Search

This work deals principally with two important issues and their interrelation: the evolution of damage in coated cemented carbide tools used for cold forming, and the assessment of mechanical behavior of cemented carbides under compressive contact loads. Damage evolution in ironing tools is qualitatively investigated by planar and cross-sectional microscopical observations. Several mechanisms of damage are identified by taking into account the circumferential distribution along the surface of the ironing tools: coating wear and/ fracture, plastic deformation and fracture of the substrate. In the context of damage location in the substrate, the emphasis is put on the study of damage following all processing steps in substrate and coating preparation. This analysis aims at determining whether the substrate is weakened or not by the mechanical and chemical pre-treatment prior to coating. Spherical indentation testing is used as an experimental procedure to reproduce the compression stress state during cold forming. Large indenter tips (300 microns) are employed for testing cemented carbides in an effort to minimize the effects of material structural inhomogeneities. The behavior of cemented carbides under compressive loads is governed by the reversibility of deformation induced by indentation. A unique feature revealed by analysis of experimental load-displacement indentation curves is the positive energy balance, in that a gain of energy is evidenced after a loading-unloading cycle. Two hypotheses are developed for interpreting such a behavior: the strain induced martensitic transformation of the cobalt ductile phase and the thermal residual stresses in the composite. Substrate weakening subsequent to mechanical and chemical pretreatment prior to coating is expected to have a particular role on the indentation behavior of cemented carbides. The impact of substrate damage on the indentation parameters is identified by a parametric study using finite element simulations in which the damage is considered uniform. In reality, the damage distribution over the surface is rather irregular. In this context, a statistical approach is used for assessment of the material indentation parameters. The failure of coating/substrate systems under compressive contact loads is investigated by finite element simulations. The first failure events are identified by considering that system failure occurs either by substrate plastic deformation or coating fracture. The response of coated systems to spherical indentation is synthesized in damage diagrams which can predict whether the coating or the substrate fails first. The possibility for using such diagrams for assessing the evolution of damage in coated cemented carbide tools is highlighted.

The present work deals with the determination of mechanical properties of materials using instrumented indentation techniques. The main advantage of indentation testing is that it is relatively simple to perform and does not need any special and time-consuming sample preparation. Furthermore, the tests can be performed on very small volumes and samples. However, the determination of mechanical properties from simple indentation tests is not as straightforward. This is mainly due to the many nonlinearities encountered (large and non-homogeneous deformations, contact involving friction, plasticity) which prevent a complete analytical treatment. In this work, two inverse methods are presented which permit to determine the stress-strain behaviour of materials from the analysis of the indentation response which could be applied both at the micro- and nanometre range. The first method is based on the use of conical indenters characterized by different included half-angles θ. The inverse algorithm developed in this work permits to attribute a representative strain value er to each indenter geometry. The corresponding representative stress value εr is calculated from the curvature of the loading curve. In this manner, it is possible to construct a stress-strain curve in a point-wise manner. The second method relies on finite element simulations which are controlled by an external optimization program. The materials stress-strain behaviour is assumed to follow a given constitutive law and the parameters of the constitutive law are then adjusted in a manner such as to reconstruct the experimental data as accurately as possible. The two stress-strain methods are validated on a series of bulk Al and Cu based alloys as well as UV LIGA Ni based coatings. For this, the stress-strain curves as well as the relevant tensile parameters (Young's modulus E, yield stress σy, hardening coefficient n) obtained experimentally are compared to the stress-strain curve and material parameters obtained with the two inverse methods. Both methods have permitted to estimate the stress-strain behaviour of bulk metallic materials tested here. The respective advantages and disadvantages of the two methods are put into evidence and possible further extensions and improvements are discussed in the final part of this work.

Die vorliegende Arbeit gründet auf einer neuartigen analytischen Theorie zum HERTZschen Kontakt am geschichtet aufgebauten Halbraum von N. SCHWARZER (1998). Der Kugeleindruckversuch in Schichtsysteme kann damit im linear-elastischen Bereich exakt beschrieben werden, was bereits erfolgreich zur Bestimmung des Elastizitätsmoduls von ultradünnen Schichten ausgenutzt wurde. Mit der Theorie kann aber auch das Spannungsfeld im Grenzfall elastischer Deformation, zu Beginn des Materialversagens analysiert werden. Über die Kenntnis kritischer Spannungswerte ließen sich damit Versagensmechanismen identifizieren. So konnte kürzlich für Schichten im System Si3N4/SiO2 gezeigt werden, dass die VON MISES-Vergleichsspannung im Versagensfall kritische Werte annimmt. Dies ist wiederum typisch für das plastische Versagen, womit erstmals die Fließpannung von Dünnschichten bestimmt wurde. Im Rahmen dieser Arbeit sollte dieses Ergebnis an magnetronsputterten Schichten vom B4C-Target bestätigt werden, wobei die mechanischen Eigenschaften über die Gaszusammensetzung N2/Ar variiert wurden. Die Nanoindentermessungen mit dem UMIS 2000 erfolgten an Einzelschichten auf Silizium und Quarzglas sowie an einem speziellen Dreischichtsystem. Die Analyse zeigt, dass die plastische Deformation nicht in allen Fällen zum Versagen führte und bei geeignetem Schichtaufbau einstellbar ist. Ferner zeigt die rasterkraft- und rasterelektronenmikroskopische Schadensbildanalyse, dass das Versagen unter bestimmten Bedingungen wahrscheinlich durch star cracks verursacht wurde. Der zweite Teil der Arbeit beschäftigt sich mit einer anderen Problematik. Ausgangspunkt dafür bildet eine Weiterentwicklung der theoretischen Methode durch SCHWARZER (2001), mit der die Anwendung der elastischen Kugeleindruckmethode an begrenzten oder real geformten Festkörperoberflächen denkbar ist. Konzeptionell wird dabei die existierende Lösung für den geschichteten Halbraum auf geeignete krummlinige Koordinaten transformiert. Zur Überprüfung des Beschreibungskonzepts wurde die Lösung für das spezielle Kontaktproblem am Viertel- und Achtelraum entwickelt. Dabei war allerdings bereits wieder eine Annahme nötig, deren Überprüfung Gegenstand dieser Arbeit ist. Hierzu wurden zwei unabhängige Indenterexperimente in der Nähe von Kanten vorgenommen, die zunächst widersprüchliche Ergebnisse lieferten. Weiterführende Untersuchungen zeigen aber einerseits, dass die Empfindlichkeit einer der beiden Messmethoden nicht ausreichend ist. Andererseits liefern die Zusatzuntersuchungen Vorschläge zu Verbesserung der weiterentwickelten Theorie.