Magnetic Force Microscopy Studies of Patterned Magnetic Structures

Transcription

1 3420 IEEE TRANSACTIONS ON MAGNETICS, VOL. 39, NO. 5, SEPTEMBER 2003 Magnetic Force Microscopy Studies of Patterned Magnetic Structures Xiaobin Zhu and Peter Grutter Abstract Magnetic force microscopy is a very powerful tool for studying magnetic nanostructures. In this paper, we demonstrate that magnetic force microscopy is a tool for imaging, manipulating, characterizing magnetization switching and switching-field variability, and studying magnetostatic coupling. Index Terms Hysteresis loop, magnetic force microscopy, magnetic nanostructures, magnetostatic coupling. I. INTRODUCTION RECENT state-of-the-art technology such as electron-beam [1], nanoimprint [1], and interferometric lithography [2], allows for patterning of magnetic nanostructures as small as 10 nm. Unique magnetic properties can thus be tailored by controlling particle shape, size, and spacing [1], [3]. The study of patterned magnetic particles is of great interest due to the fundamental research issues and their potential applications in ultrahigh density storage and magnetoresistive random access memory (MRAM) [1], [2], [4]. This booming interest requires imaging these small structures. The magnetic force microscope (MFM), which has been successfully used in the magnetic storage community for characterizing reading and writing heads and recording media, allows imaging with the spatial resolution of nm [5]. Due to this high spatial resolution and high sensitivity, MFM is an ideal tool for studying the magnetic structures and magnetization reversal of submicron sized magnets [6] [8]. This paper reviews the MFM study of lithographically patterned magnetic particles at McGill University. An emphasis is put on some quantitative information that can be extracted from various examples. II. EXPERIMENTAL TECHNIQUES In MFM, the magnetic contrast is obtained through detecting the force gradient between a ferromagnetic tip and the magnetic sample by amplitude, phase or frequency detection techniques. The principle, instrumentation, and applications of MFM can be found in several review papers [5]. An important issue related to Manuscript received January 9, This work was supported in part by the Le Fonds Pour la Formation de Chercheurs et I Aid a la Recherche de las Province de Quebec (FCAR), the National Sciences and Engineering Research Council of Canada (NSERC), and the Canadian Institute for Advanced (CIAR). X. Zhu is with the Department of Physics, McGill University, Montreal, QC H3A 2T8, Canada, and also with the Department of Physics, University of Alberta, Edmonton, AB T6G 2J1, Canada ( xzhu@ualberta.ca). P. Peter is with the Department of Physics, McGill University, Montreal, QC H3A 2T8, Canada ( grutter@physics.mcgill.ca). Digital Object Identifier /TMAG Fig. 1. Operation modes of MFM: (a) constant frequency shift mode; (b) tapping/lift mode; and (c) constant height mode. the MFM imaging is the separation of the topography and magnetic contrast. Different operation modes, as shown in Fig. 1, can result in different levels of convolution. In the conventional operation mode (constant frequency shift mode) [5], the cantilever resonance frequency is maintained constant by changing the tip-sample separation, as shown in Fig. 1(a). The image obtained in this mode can have a substantial convolution between the magnetic and the topographic signal, especially when a large bias voltage between the MFM tip and sample is applied. The tapping/lift mode can efficiently separate the topography contrast and magnetic signal [9]. As schematically shown in Fig. 1(b), the sample is scanned twice in this mode. The sample topography is obtained in the tapping mode scan using the cantilever oscillation amplitude as a feedback signal. The magnetic contrast is subsequently obtained in the lift mode scan by monitoring the cantilever s frequency or phase shift upon rescanning the previously measured topography with a user controlled height offset. The typical problem associated with the tapping/lift mode scan is that the MFM-tip stray field can produce substantial distortion to the sample magnetic structures during the tapping scan. This is because the MFM-tip stray field close to the tip end is substantial, even for a conventional low magnetic moment tip [10]. Operating MFM in the constant height mode, as schematically shown in Fig. 1(c), can substantially reduce the MFM-tip stray-field-induced irreversible distortion. In this mode, instead of tracking the sample s topography, the tip is scanned across /03$ IEEE

