Historical notes

Laser is an acronym for Light Amplification by Stimulated Emission of Radiation. The operating principles of the laser were originally elucidated for devices operating at microwave frequencies (masers). The maser was invented by Gordon, Zeiger and Townes (1955), who used an inverted population between the excited vibrational levels of ammonia. Extension of the wavelength range of the maser into the optical regime was proposed by Schawlow and Townes (1958), whose various laser schemes all comprised a gain medium, an excitation source to pump atoms or ions in the gain medium into higher energy levels and a mirror feedback system to enable one or multiple passes of the emitted radiation through the laser medium. The special qualities that distinguish laser light from other optical sources include extreme brightness, monochromaticity, coherence and directionality. Also the laser output is linearly polarized and very frequency stable. After development over forty years modern lasers operate over wavelength ranges from the mid-infrared, through the visible and beyond into the ultraviolet and vacuum ultraviolet ranges.

The first operational laser used synthetic rubies, corundum crystals containing ∼0.1 wt.% of Cr2O3, as the gain medium, pumped with white light from a helical flashlamp to oscillate on the sharp R-line at a wavelength of 693.4 nm [Maiman (1960)]. Soon afterwards the He–Ne laser was operated on the 3s → 2p (632.8 nm), 2s → 2p (1.15 μm) and 3s → 3p (3.39 μm) transitions of atomic Ne.

The successful launch of the alexandrite laser by Allied Chemicals Inc. and the elucidation of essential design parameters [Walling et al. (1979), (1980)] spawned rapid growth of research into Cr3+-based lasers [Caird and Payne (1991)]. Possible alternative gain media to alexandrite included the Cr3+-doped garnets [Struve and Huber (1985)]. However, the development of Cr3+ : colquiriite lasers at Lawrence Livermore National Laboratory [Payne et al. (1988a), (1989a)] deflected attention away from the Cr3+ garnets, these mixed fluoride gain media being as efficient as alexandrite and almost as broadband as Ti-sapphire. Two distractions from the dominance of Cr3+ ion broadband tunable lasers were the inventions of the Ti-sapphire (Ti3+ : Al2O3) [Moulton (1982a,b)] and the Cr4+ : forsterite [Petricevic et al. (1988)] lasers. At present Ti3+ : Al2O3 and Nd3+ : YAG are the market leaders in solid state laser production against which new developments are assessed. Their pre-eminence derives in part from the quality and quantity of laser rods that can be produced at modest cost. Both materials have excellent photothermal and thermomechanical properties, and are robust components under laser operating conditions. However, despite much spectroscopic research Ti-sapphire is still the only usable Ti3+-activated solid state laser. In contrast, operation of several 3d2-ion doped lasers have been reported giving broadband tunability at near-infrared wavelengths (1.0–1.7 μm) which have potential applications in optical communications, medical sciences and on remote sensing LIDAR platforms.

Almost four decades have passed since Maiman invented the ruby (Cr3+:Al2O3) laser. The intervening years have witnessed many major achievements and much innovation, culminating in a plethora of devices that impinge in many walks of life. Fundamental studies and technology have been pursued with vigour, separately and in tandem, with resulting applications in materials and device processing, optical, communication and information sciences, and medical and paramedical sciences. Such applications have the potential to materially transform our everyday living at home, at work and at leisure. Optical devices that were once regarded as revolutionary but which are taken now for granted include the fibre laser, responsible for dramatically improving the quality and cost of intercontinental telephone calls, and the compact disc, now the ubiquitous storage medium for optical information. The scope of optical materials discussed in this monograph is very broad, indicating the potential sweep of exciting and novel developments in the short and long term futures.

Ruby lasers of the kind invented by Maiman, and discussed in this monograph, operate on the Cr3+ R-line emission near 694 nm: they are pumped by the ruby's absorption of some of the visible light from a xenon flashlamp. The He–Ne laser was reported soon thereafter, since when laser action has been observed on many thousands of transitions in the gas phase, including 200 or so on neutral Ne alone.

