Chapter 1 Introduction
 
 
 

The development of thin film and multilayer structures in nanometric scale has been very important in recent years due to the improvement of different deposition techniques. In principle, any technique in which the material is produced by means of "atom by atom" construction can be used to produce thin film and multilayers. Such processes include: chemical vapor deposition, electrochemical deposition and physical vapor deposition. The physical vapor deposition technologies used in synthesizing layered structures are based on the evaporation or sputtering of a target in a vacuum or in noble gas discharges sustained under low pressures. Evaporation is a process of thermal vaporizing of the target in which the heating is carried out at a low pressure. Sputtering is a process whereby material is ejected from the surface of a target material as a result of momentum and energy transfer in a bombardment of the surface by energetic particles. A breakthrough in sputtering was the invention of the magnetron sputtering source. For conventional diodes, it is very difficult to isolate the deposition surface from plasma generated species other than the depositing species, thus results in ion bombardment mixing of the constituents that can extend several nanometers into the sample. Therefore, it is difficult to synthesize a thin film or multilayer structure with compositionally abrupt interfaces using conventional sputtering sources. Magnetron sources have a magnet structure behind the sputter source plate supporting a fringing field that produces a trap for the secondary electrons produced in the noble gas sputter ionization. These electrons are trapped at the target surface thereby trapping the plasma, and increasing the plasma density at the target surface. It is this trapping mechanism that allows the isolation of the sputtering plasma from the deposition surface, it increases the deposition rates by a factor of ten larger than conventional diodes. These advantages make it possible to make high quality multilayers. Eric E. Fullerton et al have reported achievement of epitaxy in the sputtering of Fe/Cr superlattices onto MgO single-crystalline substrates with a dc magnetron sputtering source.

Nanometric scale structures are studied in numerous fields of science for their optical, magnetic, electrical, or superconducting properties, as well as for a wide range of technological applications .

For example, the permalloy (Ni/Fe) , in an integrated thin film form, is used to make the soft magnet core in the coil of the magnet recording head. The writing of the information in magnetic recording is performed with the help of the recording head's magnetic field. The local magnetization of the media is changed according to the information pattern. The reading process is based on the signal induced by the magnetized bits at the head's coil. For faster memories a high speed recording head is essential. The head should have a high magnetization, low coercivity, high permeability and zero magnetostriction. Permalloy is the traditional soft magnetic material for a recording head because of its unique magnetic properties, although several new material have been proposed lately.

Metallic multilayers have been successfully used as dispersion elements for soft x-ray applications They were used to design normal incidence optics for x-ray microscopes and telescopes with substantially higher resolution than is possible with grazing incidence optics. Multilayers of a sequence of ferromagnetic-nonmagnetic layers are extensively used in neutron scattering technology, as monochromators, polarizers and reflectors. The interest in application of magnetic multilayers increased suddenly in the late 80's when the first reports on the exceptional properties of these artificially structured materials emerged, such as perpendicular anisotropy, giant magnetoresistance, magneto-optical activity . The interest is motivated by the prospective use of magnetic multilayers as magnetic/magnetooptic recording materials and sensors of magnetic fields. The giant magnetoresistance (GMR) effect is a large percentage change of the electrical resistivity in the presence of a magnetic field. It has been observed in some multilayers composed of ferromagnet (Fe, Co, Ni, permalloy) and nonmagnetic metal (Al, Cu, Ag, Au, Cr, Pd, Ta, Pt) layers. As an example, a GMR value of 150% was reported for the [Fe(14A)/Cr(8A)]50 (100) superlattice at 4.2 K. Due to This effect, the magnetic multilayers with GMR become a good candidate for the next generation recording heads. The multilayer head offers a much higher signal-to-noise ratio and better overall performance for high density recording.

