Introduction

Metallic superlattices and multilayers have been studied for more than 60 years [1]. However, until recently those investigations have had little impact on magnetism research. The breakthrough occurred during the early 1970’s, when the considerable advances in vacuum technologies resulted in major progress being made in modern deposition methods. Using a variety of techniques, it has become possible to manipulate atomic planes, lines of atoms and small clusters to create new man made materials engineered on an atomic scale.
One of these modern deposition methods is Molecular Beam Epitaxy (MBE). Originally, MBE was reserved for semi-conductor production. But in the late 1980’s it was adapted to the preparation of magnetic multilayer films. It has revitalized basic research in magnetism [2, 3]. The MBE technique can be used to grow high purity epitaxial films. It has enabled the production of magnetic films which are artificially structured on the nanometer scale. In particular, it is possible to create layered systems with variable thicknesses made of single crystals or polycrystalline ferromagnetic, anti ferromagnetic or paramagnetic materials. For example, multilayer films can be fabricated so that their properties are dominated by the interfaces contribution and structurally adjusted to meet desired characteristics [4]. The MBE synthesized structures display many novel properties compared to bulk materials. Since the physical dimensions of the layers are comparable to certain characteristic length scales, such as the electron wavelength or the length of exchange interaction, it is possible to engineer devices with specific properties [5]. Moreover, much of modern condensed matter materials physics, basic and applied, is now based on the development of artificial materials in unusual configurations. Two important technological interests are the miniaturization of magnetic read heads and new methods for increasing magnetic storage density. In addition, the excellent crystallographic structures and well defined dimensions imply that it is feasible to build theoretical models whose predictions can be compared to the experimental data with good accuracy.

Another important application is that of increasing the energy product (BH) of permanent magnet systems. Maximizing this parameter requires identifying materials with a high saturation magnetization M_sat and a high coercivity B_C in excess of B_C>p M_sat [6]. Moreover, the desirable qualities of a permanent magnet material are a high remanence M_rem, and nearly linear 2^nd quadrant B(H) characteristics. Fig. 1.1 reviews the development of permanent magnets with increasing (BH)_max over the last century.
Alnico materials, developed in the 1930’s were the first modern permanent compounds offering considerable magnetic hardness over the magnetic steels previously available. The manufacturing process consists in precipitating elongated Fe-Co ferromagnetic particles throughout the matrix of Al-Ni. Their properties rely on the shape anisotropy associated with the particles. They are characterized by a (BH)_max of 1.5-7.5 MGOe, and display excellent corrosion resistance. Due to their high Curie temperature (~850°C), they are still used for certain applications nowadays. Nevertheless, they possess a low coercive force, meaning that they can be easily demagnetized, and must therefore be handled properly. In the mid 1950’s, ferrites or ceramic magnets became commercially available. They are fine particle magnets produced by powder metallurgical methods. They exhibit high coercivities, nearly linear demagnetization curves, and a maximum energy product (BH)_max of 1-3.5 MGOe. The magnetism of ferrites is founded in the high magneto crystalline shape anisotropy of the particles. They constitute the most commercially important permanent magnets because of their low cost, and also because they are particularly well suited for many applications, including electric motors and capacitors [7].


 

Figure 1.1 Progress in permanent magnets, with large maximum energy product (BH)_max, as a function of year.

However, it became apparent in the 1960’s that attempts to further enhance or improve the magnetic properties of ferrites and alnico magnets were exhausted. The search then began for other materials, with high uniaxial magneto-crystalline anisotropy, high coercivity, and high saturation magnetization.
Rare-Earth / Transition-Metal (RE-TM) intermetallic compound magnets represented the most promising candidates, and advances in the development of these magnetic materials over the last 40 years have had a profound and far-reaching impact on magnetic devices. Because RE metals have a low Curie temperature T_C (generally below ambient temperature), they are combined with elements that exhibit both good magnetic properties and high T_C (greater than 500-600 K), such as the transition elements of iron, cobalt or nickel. The RE-TM alloy magnets that have received the most attention include alloys of SmCo₅ and Sm₂Co₁₇. They exhibit a maximum energy product of 20 MGOe and 30 MGOe respectively. However, the cost and availability of the principle constituents in SmCo based materials limit their commercial success. So in the 1980’s a considerable effort was expended to replace scarce Co with abundant Fe in combination with RE metals. This, ultimately, leads to the development of NdFeB based magnets, with (BH)_max~40 MGOe. The Nd₂Fe₁₄B compound delivers the highest performance ever achieved industrially to date, i.e. a maximum energy product of 45 MGOe.

