Creation and directional motion of chiral spin textures induced by electric fields
Magnetization dynamics driven by an electric field (EF) could provide long-term benefits to information technologies because of its ultralow power consumption Ohno_Nature2000 ; Ohno_NM2010 , on which considerable efforts have been made including: magnetic anisotropy modification Chiba_Nature2008 ; Maruyama_NN2009 ; Yoichi_APE2009 ; Kihiro_APE2013 ; Miwa_NC2017 , magnetization switching Yoichi_APE2009 ; Chiba_Science2003 ; Shiota_NN2012 ; Wang_NM2012 ; Dohi_AIPA2016 ; Kanai_APL2012 , domain structure modification Haruka_APE2016 ; Ando_APL2016 , and domain wall motion Michihiko_JJAP2006 ; Schellekens_NC2012 ; Chiba_NC2012 ; Bauer_APL2012 ; Haruka_JJAP2013 . Besides, it has been found that the Dzyaloshinskii-Moriya interaction (DMI) in interfacially asymmetric multilayers consisting of a ferromagnet and a heavy metal can stabilize novel magnetic structures, such as chiral domain walls, skyrmions, and chiral bubbles Nagaosa_NN2013 ; Wiesendanger_NRM2016 ; Fert_NRM2016 ; Jiang_Review2017 . These chiral spin textures can also be controlled by an EF Hsu_NN2017 ; Schott_NL2017 , and hold promise for building low-power information processing devices Nagaosa_NN2013 ; Wiesendanger_NRM2016 ; Fert_NRM2016 ; Jiang_Review2017 . In this Letter, we present experimental results on the EF-induced creation and directional motion of chiral spin textures in a magnetic multilayer with an artificial thickness gradient and interfacial DMI at room temperature, where the EF is applied at the Pt/dielectric interface. We find that the EF-induced directional motion of the chiral domain wall is accompanied with the creation of chiral bubbles at a certain external magnetic field. Our findings may provide opportunities for developing novel chiral spin textures-based information storage devices with ultralow power consumption.
EF-induced spintronic phenomena could offer great benefits to the information-related industries, since they can be harnessed for building information processing devices with ultralow power consumption Ohno_Nature2000 ; Ohno_NM2010 . Especially, the modification of magnetic parameters, such as the magnetic anisotropy Chiba_Nature2008 ; Maruyama_NN2009 ; Yoichi_APE2009 ; Kihiro_APE2013 ; Miwa_NC2017 , plays an essential role in realizing magnetization switching Yoichi_APE2009 ; Chiba_Science2003 ; Shiota_NN2012 ; Wang_NM2012 ; Kanai_APL2012 ; Chiba_APL2013 , domain structure modification Dohi_AIPA2016 ; Haruka_APE2016 ; Ando_APL2016 , and domain wall motion Michihiko_JJAP2006 ; Schellekens_NC2012 ; Chiba_NC2012 ; Bauer_APL2012 ; Haruka_JJAP2013 . Additionally, recent experiments on magnetic asymmetric multilayers show that the antisymmetric DMI, i.e., the noncollinear exchange interaction, can be induced at the ferromagnet/heavy metal interface Nagaosa_NN2013 ; Wiesendanger_NRM2016 ; Fert_NRM2016 ; Jiang_Review2017 , which can further stabilize novel chiral spin textures including chiral domain walls, skyrmions, and chiral bubbles. These chiral spin textures provide emerging opportunities for developing low-power information processing and high-density storage technologies Nagaosa_NN2013 ; Wiesendanger_NRM2016 ; Fert_NRM2016 ; Jiang_Review2017 . A number of representative examples can be found based on the manipulation of chiral domain walls, skyrmions, and chiral bubbles Nagaosa_NN2013 ; Wiesendanger_NRM2016 ; Fert_NRM2016 ; Jiang_Review2017 . In any domain wall-based applications, the creation and motion of domain walls are indispensable processes, as information can be encoded by a sequence of movable domain walls representing binary bits Wiesendanger_NRM2016 ; Fert_NRM2016 . Theoretical and experimental studies have suggested that chiral spin textures can be created and controlled in multiple different ways, such as by applying magnetic fields or electric currents Nagaosa_NN2013 ; Wiesendanger_NRM2016 ; Fert_NRM2016 ; Jiang_Review2017 ; Junya_JPDAP2011 .
