Optimization of Spray Parameters and Corrosion Properties of Plasma-Sprayed Cr₂O₃ Coatings Using Response Surface Methodology

Abstract

In this study, the experimental design of response surface methodology was used to explore the interaction between spraying parameters to obtain an optimized process to reduce the porosity of the coating, and to prepare an excellent chromium oxide coating. The order of the single parameter affecting porosity is as follows: power > main gas > spraying distance > carrier gas flow. This study found that the spraying process with the lowest porosity of the chromium oxide coating is as follows: power of 625 W, stand-off distance of 105 mm, primary gas of 42.5 lpm, carrier gas flow of 5 lpm, and feed powder delivery rate of 35 g/min. The EDS results show that the Cr and O elements in the coating with the lowest porosity are uniformly distributed, while for the coating with the highest porosity, the elements are unevenly distributed to a certain extent, which is caused by the unevenness of the structure caused by the structure defects. The corrosion current density of chromium oxide coating VI (low porosity) is 4.34 × 10⁻⁶ A, whereas that of chromium oxide coating IV (high porosity) is 1.862 × 10⁻⁵ A. On the coating with the highest porosity, the corrosion activity is dominant, while the minimum porosity of coating is the smallest.

1. Introduction

In recent years, material surface engineering was developed to enhance the durability of conventional metal surfaces against degradation caused by corrosion. This involved either altering the microstructure and composition of the surface layer or creating a wear-resistant layer to withstand external damage. By applying surface modification technologies, such as coatings with wear-resistant and corrosion-resistant properties, protection was achieved without modifying the material itself. This approach proved beneficial as it imparted special performance characteristics to materials, reducing friction (wear) and corrosion, minimizing material consumption, and lowering costs. Consequently, it gained widespread application prospects and attracted significant attention from scholars both domestically (China) and internationally. Numerous research institutions globally explored various new technologies and methods to improve the surface properties of materials. Surface engineering technology emerged as a critical direction in the advancement of materials science. Thus, investigating protection technologies for the corrosion of materials is of considerable importance.

At present, commonly used surface technologies for coating preparation in industry include laser cladding, weld overlays, plasma spraying, thermal spraying, electrochemical methods, chemical vapor deposition, ion plating, and sputtering. These technologies exhibit distinct characteristics and are widely applied across various fields. Typically, surface engineering technologies focus on protective coating techniques to achieve more corrosion-resistant and abrasion-resistant coatings. This approach is particularly advantageous as it allows the replacement of expensive high-grade stainless steel alloys with relatively cheaper carbon steel alloys, making surface coatings highly favorable. Therefore, material surface engineering plays a significant role, ranging from decorative and esthetic functions to protecting substrates from wear, corrosion, and other forms of material degradation [1].

In marine structures, shipbuilding industries, and other engineering applications where Q235 steel is used, it faces the challenges of corrosion and wears, hence need for surface modification. Machine parts and components made from Q235 steel are constantly exposed to various environmental conditions during their service life. In most applications, they encounter abrasive wear accompanied by corrosive media with extreme temperature changes (offshore mining), corrosive environments with biofouling, and corrosive gasses with high temperatures (oil and gas transport systems), etc. A good surface coating with an excellent adhesion bond between the coating/substrate interface and cohesion between the build-up layers is needed for steel that is used in these conditions, hence the application of chromium oxide plasma-sprayed coatings. The selection of chromium oxide coatings with reasonable small porosity levels has a guiding significance for its wide range application.

Chromium oxide coatings, one of the hardest oxides, exhibit a low frictional coefficient, high wear resistance, and corrosion resistance. They have superb chemical inertness, mechanical strength, high hardness, and toughness. These qualities make chromium oxide a popular candidate in research and industrial application for corrosion protection, chemical resistance, wear, and abrasive resistance. Chromium coatings produced by atmospheric plasma spraying have proven effective for their extreme hardness, good adhesion, high fracture toughness, homogenous structure, and uniform thickness. They are, however, faced with the challenges of porosity and microcracks. The quality of chromium oxide coating plays a decisive role in matrix protection, and porosity is an important indicator for evaluating coating quality. Excessive pores in the coating reduce the bonding strength between the coating and the substrate and inside the coating, which makes the coating difficult to adapt to harsh environments, such as high temperature and strong corrosion, reducing its application range and service life.

To achieve better service results, a systematic study of the chromium oxide coating is required to select the optimum porosity and improve the coating quality to cope with the above-mentioned environments. Chromium oxide is, however, challenging to spray due to its high melting point, low thermal conductivity, and tendency to vaporize in high temperatures. The vaporization of chromium oxide during spraying creates extremely fine dust particles that gather on the workpiece and are trapped inside the coating layers reducing the cohesion and mechanical properties of the coating. Thus, this study was carried out to optimize the quality of the atmospheric plasma-sprayed chromium oxide coating by optimizing the spraying process parameters to attain minimum porosity and maximum hardness using research surface methodology (RSM) and design of experiments (DOE). The spraying power level was typically controlled through the plasma spray controller by adjusting the spraying current or the supplied voltage. In most studies on chromium oxide coatings, it was referred to as either spraying current or power. Typical current and power ranges reported in the majority of studies were between 500 A and 700 A [2,3,4,5]. The spraying current or power influenced the flame temperature and particle velocity, thereby affecting the plasma generated to melt the feedstock powder, as investigated by Ramachandran et al. [6] and Markus et al. [7]. Both studies demonstrated that higher power levels or currents led to an increase in plasma temperature and particle velocity, ensuring complete melting of the feedstock particles. However, high particle velocity was found to impact splat formation, with excessively high temperatures causing splat distortion and limited contact with the substrate. An optimal temperature was identified as critical for achieving quality splat formation and bonding. Salazar et al. [8] investigated the effects of high and low power on chromium coatings. They found that high power led to a high volume of hard phases, resulting in coatings with high hardness values and excellent wear resistance under high applied loads. Conversely, low power promoted the formation of high volumes of spinel (tough phases), which were associated with high toughness values and optimal wear responses under low loads. Simanchal et al. [9] also examined the influence of arc current on the elastic modulus and fracture toughness of chromium oxide coatings. They observed that increasing the arc current correlated with an increase in partially molten particles due to the high particle velocity, which caused many particles to be blown away before complete melting, i.e., reduced dwell time. This phenomenon led to variations in fracture toughness. An optimal current value was determined to be essential for achieving well-splatted coatings with optimal quality.