2 ZHU AND GRUTTER: MAGNETIC FORCE MICROSCOPY STUDIES OF PATTERNED MAGNETIC STRUCTURES 3421 the surface at a predetermined constant height while the cantilever frequency shift is monitored. Unavoidable sample tilt is compensated by a tilt correction hardware [10]. Constant height mode has the best signal-to-noise ratio (no feedback noise) and the potential of increased scan speeds; however, the disadvantage of this mode is that a flat sample is needed, as the minimal tip-sample separation is determined by the sample s roughness. Tapping/lift mode and constant height mode are complimentary in acquiring good images. For magnetic particles with a large coercivity field, the tapping/lift mode can be confidently applied; however, for magnetic particles with small switching fields, the constant height mode should be adopted, since the latter produces smaller distortion, and the distortion can be easily detected [10]. The images in this paper were all acquired in the constant height mode. III. VERSATILITY OF MFM A. MFM as a Tool for Imaging The primary function of MFM is to obtain sample magnetic domain structures. In the literature, MFM has been successfully used to obtain the domain structures in lithographically patterned particles [6], [7], [11], [12]. The particle size, magnetocrystalline anisotropy and shape have a strong influence on the magnetic structure [13]. Elongated sub-100 nm particles can form a single-domain state. For example, the black-white contrast in Fig. 1(a) indicates that the particle forms a single-domain state. The black spot indicates the tip is repulsive to the particle, while the white spot indicates it is attractive. As the particle size becomes bigger, the domain patterns such as a flux-closure domain configuration will be formed [13]. If the magnetic particles are disk shaped [12] or elliptical with a small aspect ratio [11], a magnetic vortex state is formed. Such structures have been observed by Lorentz electron microscopy [14], MFM [11], [12], [15] and even spin-polarized scanning tunneling microscopy [16]. Fig. 2(b) (d) shows such a vortex structure of a Permalloy disk with size 700 nm [15]. In Fig. 2(b), the MFM imaging shows the particle forms a low magnetic contrast state with a bright spot in the center, which can be clearly seen in the reduced scan area in Fig. 2(c). This bright spot is a vortex core, which can be proved by imaging in the presence of an external magnetic field. The vortex core moves closer to the edge perpendicularly to the magnetic field direction, as shown in Fig. 2(d). The existence of the vortex core leads to irreproducibility of magnetization switching [17]. Patterning the magnetic disks into a ring shape can avoid the formation of vortex cores [18]. Similar to a magnetic disk, the low magnetic contrast state of a ring is energetically stable, where the magnetic moments rotate circularly around the center of the ring. Beside this state, there exists another energetically stable state in a magnetic ring. This state can be formed at remanence after a large magnetic field is applied [18]. If a large in-plane magnetic field is applied to the ring, the ring will form a single-domain state. As the magnetic field is reduced, domain walls will be nucleated in the ring, and this magnetic moment state is often called the onion state. Depending on the size, thickness and materials of a ring, there are Fig. 2. (a) Single-domain structure of a Ni particle, the particle size: 200 nm2 70 nm215 nm. (b) Vortex structure in a Permalloy disk, the disk diameter 700 nm. (c) Zoom in (140 nm) of (b). (d) Vortex core displacement in the presence of the external magnetic field of 23 Oe. (e) The onion state of a permalloy ring with ring diameter 700 nm. (f) Magnetic moment configuration of (e) obtained through micromagnetic simulation of OOMMF code. The arrow in (d) indicates the magnetic field direction. Fig. 3. (a) Mixture of parallel states and antiparallel states. (b) Two different antiparallel states. Particles size: 550 nm long, 70 nm wide, with composition of NiFe (6 nm)/cu(3 nm)/co(4 nm). Grayscale: (a) 1 Hz and (b) 0.1 Hz. Top: the magnetic moment configuration configurations in each magnetic layers. two possible moment configurations in the domain wall at the onion state : one forms a transverse domain wall, and the other forms a flux-closure domain configuration [19]. Careful MFM imaging can distinguish these two configurations [19]. For example, Fig. 2(e) shows a transverse domain wall configuration of a Permalloy ring. The corresponding micromagnetic simulation of this ring is shown in Fig. 2(f). Besides characterizing the shape influence on magnetic structures, MFM is also suitable to identify the magnetic moment configurations of patterned magnetic multilayers [20]. For example, for elongated sub-100-nm-wide pseudo-spin-valve elements composed of two different magnetic layers separated by a nonmagnetic spacer, the magnetic moment in each layer forms a single-domain state. The magnetic moment configuration of the magnetic layers therefore can either be parallel or antiparallel, as schematically shown on the top of Fig. 3. For the parallel state, MFM imaging shows a strong bipolar contrast, while the stray field above the element almost cancels in the antiparallel state, resulting in a weak contrast. The mixture of parallel state and antiparallel state is shown in Fig. 3(a), while Fig. 3(b) clearly shows the two different antiparallel states coexisting.