However, methods based on an electric current may be energy consuming and produce significant Joule heating, which impedes the integration into actual circuits with nanoscale dimensions. Therefore, as the Joule heating effect can be significantly suppressed under an EF, the use of an EF appears to be an efficient and robust method for creating, driving, and controlling magnetic structures Ohno_NM2010 ; Hsu_NN2017 ; Schott_NL2017 . For example, the EF-controlled or EF-assisted motion of a domain wall has already been realized Michihiko_JJAP2006 ; Schellekens_NC2012 ; Chiba_NC2012 ; Bauer_APL2012 ; Haruka_JJAP2013 , which suggests that the domain wall velocity can be changed by the EF (applied voltage V) effectively. The domain structure can also be significantly modified by the EF ( V) Ando_APL2016 . Most recently, it was found that a local EF ( V) can be used to create a skyrmion from the ferromagnetic background at low temperature Hsu_NN2017 . The creation and annihilation of chiral bubbles at room temperature by the EF ( V) have also been demonstrated Schott_NL2017 .
In this Letter, we experimentally demonstrate the EF-induced creation and directional motion of a chiral domain wall in a magnetic multilayer with an artificial thickness gradient and interfacial DMI at room temperature, where the EF is directly applied at the metal (Pt)/dielectric (SiO) interface. Note that the metal/dielectric interface has recently been suggested as a unique system to provide a large EF-induced magnetic anisotropy change due to the electric quadrupole induction Miwa_NC2017 , which differs from the case where the magnetic anisotropy is modified by the EF-induced charge doping at the ferromagnet/dielectric interface Maruyama_NN2009 ; Yoichi_APE2009 ; Kihiro_APE2013 . We found that the motion of the chiral domain wall is accompanied with the creation of chiral bubbles Jiang_Science2017 at a certain out-of-plane magnetic field, which acts as a new method for creating chiral bubbles. Micromagnetic simulations help to understand the experimental observations, which indicate that the spatially varied anisotropy profile generated by the thickness gradient, the EF-induced anisotropy change, as well as the DMI play essential roles in the observed domain wall dynamics. Our results are useful for the development of new EF-controlled low-power information devices.
Figure 1a shows the experimental setup, where a [Pt ( nm)/CoNi ( nm)/Pt ( nm)/CoNi ( nm)/Pt ( nm)] multilayer is sandwiched between the indium tin oxide (ITO)/SiO bilayer and the glass substrate (cf. Methods). Due to the broken inversion symmetry at the ferromagnet/heavy metal interfaces Wiesendanger_NRM2016 ; Fert_NRM2016 ; Jiang_Review2017 ; Chen_NC2013 , certain interfacial DMIs are expected to be generated at the Pt/CoNi interfaces. The multilayer structure is fabricated to have a -m-long thickness gradient from the left edge along the direction, which means that the total thickness of the multilayer decreases from nm at m to zero at m. As shown in Fig. 1b, we first measure the out-of-plane hysteresis loop at different locations with different thicknesses using the magneto-optical Kerr effect (MOKE) technique at room temperature. The horizontal distance between the left edge and the Kerr laser spot is defined as . Note that the diameter of the laser spot is determined to be m. The decreasing thickness towards the left edge results in a decreasing saturation magnetization and remanent magnetization . It can be seen that the value of increases with and reaches a constant value when m (cf. Fig. 1b inset), which indicates that the thickness gradient leads to a transition region near the left edge, where the perpendicular magnetic anisotropy (PMA) is reduced, while the region without the thickness gradient has almost constant PMA. The presence of the DMI naturally results in the tilted edge magnetization Wiesendanger_NRM2016 ; Fert_NRM2016 ; Jiang_Review2017 , which further leads to the formation of an in-plane domain near the left edge ( m) due to the reduced PMA in the vicinity of the left edge (cf. Supplementary Fig. 1).
