Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
APPLIED PHYSICS LETTERS 104, 082403 (2014) Electric-field-induced spin wave generation using multiferroic magnetoelectric cells Sergiy Cherepov,1 Pedram Khalili Amiri,1 Juan G. Alzate,1 Kin Wong,1 Mark Lewis,1 Pramey Upadhyaya,1 Jayshankar Nath,1 Mingqiang Bao,1 Alexandre Bur,2 Tao Wu,2 Gregory P. Carman,2 Alexander Khitun,3 and Kang L. Wang1 1 Department of Electrical Engineering, University of California, Los Angeles, California 90095, USA Department of Mechanical and Aerospace Engineering, University of California, Los Angeles, California 90095, USA 3 Department of Electrical Engineering, University of California, Riverside, California 92521, USA 2 (Received 27 August 2013; accepted 10 January 2014; published online 25 February 2014) In this work, we report on the demonstration of voltage-driven spin wave excitation, where spin waves are generated by multiferroic magnetoelectric (ME) cell transducers driven by an alternating voltage, rather than an electric current. A multiferroic element consisting of a magnetostrictive Ni film and a piezoelectric [Pb(Mg1/3Nb2/3)O3](1x)–[PbTiO3]x substrate was used for this purpose. By applying an AC voltage to the piezoelectric, an oscillating electric field is created within the piezoelectric material, which results in an alternating strain-induced magnetic anisotropy in the magnetostrictive Ni layer. The resulting anisotropy-driven magnetization oscillations propagate in the form of spin waves along a 5 lm wide Ni/NiFe waveguide. Control experiments confirm the strain-mediated origin of the spin wave excitation. The voltage-driven spin wave excitation, demonstrated in this work, can potentially be used for low-dissipation spin wave-based logic and C 2014 AIP Publishing LLC. [http://dx.doi.org/10.1063/1.4865916] memory elements. V The scaling of semiconductor-based complementary metal-oxide-silicon (CMOS) logic devices is facing significant difficulties, due to the increased static power dissipation of CMOS transistors at small length scales.1–3 Hence, there is an increasing interest in developing technologies that can enable beyond-CMOS computing.4–6 Spintronic logic technologies4,5,7,8 have emerged as leading candidates for this purpose, given their combination of high speed, endurance, potentially improved variability issues, and zero standby power due to nonvolatile operation.6 One of the most important requirements for any beyond-CMOS logic device is a low dynamic power consumption per operation.6,7,9 Spin waves (magnons) have been proposed as a potential means to enable beyond-CMOS computation devices, where propagation and interference of waves in a magnonic waveguide7,10–13 are utilized for information transfer and processing. The potential energy efficiency of spin wave based devices is associated with the wave nature of the information carrier which, unlike charge-based devices, does not necessarily suffer from ohmic losses.6,7,14 To maintain low energy consumption of magnonic devices, however, the spin wave generation element needs to be energy-efficient as well, so as to minimize the dynamic power dissipation. However, the commonly used current-driven transducers such as inductive antennas15 and spin transfer torque16 elements require large currents to operate, resulting in high power dissipation and scaling issues due to intrinsic ohmic losses. To reduce the power dissipation associated with magnetization excitation in spin wave logic, in this work, we propose and demonstrate spin wave transducers where spin waves are generated by alternating voltages (i.e., electric fields) rather than currents. We refer to such devices as magnetoelectric 0003-6951/2014/104(8)/082403/5/$30.00 (ME) cells. Instead of single phase multiferroics,17–19 which often provide rather small magnetoelectric coupling, in this work a multiferroic heterostructure, i.e., a magnetostrictive/ piezoelectric bi-layer is used for this purpose, having a layer structure similar to those previously reported in Refs. 20–22. In this structure, the electric-field-induced strain in a piezoelectric [Pb(Mg1/3Nb2/3)O3](1x)–[PbTiO3]x (PMN-PT) substrate converts to magnetization excitations using the magnetostrictive properties of a Ni layer. It has been shown that quasi-static electric fields on the order of 60.8 MV/m applied across a PMN-PT ferroelectric layer can