In any study or actual coating exercise, it is difficult to control all the parameters because of time constraints and the interdependence of these parameters. Usually, a compromise is reached that simultaneously allows one to realize an effective coating for application. The spraying power level is usually controlled from the plasma spray controller by regulating the spraying current or the voltage supplied. In most of the studies on chromium oxide coating, it is referred to either as Spraying Current or Power. In both cases, they found that high power levels or current caused an increase on plasma temperature and velocity of the particles. This ensured the complete melting of the feedstock particles. An optimal temperature was found to be effective for quality splat formation and bonding. Markus et al. [7] found out that there was 60% of the unmolten particles for cold conditions, 20% for the medium condition, and only 2% for the hot conditions.

The stand-off distance (SOD) is the distance between the spray gun and the work piece. The SOD influences the final quality of coating produced. Sui et al. suggested [10] that the stand-off distance directly affects the melting of the feedstock particles. Long stand-off distance guarantees the complete melting of feedstock, hence uniformity in microstructure of the plasma-sprayed coatings. Conversely, the already molten particles can re-solidify inflight if the stand-off distance is too long. This could lead to non-melted particles being lodged into coating [11]. Shorter SOD occasioned fully molten and semi-molted particles which developed coating which were bimodal in nature [12,13].

The purpose of a carrier gas is to channel and direct the feedstock powder to the plasma gun. The center of the plasma jet is the hottest part of the plasma, with the maximum velocity and is also the viscous section of the plasma. When the powder enters this region there is complete melting [14]. The commonly used carrier gas is argon. Some researchers found out that a high gas flow rate reduces the coating porosity.

As one of the widely used plasma-sprayed coatings, chromium oxide coatings have excellent properties [15]. Nevertheless, plasma-sprayed chromium oxide coatings also suffer from porosity defects, which are considered to have a negative impact on properties like hardness, strength, wear resistance, and corrosion resistance. It is thus necessary to reduce porosity hence improve the service life of chromium oxide coatings. Response surface methodology (RSM) was widely employed to optimize the preparation parameters of coatings using Minitab 19 software [16], where multiple quadratic regression equations were used to determine the optimal values of predicted responses and corresponding factors. The central composite design (CCD) was found to be suitable for multi-factor, multi-level experiments involving continuous variables, as it provided better surface fitting compared to other methods. RSM was concluded to be an accurate and widely accepted experimental design method for optimizing preparation parameters. Consequently, RSM was adopted in this study to investigate the effects of parameters on the corrosion property of APS (Atmospheric plasma spraying) Cr₂O₃ coatings.

2. Materials and Methods

The base metal used in this study was Q235B steel. The substrate material was cut into required sizes using Electrical Discharge Machining (EDM, ZhengGuang Electrical Technology Co., Ltd. Suzhou, China). The samples were then cleaned in an alcohol bath in an ultrasonic cleaner to remove grease and dirt that were accumulated during the cutting process. After cutting the substrates from the bigger pieces, they were cleaned ultrasonically in an alcohol bath to remove grease and oil that was used as a coolant during the cutting process.

Commercially available Cr₂O₃ powder (Metco 6156, Oerlikon Metco, Westbury, NY, USA) was used as the coating material for production. The powder had a typical nominal particle sizes distribution in the range of −35 to +15 μm. The feedstock powder was manufactured by sintering and crushing which resulted in blocky and smooth particles with irregular and angular shapes. A commercial atmospheric plasma spraying thermal coating system (UniCoatProTM, Oerlikon Metco, Westbury, NY, USA), supplied by Oerlikon Metco, Pfaffikon, Switzerland) was used for this experimental work. On the basis of the suggested spraying parameters that come with the METCO equipment, this paper further optimizes the spraying parameters for Cr₂O₃, as shown in

Table 1. Results of the parameter range investigation.

			Levels
Factors	Notations	Units	Lowest (−2)	Low (−1)	Middle (0)	High (+1)	Highest (+2)
Power	P	W	575	600	625	650	675
Primary gas flow rate	G	lpm	30 (3.5)	35 (4.2)	42.5 (4.75)	50 (5.5)	55 (6)
Stand-off distance	S	mm	80	95	105	125	130
Carrier gas flow rate	C	lpm	3	4	5	6	7

.

After the investigation of all terms, achievable parameter boundaries were selected to obtain practical high-quality coatings. The second-order central composite rotatable design was recognized as the most effective method in RSM, allowing the determination of empirical relationships for the response surface with a minimal number of experimental runs while maintaining accuracy. Given the wide range of factors, a five-factor, five-level central composite design matrix was utilized to optimize the experimental conditions.

Table 1 summarizes the ranges of the factors considered 0, while

Table 2. Experimental matrix design.