Figure 2 shows the room-temperature out-of-plane hysteresis loops measured by the MOKE technique with m for different amplitudes of the EF. It shows a slight reduction in coercivity when a gate voltage is applied, either positive ( V) or negative ( V). Here, the negative is defined as an electron accumulation at the top surface of the multilayer. When a larger is applied ( V), the measured Kerr signal shows a significantly reduced as well as a deformation of the hysteresis loop, which indicates a decrease of the PMA in the sample. The inset of Fig. 2 shows as a function of . It can be seen that when V, the PMA of the sample significantly reduces, while for V, the PMA basically does not vary with , implying that the negative induced electron accumulations at the Pt/SiO interface could result in the reduced PMA. In the following, we focus on the case of low negative ( V), which is comparable to or smaller than the used in recent studies (cf. Refs. Michihiko_JJAP2006, ; Schellekens_NC2012, ; Chiba_NC2012, ; Bauer_APL2012, ; Haruka_JJAP2013, ; Hsu_NN2017, ; Schott_NL2017, ).
Figure 3 shows the top-view MOKE images of the region with the thickness gradient ( m), where the initial out-of-plane magnetization is pointing along the direction. In Fig. 3a, we first apply the in the presence of an external magnetic field of mT (cf. Supplementary Movie 1). It shows that for V no domain wall is created and the sample remains in the initial magnetization configuration. For V two chiral domain walls are created in the vicinity of the left edge, and the right chiral domain wall moves towards the direction, resulting in the formation of a spin-down (pointing at the direction) domain. When is increased to V, the right chiral domain wall moves further rightwards and the displacement of the chiral domain wall reaches about m. Indeed, the spin-down domain expands when the right chiral domain wall moves rightwards. It is worth mentioning that during the EF-induced motion of the chiral domain wall, we observe the creation of many circular domain wall structures near the right domain wall (cf. the green box in Fig. 3a), of which the diameters are about m. These circular domain wall structures also move with the right chiral domain wall when is varied. When is reduced back to V from V, the right chiral domain wall moves towards the direction. Meanwhile, many circular domain wall structures remain as stable structures in the vicinity of the right chiral domain wall, which indicates the stability and rigidity of these circular domain wall structures. Hence, based on the observation as well as the possible presence of DMI at the Pt/CoNi interfaces Chen_NC2013 , these circular domain wall structures may well be topologically non-trivial chiral bubbles.
When the is further reduced to V, the right chiral domain wall moves very close to the left chiral domain wall, and the spin-down domain shrinks and almost disappears. However, chiral bubbles still exist due to their topologically protected stabilities. Figure 3b and 3c show the EF-induced creation and motion of the chiral domain wall at mT and mT (cf. Supplementary Movies 2 and 3). It shows an almost identical EF-induced domain wall motion compared to Fig. 3a, yet without the accompanying creation of chiral bubbles. The reason could be that a certain magnetic field applied pointing along the -direction is helpful for increasing the size and thus the stability of the chiral bubble, of which the core magnetization is aligned along the -direction in the given sample. For comparison purpose, we also study the magnetic field-induced motion of the chiral domain wall in Fig. 3d by applying an increasing magnetic field along the direction (cf. Supplementary Movie 4), where the magnetization is pointing along the direction initially. It is found that the magnetic field-induced motion of the chiral domain wall is in stark contrast to that induced by the EF. Moreover, no chiral bubbles are found during the chiral domain wall motion.
In order to demonstrate the repeatability of the EF-induced chiral domain wall motion, we adjust the Kerr laser spot to focus on a certain location ( m), where the chiral domain wall is able to reach the Kerr laser spot area, and then apply an alternating , as depicted in Fig. 4a. The maximum and minimum amplitudes of are set as V and V, respectively. The frequency of the applied alternating is set as Hz. At a large , the right chiral domain wall moves forwards, resulting in a spin-down domain underneath the Kerr laser spot. Similarly, at a small , the right chiral domain wall moves backwards, resulting in a spin-up domain underneath the Kerr laser spot. Therefore, the observed Kerr signal is oscillating, as a response to the repetitive motion of the chiral domain wall driven by the alternating , as shown in Fig. 4b. This demonstrates the repeatability of the EF-induced chiral domain wall motion as well as the possibility of using the EF-induced chiral domain wall motion in building binary memory devices.
