Fresenius Z Anal Chem (1989) 333 : 781
Plenary lectures Surface and microrange analysis
Applications of synchrotron radiation in surface analysis B. Lengeler Institut fiir Festk6rperforschung, Kernforschungsanlage Jiilich, D-5170 Jiilich, Federal Republic of Germany
Miiglichkeiten der Oberflfichenanalytik mit Synchrotronstrahlung Synchrotron radiation emitted by relativistic electrons in storage rings has outstanding properties. It has high intensity and high brightness, it is extremely well collimated, it is tunable from a few eV to many ten keV and it is linearly polarized in the plane of the storage ring. These extraordinary properties have greatly influenced the development of surface analytical spectroscopies using electromagnetic radiation in the vacuum ultraviolet and in the X-ray range. Due to their relatively large penetration depths in matter, hard X-rays (above a few keV photon energy) are not inherently surface sensitive but they can be made so in the regime of total external reflection. When hard X-rays fall under grazing incidence (below about 0.5 ~ on a flat surface they are totally reflected and the X-ray field penetrates into the surface only by
High resolution transmission electron microscopy (HRTEM) R. Gruehn Institut fiir Anorganische und Analytische Chemic der Universitgt, Heinrich-Buff-Ring 58, D-6300 GieBen, Federal Republic of Germany
Hoehaufliisende Durchstrahlungs-Elektronenmikroskopie High resolution transmission electron microscopy (HRTEM) can be used with crystalline solids to obtain direct images of small structural groups comprising a few coordination polyhedra with resolution nearly down to atomic scale ("lattice imaging"). More exact knowledge of the conditions required for direct imaging, as well as improvements in the instruments themselves, have now made it possible to examine very small defect regions (microdomains), faults in the stacking sequence of structural groups or atom layers (planar or Wadsley defects), and isolated defects in narrowly delimited areas that may actually be below the dimensions of the unit cell. The structural principle of the very smallest ordered regions can even be determined when X-ray structure analysis proves unable to do this. For objects of this kind transmission electron microscopy can be a powerful tool and structural information will become avail-
a distance between 2 and 7 nm, depending on the density of the reflecting medium. In this regime all X-ray techniques like Xray absorption, fluorescence or diffraction are surface sensitive in the sense that the signal originates in a layer 2 to 7 nm below the surface. When the angle of incidence is gradually increased above the angle of total reflection the probing depth increases gradually and depth profiles can be determined. A number of examples from X-ray absorption spectroscopy, fluorescence analysis and X-ray diffraction will illustrate the possibilities. The advantages and disadvantages of synchrotron radiation compared to X-rays from a tube will be discussed. For soft X-rays (below a few keV) and for electromagnetic radiation in the vacuum ultraviolet the interaction of the radiation with matter is most conveniently detected via the electrons ejected from the surface. Since these electrons have only a small penetration depth (of 1 nm to a few ten nm) all X-ray techniques in this regime are inherently surface sensitive. Since grazing incidence is not needed, the surface analysis can be easily given microprobe character. A number of examples from photoemission, Auger electron spectroscopy and from Surface-EXAFS will be discussed. Fresenius Z Anal Chem (1989) 333:781 9 Springer-Verlag 1989
able if the resolution of the instrument is in the order of atomic distances or at least equal to the size of interesting building units. The resolution of a transmission electron microscope depends on the acceleration voltage and on the spherical aberration constant C~ of the objective lens. Additional instrumental parameters are important, too. Our microscope, a Philips EM 400 HMG, has a maximum acceleration voltage of 120 kV and is ec[uipped with a LaB6 cathode. Thus a resolution of 3 . 0 3.5 A can be obtained. Particularly suitable for investigations with H R T E M are crystal structures whose building elements can be projected parallel to a crystallographic axis onto a plane in such a way that at high magnification and with high resolution, structural details can be observed and the contrast is as high as possible. Samples are prepared for microscopy by producing fragments with thin edges, e.g. by mechanical means (crushing in an agate mortar). The crystal pieces are brought onto a copper grid covered with a holey foil and are then orientated with one crystallographic axis parallel to the electron beam using the diffraction mode of the microscope. Now a pile of parallel lattice planes corresponding to a small piece of a crystal is orientated perpendicular to the electron beam of the microscope. The resulting image may under certain imaging conditions look like a projection of the structure but one should keep in mind that the way of image formation is not the same as in an usual microscope. The objective lens combines the
