Magnetic properties of hybrid nanostructures 2

Chapter Fabrication of Magnetic Gradient Nanomaterial by Combined Nanocluster Beam Deposition and Colloidal Crystals Patterning The use of reactive-ion-etched (RIE) colloidal crystals as templates for creating hollow magnetic nanostructures is demonstrated for Ni and FePt. By controlling the number of layers of colloidal crystals and the RIE conditions, two-dimensional ordered arrays of magnetic micro/nanostructures with different designed morphologies, e.g. nanodot arrays, flower-bud arrays, hexagonal arrays, hollow hemispherical caps arrays, etc., are fabricated. In addition, the monolayer of colloidal crystals coated with an underlayer was used as a topographic pattern for nanocluster beam deposition of a magnetic alloy, FePt, creating a magnetic gradient nanomaterial consisting of periodic FePt hollow hemispherical arrays. The combined effect of composition modulation of the FePt nanoclusters mediated by the underlayer and the unique hollow morphology of the hemispherical caps were investigated by magnetic force microscopy (MFM) and vibrating sample magnetometry. We have demarcated the regime where the quasi-single domain/bubble domain or multidomain flux-closure states manifest by the variation of the underlayer on the microspheres from Fe/Co to Ag. This versatile method of creating magnetic gradient nanomaterial by patterning of colloidal crystals provides a fundamental platform for the investigation of dynamic phenomena of patterned ferromagnetic elements at the nanoscale, and has great potential for novel magnetic nanodevice fabrication. 53 3.1 Introduction The rapid advances in fabricating nanostructures with controlled size and shape offered by modern lithography techniques have triggered increased research in magnetic nanostructures. Well-defined magnetic nanostructures have motivated intensive research activities not only for their fundamental scientific interest, which is derived from their size-dependent properties, but also for their potential applications which include nonvolatile magnetic random access memory, field programmable spin logic, patterned media for ultra high density data storage, biomedicine, and functional building blocks for nanoscale devices. In colloidal lithography (CL), colloidal arrays are used as lithographic masks or templates to fabricate nanostructures, 5, and have the advantages of being an inexpensive, inherently parallel, high-throughput nanofabrication technique. The earliest CL methods utilized the interstices of the close-packed single and double layers of monodispersed colloidal particles for the infiltration of metals or organic materials to form hexagonal arrays in a single layer or triangular arrays in a double layer. To allow better control over the size and shape of nanostructures, techniques including angle-resolved deposition, dry-etching conditions, as well as reactive ion etching (RIE) method were introduced. By changing the RIE conditions such as the etching time, etchant composition, and tilt angle, the size, shape, and feature spacing of the colloidal mask layer and the subsequent magnetic metal dot arrays can be controlled. 10 54 Colloidal crystal assemblies have been used as masks or templates for fabricating ferromagnetic nanostructures. The periodicity of photonic crystals, besides having a direct influence on the photonic band gap, is also useful for patterning nanostructures, which can have implications on the magnetic properties via inter-particle magnetostatic interactions. An enhanced coercivity relative to that of the continuous film was reported for 3D ordered nickel replicas of colloidal crystals fabricated by electrodeposition into colloidal crystal templates, 11 as well as for Co/Pt multilayer patterned by colloidal lithography (CL) via a RIE process. 12 Zhong et al. used high resolution MFM to characterize the magnetic fine structures of in-plane magnetized FePt dots patterned by nanosphere lithography. 