effect of defects magnetocrystalline anisotropy and shape anisotropy on magnetic structure of iron thin films by magnetic force microscopy

Effect of defects, magnetocrystalline anisotropy, and shape anisotropy on magnetic structure of iron thin films by magnetic force microscopy Ke Xu, Daniel K Schreiber, Yulan Li, Bradley R Johnson, and[.]

Effect of defects, magnetocrystalline anisotropy, and shape anisotropy on magnetic structure of iron thin films by magnetic force microscopy Ke Xu, Daniel K Schreiber, Yulan Li, Bradley R Johnson, and John McCloy Citation: AIP Advances 7, 056806 (2017); doi: 10.1063/1.4976580 View online: http://dx.doi.org/10.1063/1.4976580 View Table of Contents: http://aip.scitation.org/toc/adv/7/5 Published by the American Institute of Physics AIP ADVANCES 7, 056806 (2017) Effect of defects, magnetocrystalline anisotropy, and shape anisotropy on magnetic structure of iron thin films by magnetic force microscopy Ke Xu,1 Daniel K Schreiber,2 Yulan Li,2 Bradley R Johnson,2 and John McCloy1,a Washington State University, P.O BOX 642920, Pullman, Washington 99164, USA Northwest National Laboratory, 902 Battelle Boulevard, Richland, Washington 99352, USA Pacific (Presented November 2016; received 22 September 2016; accepted 13 November 2016; published online 10 February 2017) Microstructures of magnetic materials, including defects and crystallographic orientations, are known to strongly influence magnetic domain structures Measurement techniques such as magnetic force microscopy (MFM) thus allow study of correlations between microstructural and magnetic properties The present work probes effects of anisotropy and artificial defects on the evolution of domain structure with applied field Single crystal iron thin films on MgO substrates were milled by Focused Ion Beam (FIB) to create different magnetically isolated squares and rectangles in [110] crystallographic orientations, having their easy axis 45◦ from the sample edge To investigate domain wall response on encountering non-magnetic defects, a 150 nm diameter hole was created in the center of some samples By simultaneously varying crystal orientation and shape, both magnetocrystalline anisotropy and shape anisotropy, as well as their interaction, could be studied Shape anisotropy was found to be important primarily for the longer edge of rectangular samples, which exaggerated the FIB edge effects and provided nucleation sites for spike domains in non-easy axis oriented samples Center holes acted as pinning sites for domain walls until large applied magnetic fields The present studies are aimed at deepening the understanding of the propagation of different types of domain walls in the presence of defects and different crystal orientations © 2017 Author(s) All article content, except where otherwise noted, is licensed under a Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/) [http://dx.doi.org/10.1063/1.4976580] I INTRODUCTION It is well-known that microstructures of magnetic materials, including defects and crystallographic orientations, are known to strongly influence magnetic domain structures.1,2 In ferromagnetic materials, the total free energy in the system is dependent on the magnetization distribution in the material, and is the sum of the energy components associated with magnetocrystalline anisotropy, exchange, magnetostatic, magnetic external field, and elasticity.3 Thus, a wide variety of diverse domain structures are expected in ferromagnetic materials Technologically, the manipulation of magnetic domain walls (DW) in ferromagnetic materials has attracted massive interest due to their application in hard disk drives, DW based memories, and spintronic devices Consequently, the theoretical analysis4 as well as experimental testing5,6 of DWs has been widely evaluated in the presence of defects For example, Batista et al.5 observed the pinning and deformation of a Bloch wall in ferrite iron matrix with applied external magnetic field In addition, effects of sample shapes and crystalline orientations on domains have been examined to investigate the magnetization reversal and micromagnetic configurations.7 a john.mccloy@wsu.edu 2158-3226/2017/7(5)/056806/5 7, 056806-1 © Author(s) 2017 056806-2 Xu et al AIP Advances 7, 056806 (2017) To image the static and dynamic magnetic domain wall (DW) movements, magnetic force microscope (MFM) is preferred as it has the sufficient spatial resolution to see domain walls, ∼25 nm MFM samples only require relatively minor surface preparation, and it is less sensitive to sample surface characteristics and contamination A few previous studies proved the advantages of using MFM to measure domain movement in steels Zhang et al.8 applied MFM to observe local phase transformation from austenite to strain-induced martensite for 304 and 310S steels with high resolution IglesiasFreire’s group9 evaluated the domain structures of individual nanostructures and the magnetostatic interaction between nanostripes under the application of external magnetic field using Variable Field MFM (VF-MFM) We have recently studied the magnetic domain structure evolution under external magnetic field