modify the magnetization configuration in an adjacent magnetostrictive layer, inducing effective anisotropy fields more than 300 Oe.21–23 Moreover, it was recently shown that the magnetization in a magnetostrictive ferromagnetic layer can be dynamically excited using Surface Acoustic Waves (SAW).24,28 Here, we propose that a similar effect can be used for magnetization excitation by alternating strain-induced-anisotropy, created by RF voltages applied to the ME cell. In this work, we show that two voltage-driven piezoelectric transducers placed in proximity to each other on a common ferromagnetic bus can be used for generation and detection of the propagating spin wave. The capacitive ME cells, considered here, do not suffer from Ohmic losses intrinsic to current-driven spin wave excitation mechanisms. Hence, the power dissipation of these transducers is determined only by the negligible AC current flow during operation, caused by RF driving of the device capacitance. The ME cell capacitance, i.e., power dissipation due to AC current is dramatically reduced with further scaling down of the device. Figure 1(a) shows a schematic of the performed experiment as well as a microscopic picture of the studied devices. 104, 082403-1 C 2014 AIP Publishing LLC V 082403-2 Cherepov et al. A 5 lm wide ferromagnetic stripe (spin wave bus) was lithographically defined from the ferromagnetic bilayer consisting of Ni (20 nm) and Ni80-Fe20 (20 nm) layers, sputtered onto a single crystal piezoelectric PMN-PT, (x 32%), cut in the (110) orientation. The Ni layer provides magnetostrictive properties while the soft Ni80Fe20 layer provides a low damping for spin wave propagation. The 200 nm thick aluminum loop antennas and ME cells (see Fig. 1(a)), used for spin wave excitation and detection were designed as symmetric shorted and open coplanar transmission lines, respectively. They were built using 2 lm wide Al stripes and patterned on top of the 200 nm SiO2 insulation layer using standard lift off techniques, as shown in Figure 1 (inset). Antennas and ME cell structures were separated by 2 lm edge-to-edge distance. The ME cell electrodes provide direct electrical contact to the piezoelectric PMN-PT substrate via dedicated openings in the insulating SiO2 layer, as schematically shown in Figure 1(a). The microwave loop antennas, driven by an applied RF current, were designed to inductively excite magnetostatic spin waves,25,26 with a calculated wavelength of 8 lm. The spin wave generation by the voltage driven ME cell can be understood in the following way: an RF voltage is applied to the piezoelectric PMN-PT substrate through the opening in the insulating SiO2 layer. The alternating electric field, localized between the electrodes, generates oscillating strain in the piezoelectric substrate underneath the spin wave bus, which, in turn, leads to magnetoelastic spin wave excitation in the magnetostrictive Ni layer due to alternating strain induced anisotropy. The generated spin wave propagates within the Ni/NiFe spin wave bus to the other terminal of the device. Spin wave propagation up to 40 lm was observed in our devices. We used a vector network analyzer (VNA) to perform propagating spin wave spectroscopy (PSWS) measurements27 as shown in Figure 1(a). The two-port measurements were performed with each port of the VNA connected to the loop antenna and ME cell, respectively, using broadband surface probes. The recorded scattering parameters carry information about magnetization excitation (S11 and S22) and bi-directional spin wave propagation (S21 and S12) in the studied device, as schematically shown in Figure 1(b). The two-port measurements were performed with external magnetic fields in a range of 6500 Oe, applied along the x axis. At each magnetic field step, the magnetization response was probed in a frequency range of 1–10 GHz using the VNA operating at 5dBm power. The excitation amplitude applied to the ME cell contacts was kept below the de-poling field of the PMN-PT, which was found to be around 1.6 kV/cm. The saturation magnetization of the ferromagnetic bilayer, used as spin wave bus in the studied devices, was found to be 0.72 T as measured using superconducting quantum interference device (SQUID) magnetometry. Figure 2 shows the measured experimental reflection and transmission signals as a function of applied magnetic field H and frequency