Magnetization Ratchet in Cylindrical Nanowires

Transcription

1 ASSOCIATED CONTENT *S Supporting Information Magnetization Ratchet in Cylindrical Nanowires Cristina Bran* 1, Eider Berganza 1, Jose A. Fernandez-Roldan 1, Ester M. Palmero 1, Jessica Meier 1, Esther Calle 1, Miriam Jaafar 1, Michael Foerster 2, Lucia Aballe 2, Arantxa Fraile Rodriguez 3,4, Rafael P. del Real 1, Agustina Asenjo 1, Oksana Chubykalo- Fesenko 1 and Manuel Vazquez* 1 1. Institute of Materials Science of Madrid, CSIC Madrid. Spain 2. ALBA Synchrotron Light Facility, CELLS Barcelona. Spain 3. Departament de Física de la Matèria Condensada, Universitat de Barcelona Barcelona. Spain 4. Institut de Nanociència i Nanotecnologia (IN2UB). Universitat de Barcelona Barcelona. Spain Electrochemical Synthesis of Nanowires FeCo/Cu multilayer nanowires were grown into the pores of anodic aluminum oxide (AAO) membranes. These templates were previously synthesized by hard anodization [1] on high purity Al foils using oxalic acid (0.3M) as electrolytic bath at and keeping the temperature constant at 0º C. During the anodization, a constant voltage of 80 V was first applied for 500 s to produce a protective aluminum oxide layer at the surface of the disc which avoids breaking or burning effects during subsequent hard anodization [1, 2]. After that, the voltage was steadily increased (0.08 V s-1) up to 130 V and kept constant for 2700 seconds. FeCo/Cu nanowires were grown into the 120 nm diameter pores of AAO templates, at room-temperature, by pulsed electrodeposition [3,4] from sulfate-based electrolytes containing 0.12 M CoSO4 7H2O M CuSO4 5H2O M FeSO4 7H2O M H3BO M C6H8O6 [3]. The ph value was kept constant at about 3.0. Different potentials were used during the nanowires growth: 1.8 V for 30 seconds for FeCo layer and 0.6 V for 45 seconds for Cu layer, both versus Ag/AgCl reference electrode. The nanowires, released by chemical etching from the

2 template, were measured individually. The alumina template was dissolved using an aqueous solution composed by chromic oxide and phosphoric acid. Magnetic Force Microscopy The surface magnetization reversal process was investigated under in-situ applied magnetic field along the nanowires (up to a maximum available field of ± 700 Oe) by using an advanced MFM operational mode, as detailed elsewhere [5, 6]. The MFM scan is done repeatedly along the same line. While the image is recorded, the applied field sweeps between -700 and 700 Oe in tens of seconds in such a way that the changes in magnetization are tracked and the magnetization reversal process is visualized. Figure S1. (a) Sketch of the area imaged in (b) corresponding to the end of the nanowire with shorter segment. Inside the dashed square, small jumps are observed where magnetization reversal propagates segment by segment from shorter to longer segments. Figure S1 provides further evidence of the segment-by-segment reversal starting from the shorter segments as confirmed after imaging several nanowires. This effect is noticed due to accumulation of magnetic charges at the FeCo/Cu interfaces. Figure S1b shows a non-standard MFM image of the end with shorter segments. Note that a strong contrast arises from the high stray field of the nanowire end. Attention should be paid to the subtle stair-like contrast propagating to the right between 340 Oe and 370 Oe region inside the dashed square. This dark contrast corresponds to a configuration similar to the head-to-head domain wall. After a simple analysis of the critical fields and preferential sites in a particular nanowire we conclude certain randomness in the magnetization reversal of this chain of