for defect-free single crystal Fe thin films magnetized in the easy axis direction by MFM.10 In this paper, we further address the study of in-situ formation and propagation of DWs under external magnetic field, herein considering the effects of shape and magnetocrystalline anisotropy and their interaction Furthermore, the effect of non-magnetic defects on DW nucleation and DW movement are also examined II EXPERIMENTAL METHODS A Sample preparation and microstructure characterization Thin (∼250 nm) Fe films with cm2 area were created using molecular beam epitaxy (MBE), as described elsewhere.11 A dual-beam scanning electron microscope (SEM) FIB (FEI Quanta 650, Hillsboro, OR) was used to create 1-µm deep and 5-µm wide trenches in the Fe thin film to create multiple small geometric samples that could be tested in the MFM.12 The physical trenches in the thin films magnetically isolated each sample from the rest of the Fe film and created boundary conditions for magnetic domain walls in the sample areas The samples were 10 àm ì 10 µm squares and µm × 10 µm rectangles in lateral size, with [110] crystallographic orientations These areas in each film were at least 10 µm apart to prevent any magnetostatic interaction between them.9 Furthermore, FIB was also used to introduce controlled defects by milling a 150 nm diameter hole in the center of some samples to study domain wall response when they encounter these defects The FIB technique is a fast and reliable method to produce specimens, but radiation damage and ion implantation during the milling process can cause changes in magnetic properties, such as the reduction of saturation magnetization along the edges or the introduction of magnetic pinning defects.13,14 In our case, the specimens are large enough15 so that the effect of FIB damage is reflected only in the formation of spike domains along the specimen edges, but the overall domain structures remain intact B Magnetic measurement Magnetic domain structures were measured with MFM using an Asylum MFP-3D with a variable field module (VFM) attachment, allowing application of an external field up to ±8000 Gauss High magnetic coercivity (H c ) CoPt/FePt MFM tips (ASYMFMHC, 32 nm radius of curvature) were used to allow sufficient spatial resolution as well as good magnetic field sensitivity These tips were preferred for this study because their high H c (>5000 Oe) can help to minimize tip demagnetization as well as to eliminate the power dissipation in the magnetic images16 when fields were applied to the Fe samples C Phase field modeling Simulation was performed using the Gauss-Seidel approach and fast Fourier transform (FFT) to numerically solve the Landau–Lifshitz–Gilbert (LLG) equation.4 Details of the calculations can be found in Li et al.10 In addition, the numerical solution to the LLG equation requires the simulation volume to be discretized using a uniform grid To verify the model, the overall grid size was selected to be on the same length scale as the FIB-milled Fe thin film specimens To ensure convergence, simulations were iterated multiple times for each applied field level with a normalized time step ∆t*=0.1 056806-3 Xu et al AIP Advances 7, 056806 (2017) III RESULT AND DISCUSSION In previous research, magnetic domain states were studied by MFM for [100] square and [100] rectangular samples in Fe thin films, while applying an in-plane, easy axis oriented field.10 [110] samples have their easy axes 45o from the sample edge, so the domain structure of these samples are not as regular as corresponding [100] samples Fig 1(a) shows the formation and propagation of magnetic domain structure in a 10 µm x 10 µm square sample as a function of descending external magnetic field Between 2000 and 1500 G, the domain structure does not change significantly, indicating that the sample is near its saturated state in which all domain walls should be theoretically erased However, inadvertent ion implantation damage during the FIB milling process could give rise to the localized variations in demagnetization energy along the edges of the sample The consequences of these edge defects on the domain structure depend on the orientation of the field with respect to the edge In the case of the edges parallel to the field, spike domains are created, providing additional nucleation points for new domains on reducing the field to zero In the case of the edges perpendicular to the applied field direction, the FIB edge defects result in residual domain walls parallel to the field direction These can be seen in both the square specimen Fig 1(a) and the rectangular one in Fig 1(b) These features caused by the extra pinning defects along the sample edges are not fully removed even at the maximum applied field As the field decreases to 1000 G, a few domains start nucleating from the edges, and the domain with the opposite magnetization direction increases in size at the expense of the opposite domains At 400 G, two vortices form from the edges and gradually shift towards the center While the magnetization aligns primarily with the easy axes to reduce the magnetocrystalline anisotropy energy, at the sample periphery