2013 Volume 53 Issue 11 Pages 2000-2006
Since selective oxidation of alloying elements on the surface of steel products influences their surface properties, characterization of the surface oxides which can be considered as non-metallic inclusions is of great importance. In this study, X-ray photoelectron spectroscopy (XPS), X-ray diffractometry (XRD) and X-ray absorption spectroscopy (XAS) were used for characterizing the formation process of Mn oxides on the surface of annealed Fe–Mn alloys under a low partial pressure of oxygen. The results obtained by XPS showed that the enrichment and oxidation of Mn occurs on the surface of the Fe–Mn alloys annealed under low oxygen partial pressure, and Mn oxides are formed in the metallic Fe matrix in the surface layers. XAS spectra using grazing-exit X-ray fluorescence measurements showed depth-resolved information on chemical state of Mn. These Mn oxides were identified as MnO (manganosite) by grazing-incident XRD measurements. It was found using in situ XRD measurements at high temperatures that the lattice constant of MnO increased with increasing annealing temperature, which attributed to the non-stoichiometry of MnO. These oxidation characteristics of Mn in the Fe–Mn alloys are discussed on the basis of thermochemical properties of Mn.
Alloying elements, such as Si and Mn, are often added to steel or Fe based alloys to improve their mechanical properties. However, the oxidation characteristics of the alloying elements in the Fe based alloys are complicated, although fundamental oxidation processes have been discussed.1,2) Typically, reactive elements or less-noble alloying elements in Fe based alloys are selectively oxidized to form their oxides in the surface layer of the Fe based alloys on annealing at high temperatures. Such oxides can be considered to be non-metallic inclusions, since their properties are different from those of the matrix iron. If the surface of steel products containing such reactive elements is annealed in atmosphere of various oxygen partial pressures in production lines, oxides with complex morphology are formed on their surfaces depending on the species and amount of alloying elements.3,4,5,6,7,8) These oxides of alloying elements are considered to influence surface properties which are wettability between solid and liquid, galvanizing characteristics and so on.9) Therefore, characterization of formation processes of oxides of reactive elements on the surface of Fe based alloys provides information on control of the surface properties of annealed steel products. So far, it has been reported that reactive elements such as Si, Mn and Al are oxidized on the surface of Fe based alloys on annealing under various conditions.5,8,9,10,11) Conventional surface analytical methods such as X-ray photoelectron spectroscopy (XPS) and secondary ion mass spectrometry (SIMS) were used in these studies. It has been shown in these studies that the oxide layers of reactive elements like Si and Al cover the surfaces of these Fe based alloys under low partial pressure of oxygen,5,10) and these oxide layers must change the surface properties of the Fe based alloys.
Mn is, more or less, added to steel products, and it was reported that Mn is often enriched and oxidized in the surface layer of Fe based alloys on annealing under low partial pressure oxygen.4,9) Although such surface layers containing Mn oxide may influence the surface properties such as hot-dip galvanizing performance, surface enrichment and oxidation of Mn in Fe based alloys is still unclear. In order to understand the formation processes of Mn oxide in the surface layer of Fe based alloys on annealing, several analytical techniques such as compositional analysis and structural analysis should be combined. The objective of the present work is to characterize the formation of Mn oxides in the surface layer of Fe–Mn binary alloys which were annealed under low partial pressures of oxygen using various analytical methods. In particular, analysis of Mn oxides formed in the surface was the focus of this study. XPS and X-ray diffraction (XRD) were used for analyzing the chemical composition and chemical state of Mn and the structure of oxide phases in the surface layer of the binary alloys. X-ray absorption spectroscopy (XAS) using measurements of grazing-exit X-ray fluorescence from the sample surface was also used for analyzing the chemical state of Mn in the surface layers. These results were discussed on the basis of the thermochemical characteristics of elements.
Ingots of Fe–Mn binary alloys were prepared by vacuum induction melting. The chemical compositions of the alloys prepared in this work were Fe-3.72 mass% Mn and Fe-5.46 mass% Mn, which are hereafter referred to as Fe4Mn and Fe5Mn, respectively. The ingots were hot-rolled to sheets of 0.5 mm in thickness at 1273 K, and the sheets were cut in squares with an edge of about 10 mm. After sample surfaces were mechanically polished, the samples were washed with acetone in ultrasonic bath.