3 segments, since subsequent field sequences in equivalent conditions yield to different reversal processes. However, the experimental results shown in Fig. S2 demonstrate that to a certain extent, the control over the domain configuration is possible. Notice that the studied region is the same as the one presented in Fig. S1b. Figures S2a and S2b complete a minor hysteresis loop where at the maximum applied magnetic field (+/- 550 Oe) the magnetic moments of all the segments are parallel. However, when less strong maximum magnetic field (+/-500 Oe) is applied (see Figs. S2c and S2d), we achieve to stop a head-to-head configuration (dark contrast) and which moves back to the left end of the nanowire. This control is fully necessary for the development of domain wall based information storage technologies. Figure S2. (a), (b), (c) and (d) are non-standard MFM images corresponding to four branches of two subsequent minor hysteresis loops (Cycle 1 and Cycle 2) performed in the same area shown in Fig. S1b. The field sweeps between +/-550 Oe in (a) and (b) and between +/-500 Oe in (c) and (d). Red and blue lines account for the positive or negative magnetization direction of the nanowire segments XMCD-PEEM The samples are illuminated with circularly polarized X-rays at a grazing angle of 16 0 with respect to the surface, at the resonant L3 absorption edge of Co (778 ev). The

4 emitted photoelectrons used to form the surface image are proportional to the X-ray absorption coefficient and thus the element-specific magnetic domain configuration is given by the pixel-wise asymmetry of two PEEM images sequentially recorded with left- and right-handed circular polarization [7]. What is actually imaged is the projection of the local magnetization on the photon propagation vector so that ferromagnetic domains with magnetic moments parallel or antiparallel to the X-ray propagation vector appear bright or dark in the XMCD image while domains with magnetic moments at a different angle have an intermediate gray contrast. Furthermore, the particular cylindrical shape of the wires allows for a fractional amount of X-rays to be transmitted through the nanowire, generating photoemission from the Si substrate. By analyzing the circular dichroic or pseudo-magnetic contrast formed in transmission in the shadow area, information about the magnetic configuration in the bulk of the wire can then be obtained [2, 8]. For an explanatory sketch of the magnetic contrast observed in direct photoemission and transmission, the reader is referred to Figure 6 in Ref. 2. Notice that the dark contrast in transmission is equivalent to the bright one in direct photoemission, since the absorbed and transmitted X-rays are complementary. Micromagnetic modelling of the hysteresis cycle The modelled axial hysteresis loop is presented in Fig. S3 (bottom left). At the remanent state (Fig. S3 a), the nanowire is nearly uniformly magnetized longitudinally. At the indicated cross sections near the ends (L and R) of each segment, pairs of vortices nucleate with the core parallel to the saturation direction and opposite chiralities. The vortices also nucleate with opposite chiralities in adjacent magnetic segments. Under the application of small reverse fields, the overall magnetization is not largely altered with the exception of local reversible helicoidal reorientations of the magnetic moments mostly at the surface near the segment ends and the vortices slowly extend inside the segments. At H = 360 Oe (Fig. S3 b), the pair of vortices in the shortest segment annihilate each other at the outer shell and results in a skyrmion-like or core/shell configuration with reversed magnetization in an external shell along the entire segment. A similar second vortex pair annihilation occurs in the second shorter segment at higher field H = 425 Oe (Fig. S3 c) leading to the formation of a second single skyrmion tube. Unlike the previous processes, for the 3rd and 4th shorter segments a gradual partial reversal is first observed, which is followed by the irreversible formation of a pair of skyrmion tubes with opposite chiralities in each segment at H=445 Oe. This double

