- Email: [email protected]
Valence electronic structure of β -FeSi2 single crystal investigated by photoelectron spectroscopy using synchrotron radiation
Available online at www.sciencedirect.com
Physics Procedia 11 (2011) 63–66
www.elsevier.com/locate/procedia Physics Procedia 00 (2010) 000–000 www.elsevier.com/locate/procedia
Proceedings of APAC-SILICIDE 2010
Valence Electronic Structure of E-FeSi2 Single Crystal Investigated by Photoelectron Spectroscopy Using Synchrotron Radiation K. Ogawaa*, M. Sasakib, A. Ohnishib, M. Kitaurab, H. Fujimotoc, J. Azumaa, K. Takahashia and M. Kamadaa b
a Synchrotron Light Application Center, Saga University, 1 Honjo, Saga, Saga 840-8502, Japan Department of Physics, Yamagata University, 1-4-12 Kojirakawa, Yamagata, Yamagata 990-8560, Japan c Faculty of Science, Kumamoto University, 2-39-1 Kurokami, Kumamoto, Kumamoto 860-8555, Japan
Abstract Valence electronic structure of E-FeSi2 single crystal has been investigated by photoelectron spectroscopy using synchrotron radiation. The plate-type E-FeSi2 single crystal has been grown by chemical vapor transport method. The surface contamination and oxide layer were removed by Ne+ ion sputtering. The dependence of the valence band spectra on exciting photon energy shows that the strong peak at the binging energy (EBin) of 0.5 eV and the shoulder peak at EBin = 8 eV are assigned to be Fe3d and Si3p orbitals, respectively. Upon the excitation of Fe2p3/2, resonant enhancement of the valence band was observed at EBin = 3.7 eV, which is very similar to Fe metal. This fact suggests the existence of the two-hole satellite states from Fe3d orbitals.
⃝ 2010Published Published Elsevier ©c 2010 by by Elsevier B.V.B.V. keywords: E-FeSi2; single crystal; photoelectron spectroscopy; synchrotron radiation; resonant photoemission
1. Introduction E-FeSi2 attracts much interest in terms of both basic and applied researches [1]. One of the reasons is that E-FeSi2 is a promising material for photodetector in infrared region and solar cell. This is because E-FeSi2 is a semiconductor with a bandgap of about 0.8 eV (the corresponding wave length is 1.5 Pm) [2]. Moreover, Fe and Si are very abundant on earth and are less harmful to environment. So, E-FeSi2 is called as Kankyo (environmentally friendly) semiconductor. Another reason is that iron silicides exhibit various properties dependent on phase and stoichiometry. Although the beta phase of iron disilicide is semiconducting, the alpha phase is metallic [3]. Although E-FeSi2 is non-magnetic, Fe3Si is ferromagnetic [4]. However, precise physical properties are still to be studied since growth of monocrystalline samples large enough to be subject under measurements was difficult.
* Corresponding author. Tel.: +81-(0)942-82-8062; fax: +81-(0)942-82-8052. E-mail address: [email protected].
