Electrochemical Corrosion Resistance of Al₂O₃–YSZ Coatings on Steel Substrates

Abstract

Ceramic materials as coatings are known to have very good corrosion resistance properties compared to metallic or organic coatings, regardless of environmental conditions. The following samples were used for the experiments: an initial steel substrate and Al₂O₃ + YSZ (12.5%; 25% and 37.5% wt) atmospheric plasma spray-coated samples. The open circuit potential showed similar average values for all samples coated with ceramic layers, which were slightly higher than the potential of the original uncoated sample. The corrosion current densities (i_corr) of all plasma jet sputter-coated systems were very similar and significantly lower than those of the original material. Corrosion rates were much lower in the coated systems due to the chemical inertness of the ceramic coatings, particularly alumina- and zirconia-based coatings. It was observed that ceramic layers improve the corrosion resistance of the metallic material, especially at higher percentages of YSZ in the plasma spray-deposited complex layer. The porosity of the sputter-deposited layers reduced their corrosion resistance due to the contact between the electrolyte solution and the metal substrate created by the interconnection of the pores. The complex equivalent electrical circuit chosen for the analysis of the values led to results in accordance with the experimental parameters.

1. Introduction

Ceramic materials used as coatings are widely known to be superior to metallic or organic coatings for high-performance applications. Ceramic coatings have recently been widely used as thermal barriers for many applications of metallic materials [1].

From the studies presented in the literature, several results have been reported on YSZ- and alumina- based composite ceramic coatings. They were obtained by various methods (plasma jet sputtering [2], electrophoretic deposition [3], pressureless sintering [4], plasma slurry sintering [5], cathodic plasma electrodeposition [6], powder slurry sputtering [7], two-step procedure by Al coating of the YSZ layer and oxidation to Al₂O₃ [8] and sol-gel microwave infiltration [9]). Superior results were reported for their composite layers compared to alumina or YSZ layers.

In a study published in 2018, He et al. presented the results of a structural analysis and an investigation into the mechanical properties of ceramic composite layers formed by the combination of alumina and YSZ at varying weight ratios (wt%). The ratios of the plasma jet-sprayed ceramic composite layers were 10/90, 25/75, 40/60 and 55/45, yet the corrosion resistance properties in any electrolyte medium were not reported [10]. The mechanical properties of bimodal alumina and YSZ layers were also investigated by Ponnapureddy, who demonstrated their superiority over alumina layers [11].

While ceramic layers can be applied to a variety of applications, the majority of studies examining the corrosion resistance of composite ceramic layers have focused on CMAS (calcium-magnesium-aluminium-silicate) salts, particularly at high or very high temperatures relevant to their use as thermal barriers. Zhang et al. successfully produced TBC layers (20 wt% Al₂O₃–YSZ (7% wt Y₂O₃)) at a low cost and reported excellent corrosion resistance in CMAS and CMAS + NaVO₃ at high temperatures [12]. The superior corrosion resistance of alumina/YSZ coatings in CMAS was reported by Dou et al. for temperatures between 1250 and 1350 °C [13]. Furthermore, the corrosion behaviour of Al₂O₃/YSZ multi-layer coatings in Na₂SO₄–NaVO₃ salt media was evaluated by Mohammadi et al. The authors concluded that Al₂O₃ (crystallised in the cubic system) played a pivotal role in anti-corrosive protection [14]. Other experiments have demonstrated that alumina/YSZ ceramic coatings proposed for metal elements of gas turbines operating at high temperatures exhibit excellent corrosion resistance in 0.01 M K₃Fe(CN)₆/K₄Fe(CN)₆ aqueous solutions [15]. The formation of double layers of alumina and YSZ, also obtained by APS, has been demonstrated to enhance corrosion resistance in CMAS, even when subjected to prolonged exposure [1].