13 A special possibility offered by colloidal crystals templates is the creation of hollow nanostructures. Hollow magnetic spheres can exhibit interesting behavior such as possessing single-domain behavior at a larger diameter than a solid sphere, 14 as well as exhibiting anisotropic distribution of magnetization due to the curved surface. The non-conformal coating of the film over the sphere results in an anisotropic distribution of magnetization, and leads to unique magnetic switching. 15 , 16 , 17 , 18 In addition, complex magnetic domain structures can be achieved when there is competing contributions from magnetostatics, shape anisotropy and magnetocrystalline anisotropy because of minimization of the overall energy in such system. FePt-based building blocks have attracted considerable attention due to the high magnetic anisotropy of the L10-ordered phase. This makes it possible to extend the superparamagnetic limit to diameters of less than nm, which is of utmost importance in ultrahigh-density (10 Tbit/in.2) magnetic recording media. Since ferromagnetic 55 nanomaterials commonly tend to aggregate in solution, it is not easy to obtain high density coverage of nanoparticles on polymeric templates as well as assemble these coreshell structures into ordered arrays. One example of a rational design of appropriate chemical interactions between the nanoparticles and template is the assembly of FePt nanoparticles on the surfaces of spherical silica particles with controlled core–shell morphology by the layer-by-layer (LbL) techniques using oppositely charged polyelectrolytes. 19,20 The alternative approach of nanocluster beam deposition of FePt nanoclusters on colloidal crystals monolayer as a template serves to address these difficulties faced in solution methods in producing ordered arrays of hollow FePt hemispherical structures. The nanocluster beam deposition uses gas phase aggregation to separate the nucleation and growth stages in both time and space domain in order to fabricate monodispersed FePt nanoparticles. 21 The versatility of the nanocluster beam deposition technique is that it allows control of the sizes of the FePt nanoclusters by changing the deposition power, aggregation length and aggregation gas flow rate and is capable of producing well dispersed nanoclusters under surfactant-free conditions. Magnetic NPs produced by wet chemical methods are usually capped by surfactants to improve the dispersion properties in solution. To study the magnetic properties of materials dispersed in solution require recovery of sub-grams quantities of the material in powder form, with inevitable coagulation of the nanomaterials. In contrast, using the nanocluster beam deposition method, the magnetic properties can be directly measured on the substrate. 56 In this chapter, we demonstrate a versatile method to assemble magnetic nanoparticles of hollow geometries into ordered arrays by using the combined route of sputtering or nanocluster beam deposition and RIE-assisted colloidal lithography. First, we demonstrate that two-dimensional ordered arrays of Ni micro/nanostructures with different designed morphologies, e.g., nanodot arrays, flower-bud arrays, hexagonal arrays, hollow hemispherical caps arrays, etc., could be prepared by the route of the combined RIE and Colloidal Lithography method. In the second part, the RIE-modified colloidal crystals coated with Ag and Fe/Co underlayers was used as a topographic pattern for nanocluster beam deposition of a magnetic alloy, FePt, creating a magnetic gradient nanomaterial 22 that consists of periodic FePt hollow hemispherical arrays. The combined effect of composition modulation of the FePt nanoclusters mediated by the Ag and Fe/Co underlayers and the unique hollow morphology of the hemispherical cap on magnetic properties and domains of the FePt hemispherical caps was investigated by magnetic force microscopy (MFM) and vibrating sample magnetometry (VSM). 