Parametric optimization of corrosion and wear of electroless Ni-P-Cu coating using grey relational coefficient coupled with weighted principal component analysis

Abstract

Background

This research article considers optimization of the four process parameters based on corrosion and wear of electroless Ni-P-Cu coatings. The major characteristics indexes for performance selected to evaluate the processes are corrosion potential (E _corr), corrosion current density (I _corr) and wear. Among the corresponding four process parameters the first three are coating parameters, viz. concentration of nickel sulphate, concentration of sodium hypophosphite, concentration of copper sulphate and the fourth one is post deposition heat treatment temperature.

Methods

The corrosion property, i.e. E _corr and I _corr, has been studied by potentiodynamic polarization test and the wear is measured in terms of wear depth by DUCOM TR-25 multi-tribotester with block on roller arrangement.

Results

In this study, the process is intrinsically combined with multiple performance indexes so that grey relational analysis is specially adopted to determine the optimal combination of coating parameters. Moreover, the weighted principal component analysis (WPCA) is applied to evaluate the weighting values corresponding to various performance characteristics so that their relative importance can be properly and objectively described.

Conclusion

From the analysis the optimum combination of parameters for corrosion property and the optimum combination of parameters for corrosion and wear together are obtained. The chemical composition, surface morphology and phase behaviour are investigated using energy dispersive X-ray analysis, scanning electron microscopy and X-ray diffraction analysis, respectively.