of excitation. The distinctive magnetic-field-dependent signal, visible in Figures 2(a) and 2(d), corresponds to magnetization excitations created by the current driven antenna (S11) and by the voltage-driven ME cell (S22), respectively. The magnetic field dependent resonant feature, visible on S21 and S12 transmission measurements, Appl. Phys. Lett. 104, 082403 (2014) FIG. 1. (a) Schematic of the studied device: Spin wave generation and propagation measurements using a vector network analyzer were performed on the 5 lm wide Ni/NiFe bus lithographically defined on a PMN-PT piezoelectric substrate. Inset shows cross-section view of the ME cell. (b) The schematic of two-port measurements of transmission (S21 and S12) and reflection (S11 and S22) measurements between conventional loop antennas and voltage-driven magnetoelectric cells. represents a spin wave signal propagating along the spin wave bus, generated by the antenna and detected by the ME cell and vice versa, in Figures 2(b) and 2(c), respectively. Here, the transmission and reflection data are plotted after subtraction of the magnetic-field-independent background signal, which had an amplitude of about 55 dBm, at 2500 Oe bias magnetic field. The observed difference in the shape of the resonance curves obtained by antennas and ME cells (S11 and S22, respectively) is determined by the spin wave excitation mechanism, i.e., by the different (frequency-dependent) coupling efficiencies in the case of the two transducer geometries. We also performed experiments on simultaneous spin wave generation and detection between two identical ME cells, as shown in Figure 2(e). The measurements indicate that spin waves indeed can be generated and detected only by ME cells. Although the magnetization excitation by strain has been reported earlier,21 and a number of experimental and theoretical works on generation and propagation of magnetoelastic waves in ferromagnets were performed in the past,29–31 the observed results present a demonstration of voltage-driven generation of propagating spin waves in confined ferromagnetic structures which are more suitable for spin wave logic applications, and confirm that multiferroic ME cells can be used for generation of spin waves as well as for their detection. 082403-3 Cherepov et al. Appl. Phys. Lett. 104, 082403 (2014) We next analyze the ferromagnetic resonance (FMR) frequencies from S22 reflection parameter measurements on the voltage driven ME cell shown in Figure 2(d). The data show an excellent match to the analytically calculated dynamic response of the magnetization, described by the Kittel equation.32 The ferromagnetic resonance frequency is given by f2 ¼ c 2p 2 H 2 þ 4pMS H ; (1) where H is the external magnetic field, MS is the average saturation magnetization of the bilayer, and c is the gyromagnetic ratio. The position of the resonance peak extracted from the experimental S22 response is shown as black circles in Figure 2(f), while the solid back line represents the analytical fit obtained by solving Eq. (1). The fitting of the experimentally obtained resonance peak dispersion to the analytical data reveals the effective saturation magnetization of the Ni/NiFe bilayer to be MS 0.73 T, which is in excellent agreement with independent SQUID measurements performed on the same multilayer. In order to confirm the magnetoelastic nature of spin wave generation, control PSWS experiments between currentdriven loop antennas and voltage-driven (strain-mediated) ME cells were performed. The test samples have the same geometry, layout, and material parameters as the samples described above; however, their material structures were modified in a way to suppress strain-induced magnetization excitation. In a first control sample, the piezoelectric PMNPT substrate was replaced by a Si substrate. In this case, the S11 magnetic-field-dependent FMR signal was detected only from the loop antennas inductively coupled to the spin wave bus, as shown in Figure 3(a). The S22 parameter measured on the magnetoelectric cell, shown in Figure 3(b), did not have any magnetic-field-dependent signal, due to the negligible strain response from the non-piezoelectric Si substrate. A similar result was obtained on a reference sample where the spin wave bus, deposited on the piezoelectric PMN-PT substrate, had only a single non-magnetostrictive NiFe layer33,34 (no Ni