These samples were annealed in the temperature range between 773 and 973 K for 1800 s in 9.8% H2–Ar gas. Although it was difficult to control the precise pressure of oxygen, the partial pressure of oxygen was estimated to be approximately to 10–30 through 10–23 Pa in this temperature range by measuring dew points of the gas. According to the free energy of formation of the oxide of elements, Fe is not oxidized while Mn is oxidized as MnO under the annealing condition, as shown in Fig. 1 where thermochemical data given by a handbook12) were plotted. Here, it should be noted that the surfaces of annealed samples were exposed to air when they were transferred from the annealing furnace to an analytical apparatus. During the transfer, a thin native oxide layer is formed on the sample surfaces even at room temperature. The native oxide layer formed on samples is detected by the surface analytical method like X-ray electron spectroscopy, as shown later.
Ellingham diagram for oxide formation in Fe and Mn. Annealing conditions in this study are denoted as gray color in the figure.
The chemical composition and state of thin layers on sample surfaces were analyzed by an XPS apparatus (PHI-5600). The incident X-ray was monochromated Al Kα and the analysis area was about 1 mm in diameter. The XPS spectra were measured at a take-off angle of 45°. Ar ion sputtering was applied for obtaining depth profiles of XPS spectra. The sputtering rate of Ar ion beam was estimated to be 1.8 nm/min, by assuming that the sputtering rate for SiO2 was comparable to that of the Fe–Mn alloy. The chemical concentration was calculated from the intensities of XPS spectra along with the relative sensitivity factor of each element.
An XRD apparatus (Rigaku, RINT2200V) with Cu Kα incident radiation was used to identify the crystal structure of oxides formed in the surface layers of the Fe–Mn alloys. In situ XRD measurements were also performed to identify crystal structures of surface layers formed during annealing at high temperatures. An atmosphere of 9.8% H2–He was used to reduce absorption of incident and diffracted X-rays in diffraction measurements. The incident angle was varied from 0.5 to 3° to effectively detect diffraction peaks from surface layers of samples. The experimental setup of the present in situ XRD measurements is illustrated in Fig. 2. The information depth of diffraction patterns at the incident angle of 0.5° is estimated as sub-micrometer or less.
Schematic of the experimental setup for XRD measurements for analysis of the structure of surface layers. The incident angle, Ω, was small in this experiment.
In order to investigate the chemical state of Mn in the surface layers of samples annealed at different temperatures, XAS spectra using measurements of grazing-exit fluorescence intensities were obtained by using synchrotron radiation at the BL37XU in SPring-8 with the approval of the Japan Synchrotron Radiation Research Institute. The geometry for the experiment is schematically shown in Fig. 3.13) Incident X-ray beam energy by synchrotron radiation was scanned around the Mn K absorption edge (6.539 keV). Synchrotron X-ray beam, monochromatized by Si(111) double-crystal monochromators, was irradiated perpendicularly onto the sample surface after passage through a slit of size 0.5 mm × 0.5 mm. A two-dimensional pixel array detector (PILATUS II), which was positioned parallel to the sample plane, was utilzed to detect the emitted fluorescence. The distance between the irradiation point on the sample surface and the detector window was 172 mm. The detection angle resolution by one pixel array with a width of 172 μm was 1 mrad. Fluorescence X-ray intensities, which were counted at pixels on a line positioned at the constant take-off angle, were summed up to improve the signal to noise ratio. These XAS spectra involve chemical information on elements within a sub-micrometer in thickness,14) depending on the measuring conditions.
Schematic of depth-resolved XAS analysis using XRF signal measurements.
Scanning electron microscopy (SEM) was used to observe the surface morphology of samples annealed at different temperatures. Figure 4 shows SEM images of the surfaces of Fe4Mn annealed at 773, 873, and 973 K. Although the sample surface before annealing was flat, the roughness of the surface increased with increasing annealing temperature. This indicates that the oxidation of Mn in the samples causes the change in surface morphology.
SEM images of the samples annealed at (a) 773 K, (b) 873 K, and (c) 973 K.