Run	Blk	Power	Primary Gas Flow Rate	Stand-Off Distance	Carrier Gas Flow Rate
1	1	−1	−1	−1	1
2	1	−1	1	−1	−1
3	1	−1	1	1	−1
4	1	0	0	0	0
5	1	0	0	0	−2
6	1	1	1	−1	−1
7	1	0	0	0	0
8	1	−1	−1	1	−1
9	1	1	−1	1	−1
10	1	0	0	−2	0
11	1	−1	−1	1	1
12	1	0	−2	0	0
13	1	1	−1	−1	−1
14	1	0	0	0	0
15	1	1	−1	1	1
16	1	1	1	1	−1
17	1	0	0	2	0
18	1	0	0	0	2
19	1	0	2	0	0
20	1	1	1	1	1
21	1	0	0	0	0
22	1	0	0	0	0
23	1	−1	1	1	1
24	1	−2	0	0	0
25	1	1	−1	−1	1
26	1	1	1	−1	1
27	1	−1	−1	−1	−1
28	1	0	0	0	0
29	1	0	0	0	0
30	1	2	0	0	0
31	1	0	0	0	0
32	1	−1	1	−1	1

and

Table 3. Experimental matrix design with the responses used in the initial stage of this study.

S/No	Power	Ar/H2 Gas	SOD	Carrier Gas	Porosity
1	600	35/4.2	95	6	6.569
2	600	50/5.5	95	4	8.614
3	600	50/5.5	125	4	8.252
4	625	42.5/4.8	105	5	8.356
5	625	42.5/4.8	105	3	7.487
6	650	50/5.5	95	4	7.392
7	625	42.5/4.8	105	5	7.907
8	600	35/4.2	125	4	9.689
9	650	35/4.2	125	4	7.095
10	625	42.5/4.8	80	5	8.132
11	600	35/4.2	125	6	8.672
12	625	30/3.5	105	5	5.032
13	650	35/4.2	95	4	5.61
14	625	42.5/4.8	105	5	7.411
15	650	35/4.2	125	6	7.542
16	650	50/5.5	125	4	7.411
17	625	42.5/4.8	130	5	4.813
18	625	42.5/4.8	105	7	6.121
19	625	55/6	105	5	13.275
20	650	50/5.5	125	6	6.25
21	625	42.5/4.8	105	5	14.757
22	625	42.5/4.8	105	5	7.19
23	600	50/5.5	125	6	13.112
24	575	42.5/4.8	105	5	12.659
25	650	35/4.2	95	6	12.877
26	650	50/5.5	95	6	4.089
27	600	35/4.2	95	4	6.28
30	625	42.5/4.8	105	5	5.617
31	625	42.5/4.8	105	5	4.091
32	675	42.5/4.8	105	5	4.0798

illustrate the 32 sets of coded conditions employed to construct the design matrix. The first 16 experimental conditions were derived from a half-factorial experimental design matrix (2⁵⁻¹ = 16). All variables at the intermediate (0) level represented the center points, whereas the combinations of each process variable at either the lowest (−2) or the highest (+2) value with the other four variables maintained at intermediate levels constituted the star points. As a result, the 32 experimental runs facilitated the estimation of the linear, quadratic, and two-way interactive effects of the variables on the Cr₂O₃ coating deposits.

The model chosen includes the effects of the main and interaction effects of all the factors. The model selected is polynomial and is expressed as follows:

Y = b₀ + Σb_ix_i + Σb_iix_i² + Σb_ijx_ix_j

(1)

For the five factors, the selected polynomial can be expressed as follows:

Y = b₀ + b₁(P) + b₂(G) + b₃(S) + b₄(F) + b₅(C) + b₁₁(P²) + b₂₂(G²) + b₃₃(S²) + b₄₄(F²) + b₅₅(C²) + b₁₂(PG) + b₁₃(PS) + b₁₄(PF) + b₁₅(PC) + b₂₃(GS) + b₂₄(GF) + b₂₅(GC) + b₃₄(SF) + b₃₅(SC) + b₄₅(FC)

(2)

where b₀, is the average of responses while b₁, b₂, and b₂₃ are the coefficients that depend on their respective main and interaction effects of the parameters. Interdependencies of higher order were neglected. The significant effects were chosen based on an analysis of variances (ANOVA), whereas effects with noise probability higher than 5% were excluded for the further evaluation. Step by step, the effect with the highest probability was removed and the analysis was repeated until only significant effects remained. The remaining effects were normalized to the average value b₀ to allow a better comparison of the investigated properties in this study. Porosity was selected as a response for the optimization study.

A five-factor, five-level central composite rotatable design was used. The regression coefficients were calculated with the help of Minitab 17 statistical software. After determining the coefficients (at a 95% confidence level), the final empirical relationships were developed using the coefficients obtained.

The results of the ANOVA test for the responses, specifically deposition efficiency and porosity, are presented in

Table 4. Analysis of variance table for deposition efficiency.

Source	Adj. SS	Adj. MS	F-Value	p-Value Prob > F	Significance
Regression	16.7211	5.5737	12.11	<0.0001	significant
Primary Gas	7.9747	7.97469	17.35	<0.0001
C2	8.0686	8.06855	17.52	<0.0001
PS	0.8057	0.80573	1.75	0.197
Error	12.8923	0.46044
Lack of fit	12.6757	0.60361	19.51	0.000	not significant
Pure error	7	0.2166	0.03094
R2	76.46%
Adj. R2	71.80%
Pred. R2	62.14%
Adeq. Prec.	6.7856

. The acceptability of the model was validated using the ANOVA system. In this study, the model F-value and the corresponding probability values were analyzed to confirm the significance of the empirical correlations. The predominant factors exerting major and minor effects on the responses were evaluated based on the F-values. From the F-value analysis, it was determined that the predominant factors influencing the responses, in hierarchical order, were power, stand-off distance, primary gas flow rate, carrier gas flow rate, and powder feed rate. Throughout this study, the powder feed rate was maintained following a screening procedure.