57 3.2 Experimental Section 3.2.1 Formation of colloidal crystal assembles and Reactive-ion etching using CF4/O2 Densely packed two-dimensional arrays of monodisperse spherical polystyrene (PS) nanospheres (Polysciences) of diameter 600, 356 and 200 nm were formed by selfassembling a 4% (by weight) solution upon slow evaporation of a solvent under ambient conditions. The number of colloidal layers was adjusted by the PS nanospheres concentration. For single layer formation, the PS nanospheres were spin-coated onto the Si wafer. After the formation of the close-packed nanosphere mask, the shape of the PS nanospheres was tuned by RIE with a mixture of tetrafluromethane (CF4, 40 sccm) and oxygen (O2 60 sccm) in a SAMCO RIE-10N reactive ion etching unit for to 20 min. The forward RF power was typically 20 W and reflected power was 1-2 W. The addition of O2 into CF4 will increase the fluorine free radical concentration of the plasma and increase the etching rate. 23 A mixture of O2 and CF4 generate the oxyfluoride ion, which is a highly reactive etching agent for cutting the carbon-carbon bonds in the polymeric substances, 24 to produce polar low-molecular-weight fragments. Because of RIE etching through the voids of the interstices, the central part of each PS bead in the bottom layer was preferentially etched, and divides the particles of the bottom layer into three parts. The advantages of the use of RIE on colloidal crystals is that size and shapes of binary (or ternary) colloidal particle arrays can be controlled by the crystal orientation, the number of colloidal layers, and the RIE conditions. 25 , 26 An exact assessment of a faced-centered cubic (FCC) crystal structure or a hexagonal-closed 58 packing (HCP) crystal structure can be done by simple and direct SEM of the CL patterned structures. 3.2.2 Fabrication of colloidal crystal patterned Ni and FePt nanostructures by RF sputtering and FePt nanocluster beam deposition Ni, Fe, Co or Ag thin films were deposited onto the masked substrate by RF magnetron sputtering at a base pressure of × 10-7 torr. FePt nanoclusters were deposited using the Oxford Applied Research NC200-UHV Nanocluster Beam Deposition System at a base pressure of × 10-8 torr. The experimental setup comprised of a liquid nitrogen cooled nucleation chamber, an aggregation tube, and a deposition chamber. The exterior view of the assembled nanocluster beam deposition system and schematic drawing of the nanocluster beam deposition system was shown in figure 3.1. A magnetron discharge is used to generate the clusters, using standard 2" targets. DC power can be used for sputtering metals, magnetic materials and semiconductors. The magnetron has been designed specifically for high operating pressure of up to mbar in the aggregation region and for a high sputter rate. Water-cooled rare earth magnets are positioned behind the magnetron target. Liquid nitrogen can be used to cool the aggregation tube. This has the effect of stabilizing the deposition and can also reduce the mean cluster size. T-piece is provided for differential pumping. 59 In the nanoparticle-forming chamber, the Fe and Pt atoms were sputtered from the FePt alloy target. Inside a liquid nitrogen cooled aggregation tube, a rare gas, typically argon or helium, cools and sweeps the atoms and clusters from the aggregation region (a) (b) Figure 3.1 (a) Exterior of the assembled Nanocluster Beam Deposition System (b) Components of the nanocluster beam deposition source. towards an aperture. During deposition, the flow rates of Ar and He gases were fixed at 40 and 40 sccm, respectively. The Fe and Pt atoms collided with cold argon (Ar) and/or helium (He) gases flowing through a liquid nitrogen cooled aggregation tube, and condensed into FexPty nanoclusters. The rare gas cools and sweeps the atoms and clusters 60 from the aggregation region towards an aperture. The residence time of the clusters in the aggregation region can be varied by adjusting the separation between the magnetron and the aperture using a linear motion-drive. This helps to control the cluster size. The cluster size can be varied by adjusting several parameters such as the power supplied to the magnetron, the aperture size, the rate of rare gas flow, type of rare gas(es) being used, temperature of the aggregation region and distance between the magnetron and the aperture. 27 After the formation of FePt nanoclusters in the aggregation region, FePt nanoclusters would effuse through an aperture and be deposited on the substrate. The FePt samples were thermally treated in a home-made vacuum furnace at a pressure of ~10-7 Torr at a temperature of 500-700°C for 15 to 30 min, to remove the PS nanospheres by decomposition and to transform the A1 (FCC) to L10 (FCT) phase. 