Regular Article
Characterization of α-Al2O3 in Structural Isomers of Alumina Formed by Oxidation of Fe–Cr–Al Alloys
2022 年 62 巻 9 号 p. 1881-1886
詳細
2022 年 62 巻 9 号 p. 1881-1886
To understand the formation of structural isomers of alumina (Al2O3) on the surfaces of Fe–Cr–Al alloys, cathodoluminescence (CL), X-ray photoelectron spectroscopy (XPS), and X-ray diffraction (XRD) were used. The alloys were annealed under oxygen at a low partial pressure to focus on the initial formation of Al2O3. XRD revealed that corundum (α-Al2O3) was predominantly formed on the alloy surfaces after annealing at approximately 1100°C, whereas γ-Al2O3 and θ-Al2O3 were formed at 1000°C. CL peaks around 700 nm, which originated from Cr3+ ions in Al2O3, was drastically increases by rising the annealing temperatures from 1000 to 1100°C because of the increase of the chromium content in the Al2O3 scales that was revealed by XPS. EPMA revealed that the surficial α-Al2O3 had striped patterns, which were similar to those observed using CL imaging, indicating that CL can be conveniently used for imaging α-Al2O3 and analyzing the inhomogeneous distribution of α-Al2O3 on the alloys.
For the enhancement of the corrosion resistance of heat-resistant alloys at high temperatures, Al2O3 (alumina or aluminum oxide) scales are formed on alloy surfaces.1) These Al2O3 scales, which are notably formed on nickel-aluminum intermetallic compounds and nickel-based heat-resistant aluminum, act as corrosion-protective layers.2,3,4,5,6,7,8,9,10,11) Al2O3 exhibits several structural isomers at different temperatures.6,10) Some Al2O3 isomers, such as γ-Al2O3 and θ-Al2O3, show less resistance to high temperature corrosion owing to their high reactivity and unstable structure, whereas α-Al2O3 show high corrosion resistance owing to its high structural stability. Thus, characterizing the structural isomers of Al2O3 scales on alloy surfaces is vital. Furthermore, investigating the inhomogeneity that is associated with the uneven thickness and surficial distribution of the Al2O3 scales with different structures is required because inhomogeneous scales promotes corrosion and degradation of the heat-resistant alloys.
The formation of Al2O3 in iron-based alloys and heat-resistant steels has also been investigated.10,11,12,13,14,15,16,17,18,19) It is often necessary to understand and analyze the structural changes of Al2O3 on the surfaces of iron-based alloys at high temperatures because of the practical applications of these alloys. For example, X-ray diffraction, which is practically applied to on-line analysis of galvannealed coating layers,20) would be applicable to the on-site identification of a specific structural isomer of Al2O3.6,21) Additionally, cathodoluminescence (CL) and X-ray excited optical luminescence (XEOL) methods have been developed to evaluate oxide scales, and their application for on-site analysis has been proposed.17,36) These set-ups are compact, compared with X-ray diffractometer. In that study, the samples were subjected to different heat treatments, and the results demonstrated that the combination of CL and scanning electron microscope-energy dispersive X-ray (SEM-EDX) methods was effective for characterizing Al2O3 scales formed on the surface of heat-resistant steel.
A previous study demonstrated that the oxidation of a less-noble metal (such as aluminum) at high temperatures often results in the formation of a non-conductive thin layer on alloy surfaces under controlled annealing.19) In this study, X-ray photoelectron spectroscopy (XPS) and ellipsometry were used to investigate the growth of Al2O3 thin layers on Fe-based alloys. The results show that ~100 nm thick Al2O3 layers are formed on the surface of the samples after annealing at high temperatures and under oxygen at low partial pressure.19) Non-destructive ellipsometry is used to determine a more representative thickness of Al2O3, and destructive XPS depth profiles are used to provide artifactual information on the thickness of Al2O3. Non-destructive ellipsometry is generally applicable for characterizing thin Al2O3 films with a smooth homogeneous surface.22)
Based on the aforementioned studies, it is important to combine different analytical methods to understand the formation of oxides on the surfaces of iron-based alloys. The objective of this study is to characterize Al2O3 scales formed on the surfaces of practical Fe–Cr–Al alloys, which are annealed under gaseous oxygen at low partial pressures. Annealing is considered for fabricating specific Al2O3 structural isomers on the surface of Fe–Cr–Al alloys. XPS is used to analyze the surface concentration of Al2O3. CL and XRD are used to determine the specific structure of Al2O3. CL mapping and EPMA are used to visualize the distribution of Al2O3.
