Introduction

The resistance of FeCrAl alloys to high-temperature oxidation depends on the properties of the building up aluminum oxide. Good protective properties of α-Al2O3 (α-alumina) scale stem from its low volatility [1, 2], slow growth rate [2], low concentration of crystalline defects [3], and good chemical stability in high-temperature corrosion environments [4]. But, the scale may crack and spall in the course of cyclic thermal oxidation [5–7].

The protective character of α-Al2O3 scale and its adherence to the substrate can be improved through the use of reactive elements (RE), e.g., Zr, Hf, Y, Ce, La, and other rare earth elements) [1, 8–13]. The required RE concentration in the scale depends on several factors such as the chemical composition of the alloy, the amount and kind of impurities (e.g., sulfur) the latter contains, the kind of reactive element and the oxidation conditions (temperature, time, environment). The optimum RE content has been the subject of intensive research conducted in many laboratories [14–22]. According to some authors, the scale acquires the best protective properties in the presence of 10−4–10−1 wt% of RE [23]. Others suggest that the optimum RE content is below 0.05 at.% (about 0.3 wt%) [24]. The spallation resistance of the scale is observed to increase already at a Ce content of 0.05 wt% [25]. An excess of Ce may result in the acceleration of the alloy oxidation rate and in the increase in the scale susceptibility to spalling [25–27].

Reactive elements may be introduced into the alloy as alloying elements [16, 17, 19] or in the form of oxides [18, 20, 21]. Also ion implantation is employed for this purpose [28]. Sometimes reactive elements are deposited on the alloy by surface methods, such as sol–gel coating [29–31], MOCVD [32] or electrodeposition [33, 34].

The oxides of reactive elements, e.g., CeO2, affect the oxidation kinetics of metals [35]. According to the literature reports, when Ce is introduced, the rate of the high-temperature oxidation of alloys considerably decreases [36]. The presence of cerium in the coating and in the scale contributes to the outward diffusion of cations and to the inward anion transport during oxidation [35–37], eliminates voids in the region near the oxide-substrate interface and improves the adhesion of the scale to the substrate.

Research results indicate that in a temperature range of 1073–1273 K reactive elements affect the phase transformation of transient aluminum oxides of the γ, δ, and θ type to the stable α-Al2O3 phase [38].

The heat resistance of the FeCrAl alloy can be improved by producing vitreous coatings made up of SiO2 and Al2O3 on its surface. As a result of the outward diffusion of Al3+ cations and the inward diffusion of oxygen, aluminosilicates (e.g., mullite) form within the coating in the course of high-temperature oxidation [30, 31]. By introducing a reactive element (e.g., Ce) into such a coating one can further increase the heat resistance of the FeCrAl alloy.

FeCrAl-type alloys are designed for operation at temperatures as high as 1400 °C. Sometimes, they are tested and used at temperatures below 1000 °C (for instance as catalytic carriers in cars). In these conditions, α-Al2O3 scale forms much slower and its build-up and protective properties are complicated by the presence of transition aluminas [38].

The aim of the present contribution was to determine the high-temperature resistance of the FeCrAl alloy coated with SiO2–Al2O3 or SiO2–Al2O3–CeO2 (900 °C, under air). The coatings would be obtained by the sol–gel method and deposited using the dip-coating technique. The effect of the cerium content in the coatings on the rate of oxidation of the FeCrAl alloy and on the morphology of the building up Al2O3 scale was determined. Also the thermal behavior of coating materials was described, with special regard to the temperature range in which crystalline phases form and to the effect of powder sintering temperature on the material microstructure. The changes induced by a cerium addition to the aluminosilicate coatings and the contribution of the addition to an improvement in the high-temperature resistance of FeCrAl alloy were studied.

Experimental

Sol preparation

The precursors of: silicon (tetraethyl orthosilicate, Si(OC2H5)4, TEOS), aluminum (aluminum tri-sec-butoxide, Al(OC4H9)3, TBA), and cerium (cerium(III) 2,4-pentanedionate, Ce(C5H7O2)3·xH2O, Ce-acac) were used to prepare SiO2–Al2O3 and SiO2–Al2O3–CeO2 coating solutions. Each of the precursors would be separately diluted in butyl alcohol (1-butanol, Bu-1) and homogenized in an ultrasonic washer for 90 min. A mixture of TEOS and TBA solutions was the base for the SiO2–Al2O3 (sol A) coating solution. Ternary sols SiO2–Al2O3–CeO2 with a different cerium content (sols B, C, D) were obtained by adding cerium acetylacetonate (III) to sol A. Then, HNO3 (65 %) and CH3COOH (95.5 %) were added to the two kinds (binary and ternary) of sols and homogenized for 120 min. The coating solutions prepared in this way would be aged for 168 h. The composition of the sols is shown in Table 1.

