Characterisation of micron sized ferromagnetic structures fabricated by focussed ion beam and electron beam lithography

Traditionally electron bean lithography (EBL) has been used to fabricate micron and sub-micron sized devices, such as Γ and T gates for metal semiconductor devices for study within the semiconductor industry. EBL is also used for the fabrication of ferromagnetic elements for use as components in magnetic random access memory (MRAM) and read/write heads in hard disc drives (HDD). MRAM is being investigated as a direct replacement to standard semiconductor RAM as it has lower power consumption and is a non-volatile memory solution, although the areal density, at present, is not as great. Smaller read/write heads are necessary for HDD as recent advances now allow for perpendicular magnetisation (as opposed to parallel magnetisation) of films and increase the areal density to 100 Gb/inch¬¬2, four times the current value.

In this thesis, the physical and magnetic properties of such micron-sized devices that have been fabricated by focussed ion beam (FIB) lithography for comparison to those fabricated by the EBL method are discussed. In addition to this work, the physical and magnetic properties of micron sized element that have been irradiated using the 30keV gallium ion source are also discussed. Also in this thesis, the results of 10x10μm2 arrays of 50nm thick polycrystalline cobalt elements (270x270 nm2 with a 400 nm period) that are fabricated by EBL to determine if there is any magnetic superdomain structure present are discussed. Bright field imaging in a transmission electron microscope (TEM) is used to investigate the physical structure of the ferromagnets, such as the grain size, element roughness and dimensions. Additional information about the topography is measured by atomic force microscopy (AFM). The magnetic properties, such as the magnitude of the applied field at which irreversible events happen and the domain structure, are investigated by the Fresnel imaging and the differential phase contrast modes of Lorentz microscopy. A programme known as object orientated micromagnetic framework (OOMMF) is used to model the magnetic properties of such structures.

In the first chapter, the properties of the ferromagnetic materials are discussed. The discussion encompasses the general properties of bulk ferromagnets; the main energy contributions of a ferromagnetic system and how these contributions affect the formation of domains in the system; and the main types of domain walls that form in the structures discussed in this thesis. The energy contributions and how they can be used to model the magnetisation within the structures using the OOMMF programme are also discussed. Finally, the effects of the discrete dimensions on the magnetisation of the structures are discussed and a brief review on other work in this field is given.

The discussion in the second chapter focuses on the fabrication of the structures. Initially the two deposition techniques (plasma sputter coating and thermal evaporation) are outlined. Following this is a discussion on how the structures are fabricated by EBL and FIB lithography as well as a brief description of the hardware and software used. For FIB lithography, the discussion also focuses on the optimisation of the milling conditions, including simulations of the effects of the 30 keV Ga+ beam on different structures and the effect of differences in bean spot size. Also discussed in this chapter is the introduction of a buffer layer to protect the substrate from milling.

In the third chapter, the techniques used to characterise the structures are discussed. The TEM is discussed as well as the modes used, such as bright field imaging, Fresnel imaging and differential phase contrast (DPC). Other techniques that could also be used to investigate the magnetic properties of the elements and the array are briefly discussed. In the final section, AFM imaging, encompassing the hardware and theory is discussed.

In the forth chapter the preliminary investigations into the fabrication of micron-sized structures by FIB lithography and OOMMF simulations are discussed. The preliminary investigations into 1x1μm2  trenches irradiate the into continuous, 20nm thick, permalloy thin film using a nominal bean current of 10 pA are discussed. The investigation mainly focuses on the effects of varying the dose of the 30 keV Ga+ used to determine a dose range for the fabrication of permalloy elements. The permalloy is supported on a 50 nm thick SiN. It is found that, by increasing the dose, the milled depth and grain sizes increase as expected. Analysis of the trenches is by AFM and bright field TEM imaging. The effects of different beam currents (nominally 10 and 100 pA) on the trenches are also discussed. It is found that the 100pA beam attracts more contamination to the surface of the permalloy than the 10 pA beam and has a less well defined edge structure. Also discussed in this chapter are OOMMF simulations of the magnetisation within the 1000x200 and 500x500nm2 20nm thick, permalloy elements in an applied field. Several states are predicted for the 1000x200 nm2 element, it is found that the near uniform S-and C-states have the lowest energy and well defined square hysteresis loops when field is applied along the long in-plane axis. The simulation of the 500x500 nm2 elements predicts the four-domain flux structure to be the lowest energy state. Unfortunately, (even for elements with roughened edge structure) the simulation predicts a square hysteresis loop in an applied field that is not observed experimentally.