Two sheets of commercially-available ferritic stainless steel (samples A and B), namely, Fe-15%Cr-5Al and Fe-18%-2%Al alloys, were used as samples. The detailed chemical compositions of the alloys are listed in Table 1. The alloy sheets (ca. 1 mm thickness) were annealed at 800–1100°C for 0.5 h under pure hydrogen gas. Gas control equipment (hydrogen sensor and an oxygen pump-sensor) was used to maintain a low partial pressure of gaseous oxygen (mainly less than ~10−13 Pa), which was required to form Al2O3 scales on the alloy sheets.7) The surfaces of the annealed samples were initially blue and changed to dark brown, due to interference from the scales on the substrate.
| C | Si | Mn | P | S | Cr | Al | |
|---|---|---|---|---|---|---|---|
| Sample A | 0.005 | 0.21 | 0.12 | 0.024 | 0.000 | 14.95 | 4.73 |
| Sample B | 0.004 | 0.42 | 0.23 | 0.028 | 0.001 | 18.09 | 2.00 |
A customized SEM-CL system was used to obtain the CL spectra and images of the samples. Details of the SEM-CL have been previously reported.11,17,23,24,25,26) A spectrometer (QE65Pro, Ocean Optics Inc., Largo, Florida, USA) attached through an optical fiber to the SEM equipment (Mighty-8DXL, TECHNEX, Tokyo, Japan) was used to acquire the CL spectra. A parabolic mirror was used to collimate the light emitted from the samples, and plano-convex lenses that were attached to the tip of the optical fiber were used to collect the collimated light. CL images were captured using a digital mirrorless camera equipped with a zoom lens (LZH-10A-05T, Seimitu Wave Inc., Kyoto, Japan) via a quartz viewport attached to a SEM (Mighty-8DXL, TECHNEX, Tokyo, Japan). The detectable wavelength range of the camera was from 350 to 1000 nm. This detectable wavelength range was achieved by detaching a built-in filter that blocked ultraviolet and infrared light from a commercially available digital mirrorless camera (α7RII, Sony Corp., Tokyo, Japan).
XPS was used to analyze the chemical composition of the alloy surfaces. A Kratos AXIS-Ultra DLD instrument was used to obtain the XPS data; for the analysis, the incident X-rays were monochromated Al-Kα, with a primary X-ray beam size of approximately 800 μm. Argon ion sputtering was used to remove the contaminated layers on the surfaces.
XRD was used for structural identification of the surficial oxides on the samples that were annealed under gaseous oxygen at low partial pressure. For the XRD analysis, RIGAKU-RINT equipment with a Cu-Kα target was used. EPMA measurements were used to determine the constituent elements on the alloy surfaces. JEOL JXA-8530F equipment with a thermal type field-emission electron gun was used for EPMA analysis. The voltage of the primary electron beam was set to 15 kV. Although a limited number of elements was quantified, the ZAF (Z: atomic number; A: absorption; F: fluorescence excitation) correction method was used to quantitatively analyze the target elements.
XPS was used to determine the elemental compositions of the smooth alloy surfaces after argon ion sputtering of the surfaces for ~30 s to remove carbon contaminations.19) Even after the ion sputtering, some carbon still remained on the alloy surfaces owing to the surficial roughness, which was introduced by surface-finishing. Nevertheless, the reasonable surface compositions of the major elements on the surfaces of sample A and sample B after annealing at 1000°C and 1100°C were determined from the intensity of the Fe 2p, Cr 2p, Al 2p, and O 1s peaks by applying the corresponding sensitivity factors, as shown in Table 2. Aluminum and oxygen were the main elemental components of the alloy surfaces, suggesting the formation of Al2O3 on the alloy surfaces during the annealing.
