Electrodeposition of Ni matrix composite coatings with embedded CrAlY particles

Introduction

To accommodate higher firing temperatures for improved efficiency of industrial gas turbines, thermal barrier coatings have been adopted to provide effective protection for hot section components.^1,2 MCrAlY (where M = Ni, Co or a mixture of Ni and Co) overlays are one type of bond coats used extensively in the state-of-the-art thermal barrier coating systems.^3,4 To apply MCrAlY coatings on industrial gas turbine components, relatively low cost processes such as air plasma spray and high velocity oxyfuel are generally utilised due to the size of these components. ⁵ Alternative methods of making MCrAlY coatings have also been reported, ⁶ including electrolytic codeposition,^7,8 electrophoresis ⁹ and autocatalytic electroless deposition,^10,11 among which the electrolytic codeposition process appears to be a promising cost effective approach. Electrolytic codeposition (also called ‘composite electroplating’) ¹² is a process in which fine powders dispersed in a plating solution are deposited simultaneously with the metal onto the substrate to form a multiphase composite coating. This technique also offers advantages such as non-line-of-sight and the capability of producing dense coatings. ⁶ Although the electrocodeposition process has been applied to deposit a broad range of coatings for various applications, ¹³ very limited work can be found on fabrication of MCrAlY coatings. Foster et al. ⁷ and Honey et al. ⁸ explored a two-step process for making MCrAlYs, involving deposition of a composite coating consisting of a metal matrix (e.g. Ni or Co) and CrAlY based particles, followed by a diffusion heat treatment to convert the composite to the desired MCrAlY microstructure. Using a similar electrochemical process, Yang et al. ¹⁴ synthesised Ni based composite coatings with embedded Cr and Al nanoparticles.

As compared to conventional electroplating, electrolytic codeposition is a more complex coating process due to the involvement of the particles in the metal deposit. The quality of the composite coatings depends on a number of interrelated processing parameters, including the type of electrolyte, current density, pH, particle loading, agitation and post-deposition heat treatment (if necessary).^12,13 Unlike Ni based composite coatings containing hard oxide or carbide particles, ¹³ no detailed study can be found regarding the synergistic effects of the codeposition parameters for fabricating Ni/Co–CrAlY coatings. In this study, a design of experiments (DoEs) ¹⁵ approach was employed to investigate the effects of the key parameters in the electrocodeposition process on the amount of CrAlY particles incorporated in the Ni–CrAlY composite coatings. In contrast to one-factor-at-a-time experiments, statistical methods and experimental design are effective in reducing the number of experiments but still with the capability of elucidating the significance of process parameters and their interactions. An understanding of the codeposition parameters is critical for achieving the optimal composition/microstructure of the final MCrAlY coatings after heat treatment. Different cathode–anode configurations have been utilised in the electrocodeposition process. When the cathode (specimen) is placed vertically in the plating bath, as in a typical electroplating process, the resultant composite coatings often have low particle incorporation and a maximum of 30 vol.-% has been reported. ⁷ To increase the particle incorporation, sediment codeposition (SCD) is adopted in the present work, where the specimen is positioned in a horizontal plane beneath the anode to utilise the gravity effect.^16–19

Experimental

Ni 200 alloy (>99·0Ni, <0·25Cu, <0·40Fe, <0·35Mn, <0·15C, <0·35Si and <0·01S, wt-%) was used as the substrate material. Disc specimens (∼19 mm in diameter and ∼1·6 mm thick) were prepared by grinding to a 600 grit finish, grit blasting with 220-mesh Al₂O₃ grit, followed by ultrasonic cleaning in hot water and acetone. Electrocodeposition experiments were carried out in a Watts Ni plating solution ²⁰ containing 310 g L⁻¹ nickel sulphate, 50 g L⁻¹ nickel chloride, 40 g L⁻¹ boric acid and 0·2 g L⁻¹ sodium docecyl sulphate. Since prealloyed CrAlY based powders were not commercially available, the powder used in this study (Cr–37Al–1·7Y, wt-%) was fabricated via arc melting and ball milling in our laboratory, followed by sieving through 625 mesh (20 μm) screen. The mean particle diameter was 7 μm, as determined using the Malvern Mastersizer 2000 LASER diffractor at Particle Technology Labs (PTL, Downers Grove, IL, USA). The density of the CrAlY powder was 4·5 g cm⁻³, as measured by a pycnometer (AccuPyc II 1340 Gas Pycnometer). After the plating solution was mixed with deionised water in a 1000 mL glass beaker, the CrAlY powder was added slowly while the solution was stirred. Figure 1 shows the SCD set-up. Pure Ni plate (28×28×6 mm) was used as the anode, and the anode to cathode distance was maintained at 20 mm. Agitation was applied during the SCD process using a Teflon coated magnetic stirrer to keep the particles in suspension. All specimens were coated for 2 h at 50°C, and they were rinsed in water and dried after plating. The pH of the plating solution was maintained at 3·0–3·5, which was adjusted by adding sulphuric acid or sodium hydroxide when necessary.