c 2010 Published by Elsevier B.V. 1875-3892 ⃝ doi:10.1016/j.phpro.2011.01.033
64
K. Ogawa et al. / Physics Procedia 11 (2011) 63–66
Author name / Physics Procedia 00 (2010) 000–000
Therefore, most of the samples under study are thin films [2-4]. For device applications, thin films on, for example, Si substrates are favored. The quality of film such as stoichiometry, purity, crystallinity and even lattice match with substrates can alter the physical properties [5, 6]. In these terms, the investigation of high-quality single crystal is required. Recently, Hara succeeded to grow relatively large plate-type monocrystalline samples by chemical vapor transport (CVT) method whose typical size is about 1 mm wide, 6 mm long and 100 Pm thick [7]. By this method, we also grew plate-type monocrystalline samples, enabling the investigation of the physical properties. In order to clarify the physical properties of E-FeSi2 single crystal, in the present work, we shall investigate the valence electronic structure by photoelectron spectroscopy using synchrotron radiation. Photoelectron spectroscopy is a powerful method which can directly determine valence electronic structure. So far, most of the photoelectron study of E-FeSi2 was done with He lamp and only the line shapes of the observed valence band were compared with the calculated density of states (DOS) [8-12]. The agreement between experimental valence band and calculated DOS is fairly good, however the detailed information such as orbital components cannot be obtained by this method. Thus we adapted synchrotron radiation as light source. Tunable synchrotron radiation makes it possible to observe photon-energy (hv) dependence of photoemission intensity and resonant behavior of valence band upon excitation of a certain core level. The hv dependence of photoemission intensity gives the information on photoionization cross section. Moreover, the resonant behavior of valence band is a direct measure of component orbitals from specific elements [13]. By observing the resonant behavior in valence band upon exciting Fe core levels, therefore, we can evaluate the contribution of Fe to the valence band. We actually observed the resonant behavior of the valence band upon Fe2p3/2 excitation. According to these results we discuss the valence electronic structure of E-FeSi2 single crystal. 2. Experimental Monocrystalline E-FeSi2 was grown in Yamagata Univ. by CVT method [7]. The samples are plate-like in shape and their typical size was about 1 mm x 6 mm. The surface orientation is determined by back Laue diffraction to be (100). Before measurements, the sample was installed into an ultra-high-vacuum (UHV) chamber for sample preparation, and then surface oxides and contaminations were removed by Ne+ ion sputtering. Without exposing to air, the sputtered sample was transferred to another UHV chamber for photoelectron spectroscopy. The base pressure of the two UHV chambers was 2 x 10-8 Pa. The cleanliness of the sample surface was checked by photoelectron spectra taken at the photon energy (hv) of 700 and 800 eV. Ne+ ion sputtering was repeated until the photoelectron and Auger electron peaks from contaminations such as carbon and oxygen become negligibly small. Measurements of photoelectron spectroscopy were carried out at the Saga-University beamline BL13 in Saga Light Source [14]. At BL13, synchrotron radiation from a planar undulator is available and is linearly polarized in the horizontal plane. The synchrotron radiation is monochromatized by a varied-line-spacing plane-grating monochromator and impinges on the sample at the incident angle of 45º. The delivered photon energy is from 85 to 800 eV, which enables the observation of photoelectrons from not only valence band but also Fe2p. The resolving power of the monochromator is better than 10,000 when hv < 200 eV and is better than 5,000 when hv > 200 eV with the slit width of 10 Pm. The photon flux is 1 x 1010~5 x 1012 photons/s/300 mA. The size of the beam spot on sample is about 100 Pm (horizontal) x 30 Pm (vertical) which is sufficiently smaller than typical sample size. The electron spectrometer MB Scientific A-1 analyzer was used in the present experiment. The photoelectrons emitted within about ±7º from the sample normal were collected. Total energy resolution was estimated to be 18 meV at hv = 130 eV and 420 meV at hv = 700 eV determined from the Fermi edge of polycrystalline Au under typical measurement conditions. All the measurements were done at room temperature. 3. Results and discussion Figure 1 shows the photoelectron spectra from the valence band taken at hv = 130 eV and 699 eV. The intensity is normalized to the peak maximum. There is a strong peak at the binding energy (EBin) of 0.5 eV and a weak peak at 14 eV. A shoulder peak can be seen at EBin = 8 eV in the photoelectron spectrum at hv = 699 eV. The spectrum at hv = 130 eV is more peaky than that at hv = 699 eV and other features than valence band maximum are less pronounced. The change in spectral shape dependent on hv can have the information on the photoionization cross section. Yeh and Lindau calculated the photoionization cross section for each level of free atoms [15]. According to