layer). Here, the FMR response S11 is again detected only by the inductive loop antenna, which is inductively coupled to the ferromagnetic material, as seen in Figure 3(c). In this test sample, the saturation magnetization Ms of the NiFe bus was estimated from the Kittel fit of the experimental data, shown in Figure 2(f) as red triangles, and was found to be 0.78 T, larger than the Ms of the Ni/NiFe bi layer as expected. The negligible magnetostriction of the NiFe layer in this sample eliminates the strain-induced FMR response of the ME cell (S22), as shown in Figure 3(d). In both control samples, the transmission parameters S21 and S12 do not show any magnetic field dependent features. The difference in the S11 dispersion curves shown in Figure 2 and control data shown in Figure 3 can be explained by a small variation in antenna geometry and different thickness and roughness of the magnetic films, as well as by the presence of intrinsic stress during sample fabrication on the piezoelectric substrate. The performed test measurements provide a qualitative confirmation of the strain-induced nature of spin wave generation. This work shows a proof of concept of spin wave generation and detection by voltage-driven magnetoelectric cells; FIG. 2. Two-port measurements between antenna and ME cell. Magnetic field dependence of the FMR signal generated by antenna and ME cell are shown on (a) and (d), respectively. Figure (b) represents the spin wave signal detected by the ME cell, while being generated by the loop antenna. Figure (c) shows the transmission signal generated by the ME cell and detected by the loop antenna, while (e) shows the transmission signal recorded between two ME cells. Figure (f) shows the fitting to the Kittel equation for the experimental FMR data detected on the samples containing Ni/NiFe bilayer (black circles) and only NiFe (red triangles) spin wave bus films. 082403-4 Cherepov et al. Appl. Phys. Lett. 104, 082403 (2014) FIG. 3. (a) and (b): Antenna (a) and ME cell (b) absorption spectra measured on the sample, where Ni/NiFe spin wave bus is fabricated on a regular Si wafer. The voltage applied to the ME cell fabricated on the non-piezoelectric Si wafer does not create strain and hence does not lead to spin wave generation. (c) and (d): Reflection signal as a function of bias magnetic field measured using antenna (c) and ME cell (d) on the sample where only a NiFe spin wave bus was fabricated on the piezoelectric PMN-PT substrate. The absence of the magnetostrictive Ni layer eliminates spin wave generation by the voltage-driven ME cell. These experiments confirm the strain-induced nature of spin wave generation in our ME cells. however, further optimization is required for potential applications. The spin wave generation efficiency, essential for real applications, can be increased by ME cell geometry optimization, reducing mechanical clamping by the bulk substrate, and proper impedance matching on the excitation and detection elements. Full optimization also requires consideration of the magnetoelastic nature of the excited waves,29–31 which is beyond the scope of this work. Additionally, the spin wave propagation length can be enhanced by using lowloss materials such as magnetostrictive Heusler alloys35,36 or dielectric materials such as yttrium iron garnet (YIG) films15 instead of the Ni/NiFe spin wave bus, which contains a highdamping Ni37,38 layer necessary for strain-induced magnetization excitation in our experiment. In conclusion, we report on the voltage-induced strainmediated generation and detection of propagating spin waves using multiferroic magnetoelectric cells. We demonstrated that current-driven antennas, commonly used for spin wave generation in magnonic devices, can be replaced by voltagedriven ME cells based on ferroelectric-magnetostrictive bilayers. It should be noted that the amplitude of the spin waves generated in this demonstration is still rather small, and may not be suitable for direct implementation in a spin wave logic gate. However, this problem can be addressed by using materials with lower damping, improving the generation efficiency of the ME cells by geometrical and materials optimization, as well as through the use of spin wave amplifiers.39,40 The voltage-driven spin wave excitation mechanism can then be utilized in spin wave logic circuits, allowing for compact and energy-efficient logic elements which have a potential for beyond-CMOS logic and memory device realization. The authors gratefully acknowledge financial support from the DARPA Non-Volatile Logic program, by the NSF Nanosystems Engineering Research Center for Translational Applications of Nanoscale Multiferroic Systems (TANMS), and by NRI through the Western Institute of Nanoelectronics. 