The phase identification and composition were carried out by X-ray diffraction (XRD) using a Bragg–Brentano geometry operating with Cu Kα radiation with a Rigaku Ultima IV XRD diffractometer (Rigaku Ultima IV, Rigaku, Tokyo, Japan) in a range of 2θ = 15° to 120° with a step size of 0.02° and 5 s/step. The surface characteristics and morphology of the feedstock and coatings were examined using a scanning electron microscope (SEM, TM3030, Hitachi Limited, Tokyo, Japan) which was equipped with energy dispersive spectroscopy (EDS). For coatings, top surfaces and cross-section of the microstructure were studied in this exercise. Eight SEM cross-sectional images of different areas were taken for each set of coatings, and then their porosity was calculated and measured using the gray scale method and ImageJ software 1.8.0.

The corrosive environmental setting used in the current investigation was 3.5 wt. % NaCl solution medium exposed to the atmosphere. The electrochemical corrosion kinetic performance of the chromium oxide coatings and the substrate (Q235) were evaluated using the potentiodynamic polarization test with a potentiostat (Swiss Metrohm Autolab Electrochemical Workstation-Yutong China Co., Ltd., Herisau, Switzerland) at room temperature after 1 h of immersion in the electrolyte for equilibrium of the open-circuit potential (OCP). Prior to the test, the samples were rinsed in deionized water and ethanol, followed by dehydrating in warm air. A conventional flat three-electrode cell with graphite rod as the counter electrode, a saturated calomel reference electrode (SCE) and coating/substrate as the working electrode, was used. The polarization test was conducted on a sample of 1 cm² exposed to the atmosphere at a scan rate of 0.5 mV/s in the range of −250 to 800 mV, with reference to OCP [17]. Tafel plots inference method was used to obtain the corrosion potential and corrosion current densities from the polarization plots. EIS analysis was performed using the standard 3-electrode corrosion cell to study the stability, viz. barrier performance of the chromium oxide coatings. Impedance values were determined in the frequency range of 100 kHz to 10 mHz, with an applied sinusoidal perturbation of ±10 mV at OCP. The examination of the obtained spectra was accomplished using Nova 1.10 software to evaluate the electrochemical parameters. The surface of the post-corroded samples was described by SEM and EDS mapping to explore on the morphology of the corrosion products. The corrosion products were also examined by XRD.

3. Results

3.1. Optimization of Plasma Spraying Parameters for Depositing Cr₂O₃ Coatings Using RSM and Design of Experiments

A quadratic model was identified as the best fit for the data. The model was fitted using a stepwise automatic reduction algorithm to eliminate insignificant terms at a 95% significance level. The probability plot for deposition was illustrated in Figure 1, while the residual plots for deposition were depicted in Figure 2. The model demonstrated a significance level of <0.0001. The lack of fit was not statistically significant. The R2 value was 0.76, exceeding the recommended threshold of 0.6. The difference between the Adjusted R2 value and the Predicted R2 value was less than 0.2. The adequate precision value was also found to be sufficient. It was concluded that the model exhibited good performance. The residual versus fitted values plot displayed homoscedasticity, confirming that the assumptions of the linear model were satisfied.

One parameter and two interactions were identified to significantly influence the coating deposition efficiency. These included the primary gas flow rate and the interaction between carrier gas flow rate squared and power, as well as the interaction between power and stand-off distance (P*SOD). The model was expressed in terms of coded factors in Equation (4). Figure 3a illustrated the effects of power and SOD on deposition. It was evident that these factors did not exhibit a diverse impact on deposition. The specified limits resulted in uniform coating, which was considered excellent for this study, with minimal disparities. Figure 3b depicted the effects of primary gas flow rate and carrier gas flow rate on deposition. It was observed that these two factors had a significant influence on deposition. An increase in the carrier gas flow rate was found to enhance deposition efficiency, whereas an increase in the primary gas flow rate reduced deposition efficiency. This observation was consistent with the understanding that increasing the carrier gas flow rate ensured a greater quantity of feedstock particles were introduced into the plasma stream for deposition.

Deposition regression equation is as shown in the following:

Deposition = + 2.241 − 0.0532 Primary Gas + 0.03242 (C²) + 0.00012 (PS)

(3)

The determination coefficient (R2) denotes the goodness of fit for the model. For the case of deposition efficiency, determination coefficient (R2 > 0.76) indicates that less than 24% of the total variations are not explained by the experimental relationships. The explanation is deemed to be other factors that were assumed to be constant during all the experiments in this study. The value of the adjusted determination coefficient is also high, which suggests the high significance of the empirical correlations. The predicted R2 values also show good agreement with the adjusted R2 values.

Adequate precision was assessed by comparing the range of predicted values at the design points with the average prediction error. Simultaneously, a relatively low coefficient of variation indicated enhanced precision and reliability of the conducted experiments.

The porosity probability plot and residual plots for porosity are presented in Figure 4, respectively. The 95% confidence interval (CI) was clearly achieved, with the residual plots demonstrating a reduced number of outliers. The model met the criteria for application in this study. The plot of residuals versus fitted values exhibited homoscedasticity, confirming that the assumptions of the linear model were satisfied, consistent with the observations for deposition efficiency.

