2025 Volume 65 Issue 1 Pages 111-116
The identification of internal α-alumina (α-Al2O3) scales is crucial because α-Al2O3 determines the performance of heat-resistant alloys. A non-destructive method was recently developed to rapidly identify internal α-Al2O3 scales on the Ni–Cr–Al and Ni–Al alloys by obtaining their surface cathodoluminescence (CL) spectra. However, the peak intensities at 695 nm originating from α-Al2O3 considerably deviated from those estimated from the attenuation rates of overlaid scales (Cr2O3, NiO, and NiAl2O4) and the CL peak intensity at 695 nm for the α-Al2O3 scale on the Fe–Cr–Al alloy surface. In this study, factors causing the deviations in the CL peak intensity at 695 nm for the internal α-Al2O3 scales were investigated by obtaining images of the scales and their CL spectra. The surface Cr2O3 scale on the Ni–25Cr–5Al alloy exhibited a surface with the roughness of ~1 µm, causing the CL peak intensity at 695 nm to be one order of magnitude higher than the estimated value. The CL spectrum and scanning transmission electron microscopy observation revealed that θ-Al2O3 overlaid on α-Al2O3 in the internal Al2O3 scale on the Ni–14Al alloy reduced the CL peak intensity at 695 nm by two orders of magnitude.
Heat-resistant alloys can withstand oxidation and corrosion during their sustained operations at extreme temperatures and under adverse working conditions by forming continuous, adhesive, and slowly growing oxide scales, such as α-alumina (α-Al2O3), chromia (Cr2O3), and silica.1,2,3,4,5) Among them, α-Al2O3 is the most stable and protective oxide scale with high-temperature oxidation and corrosion resistances owing to its thermodynamic stability, good mechanical resistance, and low growth rate.2,6,7,8,9,10) Thus, α-Al2O3-forming heat-resistant alloys are suitable for use at temperatures of 900–1350°C,6,7,8,9,10) such as in aircraft jet engines, incinerators, and industrial turbines.5,6,8,11) α-Al2O3-forming alloys, such as β-NiAl, Fe3Al, and Fe–Cr–Al, form surface α-Al2O3 scales, whereas some alloys, such as γ´-Ni3Al and Ni–Cr–Al, develop internal α-Al2O3 scales underneath other surface oxide scales including NiO, FeOx, NiAl2O4, and Cr2O3 scales.5,9,10,12) Not only surface α-Al2O3 scales but also internal α-Al2O3 scales exhibit excellent oxidation and corrosion resistances.8,10) Therefore, identifying internal α-Al2O3 scales is critical for evaluating the performance of α-Al2O3-forming alloys.
Conventionally, internal scales on heat-resistant alloys have been identified via destructive analysis, such as by observing their cross sections using scanning electron microscopy (SEM) combined with energy- or wavelength-dispersive X-ray spectrometry (SEM–EDX or SEM–WDX) and by etching the surface with ions using Auger electron spectroscopy, X-ray photoelectron spectroscopy, and secondary ion mass spectrometry.11,13) Therefore, a nondestructive analytical method is desirable for rapidly identifying such internal scales because their sample preparation can be simplified, which reduces the analysis time. Photostimulated luminescence spectroscopy (PSLS) can be used to identify internal α-Al2O3 scales nondestructively;14,15,16,17) however, performing PSLS measurements over large areas, such as millimeter scales, is time-consuming because PSLS can only analyze in a small area (e.g., 2–5 μm16,17)) in a single measurement. Therefore, an analytical technique that can rapidly identify internal scales in large areas is preferable.
