During the last several years, interest in the study of nanostructured materials has been increasing at an accelerating rate, stimulated by recent advances in materials synthesis and characterization techniques and the realization that these materials exhibit many unique and interesting physical and chemical properties with a number of potential technological applications. As never before, magnetic materials are the key to the future of the storage industry. The recent development of thin film heads, the demand for higher density of information storage, and the emergence of completely new technologies call for entirely new types of magnetic materials with much higher magnetization and coercivity, Hc.
For the last 12 years, our group has been actively involved in the study of nanostructured magnetic materials, including nanoparticles, thin films/multilayers, and nanocrystalline and nanocomposite magnets.
Magnetic nanoparticles show a variety of unusual magnetic behaviors when compared to the bulk materials, mostly due to surface/interface effects, including symmetry breaking, electronic environment/charge transfer, and magnetic interactions. Furthermore, since nanophase particles can be as much as 50% surface material, they represent surface and interfacial material in bulk quantities, thus the new magnetism that may develop should be of practical value.
In collaboration with Kansas State University, we have been studying the magnetic and structural properties of ultrafine magnetic particles made by several techniques, including vapor deposition, sputtering, and chemical synthesis. We have prepared nearly spherical and magnetically hard Fe, Co, and Fe(Co)-B-based particles in the size of 7-20 nm that possess an effective anisotropy and coercivity one to two orders of magnitude higher than in the bulk materials (Fig. 5.1). We have shown that the highest coercivities are obtained in passivated particles that have a core/shell morphology with a metallic core surrounded by a (Fe2O3/Fe3O4 (or berthollide) oxide shell consisting of small grains. The high coercivity of the ultrafine particles cannot be explained by any of the existing models of magnetization reversal. Most of these models are based on single-phase, uniform ellipsoidal particles without magnetic shells surrounding them and without any particle interactions taken into account. The large coercivities have been attributed to an exchange interaction between the core and shell moments at the interface. This interaction also leads to an anomalous temperature dependence of coercivity and to shifted hysteresis loops in field-cooled samples and is known as unidirectional exchange anisotropy, proposed originally by Meiklejohn and Bean. Our studies have shown clearly that a large core/shell interface interaction must be taken into consideration in order to model and fit the data of temperature dependence of coercivity.
We have ample evidence to believe that the core/shell morphology is also present in the magnetic granules of Fe/SiO2 films. Our studies and those of others showed consistently higher coercivities in particles embedded in an SiO2 rather than a BN matrix (Figs. 5.2 and 5.3). Our latest Mössbauer data in Fe30(SiO2)70 and Fe30(BN)70 granular solids showed that the magnetic granules in the above matrices are not pure Fe particles as was expected, but rather a mixture of crystalline Fe and Fe-Si in SiO2 and amorphous Fe-B in BN (explaining thus the lower coercivities in the latter because of the lower anisotropy of Fe-B). However, the Mössbauer data consistently show the presence of an Fe-Si-O phase in the Fe/SiO2 films with higher Hc.
Fig. 5.1. The temperature dependence of the coercivity in passivated Fe particles.
Fig. 5.2. The dependence of the coercivity on
the metal volume fraction.
The theme that has evolved from our past work on the magnetic hysteresis behavior of small magnetic particles is that their high coercivity is due to both size and surface/interface effects, with the latter being predominant. However, it is rather difficult to separate the two effects in particles prepared by either vapor deposition or sputtering, because it is difficult to prepare pure (single phase) particles with either of these techniques. In the former, the particles must be passivated to be protected from oxidation, while in the latter (as discussed above), there is often significant alloying of Fe with Si and B.
In the last few years we have been developing a technique that would allow us to fabricate pure particles separately but in the same chamber where sputtering takes place. We have now constructed and used successfully three particle guns (one uses a resistive heater, the second uses spark erosion, and the third uses sputtering) to make nanoparticles (Figs. 5.4-5.6). This upgraded capability allows us to make composite films with pure or treated uniform magnetic nanoparticles of any material in any matrix by using simultaneous particle deposition and sputtering with a rotating substrate. We strongly believe that this development gives a significant boost to our research efforts in nanoparticles and will lead us to future discoveries of other unique materials and magnetic phenomena.