Background

Coating is a method by which an artificial surface can be generated to the outer surface of the substrate material to protect it from corrosion and wear. These are the two deteriorating phenomena which are the source of major loss for industrial machinery. These not only reduce the life of the industrial components but also increase the maintenance cost and expenditure for replacement of parts. Since corrosion and wear both occur at the surface of the substrate, they can be reduced or eliminated by surface treatment. In this respect, the metallic surface coating gives a practical solution. Electroless coating, also known as chemical or auto-catalytic coating, is a non-galvanic plating method that involves several simultaneous chemical reactions in an aqueous solution, which occur without the use of external electrical power. That makes the difference of this process with that of conventional electroplating process which requires external current source. Electroless coating process has gained wide acceptance in the market due to the excellent corrosion and wear resistance properties, and it is also good for soldering and brazing purposes (Sahoo and Das 2011). In recent days the binary electroless Ni-P coatings have become the research focus due to their more superior properties. These properties can be further improved by incorporating a third particle into that binary alloy. The choice of the third particle depends on the desired property. Ternary Ni-P coatings, such as Ni-Cu-P (Yu et al. 2002; Aal and Aly 2009), Ni-W-P (Palaniappa and Seshadri 2008; Balaraju et al. 2006a; Balaraju et al. 2006b; Roy and Sahoo 2013; Roy and Sahoo 2012), Ni-P-TiO₂ (Abdel Aal et al. 2008; Chen et al. 2010; Novakovic and Vassiliou 2009), Ni-P-Al₂O₃ (Alirezaei et al. 2007; Balaraju et al. 2006c), Ni-P-PTFE (Ramalho and Miranda 2005; Huang et al. 2003) and Ni-P-SiC (Lin et al. 2006; Jiaqiang et al. 2006), have been prepared by electroless deposition. Among these ternary Ni-P alloy coatings, the electroless Ni-Cu-P alloy presents more superior corrosion resistance and thermal conductivity than the others (Liu et al. 2010; Valova et al. 2010; Liu and Zhao 2004; Wang et al. 1992). The inclusion of Cu in electroless Ni-P alloys improves their smoothness (Balaraju and Rajam 2005), brightness (Tarozaitë and Selskis 2006; Chen et al. 2006) and corrosion resistance (Liu and Zhao 2004; Zhao et al. 2004; Armyanov and Georgieva 2007). The crystallization behaviours of Ni-P-Cu coatings on aluminium substrates were investigated by Chen and Lin (1999). A comparative study on the crystallization behaviour of electroless Ni-P and Ni-Cu-P deposits was performed by Hui-Sheng et al. (2001) and found that the crystalline temperature for the formation of Ni₃P phase is higher for Ni-Cu-P coating than Ni-P coating. It was mentioned that the addition of copper into electroless Ni-P matrix could improve the corrosion resistance of the coatings (Mallory and Hadju 1991). The corrosion study of electroless Ni-P-Cu reveals that 90% Ni-7% Cu-3% P in 50% NaOH solution was better than that of as-plated Ni-P (Wang et al. 1992). The anticorrosion properties of the Ni-Cu-P coatings in 1 M HCl, 1 M H₂SO₄ and 3% NaCl solutions were investigated by Cissé et al. (2010) using Tafel polarization curves and electrochemical impedance spectroscopy. The result showed a marginal improvement in corrosion resistance in 3% NaCl solution compared to acidic medium. As the corrosion and wear property of this coating depends on the coating parameters, the parameters can be optimized for best corrosive media, i.e. NaCl solution. Practically, both the corrosion and wear take place simultaneously; hence, the parameters can be optimized taking the effect of corrosion and wear together. The Taguchi method is a statistical approach for the purpose of designing and improving product quality. Tosun (2006) used the grey relational analysis for the determination of optimal drilling parameters with the objective of minimization of surface roughness and burr height. Deng (1982) proposed the grey system theory which has been proven to be useful for dealing with the problems with poor, insufficient and uncertain information. The grey-based Taguchi method was employed to optimize the process parameters of the submerged arc welding (SAW) in hardfacing, considering multiple weld qualities (Tarng et al. 2002). Grey relational analysis was adopted to investigate the electro discharge machining (EDM) parameters on machining Al-10% SiCp composites by Narender Singh et al. (2004). Several researchers have used grey relational method to optimize the design process parameters, but most of the researchers have selected the weighting values of the response parameters according to their own estimation during calculation of the grey relational grade. This method cannot emphasize the relative importance of the response parameters related with the experiment. The case study by Antony (2000) demonstrates the potential of multi-response optimization in industrial experiments using Taguchi's quality loss function and principal component analysis. The research of Lua et al. (2009) about the optimization problem with multiple performance characteristics using grey relational analysis presents a remedy by calculating the corresponding weighting values using principal component analysis (Hotelling 1993). The researchers have used grey relational analysis for optimizing combination of cutting parameters and principal component analysis for determining the corresponding weighting values of various performance characteristics. In this present investigation, the optimum combination of parameters for corrosion and wear, the grey relational coefficient is used and the corresponding weighing value of each performance characteristics calculated by weighted principal component analysis considering the grey relational coefficients and the effect of all responses are clubbed together into multiple performance index. The surface morphology, chemical composition and phase transformation behaviour were studied by scanning electron microscopy (SEM), energy dispersive X-ray spectroscopy (EDX) and X-ray diffraction (XRD) analyses, respectively.

Methods

Selection of parameters

The electroless coatings involve large number of process parameters which can affect the performance characteristics of the coatings. In this present study, after a large number of literature review and experimental trials, four main process parameters have been selected as input parameters. Among the four parameters the first three are coating parameters, viz. concentration of nickel sulphate (source of nickel), concentration of sodium hypophosphite (reducing agent) and concentration of copper sulphate (source of copper), and the fourth one is the post-deposition heat treatment temperature. The operating range of the parameters has been selected on the experimental basis, within which the coating can be deposited. The range of each parameter has been divided in to three equally spaced levels. The main parameters with their values are shown in Table 1. The responses are corrosion potential, corrosion current density and wear depth.

