosion resistance (corrosion current) of electroless nickel plating (ENP) (ENP) layers layers for for the the differently differently analyzed analyzed factors factors and and their levels (−, their levels (−, +). +). Table 3. Mean effect plot for S/N ratios for P content in the ENP layers obtained following the Table 3. Mean effect plot for S/N ratios for P content in the ENP layers obtained following the lower-the-better optimization lower-the-better optimization approach. approach. Factor [Ni (II)] [Reducer] [Complexing 1] [Complexing 2] Stirring T pH Factor [Ni (II)] [Reducer] − [Complexing −15.3 1] −12.6 [Complexing −14.4 2] Stirring −15.0 T pH −15.8 −13.0 −17.0 − −15.3 −+12.6 −13.5 −14.4−16.2 −−14.3 15.0 −15.8 −13.7 −13.0 −17.0 −12.9 −15.8 −11.7 + −13.5 −16.2 Difference 1.8 −14.33.6 −13.7 0.1 −12.9 1.3 −15.82.9 −11.75.4 2.8 Difference 1.8 Rank3.6 5 0.1 2 1.37 2.96 2.8 3 4 5.4 1 Rank 5 2 7 6 3 4 1 Taking into account the data displayed in Table 3, next step in the Taguchi’s meth- odology consists Taking in solving into account thea data prediction equation displayed in which in Table 3, nextastep projection of the P content in the Taguchi’s method-is ology consists in solving a prediction equation in which a projection of the obtained according to the optimum parameters sequence. In our case, the optimal combi- P content is obtained nation according of factors to thethe towards optimum lowest Pparameters content wassequence. In our case,forecasting A+B−C+D+E+F−G+, the optimal com- a value bination of of factors 3.3 wt.% towards the experiment P. A confirmation lowest P content was run the−optimized wasatA+B −G+, forecasting C+D+E+Fconditions and thea value of 3.3 resulting wt.% P.was P content A confirmation experiment 3.1 ± 0.2 wt.%, was run with in full agreement at thethe optimized conditions and prediction. the resulting P content Remarkably, was 3.1 ±effect the profound 0.2 wt.%, of pHinonfull theagreement P content with the prediction. in electroless nickel coatings Remarkably, the profound effect of pH on the P content in electroless has been described elsewhere [23]. On the contrary, the reducing agent concentration nickel coatings is has been described elsewhere [23]. On the contrary, the reducing agent concentration is not typically considered a key factor in the literature [7]. As can be seen in Figure 1a, lowering the reducing agent concentration brings a decrease in the P content (maximum S/N). Likewise, an increase in the Ni (II) concentration in solution decreases the P content in the layer. These results confirm the importance of the Ni (II)/H2 PO2 − ratio in solution, so that the higher the ratio within a certain range, the lower the P content [7]. not typically considered a key factor in the literature [7]. As can be seen in Figure 1a, low- ering the reducing agent concentration brings a decrease in the P content (maximum S/N). Materials 2021, 14, 1501 Likewise, an increase in the Ni (II) concentration in solution decreases the P content in the 6 of 18 layer. These results confirm the importance of the Ni (II)/H2PO2− ratio in solution, so that the higher the ratio within a certain range, the lower the P content [7]. The crystalline structure of NiP coatings is also strongly dependent on the P content. The crystalline Figure 2 shows the XRDstructure patterns of of the NiPeight coatings is also samples strongly derived dependent from on the P ma- the experimental content. Figure 2 shows the XRD patterns of the eight samples derived from the experimental matrix. trix. All the peaks can be indexed considering the face-centered cubic (fcc) phase of Ni, All the peaks can be indexed considering the face-centered cubic (fcc) phase of Ni, the the (111) being the preferred orientation in all cases. Besides the (111) reflection, the pat- (111) being the preferred orientation in all cases. Besides the (111) reflection, the patterns terns related to experiments 2, 5 and 8 show a sharper and better resolved (200) peak, related to experiments 2, 5 and 8 show a sharper and better resolved (200) peak, indicating indicating a higher degree of crystallinity. Accordingly, coatings derived from these ex- a higher degree of crystallinity. Accordingly, coatings derived from these experiments had periments had the lowest P contents, namely 3.2, 2.5 and 4.0 wt.%, respectively. Because the lowest P contents, namely 3.2, 2.5 and 4.0 wt.%, respectively. Because phosphorous phosphorous has very low solubility in nickel, the Ni lattice becomes severely distorted has very low solubility in nickel, the Ni lattice becomes severely distorted when both when both elements are co-deposited [24]. The broadening of the (111) peak in the XRD elements are co-deposited [24]. The broadening of the (111) peak in the XRD patterns of patterns of coatings having higher P percentages results from the increased lattice distor- coatings having higher P percentages results from the increased lattice distortion caused tion caused by the incorporation of P in the Ni matrix [25]. Note that P tends to accommo- by the incorporation of P in the Ni matrix [25]. Note that P tends to accommodate at the date at the grain boundaries, thus inhibiting grain growth, which ultimately causes the grain boundaries, thus inhibiting grain growth, which ultimately causes the observed peak observed peak broadening broadening (i.e., smaller(i.e., smallersizes) crystallite crystallite [26]. sizes) [26]. Figure 2. XRD Figure pattern 2. XRD of the pattern ofNiP coatings the NiP obtained coatings fromfrom obtained Experiments 1–8 as Experiments 1–8listed in Table as listed 1. 1. in Table 3.1.2.3.1.2. Thickness Thickness and and Surface Surface Morphology Morphology In this In this study, study, maximum maximum thickness thickness of the of the ENPENP layers layers waswas pursued pursued during during thethe optimiza- optimi- tionstage. zation stage.Unlike UnlikePPcontent, content, thethe optimization optimization of of thickness thicknessfell fellwithin withinthe thehigher-the-better higher-the- category. Figure 1b shows the main effects on thickness, for which better category. Figure 1b shows the main effects on thickness, for which temperaturetemperature andand reduc- ing agent concentration emerged as the most influential parameters. Higher concentrations reducing agent concentration emerged as the most influential parameters. Higher concen- of both nickel cations and reducing agent caused an increase in thickness, in agreement trations of both nickel cations and reducing agent caused an increase in thickness, in with the empirical rate law for auto-catalytic deposition [27]. Thus, the best configuration of agreement with the empirical rate law for auto-catalytic deposition [27]. Thus, the best parameters to achieve the maximum thickness was A+B+C−D−E−F+G−. The predicted configuration of parameters to achieve the maximum thickness was A+B+C−D−E−F+G−. thickness value was 24.1 µm, very close to the one obtained in the confirmation experiment The predicted thickness value was 24.1 µm, very close to the one obtained in the confir- (24.3 ± 1.6 µm), thereby confirming that the chosen configuration of parameters was indeed mation experiment (24.3 ± 1.6 µm), thereby confirming that the chosen configuration of optimal. parameters was indeed optimal. Temperature is the driving force in electroless plating processes and its influence on Temperature is the driving force in electroless plating processes and its influence on coating thickness can be seen in Figure 1b. For completeness, Figure 3 shows the SEM coating thickness can be seen in Figure 1b. For completeness, Figure 3 shows the SEM images of the coatings’ cross-sections for the different studied conditions. The results images of the coatings’ cross-sections for the different studied conditions. The results in- indicate