Fig. 5.4. Particle gun based on evaporation. Fig. 5.5. Particle gun based on spark erosion.
Fig. 5.6. Particle gun based on high pressure sputtering.
We have already undertaken a comprehensive and systematic study that will allow us to better understand intrinsic size effects, surface/interface effects, and the effects of intra- and inter-particle interactions on the magnetic properties of nanoparticles made by vapor deposition, first making pure particles, which we can later treat to modify their surfaces. We have also started preparing nanoparticles of rare earth metals and intermetallic compounds with high anisotropy, which have not yet been studied and are expected to show interesting and unique properties. This information is expected to be of great importance to magnetic recording media, since the drive for higher density media requires isolated particles with size below 10 nm and coercivity greater than 3 kOe. Transition metals and their alloys have a low anisotropy and become superparamagnetic below this size. We believe that the results of these studies will stimulate more theoretical research in this field that would help us better understand the magnetic properties and especially the hysteresis behavior of nanoparticles.
Nanocomposite magnets consisting of a uniform mixture of exchange coupled magnetically hard and soft phases have been extensively investigated in recent years because of their useful hard magnetic properties. High energy products and relatively high coercivities can be developed in these nanocomposite magnets. Among the advantages of these magnets are the high reduced remanence, mr (= Mr/Ms), and low material cost due to the reduction in the content of the expensive hard magnetic phase. A small grain size (10-20 nm) and a uniform mixture of the two phases is a prerequisite for exchange coupling. This coupling leads to a smooth hysteresis loop in which the individual character of the constituent phases is concealed (Fig. 5.7). The challenge of preparing a suitable microstructure is most conveniently handled through non-equilibrium metallurgical techniques such as melt-spinning, mechanical alloying, and sputter deposition.
Fig. 5.7. Hysteresis loops of Fe-rich Nd-Fe-B magnets.
Over the last several years, we have investigated different systems including Nd2(Fe-Co)14B, Pr2(Fe-Co)14B, and Sm2(Fe-Co)17Cx. Melt-spun ribbons with off-stoichiometric compositions were used to ensure a microstructure consisting of a fine mixture of hard (2:14:1, 2:17:Cx) and soft ((-(Fe-Co)) phases. We have studied the magnetic properties of these systems including their magnetization, coercivity, remanence, and Curie temperature and spin reorientation temperature as a function of the size and amount of the soft phase. The coercivity of the samples was found to decrease with the amount of soft phase, but the reduced remanence increased with values greater than 0.7. In addition, the Curie temperature of the hard phase was found to be affected by the presence of the higher Tc soft phase (Fig. 5.8).
Fig. 5.8. M vs T in Sm2Fe17Ga2Cx magnets consisting of
2:17:Cx + a-Fe phases.
Properties obtained so far in this type of magnet are inferior to those predicted by existing models, mainly because of the larger grain size of both phases and their non-uniform distribution. Additional studies are needed to find ways to correct these problems. Full exploitation of these effects should be observed in multilayers consisting of alternate stacks of soft and hard phases; however, very few such studies have been reported in thin films and multilayers because of the difficulties involved in preparing such structures with texture.
In the past few years, magnetic studies in nanostructured materials have focused on the interaction between the electron charges and magnetic spins. These studies have already led to discoveries of new and unique phenomena that are neither observable in traditional bulk materials nor explainable using classical theories. Examples include giant magnetoresistance (GMR) in multilayers and metallic granular solids, spin valves, spin injection in ferromagnet/insulator/ferromagnet sandwiches, etc. Currently in our lab we have three projects in the area of thin films: (1) Fe-O films, bilayers/multilayers, (2) magnetostrictive materials, and (3) nanocrystalline R-Co-M films for magnetic recording.
The studies on Fe-O films were designed to clarify whether the large increase of the coercivity and its observed anomalous temperature behavior in small passivated Fe (Co) particles are due to the presence of a particular oxide or are generic to all types of oxides. It is impossible to have good control of the stoichiometry and the thickness of the oxide shell by the conventional evaporation method of preparation. Typically, the oxide passivation layer was about 20 Å and was comprised of a mixture of all major oxides. In the last two years we started a comprehensive program on the fabrication and characterization of thin Fe-O films by reactive sputtering. The goal was to learn how to deposit single phase Fe-O films with controlled stoichiometry and thickness. This will enable us to prepare exchange-coupled Fe/Fe-O bilayers, which will be used to study how the properties of Fe films are affected by the presence of an exchange-coupled oxide layer with particular stoichiometry and thickness.