The corrosion behaviour of plasma spray-coated austenitic steel substrates with different ceramic layers was investigated by Kh Abd El-Aziz in a 0.5 M H₂SO₄ solution. Metal substrate samples with Al₂O₃, WC-12% Co and WC-12% Co/Al₂O₃ multilayer deposits were tested [16]. The corrosion resistance of the substrate is influenced by the type of surface coating material [8]. In contrast, oxide ceramic coatings exhibited fewer defects and superior corrosion resistance compared to carbides (WC-12% Co) [17].

Additionally, ceramic materials based on alumina or zirconia have demonstrated potential in the medical field [16]. The use of ZrO₂–CaO layers, comprising 95% ZrO₂ and 5% CaO, as a ceramic biomaterial has been demonstrated to facilitate osteoconductivity (the formation and growth of new bone material around the implant) and provide effective protection in biological electrolytic environments [18]. ZrO₂–CaO powder ceramic layers were manufactured via plasma jet spraying using a Sulzer Metco 9MCE spray rig. The electrochemical impedance experiments proved to be an effective method for studying the strength of ZrO₂–CaO layers deposited on Mg–Ca alloy during exposure in Ringer’s solution. The impedance results demonstrated notable differences that could be attributed to variations in immersion time, porosity and coating thickness. A thicker layer with enhanced compactness and density was observed to exhibit superior electrocorrosion resistance. The porous nature of the layers produced by spraying resulted in the formation of some connection paths between the corrosive environment and the base metal substrate over time [19].

In this study, the protective efficacy of ceramic coatings comprising alumina and zirconia stabilised with yttria (12.5, 25 and 37.5% YSZ powder, as reported in the study by He et al.) against a metal substrate (typically steel) in an acid rain electrolyte solution was investigated. No results have been reported on this topic.

2. Materials and Methods

The ceramic layers deposited on a metal substrate by plasma spraying are based on alumina and zirconia materials, which are considered corrosion-protective materials in comparison to conventional metallic materials [20]. The structural, phase, chemical and mechanical characteristics of these ceramic covering layers were previously presented in two studies by the main author [21,22]. An analysis of corrosion resistance was conducted on round working electrodes with a diameter of 10 mm. The samples with ceramic layers were wired and the metal surface was polished to facilitate contact with the working electrode. The tests were conducted at room temperature, specifically at 25 degrees Celsius. Three experiments were conducted for each sample, comprising the control and three samples with ceramic layers containing varying proportions of YSZ in alumina. The samples were classified as follows: Sample 1: metal substrate + Al₂O₃–ZrO₂ layer (12.5% wt), Sample 2: (metal substrate + Al₂O₃–ZrO₂ layer (25% wt), and Sample 3: metal substrate + Al₂O₃–ZrO₂ layer (37.5% wt)).

A corrosion cell containing three electrodes (a platinum electrode, a calomel-saturated electrode and the working electrode) was employed for both potentiodynamic measurements and electrochemical impedance spectroscopy determinations, as illustrated in Figure 1. The samples were positioned within the work cell using a Teflon washer with an inner diameter of 7 mm, ensuring that the surface area of the working electrode (the portion of the sample exposed to the corrosion environment) was consistent across all samples at 0.385 cm². All potential readings were taken in relation to the reference electrode. The solution employed was naturally aerated and comprised H₂SO₄ and HNO₃, with a pH level of 3, representative of an acid rain-type solution [23,24].

The following work sequence was employed as part of the measurement process:

-

Linear polarization for the Tafel method: potential range (−200) ÷ (+200) mV compared to the open circuit potential, potential scanning speed: dE/dt = 1 mV/s;

-

Cyclic polarization: potential range (−500) ÷ (+600) mV, potential scanning speed = 10 mV/s;

-

Electrochemical impedance spectroscopy measurements: working potential = open circuit potential, frequency range = 105 ÷ 0.1 Hz, current amplitude = 10 mV.

X-ray diffraction was performed with X’Pert Pro MPD equipment (Almelo, The Netherlands), Cu-X-ray tube, scan range of 20°–90°.