Ligand-field theory

Limitations of crystal-field theory

The limitations of crystal-field theory became apparent soon after its formulation by Bethe (1929). Van Vleck, who was primarily responsible for its early applications, recognized that the point-ion model on which it is based is quantitatively unreliable, and proposed an alternative formulation based on covalent bonding [Van Vleck (1935), Van Vleck and Sherman (1935)], now called ligand-field theory [Ballhausen (1962)]. Nevertheless, the popularity of crystal-field theory with adjustable parameters remains undiminished, contrary to expectation [Jørgensen (1971)]. By virtue of its elegance and relative conceptual simplicity, it continues to provide a useful framework for summarizing and interpolating empirical spectral information [Morrison (1992), Kaminskii (1996)].

The essential similarity of ligand-field theory and crystal-field theory is attributable to the underlying symmetry of the complex, and its implications for the wave functions and energy levels involved. However, ligand-field theory has the capacity to explain phenomena not contemplated in crystal-field theory, such as the nephelauxetic effect discussed in §4.4.8 and §9.6.3 and transferred hyperfine interactions [Spaeth et al. (1992)]. In addition, models based on covalency provide a deeper understanding of optical properties addressed in preceding chapters, and may ultimately yield quantitatively reliable predictions of crystal-field parameters.

Molecular orbitals

In the molecular-orbital theory of covalency [Hund (1927b), Mulliken (1928)], electrons occupy orbital wave functions which are delocalized over the entire complex consisting, for example, of a transition-metal ion and its immediate ligands. An approximate molecular orbital can be constructed as a linear combination of atomic orbitals (LCAO).

Introduction

Scanning techniques for obtaining topographical information about an object are now widely used in science and technology. Usually, a two-dimensional image is constructed from the signal generated by the scanning process, and this information is restricted to a region close to the sample surface. [5.1]. Here a well focused electron beam is scanned over the surface of the specimen and a response signal such as the emitted secondary electrons or the back-scattered electrons are recorded as a function of the coordinate point (x, y) of the beam focus on the sample surface. We emphasize that this technique essentially yields only information about the composition (atomic, chemical, or metallurgical microstructure) and the geometry of the specimen.

Some time ago, low-temperature scanning electron microscopy (LTSEM) was introduced [5.2] which by now has matured into an important new diagnostic tool. LTSEM extends the temperature of scanning electron microscopy to the regime of liquid helium and liquid nitrogen by providing the necessary sample cooling. However, more importantly, LTSEM yields information on the local electronic function and not just the local structure with high spatial resolution. In this way it has provided important input for the understanding of the physics of superconducting electronic circuits and devices. Of course, LTSEM is equally important for the evaluation and analysis of low Tc and high Tc superconductors. The principle of LTSEM utilizes the electron beam as a local heat source on the one hand, and the sensitive response of superconductors to small temperature changes on the other hand [5.2].

Introduction

The scanning tunneling microscope (STM) is the youngest member of the electron microscopy family, developed only a little over ten years ago. The STM has its own unique list of assets and capabilities to apply to the study of high Tc superconducting materials that distinguishes it from the other family members. The data obtainable by STM can duplicate, surpass, or complement those extracted by the other electron microscopes. The STM has the advantage of having a higher vertical resolution than the scanning electron microscope and can achieve atomic resolution without the extensive and potentially damaging sample preparation techniques required for transmission electron microscopy. A disadvantage is that STM measurements are limited to the near surface region. Its realm is truly the atomic-to-nanometer world of the surface.

In addition to the extremely high vertical resolution (less than a 1 Å) routinely attainable by scanning tunneling microscopy and the often limited sample preparation required as noted above, the STM's additional advantages lie in (1) its sensitivity to both local electronic and structural properties, (2) the variety of measurements possible, (3) the low, generally nondestructive, energy range in which it operates, and (4) its environmental flexibility, i.e. its ability to operate under a wide range of temperatures and atmospheric conditions.