More recently, the properties of layered magnets, at the nano-atomic level, have attracted much attention [6, 8-14]. The underlying reason for this interest stems from the work of Coey and Skomski [9] who have argued, on theoretical grounds, that nano-structured magnets with a giant energy product of 120 MGOe might be feasible, if the exchange spring mechanism in those materials could be suppressed. Exchange springs magnets are based on the interfacial coupling of soft and hard ferromagnetic nano-composites. The hard phase, usually a binary or ternary RE–TM intermetallic compound, provides high magnetic anisotropy and coercive fields. The magnetically soft phase (a TM), is pinned by the hard phase through the strong exchange coupling. The FM coupling between hard/soft phase leads to an enhancement of the magnetization, with the added benefit of reducing the overall RE content. To induce strong exchange coupling, the soft phase must have a high Curie temperature (T_C(Fe) ~ 1000 K). One of the technological advantages of exchange spring magnets is that they contain less RE content than single component hard RE-TM intermetallics, which lowers the cost of materials while improving corrosion resistance.

Commercial interests entail that the research into exchange springs magnets is mainly directed towards nano-dispersed hard and soft magnetic phase structures such as SmCo TM (TM = Co, Fe) [10, 15-19]. The magnets are fabricated by rapid quenching and subsequent annealing or mechanical alloying to form a nano-composite with randomly oriented hard grains [11, 12].
As mentioned earlier, magnetic exchange springs have been mainly investigated in ferromagnetically coupled structures, and studies of anti ferromagnetically coupled multilayer films are rarer. In this thesis, the magnetic properties of epitaxial Laves phase multilayer films DyFe2 YFe2 have been investigated. This essentially 2-D system is ideal for the investigation of the physics of magnetic exchange springs in an anti ferromagnetically coupled layered system. In particular, it avoids the structural complexities of the random two phase 3-D system discussed by Coey [6, 9].

The aim of this work is to propose a solid experimental and theoretical framework for the understanding of how to use the characteristics of the single Laves phase materials in order to synthesize RE-Fe₂-based nano-composite structures with novel magnetic properties.

The magnetization reversal mechanism of the DyFe₂ – YFe₂ multilayer films and the epitaxially strained Tb_(1-x)Dy_xFe₂, is studied. Interest of the Laves phase RE-Fe₂-type material arises from the strong exchange of the transition metal and the large magnetostriction of the RE.
To elucidate the magnetic properties of the epitaxial Laves phase multilayer and alloy films, it is important to understand first the basic of magnetism in RE Laves phase. The origin of their magnetic properties is reviewed in chapter 2.

The growth of the epitaxial single phase RE-Fe₂ films is described in chapter 3. In particular, a brief review of the MBE method is given, together with why well ordered single crystal RE-Fe₂ films can be stabilized on sapphire substrate.

In chapter 4, the epitaxial RE-Fe₂ Laves phase (RE = Dy, Y, Tb) are magnetically characterised. The effect of the strain induced by the substrate on the magnetic properties of the Laves phase films is also reported, since this has an important effect on the magnetic properties of the film [20]. Finally, the crystallographic characteristics of the bulk RE-Fe₂ material are compared to those of the epitaxially grown films.

In chapter 5, the experimental methods used to characterize the magnetic films are detailed. The design and the principle of the Vibrating Sample Magnetometer, the main investigatory technique used in this work for the magnetic measurements, are explained. Also, the preparation of the samples for the measurements is described.