OPEN IN VIEWER

1

Schematic of sediment codeposition set-up

The effects of the three key process parameters on the amount of incorporated CrAlY particles were studied using a 2³ full factorial design augmented with three centre point replicates, as shown in Table 1. The three parameters (i.e. the factors of control) were current density (A), agitation rate (B) and CrAlY particle concentration in the plating bath (i.e. the particle loading) (C). Two levels were assigned to each factor, and they were coded as −1 (low level) and +1 (high level) respectively. The addition of the centre point (coded as level 0 in Table 1) is an effective way to detect whether non-linearity or curvature exists for the response variable as a function of the factor. ¹⁵ The design matrix is given in Table 2, and a total of 11 experiments were performed. Only the upward facing surface of the specimen was evaluated due to the nature of the SCD set-up. The software package of Minitab 16 (Minitab Inc., State College, PA, USA) was used to analyse the experimental data and determine the main effects and interactions between the factors.

OPEN IN VIEWER

Table 1

Factors and levels in 2³ full factorial design with addition of centre point

	Actual values of coded levels
Factor	−1	0	+1
A: Current density/mA cm−2	20	40	60
B: Agitation rate/rev min−1	80	300	520
C: Concentration of CrAlY particles in the plating solution/g L−1	10	30	50

OPEN IN VIEWER

Table 2

Experimental matrix and results of 2³ full factorial design with addition of three centre point replicates

	Level of control factor
Experiment	A (current density)	B (agitation)	C (particle loading)	Incorporated particles/vol.-%
1	−1	−1	−1	28·5
2	+1	−1	−1	19·8
3	−1	+1	−1	41·7
4	+1	+1	−1	45·1
5	−1	−1	+1	36·7
6	+1	−1	+1	39·8
7	−1	+1	+1	34·7
8	+1	+1	+1	48·6
9	0	0	0	46·9
10	0	0	0	45·3
11	0	0	0	42·8

The Ni–CrAlY composite coatings were characterised using optical microscopy and scanning electron microscopy (SEM) equipped with energy dispersive spectroscopy. Before metallographic sample preparation, the specimens were Cu plated to improve the edge retention. To determine the volume fraction of the incorporated CrAlY particles, multiple backscattered electron images were taken from different locations along the coating cross-section, which were then processed using the ImageJ software. ²¹ The brightness and contrast of the images were adjusted by setting a proper threshold such that the particles were separated from the background. The area fraction of the CrAlY particles was determined, which was assumed equivalent to its volume fraction.

Results and discussion

Figure 2 displays the surface morphologies at low (Fig. 2a1–d1) and high (Fig. 2a2–d2) magnifications for the coatings deposited in the solution containing 10 g L⁻¹ of CrAlY particles. The agitation rate showed a strong effect on the coating surface roughness. Comparison of Fig. 2a and c (or Fig. 2b and d) indicates that a rougher surface was formed at the higher stirring speed, regardless of the applied current density level. A similar trend was also observed for the coatings deposited in the solution with 50 g L⁻¹ particle loading.

