- Email: [email protected]
The electronic structure of ion beam mixed ferromagnetic multilayered films
Journal of Electron Spectroscopy and Related Phenomena 114–116 (2001) 807–811 www.elsevier.nl / locate / elspec
The electronic structure of ion beam mixed ferromagnetic multilayered films a, a a b b c G.S. Chang *, S.H. Kim , K.H. Chae , E.Z. Kurmaev , V. Galakhov , A. Moewes , Y.P. Lee d , K. Jeong a , C.N. Whang a a
Atomic-Scale Surface Science Research Center and Institute of Physics and Applied Physics, Yonsei University, Seoul 120 -749, South Korea b Institute of Metal Physics, Russian Academy of Sciences — Ural Division, 620219 Yekaterinburg GSP-170, Russia c Department of Physics and Engineering Physics, University of Saskatchewan, Saskatchewan, Canada S7 N 5 E2 d Department of Physics, Hanyang University, Seoul 133 -791, South Korea Received 8 August 2000; accepted 9 September 2000
Abstract The electronic structure of ion beam mixed Co / Pt multilayered films was investigated using synchrotron-based soft X-ray fluorescence (SXF) and photoemission spectroscopy (PES). Co / Pt multilayered films (eight periods of Co and Pt sublayers) were prepared on Si(100) substrates by alternating electron-beam evaporation followed by ion beam mixing with 80 keV Ar ions under high vacuum. The SXF and PES measurements show that ion beam mixing can change the spectral features of Co 3d valence bands for PES spectra and the emission energies for Co L 3 SXF spectra. The similar valence-band characteristics of IBM CoPt 3 with pure Co and the low emission energy with respect to stable CoPt 3 alloy are correlated with the isolated Co 3d states with a weakened interatomic interaction due to the reduced coordination. 2001 Elsevier Science B.V. All rights reserved. Keywords: Ferromagnetic multilayer; Co / Pt; Ion beam mixing; Electronic structure; Soft X-ray fluorescence; Photoemission spectroscopy
1. Introduction Strong efforts in the technologies of ferromagnetism have been devoted to study the magnetic and magneto-optical properties of ferromagnetic / nonmagnetic multilayers and thin film alloys in order to develop magnetic data-storage media, sensors, and read heads [1,2]. As the size of those devices shrinks
*Corresponding author. Tel.: 182-2-2123-4280; fax: 182-2-3127090. E-mail address: [email protected] (G.S. Chang).
to sub-micron with ultrahigh storage density [3,4], the requirement for ferromagnetic materials with large magnetic moment and magnetic flux density are increasing. In the case of 3d transition metal / rare-earth metal systems, Co / Pt multilayered and Co–Pt alloy films have attracted a lot of attention due to their large magneto-optical (MO) Kerr rotation, perpendicular magnetic anisotropy, and good corrosion resistance [5–7], which makes these materials prospects for a new generation of storage devices [8]. It has been reported that magnetic and MO properties of Co / Pt multilayered and Co–Pt films correlate with spin hybridization of Co 3d and
0368-2048 / 01 / $ – see front matter 2001 Elsevier Science B.V. All rights reserved. PII: S0368-2048( 00 )00271-1
808
G.S. Chang et al. / Journal of Electron Spectroscopy and Related Phenomena 114 – 116 (2001) 807 – 811
Pt 5d states and they can be affected by their chemical structure [9,10]. In the present paper, we apply the ion beam mixing method to ferromagnetic Co / Pt films and present measurements of Co L 3 soft X-ray fluorescence (SXF) spectra and valence-band photo˚ / Pt(60 A) ˚ emission spectra (PES) of Co(15 A) multilayered films treated by ion beam mixing. Chae et al. [11] reported that ion beam mixed (IBM) multilayered films show different electronic structure from those of thermodynamically stable alloys. It also has been suggested that IBM multilayered films show a metastable disordered alloy phase, which is similar to that of a rapidly quenched alloy. Therefore, it is expected that Co 3d and Pt 5d valenceband characteristics are modified by ion beam mixing, which leads the change in magnetic properties of ferromagnetic thin film alloys. That is why a study of electronic structure of IBM multilayered films is of great interest.