the local magnetization aligns with the sample edge to reduce the magnetostatic energy Consequently, the magnetic domain structure at G in Fig 1(a) is the total energy balanced state of the system, forming a symmetric vortex The simulation result in Fig 1(a) confirms the validity of the measured domain structure at remanence, with the black arrows are pointing to the magnetization directions in selected domains Note that in the simulation no extra pinning sites, such as could be present from FIB, were added Since the sample edges are away from the magnetic easy axes (i.e., [110] orientation), the competition between magnetocrystalline anisotropy energy and magnetostatic energy is stronger than in the case where the sample edges are along the magnetic easy axes (i.e., [100] orientation) The spike domain formation as influenced by the FIB tends to be more pronounced in [110]-oriented as opposed to previously studied [100]-oriented samples Presumably this observation is due to the interaction between the misorientation of the edges with the easy axis in the [110] samples, thus represents a combined effect of magnetostatic and magnetocrystalline effects In Figure 1(b), the µm x 10 µm rectangular exposed to a decreasing magnetic field along the long axis from 3000 G to G is shown To minimize the magnetostatic energy that associates FIG Magnetic domain morphologies as a function of applied field for (a) square and (b) rectangular geometries The applied field directions and scale bars are indicated in the figure for both geometries Simulation image for each geometry provides the domain structures formed when the applied field is subsequently reduced to G 056806-4 Xu et al AIP Advances 7, 056806 (2017) with it, the zero field rectangular specimen in Fig 1(b) has a distinctly different domain morphology compared to the zero field square specimen in Fig 1(a), because the demagnetizing fields along each axis in the square specimen are equally distributed, and therefore it forms a symmetric domain structure at remanence In comparison, demagnetizing fields along the long axis are weaker than along the short axis for a rectangular specimen, creating the energy balanced asymmetric domain structure in Fig 1(b) at G, which has very good agreement with the simulation pattern as shown In general for nonspherical specimens, the free poles that exist outside the specimen will generate an unequal demagnetizing field in the presence of an applied magnetic field The demagnetizing field thus depends on specimen geometries, applied field direction, and easy axis orientation Beyond varying geometries and alignment between the crystallographic and magnetic axes, the next level of complexity is to introduce a controlled defect in the structure which can act as a pinning site for domain walls Such nonmagnetic defects in magnetic materials in real applications could be microstructural features such as voids or precipitates A series of MFM images, for square and rectangular specimens with holes, taken at decreasing field, are shown in Figure The initial domain pattern for the square with hole at 3000 G is shown in Fig 2(a) It has no spike domains and fewer DWs than the equivalent domain pattern in Fig 1(a) at 2000 G, because the external magnetic field energy is large enough to de-pin the DWs from the edges (compare 2000 G in Fig 2(a) which is similar to the max field shown in Fig 1(a)); however, the field is still not sufficiently large to overcome the pinning energy from the center hole, so the single DW across the hole remains As the field slowly decreases, the spike domains reform, several other DWs form above and below the hole, and the walls going through the hole not move away from it compared to the equivalent walls in Fig 1(a) As the magnetic field further reduces to 400 G, energy from the external field decreases, magnetostatic energy then increases to balance the energy in the system, and consequently the partitioned and circular vortex domain is reformed at G This zero field state agrees closely with the simulation pattern in Figure 2(a) except some minor features in the corners which can be attributed to FIB milling defects effects in the experimental measurements The initial quasi-saturated domain wall structure in Fig 2(b) for the rectangular [110] specimen with hole is similar but distinguishable from Figure 1(b), and therefore the geometries of domain structures at lower fields diverge due to different starting configurations At 2000 G, two DWs are pinned by the hole As the strength of the applied field decreases, magetostatic energy retrieves its priority in the energy balance, nucleating spike domains along the edges at 1000 G and releasing the DWs from the hole at 800 G A closure domain is formed at 400 G to further balance the energy system At G, the domain structure is distinct from Fig 1(b) and is slightly different compared to the simulated pattern The domain structural differences at remanence can be attributed to path-dependent hysteresis and hence path-dependent domain structures Therefore, one can state that the physical FIG Magnetic domain morphologies as a function of applied field for (a) square and (b) rectangular geometries A 150 nm diameter hole is located in the center of each sample The applied field directions and scale bars are indicated in the figure for both geometries Simulation image for each geometry provides the domain structures formed when the applied field is subsequently reduced to G 056806-5 Xu et al AIP Advances 7, 056806 (2017) boundary constraints, which control the shape anisotropy and magnetostatic energy, dominate at low field, but at higher field the effect of external field become significant enough to dominate the domain structure This explains the similar domain structures at high field, regardless of the presence or absence of a defect Overall, the defect acts as a pinning site which should alter the movement of domain walls, and therefore the domain structure, during the loading of magnetic field Since the domain evolution is path dependent, the domain structures will be different with and without defects, as the defects are affecting the path at low fields When domain walls are pinned by defects, they require more energy to unpin them, so a stronger external field is required for full saturation IV CONCLUSIONS In this work, MFM measurements were performed as a function of an external magnetic field on FIB-milled single crystal [110] square and [110] rectangular samples created in Fe thin films The crystal orientation study looked at the effect of magnetocrystalline anisotropy, while differences in square and rectangular samples simultaneously assess the effects of shape anisotropy Various domain structures were observed in measurements due to the energy competition (anisotropy, exchange, magnetostatic) in the sample, creating multiple metastable minima in the energy landscape This resulted in diverse domain patterns at the zero field remanence state In addition, the presence of a defect appeared to have an impact on the DW movement, therefore the domain structure at the same external magnetic field will not be the same with no-defect samples at lower magnetic field, since the consideration of pinning energy is required In fact, the energy balance inside of the sample is then broken, which alters the demagnetization, anisotropy, and exchange energies, further affecting the domain morphologies The described single crystal Fe film work is being used to stepwise validate and improve simulation models Ongoing studies are assessing domain structures in Fe in the presence of other defects resulting from high dislocation densities as well as grain boundaries of different orientations ACKNOWLEDGMENTS The authors would like acknowledge funding from the Department of Energy, Office of Nuclear Energy’s Nuclear Energy Enabling Technologies (NEET) program A portion of the research was performed at Pacific Northwest National Laboratory (PNNL), operated by Battelle Memorial Institute for the U.S DOE under contract DE-AC05- 76RL01830 A portion of this research was performed using the Environmental Molecular Sciences Laboratory (EMSL), a national scientific user facility sponsored by the DOE’s Office of Biological and Environmental Research and located at PNNL D.-G Park, J.-H Hong, I.-S Kim, and H C Kim, J Mater Sci 32, 6141 (1997) C Yu, Y E Kim, C G Kim, G P Han, and M C Paek, Mol Phys Reports 32, 76 (2001) S Chikazumi and C D Graham, Physics of Ferromagnetism (Oxford University Press, 2009) J X Zhang and L Q Chen, Acta Mater 53, 2845 (2005) L Batista, U Rabe, and S Hirsekorn, NDT&E Int 57, 58 (2013) R D Gomez, I D Mayergoyz, and E R Burke, IEEE Trans Magn 31, 3346 (1995) J Yu, U Ră udiger, A D Kent, L Thomas, and S S P Parkin, Phys Rev B 60, 7352 (1999) L Zhang, B An, S Fukuyama, T Iijima, and K Yokogawa, J Appl Phys 108, 063526 (2010) O ´ Iglesias-Freire, M Jaafar, L P´erez, O de Abril, M V´azquez, and A Asenjo, J Magn Magn Mater 355, 152 (2014) 10 Y Li, K Xu, S Hu, J Suter, D K Schreiber, P Ramuhalli, B R Johnson, and J McCloy, J Phys D Appl Phys 48, 305001 (2015) 11 Y Cao, K Xu, W Jiang, T Droubay, P Ramuhalli, D Edwards, B Johnson, and J McCloy, J Magn Magn Mater 395, 361 (2015) 12 J D Suter, P Ramuhalli, J S McCloy, K Xu, S Hu, Y Li, W Jiang, D J Edwards, A L Schemer-Kohrn, and B R Johnson, AIP Conf Proc 1650, 1476 (2015) 13 M A Basith, S McVitie, D McGrouther, J N Chapman, and J M R Weaver, J Appl Phys 110, 083904 (2011) 14 J Mayer, L A Giannuzzi, T Kamino, and J Michael, MRS Bull 32, 400 (2007) 15 S Getlawi, M R Koblischka, F Soldera, and U Hartmann, Microsc Microanal 13(S3), 358 (2007) 16 S H Liou and Y D Yao, J Magn Magn Mater 190, 130 (1998) S  ...AIP ADVANCES 7, 056806 (2017) Effect of defects, magnetocrystalline anisotropy, and shape anisotropy on magnetic structure of iron thin films by magnetic force microscopy Ke Xu,1 Daniel K Schreiber,2... field, herein considering the effects of shape and magnetocrystalline anisotropy and their interaction Furthermore, the effect of non -magnetic defects on DW nucleation and DW movement are also examined... such as magnetic force microscopy (MFM) thus allow study of correlations between microstructural and magnetic properties The present work probes effects of anisotropy and artificial defects on the