K. Ogawa et al. / Physics Procedia 11 (2011) 63–66
Author name / Physics Procedia 00 (2010) 000–000
Normalized Intensity (arb. units)
the calculation, the cross-section ratio Si3sp/Fe3d increases by the factor of 3.5 from hv = 130 eV to 699 eV. Therefore, the stronger components at hv = 130 eV can be attributed to Fe3d orbitals and the stronger components at hv = 699 eV can be considered as Si3sp orbitals. So, the strong peak at EBin = 0.5 eV is assigned to be Fe3d orbitals and the shoulder peak at EBin = 8 eV to be Si3sp components, which fairly agrees with the calculated DOSs [10]. In the partial DOS by the self-consistent full-potential linearized augmented plane wave method, most of Fe3d + orbitals locates at EBin = 1 eV. In contrast, Si3s and Si3p Ne sputtered components cover the wider energy range from 5 eV to 1.0 surface of E- FeSi2 13.5 eV and from 1 eV to 5 eV, respectively [10]. The single crystal peak maximums are found at EBin = 8 eV and 4 eV. Other calculations also give the similar DOSs [8, 9]. So, the shoulder peak at EBin = 8 eV is assigned to be Si3s. The hQ origin of the weak peak at 14 eV is still to be determined. 699 eV The possible candidate is the 2p levels of Ne atoms which 0.5 were implanted upon Ne+ ion sputtering [16]. We believe that inactive Ne atoms have no meaningful contribution to the valence band of E-FeSi2. The spectral shape near the valence band maximum is slightly different from the spectra reported so far [11, 12]. In the photoelectron 130 eV spectrum for an epitaxial E-FeSi2 film at hv = 170 eV, for 0.0 example, the intensity at EBin = 1.5 eV is more enhanced [11] whose reason is not clear. 16 14 12 10 8 6 4 2 0 - 2 Binding Energy (eV) Fig. 1. Photoelectron spectra of valence band of E-FeSi2 single crystal at hv = 130 and 699 eV. í6
[1u10 ]
Normalized Intensity (arb. units)
5
+
Ne sputtered surface of EíFeSi2 4 single crystal
Fig. 2. Photoelectron spectra of valence band of E-FeSi2 single crystal at various photon energies as listed in the figure. The intensity is normalized to incident photon flux.
In order to observe resonant behavior of the valence band, we must measure photoelectron spectra at the photon energies near the Fe2p3/2 excitation threshold. The binding energy of Fe2p3/2 is determined to be 707.2 3 x0.15 eV from the photoelectron spectrum taken at hv = 800 hv (eV) x0.2 eV. Resonant behavior of the valence band at and below 707.2 x0.7 the Fe2p3/2 excitation threshold is clearly shown in Fig. 2 706.8 2. The photoelectron intensity is divided by photon flux 706.1 705.4 at respective photon energy. The photoelectron intensity 704.8 at EBin = 3.7 eV begins to enhance from hv = 702.3 eV 1 703.6 until 706.8 eV when the peak begins to shift toward 702.3 higher binding energy. The shifted peak is FeL3VV 701.4 0 Auger peak whose full width at half maximum is 4.6 eV. At higher photon energies, the peak continues to shift 10 5 0 Binding Energy (eV) toward higher binding energy because it stays at a constant kinetic energy. The FeL3VV Auger peak exhibits no enhancement any longer. The resonant enhancement at EBin = 3.7 eV is surprising at a glance since Si3p orbitals look to enhance resonantly at the Fe2p3/2 excitation threshold. Noting that the resonant behavior is very similar to that observed for Fe metal [17], we suppose that this behavior suggests the existence of the two-hole satellite states from Fe3d orbitals. The similar resonant behavior with Fe metal also indicates a weak hybridization between Fe3d and Si3p orbitals in this energy region. This is partially supported by the above-mentioned calculation which states that between 5 and 1 eV, Si3p orbitals
65
66
K. Ogawa et al. / Physics Procedia 11 (2011) 63–66
Author name / Physics Procedia 00 (2010) 000–000
hybridize with Fe3d orbitals forming strong bonding orbitals in the Fe–Si directions. The prominent peak at EBin = 0.5 eV outside of this strong hybridization region, therefore, behaves similar to Fe metal. 4. Conclusions Valence electronic structure of E-FeSi2 single crystal has been investigated by photoelectron spectroscopy using synchrotron radiation. We observed the photoelectron spectra of the valence band. In comparison with existing experimental and theoretical studies, the spectral shape agrees reasonably. From the dependence on exciting photon energy, the strong peak at EBin = 0.5 eV and the shoulder peak at EBin = 8 eV are assigned to be Fe3d and Si3p orbitals. Upon the excitation of Fe2p3/2, resonant enhancement of the valence band was observed at 3.7 eV, which is very similar to Fe metal. This fact indicates the existence of the two-hole satellite states from Fe3d orbitals.