Recently, the author’s group developed a method that satisfied the aforementioned requirements.18,19) In this method, the cathodoluminescence (CL) spectra of compounds that emit light after being subjected to electron bombardment were obtained. Internal α-Al2O3 scales formed underneath the Cr2O3, NiO, and NiAl2O4 scales were identified in an area of 0.27 × 0.21 mm2 in a single measurement by obtaining the CL spectra of the heat-resistant Ni–Cr–Al and Ni–Al alloys surfaces.18,19,20) The CL spectra showed a peak at 695 nm, originating from the substitution of aluminum (III) (Al3+) ions with chromium (III) ions (Cr3+) in the α-Al2O3 scales.18,21,22,23,24) Although the peak intensity decreased by the absorption of light emitted from the internal α-Al2O3 scale by the overlaid scale of Cr2O3, the CL spectrum of the Ni–25Cr–5Al alloy with the 2.5-μm-thick surface Cr2O3 scale and 0.9-μm-thick internal α-Al2O3 scale (Fig. 1(a)) showed a peak at 695 nm. The intensity of this peak was one order of magnitude higher than that calculated from the attenuation rate of the peak at 695 nm for a 2.5-μm-thick Cr2O3 film deposited on sapphire (α-Al2O3) and that of the CL peak at 695 nm for the 1.0-μm-thick α-Al2O3 scale formed on the Fe–10Cr–15Al alloy surface (Fig. 1(b)).20,24) By contrast, when the CL spectrum was obtained for the Ni-14Al alloy comprising surface NiO scale and internal NiAl2O4 scale, followed by the α-Al2O3 at the bottom with 1.1, 0.8, and 0.5 μm thickness (Fig. 1(c)), the CL peak intensity at 695 nm was two orders of magnitude lower than that calculated from the attenuation rates of the peak at 695 nm for the 1.1-μm-thick NiO and the 0.8-μm-thick NiAl2O4 films deposited on sapphires, and the CL peak intensity at 695 nm for the 0.6-μm-thick α-Al2O3 scale formed on the Ni–18Al alloy surface.19) The reason for such varied intensities of the CL peak at 695 nm corresponding to internal α-Al2O3 scales on Ni–25Cr–5Al and Ni–14Al alloys remains unknown.
Herein, factors affecting the intensity of the CL peak at 695 nm originating from α-Al2O3 on Ni–Cr–Al and Ni–Al alloys were identified by analyzing the microstructures and CL spectra of the internal Al2O3 scale. The results of this research can help to evaluate the thickness of scales overlaying the internal Al2O3 scales on heat-resistant alloys and the phases in the internal Al2O3 scales.
The Ni–25Cr–5Al alloy was prepared by melting 70 mass% Ni powder (purity: 99%, Wako Pure Chemical Industries, Ltd., Osaka, Japan), 25 mass% Cr powder (purity: 99.9%, Kojundo Chemical Laboratory Co., Ltd., Saitama, Japan), and 5 mass% Al powder (purity: 99.99%, Kojundo Chemical Laboratory Co., Ltd., Saitama, Japan) in an Al2O3 crucible. The powder mixtures were heated at 1560°C for 30 min before being cooled to room temperature at a rate of 5°C min−1 under argon flow at 200 mL min−1. The Ni–25Cr–5Al alloy was annealed at 1100°C and 0.1 Pa for 12 h and cut into cubic slices of ~5 mm using a low-speed precision cutter (IsoMet LS, Buehler Ltd., Lake Bluff, IL, USA). One exposed surface of Ni–14Al slice was polished using 600-, 1200-, and 2400-grit abrasive papers, followed by 1-μm water-free diamond slurry (Aka-Poly WF 1 μm, Akasel A/S, Roskilde, Denmark). The polished slice was heated at 1000°C for 16 h in air to create a surface Cr2O3 scale and an internal Al2O3 scale. The Ni–14Al alloy was prepared by melting 86 mass% Ni powder (purity: 99%, Wako Pure Chemical Industries, Ltd., Osaka, Japan) and 14 mass% Al powder (purity: 99.99%, Kojundo Chemical Laboratory Co., Ltd., Saitama, Japan) following the same procedure as for the Ni–25Cr-5Al alloy. To create surface NiO scale as well as internal NiAl2O4 and α-Al2O3 scales, the polished slice of on the Ni–14Al alloy was heated at 1000°C for 4 h in air.