In chapter 6, the magnetic measurements of exchange coupled DyFe₂-YFe₂ multilayer films with thin YFe₂ layer, and of epitaxially strained Tb_(1-x)Dy_xFe₂ alloy films are presented and discussed. The magnetic switching behavior in both the systems is investigated, and it is shown how to engineer their coercivity. In addition, the simplest model of nucleation mechanism in single phase materials by Stoner and Wohlfarth [21] is reviewed, and it is generalized to hetero-phase structures. Theoretical predictions are then compared to the experimental data.

The formation of magnetic exchange springs in the anti-ferromagnetically coupled DyFe₂-YFe₂ multilayer films is detailed in chapter 7. It is shown that the bending field, the onset of the magnetic exchange spring, can be engineered by varying the thickness of the layers. A one-dimensional model is used to describe the magnetic spin configuration in the multilayer films. Finally, it is proved that the magnetic properties of a variety of exchange springs can be expressed in a universal form.

Given that the relative thickness of the magnetically hard and soft layers is readily controlled by the growth process, the impact of micro-structural changes upon the magnetic profile of the exchange springs DyFe₂-YFe₂ films can be investigated. In chapter 8, the effects of the magnetic exchange springs on the irreversible switching field of the superlattices are presented and discussed. It is shown that it is possible to engineer multilayer films which possess negative coercivities. In addition, it is also possible to grow films that are magnetically compensated. Practical applications for such magnetic systems are proposed.

Finally, the findings of this work are summarized.


References
[1] J. W. H. duMond and J. P. Youtz, Phys. Rev. 48: 127 (1935)
[2] M. N. Baibich, et al., Phys. Rev. Lett. 61: 2472-2475 (1998)
[3] J. Nogues and I. K. Schuller, J. Magn. Magn. Mater. 192 (2): 203-232 (1999)
A. E. Berkowitz and K. Takano, J. Magn. Magn. Mater. 200: 552-570 (1999)
[4] B. Dieny et al., Phys. Rev. B 43 (1): 1297-1300 (1991)
[5] M. Hehn, Science 272: 1782 (1996)
T. Aign, Phys. Rev. 81: 5656 (1998)
W. Weber, Phys. Rev. B 54: 4075 (1996)
[6] J. M. D. Coey, Solid State Commun. 102 (2-3): 101-105 (1997)
[7] ASM Handbook, vol. 2, Properties and Selection: Nonferrous Alloys and Special Purpose Materials, ASM International, Materials Park, OH, 1990, pp. 788-792
[8] E. F. Kneller and R. Hawig, IEEE Trans. Magn. 27: 3588-3600 (1991)
[9] R. Skomski and J. M. D. Coey, Phys. Rev. B 48 (21): 15812-15816 (1993)
R. Skomski, J. Appl. Phys. 76 (10): 7059-7064 (1994)
J. M. D. Coey and R. Skomski, Phys. Script. T49A: 315-321 (1993)
[10] T. Schrefl et al., J. Magn. Magn. Mater. 127 (3): L273-L277 (1993)
[11] L. Withanawasam et al., J. Appl. Phys. 75 (10): 6646-6648 (1994)
[12] J. Ding et al., J. Magn. Magn. Mater. 124 (L1) (1993)
[13] I. A. Al Omari and D. J. Sellmyer, Phys. Rev. B 52 (5): 3441-3447, 1995
[14] M. Shindo et al., J. Magn. Magn. Mater. 161 (L1) 1996
[15] R. Coehoorn et al., J. Magn. Magn. Mater. 80 (1): 101-104 (1989)
[16] K. O’Donnell et al., J. Appl. Phys. 76 (10): 7068-7070 (1994)
[17] J. M. Yao et al., J. Appl. Phys. 76: 7071-7073 (1994)
[18] K. Ounadjela and G. Suran, J. Appl. Phys. 63 (8): 3244-3246 (1988)
[19] S. Hirosawa et al., J. Appl. Phys. 73 (10): 6488-6490 (1993)
[20] M. Mougin et al., Phys. Rev. B 62: 9517-9531 (2000)
[21] E. C. Stoner and E. P. Wohlfarth, Phil. Trans. R. Soc. 240: 599-642 (1948)