Table 1 Chemical composition of sols
Full size table

In order to characterize the behavior of the material at high temperature, the A–D sols were gelled in air and the obtained powders were subjected to further thermal treatment.

Samples

The coating substrate material was a ferritic FeCrAl-type alloy (Baildon-Steelplant, Poland) with a chemical composition (wt%): Cr, 18.5; Al, 4.6; Si, 0.81; Mn, 0.6; Ti, 0.01; C, max. 0.05; S, max. 0.07; and Fe as the matrix. The samples were made of 0.06-mm thick foil.

Prior to coating deposition the specimens were ground with abrasive paper (grit no. 400, 600, and 800), degreased in acetone and in distilled water in an ultrasonic bath and dried in a stream of hot air.

SiO2–Al2O3 and SiO2–Al2O3–CeO2 would be deposited on the FeCrAl alloy surface using the dip-coating technique. The sample would be brought to the surface of the solution at a rate of 10 mm min−1. The alloy with the coating would be dried in air for 24 h. Then, the samples would be subjected to preliminary heat treatment under air in a furnace at a temperature of 500 °C for 3 h. The rate of temperature increase in furnace temperature was 2 °C min−1.

Test methodology:high-temperature oxidation

The oxidation of the FeCrAl alloy without and with a SiO2–Al2O3 or SiO2–Al2O3–CeO2 coating was conducted under air at a temperature of 900 °C for 100 h. The alloy resistance to high-temperature oxidation Eq. (1) and the coating’s protection effectiveness (S) in comparison with that of the uncoated FeCrAl alloy substrate Eq. (2) were determined on the basis of the relative changes in the sample mass versus time:

$$ \frac{\Updelta m}{{m_{0} }} = \frac{{m_{\text{k}} - m_{0} }}{{m_{0} }} \times 100\;\% \, $$
(1)
$$ S = \frac{{\Updelta m_{0} - \Updelta m_{\text{p}} }}{{\Updelta m_{0} }} \times 100\;\% \, $$
(2)

where m 0, the sample initial mass; m k, the sample mass after oxidation; Δm 0, the relative gain in the mass of the uncoated FeCrAl alloy; Δm k, the relative gain in the mass of the coated FeCrAl alloy.

The topography of the SiO2–Al2O3 and SiO2–Al2O3–CeO2 coatings was examined by scanning electron microscopy (JSM 5800LV made by Joel).

Characterization of powder coating materials

Phase analyses of powder gels A–D were carried out using the XRD technique and a Siemens D5000 diffractometer equipped with a copper radiation cathode. The measurements were performed in a 2θ angle range of 5–50° with a step of 0.04° for at least 3 s per step.

The thermal behavior of the dried gels was determined by combined DSC and TG. The analyses were carried out using the SETSYSTM apparatus (TG-DSC 1500; Setaram). The powdered samples, each weighing 20 mg, were placed in Pt crucibles and heated up to 1200 °C at a heating rate of 10 °C min−1 using Ar as the purge gas. The temperature and sensitivity calibration factor for the thermal experiments was determined at the melting points of NaCl (801 °C), Ca2P2O7 (1353 °C) and K2SO4 (1070 °C) and for the phase transition temperature of the latter compound (583 °C).

Results and discussion

Coating solutions

The obtained SiO2–Al2O3 and SiO2–Al2O3–CeO2 sols were transparent and monophase. They showed low and constant viscosity (about 4–5 cP; 4–5·10−3 Pa s) during no-air access storage for many weeks. The Ce(IV) content affected the sol color. As the Ce(IV) content in the mixture increased, the yellow color of the solution intensified. The quick gelation of the sols on the alloy surface, and so the formation of SiO2–Al2O3 and SiO2–Al2O3–CeO2 coatings, was the result of the hydrolysis of the precursors with moisture (water) contained in the air [30, 31, 39, 40].