The chemical compositions of samples were determined using an SEM (TM3030 Plus, Hitachi High-Technologies Co., Tokyo, Japan) equipped with a silicon drift EDX detector (Quantax70, Bruker Corp., Billerica, Massachusetts, USA) and field emission electron probe microanalyzer (FE–EPMA, JXA-8530F, JEOL Ltd., Tokyo, Japan) equipped with WDX spectrometers using the standard samples of Al2O3 and Cr2O3. The acceleration voltage and beam current for EPMA were set to 15 kV and 5 nA, respectively. The characteristic X-ray peak intensities for Al (Al Kα) and Cr (Cr Kα) were measured for 10 s. The microstructure of scales on the Ni–14Al alloy heated at 1000°C for 4 h in air was observed and characterized via a scanning transmission electron microscopy (STEM, JEM-ARM200F, JEOL Ltd., Tokyo, Japan) equipped with a silicon drift EDX detector (JED-2300, JEOL Ltd., Tokyo, Japan). The specimen for STEM was prepared using an ion slicer (EM-09100IS, JEOL Ltd., Tokyo, Japan) at an acceleration voltage of 6 kV and a glancing angle of 1.2° at room temperature. Then, final polishing was performed using a precision ion polishing system (PIPS II system, Gatan, Inc., Pleasanton, California, USA) at an acceleration voltage of 0.5 kV to remove the damaged layer from the specimen. The acceleration voltage of the STEM microscope was set to 200 kV.
The effect of Cr2O3 dissolution into the α-Al2O3 scale on its CL intensity was investigated by obtaining CL spectrum of samples prepared by heating a powder mixture of α-Al2O3 (purity: 99.99%, Wako Pure Chemical Industries, Ltd., Osaka, Japan) and Cr2O3 (purity: 99.9%, Wako Pure Chemical Industries, Ltd., Osaka, Japan) reagents at 1500°C in air. These reagent powders were mixed in several ratios using an agate mortar and pestle, and was pressed into pellets at 1 MPa. The pellets were placed on an Al2O3 plate and heated at 1500°C for 1 h in air. Phases in the pellets of α-Al2O3 and Cr2O3 powder mixture were identified by X-ray diffraction analysis (XRD, Ultima IV/SG, Rigaku Corporation, Tokyo, Japan) using a Cu–Kα line.
The CL spectra were acquired for the scales on the Ni–14Al alloy and the pellets of α-Al2O3 and Cr2O3 powder mixture using the custom SEM–CL system.25,26,27,28,29,30) The light emitted from the samples due to electron bombardment was collimated using an aluminum-coated off-axis parabolic mirror with a 0.5-mm hole at the center. The collimated light was collected using an optical spectrometer (QE65Pro, Ocean Optics Inc., Largo, Florida, USA) with a grating of HC-1 and a slit width of 25 μm, corresponding to a spectral resolution (full width half maximum, FWHM) of 3.16 nm.31) The tip of the optical fiber with the plano-convex lens was attached to the SEM instrument (Mighty-8DXL, TECHNEX, Tokyo, Japan). The central axis of the electron beam between the electron gun and the scan coil in the SEM–CL system was misaligned with the center of the aperture in the objective lens; this arrangement prevented the light emitted by the tungsten filament of the electron gun from entering the spectrometer. The acceleration voltage of the SEM was set to 17 kV.