Acknowledgements A part of this work was supported by Nanotechnology Net-work Project (Kyushu-area Nanotechnology Network) of the Ministry of Education, Culture, Sports, Science and Technology (MEXT), Japan.
References [1] Y. Maeda, Appl. Surf. Sci. 254 (2008) 6242. [2] D. Leong , et al., Nature 387 (1997) 686. [3] K. Kyllesbech, et al., Phys. Rev. B 50 (1994) 14 200. [4] T. Yoshitake, et al., Jpn J. Appl. Phys. 42 (2003) L849. [5] S.J. Clark, et al., Phys. Rev. B 58 (1998)10 389. [6]M. Suzuno, et al., J. Appl. Phys. 102 (2007) 103706. [7] Y. Hara, et al., Thin Solid Films 515 (2007) 8259. [8] N.E. Christensen, Phys. Rev. B 42 (1990) 7148. [9] S. Eisebitt, et al., Phys. Rev. B 50 (1994) 18 330. [10] Z.J. Pan, et al., Mat. Sci. Eng. B 131 (2006) 121. [11] F. Sirotti, et al., Phys. Rev. B 48 (1993) 8299. [12] J. Chrost, et al., Surf. Sci. 371 (1997) 297. [13] M. Kobayashi, et al., Phys. Rev. B 79 (2009) 205203. [14] K. Takahashi, et al., J. Electron Spectrosco. Relat. Phenom. 144–147 (2005) 1093. [15] J.J. Yeh and I. Lindau, Atomic Data Nuclear Data Tables 32 (1985) 1. [16] B.J. Waclawski and J.W. Gadzuk, Phys. Rev. Lett. 41 (1978) 583. [17] S. Hüfner, et al., Phys. Rev. B 61 (2000) 12 582.
Physics Procedia 11 (2011) 63–66
www.elsevier.com/locate/procedia Physics Procedia 00 (2010) 000–000 www.elsevier.com/locate/procedia
Proceedings of APAC-SILICIDE 2010
Valence Electronic Structure of E-FeSi2 Single Crystal Investigated by Photoelectron Spectroscopy Using Synchrotron Radiation K. Ogawaa*, M. Sasakib, A. Ohnishib, M. Kitaurab, H. Fujimotoc, J. Azumaa, K. Takahashia and M. Kamadaa b
a Synchrotron Light Application Center, Saga University, 1 Honjo, Saga, Saga 840-8502, Japan Department of Physics, Yamagata University, 1-4-12 Kojirakawa, Yamagata, Yamagata 990-8560, Japan c Faculty of Science, Kumamoto University, 2-39-1 Kurokami, Kumamoto, Kumamoto 860-8555, Japan
Abstract Valence electronic structure of E-FeSi2 single crystal has been investigated by photoelectron spectroscopy using synchrotron radiation. The plate-type E-FeSi2 single crystal has been grown by chemical vapor transport method. The surface contamination and oxide layer were removed by Ne+ ion sputtering. The dependence of the valence band spectra on exciting photon energy shows that the strong peak at the binging energy (EBin) of 0.5 eV and the shoulder peak at EBin = 8 eV are assigned to be Fe3d and Si3p orbitals, respectively. Upon the excitation of Fe2p3/2, resonant enhancement of the valence band was observed at EBin = 3.7 eV, which is very similar to Fe metal. This fact suggests the existence of the two-hole satellite states from Fe3d orbitals.