3 3422 IEEE TRANSACTIONS ON MAGNETICS, VOL. 39, NO. 5, SEPTEMBER 2003 Fig. 4. MFM images at remanence of a single square shaped ring element after applying different magnetic fields. (a) SEM image of a single square ring. (b) Remanence after saturation with the field indicated in (a). (c) 130 Oe. (d) 147 Oe. (e) 256 Oe. (f) 270 Oe. MFM tip: 30 nm CoPtCr. Lift height: 150 nm; in vacuum. A, B, C, and D symbolize the four segments. The arrows in (b) (f) indicate the magnetic moment orientation in each segment. B. Magnetization Reversal Studied by MFM One distinguishing characteristics of imaging by MFM is that it is compatible with external magnetic fields. The implementation of the MFM in the presence of controlled external magnetic fields [21], [22] can be used to identify magnetic structures and magnetization behavior of patterned particles. There are two different ways to study the magnetization behavior using MFM: imaging at remanence after a magnetic field is ramped to a certain value or imaging in the presence of an external magnetic field. Imaging in the presence of an external magnetic field is necessary when observing domain wall displacement or studying reversible magnetization switching. For example, the vortex core displacement as a function of external magnetic fields can be observed easily in a Permalloy disk [15]. Imaging at remanence can eliminate the combined effect of the tip stray field and an external magnetic field [10], [23]. This technique is very suitable to study the magnetic particles with a few distinct magnetic states, such as single-domain particles, as the exact switching field of the imaged particle can be obtained with the precision determined by the field ramping step [10]. In the following, several examples of imaging at remanence are presented. The domain wall configuration and propagation in patterned ring structures are of great interest in recent studies [17] [19], [24]. Square-shaped ring structures, as shown in Fig. 4(a), are such an example [24]. For a thin and narrow square-shaped ring, each segment of the four edges can be considered as a single domain state. Depending on the orientation of these single domain states, different domain patterns in the square can be formed. The MFM images in Fig. 4(b) (f) shows the different magnetic moment states of a Ni square-shaped ring at remanence after various external magnetic fields are applied. The arrows in the figures indicate the magnetic moment orientation in each segment. The MFM images indicate that the external magnetic field can switch the four segments individually. If the magnetic field is applied with a small angle with respect to the segments and [see the arrows in Fig. 4(a)], the segments Fig. 5. MFM image: (a) at remanence after applying 0304 Oe along the long axis of the element; (b) after applying 0510 Oe; (c) remanent hysteresis loop constructed from MFM images; and (d) hysteresis loop obtained by alternating gradient magnetometry. Particle size: 240 nm 2 90 nm 2 10 nm, lift height 100 nm. Tip: 50-nm CoPtCr; in vacuum. and are switched at much smaller magnetic fields [24] than the segments and. Depending on the magnetic field direction, the domain wall configuration patterns are controllable by showing the bright and dark spots in the figure moving from one corner to another [24]. By observing the domain configuration as a function of external magnetic field, MFM imaging allows a hysteresis loop to be constructed. The hysteresis loop can be constructed by counting the percentage of switched elements as a function of external magnetic fields. For a single-domain particle array, if the MFM-tip stray field does not irreversibly switch a particle from one state to another, the image can be performed at remanence after the magnetic field is ramped to predesignated values. If we assign 1 or 1 to each individual particle based on whether it is switched or not, a normalized hysteresis loop can therefore be constructed according to the number of switched particles as a function of external magnetic field. Fig. 5 shows an example of the hysteresis loop of a single-domain Permalloy particles with size 90 nm 240 nm. Fig. 5(a) and (b) shows the magnetic states at remanence after applying magnetic fields of 304 and 510 Oe, respectively. The remanent hysteresis loop constructed by MFM is shown in Fig. 5(c). This remanent hysteresis loop can be directly compared to the hysteresis loop obtained by alternating gradient magnetometery as shown in Fig. 5(d). The coercivities (310 Oe) obtained by both methods agree very well, which demonstrates the capability of characterizing the magnetization switching behaviors of nanomagnets by MFM. One of the advantages of obtaining a hysteresis loop from MFM images is that it builds a bridge between an individual element and the ensemble, whereas macroscopic techniques only study switching behavior of an ensemble. The study of switching-field variability among different particles is very important for fabricating magnetic particle

