IFF
Scientific Report 2001/2002

The research areas of the institute comprise (1) technologies for the integration of electroceramic materials into microelectronics and nanoelectronics, (2) dielectric and ferroelectric properties of oxide ceramics, and (3) the defect structure in the vicinity of internal and external interfaces in oxides. These areas are complementary to the research areas of the Institute for Electronic Materials 2 (IWE 2) at the Aachen Technical University (RWTH). Project groups often comprise staff members and students from both institutes.

The 32 ^nd IFF-Ferienschule which we organized in 2001 was entitled Advanced Electronic Materials for the Information Technology. This course turned out to be perfectly in time for preparation of the the new research programm Information Technology and Nanoelectronics in the frame-work of the HGF research area Key Technologies. One aspect of the new organization within the HGF will put more emphasis on the interdisciplinary of the research activities. In this respect, the course was strengthening the bridge between different institutes within the IFF and towards the ISG (Institut fiir Schichten and Grenzflachen).

The major project in the area of integration technologies and scaling effects is our HGF project 'Piccolo -Scaling Effects in Integrated Electroceramic Materials' (2000 - 2003). This project is embedded in the former information technology program PGI (Physikalische Grundlagen der Informationstechnologie) of the Research Center Rilich. Beyond the Institut far Elektrokeramische Materialien (1), the (2) Institut fiir Milcrostrukturforschung IMF headed by K. Urban, (3) the Theorie III, headed by H. Miiller-Krumbhaar, and (4) the Ion Technology (IT) group at the ISG headed by S. Mantl are involved. Several national and international universities and research centers participate in "Piccolo", too. The main focus of the proposal "Piccolo" is a fundamental as well as applied research on scaling effects of electroceramic materials. Today, typically polycrystalline films exhibit grain sizes much smaller than the feature sizes of the microelectronic devices. However, along with the sustaining trend towards further miniaturization, the decreasing feature sizes in microelectronic technology will approach the typical crystallite sizes of the perovskite-type oxide structures. Specific scaling effects are anticipated along this route, due to the long-range nature of the ferroelectric interaction of the oxides involved. The project aims at an (1) elucidation of the physical origin of these scaling effects, (2) an exploitation and extension of the limits to which the ferroelectric properties and high permittivities of the oxides involved can be used, and (3) the development of technological design rules for the integration of the perovskite-type oxides on a decreasing scale. The spectrum of designated results of the project comprises (semiquantitative) models for the superparaelectric limit of ferroelectric (FE) oxides, the dead layer at interfaces, the phase stability and segregation processes of perovskite films during annealing, the nucleation and growth of films by MOCVD, recipes for the deposition of single grain capacitors and ultrathin FE films as well as for reactive ion etching and a ferroelectric field-effect transistor (FE-FET) as a demonstrator. Scanning Probe Microscopy (SPM) techniques such as AIM and STM play an essential role in studying interrelationships on a microscopic base. Hence, the project "Piccolo" is an initiative to pursue research on the basic properties of electroceramic materials under scaling and on the relevance of these effects for the integration of perovskite-type oxides into microelectronics.

In cooperation with AIXTRON AG, our multiwafer MOCVD planetary reactor (MOCVD = Metal Organic Chemical Vapor Deposition) has been upgraded by the TriCent injector system which allows, in principle, the atomic layer-by-layer deposition of ultrathin films. This will be needed in the new Medea+ project which aims at the formation of alternative gate-oxide for future sub- 100nm main-stream FETs and is useful as well for our FEFET project.

As a supplement, we use high pressure oxygen sputtering for the deposition of ultrathin PZT film within the Piccolo project. For the patterning of the ceramic films and electrode / ceramic, film stacks, the Reactive Ion Beam Etching (RISE) technique is employed. In contrast to the situation in the standard Si and compound semiconductor technologies, dry etching processes of oxide ceramics have hardly been investigated yet and therefore represent a research area in which basic studies and industrially funded applied research can be linked in a beneficial manner. The integration processes are complemented by metallization methods based on electron beam and sputter techniques. Within this area, our studies aim at a better understanding of the processes and material parameters which govern the adhesion, the mechanical stress, the microstructure, and their effect on the dielectrical and ferroelectric properties.'

The second research area focuses on the dielectric and ferroelectric properties of oxide thin films and - for comparison - bulk ceramics, which are being investigated in Rilich as well as in Aachen. The material systems are based on compositions used for practical devices and model systems, e.g. SrTiO₃, BaTiO₃, SrBi₂Ta₂O₉, Pb(Zr,Ti)O₃. One of the research topics is the microscopic understanding of ferroelectric hysteresis including new approaches for the separation of reversible and irreversible contributions to polarisation based on the analysis of frequency-dependent small and large signals.

The polarisation studies are linked to the aging (imprint) phenomenon, i.e. the polarisation-dependent shift of the hysteresis curve with time and to the ferroelectric fatigue process, i. e. the reduction of the remanent polarisation by cycling. Both aging and fatigue processes play an important role in the operation of the novel non-volatile memories (Ferroelectric Random Access Memories, FeRAM). Impedance spectroscopy in the lower GHz regime is employed to determine the relaxation of the ferroelectric domain wall motion and to separate this contribution from the contribution of the crystal lattice. By varying the microstructure of the ceramics and by comparison between bulk ceramics and thin films, the model of Arlt will be extended with respect to the impact on 2D constraints imposed by mechanical stress due to the presence of substrates. The investigations of the ferroelectric properties are now additionally focussed on the scaling properties and nano-size effects. For dielectric ceramics, impedance spectroscopy is used to elucidate the interrelation of extrinsic losses and lattice defects. This activity includes the development and characterization of new microwave ceramics and is embedded into a cooperation with Norbert Klein's group at the ISG (Institut fiir Schichten and Grenzflachen) within the framework of a BMBF-Leitprojekt. In the case of ferroelectric materials, existing theories are further developed and extended towards a more quantitative description of the dielectric, piezoelectric, and elastic properties.

In addition, numerical finite-element-methods are used to describe the mutually coupled mechanical, thermal, and electrical properties of ceramic components such as multilayer capacitors and actuators.

Our third research area comprises the lattice disorder in the vicinity of internal interfaces (grain boundaries) and external interfaces (surfaces and electrode interfaces) and their impact on electronic and ionic (oxygen ions and protons) charge transport. In the case of acceptor and donor doped titanate ceramics, the studies are focused on the formation of space charge depletion layers at grain boundaries as well as the related potential barriers and the transport of charge carriers along and across the grain boundary barrier. A hot-pressing technology has been developed to decorate the grain boundary area with additional dopants and to study the influence of these artifical grain boundary states. The coupling of the space-charge formation and the defect equilibria at the surface in the case of donor-doped titanates led us to a comprehensive explanation for long pending questions concerning the redox kinetics.

In some material systems, it is necessary to determine the equilibrium constants of the defect reactions and the diffusion constants of the system in order to create the basis for the research on interfaces. In this respect, the comparison of bulk ceramics and thin films of the same composition is of vital interest. In thin film systems, the significant influence of the electrode metals, the unexpectedly high stability under conditions of dc-voltageinduced resistance degradation, as well as the tolerance of the lattice concerning the incorporation of nonstoichiometries represent current research topics.

Rainer Waser