The surface condition of the experimental samples was analysed by scanning electron microscopy (SEM VegaTescan LMH II, SE detector, 30 kV heating gun, TESCAN, Brno-Kohoutovice, Czech Republic) and X-ray energy spectroscopy (EDS) using an EDS detector from Bruker (Xflash 6/10, automatic mode detection and mapping, Bruker, Billerica, MA, USA).

3. Experimental Results

The surface condition of the ceramic layers is shown in Figure 2a–c for all three cases of initially mixed powders. It can be seen that compact coatings were obtained, but also with characteristic pores from the atmospheric jet spraying technique. No significant differences were observed between the coatings. In the second case, when alumina-blended powders with 25% YSZ alumina were used for spraying, a slightly higher number of pores on the surface was observed, as shown in Figure 2b.

The chemical analysis of the three ceramic coatings provided a qualitative identification of the deposited coating components (Zr, Y, O and Hf), as well as the identification of a small percentage of Fe (Figure 3—energy spectrum). The identification of the Hf element is attributed to zirconium, which is naturally a companion element and is more difficult to completely remove from the YSZ. The porous nature of the plasma-sprayed ceramic coatings and their partial bonding are confirmed by the appearance of the Fe element on the coated surface. The presence of Fe on the surface may also be due to partial melting of the steel substrate surface during deposition and its mixing with the ceramic material. The chemical elements showed a homogeneous distribution with no YSZ or Al agglomerates. The presence of iron is only confirmed in two or three areas on the surface, as shown in Figure 3.

Quantitatively,

Table 1. Chemical composition of the ceramic coatings.

Elements/ Coating	Zr %		Al %		O %		Y %		Hf %		Fe %
	wt	at	wt	at	wt	at	wt	at	wt	at	wt	at
Sample 1	5.8	1.4	45.8	38.5	47.4	62.5	0.4	0.1	0.1	0.01	0.4	0.15
Sample 2	12.3	2.97	40.3	32.9	46.3	64.8	0.6	0.2	0.05	0.01	0.4	0.14
Sample 3	17.73	4.41	34.32	28.86	46.83	66.41	0.81	0.21	0.05	0.01	0.25	0.11
EDS Error%	1.2		1.5		2.15		0.2		0.01		0.01

St.dev.: Zr: ±0.8; Al: ±1.2; O: ±1.5; Y: ±0.1; Hf: ±0.01; Fe: ±0.1.

shows an increase in Zr and Y percentages along with an increase in the YSZ percentage mixed with alumina powders.

The values given in Table 1 were obtained as the average of three surface determinations. The standard deviations of the elements showed a relatively small variation in the chemical composition of the ceramic layer. Large variations in the chemical composition, as well as the presence of pores in the ceramic layers, can affect the characteristics of the layers obtained, primarily the different mechanical properties, but also the corrosion resistance as a result of a different behaviour in these areas when in contact with an electrolyte solution. Roughness profiles of samples after APS deposition confirm the presence and spread of pores in the composed ceramic coatings [20].

An analysis of the Fe percentage on the surface can provide an indication of the ceramic layers’ porosity and their ability to cover the metal substrate. Increasing the percentage of YSZ in the pre-powder mix reduced these uncovered areas by almost half the percentage of Fe identified on the surface.

The ICDD PDF4+ 2022 database was used for qualitative phase analysis. Qualitative phase analysis indicated the presence of the following polycrystalline phases: α-alumina (rhombohedral), gamma-alumina (cubic) and ZrO₂ (tetragonal). The main peaks identified are shown in Figure 4.

The main identified compounds are a stable rhombohedral phase of alumina and a cubic form of alumina (α- and γ-Al₂O₃), together with a tetragonal metastable phase of zirconia. A resonance in the occurrence of these phases is based on the extremely high solidification rate of the coatings (∼106 K/s), characteristic of atmospheric plasma sputtering treatment. The results are in agreement with other reported works [25,26]. Part of the α-Al₂O₃ phase transforms into a solid solution of Al atoms that transition to ZrO₂ during the atmospheric plasma sputtering process, with aggregation of some Al atoms at grain boundaries forming the amorphous phase [27,28].