Table 1 Main coating parameters with their levels

Experimental design

This experimental investigation consists of four three-level input parameters; hence, with all possible combinations, a total number of (3)⁴ = 81 experiments can be carried out. To save time and cost, the number of experiments has been reduced by using Taguchi's specially developed orthogonal array (OA). The selection of OA depends on the number of individual parameters and their interaction considered for the analysis. In this study along with four individual parameters, the interactions between three coating parameters, i.e. interaction between nickel sulphate and sodium hypophosphite, sodium hypophosphite and copper sulphate, and nickel sulphate and copper sulphate, have been considered. As this is a three-level experiment, the total degrees of freedom associated with this experiment is 20. Hence a standard L₂₇ OA has been selected as this has 26 degrees of freedom which is higher than the degrees of freedom of experiment. A standard L₂₇ OA is shown in Table 2, which consists of 27 rows and 13 columns. Each row represents the combination of parameters for deposition of coating, and each column indicates the individual factors and their interactions.

Table 2 L ₂₇ orthogonal array

Results and discussion

Coating deposition

Mild steel blocks (AISI 1040) of size 20 mm × 20 mm × 8 mm are used as substrates for the deposition of electroless Ni-P-Cu coating. This particular dimension of the sample is chosen to fit the counter part of block on roller multi-tribotester apparatus. The sample is mechanically cleaned from foreign matters and corrosion products. After that, the MS sample is cleaned using distilled water. Then, a pickling treatment is given to the specimen with dilute (50%) hydrochloric acid for 1 min to remove any surface layer formed like rust followed by rinsing in distilled water and methanol cleaning. Table 3 indicates the bath composition and the operating conditions for successful coating of electroless Ni-P-Cu. Nickel sulphate is used as the source of nickel while sodium hypophosphite is the reducing agent and sodium citrate was added as complexing agent. The bath is prepared by adding the constituents in appropriate sequence. The pH of the solution is maintained around 9.5 by continuous monitoring with a pH meter. The cleaned samples are activated in palladium chloride solution at a temperature of 55°C. Activated samples are then submerged into the electroless bath which is maintained at a temperature of 85°C with the help of a hot plate cum stirrer attached with a temperature sensor also submerged in the solution. The deposition is carried out for 2 h. The range of coating thickness is found to lie around 28 to 30 μm by measuring with a digital micrometer instrument. After deposition, the samples are taken out of the bath and heat-treated according to the experimental design. Figure 1 shows the schematic diagram of coating deposition set-up.

Table 3 Electroless bath constituents

Figure 1

Electroless deposition set-up.

Wear measurement

The wear depths of heat-treated Ni-P-Cu-coated specimens are measured under non-lubricated condition using a multi-tribotester with block on roller configuration (DUCOM TR-25, Bangalore, Karnataka, India). The Ni-P-Cu-coated specimens serve as test specimens of average hardness of 42 HRc, which are held horizontally against a rotating roller coated with titanium nitride of hardness 85 HRc of 50-mm diameter × 20-mm thickness, as shown in Figure 2. As the hardness of the roller is higher than the hardness of coating, it may be assured that the wear will take place on the coating only. The wear test of each specimen is carried out for 5 min with 25 N load at a speed of 50 rpm. Dead weights are placed on the loading platform which is attached at one end of a 1:5 ratio loading lever. A linear voltage resistance transducer is used for measuring wear in terms of wear depth. It is worth noting that, in general, wear is measured in terms of wear volume or mass loss. However, in the present case, wear is expressed in terms of displacement or wear depth. Hence, to ensure that the wear measurements are accurate, the wear depth results are compared with the weight loss of the specimens and almost linear relationship is observed between the two for the range of test parameters considered in the present study.

Figure 2

Block on roller arrangement for wear test.