The intensity of the CL peak at 695 nm for the internal α-Al2O3 scale on the Ni–25Cr–5Al alloy was one order of magnitude higher than that calculated from the attenuation rate of the peak at 695 nm for a Cr2O3 film and the intensity of the CL peak at 695 nm for the surface α-Al2O3 scale on the Fe–10Cr–15Al alloy.20) The CL peak at 695 nm originated from Cr3+ substituted for Al3+ in the α-Al2O3 scale. Thus, Cr2O3 compositions in α-Al2O3 scales were first determined by performing elemental analysis using FE–EPMA on the cross section of the Ni–25Cr–5Al alloy with the 2.5-μm-thick surface Cr2O3 scale and 0.9-μm-thick internal α-Al2O3 scale (Fig. 1(a)). For comparison, the compositions of ⁓1.0-μm-thick α-Al2O3 scales on the Fe–25Al and Fe–10Cr–15Al alloy surfaces (Figs. 1(b) and 1(d))24,32) were determined by analyzing the cross sections of the alloys. The Cr2O3 compositions in α-Al2O3 scales on the Fe–25Al and Fe–10Cr–15Al alloys was successfully determined by bombarding electrons near α-Al2O3 scale surfaces because the total compositions determined from the EPMA analysis were close to 100 mass% (Table 1). However, the total composition of the internal α-Al2O3 scale on the Ni–25Cr–5Al alloy was not near 100 mass%, indicating that electrons were bombarded on Ni–25Cr–5Al alloy and/or Cr2O3 scale above the internal α-Al2O3 scale. Actually, a Ni Kα line was detected while analyzing the internal α-Al2O3 scale on the Ni–25Cr–5Al alloy. Subsequently, the SEM–EDX elemental analysis of α-Al2O3 scales on the Fe–25Al, Fe–10Cr–15Al, and Ni–25Cr–5Al alloys was performed by bombarding electrons on areas above the interfaces between α-Al2O3 scales and alloys. The results are shown in Table 1 as elemental compositions because the characteristic X-rays of Al, Cr, and Ni originated from the oxide (scale) and metal (alloy). The Cr content in the internal α-Al2O3 scale on the Ni–25Cr–5Al alloy was clearly higher than those in α-Al2O3 scales on the Fe–25Al and Fe–10Cr–15Al alloys. Based on the EPMA results for the Fe–25Al and Fe–10Cr–15Al alloys, the Cr2O3 content in the internal α-Al2O3 scale on the Ni–25Cr–5Al alloy was estimated to be ⁓1 mass%. The measured value was approximated because the Cr content differed across alloys.
| Fe–25Al | Fe–10Cr–15Al | Ni–25Cr–5Al | ||
|---|---|---|---|---|
| EPMA (WDX) | Al2O3 | 100.1 | 99.3 | 77.8 |
| Cr2O3 | 0.1 | 0.3 | 4.7 | |
| Total | 100.2 | 99.6 | 82.5 | |
| SEM–EDX | Al | 52 | 51 | 50 |
| Cr | 0 | 2 | 5 | |
| Fe | 7 | 9 | – | |
| Ni | – | – | 6 | |
| O | 40 | 33 | 37 | |
| Total | 99 | 95 | 98 | |
EPMA and SEM–EDX analysis revealed that the internal α-Al2O3 scale on the Ni–25Cr–5Al alloy contained higher amounts of Cr2O3 than the surface α-Al2O3 scales on the Fe–25Al and Fe–10Cr–15Al alloys. Further, CL spectra were acquired for α-Al2O3 containing Cr2O3 with its content of 0.1, 0.37, 0.75, 1.5, 2.3, 3.8, and 7.9 mass% to investigate the effects of dissolved Cr2O3 on its CL properties. Additionally, this measurement can determine whether the Cr2O3 content in the α-Al2O3 scale on heat-resistant alloys affects the intensity of the CL peak at 695 nm, which is important for the quantitative evaluation of these scales from CL spectra, such as thickness measurement. XRD analysis confirmed that all samples formed a single phase of corundum (Fig. S1: Supporting Information). The CL spectra showed the strongest peak at 695 nm, originating from Cr3+ in the α-Al2O3 of all samples (Fig. 2(a)). The intensity of the CL peak at 695 nm was almost constant below 1 mass% of Cr2O3, and drastically decreases with increasing Cr2O3 content (Fig. 2(b)). This indicated that the dissolution of >0.1 mass% Cr2O3 into α-Al2O3 did not enhance the intensity of the CL peak at 695 nm. Futehrmore, the intensity of the CL peak at 695 nm for the internal α-Al2O3 scale on the Ni–25Cr–5Al alloy was almost the same as those of the surface α-Al2O3 scales on the Fe–25Al and Fe–10Cr–15Al alloys because the internal α-Al2O3 scale on the Ni–25Cr–5Al alloy contained ⁓1 mass% of Cr2O3. Thus, the dissolution of Cr2O3 did not impact the intensity enhancement of the CL peak at 695 nm for the internal α-Al2O3 scale on the Ni–25Cr–5Al alloy. These results support previous findings that the intensity of the CL peak at 695 nm linearly increased with increasing thickness of the surface α-Al2O3 scale, regardless of the base alloys.24,32) This was because the peak intensitiy was almost constant when the Cr2O3 content range was below 1 mass% in the α-Al2O3 scales on heat-resistant alloys (below 1 mass%). Thus, the thickness of α-Al2O3 scales on the surface of heat-resistant alloys can be determined from a single calibration line of the intensities of the CL peak at 695 nm, regardless of the base alloys.