The porosity regression equation, as shown in Equation (4), indicated that Power (P), Primary Gas (G), stand-off Distance (SOD), and Carrier Gas (C) had a significant influence on porosity. Porosity exhibited an increase with higher levels of Power, Primary Gas flow rate, and Carrier Gas flow rate, while it decreased with an increase in stand-off Distance. The equation also identified significant interactions between Power and stand-off Distance, Power and Primary Gas flow rate, Power and Carrier Gas flow rate, Primary Gas and stand-off Distance, Primary Gas and Carrier Gas, and stand-off Distance and Carrier Gas. These interactions are clearly demonstrated in the contour plots presented in Figure 5.

Porosity = 7.0 + 0.0030 P + 0.008 G − 0.047 SOD
+ 0.01 C − 0.000264 P × G + 0.000072 P × SOD
+ 0.00049 P × C + 0.00106 G × SOD
+ 0.0132 G × C − 0.00854 SOD × C

(4)

The porosity and other intrinsic defects in coatings were influenced by the degree of feedstock particle melting within the plasma flame column and the extent of spreading upon impact with the substrate. This condition facilitated the flattening of molten particles, promoting the formation of a uniform lamellar structure. It was reported that the high-velocity impact of molten particles on the substrate surface led to greater spreading, thereby reducing the likelihood of porosity.

The porosity regression model (Equation (4)) demonstrated that the stand-off distance exerted the most significant influence on porosity, with the highest porosity observed at shorter stand-off distances. This was attributed to insufficient dwell time for complete particles melting at very short distances, resulting in partially melted zones that, upon cooling, exhibited poor adhesion and potential void formation due to mechanical interlocking of particles. The carrier gas flow rate was identified as the second most influential factor, with higher flow rates correlating with increased porosity levels. An optimal carrier gas flow rate was identified, as both excessively low and high flow rates were found to similarly elevate porosity levels.

Several correlational effects on porosity were identified. The interaction between gas flow rate and carrier gas exhibited the most significant influence in this study. Figure 5 illustrates the effects of primary gas flow rate and carrier gas flow rate on porosity. The highest porosity values were observed under two conditions: low carrier gas flow rate combined with low primary gas flow rate, and high carrier gas flow rate combined with high primary gas flow rate. At low primary gas flow rates, sufficient dwell time allowed particles to fully melt, which could have resulted in effective coating. However, this was counteracted by the low carrier gas flow rate, which prevented the molten particles from achieving the appropriate trajectory and speed necessary for adequate deposition and layering. As a result, coatings with the highest porosity were formed. Conversely, high primary gas flow rates reduced the dwell time within the plasma plume, and the high carrier gas flow rate caused partially molten particles to be ejected at high velocities, leading to poor coating surfaces. An optimal balance between these parameters was required.

The surface morphology of the as-sprayed chromium oxide coating was characterized using SEM micrographs, as shown in Figure 6. The coating consisted of a combination of partially melted and fully melted particles. Partially melted particles exhibited small spherical morphologies, whereas fully molten particles formed flattened splats, with some pores observed within the coating. Detailed analysis of the samples indicated the presence of numerous microcracks, predominantly aligned along the splat boundaries. These microcracks were associated with shrinkage mechanisms occurring during the cooling process.

The presence of microcracks and porosity was detrimental to the coating, leading to a compromise in its properties. The figure illustrates a typical top-view morphology and cross-section of one of the plasma-sprayed Cr₂O₃-based coatings before optimization. The cross-sectional view revealed that the coating was moderately dense. Additionally, CrO₂ suboxides, corresponding to the white spots, were observed, resulting from localized temperature variations during rapid cooling and solidification, as reported by Ouyang et al. [18] and Richard et al. [19]. The figure also depicts a typical top-view morphology and cross-section of one of the plasma-sprayed Cr₂O₃-based coatings before optimization. Throughout this study, the coating thickness was maintained within the range of 200–350 µm. Common defects included pores, unmolten particles, and microcracks. Figure 7 presents a graphical representation of the statistical porosity measurements of the APS Cr₂O₃ coatings obtained using ImageJ 1.8.0 software. These measurements were categorized based on the stand-off distances, i.e., 80 mm, 105 mm, and 130 mm, and designated according to the experimental run order. A general observation suggested that a stand-off distance of 105 mm was a reasonable estimate for optimization, although other spraying parameters also significantly influenced the level of porosity and other inherent microstructural flaws in the coatings. The various types of pores encountered were predominantly globular pores and microcracks, which were commonly associated with incomplete filling and infiltration of molten particles, rebound dissipation of portions of semi-molten particles, and the formation of voids due to trapped gasses. In preliminary studies, it was determined that the major process parameters influencing coating quality had a significant impact on the microstructural features.

Figure 8 illustrated the influence of the interaction between carrier gas and stand-off distance on porosity. The highest porosity percentages were observed under two specific conditions: low stand-off distance (SOD) combined with low carrier gas flow rates, and high SOD paired with the highest carrier gas flow rates. At low SOD and low carrier gas flow rates, particles traveled the shortest distance and attained the lowest impact velocity. The low impact velocity prevented the molten particles from spreading adequately to form a uniform lamellar structure, leading to the formation of poorly adhered lumps. This resulted in a heterogeneous coating containing voids, which, upon sufficient cooling, formed pores. Conversely, under conditions of high SOD and high carrier gas flow rates, the opposite phenomenon was observed; however, the outcome was similarly characterized by poor coating features. High SOD caused the molten particles to travel a longer distance, prolonging their interaction time with the air, which in some cases could induce oxidation. In this scenario, however, it primarily initiated the solidification and cooling of the particles. When these particles impacted the substrate surface, their ability to spread was restricted due to the onset of the solidification process. Therefore, a balance between these factors was consistently required to achieve an optimal and effective coating.