Polarization study

The potentiodynamic polarization tests of heat-treated Ni-P-Cu coatings are carried out using a potentiostat (Gill AC) of ACM Instruments, UK, shown in Figure 3. The corrosion parameters were measured by potentiodynamic polarization curve measurements. The tests are conducted at an ambient temperature of about 25°C with 3.5% sodium chloride solution as the electrolyte. The electrochemical cell consists of three electrodes. The coated specimen forms the working electrode which is actually the sample being interrogated. A saturated calomel electrode (SCE) forms the reference electrode which provides a stable ‘reference’ against which the applied potential may be accurately measured. A platinum electrode serves as the counter electrode which provides the path for the applied current into the solution. The design of the cell kit is such that only an area of 1 cm² of the coated surface is exposed to the electrolyte. The experimental set-up is shown in Figure 1. A settling time of 15 min is assigned before every experiment in order to stabilize the open circuit potential (OCP). The potentiostat is controlled via a PC which also captures the polarization data. Potentiodynamic polarization studies were carried out by polarizing the working electrode from the OCP to 250 mV in cathodic direction and 250 mV in anodic direction at a scan rate of 1 mV/s. The corrosion current densities (I _corr) were determined by extrapolating the straight-line section of the anodic and cathodic branches of the Tafel plots in the vicinity of the corrosion potential using the software installed in the instrument The polarization plot is obtained from the dedicated software, which also possesses a special tool in order to manually extrapolate the values of E _corr (corrosion potential) and I _corr (corrosion current density) from the plot. Each experiment has been repeated for three times, and the variation of result was within 2%. The average value has been taken for analysis. The results of wear and corrosion are shown in Table 4. The Tafel plots and the variation of wear are shown in Figure 4

Figure 3

Potentiodynamic polarization test arrangement.

Table 4 Results of corrosion and wear test

Figure 4

Tafel plots and variation of wear depths. For different compositions of Ni at different heat treatment temperatures (a) 300°C, (b) 400°C and (c) 500°C.

Characterization of coating

The characterization of the coating is necessary so that it can be made sure that the coating is properly developed. Energy dispersive X-ray analysis (EDAX Corporation, Mahwah, NJ, USA) is performed to determine the composition of the coating in terms of the weight percentages of nickel, phosphorous and copper. Figure 5 shows the EDX spectra of the coated surface. From the analysis, it is found that the coating consists of 11% P, 4% Cu and the remaining is Ni. Figure 6 shows the SEM of as-deposited and heat-treated (300°C, 400°C, 500°C) Ni-P-Cu-coated surface. A deposit coarse nodular structure without any porosity in as-deposited condition is clear. Nodular deposition in a coating depends on nucleation rate and the growth of the deposit. Nucleation rate depends on the bath constituents and the operating condition of the experiment. From the figures, it is clear that due to heating, crack appears in the coating. Figure 7 shows the image of the worn surface and the corroded surface. The phase transformation behaviour has been studied by XRD. Figure 8 shows the XRD pattern of as-deposited and heat-treated condition. From the figure, it is clear that in as-deposited condition the coating is mostly amorphous, but crystalline peaks appear after heating. The major crystalline peaks of Ni, Cu₃P, Ni₃P and Ni₃P₂ appear after heating at 400°C for 1 h.

Figure 5

EDX spectra of Ni-P-Cu-coated surface.

Figure 6

SEM images of Ni-P-Cu-coated surface. (a) As deposited, (b) heat treated at 300°C, (c) heat treated at 400°C and (d) heat treated at 500°C.

Figure 7

SEM images of the (a) worn surface and (b) corroded surface.

Figure 8

XRD pattern of Ni-P-Cu coating.

Analysis methodology and discussion

Grey relational coefficient

In this study among the three responses, a higher value of corrosion potential (E _corr) and a lower value of corrosion current density (I _corr) are desired for good corrosion resistance and obviously a lower value of wear depth has been targeted. As there is a huge difference between the average value of each response, the result obtained from the analysis considering these values may not give the correct result when the effect of all the parameters are considered together. To eliminate this effect, the result data of each response have been normalized or scaled between 0 and 1. The value 1 represents a good result and 0 represents a worse result. Here, E _corr is normalized considering the bigger the better as a higher corrosion potential indicates good corrosion resistance. The I _corr and wear depth both are normalized considering the smaller the better. Using this normalized value, the grey relational coefficients are calculated, which are explained stepwise:

Step 1: normalization

Normalization of E _corr is performed using Equation 1:

$$ \mathrm{Normalized}\ \mathrm{value}\ \mathrm{of}{E}_{\mathrm{corr}}\left({E}_j^{*}\right)=\frac{E_j-{E}_{\min }}{E_{\max }-{E}_{\min }} $$

(1)

Normalization of I _corr and wear depth is performed using Equations 2 and 3:

$$ \mathrm{Normalized}\ \mathrm{value}\ \mathrm{of}{I}_{\mathrm{corr}}\left({I}_j^{*}\right)=\frac{I_{\max }-{I}_j}{I_{\max }-{I}_{\min }} $$

(2)

$$ \mathrm{Normalized}\ \mathrm{value}\ \mathrm{of} W\left({W}_j^{*}\right)=\frac{W_{\max }-{W}_j}{W_{\max }-{W}_{\min }}, $$

(3)

where E _j  = E _corr value corresponding to the jth experiment

I _j  = I _corr value corresponding to the jth experiment

W _j  = wear value corresponding to the jth experiment

j = sequence of experimental run (j = 1, 2, 3…); as there is a total of 27 experimental runs, the maximum value of j is 27.

Step 2: grey relational generation

The grey relational coefficient (g _j ) for each response has been generated using Equation 4:

$$ {g}_j=\frac{\varDelta {R}_{\min}^{*}+ r\varDelta {R}_{\max}^{*}}{\varDelta {R}_j^{*}+ r\varDelta {R}_{\max}^{*}} $$

(4)

where${R}_j^{*}$ = the normalized response value (${E}_j^{*}$ for corrosion potential, ${I}_j^{*}$ for corrosion current and ${W}_j^{*}$ for wear depth)

$$ \varDelta {R}_j^{*}={R}_{j \max}^{*}-{R}_j^{*}, $$

${R}_{j \max}^{*}=$ the maximum value of ${R}_j^{*}$

$\varDelta {R}_{\max}^{*}$ and $\varDelta {R}_{\min}^{*}$ are the maximum and minimum values of $\varDelta {R}_j^{*}$, respectively.

r is a distinguishing coefficient, which belongs to [0, 1]. The distinguishing coefficient weakens the effect of max ΔR _max when it gets too big, enlarging the different significance of the relational coefficient. The suggested value of the distinguishing coefficient, r, is 0.5, due to the moderate distinguishing effects and good stability of outcomes. Therefore, r is adopted as 0.5 for further analysis in the present case.

The normalized values and grey relational coefficients of each response are shown in Table 5. The conventional method for finding the grey relational grade is to take the average of these grey relational coefficients, i.e. considering equal contribution of each response to the overall variation. However, the eigenvalue of a principal component gives a fairly good idea about the variance of the original variables that can be explained by the principal component. A larger eigenvalue of a principal component implies that the component's contribution in explaining the overall variation is higher. In this study, the corresponding weighting values are obtained from the principal component analysis.

Table 5 Results of grey analysis

Weighted principal component analysis

According to Antony (2000), the components with eigenvalues greater than 1 may be selected to replace the original responses. However, problems can arise in the situations where more than one eigenvalue becomes greater than 1. The weighted principal component (WPC)-based procedure (Su and Tong 1997; Liao 2006) for optimization of multi-response processes makes use of all the principal components irrespective of the eigenvalues so that the overall variation in all the responses can be completely explained. In this approach, the proportion of overall variation explained by each component is treated as the weight to combine all the principal components in order to form a multi-response performance index (MPI). Then, the best combination of the parametric settings can easily be obtained which can optimize the MPI. The procedure for calculating MPI is described stepwise:

Step 1: eigenvalue and eigenvectors and proportion of overall variance

The eigenvalue (λ) and eigenvectors (V) are calculated from Equation 5 imposing a condition ${\displaystyle \sum_{k=1}^Q{V}_k^2}=1$

$$ \left[ G-\lambda I\right]\times \left[ V\right]=0 $$

(5)

where

$$ G=\left[\begin{array}{cccc}\hfill \operatorname{var}(1)\hfill & \hfill \operatorname{cov}\left(1,2\right)\hfill & \hfill \dots \hfill & \hfill \operatorname{cov}\left(1, k\right)\hfill \\ {}\hfill \operatorname{cov}\left(2,1\right)\hfill & \hfill \operatorname{var}(2)\hfill & \hfill \dots \hfill & \hfill \operatorname{cov}\left(2, k\right)\hfill \\ {}\hfill \dots \hfill & \hfill \dots \hfill & \hfill \dots \hfill & \hfill \dots \hfill \\ {}\hfill \operatorname{cov}\left( j,1\right)\hfill & \hfill \operatorname{cov}\left( R,2\right)\hfill & \hfill \dots \hfill & \hfill \operatorname{var}\left( R, k\right)\hfill \end{array}\right] $$

is the covariance matrix of grey relational coefficients.

k is the number of quality characteristics; in this problem, the maximum value of k is 3.

The proportion of overall variance or weight is calculated using Equation 6:

$$ {W}_k=\frac{\lambda_k}{{\displaystyle \sum_{k=1}^Q{\lambda}_k}} $$

(6)

The eigenvalues, eigenvectors and proportion of overall variance considering only corrosion parameters (E _corr and I _corr) are shown in Table 6, and the corresponding values considering both corrosion and wear (E _corr, I _corr and W) are shown in Table 7.

Table 6 Results obtained considering E _corr and I _corr

Table 7 Results obtained considering E _corr , I _corr and wear

Step 2: calculation of principal components and MPI

The principal components are calculated using Equation 7:

$$ \left[ P\right]=\left[ g\right]\times \left[ V\right] $$

(7)

The MPI is calculated using Equation 8:

$$ \mathrm{MPI}={\displaystyle \sum_{k=1}^Q{P}_{j, k}\times {W}_{k,1}} $$

(8)

The principal components and MPI considering only corrosion parameters (E _corr and I _corr) are shown in Table 8, and the corresponding values considering both corrosion and wear (E _corr, I _corr and W) are shown in Table 9.

Table 8 Results obtained considering E _corr and I _corr

Table 9 Results obtained considering E _corr , I _corr and wear

Optimum combination of parameters

As the design of experiment is orthogonal, the effect of each parameter on MPI can be separated out by taking the average of same levels of each input parameter. For example, among the 27 experiments, there are 9 experiments, which include the level 1 of parameter A. Taking the average of these 9 MPI values, the mean MPI of level 1 for parameter A can be calculated. Similar procedure is applicable for other parameters. Table 10 shows the mean response table of the MPI taking only corrosion parameters (corrosion potential and corrosion current density), and Table 11 shows the same considering all the three parameters (corrosion potential, corrosion current and wear depth). Figures 9 and 10 show the main effect plots obtained from the response tables, respectively. From the plots, the optimum combination of input parameters can be obtained. As the larger value of MPI corresponds better multiple response characteristics, the optimum combination can be obtained by selecting the largest level average of each parameter. Figure 9 yields the optimum combination considering only corrosion parameters is A1B2C2D2, and Figure 10 yields the optimum combination considering corrosion parameters and wear together is A1B3C1D2.

Table 10 Response table considering E _corr and I _corr

Table 11 Response table considering E _corr , I _corr and wear

Figure 9

Main effect plot considering E _corr and I _corr .

Figure 10

Main effect plot considering E _corr , I _corr and wear.