It is desirable to measure flat surfaces of samples in XPS for quantitative surface analysis. However, XPS was applied to analyze the chemical composition and state of the surfaces of the annealed samples with a roughness in this study, as in shown in Fig. 4. Figure 5 shows the concentrations of the surfaces of Fe4Mn annealed at 773, 823, 873, and 973 K. The concentrations were calculated from intensities of Fe 2p, Mn 2p, and O 1s XPS spectra measured at the take-off angle of 45°. As these samples were exposed to air after annealing, the surface concentration of oxygen was high due to adsorption of contaminated carbon species from air.15) These XPS results indicate that the surface concentration of Mn is considerably higher than that of the bulk level in samples annealed at temperatures above 773 K. It is noted that the surface concentration of elements appears to depend on annealing temperature in complicated manner. This is because only a thin surface layer was analyzed by XPS, while the surface roughness is increased by selective oxidation of less-noble elements in a surface layer of iron based alloys.16) It is generally considered that less-noble elements such as Si and Mn are enriched in a surface layer to form their oxide with increasing annealing temperature.17) Therefore, the XPS results obtained for a thin surface layer as shown in Fig. 5 are influenced by the morphology of Mn oxide and metallic Fe.
Surface concentration calculated from XPS data for the surfaces of Fe4Mn, which were annealed at 773, 823, 873, and 973 K for 1800 s under low partial pressure of oxygen.
In order to obtain in-depth information on the chemical state of Fe and Mn in the surface layers of Fe4Mn, XPS spectra were measured along with Ar ion sputtering. Figure 6 shows Fe 2p and Mn 2p XPS spectra measured in different depth for the samples annealed at 773 K for 30 min. The intensities of Fe 2p and Mn 2p spectra were low at the sample surface, as the surface was exposed to air after annealing. The Fe 2p and Mn 2p XPS spectra became clear after Ar ion sputtering. Fe 2p XPS spectra reveal the Fe oxide state (binding energy, BE = 710.0 eV) in the outer layer. The Fe 2p XPS the spectra reveal that Fe is metallic state in the inner layer, as the peak of metallic Fe (BE = 707.0 eV) appears with increasing sputtering time. On the other hand, Mn 2p XPS spectra indicate that Mn is the oxide state (BE = 641.5 eV) irrespective of the depth under the present experimental conditions, as Mn 2p XPS spectra were not changed by sputtering. These results imply that Mn oxides are formed in metallic Fe matrix. Mn is considered to be selectively oxidized in the metallic Fe matrix, as expected from the Ellingham diagram, as shown in Fig. 1. Practically the activity of Mn in Fe–Mn alloys is lower than that in pure Mn. Although, oxygen is penetrated into the alloy to be reacted with Mn during selective oxidation of Mn in Fe–Mn alloys, reaction conditions, such as the activity of Mn and O at the metal/oxide interface and the mobility of these elements in the matrix, are very complicated.1,2) Therefore, quantitative description of the whole kinetics of selective oxidation is limited in this study.
XPS spectra for the Fe4Mn sample annealed at 773 K (a) Fe 2p and (b) Mn 2p. The samples were sputtered up to about 18 nm.
From sequential Fe 2p, Mn 2p, O 1s, and C 1s XPS spectra, the compositional depth profiles of C, O, Fe, and Mn were calculated for the surface layers Fe4Mn samples annealed at 773 and 973 K. The surface contaminated organic layer was almost removed by sputtering up to about 2 nm. Figures 7(a) and 7(b) show that the maximum concentrations of Mn were about 38 and 50 at% at 4 and 3 nm depths, respectively. The total Mn concentration in the surface layer increased with increasing annealing temperature. Here, it should be noted that the apparent atomic concentration in the depth profiles from the oxide surface are strongly influenced by selective sputtering of oxygen, though Ar ion sputtering is often applied for obtaining the depth profiles.18) Thus, it is considered that the elemental concentration in the depth profiles is deviated from the real value. Nevertheless, the present results indicate that an enrichment of Mn to the surface occurred during annealing in 9.8% H2–Ar gas in spite of 3.8 at% Mn in bulk. The distribution of Mn and O seemed to be correlative, indicating the formation of Mn oxides in the surface layer. These results coupled with the SEM observation results indicate that Mn oxide particles were formed in the metallic Fe matrix, and the amount and/or size of Mn oxide particle increased with increasing annealing temperature.