The findings from the above discussions aligned with the preliminary study. The models facilitated the identification of parameters that optimized the coating characteristics, ensuring the attainment of the lowest porosity percentage. The RSM models uncovered several interaction effects that would not have been identified if the traditional one-factor-at-a-time approach had been utilized in this study.

The model underwent validation to verify its reliability and suitability for further investigation. Based on the models, specific parameters were selected, which yielded the lowest porosity percentage among the 32 trial tests conducted.

Optimization was accomplished by selecting coatings from the screened stage that exhibited the lowest porosity percentage, followed by conducting additional tests to assess their performance under these conditions. From screening and optimization process, we came up with 5 coatings with lowest porosity level, as shown in

Table 5. Spraying parameters for optimization.

Coating	Current (A)	Primary Gas (lpm)	SOD (mm)	Carrier Gas (lpm)	Porosity (%)
I	500	30	130	7	5.61
II	500	42.5	105	5	4.81
III	650	42.5	80	5	5.03
IV	625	30	105	5	6.25
V	625	42.5	105	5	4.08

.

During the optimization process, the feed powder rate was maintained at 35.26 g/min. This rate was determined to be optimal for achieving a high-quality coating based on prior research. The objective was to produce a dense, uniform, and durable coating capable of maintaining its integrity over extended periods in both marine and general industrial environments. The initial optimization focused on the stable Cr₂O₃ coating, with the aim of creating a dense and long-lasting coating that would retain its integrity prolonged durations. The goals and significance levels for each response are evaluated and summarized in

Table 6. Cr₂O₃ optimization parameters.

Response	Goal
Porosity (%)	minimize
Corrosion resistance	maximize

.

The coatings produced after optimization were designated coatings I, II, III, IV, and V.

3.2. Microstructure of the Optimized APS Cr₂O₃ Coatings

Following the optimization process, five coatings were selected for further analysis. These coatings exhibited the lowest percentage porosity and were labeled as Coating I through Coating V. Visual inspection revealed that the surfaces were smooth and devoid of any apparent structural defects.

Figure 9 displayed the XRD patterns of the optimized APS Cr₂O₃ coatings.

All five coating groups exhibited a single Cr₂O₃ phase composition. Furthermore, no significant alterations in the phase constituents were observed despite variations in the spraying parameters. Simultaneously, the strongest diffraction peak detected in all coatings corresponded to the Cr₂O₃ (104) plane (observed at 2θ = 33.54°). This finding indicates that the crystalline grains in all five coating groups demonstrate consistent preferred orientation during growth.

The morphology of the optimized coatings, designated as Coatings I, III, IV, and V, was examined using SEM micrographs. Figure 10 illustrates the typical cross-sectional and top-view morphologies of plasma-sprayed Cr₂O₃-based coatings prepared with different optimized spraying parameters. Figure 10(a1,a2) depicts the microstructure of Coating I. Figure 10(a1) shows the cross-sectional view of the coating, which exhibited a thickness of approximately 150 µm. The primary defects observed in the oxide coatings included typical pores, microcracks, and partially melted particles. These defects were also evident in the top-view microstructure. Coating I was fabricated under the following spray parameters: current of 500 A, primary gas flow rate of 30 lpm, stand-off distance (SOD) of 130 mm, and carrier gas flow rate of 7 lpm. The resulting porosity level was 5.61%. It was evident from the optimized conditions that Coating I had the lowest spraying current among the five coatings. Despite the low current, the highest carrier gas flow rate ensured a lower porosity percentage. The presence of unmolten particles was likely attributed to the longest stand-off distance. As previously reported in the literature and supported by the developed regression model, a longer SOD increased the likelihood of particle solidification during flight before impacting the substrate. However, this effect was mitigated by the high carrier gas flow rates employed.

Figure 10(b1,b2) presents the microstructural morphology of Coating IV, which exhibited a density comparable to that of Coating I. The microstructure was densely packed and demonstrated strong adhesion to the substrate. The coating thickness ranged between 200 and 250 µm, representing the highest value among all coatings. The parameters used for its preparation were: current of 625 A, primary gas flow rate of 30 lpm, SOD of 105 mm, and carrier gas flow rate of 5 lpm. Although Coating IV displayed a densely packed cross-section, the top-view revealed a higher porosity percentage of 6.25% compared to the other coatings. This was attributed to the primary gas flow rate used. A primary gas flow rate of 30 lpm was considered the lowest in this study. It has been reported elsewhere that a low primary gas flow rate has a similar effect to increasing the spraying current, resulting in enhanced particle melting and uniform coating formation. However, in this case, it was noted that Coating IV exhibited the highest porosity percentage among all coatings.

Figure 10(c1,c2) depicted the microstructure of both the cross-section and top view of Coating III. Similarly to the previously discussed micrographs, these images also exhibited inherent typical defects such as pores, microcracks, and unmolten particles. Coating III was relatively thin compared to the others, with a thickness of less than 100 µm. The spraying conditions were as follows: current of 650 A, primary gas flow rate of 42.5 lpm, stand-off distance (SOD) of 80 mm, and carrier gas flow rate of 5 lpm. The porosity percentage achieved was 5.03%. Among the spraying conditions, two factors were identified as more dominant: the spraying current and the stand-off distance. The balancing parameters were the forming gas and the carrier gas. It was reported that a high spraying current led to sufficient and complete melting of the feedstock particles, ensuring a uniform coating with minimal unmolten particles. Conversely, a small stand-off distance was also reported to result in a uniform coating, as the melted particles were not exposed to the atmosphere for an extended period, thereby reducing their chances of resolidification. Based on these observations, Coating III represented a middle ground with moderate features, although the shortest SOD still influenced the coating properties.