⃝ 2010Published Published Elsevier ©c 2010 by by Elsevier B.V.B.V. keywords: E-FeSi2; single crystal; photoelectron spectroscopy; synchrotron radiation; resonant photoemission
1. Introduction E-FeSi2 attracts much interest in terms of both basic and applied researches [1]. One of the reasons is that E-FeSi2 is a promising material for photodetector in infrared region and solar cell. This is because E-FeSi2 is a semiconductor with a bandgap of about 0.8 eV (the corresponding wave length is 1.5 Pm) [2]. Moreover, Fe and Si are very abundant on earth and are less harmful to environment. So, E-FeSi2 is called as Kankyo (environmentally friendly) semiconductor. Another reason is that iron silicides exhibit various properties dependent on phase and stoichiometry. Although the beta phase of iron disilicide is semiconducting, the alpha phase is metallic [3]. Although E-FeSi2 is non-magnetic, Fe3Si is ferromagnetic [4]. However, precise physical properties are still to be studied since growth of monocrystalline samples large enough to be subject under measurements was difficult.
* Corresponding author. Tel.: +81-(0)942-82-8062; fax: +81-(0)942-82-8052. E-mail address: [email protected].
c 2010 Published by Elsevier B.V. 1875-3892 ⃝ doi:10.1016/j.phpro.2011.01.033
64
K. Ogawa et al. / Physics Procedia 11 (2011) 63–66
Author name / Physics Procedia 00 (2010) 000–000
Therefore, most of the samples under study are thin films [2-4]. For device applications, thin films on, for example, Si substrates are favored. The quality of film such as stoichiometry, purity, crystallinity and even lattice match with substrates can alter the physical properties [5, 6]. In these terms, the investigation of high-quality single crystal is required. Recently, Hara succeeded to grow relatively large plate-type monocrystalline samples by chemical vapor transport (CVT) method whose typical size is about 1 mm wide, 6 mm long and 100 Pm thick [7]. By this method, we also grew plate-type monocrystalline samples, enabling the investigation of the physical properties. In order to clarify the physical properties of E-FeSi2 single crystal, in the present work, we shall investigate the valence electronic structure by photoelectron spectroscopy using synchrotron radiation. Photoelectron spectroscopy is a powerful method which can directly determine valence electronic structure. So far, most of the photoelectron study of E-FeSi2 was done with He lamp and only the line shapes of the observed valence band were compared with the calculated density of states (DOS) [8-12]. The agreement between experimental valence band and calculated DOS is fairly good, however the detailed information such as orbital components cannot be obtained by this method. Thus we adapted synchrotron radiation as light source. Tunable synchrotron radiation makes it possible to observe photon-energy (hv) dependence of photoemission intensity and resonant behavior of valence band upon excitation of a certain core level. The hv dependence of photoemission intensity gives the information on photoionization cross section. Moreover, the resonant behavior of valence band is a direct measure of component orbitals from specific elements [13]. By observing the resonant behavior in valence band upon exciting Fe core levels, therefore, we can evaluate the contribution of Fe to the valence band. We actually observed the resonant behavior of the valence band upon Fe2p3/2 excitation. According to these results we discuss the valence electronic structure of E-FeSi2 single crystal. 2. Experimental Monocrystalline E-FeSi2 was grown in Yamagata Univ. by CVT method [7]. The samples are plate-like in shape and their typical size was about 1 mm x 6 mm. The surface orientation is determined by back Laue diffraction to be (100). Before measurements, the sample was installed into an ultra-high-vacuum (UHV) chamber for sample preparation, and then surface oxides and contaminations were removed by Ne+ ion sputtering. Without exposing to air, the sputtered sample was transferred to another UHV chamber for photoelectron spectroscopy. The base pressure of the two UHV chambers was 2 x 10-8 Pa. The cleanliness of the sample surface was checked by photoelectron spectra taken at the