2. Experimental Co / Pt multilayered films with eight periods of Co and Pt sublayers were prepared by alternating electron-beam evaporation on Si(100) substrates. The sublayer thickness of Co and Pt layers was chosen to ˚ respectively, to match the CoPt 3 15 and 60 A, composition. Ion beam mixing was carried out by using 80 keV Ar 1 beam at room temperature [12]. Ion dose and beam current were fixed at 1.0310 16 ions / cm 2 and 1.5 mA, respectively. The electronic valence-band structure was measured by PES at the VUV beamline at Pohang Light Source, South Korea. To obtain information about the local density of state on Co and Pt, the measurements were performed at the incident photon energies of 100 and 50 eV so that the contributions of Pt 5d and Co 3d states to corresponding photoemission spectra are minimized, respectively [13]. The spectra of IBM CoPt 3 films were compared with those of pure Co, Pt, and stable CoPt 3 alloy films prepared by electronbeam evaporation. In addition, soft X-ray fluorescence measurements of Co L 3 (valence 3d4s→2p 3 / 2 transition) were performed at the Advanced Light Source on beamline 8.0, which is described in detail
elsewhere [14]. The fluorescent radiation emitted from the sample was energy-analyzed by a highresolution grating spectrometer. The band width of the monochromator was set to 1.0 eV for photon energies between 768 and 840 eV, respectively. The Co L 3 X-ray emission spectra were recorded using a spherical grating fluorescence spectrometer (1500 lines / mm, 10 m radius). The spectrometer resolution in the first order of diffraction was 0.9 eV for Co L 3 . The MO properties were measured by using a polar MO spectroscopy with a Xe lamp (150 W: Hamamatsu, L2175) in a saturating field of 10 kOe at room temperature. The rotation angle was measured using an InGaAs photocathode (Hamamatsu, R2658).
3. Results and discussion Fig. 1 shows valence-band PES spectra of IBM CoPt 3. Since we are interested in the behavior of ferromagnetic Co 3d states in IBM CoPt 3 sample, we chose the incident photon energies of 50 eV (Fig. 1a) that the relative ratio of Co 3d photoionization cross section to Pt 5d is minimized to distinguish the local density of states of Co 3d from that of Pt 5d. In case of 100 eV photon energy, Co 3d characteristics is much stronger than those at 50 eV, because the photoionization cross-section ratio of Co 3d to Pt 5d at a photon energy of 100 eV is a factor of 8.3 larger than the Pt 5d to Co 3d ratio at 50 eV [15]. As seen in Fig. 1a, a valley at about 3 eV (arrow A) is a typical feature of the Pt 5d band [16], while the flattened spectral weight at the Fermi level (EF ) is attributed to the photon energy-dependent cross-section effects at lower photon energies [17,18]. In addition, the Pt 5d characteristic feature shifts slightly toward higher binding energy from pure Pt to CoPt 3 alloy. This reflects that the covalent interaction of Pt atoms changes under a variation of nearest neighbors surrounding the Pt site. Uba et al. [9] have reported that the relative positions of Co and Pt d bands affect the Co 3d–Pt 5d hybridization and the valence-band width is increased in the Co–Pt alloys [9]. As indicated by Fig. 1b, the energy shift is more significantly observed in the spectra dominated by Co 3d states, even though the spectral weight from
G.S. Chang et al. / Journal of Electron Spectroscopy and Related Phenomena 114 – 116 (2001) 807 – 811
Fig. 1. Valence-band PES spectra of ion beam mixed CoPt 3 films taken at photon energies of 50 (a) and 100 eV (b). A superposition in a ratio of 0.25:0.75 of the spectra of Pure Co and Pt was added in (b) for comparison.