Figure 10(d1,d2) illustrated the cross-section and top view of the morphology of Coating V. Similarly to the other cases, it exhibited typical features such as pores and microcracks in the top view and some unmolten particles in the cross-section. Among all the coatings, Coating V demonstrated the lowest porosity level of 4.08%. The spraying parameters were as follows: current of 625 A, primary gas flow rate of 42.5 lpm, stand-off distance (SOD) of 105 mm, and carrier gas flow rate of 5 lpm. The chromium oxide coating was relatively thin and compact, indicating strong adhesion to the substrate. This observation was further supported by the XRD and EDS analysis presented in the subsequent section.

The magnified SEM images of Coating II are given in Figure 11 and Figure 12. Figure 11 presented the EDS mapping results of the optimized chromium oxide for Coating II, which exhibited the highest porosity. The mapping revealed a non-uniform distribution, with numerous patches observed above the green layer. This non-uniformity indicated an uneven coating, accompanied by microstructural flaws, as previously observed in the coating morphology analysis. The aggregation of Cr elements in most regions suggested the presence of increased pores or microcracks on the coating surface. A spectrum spot was analyzed on the surface of the optimized chromium oxide coating from Coating IV (characterized by high porosity), as depicted in Figure 12. The corresponding elemental information table and EDS spectrum analysis are provided. The Cr atom content was measured at 61.61%, while the O atom content was 38.39%. These values suggested uneven packing of coating particles during the deposition process, leading to the formation of pores and unfilled voids due to inadequate flattening upon impact. Figure 13 illustrated the EDS line scan, demonstrating the distribution of Fe, Cr, and O, which exhibited a distinctive bilayer structure composed of large grains and small crystals. This figure corresponded to the optimized chromium oxide coating for Coating V, which had a low porosity level. The scan direction extended from the substrate to the coating. The results indicated a generally uniform distribution of Cr and O across the analyzed area. Fe elements were not detected until the scan reached the substrate/coating interface, signifying a well-layered coating with superior structural integrity.

3.3. Optimized Corrosion Resistance of APS Cr₂O₃ Coatings

Typical potentiodynamic polarization curves (E vs. log i) for the coatings and Q235 grade steel substrate in 3.5 wt. % NaCl solution exposed to air at room temperature were presented in Figure 14. The measured values of corrosion potential (Ecorr) and corrosion current density (icorr) for the different coatings are summarized in

Table 7. Electrochemical kinetic parameters of the various coatings estimated from the polarization curves after exposure in aerated 3.5 wt. % NaCl solution.

Sample	βa mV Decade−1	βc mV Decade−1	Ecorr, mV	Icorr, µA cm−2	CR, mpy
Coating I	208.20	340.49	−1006.33 ± 3	8.49	1.4814
Coating II	222.80	367.43	−952.56 ± 6	6.36	1.2725
Coating III	1171.4	320.64	−1029.51 ± 5	11.92	1.5976
Coating IV	116.90	432.34	−1047.5 ± 9	18.62	1.8476
Coating V	170.00	284.70	−903.63 ± 1	4.78	0.6674
Substrate	421.64	523.16	−1191.72 ± 11	26.71	3.1033

.

Figure 14 clearly demonstrates that the substrate underwent significant active corrosion across the entire range of the anodic polarization scan, in contrast to the chromium oxide-based coatings, which exhibited superior performance. Although the coatings did not display distinct passivation regions, they showed lower current density (icorr) values. The Ecorr and icorr values of the substrate (−1191.7 mV and 26.7 µA cm⁻², respectively) were observed to be more negative and higher, respectively, compared to those of the optimized chromium oxide coatings. This confirmed that the substrate metal was more prone to corrosion than the optimized coatings. The optimized chromium oxide coatings exhibited higher Ecorr and lower icorr values, as illustrated in Figure 14. The Ecorr and icorr values for Coating I, Coating II, Coating III, Coating IV, and Coating V are tabulated in

Table 8. Electrochemical impedance parameters of the optimized Cr₂O₃ coatings derived from the EIS plots obtained after exposure in 3.5 wt. % NaCl solution at OCP.

Sample	Rs, Ω cm2	Cc, µF cm−2	Rp, Ω cm2	Rct, Ω cm2	Cct, µF cm−2	Rt, Ω cm2
Coating I	5.48	121	151.42	1370.3	1000	1527.20
Coating II	4.08	58.9	131.73	1351.6	707	1487.40
Coating III	8.00	215	140.11	1068.4	3640	1216.50
Coating IV	8.10	26.4	108.72	1000.5	1590	1117.32
Coating V	8.60	31.0	162.71	1495.9	837	1667.21

.

The appropriate spraying conditions ensured a balance of regulating factors, leading to more effective coatings for most engineering applications. However, existing porosities could lead to the deterioration of the coating surface, attributed to modifications in passive films caused by compositional variations at the pore margins and the formation of localized regions that accelerated localized corrosion processes. It was noted that the presence of porosity significantly reduced Ecorr values, thereby accelerating coating corrosion. Among the coatings, Coating V exhibited the best performance, despite having a porosity level of 4.1% vol., similar to the other coatings. All five coatings were considered optimized compared to the initial 32 runs conducted during the screening and design stages of this study. It was concluded that the low porosity levels contributed to the inhibition of micro-galvanic coupling at the coating/substrate interface, particularly in porous regions, microcracks, voids, or inclusions within the coating. These findings were consistent with previous studies.