A detail structural characterization of thin films and multilayers is essential for further understanding of the physical and chemical processes such as corrosion, porosity, aging, annealing, crystallinity, strain and interdiffusion, as well as some magnetic (antiferromagnetic (AF) coupling, GMR effect) and mechanical (acoustic surface and film excitations) properties, which are characteristic of nanometric scale structures. Some of the pertinent structural questions to be answered are: Is the substrate smooth or rough ? Is the substrate clean or dirty? Is the top layer contaminated, oxidized or in any other way different from embedded layers? What is the composition and structure of the layers? Is the structure amorphous, polycrystalline or single-crystal? Are the interfaces abrupt or graded ? Abrupt interfaces are defined as interfaces where the compositional change from material A to material B occurs over one atomic distance normal to the surface. Distinguishing between rough abrupt interfaces and graded interface (due to interdiffusion or composition mix during the sputtering processes) might be difficult. Does the deposition of material A onto material B or the substrate differ from the deposition of B onto A or the substrate during synthesis? A technique that provides depth composition profile is the most appropriated to characterize thin film and multilayer structures.

Many techniques are used for the analysis of compositional structures of thin films, such as Rutherford backscattering spectroscopy (RBS), Auger electron spectroscopy (AES), x-ray photoelectron spectroscopy (XPS), x-ray fluorescence (XRF), etc.. RBS is a non-destructive and quantitative method for compositional depth analysis, especially suitable for probing heavy impurities in light background materials. However, the characterization of multilayers or alloy thin films composed of elements of similar atomic numbers (e.g. Ni/Fe, Cr/Fe) using RBS will be difficult, if not impossible, due to the overlap of the backscattering energies. Because of the limited escape depth of Auger electrons, AES and XPS must be used in conjunction with sputter-etching processes and therefore are destructive methods for depth analysis. XRF energy spectrum analysis is a widely applied technique for determination of elemental composition. In recent years a grazing incidence geometry for XRF (GIXF) has been used by many groups in compositional profiling of layered structures. Since the penetration depth of X-ray changes from a few nm to a few m m with changes of the incident angle around the critical angle of total reflection, the GIXF scan gives information about the depth concentration profile of the fluorescent probes. However, since the change in penetration depth takes place in a very narrow angular range, in practice the applicability of this effect for depth profiling is limited. If the Bragg peak intensity of the multilayer sample is comparable to that of incident beam, a standing wave condition can be setup. In that case the shape of the GIXF intensity near a Bragg peak is rather sensitive to a change in composition, thereby providing compositional information of the embedded layers.

There are other techniques used in surface analysis that provide information other than the composition profile. Morphology of thin films is studied by transmission electron microscopy (TEM) and scanning electron microscopy (SEM). Surface x-ray diffraction (XRD) and high-energy electron diffraction (HEED) are used to determine orientations in the plane of the film and their azimuthal location. Surface extended x-ray absorption fine-structure (EXAFS) measurements, such as obtained from the energy spectrum of the fluorescence excited by the x-ray of grazing incidence, gives local information about the atomic coordination and inter atomic distances.

Among all these analytical techniques, however, the potential of using x-ray reflectivity, or more specifically, resonant reflectivity, as a non-destructive method for composition depth profiling, has not been explored. It is the object of this thesis to develop this new tool in probing the composition depth profile of nanometric scale structures.

It is well known that the refractive index of a material in the x-ray wavelength region is slightly less than one. For sufficiently low angles of incidence, an x-ray beam can be specularly reflected from the surface and embedded interfaces of a thin film. The reflected beams interfere constructively and destructively, and this interference gives rise to a series of intensity fringes. Many important structural information such as sub-layer thickness, effective electron density and interfacial roughness can be obtained from these intensity curves. The x-ray reflectivity measurement have been used to study the nanometric scale structures of many different materials.

Although Compton demonstrated the phenomena of small-angle x-ray specular reflectivity in 1922, the first serious attempt to use the technique to characterize material surface is Parratt's measurements on a copper surface in 1954. However, his work was seriously limited by both the low brilliance of the x-ray beams that were available at that time as well as by the difficulty in obtaining a sufficiently smooth surface. With the developments of thin film synthesis techniques and using conventional or rotating anode x-ray sources, reflectivity measurements were carried out on a broad range of surfaces. Examples include studies of mercury and liquid-metal surfaces, of both coated and uncoated solid substrates, of surface roughness and oxide layers of sputtered polycrystalline films, and of sputter-deposited spin-valve layered structures. With the fast development of Synchrotron radiation facilities around the world, the x-rays produced by synchrotron radiation are playing a more important role in the studies of material structures. The high intensity and the energy tunability over a broad spectral range of synchrotron radiation greatly enhanced the application of x-ray specular reflectivity as a probe of interface and surface structures. It increases the range of accessible scattering angles and allows the use of a variety of energies. Some recent examples are: studies of silicon coated by organic monolayers (alkylsiloxanes), of Si1-xGex/Si heterostructures, of the surface structure of diblock copolymer films, and of CoSi2 layers in Si produced by ion-beam synthesis.