OPEN IN VIEWER

2

Low and high magnification SEM surface images of coatings deposited with 10 g L⁻¹ particle loading; current density and agitation rate were a1, a2 20 mA cm⁻², 80 rev min⁻¹; b1, b2 60 mA cm⁻², 80 rev min⁻¹; c1, c2 20 mA cm⁻², 520 rev min⁻¹; and d1, d2 60 mA cm⁻², 520 rev min⁻¹

Figure 3 shows the cross-sections of the composite coatings deposited with 10 g L⁻¹ particle loading under various combinations of current density and agitation speed. The amount of the CrAlY particles incorporated in the composite coating ranged from 20 to 45 vol.-% (Table 2). A higher stirring speed (520 rev min⁻¹) led to higher particle incorporation, up to 45 vol.-%. Figure 4 presents the cross-sections of the composite coatings deposited with 50 g L⁻¹ particle loading. As the CrAlY particle concentration in the solution was increased from 10 to 50 g L⁻¹, the amount of particles incorporated in the coating also increased, from 20–45 vol.-% to 35–49 vol.-%. It should be noted that commercial MCrAlY coatings typically contain 8–12%Al, 18–22%Cr and up to 0·5%Y (in wt-%). ²² To form an NiCrAlY coating with an Al level of ∼10 wt-%, based on the chemical composition and the density of the present CrAlY powder, ∼40 vol.-%CrAlY particles are required in the as deposited coatings. Clearly, by choosing the proper combination of deposition parameters, the required CrAlY particle incorporation could be realised. However, it is worth mentioning that voids were formed in some coatings (as pointed out by the arrows in Fig. 4b and c), leading to less dense coatings. Although the growth rate of the Ni matrix was mainly controlled by the magnitude of the current density employed in the electrodeposition process, the presence of the CrAlY particles and porosity could affect the overall thickness and uniformity of the composite coatings. As a result, the coatings containing voids appeared thicker and/or rougher than those dense coatings deposited at the same current density (e.g. Fig. 4a versus c). It should also be noted that the sharp angles formed between the ball milled CrAlY particles and the Ni matrix did not cause any cracking in the as deposited coatings. It is believed that the relatively ductile Ni matrix can relax the stress present at the sharp corners.

OPEN IN VIEWER

3

Cross-sectional images (SEM) of coatings deposited with 10 g L⁻¹ particle loading: current density and agitation rate were a 20 mA cm⁻², 80 rev min⁻¹; b 60 mA cm⁻², 80 rev min⁻¹; c 20 mA cm⁻², 520 rev min⁻¹; and d 60 mA cm⁻², 520 rev min⁻¹

OPEN IN VIEWER

4

Cross-sectional images (SEM) of coatings deposited 50 g L⁻¹ particle loading: current density and agitation rate were a 20 mA cm⁻², 80 rev min⁻¹; b 60 mA cm⁻², 80 rev min⁻¹; c 20 mA cm⁻², 520 rev min⁻¹; and d 60 mA cm⁻², 520 rev min⁻¹

In a DoE study, the ‘effect’ of a factor refers to the change in the response by a change in the level of the factor. The total number of effects in a 2³ factorial design is seven, including three main effects (A, B and C) and four interaction effects (AB, AC, BC and ABC). ¹⁵ A Pareto chart ²³ is usually used to present the relative importance of the factors and their interactions, in the order from the most to the least significant, as shown in Fig. 5. The Student's t test was performed to determine whether the effect was significant. With 95% confidence, the t value was 3·18; hence, factors and interactions with an effect size >3·18 should be considered as significant. Figure 5 suggests that the agitation rate (B) was the factor with the most significance, followed by the interaction between agitation and particle loading (BC) and then the particle loading (C). While the effect of the current density (A) itself was insignificant, interactions AB and AC were statistically significant in affecting the CrAlY particle incorporation in the deposited coating.

OPEN IN VIEWER

5

Pareto chart showing effect of factors (A, B and C) and their interactions (AB, AC and BC) on CrAlY particle incorporation

The results of analysis of variance (ANOVA) are presented in Table 3. Analysis of variance is a statistical method that partitions the total variation of a set of data into components associated with specific sources of variation for the purpose of testing a hypothesis.^15,24,25 The F ratio is generally used to determine the significance of the factor effects, which is defined as the ratio of ‘the treatment mean square’ and ‘the error mean square’. The p value is considered as the smallest level at which the data are significant, and a small value of 0·05 or 0·1 is typically used. For the current case, at the 95% confidence level, if F exceeds 10·13 or P is lower than 0·05, the factor is deemed significant. Similar to the conclusion obtained from the Pareto chart (Fig. 5), the ANOVA corroborated that the agitation rate (B) exhibited the greatest significance (F = 87·93) in affecting the particle incorporation in the Ni–CrAlY coatings, followed by the interaction between agitation and particle loading (BC). The particle loading factor (C), as well as AB and AC interactions, were statistically significant, but to a less degree. The small F value (5·87) for the current density (A) reveals that it was not a significant factor in influencing the amount of the incorporated CrAlY particles. In addition, the F value of the curvature was high, suggesting that the factors and the response variables did not follow a linear relationship. Similar non-linearity has also been reported for electrodeposited Cu coatings containing Nb particles. ²⁶