After heat treatment at 1000°C, Cr2O3 grains of ⁓1 μm covered the entire Ni–25Cr–5Al alloy surface (Fig. 3).20) Thus, the surface Cr2O3 scales on the Ni–25Cr–5Al alloy exhibited rough surfaces, indicating the presence of areas with the Cr2O3 thickness 1 μm lower than the average thickness of 2.5 μm. The attenuation rate of the CL peak at 695 nm for the 1.5-μm-thick Cr2O3 film was ⁓20 times higher than that for the 2.5-μm-thick Cr2O3 film,20) indicating that the intensity of the CL peak at 695 nm for these areas was one order magnitude higher than that for areas with the Cr2O3 thickness of 2.5 μm. Therefore, owing to the presence of the areas with the Cr2O3 thickness of 1.5 μm, the intensity of the CL peak at 695 nm for the internal α-Al2O3 scale on the Ni–25Cr–5Al alloy was higher than that calculated from the attenuation rate of the peak at 695 nm for the Cr2O3 film and the intensity of the CL peak at 695 nm for the surface α-Al2O3 scale.
The scales on the Ni–14Al alloy heated at 1000°C for 4 h comprised NiO scale on the surface, followed by NiAl2O4 and α-Al2O3 underneath it. The intensity of CL peak at 695 nm for the internal α-Al2O3 scale was two orders of magnitude lower than that calculated from the attenuation rates of the CL peak at 695 nm for NiO and NiAl2O4 films and the intensity of the CL peak at 695 nm for the surface α-Al2O3 scale on the Fe–15Al alloy.24) NiO did not affect the intensity of the CL peak at 695 nm for the internal α-Al2O3 scale because the pseudo-binary phase diagram of the Al2O3–NiO system suggested that NiO did not dissolve into the α-Al2O3 scale.33) θ-Al2O3 existed as a transient oxide during the early oxidation stage of Ni–Al alloys heated at 1000°C2,5,34,35,36) because the growth rate of θ-Al2O3 was two orders of magnitude higher than that of α-Al2O3.36) By contrast, heat-resistant alloys containing higher amounts of Cr, such as the Ni–25Cr–5Al alloy, do not normally form θ-Al2O3 scales after heat treatment at 1000°C because Cr promotes α-Al2O3 scale formation.2) The CL spectrum of θ-Al2O3 showed a peak at 685 nm17,31,37,38,39) and its luminescence intensity was one order of magnitude lower than that of α-Al2O3.38) This indicated that the presence of θ-Al2O3 in the Al2O3 scale reduced the intensity of the CL peak at 695 nm. The CL spectrum of the Al2O3 scale can distinguish between α-Al2O3 and θ-Al2O3 in the scale based on the peak position, i.e., if the peak is appearing at 695 nm (α-Al2O3) or 685 nm (θ-Al2O3). Furthermore, the θ-Al2O3 content (Cθ) can be determined from the intensities of CL peaks at 695 nm and 685 nm as follows:31,39)
| (1) |
where I685 nm and I695 nm are the emission intensities at 685 nm and 695 nm, respectively. When the CL spectrum was acquired by bombarding electron beam on the cross section of the Al2O3 scale on the Ni–14Al alloy (Fig. 1(c)), an intense peak was observed at 695 nm, corresponding to the peak of the α-Al2O3 scale on the β-NiAl alloy heated at 1100°C for 10 h (Fig. 4). In addition to the primal CL peak at 695 nm, a small peak was observed at 685 nm (an arrow in Fig. 4) that corresponded to the peak of the θ-Al2O3 scale on the β-NiAl heated at 1000°C for 1 h. The scale on the β-NiAl alloys heated at 1100°C for 10 h and at 1000°C for 1 h comprised only α-Al2O3 and θ-Al2O3, respectively as confirmed via XRD.31) The internal Al2O3 scale on the Ni–14Al alloy contained 45% of θ-Al2O3 as determined using Eq. (1). Moreover, the transformation from θ-Al2O3, to α-Al2O3 