The open circuit potential showed a similar average value for all samples coated with ceramic layers, which was slightly higher than the potential of the original uncoated sample. The corrosion current densities (i_corr) of all plasma jet sputter-coated systems are quite similar and significantly lower than those of the original material [29]. Corrosion rates are much lower for the coated systems due to the chemical inertness characteristic of ceramic coatings, particularly alumina- and zirconia-based coatings [30]. Sample 2 shows the highest corrosion rate among the ceramic-coated samples, but it is still lower, even by half, compared to the original material, as shown in

Table 2. Cyclic and linear potentiometry parameters recorded during the electro-corrosion resistance analysis process.

Material	Experimental Parameters Recorded During the Potentiometry Process
	OCP mV	E (I = 0) (mV)	icorr mA/cm2	vcorr mpy	βc (mV/dec)	βa (mV/dec)
Steel substrate	−416	−410.1	13.01	152.11	−135	318
Sample 1	−442	−448.3	1.51	17.69	−18	346
Sample 2	−445	−462.3	6.11	71.54	-	391
Sample 3	−450	−415	0.35	4.08	−109	−84

. The difference between the coated samples can be attributed to the appearance of a larger number of pores on sample 2, which contributes to the decrease in its corrosion resistance by forming communicating channels through which the electrolyte solution infiltrates until it comes into close contact with the substrate [31]. The much lower values of i_corr in comparison to v_corr indicate better protection of the samples with ceramic layers against the substrate.

A slight enhancement of the anodic reactions (βa > βc) is observed for the first sample, which is reversed to a lesser extent in sample 3 with ceramic coatings having a higher percentage of YSZ. It is recognised and demonstrated in the literature that ceramic coatings cannot eliminate the infiltration of electrolyte solutions up to the interface of the bonding layer between the metallic substrate and the ceramic layer formed over it. At the same time, and in confirmation, the anodic curves plotted in Figure 5a do not show a passivation plateau, presumably because oxide defects and inclusions in the adhesion layer prevent complete passivation and, as deduced from the very low corrosion potential, allow corrosion of the steel substrate. This suggests that the corrosion resistance of these coatings is not primarily based on passivation, as none of the coatings showed a clear active-passivation transition. Among the ceramic protective layers, Al₂O₃–37.5% YSZ appears to provide the best corrosion protection. This is attributed to the superior corrosion resistance of zirconia compared to metallic materials in general and to alumina in most cases.

The cyclic potentiometry in Figure 5b shows a generalised surface corrosion nature with a much higher current density for the uncoated sample, which is also confirmed by the intensely corroded surface of the metallic material, as shown in Figure 6a,b. Sample 2 shows a cyclic curve with a different appearance, highlighting the appearance of more intense pitting corrosion zones on the surface in the areas where the interconnected pores help the penetration of the electrolyte solution to the metal substrate, probably due to a higher number of pores in the structure [32]. Scanning electron microscopy, shown in Figure 6g, shows the pores in the ceramic layer and the formation of a cracked oxide layer as a result of surface corrosion on the ceramic material.

The best electro-corrosion behaviour was exhibited by sample 3 as a result of the higher percentage of YSZ in the ceramic layer structure. The surface condition of the materials is shown in Figure 6 at the macro scale (Figure 6a–d) through optical microscopy and at the micro scale (Figure 6e–h) through scanning electron microscopy.

A better fit of the data was achieved by replacing the double-layer capacitance by a constant-phase element, i.e., CPE, which expresses the non-ideal behaviour of the electrical double-layer capacitance (capacitance change with frequency). The equivalent circuit consists of a resistor (R_s) in series with a parallel combination of a constant-phase element (CPE), a resistor (R_ct), and a series combination of resistors and inductances (R_L, L) as shown in Figure 7. The parameter values that best match the experimental data are given in

Table 3. Parameters of the equivalent circuit for the experimental samples in the acid rain solution.