Pt states is very weak due to a small photoionization cross section of Pt 5d with respect to Co 3d states. However, the interesting feature is the contribution of Pt 5d to the spectrum of IBM CoPt 3 (arrow B and C), which does not resemble that of stable CoPt 3 alloy film but the superposed spectrum of pure Co and Pt with a factor of 0.25 and 0.75, respectively, which reflects that the Co 3d and Pt 5d electrons are in an isolated state of greatly weakened interatomic electronic interaction. During the ion beam mixing treatment, the independent motions of atoms activated by ion bombardment are quenched in a very
809
short period of the order of 10 211 s at the end of ion-solid interaction [19]. Therefore, the rapid quenching process can induce the formation of a metastable CoPt 3 phase with expanded volume and reduced coordination, which does not allow the electronic interaction between Co and Pt atom to be relaxed to an equilibrium state. To clarify this implication, we measured the distribution of Co 3d states using soft X-ray fluorescence, where the dipole selection rule gives a possibility to extract information about the partial density of Co 3d states from Pt 5d. Acker et al. [17] reported that, in case of Fe–Pt and Co–Pt alloy phase, Fe and Co majority-spin states located at higher binding energy have a strong tendency to mix with Pt 5d states, while much Fe and Co minorityspin characters positioned near and above EF as an unoccupied virtual bound state are driven to move higher above EF under the d resonance with Pt minority-spin states [17]. This means that, in case of soft X-ray fluorescence measurements of Co L 3 , the emission energy of stable CoPt 3 alloy film from valence 3d4s→2p 3 / 2 transition is higher than that of pure Co. Otherwise, if Co 3d electrons in IBM CoPt 3 have a greatly weakened interatomic electronic interaction with Pt 5d states, Co L 3 X-ray emission energy must be reduced with respect to stable CoPt 3 alloy film and resemble the 3d states of pure Co. Fig. 2 shows the Co L 3 SXF spectra of the investigated samples. These spectra correspond to dipole 3d4s→2p 3 / 2 transition (when the photon-induced vacancies in Co 2p 3 / 2 levels are filled by Co 3d and 4s electrons) mostly probing Co 3d states in the sample. As seen in Fig. 2, the spectral maximum of Co L 3 (arrow A) shifts by about 1.6 eV from pure Co (780.8 eV) to the stable CoPt 3 alloy (782.4 eV). This shift may be induced by corresponding shift of Co 2p 3 / 2 core level in CoPt 3 alloy with respect to pure Co and / or by the redistribution of Co 3d states as a result of Co 3d–Pt 5d hybridization. To check these suppositions, we also performed X-ray photoelectron spectroscopy (XPS) measurements of Co 2p levels of pure Co, stable CoPt 3 alloy, and IBM CoPt 3 films. Fig. 3 shows the obtained XPS spectra of the samples. It can be seen that no energy shift of Co 2p 3 / 2 and Co 2p 1 / 2 levels is observed for all the investigated samples. This means that one can attribute the observed changes in Co L 3 SXF spectrum of
810
G.S. Chang et al. / Journal of Electron Spectroscopy and Related Phenomena 114 – 116 (2001) 807 – 811
˚ / Pt(60 A) ˚ multilayer, ion Fig. 2. Co L 3 SXF spectra of Co(15 A) beam mixed CoPt 3 . The spectra of Pure Co and CoPt 3 alloy were added for comparison.
stable CoPt 3 alloy film with respect to that of pure Co to the redistribution of Co 3d states. It is natural that the spectrum of Co / Pt multilayered film before ion beam mixing is similar with that of pure Co, because the interfacial Co 3d–Pt 5d hybridization is overwhelmed by the metallic interaction between Co atoms in Co sublayer. Otherwise, IBM CoPt 3 shows the same trend as in the PES measurements. The spectral maximum of IBM CoPt 3 (781.1 eV) is close to that of pure Co with respect to stable CoPt 3 alloy film, even though the sample was ion-beam-mixed to form a CoPt 3 phase. As discussed above, this is a direct evidence of isolated Co 3d states and, therefore, lower Co coordination in IBM CoPt 3 film by the formation of metastable alloy phase than that of a stable one which is in thermodynamically equilibrium state. Finally, we carried out MO Kerr spectroscopy measurement as a representative MO property of Co–Pt alloy samples. Fig. 4 shows the saturated MO Kerr rotation angles (uK ) of the investigated samples.
Fig. 3. Co 2p 3 / 2 and 2p 1 / 2 XPS spectra of pure Co, CoPt 3 alloy, and ion beam mixed CoPt 3 film.
˚ / Pt(60 A) ˚ multilayer, Fig. 4. Polar MOKE spectra of Co(15 A) ion beam mixed CoPt 3 . The inset is MOKE hysteresis loops of corresponding samples taken at the wavelength of 450 nm.
G.S. Chang et al. / Journal of Electron Spectroscopy and Related Phenomena 114 – 116 (2001) 807 – 811
˚ / Pt(60 A) ˚ multilayered The small uK of Co(15 A) film come from pure ferromagnetic Co sublayers, because the enhancement of uK of Co / Pt multilayered structures can be realized only with a ˚ [20]. Otherthickness of Co sublayer below 10 A wise, Weller et al. [21] reported that the large uK of Co–Pt alloy films is correlated with strong hybridization between Co 3d and Pt 5d electrons. However, the uK of IBM CoPt 3 film increases dramatically over the wavelength range displayed and the values are comparable or larger than those of Co / Pt multilayers and Co–Pt alloy films at corresponding composition and wavelength [5,22]. This may seem to be contradictory to the results of PES and SXF measurements, which suggests that IBM CoPt 3 has a weak spin-hybridization feature. Bland et al. [23] reported that the magnetic atoms with reduced coordination such as a disordered structure or a ferromagnetic monolayer on non-magnetic substrate have the enlarged spin-splitting and the enhanced magnetic moment per Co atom under the transition of 3d electrons from spin-minority band to spinmajority one [23]. Therefore, the comparable uK of IBM CoPt 3 reflects that the reduced coordination effect compensates the weak spin-hybridization character.