In the fifth chapter, a comparison of the physical and magnetic structure of the 1000x200 and 500x500 nm2 elements fabricated by EBL and FIB lithography on to 20 nm thick, permalloy (supported on 50nm thick SiN) are discussed. This includes elements that are fabricated on 20 nm permalloy on a 4 nm copper buffer layer by FIB lithography. A comparison of the grain structure (by bright field TEM imaging) reveals an increase within the elements fabricated by FIB lithography (this appears up to the centre of the element for the permalloy-only elements at high dose) compared to unaltered grain sizes in those fabricated by EBL. The edge roughness of elements fabricated by FIB lithography is significantly increased compared to those fabricated by EBL (which is of the order of the grain size). As expected, the depth of the trench surrounding the element (anaylsed by AFM is found to increase with dose. A comparison of the magnetic structure reveals that the magnetisation path followed by the 1000x200 nm2 elements fabricated by FIB lithography (for doses >0.04 nCμm-2) is similar to that followed by the EBL elements although the magnitude of the switching field is reduced. The majority of 500x500nm2 elements fabricated by FIB lithography (again doses for >0.04 nCμm-2) also follow the same path as those fabricated by EBL. In a significant minority, however, the formation of the single vortex structure from a near uniform state proceeds through an intermediate metastable twin vortex state. For the elements fabricated by the >0.04 nCμm-2 dose, the elements appear not to be isolated from the film and are, therefore, still be influenced by the thin film.
In the sixth chapter the physical and magnetic characteristics of 2x2μm2 elements that are irradiated by several patterns are discussed. The elements are composed of a 10 nm Cu capping layer, 20 nm permalloy ferromagnetic layer and a 5nmCu buffer layer on a 50 nm SiN substrate. The increased dimensions are to reduce the contributions of the edge effects on the magnetisation within the bulk of the element. The patterns are irradiated using several doses, from light irradiation to light milling. It is observed that in the case of irradiated areas grain growth is proportional to dose, whereas one pixel wide lines do not exhibit any visible grain growth (for a nominally 10 pA beam current). DPC and Fresnel imaging are used to identify what changes, if any, result in the magnetisation within the elements that are irradiated. The irradiated sections are found to disrupt the formation of the four-domain single vortex structure within the elements in zero field and also to disrupt the magnetisation paths from that of the unirradiated element as a field is applied.

The cobalt arrays that are fabricated by EBL are discussed in the seventh chapter. Bright field TEM imaging is used to accurately measure the dimensions of the elements within the array and the dimensions are found to be close to the nominal 270x270 nm2 expected with one in-plane dimension being marginally larger. Large saturating magnetic fields (~0.7 T) are applied to two orthogonal in-plane axes by several procedures. By analysing the vortex chirality distributions within the remanent states, switching of the vortex chirality distributions with the applied field is found to occur. The switching is similar in nature to that of domains within discrete elements such as those discussed in previous chapters. The distribution of the vortices is found to be anisotropic with respect to the direction of the applied field along the in-plane axes, with the more uniform vortex chirality distributions along the axis with the larger dimensions. Arrays of reduced dimensions are simulated in OOMMF to try to understand the nature of the reversal within the arrays. The perfect array is found to have an even split of vortex chiralities distributions as a field is applied for both orthogonal in-plane axes with two “domains” forming. To stimulate a more real array, the individual elements are displaced randomly by 1 pixel; this changes the vortex chiralities distribution within the array. To gain further insight into the magnetic behaviour of the array, the stray field of several non-flux closure structures (flower C- and S-states) are studied.

In the final chapter, chapter 8, conclusions drawn from the work discussed in this thesis are presented as well as any possible future work.