| Temperature (°C) | Fe | Cr | Al | O | |
|---|---|---|---|---|---|
| Sample A | 1000 | 1.2 | 0.2 | 36.5 | 62.1 |
| 1100 | 1.3 | 1.3 | 33.0 | 64.4 | |
| Sample B | 1000 | 1.4 | 0.2 | 35.3 | 63.1 |
| 1100 | 1.4 | 0.6 | 34.8 | 63.2 |
Al2O3 exhibits several types of structural isomers, such as α-Al2O3, γ-Al2O3, λ-Al2O3, θ-Al2O3, at different annealing temperatures. We acquired CL spectra of sample surfaces annealed at different temperatures (800–1100°C) to investigate the availability of the CL spectra for identifying the structural isomers of Al2O3. The intensities and shapes of the CL spectra was drastically varied by changing the annealing temperature from 1000 to 1100°C for sample A and B (Fig. 1). In particular, the intensity of the peak around 695 nm drastically changed. This indicaties that structural transformation of Al2O3 occurred at a temperature between 1000 to 1100°C. The shape of the CL spectra of samples annealed at 1100°C was similar to corundum (α-Al2O3),17,27,28,29,30) whereas the shape of the CL spectra of samples annealed at 1000°C was similar to γ-Al2O3.31) An intense luminescence peak was observed at 695 nm for α-Al2O3, and was attributed to the radiative transition of Cr3+ ions in the octahedral crystal field of Al2O3.30) Peaks at 670 and 715 nm originated from side bands around the peak at 695 nm. Peaks around 775 nm were attributed to Fe3+ ions in the octahedral crystal field of Al2O3.29) Small peaks at 335 and 535 nm were related to intrinsic defects (e.g., oxygen vacancy).29,32) As for γ-Al2O3, peaks around 700 nm and a broad peak abound 740 nm were attributed to Cr3+ and Fe3+ ions in the octahedral crystal field of Al2O3, respectively.31) A peak at 535 nm was related to intrinsic defects.29) The peaks around 700 nm for α-Al2O3 and γ-Al2O3 commonly originated from Cr3+ ions. The chromium contents of sample surfaces increased 3–6 times by rising the annealing temperature from 1000 to 1100°C (Table 2), which is consistent with a previous report that the chromium content in α-Al2O3 was higher than that in γ-Al2O3.31) Therefore, the drastic increase of the peak intensity around 700 nm is attributed to the increase of chromium contents in Al2O3 scale because the CL intensity generally increases with increasing the activator contents until about 1%.33) We can distinguish α-Al2O3 from γ-Al2O3 by examining shapes of the CL spectra: α-Al2O3 exhibit the intense sharp peak at 700 nm, whereas γ-Al2O3 exhibit a broad peak abound 700 nm and 740 nm.
CL spectra of (a) sample A and (b) sample B annealed at 1000 and 1100°C under relatively low partial pressure (~10−13 Pa) of O2. The acquisition time for the CL spectra was 10 s.
We also acquired XRD patterns of the samples annealed at 1000 to 1100°C to confirm the occurrence of the structural transformation of Al2O3 at a temperature between 1000 to 1100°C. As shown in Fig. 2, α-Al2O3 phase was only detected for the samples annealed at 1100°C, whereas γ-Al2O3 and θ-Al2O3 phases were observed but α-Al2O3 phase was not detected for the samples annealed at 1000°C. These results support that α-Al2O3 and γ-Al2O3 were formed on the sample surfaces annealed at 1100°C and 1000°C, respectively, which is in good agreement with the results of the CL spectra. Although θ-Al2O3 phase was detected for the samples annealed at 1000°C by XRD, distinctive CL peaks related to θ-Al2O3 were not detected. This is because θ-Al2O3 exhibits CL peaks at 685 and 695 nm,34) which are close to the peaks of γ-Al2O3, and their intensities were more than one order magnitude lower than those of γ-Al2O3.35) However, θ-Al2O3 does not exhibit the broad peak around 740 nm that γ-Al2O3 shows.34) Thus, the peaks around 700 nm was slightly distinct for sample B annealed at 1000°C, compared with sample A annealed at 1000°C (Fig. 1), because XRD peak intensities of γ-Al2O3 for sample B annealed at 1000°C were less intense than those for sample A annealed at 1000°C (Fig. 2). Therefore, we might detect the presence of θ-Al2O3 phase from the shape of the peak intensity around 700 nm.
XRD patterns of samples A and B annealed at 1000 and 1100°C under relatively low oxygen partial pressure (~10−13 Pa). (Online version in color.)
The thickness of Al2O3 scales on sample A is expected to be larger than that Al2O3 scales on sample B owing to the higher Al content in sample A (Table 1). Our previous studies revealed that the intensity of the CL peak at 695 nm for α-Al2O3 linearly increased with the thickness in the 0.20–1.50 μm range.17,36) The present study revealed that intensity of CL spectrum of sample A annealed at 1100°C, which formed α-Al2O3 on the surface, was one order magnitude higher than that of sample B annealed at 1100°C (Fig. 1). Furthermore, even though the annealing temperatures were below 1100°C, sample A exhibited higher CL intensities than sample B (Fig. 1). This indicates that we can estimate the thickness of Al2O3 scales from their CL intensities, regardless of structures of Al2O3 scales.
3.2. Imaging of Distribution of α-Al2O3To determine the inhomogeneous distribution of the constituent elements on the surfaces of the annealed alloys, EPMA was conducted to map the element distribution on the alloy surfaces. The samples were annealed under relatively high oxygen partial pressure, which can produce oxide layers with a thickness enough to analyze using EPMA. More in-depth information can be obtained from EPMA because the ordinal scale is on the order of one micron in depth, compared to a few nanometers for XPS. Thus, the quantitative concentrations determined using EPMA are reflective of the thickness of the oxide scales. Figure 3 shows maps of the Fe, Cr, Al, and O concentrations on the surface of sample A annealed at 1100°C under oxygen at a partial pressure of 10−4 Pa. The elemental concentrations obtained from EPMA are displayed using pseudo-colors (red: high concentration; yellow: medium concentration; green or blue: low concentration), and a backscattered electron (COMPO) image is also displayed. These maps reveal that the concentration of aluminum and oxygen are correlated: areas with a high Al concentration corresponded to areas with a high O concentration, and vice versa. This suggests the formation of inhomogeneous Al2O3 layers on the surfaces.