that experiments 2, 3, 6 and 7 yielded the thickest coatings. In these experiments, dicate that experiments the temperature was 2, set3,in6its and 7 yielded upper the◦ C), level (90 thickest coatings. pointing In these to a higher experiments, effect of temperature on thickness than the other parameters. Theoretically, experiments 2 and 3 were the most favorable conditions toward thicker coatings. However, in the case of experiment 2, lower Ni (II) concentration and higher concentration of complexing agent 2 (see Table 1) decreased the amount of free Ni2+ cations available for reduction by H2 PO2 − . As a result, this experimental configuration reduces the effect of temperature on the plating rate, although it is still higher in comparison with experiments operating at lower temperatures. the temperature the temperature was was set set in in its its upper upper level level (90 (90 °C), °C), pointing pointing to to aa higher higher effect effect of of tempera- tempera- ture on ture on thickness thickness thanthan the the other other parameters. parameters. Theoretically, Theoretically, experiments experiments 22 and and 33 were were the the most favorable conditions toward thicker coatings. However, most favorable conditions toward thicker coatings. However, in the case of experiment 2,in the case of experiment 2, lower Ni (II) concentration and higher concentration of complexing lower Ni (II) concentration and higher concentration of complexing agent 2 (see Table 1) agent 2 (see Table 1) Materials 2021, 14, 1501 decreased the the amount amount of of free free Ni Ni2+2+ cations available for reduction by H2PO2−. As a result, 7 of 18 decreased cations available for reduction by H2PO2−. As a result, this experimental this experimental configuration configuration reduces reduces thethe effect effect ofof temperature temperature on on the the plating plating rate, rate, alt- alt- hough it hough it is is still still higher higher inin comparison comparison with with experiments experiments operating operating at at lower lower temperatures. temperatures. Meanwhile, experiments Meanwhile, experiments Meanwhile, 6 and 6 and 7 were conducted 7 were conducted experiments 6 and 7 were with the withconducted temperature the temperature with the factor factor set at set at the temperature thefactor set at the upper level (90 upper level (90 upper °C) and °C) and the pH the (90 level ◦ pH atat the C)the lower andlower the pH level level (5.2) (Table (5.2)lower at the (Table 1). The 1). The level coating (5.2)coating (Table 1).derived derived from from derived from The coating experiment 77 was experiment was the thickest. the thickest. experiment 7 wasImportantly, Importantly, the thickest. inImportantly, in this case this case thethe in concentrations concentrations of both of both Ni this case the concentrations Ni (II) (II) of both Ni (II) and H 2PO2−− were in the upper and H2PO2 were andinHthe − upper range, range, whereas whereas upper the the concentration concentration of complexing of concentration complexing agentsagents 11 2 PO 2 were in the range, whereas the of complexing agents and 22 were and were in in the the 1 andlower lower 2 were range, range, thus thus in the setting setting lower range,aa scenario scenario in which in thus setting which the amount the a scenario amount in which of the of freeamount free Ni2+ Ni 2+ of free Ni2+ cations to cations to be be reduced reduced istohigher. cationsis higher. be reduced is higher. Figure 3. SEM Figure 3. SEM images Figure 3. SEM images of the of the the cross-section cross-section images of cross-section of ENP of ENP coatings of ENP coatingsfrom obtained coatings obtained from experiment experiment obtained from experiment number number (a)number (a) 1, (b) 2, (a) (c) 3, (d) 4, (e) 5, 1, (f) 6, (g) 7 and (b) (h)2, 8.(c) 3, (d) 4, (e) 5, (f) 6, (g) 7 and (h) 8. 1, (b) 2, (c) 3, (d) 4, (e) 5, (f) 6, (g) 7 and (h) 8. Bath components Bath components concentrations Bath componentsand concentrations and working parameters concentrations working parameters and working also have also have an effect effectalso parameters an on sur- on sur- have an effect on face topography of coatings as can be seen in Figure 4. All the coatings showed the face topography of coatings as can be seen in Figure 4. All the coatings showed the typical showed the surface topography of coatings as can be seen in Figure 4. All the typical coatings nodular growthtypical nodular growth with nodular growthendings with cauliflower-like cauliflower-like with cauliflower-like endings endings randomly randomly distributed randomly distributed on the on distributed the surface surface show- on the surface show- ing differences ing showing differences in in size size anddifferences and nodules in size nodules distribution.and nodules distribution. Among Among the distribution. the different Among different factors factors ablethe able to different impact factors able to to impact surface impactpH surface topography, topography, pHsurface and the and topography, the nature and nature pHconcentration and and the natureof concentration ofand concentration complexing complexing agents agents of complexing are the are the agents are most important the onesmost important and, ones presumably, and, they presumably, act together they (i.e., act mixed together effect) most important ones and, presumably, they act together (i.e., mixed effect) on surface to- (i.e., on mixed surface effect) to- on surface pography of topography electroless of nickel electroless coatings nickel [28–30]. pography of electroless nickel coatings [28–30]. coatings [28–30]. Figure 4. Top-view 4.SEM Figure 4. imagesSEM Top-view of ENP coatings images obtained of ENP ENP from coatings experiment obtained number (a) 1, from experiment experiment (b) 2, (c) number (a)3, 1,(d) (b)4,2, 2,(e) (c)5, (f) 6, (g) 7 Figure Top-view SEM images of coatings obtained from number (a) 1, (b) (c) and (h) 8. 3, (d) 4, 3, (d) 4, (e) (e) 5, 5, (f) (f) 6, 6, (g) (g) 77 and and (h) (h) 8. 8. 3.1.3. Hardness As can be seen in Figure 5, the hardness of the coatings, obtained from the different experimental conditions, shows that experiments 2, 3, 6 and 8 led to the highest coating hardness values, with a top value of 873 ± 31 Hv0.1 . These values were near those typically exhibited by hard chromium coatings (HCC) obtained by both direct current and pulse plating, which lie in the range of 1000–1100 Hv, and also near the hardness of some HCC substitutes like electrodeposited Ni-W alloys [31,32]. These results confirmed the preemi- nent position of electroless nickel coatings as a greener alternative to HCC coatings in terms of hardness. This fact along with the main benefits of electroless plating technology (i.e., thickness homogeneity, low concentration of metallic ions, tunable properties depending hardness values, with a top value of 873 ± 31 Hv0.1. These values were near those typically exhibited by hard chromium coatings (HCC) obtained by both direct current and pulse plating, which lie in the range of 1000–1100 Hv, and also near the hardness of some HCC substitutes like electrodeposited Ni-W alloys [31,32]. These results confirmed the preemi- Materials 2021, 14, 1501 nent position of electroless nickel coatings as a greener alternative to HCC coatings in 8 of 18 terms of hardness. This fact along with the main benefits of electroless plating technology (i.e., thickness homogeneity, low concentration of metallic ions, tunable properties de- pending on P on P content, content, no need no need to apply to apply an external an external electric electric field, field, etc.) make etc.) make ENPofone ENP one theof most the most promising alternatives to HCC for functional applications. promising alternatives to HCC for functional applications. Figure 5. Hardness Figure values 5. Hardness of electroless values nickel of electroless coatings nickel obtained coatings fromfrom obtained experiments 1–8. 