782 beams of the diffraction pattern in a suitable way forming an interference figure. Mathematically this corresponds to a second Fourier transformation within the description of the imaging process. In this respect the electron microscope is used as an interferometer. Only for optimum conditions one can assume that the details of the resulting images correspond in a simple way to the projected structure. It is indispensable that samples must be thin, if possible 50-100 A thick. The microscope must be defocussed, usually one needs an underfocus near to the so called Scherzer focus. For the interpretation of the images computer simulations play an important role. In many cases the contrast of the images, which depends on the value of defocus and on the thickness of the crystal region, cannot be interpreted directly. Therefore a sufficient number of simulated images computed for different experimental parameters must be available. Niobium oxides and oxide fluorides, and also other particularly stable compounds with suitable structures known from Xray structure analysis, have in the last 10 years been used by physicists as model substances in order to perfect the theory of computer simulations of electron microscopic images. Computer methods are now in use that portray the crystal as being built up of layers (the multislice method), and that make it possible to calculate the intensity relationship in a high-resolution microscope on the basis of the dynamical theory of elec-
tron diffraction (the dynamical n-beam multislice method). The effects of defocusing, spherical and chromatic aberration, the size of the objective aperture, etc. can all be taken into account. It is particularly important that model calculations can even be applied to structural defects, to the occupation of interstitial lattice points, and so on. Only by such model calculations can one interpret complicated contrast relationships in electron micrographs with a sufficient degree of reliability. Such parallel computer investigations are the more necessary, the more accurately are the real structures to be examined. During the investigation of oxides having block structures it was found that the correspondence between the image contrast and the details of structure is nearly one to one. Therefore in some special cases it is possible to interpret images of block structures "directly". Strong limitations come from the high energy of the electron beam. At highest resolution nearly all organic materials and other sensitive substances are damaged. Only compounds having a high thermal stability and especially high melting inorganic materials can be investigated at atomic resolution. In our laboratory mainly transition metal oxides including superconducting materials and a few oxyhalides were investigated. Fresenius Z Anal Chem (1989) 333 : 781 - 782 9 Springer-Verlag 1989
Modern analytical techniques and current applications in microelectronics H. Rehme
Siemens AG, Research Laboratories, Otto-Hahn-Ring 6, D-8000 Mfinchen 83, Federal Republic of Germany Moderne Analysenmethoden und aktuelle Anwendungen fiir die Mikroelektronik
Today, technological progress is determined very significantly by developments in the field of microeleetronics. Hardly any other technology makes such extreme demands on the purity and perfection of the materials and the precision and complexity of the production processes involved. Manufacturing of microelectronics devices requires many individual processing steps, such as cleaning, film deposition, etching and doping. The production of a 4-Mbit memory chip, for instance, requires nearly 500 such single steps. The geometrical dimensions (interconnection widths, layer thicknesses, penetration depths, channel lengths etc.) involved here are in the region of 1 gm down to 10 nm. The high demands made on this technology have resulted in correspondingly high requirements on analytical techniques. Some typical problems of analysis in integrated silicon circuits are indicated in Fig. 1, which shows a schematic cross-section of an integrated circuit. Very similar and equally demanding requirements are imposed by components in optoelectronics and high-frequently electronics, which consist of complex layer structures from compound semiconductors (e. g. gallium arsenide, gallium aluminium arsenide). The very diverse objectives pursued (see A to F in Fig. 1) mean that a wide range of different analytical techniques must be available in an industrial analytical laboratory: highly-sensitive methods of trace analysis (such as ion chromatography, inductively-coupled plasma mass spectrometry, ion-drift spectrometry), microscopic methods with high and very high spatial resolution (scanning and transmission electron microscopy), methods of micron and submicron-analysis (X-ray and Auger electron microanalysis, analytical electron microscopy) as well as techniques for surface- and depth-profiling analysis (Auger
-/2--r0ssivat'o0 ^
.............................
. ~ " ~
I
"
~
~ -,J~,
-.----.
Interconnection ~
Oxide
Fig. 1. Analytical problems arising in the development phase of microelectronics devices, illustrated by a schematic drawing of an integrated circuit. A Monitoring purity and perfection of the substrate crystal; B analysis of surface contaminations after cleaning and etching steps; C characterization of the micromorphology after layer deposition and patterning; D determination of the microstructure in polycrystalline interconnections and electrodes; E characterization of interface effects after deposition and annealing steps; F measurement of the depth distribution of dopants after implantation and diffusion processes electron spectrometry, ion backscattering spectrometry, secondary ion mass spectrometry). The selective use of these methods allows us to provide quick and effective backup during the development and production of mieroelectronics devices. Analysis thus makes a significant contribution to reducing development times as well as guaranteeing the yield, quality and reliability of the products. Fresenius Z Anal Chem (1989) 333:782 9 Springer-Verlag 1989