A critical examination of statistical nanoindentation on model materials and hardened cement pastes based on virtual experiments

• Published in: Cement & concrete composites 2009-11, Vol.31 (10), p.705-714
• Main Authors: Trtik, Pavel, Münch, Beat, Lura, Pietro
• Format: Article
• Language: English
• Subjects: Cements; Close packed lattices; Composite models; Elastic modulus; Elastic properties; Focused ion beam nanotomography; Hexagonal cells; Modulus of elasticity; Nanoindentation; Phases
• ISSN: 0958-9465; 1873-393X
• DOI: 10.1016/j.cemconcomp.2009.07.001

Recent results of statistical nanoindentation testing on hardened cement pastes (HCP) reported in the literature show a multipeak response in the elastic modulus frequency plots. These peaks have been interpreted as indicating the true elastic modulus of two calcium silicate hydrate (C–S–H) phases of different densities. However, the application of statistical indentation for determining material properties of single phases in HCP appears to violate some of the basic principles of the technique. To elucidate this aspect, virtual experiments emulating statistical nanoindentation are performed and presented in this paper. They are based on 3D images of: (i) an idealized two-phase material and (ii) HCP acquired by focussed ion beam nanotomography (FIB-nt); within both 3D images, a marching object is sampled over a large number of positions. Based on the local phase composition within the marching object, a local stiffness is estimated by using simple composite models. Due to the large number of the investigated positions, the elastic modulus can be statistically evaluated in the same manner as in the real statistical indentation experiments. The results presented in this paper indicate that the homogenous C–S–H regions present in HCP are too small to cause independent and separated peaks in the elastic modulus plots. Moreover, the presence of phases other than C–S–H, namely unhydrated cement and calcium hydroxide, may also produce spurious peaks in the frequency plots. In conclusion, the presented results question the notion that the multipeak signature in statistical nanoindentation experiments on HCP can be explained only by the presence of two distinct C–S–H phases.