Having experimentally demonstrated the EF-induced motion of the chiral domain wall, we further study these observed phenomena in a qualitative manner under the framework of micromagnetics (cf. Methods). Figure 5 shows the numerical simulation results on the EF-induced creation and motion of the chiral domain wall. The geometry of the simulated sample is nm. For the purpose of modeling the effect of the thickness gradient, we first assume a linearly increasing along the direction from nm to nm (cf. Fig. 5a). We also assume a linearly increasing PMA along the direction from nm to nm. As shown in Fig. 5b, we assume that the PMA linearly increases from at nm to at nm in the absence of the EF. When the is applied, we assume that the PMA linearly increases from at nm to at nm, showing the reduced PMA induced by the EF but regardless of the exact value of . When the amplitude of is further increased, we assume that the PMA linearly increases from at nm to at nm. The detailed values of and other parameters used in the simulations are given in the Methods.
In Fig. 5c, we simulated the EF-induced creation and motion of a chiral domain wall at mT (cf. Supplementary Movie 5). At the initial state, the relaxed magnetization is pointing along the direction, except the magnetization near the left edge, where it lies in the -plane and is pointing along the direction due to the DMI and reduced PMA (cf. Supplementary Fig. 1). When the PMA is decreased due to the application of the EF (change #1), a chiral domain wall is formed and moves towards the direction. When the PMA is further decreased (change #2), the domain wall continues to move towards the direction. When the PMA is increased back to its original value (changes #3 and #4), the chiral domain wall moves towards the direction and annihilates. It can be seen that the above simulated results agree well with the experimental observed chiral domain wall creation and motion induced by the EF (cf. Fig. 3). However, during the chiral domain wall motion process, no chiral bubbles are found, which differs from the experimental results. The reason could be that both the sizes of the experimental sample and the experimentally created chiral bubbles are much larger than those in the simulation, so that a large dipolar interaction in the experimental sample can help to excite and stabilize chiral bubbles. Here, it is worth mentioning that the EF-induced creation and motion of the chiral domain wall is realizable only when the system has a certain DMI (cf. Supplementary Fig. 2), justifying the presence of DMI in our sample.
In order to numerically show the chiral bubble motion induced by the EF, we manually create a chiral bubble with the topological charge of (cf. Methods) in the sample and relax the sample to the stable/metastable state. As shown in Fig. 5d, when we decrease (changes #1 and #2) and then increase the PMA (changes #3 and #4), the chiral bubble moves forwards and backwards accordingly (cf. Supplementary Movie 6), which proves that the EF-induced PMA change can result in the motion of the chiral bubble, as both the chiral domain wall and chiral bubble are stabilized at a certain anisotropy value. The location corresponding to the certain anisotropy value shifts when the PMA profile is changed by the EF.
In conclusion, using the MOKE technique, we have realized and observed the EF-induced creation and directional motion of a chiral domain wall in a magnetic multilayer with an artificial thickness gradient at room temperature. It is found that a domain wall can be created and directionally displaced about m for a voltage as low as V. The EF-induced chiral domain wall motion is also demonstrated to be repeatable. In addition, we find that the EF-induced motion of the chiral domain wall is accompanied with the creation of chiral bubbles at a certain external out-of-plane magnetic field. Micromagnetic simulations based on the EF-induced anisotropy change are used to help understand the experimental observations, implying the importance of the thickness gradient-induced anisotropy profile, the EF-induced anisotropy change, and the DMI on the domain wall dynamics. Our findings suggest that magnetic multilayers with an artificial gradient in thickness, PMA, and DMI can be harnessed for designing EF-controlled magnetic devices with ultralow energy consumption.
Experimental details. The Pt ( nm)/CoNi ( nm)/Pt ( nm)/CoNi ( nm)/Pt ( nm) films were deposited through a -m-wide and -m-long Ni shadow mask by DC magnetron sputtering under an Ar pressure of Pa after the chamber was evacuated to a base pressure of about Pa. A thickness gradient at the edges of the magnetic stripe was formed due to the thickness of the stencil mask. One side of the magnetic stripe was connected with Ta ( nm)/Cu ( nm)/Ta ( nm). A -nm-thick SiO film was deposited by RF sputtering under Ar + O pressure of Pa to cover the magnetic stripe. ITO pad were deposited above the SiO and magnetic stripe by DC magnetron sputtering under Ar + O pressure of Pa.