XPS depth profiles for Fe4Mn annealed at (a) 773 K and (b) 973 K. The Mn composition increased with increasing annealing temperature. The distribution of Mn is correlated with that of O.
As examples of different diffraction peaks from surfaces of Fe–Mn alloys, Fig. 8 shows X-ray diffraction patterns of Fe4Mn treated by different conditions as samples (a), (b), and (c). The incident angle of X-ray was 0.5° in these measurements. Sample (a) was annealed at 773 K under a low partial pressure of oxygen and sample (b) was subsequently annealed at 673 K in air. Sample (c) was annealed at 673 K in 9.8% H2–He gas after annealing in air. The reference data for Fe and relevant oxides are also shown in Fig. 8. Depending on the partial pressure of oxygen, diffraction peaks from metallic body-centered cubic iron (Fe), hematite (α-Fe2O3) or magnetite (Fe3O4) were observed in these patterns. As an oxide of Mn, the diffraction peaks assigned to MnO (manganosite) were detected in the pattern of sample (a).
XRD patterns for (a) Fe4Mn annealed at 773 K under low partial pressure of oxygen, (b) Fe4Mn subsequently annealed at 673 K in air, and (c) Fe4Mn finally annealed at 673 K in 9.8% H2–He gas.
Figure 9 shows the X-ray diffraction patterns for Fe4Mn annealed at (a) 773 K, (b) 873 K, and (c) 973 K. These diffraction patterns were measured at X-ray incident angles from 0.5 to 3.0°. With decreasing incident angle, the diffraction intensity of Fe (110) at about 45° decreases, while the intensities of MnO (111) at about 35° and MnO (200) at about 41° does not decrease. This indicates that MnO was formed in the surface layer of this sample. The diffraction intensity of Mn (111) increases with increasing temperature, indicating that the amount of MnO increases with increasing temperature. This is consistent with the XPS depth profile results shown in Fig. 7, though only the chemical composition of thin surface layers was analyzed by XPS.
XRD patterns for Fe4Mn annealed at (a) 773 K, (b) 873 K, and (c) 973 K, which were measured at different incident X-ray angles shown in Fig. 2. The pattern was obtained at incident angles from 0.5 to 3.0°.
As Fe4Mn annealed under low partial pressure of oxygen appears to be composed of Fe and MnO from X-ray diffraction patterns, the lattice constants of the Fe matrix and MnO can be analyzed from diffraction peak positions measured at high temperatures. Figure 10 shows lattice constant of Fe and MnO as a function of annealing time, which were estimated from data of in situ XRD measurements for Fe4Mn and Fe5Mn annealed at 973 K. These results indicate that the lattice constant of metallic Fe estimated from the Fe (110) peak was almost unchanged through the XRD measurements at 973 K. On the other hand, the lattice constant of MnO estimated from the MnO (111) peak was smaller than that of reference material MnO in the short annealing duration. Therefore, the lattice constant of MnO formed in the metallic Fe matrix increased with annealing duration. It is considered that this characteristic lattice constant change is attributed to non-stoichiometry of MnO, which reveals a compositional width depending on temperature as presented in thermochemical data19) and phase diagram.20) When annealing time is short, the amount of oxygen penetrating to the sample surface layer is small. Then, the value x in MnOx is small and the lattice constant is small. If annealing time becomes long, the x value in MnOx and the lattice constant of MnOx increase.
Lattice constants of (a) Fe (110) and (b) MnO (111) for Fe4Mn and Fe5Mn annealed at 973 K versus annealing time. Reference data for MnO at 973 K were shown.
Here, it is noted that the lattice constant of MnO formed in the surface layer Fe5Mn increases earlier than the case of Fe4Mn as shown in Fig. 11. Figure 11 illustrates the schematics of changes of x value in MnOx during annealing. Because the amount of Mn diffusing from the inner side in high Mn alloy is higher than in low Mn alloy, the amount of MnO formed in the surface layer may be higher in the high Mn alloy. When oxygen do not penetrate deeply into the surface layer, the lattice constant of MnO formed in the surface layer measured by XRD depends on Mn concentration. On the other hand, when the annealing time was long, oxygen penetrated deeply, resulting in a small dependence of the lattice constant on the bulk Mn concentration. Thus, it is considered that the difference in the oxidation processes between Fe4Mn and Fe5Mn was observed in the present XRD measurements, although it was difficult to obtain information on whole depth profiles by surface sensitive XPS.