1–8. experiments Figure Figure 1c indicates 1c indicates that that the most the most influential influential parameters parameters to secure to secure highhigh hardness hardness werewere pH and concentration of complexing agent 2, followed by temperature pH and concentration of complexing agent 2, followed by temperature and reducing agent and reducing agent concentration to a lesser extent. The optimum configuration concentration to a lesser extent. The optimum configuration of parameters to maximize of parameters to maximize hardness hardness of coatings of coatings was A+B+C+D+E−According was A+B+C+D+E−F+G+. F+G+. According to this parametric to this parametric setup, thesetup, pre- the prediction diction equationequation yielded yielded a hardness a hardness value of value 870 Hv of0.1870 Hv0.1 . fabricated . Coatings Coatings fabricated in the in the confir- confirmation experiment had a hardness of 863 ± 27 Hv mation experiment had a hardness of 863 ± 27 Hv0.1, thereby validating the model. 0.1 , thereby validating the model. Figure Figure 5 shows 5 shows thatthat thethe coatings coatings from from experiments experiments 3 and8 8exhibit 3 and exhibitthe thehighest highesthard- hardness ness values among all samples. Interestingly, these coatings also feature the highest P contents values among all samples. Interestingly, these coatings also feature the highest P con- tents(see (seeTable Table2). 2). The influence of The influence ofcomplexing complexingagentsagentsonon thethe P content P content hashas beenbeen reported reported elsewhere elsewhere for ENP for ENP layers layers due todue to their their ability ability for controlling for controlling the freetheNi free Ni2+ concentration 2+ concentration in in the electrolyte the electrolyte whichaffects which lately lately affects the composition the composition of the coating of the coating [33]. However, [33]. However, in our in our case, several complexing agents are playing a role simultaneously and their effect case, several complexing agents are playing a role simultaneously and their effect cannot cannot be straightforwardly deconvoluted. Complexing agents 1 and 2 can be regarded be straightforwardly deconvoluted. Complexing agents 1 and 2 can be regarded as “real” as “real” complexing agents, whereas other chemicals present in the electrolyte, like short complexing agents, whereas other chemicals present in the electrolyte, like short chained chained organic acids, act mainly as pH buffers. In double-complexed systems, it has been organic acids, act mainly as pH buffers. In double-complexed systems, it has been re- reported that higher ratios of complexing agents can lead to higher P percentages in the NiP ported that higher ratios of complexing agents can lead to higher P percentages in the NiP layers [34]. Experiment 3 had a (complexing 1/complexing 2) molar ratio of 0.236, whereas layers [34]. Experiment 3 had a (complexing 1/complexing 2) molar ratio of 0.236, whereas that for experiment 8 was of 0.083. Besides differences in the concentration of Ni (II) and that for experiment 8 was of 0.083. Besides differences in the concentration of Ni (II) and complexing agents, both systems exhibited similar free2+Ni2+ ions concentration. Namely, complexing2+agents, both systems exhibited similar free Ni ions concentration. Namely, free Ni concentration was 6.8 × 10−5 and 3.8 × 10−5 for experiments 3 and 8, respectively. Accordingly, experiment 8 yielded coatings with lower P content and enhanced crystallinity (Figure 2) in comparison with experiment 3, whose coatings featured higher P content and were mechanically harder (Figure 5). 3.1.4. Wear Resistance Evaluation of the mechanical performance of ENP coatings was completed with the analysis of their wear resistance. Figure 6 gathers the 3D images of worn surfaces from NiP coatings subject to wear tests. As can be inferred from Figure 6c,g, coatings from experiments 3 and 7 displayed the highest worn volumes among all samples. For the sake tent and were mechanically harder (Figure 5). 3.1.4. Wear Resistance Evaluation of the mechanical performance of ENP coatings was completed with the Materials 2021, 14, 1501 analysis of their wear resistance. Figure 6 gathers the 3D images of worn surfaces 9from of 18 NiP coatings subject to wear tests. As can be inferred from Figure 6c,g, coatings from ex- periments 3 and 7 displayed the highest worn volumes among all samples. For the sake of of comparison, comparison,the theworn wornvalues, values,together with together thethe with P content, thickness, P content, hardness, thickness, and cor- hardness, and rosion rates are listed in Table 4. corrosion rates are listed in Table 4. Figure Figure 6. 6. Confocal Confocal microscopy microscopy images images of of 3D 3D wear wear tracks tracks of of coatings coatings obtained obtained from from experiments experiments 1–8 1–8 (a–h) (a–h) and and subject subject to to wear tests. wear tests. Table 4. Summary of the values of the properties of the NiP coatings obtained from the eight ex- Table 4. Summary of the values of the properties of the NiP coatings obtained from the eight experiment setups. periment setups. P PThicknessThickness Hardness Hardness Worn WornVolume Volume Corrosion Corrosion Rate Rate L8 L8 3 2 wt.% wt.% µm µm Hv0.1Hv0.1 10−3−3)) mm (×(×10 mm3 µA/cm µA/cm2 1 7.1 ± 0.1 1 0.1 ± 0.111.6 ± 0.1 7.1 ±11.6 714 ±714 32 ± 32 0.049±±0.005 0.049 0.005 1.20 ± 1.20 0.04 ± 0.04 2 3.2 ± 0.2 2 0.2 ± 0.316.5 ± 0.3 3.2 ±16.5 853 ±853 23 ± 23 0.068±±0.012 0.068 0.012 0.30 ± 0.30 0.01 ± 0.01 3 7.8 ± 0.1 3 0.1 ± 1.624.4 ± 1.6 7.8 ±24.4 860 ±860 7 ±7 0.230±±0.085 0.230 0.085 0.50 ± 0.50 0.02 ± 0.02 4 6.4 ± 0.1 13.1 ± 0.3 792 ± 7 0.051 ± 0.003 0.40 ± 0.01 4 6.4 ± 0.1 13.1 ± 0.3 792 ± 7 0.051 ± 0.003 0.40 ± 0.01 5 2.2 ± 0.3 11.1 ± 0.2 788 ± 10 0.050 ± 0.004 0.20 ± 0.01 5 2.2 ± 0.3 11.1 ± 0.2 788 ± 10 0.050 ± 0.004 0.20 ± 0.01 6 6.7 ± 0.1 20.3 ± 0.3 834 ± 5 0.045 ± 0.008 1.50 ± 0.06 7 8.5 ± 0.3 6 6.7 ±25.6 0.1 ± 1.020.3 ± 0.3 777 ±834 28 ± 5 0.150 ±±0.021 0.045 0.008 1.50 ± 0.06 0.40 ± 0.03 8 4.0 ± 0.1 7 8.5 ±15.4 0.3 ± 0.425.6 ± 1.0 873 ±777 31 ± 28 0.150±±0.005 0.045 0.021 0.40 ± 0.30 ± 0.03 0.02 8 4.0 ± 0.1 15.4 ± 0.4 873 ± 31 0.045 ± 0.005 0.30 ± 0.02 Upon looking at Figure 1d, it was concluded that temperature, and the concen- trations of complexing agent 2 and reducing agent were the most influential param- eters for hardness. According to this, the optimal configuration of parameters was A+B−C−D+E−F−G−. The prediction equation rendered a worn volume value of 0.045 × 10−3 mm3 . The confirmation experiment analysis demonstrated that the resulting coating was characterized by a soft wear track. The worn volume determined from the 3D image was 0.045 × 10−3 mm3 , again confirming the model. Smoothness and the characteristic cauliflower-like microstructure of electroless nickel coatings greatly account for their good performance in wear, which is due to the self-lubricant character of NiP coatings [35]. The conditions optimized for maximum wear resistance simultaneously promoted a higher A+B−C−D+E−F−G−. The prediction equation rendered a worn volume value of 0.045 × 10−3 mm3. The confirmation experiment analysis demonstrated that the resulting coating was characterized by a soft wear track. The worn volume determined from the 3D image was 0.045 × 10-3 mm3, again confirming the model. Smoothness and the characteristic cauli- Materials 2021, 14, 1501 flower-like microstructure of electroless nickel coatings greatly account for their good 10 per- of 18 formance in wear, which is due to the self-lubricant character of NiP coatings [35]. The conditions optimized for maximum wear resistance simultaneously promoted a higher degree degree of of crystallinity crystallinityofofthe thecoatings, coatings,which whichexhibited exhibitedcrystallite sizes crystallite in in sizes thethe range of 5– range of 10 nm. The presence of small crystallites prevent movement of dislocations and 5–10 nm. The presence of small crystallites prevent movement of dislocations and other other de- formation deformationmechanisms mechanisms involved in plastic involved deformation in plastic due to deformation theto due higher number the higher of atoms number of placed at grain boundaries and triple junctions [9,36]. atoms placed at grain boundaries and triple junctions [9,36]. 