The corrosion susceptibility ranking of the coatings, from highest to lowest, was as follows: Coating IV > Coating III > Coating I > Coating II > Coating V. This ranking aligned with the earlier discussion that coatings with higher porosity percentages were more susceptible to corrosion due to the presence of pores, which allowed the permeation of corrosive media to reach the substrate metal. All the optimized plasma-sprayed chromium oxide coatings demonstrated superior corrosion resistance, characterized by lower icorr and nobler Ecorr values compared to the substrate metal. The enhanced corrosion resistance of the optimized coatings was attributed to differences in porosity levels, i.e., the density and homogeneity of the coatings, resulting from the optimized spraying conditions.

EIS measurement was taken at OCP on all the coatings as well as substrate in a freely exposed 3.5 wt. % NaCl solution at room temperature to analyze the detailed film formation behavior and the effect of corrosive solution penetration. The electrical equivalent circuits adopted for the simulation of EIS plots are shown in Figure 15.

In the circuit, solution resistance is Rs, pore resistance is Rp originating from the hindrances to the ionic conduction paths in the coating, Rct is the resistance related to the charge transfer phenomena at the coating/substrate interface, Cc the coating capacitance and Cct is the double layer capacitance. The constant phase element (CPE) used in the circuit in place of pure capacitor was due to the inhomogeneous and porous nature of the coatings. The Nyquist and Bode plots are shown in Figure 15 and Figure 16, respectively. The presence of two consecutive semicircles in the Nyquist plots and the two inflection points on the corresponding Bode phase plots for all the optimized coatings suggests the associated time constants. The low frequency loop could be assigned to the corrosion reaction process while the high-frequency loop indicates the defectiveness in the coatings.

The corrosion parameters estimated with superior fitting quality with goodness of fit in the range of 10–3-10–4 using the equivalent circuits which are reported in Table 8. In this study, the solution resistance did not have any considerable impact on the corrosion processes hence its small variation could be neglected. As noted earlier from other literature the pore resistance and coating capacitance shows the degree of resistance provided by the coating for the corrosive solution to percolate through the coating layers. Thus, coating with more dense and uniformly packed structure had larger Rp value and lower Cc value. This is evident from fitting parameter Table 8. This is attributed to the better resistance to ionic conduction path to the corrosive solution permeating, as the percentage corrosion decreases from 6.3% vol. Coating IV to 4.1% vol. Coating V. It is also important to note that the chromium oxide nature provided the passivation film which was advantageous in inhibiting corrosion activities.

To better understand the extent of corrosion effects on the coatings, the corroded surfaces of all coatings were further investigated. SEM micrographs of the surface of the optimized chromium coatings after the corrosion test are presented in Figure 17. It was observed that the extent of cracking in the coatings increased with higher levels of porosity. Coating IV (Figure 17d) exhibited the least corrosion inhibition effectiveness, characterized by the highest Icorr and the least noble behavior among the optimized coatings, as well as the lowest corrosion resistance. The widespread microcracks on the corroded surfaces of Coating IV, and to a lesser extent on the other coatings, were attributed to factors preceding the access of the corrosive solution through the coating layer. These factors included (1) the compactness and adhesive bond of the coating, determined by the percentage porosity level, and (2) the crystallinity of the coating. Additionally, the cracking of the corroded surface was associated with corrosion byproducts/rust formed on the coating surfaces, which were analyzed using XRD after the potentiodynamic corrosion test. The peaks identified in all coatings were associated with eskolaite (Cr₂O₃) and traces of sodium chlorate (NaClO₄). In cases where localized corrosion was suspected, the presence of FeO-(OH) was detected in the form of rust. Since chromium oxide forms a passivation layer, corrosion activities were minimal except in areas where porosity posed a significant challenge. In the salt environment, a less stable phase of iron compound, particularly β-FeOOH, was formed, especially through pores, leading to pitting. This less stable phase was expected to form under saline conditions due to the equilibrium reached by the structures, facilitated by the presence of chloride ions from sodium chloride.

4. Conclusions

Using scanning electron microscope, XRD phase analysis characterization means electrochemical test on APS Cr₂O₃ coating; a series of experiments and theoretical studies were carried out to ascertain corrosion performance from these optimized coatings through the discussions outlined. The conclusions are as follows:

(1) The order of factors affecting formation of porosity in Cr₂O₃ coatings were Power > Primary Gas > Standoff Distance > carrier gas. The coatings produced from this arrangement had good metallurgical bonding with uniform morphology and structure with minimum cracks and pores.

(2) The presence of pores and cracks were reported to be highest for coating IV and least for coating V. (Coating I—5.6%, Coating II—4.8%, Coating III—5.0%, Coating IV—6.3%, and Coating V—4.1%). These were measured by using statistical methods with the help of ImageJ software. Coating V had optimal process parameters of: P(625 W); G(42.5 lpm); S(105 mm); C(5 lpm) and F(35 gpm) whereas coating V had P(625 A); G(30 lpm); S(105 mm); C(5 lpm) and F(35 gpm). These two coatings had almost similar process parameters except the gas flow rate. Higher gas flow rate helped with the transportation of molten particles necessitating superior coating with low porosity level.