The STM's roots lie in electron vacuum tunneling spectroscopy. In the context of measuring electronic properties, it is more correctly described as a spectrometer, for it is the electronic properties of the surface that are being probed in the STM experiment. The correspondence between the electronic and topographic properties is responsible for the microscope label.

Introduction

The electric transport properties of a high temperature superconductor are largely determined by the absence or presence of high-angle grain boundaries and their arrangement within the material. Ultimately, the grain boundary properties are governed by the grain boundary structure and composition at the atomic level. An important goal of electron microscopy investigations is to establish correlations between the electrical transport behavior and the structure and chemical composition of grain boundaries. In this chapter we examine, via the example of YBa2Cu3O7–x (YBCO) grain boundaries, grain boundary structures in high Tc materials and their influence on grain boundary transport properties. The weak-link behavior of high-angle grain boundaries will be discussed in view of results from structural investigations at different length scales ranging from the macroscopic to the mesoscopic and microscopic down to the atomic scale.

Polycrystals of YBa2Cu3O7–x typically can carry critical currents that are two orders of magnitude lower than the critical current densities in corresponding single crystals. This reflects the average reduction of the critical current due to high-angle grain boundaries. Thus, high-angle grain boundaries present a considerable impediment to high current applications of high Tc materials. Conversely, the weak-link nature of grain boundaries can be of considerable value when applied to the design of microelectronic devices, such as superconducting quantum interference devices (SQUIDS). In fact, one of the first commercial applications of high Tc materials was based on the function of grain boundaries as Josephson junctions [10.1].

Following Chaudhari et al. [10.2], who measured the superconducting properties of individual grain boundaries in thin films, Dimos et al. [10.3, 10.4] demonstrated a strong correlation between the critical current densities across the grain boundaries and the grain boundary misorientations for several grain boundary geometries, which included tilt, twist, and general grain boundaries.

Introduction

It is clear that electron microscopy is not the most favourable technique for structure determination of new (superconducting) phases; X-ray diffraction and particularly neutron diffraction do a far better job in the ab initio structure determination. Electron microscopy and electron diffraction are extremely powerful however to determine the local structure; i.e. to detect deviations from the average structure, as determined by X-rays or neutrons. In this way several new phases have been first identified by electron microscopy; some of them have been later made into bulk superconductors. In other cases the identification of isolated defects in an existing material have inspired chemists to produce new superconducting materials; this was, for example, the case for the occurrence of double HgOδ layers in a one-layer Hg-1223 superconductor.

In the first part of this contribution we will focus on the well known YBa2Cu3O7–δ superconductor; this material allows a large number of substitutions without drastically altering its structural aspects, but with sometimes completely different physical properties. In the second part, we will concentrate on the more recent Hg-based superconductors and illustrate the extreme importance of the different electron microscopy techniques in the development of new superconducting compounds.

Oxygen vacancy order in the CuO plane of YBa2Cu3O7–δ

From a microstructural point of view, YBa2Cu3O7–δ is an interesting compound. It can assume variable oxygen contents (0 ≤ δ ≤ 1), ordered into various ordering schemes as observed abundantly by electron microscopy [7.1–7.16], and more recently by X-ray [7.17–7.19] and neutron diffraction [7.3]. Another important feature of the compound is its susceptibility to elemental substitutions, resulting again in a variety of oxygen ordered phases.

Introduction

This chapter is intended as a convenience to those readers actively engaged in the investigation of high Tc superconductors by transmission electron microscopy (TEM). A future possible application of the newly discovered high Tc superconductors is their use in electronic devices. The electrical properties of a device strongly depend on their microstructure, since grain boundaries in these materials can behave as weak links as reported by Dimos et al. [4.1]. Therefore, TEM is an important tool in the study of the relationship between the microstructure and the electrical properties.