3.2.3 Characterization of magnetic nanostructures by VSM and MFM Hysteresis loops were measured using a Lakeshore Vibrating Sample Magnetometry (VSM). MFM measurements were performed in order to determine the remanent magnetic domain structures of the hemispherical caps. This is an established magnetic imaging technique which gives a measure of the vertical component of the stray field ensuing from the cap. A Digital Instruments (Veeco Metrology Group) Dimension™ 3100 with a Nanoscope IV controller was used in Tapping and Lift mode™, which allows the imaging of topography and magnetic information in a single scan. CoCr-coated silicon tips magnetized along the tip axis (perpendicular to the sample) 61 surface were used. The topographical information was obtained in the tapping mode scan and the phase variation was recorded at a lifted height after each complete topographical line scan. 62 magnetization and thus enables the phenomena of switching to be deconvoluted from the hysteresis measurement, which generally includes a reversible component. There are two principle remanence curves; the isothermal remanence (IRM) and the DC demagnetization curve (DCD). The IRM is measured after the application and removal of a field with the sample initially demagnetized. The DCD is measured from the saturated state by application of increasing demagnetizing fields. Both are illustrated schematically in Figure 4.1. These remanence curves are of importance because they yield the true SFD for the material. The former approach is used to obtain the remanence curve for the AFC CoCrPt nanodots. 4.2.3.2 Determination of SFD Figure 4.2 Determining SFD from hysteresis loop of a magnetic recording medium. The squareness ratio, S* which is related to the slope at Hc, and SFD are of particular importance in characterizing the magnetic properties of magnetic media. S* is related to the slope of the hysteresis loop at Hc, 89 dM dH = Hc Mr H c (1 − S * ) (4.1) This is known as the Williams-Comstock construction. The SFD = ΔH/Hc where ΔH is the full width at half maximum of the differentiated curve dM/dH (as illustrated in Figure 4.2) can be thought of as a distribution function of the number of units reversing at a certain field. For a particulate medium without collective behavior, the SFD has a close relation to particle size distribution because differently sized and shaped particles will reverse at different field strengths. Media with high Hc and small SFD are desirable for high density recording. 90 4.3 Results and Discussions 4.3.1 Schematic of AFC structure Figure 4.3 (a) Schematic illustration of the array of magnetic dots investigated for the effect of demagnetizing interactions through magnetic force microscopy (b) Layer structure of AFC patterned media. Figure 4.3a shows the schematic of the patterned AFC magnetic bits fabricated by e-beam lithography, the detailed description of the AFC layer structure is given in the table in figure 4.3b. If the layer on top and the bottom were antiferromagnetically coupled as schematically depicted in Figure 4.3a, then the Mr of one layer will be partially cancelled by the other. This will lead to a reduced Mr, which is expected to improve stability and reduce SFD. 4.3.2 Hysteresis loops of unpatterned thin film 91 Figure 4.4 shows the hysteresis loops measured by MOKE for the as-deposited AFC film for different top layer thickness. The kink arises from the magnetization reversal of the bottom layer. For unpatterned films, it was observed that the AFC coupling (indicated by the presence of kinks) near remanence state occurs only when the thickness of stabilizing layer is smaller than 5.5 nm. For larger t2, both layers have their magnetizations parallel at remanence state which is not desirable for reducing the interactions. Figure 4.4 Hysteresis loops for unpatterned AFC films with different stabilizing layer thicknesses. 