3.1.5. Corrosion Resistance Corrosion Corrosion resistance resistanceoptimization optimizationwas wasbased basedonon the thelower-the-better lower-the-better criteria. TheThe criteria. ca- thodic cathodic andandanodic anodic polarization polarizationcurves curvesofofthe thedifferent differentcoatings coatingsare aredisplayed displayedin in Figure Figure 7. 7. Significant variations were observed in the corrosion potential potential (E (Ecorr corr)) values, values, with with differ- differ- ences higher higherthan than100 100mV, mV, suggesting suggesting thethe possibility possibility to prepare to prepare coatings coatings with variable with variable noble noble character character depending depending on theon the experimental experimental conditions. conditions. Nevertheless, Nevertheless, Ecorr Evalues corr values do do not give by itself information about the kinetics of the corrosion phenomena not give by itself information about the kinetics of the corrosion phenomena and, there- and, therefore, the corrosion fore, rate obtained the corrosion ratefrom icorr by obtained Tafelicorr from extrapolation is regarded as by Tafel extrapolation is aregarded better indicator [37]. as a better In this sense, indicator [37].coatings obtained In this sense, from experiments coatings obtained from 2, 5experiments and 8 showed 2, the lowest 5 and corrosion 8 showed the rate among lowest all therate corrosion samples among (Table all the4).samples The cathodic (Tablebranches indicatebranches 4). The cathodic that theirindicate Ecorr values that are inEthe their order of corr values experiments are in the order8 of< 2experiments < 5. 8 < 2 < 5. Figure 7. (a) Cathodic and (b) anodic branches of the polarization polarization curves for the various ENP coatings obtained through experiments 1–8. Overall, the Overall, the cathodic cathodicbranches branchesshown shownininFigure Figure7a7aindicate indicate a very a very similar similar behavior behavior in in all all samples samples except except forfor thecoating the coatingderived derivedfrom fromexperiment experiment7.7.In Inthis this case, case, an extended transition zone transition zone in in the the region region from −0.35toto−0.70 from −0.35 −0.70VVvs.vs. Ag/AgCl Ag/AgCl waswas observed observed before before the the diffusion-controlled diffusion-controlled region. region. Meanwhile, Meanwhile, the other the other samplessamples all showed all showed a diffusion- a diffusion-con- controlled trolled region region fromfrom−0.5−0.5 V vs.V vs. Ag/AgCl. Ag/AgCl. At potentials At potentials moremore negative negative −1.0−V thanthan 1.0vs. V vs. Ag/AgCl, a change in the slope of the cathodic branch is observed, Ag/AgCl, a change in the slope of the cathodic branch is observed, indicating the begin-indicating the beginning ning of hydrogen of hydrogen evolutionevolution as described as described elsewhere elsewhere [38]. [38]. The anodic branches indicate that coatings resulting from experiments 2 and 5 are among those with higher corrosion resistance (see Figure 7b). In particular, coatings from experiments 2 and 5 showed more positive Ecorr values than the coating from experiment 8, although all of them exhibit similar anodic slopes giving rise to comparable corrosion rates (see Table 4). In the three cases, the curves exhibited a pseudo-passive region for values of +150 mV vs. Ecorr . The sample from experiment 7 displayed a similar behavior, although it showed a slightly higher corrosion rate. Experiments 2, 5 and 8 corresponded to coatings with high crystallinity (Figure 2), characterized by the occurrence of the (200) plane and the (111) preferred orientation of the fcc nickel phase. High corrosion resistance has already been reported in aerated NaCl media for nanocrystalline NiP coatings irrespective of the P content, with corrosion rates comparable to those obtained in this study for the best performing coatings [38]. Except for experiment 7 derived coating, samples with values of +150 mV vs. Ecorr. The sample from experiment 7 displayed a similar behavior, although it showed a slightly higher corrosion rate. Experiments 2, 5 and 8 corresponded to coatings with high crystallinity (Figure 2), characterized by the occurrence of the (200) plane and the (111) preferred orientation of the fcc nickel phase. High corrosion resistance Materials 2021, 14, 1501 has already been reported in aerated NaCl media for nanocrystalline NiP coatings 11 of 18 irre- spective of the P content, with corrosion rates comparable to those obtained in this study for the best performing coatings [38]. Except for experiment 7 derived coating, samples withP low low P contents contents (<4 wt.% (<4 P) wt.% P) showed showed lower lower corrosion corrosion rates, which rates, which is directly is directly linkedlinked with with their crystalline structure. The transition from nanocrystalline their crystalline structure. The transition from nanocrystalline to mixed nanocrystalline– to mixed nanocrystal- line–amorphous amorphous structure structure corresponds corresponds to coatings to coatings with medium with medium wt.% P wt.% P and negatively and negatively affects the corrosion performance. This is due to the presence of smaller crystalscrystals affects the corrosion performance. This is due to the presence of smaller which which create createactive more moresites active prone sites to prone to corrosion corrosion attackgrain attack through through grain and and nodule nodule boundaries. boundaries. Moreover, Moreover, P enrichment at the outer coatings’ surface can occur under P enrichment at the outer coatings’ surface can occur under anodic polarization conditions, anodic polarization conditions, which which is caused by is caused by nickel preferential preferential nickelaccording dissolution dissolution to according to previous previous studies stud- [14,39,40]. ies [14,39,40]. Yet, this effect was Yet, this effect was not observed in our case.not observed in our case. Accordingto According toFigure Figure1e, 1e,the theanalysis analysisof ofthe themost mostinfluential influentialparameters parametersrevealed revealedthat that higher pH, higher pH, lower lower stirring stirring rate rateand andhigher higherreducing reducingagent agentconcentration concentrationled ledtotocoatings coatings characterized by lower characterized lower corrosion corrosionrates. rates.Complexing Complexing agents agentsandandstabilizers present stabilizers in ENP present in bathsbaths ENP havehave beenbeen described as corrosion described performance as corrosion performance modifiers [41,42]. modifiers However, [41,42]. However, the con- the centration of the concentration complexing of the complexing agents agents had had little influence little influence onon thethe corrosion corrosion performance performance of the of thestudied studied NiP NiP coatings coatings according according toto thetheS/N S/N ratio ratiovalues valuesin in Figure Figure1e.1e. Onthe On theother otherhand, hand,the thenotorious notoriousinfluence influenceofof thethe stirring stirring rate rate ononthethe corrosion corrosion re- resis- sistance tance of coatings of the the coatings was was somewhat