4. Conclusion The electronic structure of ion-beam-mixed ferromagnetic Co / Pt multilayered films was investigated using synchrotron-based photoemission spectroscopy and soft X-ray fluorescence spectroscopy. The valence-band characteristics of IBM CoPt 3 resemble that of pure Co and the emission energy from Co 3d4s→2p 3 / 2 transition is reduced with respect to stable CoPt 3 alloy. The overall results are correlated with the isolated Co 3d states with a weakened interatomic interaction due to the reduced coordination.
Acknowledgements This work was supported by the Brain Korea 21 Project and the Korea Science and Engineering Foundation through Grant No. 995-0200-003-2.
811
References [1] M. Kaneko, J. Magn. Magn. Mater. 148 (1995) 351. [2] W.B. Zeper, F.J.A.M. Greidanus, P.F. Carcia, C.R. Fincher, J. Appl. Phys. 65 (1989) 4971. [3] R.L. White, J. Magn. Magn. Mater. 200 (1999) 616. [4] K. O’Grady, H. Laidler, J. Magn. Magn. Mater. 209 (2000) 1. ¨ H. Brandle, ¨ [5] D. Weller, W. Reim, K. Sporl, J. Magn. Magn. Mater. 93 (1991) 183. [6] C.-J. Lin, G.L. Gorman, C.H. Lee, R.F.C. Farrow, E.E. Marinero, H.V. Do, H. Notarys, C.J. Chien, J. Magn. Magn. Mater. 93 (1991) 194. [7] Y. Ochiai, S. Hashimoto, K. Aso, IEEE Trans. Magn. 25 (1989) 3755. [8] K. Sato, H. Ikekame, Y. Tosaka, K. Tsuzukiyama, I. Togami, M. Fujisawa, J. Magn. Magn. Mater. 126 (1993) 572. [9] S. Uba, L. Uba, A.N. Yaresko, A.Y. Perlov, V.N. Antonov, R. Gontarz, Phys. Rev. B 53 (1996) 6526. [10] M. Lairson, B.M. Clemens, Appl. Phys. Lett. 63 (1993) 1438. [11] K.H. Chae, Y.S. Lee, S.M. Jung, Y. Jeon, M. Croft, C.N. Whang, Nucl. Instrum. Methods B 106 (1995) 60. [12] G.S. Chang, K.H. Chae, C.N. Whang, Y.P. Lee, J. Magn. Soc. Jpn. 23 (1999) 646. [13] J.J. Yeh, I. Lindau, Atom. Data Nucl. Data Tables 32 (1985) 1. [14] J.J. Jia, T.A. Callcott, J. Yurkas, A.W. Ellis, F.J. Himpsel, ¨ M.G. Samant, J. Stohr, D.L. Ederer, Rev. Sci. Instrum. 66 (1995) 1394. [15] G.S. Chang, C.N. Whang, J.Y. Rhee, Y.P. Lee, J. Appl. Phys. 84 (2000) 1775. ¨ [16] N.V. Smith, G.K. Wertheim, S. Hufner, M.M. Traum, Phys. Rev. B 10 (1974) 3197. [17] J.F. van Acker, P.J.W. Weijs, J.C. Fuggle, K. Horn, H. Haak, K.H.J. Buschow, Phys. Rev. B 43 (1991) 8903. ¨ G. Apai, F.R. McFeely, R.S. Williams, [18] P.S. Wehner, J. Stohr, D.A. Shirley, Phys. Rev. B 14 (1976) 2411. [19] W.L. Johnson, Y.Y. Cheng, M. van Rossum, M.-A. Nocolet, Nucl. Instrum. Methods B 7–8 (1985) 657. [20] S. Hashimoto, Y. Ochiai, K. Aso, J. Appl. Phys. 66 (1989) 4909. [21] D. Weller, H. Brandle, G. Gorman, C.-J. Lin, H. Notarys, Appl. Phys. Lett. 61 (1992) 2726. [22] W.B. Zeper, F.J.A.M. Greidanus, P.F. Carcia, C.R. Fincher, J. Appl. Phys. 65 (1989) 4971. [23] J.A.C. Bland, B. Heinrich (Eds.), Ultrathin Magnetic Structures, Vol. 1, Springer, New York, 1994, Chapter 2.