	Rs Ohm/cm2	CPE		Rct Ohm/cm2	RL Ohm/cm2	L H/cm2	CPE
		Q Ssn/cm2	n				Q Ssn/cm2	n
Initial	5.18	0.003451	0.76	1.68	0.56	0.1357
Proba 1	31.93	0.0005059	0.645	59	-	1042	0.030	1
Proba 2	11.87	0.0001783	0.57	28	-	2.83	0.01	0.8
Proba 3	57.33	0.9390	-	50	737	1850

.

Impedance spectra measured by spectroscopy can be displayed as complex plane plots, also known as Nyquist plots and/or Bode plots. The Nyquist plot shows the impedance data as a complex variable separated into real, Z_re, and imaginary, Z_im, parts, expressed in Ω cm². The Bode plot instead shows the frequency dependence of the absolute magnitudes of the impedance modulus, |Z|, and the phase angle, Φ. The advantages of this procedure are that data for all measured frequencies are displayed and a wide range of impedance values can be displayed simultaneously. The frequency dependence of log|Z| and Φ in Nyquist and Bode impedance plots obtained for samples coated at different exposures in the test electrolyte is shown in Figure 8. Selected examples of the Nyquist and Bode impedance plots obtained for the first coated samples at both exposures in the testing electrolyte solution are shown in Figure 8.

The Nyquist plot of the base material, as shown in Figure 8a, without the ceramic coating, shows a capacitive arc at high and intermediate frequencies, followed by an inductive arc at low frequencies for room temperature immersion in the electrolyte solution. An inductive arc is usually associated with metal dissolution processes, in this case, substrate rusting. The plot sizes and Nyquist curves are much larger for the samples coated with ceramic layers (especially samples 1 and 3), as shown in Figure 8c,e,g, which represents a significant increase in the corrosion resistance of this system and confirms the results of the linear potentiometer tests.

The Bode spectrum obtained at open circuit potential for the initial uncoated sample in rainwater solution is shown in Figure 8b. Lower impedance values were measured, and the inductive behaviour of the material was observed in the low frequency range. The Bode plot of frequency versus impedance shows that the impedance values decrease slowly, indicating the dissolution of the corrosion products in the solution. The slow but continuous decrease in impedance indicated an intensification of the corrosion of the initially uncoated steel. Analysis of the Bode spectra in terms of the equivalent circuit (EC) showed the values of the impedance parameters. The EIS spectrum obtained for the initial uncoated material matches very closely with the circuit shown in Figure 7. The experimental values showed good agreement with an error of less than 5%. The results obtained are shown in Table 3. The electrolyte solution continuously passes through the porous Al₂O₃ ceramic layers with YSZ to react with the steel substrate. Over time, the ceramic layers act as an effective cathode for hydrogen release. Since R_L is inversely proportional to the corrosion rate, an increase in R_L, as shown in Table 3, indicates that the corrosion resistance of the ceramic coatings on the metal substrate has become stronger.

Figure 9 shows the main elemental spectra identified on the corroded samples. On the surface of the initial sample, the formation of corrosion compounds based on oxides, hydroxides, carbonates, nitrates or nitrides and sulphides was observed structurally—as shown in Figure 6—and chemically—as shown in Figure 6 and Figure 9—as a result of the metal interacting with the electrolytic solution.

On the samples coated with ceramic layers, i.e., samples 1–3 in

Table 4. Chemical composition of the surface of the experimental samples (initial and samples 1–3) after the electro-corrosion resistance test (StDev. O: ±2.5; Al: ±1.5; Zr: ±0.5; Fe: ±0.1; Y: ±0.1; N: ±0.1; S: ±0.1 and C: ±2.2).