somewhat unforeseen. unforeseen. UponUpon analyzing analyzing the interactions the interactions between be- factors, Ni (II) and tween factors, Ni (II)complexing agent 2agent and complexing concentrations 2 concentrationsshowed a strong showed interaction a strong with interaction regard to the to with regard corrosion rate asrate the corrosion suggested by the by as suggested x-shaped plot depicted the x-shaped in Figure plot depicted in 8b, which Figure 8b, indicates that the stirring effect could be poisoned by this interaction. which indicates that the stirring effect could be poisoned by this interaction. Moreover, Moreover, the in- teraction between factorsfactors A and A B (i.e., − ), although the interaction between and concentrations B (i.e., concentrationsof Ni (II)ofand H2 PO Ni (II) and 2 H 2PO2−), alt- present, was notwas hough present, considered significant not considered according significant to common according knowledge to common in electroless knowledge in elec- plating troless (Figure plating 8a). (Figure 8a). Figure8.8.Interaction Figure Interactionplots plotsfor forsome someofofthe thefactors factorsconsidered consideredininthe theLL88matrix matrixfor for(a) (a)Ni Ni(II) (II)and andreducing reducingagent, agent,(b) (b)Ni Ni(II) (II) and andcomplexing complexingagent agent2,2,and and(c) (c)reducing reducingagent agentand andcomplexing complexingagent agent2 2concentrations. concentrations. InInorder ordertotoelucidate the elucidate real the contribution real contributionof the interaction of the between interaction the concentrations between the concentra- of tions of Ni (II) (factor 1) and complexing agent 2 (factor 4), a dedicated Taguchistudy Ni (II) (factor 1) and complexing agent 2 (factor 4), a dedicated Taguchi studywas was conducted using an L4 matrix, which allowed the study of two factors conducted using an L4 matrix, which allowed the study of two factors at two differentat two different levels levels in in aa separate separate manner, manner,including includingthetheeffect effectof of their their interaction interaction on on the the corrosion corrosion per- performance (Table formance (Table 5). 5). Table 5. Experimental matrix derived from dedicated L4 OA. Factors and Experiments A Ni (II) g/L B Reducer g/L Interaction A-B 1 4.8 7 - 2 4.8 12 - 3 6.5 7 - 4 6.5 12 - The effect plot of the different factors, and their corresponding interaction, evaluated in the dedicated Taguchi study can be shown in Figure 9. 1 4.8 7 - 1 4.8 7 - 2 4.8 12 - 2 4.8 12 - 3 6.5 7 - 3 6.5 7 - 4 6.5 12 - 4 6.5 12 - Materials 2021, 14, 1501 12 of 18 The effect plot of the different factors, and their corresponding interaction, evaluated The effect plot of the different factors, and their corresponding interaction, evaluated in the dedicated Taguchi study can be shown in Figure 9. in the dedicated Taguchi study can be shown in Figure 9. Figure 9. Main effect plot for a dedicated Taguchi design to determine the influence of Ni (II) and complexing agent 2 Main effect Figure 9. Main effect plot plot for for a dedicated dedicated Taguchi design to determine the influence of Ni (II) and complexing agent 2 concentrations on the corrosion performance of ENP coatings. concentrations on the corrosion performance of ENP coatings. The Thedetailed detailedanalysis analysisofof ofthis thisinteraction interactionrevealed revealedthat thatititwas wasindeed indeednot notsignificant significant The detailed analysis this interaction revealed that (Figure (Figure 10), 10),thus thus validating validating our our previous previous outcome outcome that thatstirring stirringwas waskey keyto toachieve achieve the the (Figure 10), thus validating our previous outcome that stirring was key lowest corrosion lowest corrosion rate. rate. lowest corrosion rate. Figure10. Figure Interaction 10.Interaction plotplot for for a dedicated a dedicated L4 matrix L4 matrix to determine to determine the influence the influence of and of Ni (II) Ni (II) com- and Figure 10. Interaction complexing plot for a dedicated L4 matrixperformance to determine the theinfluence of Ni (II) and com- plexing agentagent concentrations concentrations on theon the corrosion corrosion performance of theofENP ENP coatings. coatings. plexing agent concentrations on the corrosion performance of the ENP coatings. Oncethe Once theeffect effectofofall allfactors factorswas wasconfirmed, confirmed,the theprediction predictionequation equationfor forcorrosion corrosion Once was resistance the effect defined of as allA+B+C factors−was D+E+Fconfirmed, − G+, the prediction yielding a equation theoretical value for for corrosion corrosion resistance was defined as A+B+C−D+E+F−G+, yielding a theoretical value for corrosion resistance rateofof0.3 was 0.3µA/cm µA/cm defined 2 .A as A+B+C−D+E+F−G+, confirmationexperiment yielding experimentconducted a theoretical conductedunder underthe value theoptimal for corrosion optimalparameters parameters rate 2. A confirmation rate of 0.3 resulted sequence µA/cm2. in A confirmation coatings withexperiment a corrosionconducted rate of 0.25 under ± 0.05theµA/cm optimal 2 , parameters confirming sequence resulted in coatings with a corrosion rate of 0.25 ± 0.05 µA/cm ,2 confirming the 2 sequence resulted the predicted value inderived coatingsfromwiththe a corrosion model. Thisrateresult of 0.25 ± 0.05 also µA/cmthe confirms , confirming importancethe of predicted value derived from the model. This result also confirms the importance of stir- predicted stirring invalue derived electroless fromplating. nickel the model. It isThis result also advisable confirms to stir the importance the solution while plating, of stir- in ring in electroless nickel plating. It is advisable to stir the solution while plating, in order ring orderintoelectroless nickel plating. avoid hydrogen pittingItand is advisable to stir the the occurrence solution although of patterns, while plating, in order too vigorous to avoid hydrogen pitting and the occurrence of patterns, although too vigorous agitation to avoid hydrogen agitation pitting can negatively andthe affect theplating occurrence of patterns, process although [43]. Moderate too vigorous stirring rates are agitation therefore can negatively affect the plating process [43]. Moderate stirring rates are therefore recom- can negatively to recommended affect thehigher secure platingdiffusion process rates, [43]. Moderate conveyed stirring rates are by convection, in therefore recom- order to facilitate mended to secure higher diffusion rates, conveyed by convection, in order to facilitate the mended to secure higher diffusion rates, conveyed by convection, in order to the arrival of fresh reactants to the substrate through the diffusion layer. On the other hand, facilitate the arrival of fresh reactants to the substrate through the diffusion layer. On the other hand, arrival higher of fresh reactants agitation regimesto thelead can substrate to NiPthrough coatingsthe diffusion layer. characterized On the other by increased hand, corrosion rates in chloride media [44]. Suitable agitation promotes, in acid electrolytes, the increase in pH in the diffusion layer which, in turn, facilitates the incorporation of low amounts of P in the layers as well as the growth of more compact and smoother coatings [7]. As a result, compact NiP layers with fewer defects are obtained, which ultimately causes a better response in electrochemical corrosion tests in chloride media. 3.2. Multivariate Optimization The design of novel coatings is typically aimed at enhancing a given property without considering the effect that the experimental parameters might have on other properties. In this study, the most favorable set of factors have been independently determined following a univariate optimization approach in the preceding section, which is summarized in (Table 6). 