4.3.3 Reduction in SFD with AFC Figure 4.5 shows the MFM images captured in their remanence states after the application of a magnetic field (the magnitude of the applied field is indicated in the images), for different t2. The images on the top shows the point at which the bits start to switch while the images on the bottom shows the point when almost all the bits are switched. If the applied field is high enough such that the magnetization direction of the 92 bottom layer is aligned with the magnetic field when the field is removed, the dot is said to have “switched” and it appears as a bright spot in the MFM image, otherwise it adopts the other antiferromagnetically coupled configuration at remanence and the “unswitched” dot appear as a dark spot. Figure 4.6 shows the remanence curves obtained by plotting the percentage of dots switched at each applied field with 60 nm and 40 nm spacing. The AFC magnetic configurations of the dark and bright spots are shown in the inset of Figure 4.6a. Figure 4.5 μm by μm MFM images of 40 nm spaced magnetic bits for a stabilizing thickness t2 of (a) nm, (b) 3.0 nm, (c) 5.3 nm and (d) 6.7 nm. 93 Figure 4.6 Switching field distribution of 40 and 60 nm spaced bits for different values of stabilizing layer thickness t2 of (a) nm, (b) 3.0 nm, (c) 5.3 nm and (d) 6.7 nm. Inset shows the direction of the dipole moments for the bright and dark spots in the MFM images and the corresponding layer structures. Figure 4.7. Derivative curves of AFC patterned media for different values of stabilizing layer thickness t2 of (a) nm, (b) 3.0 nm, (c) 5.3 nm and (d) 6.7 nm with 40 nm spacing. 94 Figure 4.7 shows the SFD determined by differentiating the remanence curves as a function of t2 and bit spacing which shows a peak at the median switching field HSW. In the ideal case for an isolated AFC entity the bits should align perfectly antiparallel in the remanent state and in applied field until the applied field is large enough to overcome the exchange energy and anisotropy energy for reversal to occur over a sharp field. However, the SFD of a single entity in a patterned medium is affected by a combination of intrinsic magnetic properties, including exchange coupling between the ferromagnetic layers, magnetostatic coupling with its neighbors, 15 , 16 as well as extrinsic properties such as variations in the stabilizing layer thickness, spacing between neighbors. 17 HSW was increased with t2, for both 40 and 60 nm spacing (not shown here). The SFD was found to be narrower and sharper in the case of 60 nm than 40 nm spacing, which is believed to be due to stray field influence from the neighboring bits as the spacing between the bits is reduced to 40 nm. A big difference was not seen in the absolute width of the SFD for different stabilizing layer thicknesses, for the same spacing. Figure 4.8. (a) Stability of the bits, indicated by ΔH/H and (b) coercivity for AFC patterned media as a function of stabilizing layer thicknesses for 40 and 60 nm spaced 95 bits. 4.3.4 Improvement in stability and coercivity with AFC Figure 4.8a summarizes the effect of stabilizing layer thickness t2 on the stability of the patterned media apparent from the expression ΔH/H, which is the full width at half maximum of the switching field distribution divided by the mean switching field. As mentioned earlier, magnetostatic interactions with neighboring islands lead to a widening of the SFD and set a limit on the minimum SFD for the more closely spaced bits (i.e. 40 nm spaced bits). This effect is most evident for the 40 nm spaced single layer structure without AFC, which has the highest ΔH/H of 1.04. Increasing t2 increases the switching field and reduces the demagnetizing effect of neighboring islands due to reduced Mr in AFC structures and probably an increased thermal stability in AFC structures. ΔH/H decreases with t2 and saturates at t2=5.3 nm for the 40 nm bits, whereas ΔH/H does not change much with t2 for the 60 nm spaced bits, implying that the interaction between bits is almost negligible at 60 nm spacing. Thus ΔH/H is higher for the 40 nm than 60 nm spaced bits but this difference in ΔH/H between 40 nm and 60 nm spaced bits decreases with t2, indicating the effectiveness of AFC design at high densities. From the MOKE measurement of the unpatterned films, AFC is clearly induced by the Ru layer for a stabilizing layer thickness below 5.5 nm. However, the