Buckling of bridging fibres in composites

In brittle materials such as concrete and ceramics, fibre reinforcement has been widely accepted as an effective way of improving their strength and toughness. In addition, a notable pseudo strain-hardening phenomenon can contribute to a significantly enhanced ductility of the composite when an adequately designed fibre system is used. This condition was first proposed by Aveston et al. [1], and later extended by Marshall et al. [2] for continuous aligned fibre reinforced brittle matrix composites. More recently, further extensions to randomly oriented discontinuous fibre reinforced composites have been presented [3, 4]. Upon satisfying the conditions described in the above mentioned micromechanical models, the ultimate tensile strains of the composites are usually improved by two orders of magnitude [5-7]. Total fracture energy reaching 35 kJ/m 2 was also reported for a 2% polyethylene fibre reinforced cement paste [8]. This kind of ductile fracture resembles metal instead of brittle materials. The pseudo strain-hardening behaviour of fibre reinforced brittle matrix composites is associated with multiple cracking, and results from adequate stress transfer capability of bridging fibres. Studies are typically conducted under monotonic tensile loading only. In reality, composites are usually subject to cyclic loads. As a preliminary report of ongoing research on the cyclic behaviour of pseudo strain-hardening cementitious composites, we present initial findings on buckling of bridging nylon fibres across fracture planes in a cement composite after complete unloading in tension. An analytic model is also proposed to account for this buckling phenomenon. Type I Portland cement, silica fume and superplasticizer were used to form the cement paste with water/cementitious ratio of 0.27. Discontinuous nylon fibres (Lr = 21 mm, dr = 25/+m, and Ef = 5.2 GPa) were used to reinforce the paste at a volume fraction of 2%. Tensile coupon specimens of size 304.8 × 76.2 x 12.7 mm were prepared and tested under direct tension in a servo hydraulic tester. Detailed mix proportions and testing procedures can be found elsewhere [9]. Tensile stressstrain curves were recorded. An optical microscope with 50 times magnification was used to examine the bridging fibres after the specimen was completely unloaded. The stress-strain curve is shown in Fig. 1, where four peaks are identified, corresponding to four multiple cracks which occurred within a length of