OPEN IN VIEWER

Table 3

Analysis of variance (ANOVA) for CrAlY particle incorporation in composite coatings

Source	Sum of squares	Degrees of freedom	Mean square	F ratio	p value
Current density (A)	17·11	1	17·11	5·87	0·094
Agitation rate (B)	256·51	1	256·51	87·93	0·003
Particle loading (C)	76·26	1	76·26	26·14	0·014
A×B	65·55	1	65·55	22·47	0·018
A×C	62·16	1	62·16	21·31	0·019
B×C	125·61	1	125·61	43·06	0·007
Curvature	144·48	1	144·48	49·53	0·006
Error	8·751	3	2·917
Total	756·44	10

The mechanism for codeposition of particles into a metal deposit has been studied extensively and documented in the literature.^12,13,27,28 It is generally believed that five consecutive steps are engaged during the electrolytic codeposition process: ¹³

During electrocodeposition, bath agitation serves two purposes,^12,13 namely, to keep the particles suspended in the solution and to transport the particles to the cathode surface. It has been agreed that increased agitation normally enhances the particle incorporation in the metal deposit because particle transfer from the bulk of the electrolyte to the cathode surface is boosted. However, excessive agitation may decrease particle incorporation, for the vigorous hydrodynamic forces in the electrolyte can prevent the particles from being entrapped in the metal deposit. Owing to the relatively large size of the CrAlY particles (∼7 μm) used in this DoE study, agitation played the most important role in particle incorporation since sufficient agitation was crucial in keeping the CrAlY particles in suspension.

The CrAlY particle loading was another important factor in affecting the particle incorporation in the composite coating, which increased significantly when the particle loading was raised from 10 to 50 g L⁻¹; such a trend has been observed for a variety of codeposited systems.^26,28 For plating solutions with a low particle concentration, the particle incorporation is limited by the supply of particles to the cathode surface by agitation and diffusion. Increasing the overall amount of particles in the electrolyte enhances the probability of particles to reach the electrochemical double layer at the cathode and thus improves particle incorporation. However, too high a particle concentration can increase the chance of collisions between particles and thus have a negative impact on particle incorporation. ²⁶ For the present set-up and coating system, a combination of high agitation speed and high particle loading appeared to be beneficial for increasing the CrAlY particle incorporation.

Even though current density is a key parameter in traditional electroplating, its effect on particle incorporation varies, depending on the codeposited coating system. Different types of relationships have been observed,^28,29 i.e. the particle content in the composite coatings either increases or decreases continuously with the current density, or exhibits one or multiple peaks as a function of current density. In this study, while the current density was relatively insignificant with regard to particle codeposition, it had an effect on the overall coating quality by affecting the porosity and thickness uniformity. If the rate of metal deposition cannot keep up with the speed of particle settlement, there is a tendency to form less dense coatings, as observed in the coating deposited with 50 g L⁻¹ of CrAlY particles (Fig. 4c).

Conclusion

Prealloyed CrAlY particles were codeposited into the Ni coating via a sediment codeposition process in a Watts bath. A DoE study was conducted to understand the effects of deposition parameters, including current density, agitation speed and concentration of CrAlY particles in the plating bath, on the volume fraction of CrAlY particles in the composite coatings. The results indicate that the agitation rate had the most significant effect, followed by the particle loading, whereas the current density exerted very minimal influence. For the present codeposition configuration and coating system, a combination of high agitation speed and high particle loading was beneficial for enhanced particle incorporation, with the maximum amount of embedded CrAlY particles being ∼49 vol.-%. The DoE study also suggests a non-linear relationship between the three deposition parameters and the amount of incorporated CrAlY particles in the coating.