3.2. Multivariate Optimization The design of novel coatings is typically aimed at enhancing a given property with- out considering the effect that the experimental parameters might have on other proper- ties. In this study, the most favorable set of factors have been independently determined Materials 2021, 14, 1501 following a univariate optimization approach in the preceding section, which is summa- 13 of 18 rized in (Table 6). Table 6. Summary of the univariate optimization results for the features or properties considered Table 6. Summary of the univariate in this study. optimization results for the features or properties considered in this study. Feature Feature Units Units Equation Equation Predicted Predicted Experimental Experimental P content P wt.% content wt.% A+B−C+D+E+F−G+ A+B−C+D+E+F−G+ 3.3 3.3 3.13.1 ± 0.2 ± 0.2 Thickness Thickness µm A+B+Cµm A+B+C−D−E−F+G- −D−E−F+G- 24.1 24.1 24.3 24.3 ± 1.6 ± 1.6 Hardness HvHardness 0.1 Hv0.1 −F+G+ A+B+C+D+E A+B+C+D+E−F+G+870 870 ± 27 863863 ± 27 (×10−3 ) mm3 −C−3−D+E+F Worn volume Worn2volume A+B (×10 ) mm3 −G − 0.045 A+B−C−D+E+F−G− 0.045 0.045 ± 0.005 0.045 ± 0.005 Corrosion rate µA/cm A+B+C−D+E+F−G+ 0.30 0.25 ± 0.03 Corrosion rate µA/cm2 A+B+C−D+E+F−G+ 0.30 0.25 ± 0.03 On the On the other other hand, hand, in in multivariate multivariate optimization, optimization, moremore than than one feature is one feature considered is considered in the in the same same optimization optimization run, run, thus thus allowing allowing for for aa more more realistic realistic coating coating design. design. This This isis of of utmost importance in real world applications, where coatings are exposed utmost importance in real world applications, where coatings are exposed to more than to more than one degradation one degradationmechanism mechanism(i.e.,(i.e.,wear-corrosion, wear-corrosion, fatigue-corrosion, fatigue-corrosion, etc.) etc.) during during theirtheir ser- vice life [45–47]. Hence, multivariate optimization was run for all the features that were service life [45–47]. Hence, multivariate optimization was run for all the features that were considered separately considered separately in in the the previous previous univariate univariate optimization. optimization. Several approaches for multivariate optimization based Several approaches for multivariate optimization based onon the the Taguchi Taguchi methodology methodology exist such exist such asas grey grey relational relational analysis, analysis, genetic genetic algorithms, algorithms, etc.etc. [20,48,49]. [20,48,49]. The The multivariate multivariate optimization workflow used in this study is shown optimization workflow used in this study is shown in Figure 11.in Figure 11. Figure 11. Multivariate Figure 11. Multivariate optimization optimization workflow used in workflow used in this this study. study. The relation between different factors should be first evaluated to check whether they can be expressed as a lineal combination, which permits to reduce the dimension of the original matrix. Several methods can be used for reduction of data dimensionality. Of them, the calculation of correlation coefficient was selected in this study. The correlation coefficient r is given by Equation (1), where n corresponds to the number of samples, and xi and yi to the values of the different factors involved in the comparison. n·(∑ni=1 xi ·yi ) − (∑ni=1 xi )·(∑ni=1 yi ) r= q q (1) n ∑ni=1 x2i − (∑ni=1 xi )2 · n ∑ni=1 y2i − (∑ni=1 yi )2   An r = 0 means that there is not a lineal correlation between factors, r = 1 means perfect positive correlation between factors, and 0 < r < 1 indicates a certain degree of lineal correlation. The same is true for negative correlations, for which r = −1. Using the Materials 2021, 14, 1501 14 of 18 values shown in Table 4, and applying Equation (1), the resulting correlation matrix is (Equation (2)): r11 r12 r13 r14 r15 r16 1 0.634 0.281 0.670 0.495 0.442       r21 r22 r23 r24 r25 r26     0.634 1 0.364 0.201 0.741 0.055    r31 r32 r33 r34 r35 r36   0.281 0.364 1 0.410 0.211 0.302  r= =  (2)   r41 r42 r43 r44 r45 r46   0.670 0.201 0.410   1 0.489 0.480    r51 r52 r53 r54 r55 r56   0.495 0.741 0.211 0.489 1 0.232  r61 r62 r63 r64 r65 r66 0.442 0.055 0.302 0.480 0.232 1 The matrix depicted in Equation (2) shows that the highest value for the correlation coefficient among the different factors correspond to position r25 = 0.741 for thickness and worn volume features. Yet, although there is apparently some degree of correlation, the value does not support the existence of a lineal combination. Thus, it was not possible to reduce data dimensionality as there were no lineal combinations of variables. According to the procedure described in Figure 11, data resulting from the univariate optimization were transformed into a normal distribution using the Zscore, as shown in Equation (3), where x is the value of each feature in the different experiments, µ is the average of all the measurements for that feature, and σ the standard deviation. (x − µ) Z= (3) σ A synthesis indicator (SI) was defined for the different features depending on the specific weight a definite feature would have on the overall response. Thus, according to this research, minimum wear volume and lowest corrosion resistance were the features with the highest impact in the response derived from the multivariate optimization, followed by the P content, thickness and hardness of the coatings. As the optimization of wear and corrosion resistance fell within the lower-the-better category, the sign of these features was negative. The SI values for the different properties are listed in Table 7. Table 7. Synthesis indicator (SI) values employed in the multivariate optimization. SI P Thickness Hardness Wear Corrosion value −0.2 0.1 0.1 −0.35 −0.25 By applying the SI to the obtained values, it was possible to elucidate the main effects on wear and corrosion resistance in an analogous manner to the Taguchi methodology used in the univariate optimization, but considering the two properties simultaneously. Optimal factors were achieved in positions for which the overall effect is maximized. Table 8 shows the main effect values, their ranking within the factors considered, and the optimal level (+, −) for each factor. Table 8. Main effect chart on wear and corrosion resistance of ENP coatings for multivariate optimization. Factor [Ni (II)] [Reducer] [Complexing 1] [Complexing 2] Stirring T pH − −0.131 0.080 0.057 −0.251 −0.204 0.186 −0.258 + 0.131 −0.080 −0.057 0.251 0.204 −0.186 0.258 Main effect 0.263 0.160 0.115 0.502 0.408 0.372 0.516 Rank 5 6 7 2 3 4 1 Optimal level + − − + + − + Table 8 demonstrates that the most influential parameters to secure coatings with the lowest wear volume and highest corrosion resistance were pH, complexing agent 2 concentration, stirring and temperature, followed by Ni (II), reducing agent and complexing agent 1 concentrations. Thus, the optimized factors arrangement for multivariate analysis was A+B−C−D+E+F−G+. In order to obtain the prediction equation for each individual Materials 2021, 14, 1501 15 of 18 feature, the results were brought back into the univariate Taguchi matrix, where the new configuration of factors was implemented. The final projected values for each feature are shown in Table 9. Table 9. Predicted values for all the properties subject to multivariate optimization. Feature Units Predicted P content wt.