began at the boundary between θ-Al2O3 and the β-NiAl alloy,2) indicating that most α-Al2O3 was present within the internal Al2O3 scale, further reducing the intensity of the CL peak at 695 nm. The STEM observation revealed the presence of needle-like grains, which are a typical shape of metastable alumina of γ-Al2O3 and θ-Al2O35,34,36,40,41) near the boundary between the Al2O3 scale and NiAl2O4 scale (Fig. 5(a)). Further, the NiAl2O4, Al2O3, and Ni–14Al phases were confirmed via EDX analysis; however, a NiO phase was not confirmed because the NiO scale was sputtered by argon ions during the preparation of the specimen using a precision ion polishing system. Furthermore, some needle-like grains (e.g., areas A and B in Fig. 5(b)) were confirmed to be Al2O3 because aluminum and oxygen were only detected in grains via EDX mapping (Figs. 5(b)–5(e)). This confirmed the presence of metastable alumina at the boundary between the Al2O3 and NiAl2O4 scales. Therefore, the formation of θ-Al2O3 and presence of α-Al2O3 in an area within the internal Al2O3 scale decreased the intensity of the CL peak at 695 nm by two orders of magnitude compared with that estimated from the attenuation rates of the CL peak at 695 nm for the NiO and NiAl2O4 films and the intensity of the CL peak at 695 nm for the surface α-Al2O3 scale.
Herein, factors affecting the intensity of the CL peak at 695 nm originating from α-Al2O3 in internal Al2O3 scales on the Ni–Cr–Al and Ni–Al alloys were determined via surface observation, elemental analysis, and CL analysis. The surface observation of scales on the Ni–25Cr–5Al alloys heated at 1000°C revealed the presence of an area where the Cr2O3 scale thickness was 1 μm lower than the average thickness (2.5 μm), which caused the intensity of the CL peak at 695 nm to be one order of magnitude higher than that calculated from the attenuation rate of the CL peak at 695 nm for Cr2O3 and the intensity of the CL peak at 695 nm for the surface α-Al2O3 scale. The CL spectra of α-Al2O3 containing various Cr2O3 contents revealed that the intensity of the CL peak at 695 nm was almost constant below 1 mass% of Cr2O3. This confirmed that the thickness of surface α-Al2O3 scales on heat-resistant alloys can be determined from a single calibration line between the intensity of the CL peak at 695 nm and the thickness of the α-Al2O3 scale, regardless of base alloys. The CL spectrum of the internal Al2O3 scale on the Ni–14Al alloy heated at 1000°C for 4 h demonstrated that the scale contained 45% of θ-Al2O3. The STEM observations revealed the presence of the metastable Al2O3 scale near the boundary between the Al2O3 and NiAl2O4 scales. These results indicate that the presence of α-Al2O3 underneath θ-Al2O3 decreased the intensity of the CL peak at 695 nm by two orders of magnitude than that calculated from the attenuation rates of the peak at 695 nm for the NiO and NiAl2O4 films and the intensity of the CL peak at 695 nm for the surface α-Al2O3 scale.
The author has no conflicts of interest to declare that are relevant to the content of this article.
XRD patterns of α-Al2O3 containing Cr2O3 heated at 1400°C. This material is available on the Website at https://doi.org/10.2355/isijinternational.ISIJINT-2024-309.
Financial support for the present study was provided by JSPS KAKENHI Grant Number 22H01837 and SDGs Research Project of Shimane University. The author thanks Mr. Issei Narita of Tohoku University for helping to perform the EPMA analysis.