reversal of the top layer occurs in the first quadrant at positive field for t2=3.0 nm, and shifts to negative field and occurs in the second quadrant for t2=6.7 nm. The reduction in the Mr for t2=5.3 nm, where the bits are in a stable antiparallel configuration at remanence, probably accounts for the lowest ΔH/H of 0.23. Exceeding a certain optimal stabilizing layer thickness, the anisotropy 96 energy of the thicker stabilizing layer may exceed the interface coupling energy, making it more difficult to switch the magnetization of the top layer. An antiparallel coupling may not happen at zero fields for t2=6.7nm and the reduction of SFD via AFC becomes less efficient. It was also noticed from the MFM and hysteresis loops that Hc in the patterned media is higher than that of the continuous film. While the continuous film remanent coercivity (Hc), defined as the reverse field required to reduce the Mr to zero, is typically kOe from AGM measurements, the 60 nm diameter bits have a measured Hc of 4–8 kOe from MFM measurements. Figure 4.8b shows that Hc increases with t2 and is higher for the 60 nm spaced bits. The introduction of the stabilizing layer has increased the Hc of the films, and the nanodot array is slightly more stable at a spacing of 60 nm than at 40 nm, which is consistent with the trend of ΔH/H in Figure 4.6a. The continuous film reverses by nucleation of a low anisotropy volume followed by rapid domain wall propagation, whereas the small islands reverse by rotation, thus their Hc is determined by the island anisotropy. 18 The increase in Hc of patterned nanodot arrays is a result of a more coherent rotational switching process for the islands. This enhancement of Hc, together with the reduction in Mr by the anti-parallel coupling, allows a lower SFD to be obtained in AFC patterned media. 4.3.5 Effect of stabilizing layer coercivity on AFC Figures 4.9a and 4.9b show AGM hysteresis curves for AFC structures with different deposition pressures of the top magnetic layer. The coercivity of the top layer determined from minor hysteresis loop shown in Figure 4.9, was found to be 76.6 and 97 24.7 Oe at deposition pressures of 0.5 and 10 Pa respectively. This reduction of the coercivity of top layer at high deposition pressure is mainly due to grain segregation leading to almost superparamagnetic state. Hex M Figure 4.9 Hysteresis loop for AFC media with deposition pressure of top layer of (a) 0.5 Pa, and (b) 10 Pa. 4.3.6 AFC for different stabilizing layer coercivities Antiferromagnetic coupling induced by the Ru layer and the reduction in remanent moment was observed in the continuous films at all deposition pressures. This is possible because the coercivity of top layer (stabilizing layer) was smaller than exchange coupling field (Hex) for all samples at different deposition pressures. This helps to keep the magnetization of the stabilizing layer in an anti-parallel state with respect to the bottom recording layer and thereby making the kink lie in first quadrant. 19 98 Figure 4.10 Selected MFM images for the AFC structures at pressure of Pa for top layer at remanent states Figure 4.10 shows MFM images for case of 60 nm diameter dots and with 40 nm spacing at a deposition pressure of Pa measured at remanent states after the application of the indicated magnetic field. It can be seen that the magnetizations of dots with different anisotropy fields switch at different applied fields. a b Figure 4.11 (a) Percentage of switched dots and (b) SFD for dots with 60 nm diameter and 40 nm spacing. The top layer with nm thickness was deposited at the indicated Arpressures. 99 Hc High V Low P Low V High P t (a) High P (b) Low P Hex>Hc Hex[...]... Chem Mater 20 07, 19, 25 66 21 Qiu, J M and Wang, J P Appl Phys Lett 20 06, 88, 1 925 05 22 Ulbrich, T C.; Makarov, D.; Hu, G.; Guhr, I L.; Suess, D.; Schrefl, T and Albrecht, M Phy Rev Lett 20 06, 96, 07 720 2 23 Min J.-H.; Hwang, S.-W.; Lee, G.