The electronic structure of ion beam mixed ferromagnetic multilayered films a, a a b b c G.S. Chang *, S.H. Kim , K.H. Chae , E.Z. Kurmaev , V. Galakhov , A. Moewes , Y.P. Lee d , K. Jeong a , C.N. Whang a a
Atomic-Scale Surface Science Research Center and Institute of Physics and Applied Physics, Yonsei University, Seoul 120 -749, South Korea b Institute of Metal Physics, Russian Academy of Sciences — Ural Division, 620219 Yekaterinburg GSP-170, Russia c Department of Physics and Engineering Physics, University of Saskatchewan, Saskatchewan, Canada S7 N 5 E2 d Department of Physics, Hanyang University, Seoul 133 -791, South Korea Received 8 August 2000; accepted 9 September 2000
Abstract The electronic structure of ion beam mixed Co / Pt multilayered films was investigated using synchrotron-based soft X-ray fluorescence (SXF) and photoemission spectroscopy (PES). Co / Pt multilayered films (eight periods of Co and Pt sublayers) were prepared on Si(100) substrates by alternating electron-beam evaporation followed by ion beam mixing with 80 keV Ar ions under high vacuum. The SXF and PES measurements show that ion beam mixing can change the spectral features of Co 3d valence bands for PES spectra and the emission energies for Co L 3 SXF spectra. The similar valence-band characteristics of IBM CoPt 3 with pure Co and the low emission energy with respect to stable CoPt 3 alloy are correlated with the isolated Co 3d states with a weakened interatomic interaction due to the reduced coordination. 2001 Elsevier Science B.V. All rights reserved. Keywords: Ferromagnetic multilayer; Co / Pt; Ion beam mixing; Electronic structure; Soft X-ray fluorescence; Photoemission spectroscopy
1. Introduction Strong efforts in the technologies of ferromagnetism have been devoted to study the magnetic and magneto-optical properties of ferromagnetic / nonmagnetic multilayers and thin film alloys in order to develop magnetic data-storage media, sensors, and read heads [1,2]. As the size of those devices shrinks
*Corresponding author. Tel.: 182-2-2123-4280; fax: 182-2-3127090. E-mail address: [email protected] (G.S. Chang).
to sub-micron with ultrahigh storage density [3,4], the requirement for ferromagnetic materials with large magnetic moment and magnetic flux density are increasing. In the case of 3d transition metal / rare-earth metal systems, Co / Pt multilayered and Co–Pt alloy films have attracted a lot of attention due to their large magneto-optical (MO) Kerr rotation, perpendicular magnetic anisotropy, and good corrosion resistance [5–7], which makes these materials prospects for a new generation of storage devices [8]. It has been reported that magnetic and MO properties of Co / Pt multilayered and Co–Pt films correlate with spin hybridization of Co 3d and
0368-2048 / 01 / $ – see front matter 2001 Elsevier Science B.V. All rights reserved. PII: S0368-2048( 00 )00271-1
808
G.S. Chang et al. / Journal of Electron Spectroscopy and Related Phenomena 114 – 116 (2001) 807 – 811
Pt 5d states and they can be affected by their chemical structure [9,10]. In the present paper, we apply the ion beam mixing method to ferromagnetic Co / Pt films and present measurements of Co L 3 soft X-ray fluorescence (SXF) spectra and valence-band photo˚ / Pt(60 A) ˚ emission spectra (PES) of Co(15 A) multilayered films treated by ion beam mixing. Chae et al. [11] reported that ion beam mixed (IBM) multilayered films show different electronic structure from those of thermodynamically stable alloys. It also has been suggested that IBM multilayered films show a metastable disordered alloy phase, which is similar to that of a rapidly quenched alloy. Therefore, it is expected that Co 3d and Pt 5d valenceband characteristics are modified by ion beam mixing, which leads the change in magnetic properties of ferromagnetic thin film alloys. That is why a study of electronic structure of IBM multilayered films is of great interest.