% 2.2 Thickness µm 10.8 Hardness Hv0.1 842 Worn volume (×10−3 ) mm3 0.045 Corrosion rate µA/cm2 0.02 According to the predicted values, lower phosphorus content and lower corrosion rates could be achieved through a multivariate optimization, thus improving the results achieved in the univariate optimization (cf. Table 5). It should be pointed out that a different arrangement of factors was obtained compared to the univariate optimization, thus demonstrating the suitability of the multivariate approach for the optimization of this kind of multifactorial systems. Results of the Confirmation Tests P content, thickness, hardness, wear and corrosion resistance were evaluated on coat- ings obtained from the formulation and working parameters derived from the multivariate optimization and shown in Table 7. Phosphorous content evaluated by EDX amounted to 2.2 ± 0.2 wt.%, thereby confirming the predicted value. Mean coating thickness was 12.1 ± 0.4 µm (Figure 12a), which was slightly higher than the projected value, but in the range of the model considering the standard deviation of the measurement. Figure 12. Confirmation test results for (a) thickness, (b) wear and (c) corrosion performance of coatings obtained from multivariate optimization. Hardness measurements yielded a value of 851 ± 27 Hv0.1 , which was again in full agreement with the prediction. The worn volume determined after a tribological test was 0.068 ± 0.010 × 10−3 mm3 (Figure 12b). This value was slightly higher than the predicted value but of the same order of magnitude, and it was also higher than the best value achieved in the univariate optimization (cf. Table 5). Note that this is a global optimization and therefore some features may not reach their optimal values. Finally, the corrosion resistance of the coatings produced for the confirmation test was assessed by electrochemical techniques. The polarization curve (cathodic plus anodic branches) of the coating obtained in the confirmation experiment is shown in Figure 12c. The corrosion rate was 0.070 ± 0.02 µA/cm2 , which was moderately higher than the 0.020 µA/cm2 value predicted by the model. Nevertheless, it is worth pointing out that the value obtained after the multivariate optimization is one order of magnitude lower than the values obtained in the univariate optimization (cf. Table 4). Despite the fact that Materials 2021, 14, 1501 16 of 18 corrosion rate was not the feature with the highest specific weight in the optimization run, the experimental setup derived from the multivariate optimization had a positive effect on the corrosion performance of the coatings. In particular, higher stirring rate and lower plating rate (favored by a lower temperature) boosted the corrosion resistance, as observed in the univariate optimization (see Table 5). Low plating rate allows the growth of more compact electroless nickel coatings with lesser defects, thereby improving the corrosion performance of the NiP coatings. On the other hand, higher pH promotes the incorporation of lower amounts of P in the nickel matrix, which is beneficial for the mechanical properties. Table 10 summarizes the predicted and experimental results obtained from the multi- variate optimization. The experimental values match very well with the predicted values, validating the methodology proposed for the optimization of complex electroless nickel formulations. Remarkably, the optimized solution renders NiP coatings with three times lower corrosion rates than the value obtained in the univariate optimization, without compromising hardness and wear performance. Table 10. Predicted and experimental values of ENP coatings subject to multivariate optimization. Feature Units Predicted Experimental P content wt.% 2.2 2.2 ± 0.2 Thickness µm 10.8 12.1 ± 0.4 Hardness Hv0.1 842 851 ± 27 Wear (×10−3 mm3 ) 0.045 0.068 ± 0.010 Corrosion (µA/cm2 ) 0.02 0.07 ± 0.02 4. Conclusions The Taguchi method has been adopted for the optimization of a complex multicompo- nent electroless nickel formulation. In this study, two approaches have been employed for optimizing different features or properties, first individually (univariate approach) and later globally (multivariate approach). In either case, it has been possible to obtain valuable information from a reduced number of experiments. The P content, thickness, hardness, wear volume and corrosion resistance were chosen as the optimizable features. From the compositional point of view, the role of complexing agent 2 in modulating hardness and wear resistance of the coatings needs to be highlighted. On the other hand, pH and stirring rate have been determined as the most influential parameters for achieving coatings with low corrosion rates. Moreover, lower operation temperature favors the growth of more compact and defect-free NiP coatings. As opposed to univariate optimization, multivariate optimization allows improving more than one feature at the same time. In this study, wear and corrosion resistance were selected as features with the highest specific weight in the multivariate optimization response. The optimal experimental setup derived from the multivariate optimization rendered a configuration of factors which was different from any of the combinations obtained in the univariate optimization. Moreover, it has been possible to obtain coatings exhibiting three times lower corrosion rates in comparison to the best result achieved in the univariate optimization stage for this feature. Surprisingly, corrosion resistance was not the feature with the highest specific weight on the overall response but was best optimized under the given set of factors. Accurate design of experiments is as a powerful tool for the optimization of complex electroless nickel electrolytes, providing useful information with a limited number of trials, making this methodology both time and cost effective. The proposed univariate and multivariate optimization methodologies could be extrapolated to other complex systems in which bath components (and their concentrations) and experimental conditions play important roles on the performance of the coatings. Author Contributions: Conceptualization, A.S.-P., E.G.-L.; methodology and investigation, I.U. and A.S.-P.; formal analysis, A.S.-P., E.G.-L., J.S., E.P.; writing—original draft preparation, A.S.-P., E.G.- Materials 2021, 14, 1501 17 of 18 L.; writing—review and editing, A.S.-P., E.G.-L., J.S., E.P.; supervision, E.P., E.G.-L., J.S.; project administration, A.S.-P., E.G.-L.; funding acquisition, E.P., E.G.-L., J.S. All authors have read and agreed to the published version of the manuscript. Funding: This work was partially supported by the Spanish Government (MAT2017-86357-C3- 1-R, MAT2017-86357-C3-2-R projects), Ministerio de Ciencia e Innovación of the Spanish Gov- ernment (project SURF-ERA, EXP-00137314/CER-20191003) and by the Departamento de Desar- rollo Económico y Competitividad of the Basque Government (ELKARTEK, KK-2019/00077, KK- 2020/00108). Partial financial support from the Generalitat de Catalunya (2017-SGR-192) is also acknowledged. Institutional Review Board Statement: Not applicable. Informed Consent Statement: Not applicable. Data Availability Statement: The data presented in this study are available on request from the corresponding author. Conflicts of Interest: The authors declare no conflict of interest. References 1. Brenner, A.; Riddell, G. Deposition of nickel and cobalt by chemical reduction. J. Res. Natl. Inst. Stand. Technol. 1947, 39, 385. [CrossRef] 2. Sudagar, J.; Lian, J.; Sha, W. Electroless nickel, alloy, composite and nano coatings-A critical review. J. Alloy. Compd. 2013, 571, 183–204. [CrossRef] 3. Diegle, R.B.; Sorensen, N.R.; Clayton, C.R.; Helfand, M.A.; Yu, Y.C. An XPS Investigation into the Passivity of an Amorphous Ni-20P Alloy. J. Electrochem. Soc. 1988, 135, 1085–1092. [CrossRef] 4. Elsener, B.; Crobu, M.; Scorciapino, M.A.; Rossi, A. Electroless deposited Ni–P alloys: Corrosion resistance mechanism. J. Appl. Electrochem. 2008, 38, 1053–1060. [CrossRef] 5. Zhang, H.; Zou, J.; Lin, N.; Tang, B. Review on electroless plating Ni–P coatings for improving surface performance of steel. Surf. Rev. Lett. 2014, 21, 1–13. [CrossRef] 6. Zhang, B. Amorphous and Nano Alloys Electroless Depositions; Elsevier: Amsterdam, The Netherlands, 2016. 