4 ZHU AND GRUTTER: MAGNETIC FORCE MICROSCOPY STUDIES OF PATTERNED MAGNETIC STRUCTURES 3423 Fig. 6. Switching-field distribution of widely separated Permalloy disk arrays with the particle size 1:2 m 2 0:2 m 2 30 nm. 600 different individual particles are studied. array with a uniform switching field. With the ability of obtaining precise switching fields of different individual particles, MFM makes it possible to conduct comparison among different particles and to build connections between the magnetization switching and particle microstructures. For example, Fig. 6 shows the switching-field distribution of a widely separated Permalloy particle array [10]. For this particle array, a broad switching-field distribution among different particles is observed. This switching-field distribution is due to switching-field difference of different individual particles. The broadness of switching-field distribution is a common observation for a variety of lithographically patterned structures. Although the elements appear to be identical, they have differences in edge roughness and microstructure, and even modest variability in microstructure or roughness can cause a wide range in the switching field. Above examples demonstrate that MFM combined with an external magnetic field is a unique technique for studying nanoparticle arrays as it can obtain the image of individual particles while studying the collective magnetization behavior. C. Local Hysteresis Loop Technique The investigation of the switching behavior of an individual magnetic element is essential to understand the switching mechanism. MFM imaging in the presence of an external magnetic field is a conventional way to characterize it. However, in some cases, it is possible to characterize the switching field and switching mechanisms of individual particles without imaging. We call this technique a local hysteresis loop technique [15], [20]. As schematically shown in the inset of Fig. 7(a), to perform this measurement, the MFM tip is located at the end of an element, and the cantilever frequency shift is monitored while sweeping the external field. The cantilever frequency shift is proportional to the force gradient between the tip and the sample, and is thus a measure of the local sample moment. This technique is ideally suited to study the particles of submicron size, especially single-domain particles. Fig. 7 shows local hysteresis loops of a pseudo-spin-valve particle (NiFe Cu Co) in the same particle array shown in Fig. 3 [20]. Fig. 7(a) is a minor hysteresis loop of the pseudo-spin-valve element as the magnetic field is swept in the range of 250 Oe. In this field range, only the magnetically soft layer (NiFe) is switched back and forth. The magnetic layers form parallel or antiparallel states, as indicated on the top of Fig. 3. The parallel state gives rise to large frequency shift of cantilever, while the antiparallel state only induces little change of cantilever resonance frequency. Two abrupt frequency jumps in the figure indicates such switching. If the magnetic field, however, is swept in a larger range 550 Oe, both the magnetically hard (Co) and magnetically soft layer (NiFe) can be switched. Two additional frequency shift jumps at larger fields corresponding to the switching of Co layer appear in the major hysteresis loop, as indicated in Fig. 7(b). The hysteresis loops clearly show that the switching of either layer occurs at different magnetic fields, and demonstrate that two layers are coupled by showing asymmetric switching fields in Fig. 7(a). Advantages of this technique include easy characterization of the switching behavior of individual particles [15], [20] and easy comparison among different particles [20]. D. Magnetostatic Interaction and Coupling Magnetic particles will be strongly coupled, if they are placed close to each other. Such coupling gives rise to a change of the switching field, switching mechanism, initial susceptibility, and switching field distribution. Cowburn et al. demonstrates that coupled single-domain magnetic particles could be used to propagate information [25]. In a simple picture, the particles with low coercivities will be switched at smaller magnetic field. The switched particles, however, can act as triggers for the switching of