2. Experimental Co / Pt multilayered films with eight periods of Co and Pt sublayers were prepared by alternating electron-beam evaporation on Si(100) substrates. The sublayer thickness of Co and Pt layers was chosen to ˚ respectively, to match the CoPt 3 15 and 60 A, composition. Ion beam mixing was carried out by using 80 keV Ar 1 beam at room temperature [12]. Ion dose and beam current were fixed at 1.0310 16 ions / cm 2 and 1.5 mA, respectively. The electronic valence-band structure was measured by PES at the VUV beamline at Pohang Light Source, South Korea. To obtain information about the local density of state on Co and Pt, the measurements were performed at the incident photon energies of 100 and 50 eV so that the contributions of Pt 5d and Co 3d states to corresponding photoemission spectra are minimized, respectively [13]. The spectra of IBM CoPt 3 films were compared with those of pure Co, Pt, and stable CoPt 3 alloy films prepared by electronbeam evaporation. In addition, soft X-ray fluorescence measurements of Co L 3 (valence 3d4s→2p 3 / 2 transition) were performed at the Advanced Light Source on beamline 8.0, which is described in detail
elsewhere [14]. The fluorescent radiation emitted from the sample was energy-analyzed by a highresolution grating spectrometer. The band width of the monochromator was set to 1.0 eV for photon energies between 768 and 840 eV, respectively. The Co L 3 X-ray emission spectra were recorded using a spherical grating fluorescence spectrometer (1500 lines / mm, 10 m radius). The spectrometer resolution in the first order of diffraction was 0.9 eV for Co L 3 . The MO properties were measured by using a polar MO spectroscopy with a Xe lamp (150 W: Hamamatsu, L2175) in a saturating field of 10 kOe at room temperature. The rotation angle was measured using an InGaAs photocathode (Hamamatsu, R2658).
3. Results and discussion Fig. 1 shows valence-band PES spectra of IBM CoPt 3. Since we are interested in the behavior of ferromagnetic Co 3d states in IBM CoPt 3 sample, we chose the incident photon energies of 50 eV (Fig. 1a) that the relative ratio of Co 3d photoionization cross section to Pt 5d is minimized to distinguish the local density of states of Co 3d from that of Pt 5d. In case of 100 eV photon energy, Co 3d characteristics is much stronger than those at 50 eV, because the photoionization cross-section ratio of Co 3d to Pt 5d at a photon energy of 100 eV is a factor of 8.3 larger than the Pt 5d to Co 3d ratio at 50 eV [15]. As seen in Fig. 1a, a valley at about 3 eV (arrow A) is a typical feature of the Pt 5d band [16], while the flattened spectral weight at the Fermi level (EF ) is attributed to the photon energy-dependent cross-section effects at lower photon energies [17,18]. In addition, the Pt 5d characteristic feature shifts slightly toward higher binding energy from pure Pt to CoPt 3 alloy. This reflects that the covalent interaction of Pt atoms changes under a variation of nearest neighbors surrounding the Pt site. Uba et al. [9] have reported that the relative positions of Co and Pt d bands affect the Co 3d–Pt 5d hybridization and the valence-band width is increased in the Co–Pt alloys [9]. As indicated by Fig. 1b, the energy shift is more significantly observed in the spectra dominated by Co 3d states, even though the spectral weight from
G.S. Chang et al. / Journal of Electron Spectroscopy and Related Phenomena 114 – 116 (2001) 807 – 811
Fig. 1. Valence-band PES spectra of ion beam mixed CoPt 3 films taken at photon energies of 50 (a) and 100 eV (b). A superposition in a ratio of 0.25:0.75 of the spectra of Pure Co and Pt was added in (b) for comparison.