7. Riedel, W. Electroless Nickel Plating; ASM International: Cleveland, OH, USA, 1991. 8. Gawrilov, G.G. Chemical (Electroless) Nickel Plating; Portcullis Press Ltd.: London, UK, 1979. 9. Yanhai, C.; Hengyang, C.; Zhencai, Z.; Yong, Z.; Yuxing, P. Tribological Behavior of Ni-P Deposits on Dry Condition. Rare Met. Mater. Eng. 2014, 43, 11–16. [CrossRef] 10. Taheri, R.; Oguocha, I.; Yannacopoulos, S. The tribological characteristics of electroless NiP coatings. Wear 2001, 249, 389–396. [CrossRef] 11. Mimani, T.; Mayanna, S. The effect of microstructure on the corrosion behaviour of electroless Ni P alloys in acidic media. Surf. Coat. Technol. 1996, 79, 246–251. [CrossRef] 12. Shen, Y.; Xue, W.; Liu, Z.; Zuo, L. Nanoscratching deformation and fracture toughness of electroless Ni–P coatings. Surf. Coat. Technol. 2010, 205, 632–640. [CrossRef] 13. Hamada, A.; Sahu, P.; Porter, D. Indentation property and corrosion resistance of electroless nickel–phosphorus coatings deposited on austenitic high-Mn TWIP steel. Appl. Surf. Sci. 2015, 356, 1–8. [CrossRef] 14. SriBalaji, M.; Arunkumar, P.; Babu, K.S.; Keshri, A.K. Crystallization mechanism and corrosion property of electroless nickel phosphorus coating during intermediate temperature oxidation. Appl. Surf. Sci. 2015, 355, 112–120. [CrossRef] 15. Farzaneh, A.; Ehteshamzadeh, M.; Mohammadi, M. Corrosion performance of the electroless Ni–P coatings prepared in different conditions and optimized by the Taguchi method. J. Appl. Electrochem. 2010, 41, 19–27. [CrossRef] 16. Khosrow, D. Quality Control, Robust Design, and the Taguchi Method; Springer: Berlin, Germany, 1988. 17. Bonin, L.; Castro, C.; Vitry, V.; Hantson, A.-L.; Delaunois, F. Optimization of electroless NiB deposition without stabilizer, based on surface roughness and plating rate. J. Alloy. Compd. 2018, 767, 276–284. [CrossRef] 18. American Iron and Steel Institute (AISI). AISI 1010 Carbon Steel. AISI 1010; American Iron and Steel Institute: Washington, DC, USA, 1908. 19. Ishrat, S.I.; Khan, Z.A.; Siddiquee, A.N.; Badruddin, I.A.; AlGahtani, A.; Javaid, S.; Gupta, R. Optimising Parameters for Expanded Polystyrene Based Pod Production Using Taguchi Method. Mathematics 2019, 7, 847. [CrossRef] 20. Yang, W.; Tarng, Y. Design optimization of cutting parameters for turning operations based on the Taguchi method. J. Mater. Process. Technol. 1998, 84, 122–129. [CrossRef] 21. Bruker EVA, Version 2020. Software for the Analysis of 1D and 2D X-ray Datasets Including Visualization, Data Reduction, Phase Identification and Quantification, Statistical Evaluation; Bruker: Billerica, MA, USA, 2020. 22. Shahavi, M.H.; Hosseini, M.; Jahanshahi, M.; Meyer, R.L.; Darzi, G.N. Clove oil nanoemulsion as an effective antibacterial agent: Taguchi optimization method. Desalination Water Treat. 2015, 57, 18379–18390. [CrossRef] Materials 2021, 14, 1501 18 of 18 23. Chen, K.; Liu, C.; Whalley, D.C.; Hutt, D.A.; Li, J.F.; Mannan, S.H. Electroless Ni-W-P alloy coatings as barrier metallisations for liquid solder interconnects. In Proceedings of the 2006 1st Electronic Systemintegration Technology Conference, Dresden, Germany, 5–7 September 2006. 24. Keong, K.G.; Sha, W. Crystallisation and Phase Transformation Behaviour of Electroless Nickel-Phosphorus Deposits and Their Engineering Properties. Surf. Eng. 2002, 18, 329–343. [CrossRef] 25. Palaniappa, M.; Seshadri, S. Friction and wear behavior of electroless Ni-P and Ni-W-P alloy coatings. Wear 2008, 265, 735–740. [CrossRef] 26. Shu, X.; He, Z.; Wang, Y.; Yin, L. Mechanical properties of Ni-based coatings fabricated by electroless plating method. Surf. Eng. 2020, 36, 944–951. [CrossRef] 27. Popov, K.I.; Djokić, S.S.; Nikolić, N.D.; Jović, V.D. Morphology of Electrochemically and Chemically Deposited Metals; Springer International Publishing: Berlin, Germany, 2016. 28. Huang, Y.; Cui, F. Effect of complexing agent on the morphology and microstructure of electroless deposited Ni-P alloy. Surf. Coat. Technol. 2007, 201, 5416–5418. [CrossRef] 29. Rahman, A.; Jayaganthan, R. Effect of pH values on nanostructured Ni-P films. Appl. Nanosci. 2015, 5, 493–498. [CrossRef] 30. Małecki, A.; Micek-Ilnicka, A. Electroless nickel plating from acid bath. Surf. Coat. Technol. 2000, 123, 72–77. [CrossRef] 31. Imaz, N.; Ostra, M.; Vidal, M.; Díez, J.; Sarret, M.; García-Lecina, E. Corrosion behaviour of chromium coatings obtained by direct and reverse pulse plating electrodeposition in NaCl aqueous solution. Corros. Sci. 2014, 78, 251–259. [CrossRef] 32. Imaz, N.; Diez, J.A.; Ostra, M.; Sarret, M.; García-Lecina, E. Advanced Ni–W coatings obtained by the combination of pulse plating and chemometric techniques. Trans. IMF 2014, 92, 305–315. [CrossRef] 33. Ashtiani, A.A.; Faraji, S.; Iranagh, S.A.; Faraji, A.H. The study of electroless Ni–P alloys with different complexing agents on Ck45 steel substrate. Arab. J. Chem. 2017, 10, S1541–S1545. [CrossRef] 34. Cui, G.; Li, N.; Li, D.; Chi, M. Study of Optimized Complexing Agent for Low-Phosphorus Electroless Nickel Plating Bath. J. Electrochem. Soc. 2005, 152, C669–C674. [CrossRef] 35. Sahoo, P.; Das, S.K. Tribology of electroless nickel coatings–A review. Mater. Des. 2011, 32, 1760–1775. [CrossRef] 36. Apachitei, I.; Duszczyk, J. Autocatalytic nickel coatings on aluminium with improved abrasive wear resistance. Surf. Coat. Technol. 2000, 132, 89–98. [CrossRef] 37. Wang, L.L.; Chen, H.J.; Hao, L.; Lin, A.; Gan, F.X. Electrochemical corrosion behavior of electroless Ni-P coating in NaCl and H2SO4 solutions. Mater. Corros. 2010, 62, 1003–1007. [CrossRef] 38. Crobu, M.; Scorciapino, A.; Elsener, B.; Rossi, A.; Scorciapino, M.A. The corrosion resistance of electroless deposited nano- crystalline Ni–P alloys. Electrochim. Acta 2008, 53, 3364–3370. [CrossRef] 39. Zhang, Y.Z.; Yao, M. Studies of Electroless Nickel Deposits with Low Phosphorus Content. Trans. IMF 1999, 77, 78–83. [CrossRef] 40. Lo, P.-H.; Tsai, W.-T.; Lee, J.-T.; Hung, M.-P. Role of phosphorus in the electrochemical behavior of electroless Ni-P alloys in 3.5 wt.% NaCl solutions. Surf. Coat. Technol. 1994, 67, 27–34. [CrossRef] 41. Van Gool, A.P.; Boden, P.J.; Harris, S.J. Corrosion Behaviour of some Electroless Nickel—Phosphorus Coatings. Trans. IMF 1987, 65, 108–114. [CrossRef] 42. Bahramian, A.; Eyraud, M.; Vacandio, F.; Knauth, P. Improving the corrosion properties of amorphous Ni-P thin films using different additives. Surf. Coat. Technol. 2018, 345, 40–52. [CrossRef] 43. Dearden, J. Electroless Plating—Its Applications in Resistor Technology. Electrocompon. Sci. Technol. 1976, 3, 103–111. [CrossRef] 44. Zeller, R.L. Electrochemical Corrosion Testing of High Phosphorus Electroless Nickel in 5% NaCl. Corrosion 1991, 47, 692–702. [CrossRef] 45. Flis, J.; Duquette, D.J. Effect of Phosphorus on Anodic Dissolution and Passivation of Nickel in Near-Neutral Solutions. Corrosion 1985, 41, 700–706. [CrossRef] 46. Meng, M.; Leech, A.; Le, H. Mechanical properties and tribological behavior of electroless Ni–P–Cu coatings on corrosion-resistant alloys under ultrahigh contact stress with sprayed nanoparticles. Tribol. Int. 2019, 139, 59–66. [CrossRef] 47. Panja, B.; Sahoo, P. Tribo-corrosion Behavior of Electroless Ni-P Coatings in Alkaline Corrosive Environment. Port. Electrochim. Acta 2014, 32, 303–313. [CrossRef] 48. Lee, J.-H.; H, W.; B, Y.C.; T, C.K.; Y, H.C. Parametric Optimization for Electroless Ni–P on Low Carbon Steel. J. Comput. Theor. Nanosci. 2016, 13, 1640–1648. [CrossRef] 49. Jeyapaul, R.; Shahabudeen, P.; Krishnaiah, K. Simultaneous optimization of multi-response problems in the Taguchi method using genetic algorithm. Int. J. Adv. Manuf. Technol. 2006, 30, 870–878. [CrossRef]