their nearest-neighbor particles. This coupling effect can spatially visualized through careful a MFM imaging. In a closely packed Permalloy disk array arranged in a square lattice [15], we found that the magnetic coupling induces switching-field anisotropy and the switched disks form chain-like structures [15]. This type of coupling can be visualized more clearly in a truncated chain structure [26]. The switching mechanism of a submicron Permalloy disk is a vortex nucleation process (single-domain state to vortex state) and a vortex annihilation process (vortex state to single-domain state) [15]. For the nucleation process, we found that switched elements initiated at the edge of a chain, while the annihilation occurs at the center of a chain. The switched elements form a chain-like structure [26]. Fig. 8 shows the coupling during the annihilation process, starting from the initial vortex state, if the magnetic field of 425 Oe is applied along the disk chain. One of the disks in one chain switches to a single-domain state, as shown in Fig. 8(a). At a larger field of 440 Oe, two additional disks in this chain are switched, as shown in Fig. 8(b). Fig. 8(c) shows the images after a magnetic field of 455 Oe is applied. It can be clearly seen that the switched disks form a chain-like structure. Ideally, the particles in the center of the disk chain is annihilated first, as its neighbor disk applies the highest magnetic field to it. The switched disks themselves produce larger effective fields to their neighboring disks than before, and help their neighboring disks to expel the vortices. Fig. 8(d) shows that all the elements in two chains have almost switched, while there is no element switched in other chains within the scan area.

5 3424 IEEE TRANSACTIONS ON MAGNETICS, VOL. 39, NO. 5, SEPTEMBER 2003 Fig. 7. Hysteresis loop of a single pseudo-spin-valve element of Py(6 nm)/cu(3 nm)/co(4 nm) with size 550 nm 2 70 nm. (a) Minor loop. (b) Major loop. The inset schematically shows the local hysteresis loop technique. Fig. 8. MFM images of Permalloy disk chains after applying the magnetic field of: (a) 425 Oe (b) 440 Oe; (c) 455 Oe; and (d) 489 Oe. Images were taken at magnetic field of 250 Oe. The diameter of the Permalloy disks is 700 nm, and the thickness is 40 nm. The separation between the disks in chains is 100 nm. The switched particles are circled. E. MFM-Tip Stray Manipulation It is well known that MFM tip stray field can produce distortion to submicron sized magnetically soft particles. A severe distortion is that the tip stray field can directly flip a particle moment state [27]. This distortion can used to locally control the sample magnetization states, i.e., to write a bit in patterned magnetic media [1]. The prerequisite of this process is that the effective tip stray field is larger than the switching field of the particle [1]. The writing and reading process can be performed using two MFM tips: the writing tip with a large magnetic moment and the reading tip with very small magnetic moment [1]. Since the MFM-tip stray field applied to particles is tip-sample separation dependent, alternatively, it is possible to use one tip to accomplish the writing and reading process. A suitable MFM-tip moment needs to be selected to accomplish this. For a large Fig. 9. (a), (b) Two MFM images of the same magnetic particles. Tip: 50-nm CoPtCr, lift height h=120nm. Schematic diagram of particle and tip position is shown on the left side of the images. Experimental procedure of performing the writing and reading process is shown at the bottom of the images. tip-particle separation, MFM can be used to read the magnetic moment states of the particle, while at smaller tip sample separation, the tip stray field can be used to control the particle magnetic moment state. This process is demonstrated in Fig. 9 through a Permalloy particle with size nm. For a large tip-sample separation (120 nm) we find that the particle forms a single-domain state. For a small tip-sample separation (60 nm) by positioning the tip at and, respectively, the particle moments can be switched as schematically shown at the bottom of the image. The resulting moment state can be imaged at a large tip-particle separation (120 nm) as shown in Fig. 9(a) and (b). This local tip-manipulation mode is very useful. Besides reading and writing bits in magnetic storage media, it is also possible to use the MFM-tip stray field as a local input signal for a magnetic logical device [25]. The recent technique of operating more than 1000 atomic force microscopes in parallel makes it promising to integrate MFM-tip arrays into a real device [28].