To obtain a TEM sample representative of the as-received sample is not only a technical problem but also a problem of understanding the sample preparation process. Unfortunately, the solution is often strongly dependent on the materials being prepared. For instance, high Tc superconductors easily react with moisture and degrade during sample preparation. Moreover, they easily become amorphous during ion milling. For high Tc thin films, the films are usually softer than the substrate, thus have a much higher ion-milling rate. Taking precautions against these kinds of difficulties makes sample preparation for high Tc superconductors relatively difficult.

There have been a number of review articles on TEM sample preparation techniques [4.2–4.5]. TEM samples of high Tc superconductors are mostly prepared either by crushing, cleaving or ion milling. These methods will be dealt with in Sections 4.2 and 4.3. Minor details frequently determine the success of a technique. To illustrate this, we will describe the normal preparation procedure for both techniques while we will concentrate on some ‘tricks’ to obtain a good TEM sample reliably and fast. The method described here will mainly be focused on cross-section sample preparation.

Introduction

In conventional electron microscopy, specimens are observed using the intensity of an electron beam. However, in electron holography [2.1], the phase as well as the intensity of the electron beam transmitted through a specimen is first recorded on film as an interference pattern, which is called a ‘hologram’. The illumination of a laser beam onto this hologram then produces an optical image of a specimen in three dimensions. To be more exact, the wavefronts of the scattered electron beam are reproduced as wavefronts of a laser beam. Although the optical wavelength is 105 times larger than the electron wavelength, the two wavefronts are otherwise alike.

Once the image is completely transferred from inside the electron microscope onto an optical bench, versatile optical techniques can be used for electron optics. For example, the effect of the spherical aberration in the objective lens of the electron microscope can be compensated for to improve the resolution which was the original objective for which Gabor invented holography [2.1]. This is done by optically adding aberration with an opposite sign. The phase distribution of the electron beam can also be drawn on an electron micrograph by using an optical interferometer in the optical reconstruction stage of electron holography [2.2]. An electron microscope with an electron biprism [2.3] can provide an interferogram, but not a contour map nor a phase-amplified interference micrograph.

These possibilities were opened up by the development of a ‘coherent’ field-emission electron beam [2.4] which is indispensable for forming high-quality electron holograms.

Discovered just over a hundred years ago, the ubiquitous electron now forms the basis for a remarkably large range of characterization tools. Surface roughness and morphology, local atomic and electronic structure, vortex motion and superconducting properties can all be imaged thanks to the electron. Being light in mass, samples withstand appreciable irradiation without destruction. Carrying a charge, electrons can be accelerated to high energies and focussed to form transmission images or fine probes, which enables the interior of bulk samples or thin films to be investigated. Electrons may be scattered elastically to provide images of defects and interfaces at atomic resolution, or inelastically, facilitating spectroscopic studies of electronic structure in the vicinity of individual defects or interfaces. Low energy electrons, guided by a metal probe, form the basis for scanning tunneling microscopy, revealing insights into the atomic and electronic structure of surfaces.

This book presents the entire range of electron-based microscopies as applied to high Tc superconductors, scanning electron microscopy, transmission electron microscopy and scanning tunneling microscopy. Introductory chapters cover the basics of high-resolution transmission electron microscopy and microanalysis by scanning transmission electron microscopy. One chapter deals in detail with the difficult procedures of specimen preparation. Other chapters deal with imaging techniques specific to superconductors, the imaging of vortices by electron holography and the mapping of weak links by low temperature scanning electron microscopy. Several chapters deal with specific applications to subjects such as grain boundaries, thin films and device structures.

Introduction

It is clear from the wealth of transport measurements involving both thin films [11.1–11.3] and bulk materials [11.4] that grain boundaries have a strong effect on the transport properties of high Tc materials. Perhaps the most likely source of this effect is the small superconducting coherence length (5–15 Å), which makes the high Tc superconductors extremely sensitive to defects that distort the perfect crystal structure. Electron microscopy is the only experimental technique capable of analyzing these defects on the scale of the coherence length. In particular, the scanning transmission electron microscope (STEM) allows both Z-contrast imaging and electron energy loss spectroscopy (EELS) to be performed with atomic resolution (∼2.2 Å). By using these techniques it is possible to relate the changes in hole concentration, measured from the energy loss spectrum, with defined atomic locations at grain boundaries observed in the image [11.5,11.6]. Such a combination of techniques therefore provides insight into the structure–property relationships of grain boundaries at the fundamental atomic level. In this chapter, the application of these techniques to investigate the dramatic changes in carrier concentrations associated with grain boundaries in YBa2Cu3O7–δ will be discussed.

