- Email: [email protected]
Characterization of magnetic tunnel junctions using IETS
Journal of Magnetism and Magnetic Materials 198}199 (1999) 152}154
Characterization of magnetic tunnel junctions using IETS R.J.M. van de Veerdonk *, J.S. Moodera, W.J.M. de Jonge Department of Applied Physics and Research School COBRA, Eindhoven University of Technology, Den Dolech 2, 5600 MB Eindhoven, Netherlands Francis Bitter Magnet Laboratory, Massachusetts Institute of Technology, 170 Albany Street, Cambridge, MA 02139, USA
Abstract In this paper inelastic electron tunneling spectroscopy measurements are presented on tunnel junctions containing both magnetic and nonmagnetic electrodes. Magnon excitations have been found near zero voltage; phonons are observed between 30 and 110 mV. Therefore, a model description of the transport properties of magnetic tunnel junctions at elevated temperatures or non-zero bias voltage should include both inelastic contributions. 1999 Elsevier Science B.V. All rights reserved. Keywords: Tunneling; Magnetoresistance; Inelastic electron tunneling spectroscopy
Tunnel junctions containing two ferromagnetic electrodes separated by an insulator show a sizable junction magnetoresistance (JMR) e!ect depending on the relative orientation of the electrode magnetizations [1]. The magnitude of the JMR at zero temperature and zero voltage is determined by the polarization of the tunneling electrons and can be understood by Julliere's model [2}4]. However, the decrease of the JMR at elevated temperatures and non-zero bias voltage is not yet understood. It has been suggested that impurity, phonon, and magnon assisted tunneling processes contribute to the decreasing JMR [5}7]. We have performed a detailed study on the temperature dependence of the transport properties of tunnel junctions [8,9], with the goal to provide an experimental basis for an improved model description. In this paper inelastic electron tunneling spectroscopy (IETS) [10] is used to show that inelastic phonon and magnon assisted contributions to the tunnel current are signi"cant. The junctions used in this study have been fabricated using shadow evaporation onto glass substrates, similar to the method described in Ref. [1]. This results in
* Corresponding author. Tel.: #31-40-274-4089; fax: #3140-274-4282; e-mail: [email protected].
a cross-strip geometry junction with an area of &200;300 lm. The tunnel barriers are formed as a native oxide/nitride on 10 nm-thick thermally evaporated Al strips using an in-situ O or N DC glow dis charge. The 20 nm-thick counterelectrodes were e-gun evaporated Co, or thermally evaporated Ni Fe , Cu, Al, or Au strips. The resistance of the studied junctions was of the order of a few k). In Figs. 1 and 2 the di!erential conductance (G" dI/d<) and the inelastic electron tunneling spectra, i.e., the derivative of the di!erential conductance versus voltage (dG/d<), of some of the tunnel junctions are shown. The results for the Al O (Fig. 1) and AlN (Fig. 2) barriers are very similar. In Fig. 1a and Fig. 2a one of the electrodes is magnetic, while in Fig. 1b and Fig. 2b both electrodes are nonmagnetic. In all cases no "ne-structure is observed in the IETS measurements at high voltages, where contributions from water and carbon contaminations would show up [10]. This indicates that the barrier preparation process is very clean. A peak around <"115 mV dominates the IETS measurements, which has been identi"ed as an Al}O stretching mode for the Al O barrier [10] and as a TO phonon mode for the AlN barrier [11]. This large inelastic peak shows up as a clearly visible knee in the G}< curves, which becomes even better resolved when the sample is
0304-8853/99/$ } see front matter 1999 Elsevier Science B.V. All rights reserved. PII: S 0 3 0 4 - 8 8 5 3 ( 9 8 ) 0 1 0 5 5 - 5
R.J.M. van de Veerdonk et al. / Journal of Magnetism and Magnetic Materials 198}199 (1999) 152}154
Fig. 1. Conductance and IETS measurements at ¹"4.2 K of (a) an Al}Al O }Ni Fe junction (Co showed similar results) and (b) an Al}Al O }Au junction (Cu and Al showed similar results). The AC voltage excitation was &1 mV for the G}< \ and &10 mV for the IETS measurements. The vertical \ dashed lines indicate 115 mV. The dotted line in (b) illustrates the elastic part of the IETS measurement.