photon energy (hv) of 700 and 800 eV. Ne+ ion sputtering was repeated until the photoelectron and Auger electron peaks from contaminations such as carbon and oxygen become negligibly small. Measurements of photoelectron spectroscopy were carried out at the Saga-University beamline BL13 in Saga Light Source [14]. At BL13, synchrotron radiation from a planar undulator is available and is linearly polarized in the horizontal plane. The synchrotron radiation is monochromatized by a varied-line-spacing plane-grating monochromator and impinges on the sample at the incident angle of 45º. The delivered photon energy is from 85 to 800 eV, which enables the observation of photoelectrons from not only valence band but also Fe2p. The resolving power of the monochromator is better than 10,000 when hv < 200 eV and is better than 5,000 when hv > 200 eV with the slit width of 10 Pm. The photon flux is 1 x 1010~5 x 1012 photons/s/300 mA. The size of the beam spot on sample is about 100 Pm (horizontal) x 30 Pm (vertical) which is sufficiently smaller than typical sample size. The electron spectrometer MB Scientific A-1 analyzer was used in the present experiment. The photoelectrons emitted within about ±7º from the sample normal were collected. Total energy resolution was estimated to be 18 meV at hv = 130 eV and 420 meV at hv = 700 eV determined from the Fermi edge of polycrystalline Au under typical measurement conditions. All the measurements were done at room temperature. 3. Results and discussion Figure 1 shows the photoelectron spectra from the valence band taken at hv = 130 eV and 699 eV. The intensity is normalized to the peak maximum. There is a strong peak at the binding energy (EBin) of 0.5 eV and a weak peak at 14 eV. A shoulder peak can be seen at EBin = 8 eV in the photoelectron spectrum at hv = 699 eV. The spectrum at hv = 130 eV is more peaky than that at hv = 699 eV and other features than valence band maximum are less pronounced. The change in spectral shape dependent on hv can have the information on the photoionization cross section. Yeh and Lindau calculated the photoionization cross section for each level of free atoms [15]. According to
K. Ogawa et al. / Physics Procedia 11 (2011) 63–66
Author name / Physics Procedia 00 (2010) 000–000
Normalized Intensity (arb. units)
the calculation, the cross-section ratio Si3sp/Fe3d increases by the factor of 3.5 from hv = 130 eV to 699 eV. Therefore, the stronger components at hv = 130 eV can be attributed to Fe3d orbitals and the stronger components at hv = 699 eV can be considered as Si3sp orbitals. So, the strong peak at EBin = 0.5 eV is assigned to be Fe3d orbitals and the shoulder peak at EBin = 8 eV to be Si3sp components, which fairly agrees with the calculated DOSs [10]. In the partial DOS by the self-consistent full-potential linearized augmented plane wave method, most of Fe3d + orbitals locates at EBin = 1 eV. In contrast, Si3s and Si3p Ne sputtered components cover the wider energy range from 5 eV to 1.0 surface of E- FeSi2 13.5 eV and from 1 eV to 5 eV, respectively [10]. The single crystal peak maximums are found at EBin = 8 eV and 4 eV. Other calculations also give the similar DOSs [8, 9]. So, the shoulder peak at EBin = 8 eV is assigned to be Si3s. The hQ origin of the weak peak at 14 eV is still to be determined. 699 eV The possible candidate is the 2p levels of Ne atoms which 0.5 were implanted upon Ne+ ion sputtering [16]. We believe that inactive Ne atoms have no meaningful contribution to the valence band of E-FeSi2. The spectral shape near the valence band maximum is slightly different from the spectra reported so far [11, 12]. In the photoelectron 130 eV spectrum for an epitaxial E-FeSi2 film at hv = 170 eV, for 0.0 example, the intensity at EBin = 1.5 eV is more enhanced [11] whose reason is not clear. 16 14 12 10 8 6 4 2 0 - 2 Binding Energy (eV) Fig. 1. Photoelectron spectra of valence band of E-FeSi2 single crystal at hv = 130 and 699 eV. í6
[1u10 ]
Normalized Intensity (arb. units)
5
+
Ne sputtered surface of EíFeSi2 4 single crystal
Fig. 2. Photoelectron spectra of valence band of E-FeSi2 single crystal at various photon energies as listed in the figure. The intensity is normalized to incident photon flux.