Simulation details. The micromagnetic simulation is carried out by using the 1.2a5 release of the Object Oriented MicroMagnetic Framework (OOMMF) OOMMF . The simulator used a number of the standard OOMMF extensible solver (OXS) objects for modeling different micromagnetic energy terms. The OXS extension module for the interface-induced DMI is also used in our simulation. The three-dimensional (3D) time-dependent magnetization dynamics are described by the Landau-Lifshitz-Gilbert (LLG) equation, given as
where is the magnetization, is the saturation magnetization, is the time, is the absolute gyromagnetic ratio, and is the Gilbert damping constant. is the effective field, where the average energy density contains the Heisenberg exchange, DMI, PMA, applied magnetic field, and demagnetization energy terms, which reads
where , , and denote the Heisenberg exchange, PMA, and DMI constants, respectively. is the applied external magnetic field, and is the demagnetization field. is the out-of-plane Cartesian component of the magnetization . is the vacuum permeability constant, and is the unit surface normal vector. In the simulation, we treat the model as a single layer, where the lateral cell size is set as nm. The saturation magnetization is measured to be kA m for the region without the thickness gradient. Other simulation are: , pJ m, mJ m, and MJ m. In our simulations, the effect of the EF is assumed to induce a change of the PMA. The topological magnetization textures are characterized by the topological charge given as
where is the reduced magnetization. The topological charge is also referred to as the skyrmion number. A topologically trivial bubble has , while a topologically non-trivial ground-state skyrmion has .
- (1) Ohno, H. et al. Electric-field control of ferromagnetism. Nature 408, 944-946 (2000).
- (2) Ohno, H. A window on the future of spintronics. Nat. Mater. 9, 952-954 (2010).
- (3) Chiba, D. et al. Magnetization vector manipulation by electric fields. Nature 455, 515-518 (2008).
- (4) Maruyama, T. et al. Large voltage-induced magnetic anisotropy change in a few atomic layers of iron. Nat. Nanotech. 4, 158-161 (2009).
- (5) Yoichi, S. et al. Voltage-assisted magnetization switching in ultrathin FeCo alloy layers. Appl. Phys. Express. 2, 063001 (2009).
- (6) Kihiro, Y. et al. Electric field modulation of magnetic anisotropy in MgO/Co/Pt structure. Appl. Phys. Express. 6, 073004 (2013).
- (7) Miwa, S. et al. Voltage controlled interfacial magnetism through platinum orbits. Nat. Commun. 8, 15848 (2017).
- (8) Chiba, D., Yamanouchi, M., Matsukura, F. & Ohno, H. Electrical manipulation of magnetization reversal in a ferromagnetic semiconductor. Science 301, 943 (2003).
- (9) Shiota, Y. et al. Induction of coherent magnetization switching in a few atomic layers of FeCo using voltage pulses. Nat. Mater. 11, 39-43 (2012).
- (10) Wang, W.-G., Li, M., Hageman, S. & Chien, C. L. Electric-field-assisted switching in magnetic tunnel junctions. Nat. Mater. 11, 64-68 (2012).
- (11) Dohi, T. et al. Effect of electric-field modulation of magnetic parameters on domain structure in MgO/CoFeB. AIP Adv. 6, 075017 (2016).
- (12) Kanai, S. et al. Electric field-induced magnetization reversal in a perpendicular-anisotropy CoFeB-MgO magnetic tunnel junction. Appl. Phys. Lett. 101, 122403 (2012).
- (13) Haruka, K. et al. Switching local magnetization by electric-field-induced domain wall motion. Appl. Phys. Express. 9, 063004 (2016).
- (14) Ando, F. et al. Modulation of the magnetic domain size induced by an electric field. Appl. Phys. Lett. 109, 022401 (2016).
- (15) Michihiko, Y., Daichi, C., Fumihiro, M. & Hideo, O. Current-assisted domain wall motion in ferromagnetic semiconductors. Jpn. J. Appl. Phys. 45, 3854 (2006).
- (16) Schellekens, A. J. et al. Electric-field control of domain wall motion in perpendicularly magnetized materials. Nat. Commun. 3, 847 (2012).