Morphology of FeCrAl alloy surface after high-temperature oxidation

SEM examinations show a close similarity between the topographies of all the FeCrAl samples after high-temperature oxidation at 900 °C for 100 h (Fig. 1). This applies to both the alloy alone and the alloy covered with two- and three-component oxide coatings. The scale is cracked to a different degree. The fewest cracks occur on the FeCrAl alloy while the most cracks occur on the alloy covered with the ternary SiO2–Al2O3–CeO2 coating. Also the largest number of segregations occur on the alloy with the latter coating. From an analysis of the literature on the subject, it appears that this may be the θ-Al2O3 phase [1, 41–48], and the previous studies [30, 31, 49] by the authors indicate the γ-Al2O3 form.

Fig. 1
figure 1

Surface of FeCrAl alloy after high-temperature oxidation at 900 °C for 100 h without coating (a); with SiO2–Al2O3 layer (b); with SiO2–Al2O3–CeO2 coatings obtained from sol B (c); and with SiO2–Al2O3–CeO2 coatings obtained from sol D (d)

Full size image

Oxidation of FeCrAl alloy and alloy with SiO2–Al2O3 or SiO2–Al2O3–CeO2 coating at 900 °C

The samples made of FeCrAl alloy foil without and with a SiO2–Al2O3 or SiO2–Al2O3–CeO2 coating do not deform during isothermal high-temperature oxidation. This indicates that the stresses arising at the substrate-scale-coating interfaces are not too high.

In order to determine the behavior of the FeCrAl alloy and that of the latter alloy with sol–gel coatings during oxidation at 900 °C the change in mass as function of time was measured (Fig. 2).

Fig. 2
figure 2

Relative change in mass of uncoated FeCrAl alloy and alloy covered with SiO2–Al2O3 or SiO2–Al2O3–CeO2 coating after oxidation under air at 900 °C

Full size image

The samples gain in mass is in accordance with the parabolic oxidation rate. The deposition of the vitreous coatings on the alloy results in a reduction in the rate of oxidation of the substrate. It can be assumed that the coatings act as barriers.

The uncoated alloy showed the largest mass increment and so the lowest resistance to oxidation. The SiO2–Al2O3 sol–gel coating clearly improves the behavior of the samples. In the case of the three-component coatings with CeO2, the mass increment depends on the cerium content. The presence of a large amount of cerium (sol B, Si:Al:Ce = 3:1:0.1) results in a deterioration in oxidation resistance in comparison with the alloy covered with the SiO2–Al2O3 coating. This has also been observed for other reactive elements by other researchers [1–4, 7–17, 19, 20, 25–29]. A small amount of cerium in the coating (sols C and D, Si:Al:Ce = 3:1:0.01 and 3:1:0.001) reduces the rate of oxidation of the alloy, also in comparison with the sample covered with the SiO2–Al2O3 (sol A) coating.

The protection effectiveness of the coatings during high-temperature oxidation ranges from 60 to 74 % (Table 2) and remains at a constant level over time (100 h), which indicates their high stability.

Table 2 Protection effectiveness of coatings in comparison with uncoated FeCrAl substrate
Full size table

Characterization of coating materials

The phase composition of the gelled and sintered coating materials was determined by the XRD method. The materials are rich in silica (86 mol% SiO2; 14 mol% Al2O3) and their composition corresponds to the spinodal region in the binary SiO2–Al2O3 system in a wide temperature range (~700–1250 °C) [50]. The XRD patterns of the samples subjected to thermal treatment at 800 °C (5 h), 900 °C (100 h), and 1100 °C (10 h) are shown in Fig. 3. An analysis of the XRD patterns of the sols dried at room temperature and then annealed at 500° for 3 h showed their noncrystalline character. Also the samples with a low Ce content (sols C and D) had an amorphous character. Whereas the thermal treatment of sample B resulted in the crystallization of cerium oxide. The presence of crystalline CeO2 in samples C and D was not confirmed by the XRD method—probably due to the low cerium oxide content in the gels (the amount undetectable by XRD method).