EPMA maps of Fe, Cr, Al, and O, and backscattering (COMPO) image of sample A annealed at 1100°C under oxygen at a partial pressure of 10−4 Pa. (Online version in color.)
Sample A was also annealed at 1100°C under relatively low oxygen partial pressure (~10−3 Pa) to invesitage the influence of the oxygen partial pressure on the inhomogeneous distribution of Al2O3. Figure 4 shows the maps of the Fe, Cr, Al, and O concentrations on the surface of sample A (the same scale in Fig. 3 was used), along with a backscattered electron image. These results clearly indicate the formation of inhomogeneous Al2O3 layers on the alloy surfaces as well as sample A annealed at an oxygen partial pressure of 10−4 Pa. Additionally, decidedly low concentrations of iron and chromium was observed, compared with sample A annealed at an oxygen partial pressure of 10−4 Pa (Fig. 3), revealing that thicker α-Al2O3 scales were formed by increasing the oxygen partial pressure from 10−4 to 10−3 Pa.
EPMA maps of Fe, Cr, Al, and O, and backscattering (COMPO) image of sample A annealed at 1100°C under oxygen at a partial pressure of 10−3 Pa. (Online version in color.)
Finally, CL imaging was performed for sample A annealed at 1100°C under oxygen at respective partial pressures of 10−4 Pa and 10−3 Pa, as shown in Fig. 5. In the CL images, brighter contrast of sample A indicates thicker Al2O3 layers because the same acquisition time was used for the CL images of the two samples. Thus, the thickness of sample A annealed at the oxygen partial pressure of 10–3 Pa was larger than that annealed at the oxygen partial pressure of 10−4 Pa, which is completely consistent with the results obtained by EPMA. These images also indicate that α-Al2O3 were not homogeneously distributed on the alloy surfaces. Ideally, further studies of the spectra should be conducted because it is routine to perform several investigations of the CL spectra of Al2O3.37,38,39,40,41) Nevertheless, because the use of the present CL equipment is relatively convenient, CL mapping can be used to observe the inhomogeneous distribution of corundum (α-Al2O3) rapidly.
CL maps of sample A annealed at 1100°C under oxygen at a partial pressure of (a) 10−4 Pa and (b) 10−3 Pa. The acquisition time for the CL images was 1 s. (Online version in color.)
CL, XPS, XRD, and EPMA were used to characterize Al2O3 formed on the surfaces of Fe–Cr–Al alloys annealed under relatively low oxygen partial pressure. Specifically, the CL spectra were used to identify the structural isomers of Al2O3 formed on the surfaces of Fe–Cr–Al alloys. This approach is suitable for understanding the formation of corundum (α-Al2O3), which is deposited as corrosion-resistant and protective oxide scales at high temperatures. The main conclusions from the CL, XPS, XRD, and EPMA analyses are as follows:
(1) XPS showed that alumina or aluminum oxide (Al2O3) scales were formed on the alloy surfaces after annealing under oxygen at relatively low oxygen partial pressure. The chromium contents in the Al2O3 scales increased by rising the annealing temperature from 1000°C to 1100°C.
(2) CL and XRD analyses showed that the Al2O3 scales on the alloy surfaces changed from γ-Al2O3 and θ-Al2O3 to α-Al2O3 by rising the annealing temperature from 1000°C to 1100°C.
(3) The CL spectrum of the alloys annealed at 1100°C exhibited an intensity peak around 700 nm, which originates from Cr3+ ions in Al2O3. The peak intensity increased one order magnitude by rising the annealing temperature from 1000°C to 1100°C because of the increase of the chromium contents in Al2O3.
(4) The EPMA maps of aluminum and oxygen are consistent with the contrast obtained from the CL images.
These results indicate that CL imaging will contribute to the observation of the inhomogeneous distribution of corundum (α-Al2O3). Furthermore, CL experiments can be relatively conveniently executed.
The authors acknowledge financial support from the Iron and Steel Institute of Japan (ISIJ) through the Research Group of Non-Destructive/On-Site Analysis. Gratitude is also extended to Nippon Steel Stainless Steel Corporation for providing the stainless steel used in this study. This work was partially supported by the Cooperative Research Projects of Research Institute for Electrical Communication, Tohoku University and Global Institute for Materials Research, Tohoku University. They would like to thank to Ms. K. Omura, Mr. T. Tanno, and Mr. M. Tashiro for their technical support with XPS, XRD, and EPMA.