In order to observe resonant behavior of the valence band, we must measure photoelectron spectra at the photon energies near the Fe2p3/2 excitation threshold. The binding energy of Fe2p3/2 is determined to be 707.2 3 x0.15 eV from the photoelectron spectrum taken at hv = 800 hv (eV) x0.2 eV. Resonant behavior of the valence band at and below 707.2 x0.7 the Fe2p3/2 excitation threshold is clearly shown in Fig. 2 706.8 2. The photoelectron intensity is divided by photon flux 706.1 705.4 at respective photon energy. The photoelectron intensity 704.8 at EBin = 3.7 eV begins to enhance from hv = 702.3 eV 1 703.6 until 706.8 eV when the peak begins to shift toward 702.3 higher binding energy. The shifted peak is FeL3VV 701.4 0 Auger peak whose full width at half maximum is 4.6 eV. At higher photon energies, the peak continues to shift 10 5 0 Binding Energy (eV) toward higher binding energy because it stays at a constant kinetic energy. The FeL3VV Auger peak exhibits no enhancement any longer. The resonant enhancement at EBin = 3.7 eV is surprising at a glance since Si3p orbitals look to enhance resonantly at the Fe2p3/2 excitation threshold. Noting that the resonant behavior is very similar to that observed for Fe metal [17], we suppose that this behavior suggests the existence of the two-hole satellite states from Fe3d orbitals. The similar resonant behavior with Fe metal also indicates a weak hybridization between Fe3d and Si3p orbitals in this energy region. This is partially supported by the above-mentioned calculation which states that between 5 and 1 eV, Si3p orbitals
65
66
K. Ogawa et al. / Physics Procedia 11 (2011) 63–66
Author name / Physics Procedia 00 (2010) 000–000
hybridize with Fe3d orbitals forming strong bonding orbitals in the Fe–Si directions. The prominent peak at EBin = 0.5 eV outside of this strong hybridization region, therefore, behaves similar to Fe metal. 4. Conclusions Valence electronic structure of E-FeSi2 single crystal has been investigated by photoelectron spectroscopy using synchrotron radiation. We observed the photoelectron spectra of the valence band. In comparison with existing experimental and theoretical studies, the spectral shape agrees reasonably. From the dependence on exciting photon energy, the strong peak at EBin = 0.5 eV and the shoulder peak at EBin = 8 eV are assigned to be Fe3d and Si3p orbitals. Upon the excitation of Fe2p3/2, resonant enhancement of the valence band was observed at 3.7 eV, which is very similar to Fe metal. This fact indicates the existence of the two-hole satellite states from Fe3d orbitals.
Acknowledgements A part of this work was supported by Nanotechnology Net-work Project (Kyushu-area Nanotechnology Network) of the Ministry of Education, Culture, Sports, Science and Technology (MEXT), Japan.
References [1] Y. Maeda, Appl. Surf. Sci. 254 (2008) 6242. [2] D. Leong , et al., Nature 387 (1997) 686. [3] K. Kyllesbech, et al., Phys. Rev. B 50 (1994) 14 200. [4] T. Yoshitake, et al., Jpn J. Appl. Phys. 42 (2003) L849. [5] S.J. Clark, et al., Phys. Rev. B 58 (1998)10 389. [6]M. Suzuno, et al., J. Appl. Phys. 102 (2007) 103706. [7] Y. Hara, et al., Thin Solid Films 515 (2007) 8259. [8] N.E. Christensen, Phys. Rev. B 42 (1990) 7148. [9] S. Eisebitt, et al., Phys. Rev. B 50 (1994) 18 330. [10] Z.J. Pan, et al., Mat. Sci. Eng. B 131 (2006) 121. [11] F. Sirotti, et al., Phys. Rev. B 48 (1993) 8299. [12] J. Chrost, et al., Surf. Sci. 371 (1997) 297. [13] M. Kobayashi, et al., Phys. Rev. B 79 (2009) 205203. [14] K. Takahashi, et al., J. Electron Spectrosco. Relat. Phenom. 144–147 (2005) 1093. [15] J.J. Yeh and I. Lindau, Atomic Data Nuclear Data Tables 32 (1985) 1. [16] B.J. Waclawski and J.W. Gadzuk, Phys. Rev. Lett. 41 (1978) 583. [17] S. Hüfner, et al., Phys. Rev. B 61 (2000) 12 582.