cooled to ¹"1.1 K (not shown). Closer examination of the IETS measurements reveals that the peak around 115 mV has a low voltage shoulder extending to about 30 mV [10,11]. This is best seen in Fig. 1b and Fig. 2b, where a third order polynomial "t through the high voltage data indicates the elastic tunnel contribution. Near zero voltage an anomaly is observed in the G}< curves for the junctions in which the top electrode is magnetic (Fig. 1a and Fig. 2b), while the anomaly is absent when the top electrode is nonmagnetic (Fig. 1b and Fig. 2b). In the IETS measurements, this anomaly results in a large feature for which the sharpness is limited by the modulation voltage only. Since the anomaly is only observed in the presence of a magnetic electrode, this feature is tentatively ascribed to magnon excitations. The temperature dependence of the peak around 115 mV in the IETS measurements also seems to indicate a magnon contribution at that voltage, similar to an earlier observation of magnons in NiO by Tsui et al. [8,12]. The dip at zero voltage in the G}< curves for the junctions with magnetic electrodes is signi"cant and indicates that the inelastic magnon contribution to the tunnel current cannot be neglected in the description of the transport properties of magnetic tunnel junctions. Since magnons can also be excited thermally (room temperature corresponds to an energy of 25 meV), magnon excitations will contribute to the tunnel current at ele-
153
Fig. 2. Conductance and IETS measurements at ¹"4.2 K of (a) an Al}AlN}Co junction (Ni Fe showed similar results) and (b) an Al}AlN}Al junction (Au and Cu showed similar results). The AC voltage excitation was &1 mV for the G}< \ and &10 mV for the IETS measurements. The vertical \ dashed lines indicate 115 mV. The dotted line in (b) illustrates the elastic part of the IETS measurement.
vated temperatures and zero bias voltage. Indeed, we found that using mainly magnon excitations, the decreasing JMR of magnetic tunnel junctions at elevated temperatures can be explained [8,9]. Likewise, the occurrence of the knee in the G}< curves illustrates that the inelastic phonon contribution to the tunnel current is also non-negligible. However, since the energy for phonon excitations corresponds to temperatures well above room temperature, the e!ect of the phonons shows up mainly in the voltage dependence. In conclusion, we performed IETS measurements on tunnel junctions containing both magnetic and non-magnetic electrodes. No indications of carbon or hydrogen contaminations have been found, indicating that the barrier preparation is very clean. Signi"cant contributions to the tunnel current are observed from inelastic magnon and phonon excitations. Therefore we believe that both magnon and phonon contributions must be included in an improved description of the temperature and voltage dependence of the magnetoresistance of magnetic tunnel junctions. R.V. acknowledges a travel grant from the Dutch Technology Foundation (NWO). This work was supported in part by the EUT/Philips Research collaboration, ESPRIT Research Project No. 20 027, &Novel Magnetic Nanodevices of arti"cially layered Materials (NM)', NSF Grant No. DMR 9423013, and ONR Grant No. N00014-92-J-1847.
154
R.J.M. van de Veerdonk et al. / Journal of Magnetism and Magnetic Materials 198}199 (1999) 152}154
References [1] J.S. Moodera, L.R. Kinder, T.M. Wong, R. Meservey, Phys. Rev. Lett. 74 (1995) 3273. [2] M. Julliere, Phys. Lett. 54A (1975) 225. [3] S. Maekawa, U. GaK phert, IEEE Trans. Magn. MAG-18 (1982) 707. [4] J.C. Slonczewski, Phys. Rev. B 39 (1989) 6995. [5] A.M. Bratkovsky, Phys. Rev. B 56 (1997) 2344. [6] S. Zhang, P.M. Levy, A.C. Marley, S.S.P. Parkin, Phys. Rev. Lett. 79 (1997) 3744. [7] J. Inoue, S. Maekawa, unpublished results.
[8] J.S. Moodera, J. Nowak, R.J.M. van de Veerdonk, Phys. Rev. Lett. 80 (1998) 2941. [9] C.H. Shang, J. Nowak, R. Jansen, J.S. Moodera, Phys. Rev. B 58 (1998) R2917. [10] E.L. Wolf, in: R.J. Elliot, J.A. Krumhansl, W. Marshall, D.H. Wilkinson (Eds.), Principles of Electron Tunneling Spectroscopy, vol. 71 of International Series of Monographs on Physics, Oxford University Press, New York, 1985. [11] O.I. Shklyarevskii, I.K. Yanson, V.D. Zaporozhskii, Solid State Commun. 14 (1974) 327. [12] D.C. Tsui, R.E. Dietz, L.R. Walker, Phys. Rev. Lett. 27 (1971) 1729.