Schematic of MnOx formed in Fe–Mn alloy during annealing. In a short annealing duration, shallow oxygen penetration into the surface layer resulted in MnOx with a lower x value. However, for a longer annealing duration, MnOx with higher x value was formed.
Furthermore, it should be remarked that there are many factors influencing the kinetics of selective oxidation in alloys.1,2,16) The kinetics of selective oxidation is influenced not only by outer diffusion of a less-noble element from the inner side of the alloy and penetration of oxygen from atmosphere but also by nucleation and growth of oxide and so on. Therefore, it is difficult to establish a practical model to describe selective oxidation of the less-noble element based on a diffusion process of a specific element. Even if a model for the selective oxidation is proposed, it is noted that diffusion parameters of relevant elements at a given temperature depend on the alloy composition and diffusion paths such as bulk and interfacial boundaries. Such diffusion parameters have not been available for experimental limitation. In addition, it is also noted to hardly obtain information on changes in the microstructure of the interfaces between the alloy matrix and oxides during annealing in the present experiments.
3.3. Non-destructive Chemical State Analysis by X-ray Absorption SpectroscopyGrazing-exit XAS spectra using the normalized X-ray fluorescence intensity profiles of as-prepared Fe4Mn and annealed Fe4Mn at 773 K are shown in Fig. 12. In the present XAS measurements, the X-ray fluorescence intensity measured in the detector includes intensities of elastic scattered X-ray as well as the Mn K fluorescence. Here, the fluorescence intensity at the energies above the Mn K absorption edge was obtained by removing the background intensity extrapolated from the intensity measured at the energies below the absorption edge. The X-ray fluorescence intensity detected at each array is composed of different weighted contributions from layers at different depths in near-surface region of the sample. Therefore, the fluorescence intensity profiles measured at the pixel array located at lower take-off angle contain chemical information from an outer surface layer.
Grazing-exit XAS spectra of (a) as-prepared Fe4Mn, and (b) Fe4Mn annealed at 773 K. (mrad = 0.057°)
The intensity arising from deeper layer becomes higher as the take-off angle is increased. In the fluorescence intensity profiles of the as-prepared sample, all intensity profiles or XAS spectra are similar to those for body centered cubic (bcc) metal, as shown in Fig. 12(a). Therefore, these results indicate that Mn atoms occupy the lattice site in bcc Fe–Mn alloy irrespective of the depth from the surface. On the other hand, referring to the Mn K XANES spectra,21) characteristic spectral features of MnO are observed in the spectra measured at low exit angles for the annealed sample at 773 K, as shown in Fig. 12(b). These results are consistent with the XPS and XRD results shown before.
Different analytical techniques have been used for characterize surfaces layers of Fe–Mn alloys annealed under a low partial pressure of oxygen. The SEM observation showed that the surface became rough by annealing, which may be caused by formation of oxides in the Fe–Mn alloy matrix. The results by XPS and XAS showed that Mn is oxidized in the surface layers by annealing at high temperatures under a low partial pressure of oxygen. The structural results obtained by grazing incident XRD indicated that the non-stoichiometric MnO was formed during annealing. The formation and non-stoichiometry of MnO particles in the surface layers was discussed on the basis of the thermochemical properties of elements.
The authors would like to express sincere to Professor R. Inoue and Professor K. Wagatsuma for their valuable discussion about this study. They are also grateful to Mr. M. Itoh for the help for measurements of XPS analysis. The synchrotron radiation experiments were performed at the SPring-8 with the approval of the Japan Synchrotron Radiation Research Institute (JASRI). The authors would like to acknowledge Dr. T. Uruga, H. Tanida, H. Toyokara, Y. Terada and Y. Takagaki of JASRI for considerable aid in the experiment at SPring-8. This work was also supported by the foundation of Iron and Steel Institute of Japan.