5 skyrmion tube configuration collapses at H=460 Oe. At this stage the local magnetization reverses in both segments inducing the reversal along the remaining longer segments. As the applied field further increases, the reversal is accomplished with the gradual reduction of the cores of the skyrmion tubes in the two shortest segments and their collapse at 750 and 970 Oe, respectively. Figure S3. Bottom left, modeled hysteresis loop and a zoom view of the irreversibility area in the inset. Figures (a)-(e) show the longitudinal component of magnetization, m x, at the respective points marked in the loop. Cross sections close to the Left (L) and Right (R) ends of each individual segments are displayed colored according to the scale bar. Their chirality is denoted by pale orange (leftwards/counter Clockwise) and pale blue (rightwards/clockwise) arrows. Vertical green arrows point the Cu segments where the reversal process is pinned. According to the simulations, and apart from the indicated local reversible magnetization rotations, the remagnetization proceeds in a stepwise propagation of irreversible processes starting at the shorter segments. Remarkably, although started first, the remagnetization in shorter segments is the last one to be completed. From the state d to the state e magnetization changes occur in one field step although dynamically they again propagate unidirectionally as indicated in the main step. The pinning mechanism observed here in first four segments is similar to the corkscrew -pinning observed in FeCo nanowires with modulations in diameter [9] which is characterized by the helicoidal tubes formed by the 3D skyrmions in each segments. Unlike the case of the geometrical modulation, the compositional modulation introduced by the narrow Cu spacers here produces a less pronounced helicity of the

6 skyrmion tubes (smaller displacement of the core from the nanowire axis), especially for the short segments which exhibit almost straight skyrmion tubes. The present result confirms that the corkscrew mechanism is not exclusive from the geometrical modulation and can be obtained by alternative strategies that introduce local modifications of the magnetostatic energy. REFERENCES [1] Vega V.; Garcia J.; Montero-Moreno J.M.; Hernando B.; Bachmann J.; Prida V.M. and Nielsch K. Unveiling the Hard Anodization Regime of Aluminum: Insight into Nanopores Self-Organization and Growth Mechanism. ACS Applied Materials and Interfaces 2015, [2] Bran, C.; Berganza, E.; Palmero, E. M.; Fernandez-Roldan, J. A.; Del Real, R. P.; Aballe, L.; Foerster, M.; Asenjo, A.; Fraile Rodríguez, A.; Vazquez, M. Spin Configuration of Cylindrical Bamboo-like Magnetic Nanowires. J. Mater. Chem. C 2016, 4, [3] Palmero E.M.; Bran C.; del Real R.P.; Magen C. and Vazquez. Structural and Magnetic Characterization of FeCoCu/Cu Multilayer Nanowire Arrays M. IEEE Magnetics Letters 2014, 5, [4] Bran C.; Ivanov Yu.P.; Kosel J.; Chubykalo-Fesenko O. and Vazquez M. Co/Au Multisegmented Nanowires: a 3D Array of Magnetostatically Coupled Nanopillars. Nanotechnology 2017, 28, [5] Jaafar, M.; Serrano-Ramon, L.; Iglesias-Freire, O.; Fernández-Pacheco, A.; Ricardo Ibarra, M.; De Teresa, J. M.; Asenjo, A. Hysteresis Loops of Individual Co Nanostripes Measured by Magnetic Force Microscopy. Nanoscale Res. Lett. 2011, 6, [6] Berganza, E.; Bran, C.; Jaafar, M.; Vazquez, M.; Asenjo, A. Domain Wall Pinning in FeCoCu Bamboo-like Nanowires. Sci. Rep. 2016, 6, [7] Aballe, L.; Foerster, M.; Pellegrin, E.; Nicolas, J.; Ferrer, S. The ALBA Spectroscopic LEEM-PEEM Experimental Station: Layout and Performance. J. Synchrotron Radiat. 2015, 22, [8] Kimling J.; Kronast F.; Martens S.; Bohnert T.; Martens M.; Herrero-Albillos J.; Tati-Bismaths L.; Merkt U.; Nielsch K. and Meier G. Photoemission Electron Microscopy of Three-dimensional Magnetization Configurations in Core-Shell Nanostructures. Physical Review B 2011, 84, [9] Fernández-Roldán, J. Á.; Perez del Real, R.; Bran, C.; Vázquez, M.; Chubykalo-Fesenko, O. Magnetization Pinning in Modulated Nanowires: From Topological Protection to the corkscrew Mechanism. Nanoscale 2018, 10,