By increasing the amount of O2 during deposition, films with increased extent of oxidation were formed. It was found that the stoichiometry goes through the following sequence: a-Fe, amorphous/nanocrystalline film, a-Fe + FeO, off-stoichiometric FexO, FeO + Fe3O4, single-phase Fe3O4, Fe3O4 + g-Fe2O3, Fe3O4 +a-Fe2O3 and single-phase a-Fe2O3.
The most interesting result was the observation of anomalous ferromagnetic type behavior in off-stoichiometric FexO and single-phase a-Fe2O3 films. These two oxides are well established to be antiferromagnetic in bulk. Extensive magnetic, structural, and Mössbauer studies revealed that the reason for the strong magnetic behavior in FexO films was the presence of clusters of Fe3+ ions tetrahedrally coordinated in an fcc closed packed oxygen matrix. In the case of a-Fe2O3 films, the unusual magnetic properties were found to be a consequence of the presence of a significant amount of uncompensated surface moments.
Studies on the thickness dependence of stoichiometry and magnetic properties of Fe-O films are currently underway. Future studies will include exchange-coupled bilayers/multilayers with a particular type of Fe-oxide.
Magnetostrictive materials have been of great scientific and technological importance for many years. The magnetostrictive phenomenon refers to the deformation of a material in a magnetic field. In 1975, the discovery of new magnetostrictive materials, for example, Terfenol-D (Tb0.3Dy0.7Fe2), with enhanced performance and interesting physics renewed interest in magnetostriction.
In the present high technology era, a great need exists for microsystems capable of performing functions that are not fulfilled by existing integrated electronic circuits. In many cases, magnetostrictive films could allow such functions to be readily obtained. For example, in the first class of applications, they could constitute the driving elements of micro-actuators (micro-robots, -pumps or -motors), which are in high demand by industry. A second class of applications would involve the magnetic control of elastic properties or reciprocally the stress/strain dependence of magnetic permeability, to develop various electronic devices like a resonator with magnetically adjustable frequency and stress-controlled inductance.
In all of these applications, the system efficiency depends critically on the relative change in length (l=Dl/l) as a function of the applied field. Values of Dl/l on the order of 200 x 10-6 with a field smaller than 200 Oe are desired. This is not possible with existing materials, which require saturation fields of tens of kOe. These stringent requirements demand new materials or materials with improved structures. Nanostructured materials are potential candidates that can lead to the desired properties through atomic engineering. Such properties can be obtained in amorphous or nanocrystalline alloys with reduced anisotropy and in multilayers with alternate stacks of high magnetostrictive materials and soft magnetic materials that are exchange coupled. Research on magnetostriction in nanostructured materials is currently being carried out mostly in Europe and Japan. There is very little work done in the United States. If this trend continues, the United States will become totally dependent on foreign sources for this highly valued technology. On the other hand, since present studies on high magnetostrictive materials are at an embryonic stage, we expect our efforts to have an immediate and enormous reward.
With the recent advances in deposition techniques and our current understanding of nanostructured materials, it is quite feasible that giant magnetostriction with small anisotropy (and therefore a small saturation field) can be obtained in nanostructured materials. Our preliminary results on amorphous Tb33Fe67/Fe80B20 multilayers (Fig. 5.9) have already shown very promising results with high transverse magnetostriction of 600 x 10-6 obtained with a small external field around 2 kOe.
This work is supported by NSF-DMR and the U.S. Army Research Office.
Fig. 5.9. Magnetostriction in Tb-Fe/Fe-B multilayers.
1. Magnetic Nanoparticles
- Berkowitz (University of California at San Diego, La Jolla)
- Chien (Johns Hopkins)
- Majetich (Carnegie Mellon University)
- Awschalom (University of California at Santa Barbara)
2. Nanocomposite Magnets
- Magnequench International
- Sellmyer (University of Nebraska)
3. Magnetostrictive Materials
- Walser (University of Texas)
- Wutting (University of Maryland)
- O'Handley (MIT)