None of the reflectivity measurements in the above mentioned studies are used in compositional profile determination. Since the reflected x-ray intensity is elastically scattered by the electrons in the sample, the x-ray reflectivity is not directly related to the atomic structures and therefore not an element sensitive technique. It is sensitive only to the electron density and the linear absorption of the materials. The electron density and linear absorption are proportional to the atomic density and the atomic number. For a multi- elemental structure, if the composition profile is given, the electron density and absorption distribution and thereby the reflectivity can be calculated. Unfortunately, the converse of this statement is not necessarily true. The inverse problem cannot usually be uniquely solved owing to the limited range of scattering data and the nonlinear nature of the inverse transformation, and more importantly, the lack of phase information lost in the reflected intensity measurements (This is the same as the "phase problem" in crystallography). Therefore, even though some other approaches have been suggested, a least-squares refinement analysis is almost the only practical way to determine the structural parameters. For conventional single energy reflectivity measurements, the ambiguity in the interpretation of the data curve is twofold. First, a single reflectivity data curve may correspond to different density profiles. Second, for a structure of more than two components, the composition profile can not be determined uniquely from the given density profile ( In principle, the composition profile of a two-component system can be defined if the electron density and the linear absorption are taken as independent parameters in the least squares process. However, sometimes the linear absorption, as a fitting parameter, is not as sensitive as the electron density (e.g. when the x-ray energy is lower than the absorption edges of both elements), therefore the composition profile thereby derived is usually not reliable.). So, generally speaking, the composition profile can not be uniquely determined from a single energy reflectivity measurement.

In this thesis, we will demonstrate that with measurements of x-ray reflectivity at right selected energies ( we call it resonant x-ray reflectivity, RXR), one can determine the composition profile of the nanometric scale structures. This non-destructive technique has a spatial resolution of a few angstroms and is specially useful in structural studies of magnetic materials composed of elements with similar atomic number.

It is well known that the effective electron density r e and the linear absorption m are strong energy dependent variables near the absorption edges of the elements. The dispersion behavior is related to the specific atomic structure. The r e and m contrast between the chemical components in the layered structures varies as the energy of the incident x-ray changes. The reflectivity measured at selected energies ( below and above, close and away from the respective absorption edge of the constitutional elements) then shows different features and gives information about the compositional structure of the samples under study.

It is interesting to compare RXR with the "anomalous scattering" technique used in crystallography. To obtain the phase information from the diffraction pattern of a crystal structure, an isomorphous derivative of the crystal is often introduced. The derivative is prepared by inclusion of an extra atom, preferably a heavy atom such as platinum or mercury, into the native crystal. Comparison of the two diffraction patterns gives the needed phase information. The same information can, however, be obtained by varying the photon energy of the x-ray beam. By tuning the beam across the absorption edge of the heavy atom, an effect similar to making an isomorphous derivatives can be obtained. Above the edge, the atom absorbs and appears not to be in the structure and below it scatters x-rays into the diffraction pattern instead of absorbing and is thus "in" the structure. This technique is called " anomalous scattering". The similarity between RXR and anomalous scattering is obvious. The difference between the two techniques is: instead of scattered by individual atoms, the specular reflectivity intensity is scattered by the discontinuities at interfaces of the layered structures. The optical contrast between the two sides of the interfaces is the main factor affecting the line shape of the reflectivity curves. Tuning the x-ray energy across the absorption edge of a layer will not make that layer disappear but more visible since the contrast in the linear absorption is enhanced. Further more, by tuning the x-ray energy close and away from the edges will also change ( or even inverse) the contrast of the effective electron densities (or the atomic scattering factors) and thus change the reflectivity curves. All these changes are related to the composition profile of the layered structures and consequently, make RXR an elemental sensitive technique.