Pt states is very weak due to a small photoionization cross section of Pt 5d with respect to Co 3d states. However, the interesting feature is the contribution of Pt 5d to the spectrum of IBM CoPt 3 (arrow B and C), which does not resemble that of stable CoPt 3 alloy film but the superposed spectrum of pure Co and Pt with a factor of 0.25 and 0.75, respectively, which reflects that the Co 3d and Pt 5d electrons are in an isolated state of greatly weakened interatomic electronic interaction. During the ion beam mixing treatment, the independent motions of atoms activated by ion bombardment are quenched in a very
809
short period of the order of 10 211 s at the end of ion-solid interaction [19]. Therefore, the rapid quenching process can induce the formation of a metastable CoPt 3 phase with expanded volume and reduced coordination, which does not allow the electronic interaction between Co and Pt atom to be relaxed to an equilibrium state. To clarify this implication, we measured the distribution of Co 3d states using soft X-ray fluorescence, where the dipole selection rule gives a possibility to extract information about the partial density of Co 3d states from Pt 5d. Acker et al. [17] reported that, in case of Fe–Pt and Co–Pt alloy phase, Fe and Co majority-spin states located at higher binding energy have a strong tendency to mix with Pt 5d states, while much Fe and Co minorityspin characters positioned near and above EF as an unoccupied virtual bound state are driven to move higher above EF under the d resonance with Pt minority-spin states [17]. This means that, in case of soft X-ray fluorescence measurements of Co L 3 , the emission energy of stable CoPt 3 alloy film from valence 3d4s→2p 3 / 2 transition is higher than that of pure Co. Otherwise, if Co 3d electrons in IBM CoPt 3 have a greatly weakened interatomic electronic interaction with Pt 5d states, Co L 3 X-ray emission energy must be reduced with respect to stable CoPt 3 alloy film and resemble the 3d states of pure Co. Fig. 2 shows the Co L 3 SXF spectra of the investigated samples. These spectra correspond to dipole 3d4s→2p 3 / 2 transition (when the photon-induced vacancies in Co 2p 3 / 2 levels are filled by Co 3d and 4s electrons) mostly probing Co 3d states in the sample. As seen in Fig. 2, the spectral maximum of Co L 3 (arrow A) shifts by about 1.6 eV from pure Co (780.8 eV) to the stable CoPt 3 alloy (782.4 eV). This shift may be induced by corresponding shift of Co 2p 3 / 2 core level in CoPt 3 alloy with respect to pure Co and / or by the redistribution of Co 3d states as a result of Co 3d–Pt 5d hybridization. To check these suppositions, we also performed X-ray photoelectron spectroscopy (XPS) measurements of Co 2p levels of pure Co, stable CoPt 3 alloy, and IBM CoPt 3 films. Fig. 3 shows the obtained XPS spectra of the samples. It can be seen that no energy shift of Co 2p 3 / 2 and Co 2p 1 / 2 levels is observed for all the investigated samples. This means that one can attribute the observed changes in Co L 3 SXF spectrum of
810
G.S. Chang et al. / Journal of Electron Spectroscopy and Related Phenomena 114 – 116 (2001) 807 – 811
˚ / Pt(60 A) ˚ multilayer, ion Fig. 2. Co L 3 SXF spectra of Co(15 A) beam mixed CoPt 3 . The spectra of Pure Co and CoPt 3 alloy were added for comparison.
stable CoPt 3 alloy film with respect to that of pure Co to the redistribution of Co 3d states. It is natural that the spectrum of Co / Pt multilayered film before ion beam mixing is similar with that of pure Co, because the interfacial Co 3d–Pt 5d hybridization is overwhelmed by the metallic interaction between Co atoms in Co sublayer. Otherwise, IBM CoPt 3 shows the same trend as in the PES measurements. The spectral maximum of IBM CoPt 3 (781.1 eV) is close to that of pure Co with respect to stable CoPt 3 alloy film, even though the sample was ion-beam-mixed to form a CoPt 3 phase. As discussed above, this is a direct evidence of isolated Co 3d states and, therefore, lower Co coordination in IBM CoPt 3 film by the formation of metastable alloy phase than that of a stable one which is in thermodynamically equilibrium state. Finally, we carried out MO Kerr spectroscopy measurement as a representative MO property of Co–Pt alloy samples. Fig. 4 shows the saturated MO Kerr rotation angles (uK ) of the investigated samples.
Fig. 3. Co 2p 3 / 2 and 2p 1 / 2 XPS spectra of pure Co, CoPt 3 alloy, and ion beam mixed CoPt 3 film.
˚ / Pt(60 A) ˚ multilayer, Fig. 4. Polar MOKE spectra of Co(15 A) ion beam mixed CoPt 3 . The inset is MOKE hysteresis loops of corresponding samples taken at the wavelength of 450 nm.
G.S. Chang et al. / Journal of Electron Spectroscopy and Related Phenomena 114 – 116 (2001) 807 – 811
˚ / Pt(60 A) ˚ multilayered The small uK of Co(15 A) film come from pure ferromagnetic Co sublayers, because the enhancement of uK of Co / Pt multilayered structures can be realized only with a ˚ [20]. Otherthickness of Co sublayer below 10 A wise, Weller et al. [21] reported that the large uK of Co–Pt alloy films is correlated with strong hybridization between Co 3d and Pt 5d electrons. However, the uK of IBM CoPt 3 film increases dramatically over the wavelength range displayed and the values are comparable or larger than those of Co / Pt multilayers and Co–Pt alloy films at corresponding composition and wavelength [5,22]. This may seem to be contradictory to the results of PES and SXF measurements, which suggests that IBM CoPt 3 has a weak spin-hybridization feature. Bland et al. [23] reported that the magnetic atoms with reduced coordination such as a disordered structure or a ferromagnetic monolayer on non-magnetic substrate have the enlarged spin-splitting and the enhanced magnetic moment per Co atom under the transition of 3d electrons from spin-minority band to spinmajority one [23]. Therefore, the comparable uK of IBM CoPt 3 reflects that the reduced coordination effect compensates the weak spin-hybridization character.