1 G. E. Moore, Proceedings of the IEEE 86, 82 (1998). T. N. Theis and P. M. Solomon, Proceedings of the IEEE 98, 2005 (2010). 3 The International Technology Roadmap For semiconductors, 2011. 4 D. A. Allwood, G. Xiong, C. C. Faulkner, D. Atkinson, D. Petit, and R. P. Cowburn, Science 309, 1688 (2005). 5 A. Imre, G. Csaba, L. Ji, A. Orlov, G. H. Bernstein, and W. Porod, Science 311, 205 (2006). 6 K. L. Wang and P. K. Amiri, SPIN 02, 1250009 (2012). 7 A. Khitun, M. Bao, and K. L. Wang, J. Phys. D: Appl. Phys. 43, 264005 (2010). 8 B. Behin-Aein, D. Datta, S. Salahuddin, and S. Datta, Nat. Nano 5, 266 (2010). 9 D. E. Nikonov and I. A. Young, IEEE Int. Electron Dev. Meet. 2012, 25.4.1. 10 T. Schneider, A. A. Serga, B. Leven, B. Hillebrands, R. L. Stamps, and M. P. Kostylev, Appl. Phys. Lett. 92, 022505 (2008). 11 K.-S. Lee and S.-K. Kim, J. Appl. Phys. 104, 053909 (2008). 12 Y. Wu, M. Bao, A. Khitun, J.-Y. Kim, A. Hong, and K. L. Wang, J. Nanoelectron. Optoelectron. 4, 394 (2009). 13 M. Bao, K. Wong, A. Khitun, J. Lee, Z. Hao, K. L. Wang, D. W. Lee, and S. X. Wang, Europhys. Lett. 84, 27009 (2008). 14 Y. Kajiwara, K. Harii, S. Takahashi, J. Ohe, K. Uchida, M. Mizuguchi, H. Umezawa, H. Kawai, K. Ando, K. Takanashi, S. Maekawa, and E. Saitoh, Nature 464, 262 (2010). 15 A. A. Serga, A. V. Chumak, and B. Hillebrands, J. Phys. D: Appl. Phys. 43, 264002 (2010). 16 S. I. Kiselev, J. C. Sankey, I. N. Krivorotov, N. C. Emley, R. J. Schoelkopf, R. A. Buhrman, and D. C. Ralph, Nature 425, 380 (2003). 17 R. de Sousa and J. E. Moore, Appl. Phys. Lett. 92, 022514 (2008). 18 P. Rovillain, R. De Sousa, Y. Gallais, A. Sacuto, M. A. Measson, D. Colson, A. Forget, M. Bibes, A. Barthelemy, and M. Cazayous, Nature Mater. 9, 975 (2010). 19 D. Mills and I. Dzyaloshinskii, Phys. Rev. B 78, 184422 (2008). 20 T. Wu, A. Bur, P. Zhao, K. P. Mohanchandra, K. Wong, K. L. Wang, C. S. Lynch, and G. P. Carman, Appl. Phys. Lett. 98, 012504 (2011). 21 C.-J. Hsu, J. L. Hockel, and G. P. Carman, Appl. Phys. Lett. 100, 092902 (2012). 22 T. Wu, A. Bur, K. Wong, J. L. Hockel, C.-J. Hsu, H. K. D. Kim, K. L. Wang, and G. P. Carman, J. Appl. Phys. 109, 07D732 (2011). 23 J. L. Hockel, A. Bur, T. Wu, K. P. Wetzlar, and G. P. Carman, Appl. Phys. Lett. 100, 022401 (2012). 24 M. Weiler, L. Dreher, C. Heeg, H. Huebl, R. Gross, M. S. Brandt, and S. T. B. Goennenwein, Phys. Rev. Lett. 106, 117601 (2011). 25 R. W. Damon and J. R. Eshbach, J. Phys. Chem. Solids 19, 308 (1961). 26 R. E. De Wames and T. Wolfram, J. Appl. Phys. 41, 987 (1970). 27 M. Bailleul, D. Olligs, and C. Fermon, Appl. Phys. Lett. 83, 972 (2003). 28 L. Dreher, M. Weiler, M. Pernpeintner, H. Huebl, R. Gross, M. S. Brandt, and S. T. B. Goennenwein, Phys. Rev. B 86, 134415 (2012). 2 082403-5 29 Cherepov et al. S. M. Rezende and F. R. Morgenthaler, J. Appl. Phys. 40, 537 (1969). R. E. Camley, J. Appl. Phys. 50, 5272 (1979). 31 P. R. Emtage, Phys. Rev. B 13, 3063 (1976). 32 C. Kittel, Phys. Rev. 73, 155 (1948). 33 J. Lo, C. Hwang, T. C. Huang, R. Campbell, and S. Jose, IEEE Trans. Magn. 23, 3065 (1987). 34 J. P. Reekstin, E. N. Mitchell, G. I. Lykken, G. D. Babcock, I. W. Wolf, T. S. Crowther, T. R. Long, W. Metzdorf, and R. H. Nelson, J. Appl. Phys. 38, 1449 (1967). 35 S. Li, M. Liu, J. Lou, S. Beguhn, J. Wu, J. Qiu, J. Lin, Z. Cai, Y. Hu, F. Xu, J.-G. Duh, and N. X. Sun, J. Appl. Phys. 111, 07C705 (2012). 30 Appl. Phys. Lett. 104, 082403 (2014) 36 T. Sebastian, Y. Ohdaira, T. Kubota, P. Pirro, T. Bra, K. Vogt, A. A. Serga, H. Naganuma, M. Oogane, Y. Ando, and B. Hillebrands, Appl. Phys. Lett. 100, 112402 (2012). 37 M. Oogane, T. Wakitani, S. Yakata, R. Yilgin, Y. Ando, A. Sakuma, and T. Miyazaki, Jpn. J. Appl. Phys., Part 1 45, 3889 (2006). 38 J. Walowski, M. D. Kaufmann, B. Lenk, C. Hamann, J. McCord, and M. M€ unzenberg, J. Phys. D: Appl. Phys. 41, 164016 (2008). 39 M. Bao, A. Khitun, Y. Wu, J.-Y. Lee, K. L. Wang, and A. P. Jacob, Appl. Phys. Lett. 93, 072509 (2008). 40 A. Khitun, D. E. Nikonov, and K. L. Wang, J. Appl. Phys. 106, 123909 (2009).