Fig. 3
figure 3

The XRD patterns of coating materials sintered at different temperatures

Full size image

The powdered gels were analyzed by the DSC/TG methods. As an example, Fig. 4 shows the heating curve for gel C. The shape of the DSC and TG curves for the other samples is similar. In a temperature range up to about 550° a single-step decrement (amounting to 28–30 wt%.) in the mass of the preparation was registered on the TG curves for the samples. This decrement was accompanied by an endothermic thermal effect on the DSC curves, with a maximum at a temperature of 500, 460, and 420 °C for, respectively, sol B, C, and D. The mass decrement registered in this range of temperatures can be ascribed to the dehydration of the gel and to the decomposition of its components (alcoholates).

Fig. 4
figure 4

The DSC/TG curves of heating of gel C

Full size image

At a temperature of 1060 °C, a weak exothermic effect was observed on the DSC curves. Unfortunately, an analysis of the XRD samples sintered at 1100 °C (Fig. 3) did not confirm the presence of a new crystalline phase, probably due to the XRD method low sensitivity in detecting small amounts of phases. However, the available literature data indicate that the registered exothermic effect can be ascribed to the crystallization of the original mullite or crystobalite [51–54].

Morphology of powders obtained from A to D sols after sintering at 900 °C

The morphology of the powders obtained from sols A to D after sintering at 900 °C for 100 h varies (Fig. 5). The SiO2–Al2O3 grains obtained from sol A have an irregular shape, a smooth surface (Fig. 5a and characteristics of the vitrified phase. As the cerium oxide content in the material increases, the powder grains become smaller and agglomerate to a considerable degree and the percentage of the fraction characterized by smooth surface decreases (Fig. 5b–d). At the same time, crystalline cerium oxide was found to be present only in the preparation obtained from sol B (with the highest CeO2 content) (Fig. 3). The other powders (gels A, C and D) had an amorphous character. The authors’ earlier studies [50] on the effect of the thermal treatment of aluminosilicate gels (with different Si:Al molar ratios) on their morphology showed that an increase in the amount of the crystalline fraction in the preparations results in an increase in the specific surface area of the powders, and thus in an increase in their porosity. These are undesirable effects if such materials are to be used for coatings which are to protect alloys against high-temperature corrosion.

Fig. 5
figure 5

Morphology of SiO2–Al2O3 and SiO2–Al2O3–CeO2 gels after heat treatment at 900 °C for 100 h obtained from sol A (a), sol B (b), sol C (c), and sol D (d)

Full size image

Conclusions

On the basis of the results of the studies one can conclude that

  1. (1)

    Single-layer protective SiO2–Al2O3 or SiO2–Al2O3–CeO2 coatings obtained by the sol–gel method and deposited on the FeCrAl alloy using the dip-coating technique improve the substrate resistance to high-temperature oxidation. The presence of ternary coatings on the metallic substrate results in different gains in the mass of the alloy depending on the cerium oxide content. Because of the higher cerium content in the coating obtained from sol B (Si:Al:Ce = 3:1:0.1), the sample resistance is worse than that of the alloy with coating A (SiO2–Al2O3) and with coatings C–D. As the amount of cerium in the coating decreases (sols C–D with Si:Al:Ce molar ratios of respectively 3:1:0.01 and 3:1:0.001), the rate of oxidation of the alloy relative to that of the alloy covered with the SiO2–Al2O3 coating also decreases.

  2. (2)

    The samples gain in mass is in accordance with the parabolic oxidation rate. The fact that the samples (in the form of thin foils) do not undergo deformation when subjected to high-temperature oxidation indicates that the stresses arising at the substrate-scale-coating interfaces are not too high.

  3. (3)

    The protection effectiveness of the coatings ranges from 60 to 74 % in the course of 100 h of high-temperature oxidation.

  4. (4)

    During high-temperature oxidation at 900 °C characteristic segregations form on the surface of the FeCrAl alloy. The largest number of such segregations occur on the alloy covered with the three-component SiO2–Al2O3–CeO2 coating. The segregations may constitute a θ-Al2O3 or γ-Al2O3 phase.

  5. (5)

    The fraction of a crystalline phase (CeO2) in the powders obtained from the sols through sintering at 900 °C affects the morphology of the powders. As the amount of cerium oxide in the material increases, the powder grain size decreases considerably whereby the material specific surface increases. The increase in the surface development of the powders may be responsible for the worse protective properties of the coatings produced from the three-component sols with the highest CeO2 content.