Characterization of magnetic tunnel junctions using IETS R.J.M. van de Veerdonk *, J.S. Moodera, W.J.M. de Jonge Department of Applied Physics and Research School COBRA, Eindhoven University of Technology, Den Dolech 2, 5600 MB Eindhoven, Netherlands Francis Bitter Magnet Laboratory, Massachusetts Institute of Technology, 170 Albany Street, Cambridge, MA 02139, USA
Abstract In this paper inelastic electron tunneling spectroscopy measurements are presented on tunnel junctions containing both magnetic and nonmagnetic electrodes. Magnon excitations have been found near zero voltage; phonons are observed between 30 and 110 mV. Therefore, a model description of the transport properties of magnetic tunnel junctions at elevated temperatures or non-zero bias voltage should include both inelastic contributions. 1999 Elsevier Science B.V. All rights reserved. Keywords: Tunneling; Magnetoresistance; Inelastic electron tunneling spectroscopy
Tunnel junctions containing two ferromagnetic electrodes separated by an insulator show a sizable junction magnetoresistance (JMR) e!ect depending on the relative orientation of the electrode magnetizations [1]. The magnitude of the JMR at zero temperature and zero voltage is determined by the polarization of the tunneling electrons and can be understood by Julliere's model [2}4]. However, the decrease of the JMR at elevated temperatures and non-zero bias voltage is not yet understood. It has been suggested that impurity, phonon, and magnon assisted tunneling processes contribute to the decreasing JMR [5}7]. We have performed a detailed study on the temperature dependence of the transport properties of tunnel junctions [8,9], with the goal to provide an experimental basis for an improved model description. In this paper inelastic electron tunneling spectroscopy (IETS) [10] is used to show that inelastic phonon and magnon assisted contributions to the tunnel current are signi"cant. The junctions used in this study have been fabricated using shadow evaporation onto glass substrates, similar to the method described in Ref. [1]. This results in
* Corresponding author. Tel.: #31-40-274-4089; fax: #3140-274-4282; e-mail: [email protected].
a cross-strip geometry junction with an area of &200;300 lm. The tunnel barriers are formed as a native oxide/nitride on 10 nm-thick thermally evaporated Al strips using an in-situ O or N DC glow dis charge. The 20 nm-thick counterelectrodes were e-gun evaporated Co, or thermally evaporated Ni Fe , Cu, Al, or Au strips. The resistance of the studied junctions was of the order of a few k). In Figs. 1 and 2 the di!erential conductance (G" dI/d<) and the inelastic electron tunneling spectra, i.e., the derivative of the di!erential conductance versus voltage (dG/d<), of some of the tunnel junctions are shown. The results for the Al O (Fig. 1) and AlN (Fig. 2) barriers are very similar. In Fig. 1a and Fig. 2a one of the electrodes is magnetic, while in Fig. 1b and Fig. 2b both electrodes are nonmagnetic. In all cases no "ne-structure is observed in the IETS measurements at high voltages, where contributions from water and carbon contaminations would show up [10]. This indicates that the barrier preparation process is very clean. A peak around <"115 mV dominates the IETS measurements, which has been identi"ed as an Al}O stretching mode for the Al O barrier [10] and as a TO phonon mode for the AlN barrier [11]. This large inelastic peak shows up as a clearly visible knee in the G}< curves, which becomes even better resolved when the sample is
0304-8853/99/$ } see front matter 1999 Elsevier Science B.V. All rights reserved. PII: S 0 3 0 4 - 8 8 5 3 ( 9 8 ) 0 1 0 5 5 - 5
R.J.M. van de Veerdonk et al. / Journal of Magnetism and Magnetic Materials 198}199 (1999) 152}154
Fig. 1. Conductance and IETS measurements at ¹"4.2 K of (a) an Al}Al O }Ni Fe junction (Co showed similar results) and (b) an Al}Al O }Au junction (Cu and Al showed similar results). The AC voltage excitation was &1 mV for the G}< \ and &10 mV for the IETS measurements. The vertical \ dashed lines indicate 115 mV. The dotted line in (b) illustrates the elastic part of the IETS measurement.
