Nanometre-indentation in materials with a high hardness over modulus ratio. I

A major problem in depth sensing indentation experiments is the determination of the real area of contact between the indenter and the sample, as it determines the measured hardness and Young's modulus. The current method of linear extrapolation of the first part of the unloading curve for the estimation of a 'pastic' indentation depth and from this the contact area (after Loubet 1984, Doerner and Nix 1986) is not very reliable especially with materials possessing a high ratio of hardness to Young's modulus. These materials exhibit only a very short or vanishing linear part of the unloading curve and are therefor subject to a larger statistic (from curve fitting) and probably also systematic error (fom the flat punch approximation) than for example metals. Also the inevitable deviation of the tip of the pyramid from the ideal triangular shape causes problems and has to be corrected with an experimentally determined function. Indentation experiments were carried out on several hard materials (Al2O3, Si, Si3N4, glass et al.) with an apparatus developed by Wierenga and Franken (1984) and compared to models of the unloading process (flat punch, cone, parabolid of revolution) for which analytical solutions in a closed form exist (Sneddon 1965). In a companion paper (NANOMETER-INDENTATION IN MATERIALS WITH A HIGH HARDNESS OVER MODULUS RATIO. II. Finite Element Simulation) the results of three dimensional computer simulations of the indentation process using the finite element method are described.