- (17) Chiba, D. et al. Electric-field control of magnetic domain-wall velocity in ultrathin cobalt with perpendicular magnetization. Nat. Commun. 3, 888 (2012).
- (18) Bauer, U., Emori, S. & Beach, G. S. D. Voltage-gated modulation of domain wall creep dynamics in an ultrathin metallic ferromagnet. Appl. Phys. Lett. 101, 172403 (2012).
- (19) Haruka, K. et al. Direct observation of domain wall motion in Co/Pt wire under gate electric field. Jpn. J. Appl. Phys. 52, 070206 (2013).
- (20) Nagaosa, N. & Tokura, Y. Topological properties and dynamics of magnetic skyrmions. Nat. Nanotech. 8, 899-911 (2013).
- (21) Wiesendanger, R. Nanoscale magnetic skyrmions in metallic films and multilayers: a new twist for spintronics. Nat. Rev. Mater. 1, 16044 (2016).
- (22) Jiang, W. et al. Chiral magnetic skyrmions in thin films. Preprint at https://arxiv.org/abs/1706.08295 (2017).
- (23) Fert, A., Reyren, N. & Cros, V. Magnetic skyrmions: advances in physics and potential applications. Nat. Rev. Mater. 2, 17031 (2017).
- (24) Hsu, P.-J. et al. Electric-field-driven switching of individual magnetic skyrmions. Nat. Nanotech. 12, 123-126 (2017).
- (25) Schott, M. et al. The skyrmion switch: turning magnetic skyrmion bubbles on and off with an electric field. Nano. Lett. 17, 3006-3012 (2017).
- (26) Chiba, D., Ono, T., Matsukura, F. & Ohno, H. Electric field control of thermal stability and magnetization switching in (Ga,Mn)As. Appl. Phys. Lett. 103, 142418 (2013).
- (27) Junya, S., Gen, T. & Hiroshi, K. A brief review of field- and current-driven domain-wall motion. J. Phys. D: Appl. Phys. 44, 384004 (2011).
- (28) Jiang, W. et al. Blowing magnetic skyrmion bubble. Science 349, 283-286 (2015).
- (29) Chen, G. et al. Tailoring the chirality of magnetic domain walls by interface engineering. Nat. Commun. 4, 2671 (2013).
- (30) Donahue, M. J. & Porter, D. G., OOMMF User’s Guide, Version 1.0, Interagency Report NO. NISTIR 6376, National Institute of Standards and Technology, Gaithersburg, MD (1999) http://math.nist.gov/oommf.
X.Z. was supported by JSPS RONPAKU (Dissertation Ph.D.) Program. M.E. acknowledges the support by the Grants-in-Aid for Scientific Research from JSPS KAKENHI (Grant Nos. JP17K05490, 25400317 and JP15H05854), and also the support by CREST, JST (Grant No. JPMJCR16F1). W.J. was supported by the National Key R&D Program of China (Grant Nos. 2017YFA0206200 and 2016YFA0302300), the 1000-Youth Talent Program of China, and the State Key Laboratory of Low-Dimensional Quantum Physics. Work carried out at Tsinghua University was also supported in part by the Beijing Advanced Innovation Center for Future Chip (ICFC). Y.Z. acknowledges the support by the National Natural Science Foundation of China (Grant No. 11574137) and Shenzhen Fundamental Research Fund (Grant No. JCYJ20160331164412545). X.L. acknowledges the support by the Grants-in-Aid for Scientific Research from JSPS KAKENHI (Grant Nos. 17K19074, 26600041 and 22360122).
X.L. conceived the idea. X.L., Y.Z. and A.M. coordinated the project. C.M., Y.Y. and X.L. performed the experiments. X.Z. and J.X. carried out the simulations. X.Z. and X.L. drafted the manuscript and revised it with input from M.E., W.J. and Y.Z. All authors discussed the results and reviewed the manuscript. C.M. and X.Z. contributed equally to this work.
Supplementary Information accompanies this paper at <this http URL>. Correspondence and requests for materials should be addressed to X.L. and Y.Z.
Competing financial interests
The authors declare no competing financial interests.
The data that support the findings of this study are available from the corresponding authors upon reasonable request.