cooled to ¹"1.1 K (not shown). Closer examination of the IETS measurements reveals that the peak around 115 mV has a low voltage shoulder extending to about 30 mV [10,11]. This is best seen in Fig. 1b and Fig. 2b, where a third order polynomial "t through the high voltage data indicates the elastic tunnel contribution. Near zero voltage an anomaly is observed in the G}< curves for the junctions in which the top electrode is magnetic (Fig. 1a and Fig. 2b), while the anomaly is absent when the top electrode is nonmagnetic (Fig. 1b and Fig. 2b). In the IETS measurements, this anomaly results in a large feature for which the sharpness is limited by the modulation voltage only. Since the anomaly is only observed in the presence of a magnetic electrode, this feature is tentatively ascribed to magnon excitations. The temperature dependence of the peak around 115 mV in the IETS measurements also seems to indicate a magnon contribution at that voltage, similar to an earlier observation of magnons in NiO by Tsui et al. [8,12]. The dip at zero voltage in the G}< curves for the junctions with magnetic electrodes is signi"cant and indicates that the inelastic magnon contribution to the tunnel current cannot be neglected in the description of the transport properties of magnetic tunnel junctions. Since magnons can also be excited thermally (room temperature corresponds to an energy of 25 meV), magnon excitations will contribute to the tunnel current at ele-
153
Fig. 2. Conductance and IETS measurements at ¹"4.2 K of (a) an Al}AlN}Co junction (Ni Fe showed similar results) and (b) an Al}AlN}Al junction (Au and Cu showed similar results). The AC voltage excitation was &1 mV for the G}< \ and &10 mV for the IETS measurements. The vertical \ dashed lines indicate 115 mV. The dotted line in (b) illustrates the elastic part of the IETS measurement.
vated temperatures and zero bias voltage. Indeed, we found that using mainly magnon excitations, the decreasing JMR of magnetic tunnel junctions at elevated temperatures can be explained [8,9]. Likewise, the occurrence of the knee in the G}< curves illustrates that the inelastic phonon contribution to the tunnel current is also non-negligible. However, since the energy for phonon excitations corresponds to temperatures well above room temperature, the e!ect of the phonons shows up mainly in the voltage dependence. In conclusion, we performed IETS measurements on tunnel junctions containing both magnetic and non-magnetic electrodes. No indications of carbon or hydrogen contaminations have been found, indicating that the barrier preparation is very clean. Signi"cant contributions to the tunnel current are observed from inelastic magnon and phonon excitations. Therefore we believe that both magnon and phonon contributions must be included in an improved description of the temperature and voltage dependence of the magnetoresistance of magnetic tunnel junctions. R.V. acknowledges a travel grant from the Dutch Technology Foundation (NWO). This work was supported in part by the EUT/Philips Research collaboration, ESPRIT Research Project No. 20 027, &Novel Magnetic Nanodevices of arti"cially layered Materials (NM)', NSF Grant No. DMR 9423013, and ONR Grant No. N00014-92-J-1847.
154
R.J.M. van de Veerdonk et al. / Journal of Magnetism and Magnetic Materials 198}199 (1999) 152}154
References [1] J.S. Moodera, L.R. Kinder, T.M. Wong, R. Meservey, Phys. Rev. Lett. 74 (1995) 3273. [2] M. Julliere, Phys. Lett. 54A (1975) 225. [3] S. Maekawa, U. GaK phert, IEEE Trans. Magn. MAG-18 (1982) 707. [4] J.C. Slonczewski, Phys. Rev. B 39 (1989) 6995. [5] A.M. Bratkovsky, Phys. Rev. B 56 (1997) 2344. [6] S. Zhang, P.M. Levy, A.C. Marley, S.S.P. Parkin, Phys. Rev. Lett. 79 (1997) 3744. [7] J. Inoue, S. Maekawa, unpublished results.
[8] J.S. Moodera, J. Nowak, R.J.M. van de Veerdonk, Phys. Rev. Lett. 80 (1998) 2941. [9] C.H. Shang, J. Nowak, R. Jansen, J.S. Moodera, Phys. Rev. B 58 (1998) R2917. [10] E.L. Wolf, in: R.J. Elliot, J.A. Krumhansl, W. Marshall, D.H. Wilkinson (Eds.), Principles of Electron Tunneling Spectroscopy, vol. 71 of International Series of Monographs on Physics, Oxford University Press, New York, 1985. [11] O.I. Shklyarevskii, I.K. Yanson, V.D. Zaporozhskii, Solid State Commun. 14 (1974) 327. [12] D.C. Tsui, R.E. Dietz, L.R. Walker, Phys. Rev. Lett. 27 (1971) 1729.