-R and Moon, S H J Vac Sci Technol., A 20 03, 21 , 120 3 80 24 Kogoma, M.; Kasai, H.; Takahashi, K.; Moriwaki, T and Okazaki, S J Phys D: Appl Phys 1987, 20 , 147 25 Jiang,... Chem Mater 1999, 11, 21 32 26 Choi D G.; Yu H K.; Jang S G and Yang S M., J Am Chem Soc 20 04, 126 , 7019 27 Deng, S.; Fan, H M.; Zhang, X.; Loh, K P.; Cheng, C L.; Sow, C H and Foo, Y L Nanotechnology 20 09, 20 , 175705 28 Zhang, G.; Wang, D Y and Helmuth, M Angew Chem Int Ed 20 05, 44, 7767 29 Seki, T.; Shima, T.; Yakushiji, K.; Takanashi, K.; Li, G Q and Ishio, S IEEE Trans Magn 20 05, 41, 3604 30 Russek,... IEEE Trans Magn 20 00, 36, 29 90 31 Shinjo, T.; Okuno, T.; Hassdorf, R.; Shigeto, K and Ono, T Science 20 00, 28 9, 930 32 Lua, S Y H.; Kushvaha, S S.; Wu, Y H.; Teo, K L and Chong, T C App Phys Lett 20 08, 93, 122 504 33 Wachowiak, A.; Wiebe, J.; Bode, M.; Pietzsch, O.; Morgenstern, M and Wiesendanger, R Science 20 02, 29 8, 577 34 Zeng, H.; Li, J.; Wang, Z L.; Liu, J P and Sun, S Nano Lett 20 04, 4, 187 35... into 10 μm by 10 μm arrays of 60×60 nm square islands with a thickness of approximately 20 nm and spacings of 40 nm and 60 nm The size and spacings were chosen such that the effects of magnetostatic interaction between neighboring bits will be significant enough to influence the SFD 87 4 .2. 3 Characterization of magnetic nanostructures by MOKE and MFM The magnetic properties of the continuous films were... topmost 2 layers (Figure 3.2c (v)) These demonstrate that RIE-assisted colloidal crystal lithography is a versatile technique for creating magnetic nanostructures of different morphologies (a) 1 M/Ms M/Ms 1 0 (b) 0 in out in out -1 -2 -1 -1 0 H (kOe) 1 2 -2 -1 0 1 2 H (kOe) Figure 3.3 Hysteresis loops of (a) colloidal crystal patterned Ni and (b) 30nm thick Ni film on Si, taken in in-plane and out -of- plane... for the investigation of dynamic phenomena of patterned ferromagnetic elements at the nanoscale 78 References 1 Dai, J B.; Tang, J K.; and Hsu, S T J of Nanoscience and Nanotechnology 20 02, 2 281 2 Allwood, D A.; Xiong, G; Faulkner, C C.; Atkinson, D.; Petit, D.; Cowburn, R P Science, 20 05, 309, 1688 3 Hamann, H F.; O'Boyle, M.; Martin, Y C.; Rooks, M and Wickramasinghe, K Nat Mat 20 06, 5 383 4 Berry,... crystals are summarized in Figure 3 .2 Figure 3.2a shows the morphology of a single layer of nanospheres that had undergone CF4/O2 etching and Ni deposition CF4/O2 etching widened the interstices and flattened the nanospheres into a biconvex shape profile The schematic in figure 3.2a (i) shows the process of creation of the biconvex nanostructure featured in figure 3.2a (ii) and (iii) Diffusion during... process can also be applied to pattern a magnetic alloy FePt, similar to patterning Ni In this part of the work, we focused our attention on a monolayer of colloidal crystals The schematic in Figure 3.4 shows the process of creating magnetic hemispherical caps of FePt The monolayer of colloidal crystals array was first subjected to RIE using a mixture of O2/CF4 The size of the nanospheres is reduced, the... Curtis, S G J Phys D 20 03, 36, R198 5 Haynes, C L and Van Duyne, R P J Phys Chem B 20 01, 105, 5599 6 Yi, D K and Kim, D Y Chem Commun 20 03, 8, 9 82 7 Haynes, C L.; McFarland, A D.; Smith, M T.; Hulteen, J C and Van Duyne, R P J Phys Chem B 20 02, 106, 1898 8 Kuo, C W.; Shiu, J Y and Chen, P Chem Mater 20 03, 15 29 17 9 Choi, D G.; Yu, H K.; Jang, S G and Yang, S M J Am Chem Soc 20 04, 126 , 7019 10 Tan, B... K J Phys Chem B 20 04, 108 18575 11 Eagleton, T S and Searson, P C Chem Mater 20 04, 16, 5 027 79 12 Choi, D G.; Kim, S.; Jang, S G.; Yang, S M.; Jeong, J R and Shin, S C Chem Mater 20 04, 16, 420 8 13 Zhong, H.; Guido, T.; Wu, P; Andreas, D.; Wei, D and Yuan, J Nanotechnology 20 08, 19 095703 14 Beleggia, M.; Lau, J W.; Schofield, M A.; Zhu, Y.; Tandon, S and De Graef, M J Magn Magn Mater 20 06, 301, 131 . deposition of a magnetic alloy, FePt, creating a magnetic gradient nanomaterial 22 that consists of periodic FePt hollow hemispherical arrays. The combined effect of composition modulation of the. magnetic nanostructures is demonstrated for Ni and FePt. By controlling the number of layers of colloidal crystals and the RIE conditions, two-dimensional ordered arrays of magnetic micro /nanostructures. layers of colloidal crystals are summarized in Figure 3 .2. Figure 3.2a shows the morphology of a single layer of nanospheres that had undergone CF 4 /O 2 etching and Ni deposition. CF 4 /O 2