Imaging and microanalysis of boundary structures

The specimens used in this study are laser ablated thin films of YBa2Cu3O7–δ (YBCO) [11.7] grown on SrTiO3 bicrystal substrates. This method of boundary preparation has been chosen as it is known to produce clean grain boundaries, i.e. free from impurity segregation or copper enrichment, in which the oxygen stoichiometry can be well controlled through annealing. Such specimens therefore contain the inherent limits to the superconducting properties of YBCO grain boundaries, which can only be degraded further by copper enrichment and oxygen deficiency.

Introduction

Transmission electron energy loss spectroscopy (EELS) consists of measuring energy loss dispersion of inelastically scattered high-energy electrons transmitted through a thin film. The high-energy electrons, which interact with the electrons in the solid, lose a certain amount of energy and transfer momentum to the solid. Because of the energy and momentum conservation rules, energy loss and the corresponding momentum-transfer (q) of the probed electron represents the energy and the momentum of the electronic excitations in solids.

Although some optical techniques, such as soft X-ray absorption and optical reflectance measurements, provide comparative information about solids with higher energy resolution, EELS enjoys several unique advantages over optical spectroscopies. First of all, unlike optical reflectance measurements which are sensitive to the surface condition of the sample, the transmitted EELS represents the bulk properties of the material. Secondly, EELS spectra can be measured with q along specific controllable directions and thus, can be used to study the dispersion of plasmons, excitons, and other excitations [8.1–8.5]. Such experiments offer both dynamics as well as symmetry information about the electronic excitations in solids. In addition, the capability to probe the electronic structure at finite momentum-transfer also allows one to investigate the excited monopole or quadrupole transitions, which cannot be directly observed by conventional optical techniques limited by the dipole selection rule.

Because of the significant energy spread of conventional TEM electron sources (e.g. LaB6, W-hairpin filaments with ΔE ∼ 1–2 eV), EELS measurements to investigate the electronic structure of solids have been generally limited to dedicated electron energy loss spectrometers with energy resolutions ∼0.1 eV.

Introduction

High-resolution transmission electron microscopy (HRTEM) has been widely and effectively used for analyzing crystal structures and lattice imperfections in various kinds of advanced materials on an atomic scale. This is especially the case for high Tc superconductors (HTSCs). The most characteristic feature in crystal structures of HTSCs is that there is a common structural element, a CuO2 plane, in which superconductive carriers (positive holes or electrons) are transported. The remaining part, sandwiching the CuO2 planes, accommodates additional oxygen atoms or lattice defects to provide carriers to the CuO2 planes. This is known as the charge reservoir. The transition temperature between superconductive and non-superconductive states, Tc, strongly depends on the concentration of carriers in CuO2 planes and the number of CuO2 planes. Any charge reservoir is composed of some structural elements, including lattice defects. An aim of HRTEM is to clarify the structure of the charge reservoirs. Additionally, a variety of microstructures strongly affect the critical current density, Jc, since they closely relate to the weak link at boundaries between superconductive grains as well as to the pinning of magnetic fluxoids. The characterization of point defects, dislocations, stacking faults, precipitates, grain boundaries, interfaces and surface structures is another important aim of HRTEM. In this chapter, we describe some fundamental issues in analyzing crystal structures and microstructures in HTSCs by HRTEM.

Theoretical background for HRTEM

HRTEM images closely depend not only on some optical factors in the imaging process by the electron lens, but also on a scattering process of the electrons incident on the crystal specimen [1.1]. This section describes the electron-optical background for HRTEM.