Significance of parameters on MPI

The response table also reveals the significance of each individual factor. In the response tables, the maximum deviation of each parameter is listed in the right column. It is obtained by subtracting the lowest mean MPI from the largest mean MPI value among the three levels of any parameter. The parameter has huge impact on the multiple responses, which has maximum deviation. From the tables, it is clear that parameter A, i.e. concentration of nickel sulphate, and parameter C, i.e. concentration of copper sulphate, have positive impact on the corrosion and wear property. The effect of nickel is highly dominant for both the cases, but the effect of copper is higher when only corrosion parameters are considered. It has been seen that due to heat treatment the structure of the coating transforms into crystalline. The coating becomes hard mainly due to the formation of the nickel phosphide structure at 400°C, and thus, improved wear resistance is achieved at this stage along with the corrosion. According to Hui-Sheng et al. (2001), after heating 500°C for 1 h, the metastable phase Ni₅P₂ transforms completely to stable Ni₃P phase. It leads to harder and wear resistant coating due to crystallization which leads to more corrosive prone surface. Thus, 500°C may not be the optimum heat treatment temperature. Hence, this present analysis has a good agreement with this result. The results of analysis of variance (ANOVA) considering the corrosion parameters (E _corr and I _corr) and also considering the corrosion parameters and wear together are shown in Tables 12 and 13, respectively. The tables reveal that the percentage contribution of nickel is highest for both the conditions. Along with this the ANOVA results also focus on the significance of the interaction of parameters on the responses. It is clear from both the ANOVA tables that the percentage contribution of the interaction between nickel and copper is highest among the three interactions.

Table 12 ANOVA table considering E _corr and I _corr

Table 13 ANOVA table considering E _corr , I _corr and wear

Confirmation test

To validate the result obtained from the analysis, a confirmation test was carried out with the optimum combination of parameters. Coatings are developed with the optimum combination of parameters obtained from optimization analysis, viz. A1B2C2D2 for corrosion optimization and A1B3C1D2 for combined corrosion and wear optimization. These coatings are then subjected to corrosion and wear tests. The results of these tests are compared with the tests on coatings developed with mid-level combination of parameters, i.e. A2B2C2D2. It is because with this combination the bath is most stable for a long time, and maximum thickness of coating can be achieved. However, the aim is to find out the best quality coating against corrosion and wear. Hence, a comparison between the mid-level result and the optimum level results has been carried out. The result of the confirmation test is tabulated in Table 14. From the table, it is clear that at optimum condition for corrosion, the value of corrosion potential is improved by 49%, while the value of corrosion current decreases by 84%. For combined corrosion and wear optimization case, the value of corrosion potential is improved by 7%, while the value of corrosion current decreases by 76% and wear depth decreases by 40%. Thus, the optimum combination of parameters yields a better coating. The polarization curves for both the optimum conditions and mid-level combination are shown in Figure 11. The improvement of corrosion resistance of the coatings obtained from the optimum combination of parameters is clearly seen in these plots since corrosion potential increases and corrosion current decreases from the mid-level combination.

Table 14 Results of confirmation test

Figure 11

Tafel plots. Mid-level combination (1), optimum combination considering corrosion and wear together (2) and optimum combination considering only corrosion (3).

Conclusions

The electroless ternary Ni-P-Cu coating has been developed on mild steel substrate by varying four input design parameters, namely concentration of nickel source (nickel sulphate), concentration of reducing agent (sodium hypophosphite), concentration of copper source (copper sulphate) and post-deposition heat treatment temperature. The design of experiment was done by Taguchi L₂₇ OA with 27 experimental runs. The wear depth of the heat-treated coatings was measured with a multi-tribotester instrument with block on roller configuration. The polarization (corrosion) tests were carried out using a potentiostat instrument. By extrapolating the Tafel plot, the corrosion current density and the corrosion potential were measured. Then, the grey analysis together with weighted principal component analysis is successfully employed for finding out the optimal combinations of the design process parameters of electroless Ni-P-Cu coatings for better value of polarization test and also considering the polarization and wear test together. Confirmation tests were carried out for both the cases to validate the experimental value. The energy dispersive X-ray analysis shows that it is a pure ternary coating consisting of nickel phosphorous and copper; the surface morphology and phase transformation behaviour have been studied by SEM and XRD analyses, respectively.