Sample	Fe% wt     at		C% wt     at		O% wt     at		N% wt     at		S% wt     at		Al% wt     at		Zr% wt     at		Y% wt     at		Others wt
Initial	80.79	57.25	5.37	17.7	8.44	20.87	0.21	0.59	0.44	0.54	--	--	--	--	--	--	Cu: 3.74; Mn:1.02
Sample 1	0.8	0.3	-	-	47.6	64.4	0.5	0.8	-	-	39.6	31.7	10.76	2.6	0.8	0.2	-
Sample 2	2.2	0.9	-	-	47.1	62.7	0.2	0.2	-	-	43.9	34.6	5.9	1.4	0.7	0.2	-
Sample 3	0.95	0.4	-	-	47.8	62.8	0.3	0.4	-	-	44.9	35	5.3	1.2	0.7	0.2	-
EDS Er.	2.1		2.5		3		0.2		0.1		1.1		0.25		0.05		0.25

, the same elements were identified on the surface, except for sulphur, which was not observed on the ceramic surfaces, with a very high percentage of oxygen, mainly due to the ceramic layer and, to a lesser extent, to compounds formed on the surface, especially for sample 2, as shown in Figure 10c, on which several iron-based compounds can be observed. For the samples with ceramic coatings, there is a decrease in the Fe percentage identified on the surface after the electro-corrosion test and an increase in the YSZ percentage in the deposited ceramic layer. Figure 10 also shows that the surface of sample 3 is less corroded than that of the other two samples with deposits. The identification of a lower percentage of Fe on sample 3′s surface may be due to its lower porosity and also explains the better corrosion resistance behaviour of this sample.

At low pH, the corrosion mechanism is dependent not only on the hydrogen ion concentration but also on the ions present (SO₄²⁻, NO₃⁻). Figure 10a shows the presence of a uniform layer of iron oxides in the original sample, as well as the presence of isolated compounds based on sulphur or nitrogen that have migrated from the acid rain solution to the surface of the steel. The samples with deposits, shown in Figure 10b–d, do not show any corrosion layer on the surface, but the presence of some compounds can be observed only in certain areas. The ceramic layer protects the metal surface; it generally interacts very little with the electrolyte solution, and the detection of compounds other than those in the ceramic layer may be attributed to their remaining in the pores of the ceramic layer [15,33,34,35].

The composition of the bond coat is very important in the corrosion mechanism. The ions from the electrolyte solution infiltrate the porous microstructures of the Al₂O₃–ZrO₂ layer_. The main reactions are as follows:

Al − 3e⁻→Al³⁺

(1)

O² + 2H₂O + 4e⁻→4OH⁻

(2)

Al³⁺ + 3OH⁻→Al(OH)₃

(3)

Given the low corrosion rate, the oxidation rate of the bonding coating is low at room temperature.

The detection of carbon can be related to the SEM chamber contamination and can be considered as EDS error.

For sample 2, in Figure 10c, a more affected surface of the ceramic layer is observed, accompanied by the presence of macropores and iron compounds on the ceramic surface. This is a result of the electrolyte solution coming into contact with the metal substrate through the interconnected pores of the ceramic layer. To avoid the interconnectivity of the pores in the ceramic layer and to cover 100% of the substrate surfaces, the deposition technique can be adapted [36,37,38]. The elemental distribution results, as shown in Figure 10, confirm the results obtained by linear and cyclic potentiometry.

4. Conclusions

After analysing the experimental results obtained, the following observations were made:

-

The percentages of alumina and zirconia stabilised with yttria influence the porosity of the produced ceramic layers;

-

The interconnection of the pores leads to the surface of the metallic substrate being in contact with the electrolyte solution, a fact that leads to the formation of corrosion compounds based on hydroxides and ferrous oxides, reaching the surface of the ceramic layer;

-

Ceramic coatings improve the corrosion resistance of the metallic material, especially at higher percentages of YSZ in the complex layer deposited by plasma spraying;

-

The increased percentage of YSZ in the composite layers increases the resistance to electrochemical corrosion in acidic solutions such as acid rain, with the Al₂O₃–37.5% YSZ sample having an i_corr thirty-seven times lower than the corrosion current of the metal substrate.