6 ZHU AND GRUTTER: MAGNETIC FORCE MICROSCOPY STUDIES OF PATTERNED MAGNETIC STRUCTURES 3425 IV. SUMMARY AND CONCLUSION MFM is well suited to characterize lithographically patterned nanomagnets. In the imaging mode, MFM can be used to characterize the magnetic structures. MFM is compatible with external magnetic fields, which makes it ideal to study the magnetization reversal of magnetic particles, and obtain precise switching fields of particles. By counting the percentage of switched particles as a function of the external magnetic field, a hysteresis loop comparable with macroscopic magnetometery measurements can be obtained through imaging at remanence or at a suitable magnetic field [15], [29]. The advantage of studying magnetization reversal by MFM over macroscopic measurement is that it directly determines the individual particle differences and studies switching-field distributions. MFM is able to study local hysteresis loop by monitoring the cantilever status versus external magnetic fields, which can be used to study switching mechanism of individual elements, and is potentially ideal to study the switching-field distribution of an individual particles, e.g. thermal switching. The magnetic field emanating from the tip can be used to apply a localized magnetic field to a sample, and can also locally switch the magnetization of a particle. ACKNOWLEDGMENT The authors would like to thank V. Metlushko, B. Ilic, J. Beerens, Y. Hao, F. J. Castano, S. Haratani, B. Vogeli, C. A. Ross, and H. I. Smith for fabricating the samples. REFERENCES [1] S. Y.Stephen Y. Chou, Patterned magnetic nanostructures and quantized magnetic disks, Proc. IEEE, vol. 85, pp , Apr [2] C. A. Ross, H. I. Smith, T. Savas, M. Schattenburg, M. Farhoud, M. Hwang, M. Walsh, and R. J. Ram, Fabrication of patterned media for high density magnetic storage, J. Vac. Sci. Technol. B, vol. 17, pp , Nov./Dec [3] R. P. Cowburn, Property variation with shape in magnetic nanoelements, J. Phys. D: Appl. Phys, vol. 33, p. R1-16, Jan [4] S. A. Wolf, D. D. Awschalom, R. A. Buhrman, J. M. Daughton, S. von Molnar, M. L. Roukes, A. Y. Chtchelkanova, and D. M. Treger, Spintronics: a spin-based electronics vision for the future, Science, vol. 294, pp , Nov [5] P. Grütter, H. J. Mamin, and D. Rugar, Scanning Tunnelling Microscopy II. Berlin: Springer, 1992, pp [6] G. A. Gibson, J. F. Smyth, S. Schultz, and D. P. Kern, Observation of the switching fields of individual Permalloy particles in nanonlithographic arrays via magnetic force microscopy, IEEE Trans. Magn., vol. 27, pp , Nov [7] M. Lederman, G. A. Gibson, and S. Schultz, Observation of thermal switching of a single ferromagnetic particles, J. Appl. Phys., vol. 73, pp , May [8] E. D.E. Dan Dahlberg and J.