4. Conclusion The electronic structure of ion-beam-mixed ferromagnetic Co / Pt multilayered films was investigated using synchrotron-based photoemission spectroscopy and soft X-ray fluorescence spectroscopy. The valence-band characteristics of IBM CoPt 3 resemble that of pure Co and the emission energy from Co 3d4s→2p 3 / 2 transition is reduced with respect to stable CoPt 3 alloy. The overall results are correlated with the isolated Co 3d states with a weakened interatomic interaction due to the reduced coordination.
Acknowledgements This work was supported by the Brain Korea 21 Project and the Korea Science and Engineering Foundation through Grant No. 995-0200-003-2.
811
References [1] M. Kaneko, J. Magn. Magn. Mater. 148 (1995) 351. [2] W.B. Zeper, F.J.A.M. Greidanus, P.F. Carcia, C.R. Fincher, J. Appl. Phys. 65 (1989) 4971. [3] R.L. White, J. Magn. Magn. Mater. 200 (1999) 616. [4] K. O’Grady, H. Laidler, J. Magn. Magn. Mater. 209 (2000) 1. ¨ H. Brandle, ¨ [5] D. Weller, W. Reim, K. Sporl, J. Magn. Magn. Mater. 93 (1991) 183. [6] C.-J. Lin, G.L. Gorman, C.H. Lee, R.F.C. Farrow, E.E. Marinero, H.V. Do, H. Notarys, C.J. Chien, J. Magn. Magn. Mater. 93 (1991) 194. [7] Y. Ochiai, S. Hashimoto, K. Aso, IEEE Trans. Magn. 25 (1989) 3755. [8] K. Sato, H. Ikekame, Y. Tosaka, K. Tsuzukiyama, I. Togami, M. Fujisawa, J. Magn. Magn. Mater. 126 (1993) 572. [9] S. Uba, L. Uba, A.N. Yaresko, A.Y. Perlov, V.N. Antonov, R. Gontarz, Phys. Rev. B 53 (1996) 6526. [10] M. Lairson, B.M. Clemens, Appl. Phys. Lett. 63 (1993) 1438. [11] K.H. Chae, Y.S. Lee, S.M. Jung, Y. Jeon, M. Croft, C.N. Whang, Nucl. Instrum. Methods B 106 (1995) 60. [12] G.S. Chang, K.H. Chae, C.N. Whang, Y.P. Lee, J. Magn. Soc. Jpn. 23 (1999) 646. [13] J.J. Yeh, I. Lindau, Atom. Data Nucl. Data Tables 32 (1985) 1. [14] J.J. Jia, T.A. Callcott, J. Yurkas, A.W. Ellis, F.J. Himpsel, ¨ M.G. Samant, J. Stohr, D.L. Ederer, Rev. Sci. Instrum. 66 (1995) 1394. [15] G.S. Chang, C.N. Whang, J.Y. Rhee, Y.P. Lee, J. Appl. Phys. 84 (2000) 1775. ¨ [16] N.V. Smith, G.K. Wertheim, S. Hufner, M.M. Traum, Phys. Rev. B 10 (1974) 3197. [17] J.F. van Acker, P.J.W. Weijs, J.C. Fuggle, K. Horn, H. Haak, K.H.J. Buschow, Phys. Rev. B 43 (1991) 8903. ¨ G. Apai, F.R. McFeely, R.S. Williams, [18] P.S. Wehner, J. Stohr, D.A. Shirley, Phys. Rev. B 14 (1976) 2411. [19] W.L. Johnson, Y.Y. Cheng, M. van Rossum, M.-A. Nocolet, Nucl. Instrum. Methods B 7–8 (1985) 657. [20] S. Hashimoto, Y. Ochiai, K. Aso, J. Appl. Phys. 66 (1989) 4909. [21] D. Weller, H. Brandle, G. Gorman, C.-J. Lin, H. Notarys, Appl. Phys. Lett. 61 (1992) 2726. [22] W.B. Zeper, F.J.A.M. Greidanus, P.F. Carcia, C.R. Fincher, J. Appl. Phys. 65 (1989) 4971. [23] J.A.C. Bland, B. Heinrich (Eds.), Ultrathin Magnetic Structures, Vol. 1, Springer, New York, 1994, Chapter 2.