-G.Jian-Gang Zhu, Micromagnetic microscopy and modeling, Phys. Today, pp , Apr [9] Tapping Lift Mode: Trademark of Digital Instruments, Santa Barbara, CA. [10] X. Zhu, P. Grütter, V. Metlushko, and B. Ilic, Magnetic force microscopy study of electron-beam-patterned soft Permalloy particles: technique and magnetization behavior, Phys. Rev. B., vol. 66, pp , Jul [11] A. Fernandez and C. J. Cerjan, Nucleation and annihilation of magnetic vortices in submicron-scale Co dots, J. Appl. Phys., vol. 87, pp , Feb [12] T. Shinjo, T. Okuno, R. Hassdorf, K. Shigeto, and T. Ono, Magnetic vortex core observation in circular dots of Permalloy, Science, vol. 289, pp , Aug [13] A. Hubert and R. Schäfer, Magnetic Domains: The Analysis of Magnetic Microstructures. Berlin Heidelberg, Germany: Springer-Verlag, [14] M. Schneider, H. Hoffman, and J. Zweck, Lorentz microscopy of circular feromagnetic permalloy nanodisks, Appl. Phys. Lett., vol. 77, pp , Oct [15] X. Zhu, P. Grütter, V. Metlushko, and B. Ilic, Magnetic structures and configurational anisotropy of close packed Permalloy dot array, Appl. Phys. Lett., vol. 80, pp , Jun [16] A. Wachowiak, J. Wiebe, M. Bode, O. Pietzsch, M. Morgenstern, and R. Wiesendanger, Direct observation of internal spin structure of magnetic vortex cores, Science, vol. 298, pp , Oct [17] J.-G. Zhu, Y. Zheng, and G. A. Prinz, Ultrahigh density vertical magnetoresistive random access memory, J. Appl. Phys., vol. 87, pp , May [18] J. Rothman, M. Klaui, L. Lopez, C. A. F. Vaz, A. Bleloch, J. A. C. Bland, Z. Cui, and R. Speaks, Observation of a bi-domain state and nucleation free switching in mesoscopic ring magnets, Phys. Rev. Lett., vol. 86, pp , Feb [19] X. Zhu, P. Grutter, V. Metlushko, and B. Ilic, Magnetic structures and magnetization reversal of patterned Permalloy rings, Phys. Rev. B, submitted for publication. [20] X. Zhu, P. Grütter, Y. Hao, F. J. Castano, S. Haratani, C. A. Ross, H. I. Smith, B. Vogeli, and H. I. Smith, Magnetization switching in 70 nm wide pseudo-spin valve nanoelements, J. Appl. Phys., vol. 93, pp , Jan [21] R. Proksch, E. Runge, P. K. Hansma, and S. Foss, High fieldmagnetic force microscopy, J. Appl. Phys., vol. 78, pp , Sept [22] R. D. Gomez, E. R. Burke, and I. D. Mayergoyz, Magnetic imaging in the presence of external fields: technique and applications, J. Appl. Phys., vol. 79, pp , Apr [23] M. Kleiber, F. Kummerlen, M. Lohndorf, A. Wadas, D. Weiss, and R. Wiesendanger, Magnetization switching of submicrometer Co dots induced by a magnetic force microscope tip, Phys. Rev. B, vol. 58, pp , Sept [24] X. Zhu, P. Grütter, V. Metlushko, and B. Ilic, Control of domain patterns in Nickel square rings, J. Appl. Phys., vol. 93, pp , June [25] R. P. Cowburn and M. E. Welland, Room temperature magnetic quantum cellular automata, Science, vol. 287, pp , Feb [26] X. Zhu, V. Metlushko, B. Ilic, and P. Grutter, Direct observation of magnetostatic coupling of Permalloy chain structures, IEEE Trans. Magn., vol. 39, pp , Sept [27], Systematic study of magnetic tip induced magnetization reversal of e-beam patterned Permalloy particles, J. Appl. Phys., vol. 91, pp , May [28] M. I. Lutwyche, M. Despont, U. Drechsler, U. Dürig, W. Häberle, H. Rothuizen, R. Stutz, R. Widmer, G. K. Binnig, and P. Vettiger, Highly parallel data storage system based on scanning probe arrays, Appl. Phys. Lett., vol. 77, pp , Nov [29] X. Zhu, P. Grütter, V. Metlushko, Y. Hao, F. J. Castano, C. A. Ross, B. Ilic, and H. I. Smith, The construction of hysteresis loops of single domain elements and coupled Permalloy ring arrays by magnetic force microscopy, J. Appl. Phys., vol. 93, pp , June 2003.