J. Pokluda et alii, Frattura ed Integrità Strutturale, 34 (2015) 142-149; DOI: 10.3221/IGF-ESIS.34.15 143 K EYWORDS . mode II and mode III cracks, ferritic-pearlitic steel, pearlitic steel, micromechanism, mode I branching I NTRODUCTION bservation of crack paths on a microscopic level enables us to reveal local loading modes at the crack tip under various kinds of remote loading. In the recent works [1 – 3], such an analysis was done for pure mode II and mode III loading which led to a proposal of relationship for intrinsic resistance to crack propagation under pure mode II near- threshold loading and a verification of mode-I-branching criterion. This paper broadens the experimental data by presenting results obtained for steels with two different microstructures, which are compared with other previously tested materials. Identification of the effect of microstructure of multiphase materials contributes to the research of shear-mode crack growth micromechanisms. The influence of microstructure on the shear-mode fatigue crack growth is divided here into two factors: the crystal lattice type and the presence of different phases. The first factor was previously studied on four single-phase metals, namely the ARMCO iron, titanium, nickel and stainless steel [3]. Crystal lattices of the first three materials have sufficient numbers of slip systems that enable the crack propagation described by the emission dislocation model [2]. In this model, the crack growth mechanism is realized as cyclic blunting and re-sharpening of the crack tip, during which (shielding) dislocations are emitted from the crack tip and some of them come back to the tip during the unloading phase. Emission of dislocations from the tip of a mode II loaded crack is easy when conveniently oriented slip systems are available. Results from the first three materials show a good correlation between the availability of slip planes and the measured angles of deflected remote mode II cracks with respect to the original crack plane. In pure ferrite (ARMCO iron) the dense set of slip planes of the bcc lattice enables a selection of the slip systems oriented almost parallel to the applied shear stress. This is probably the reason why the shear cracks in this material propagate almost coplanarly (small deflection angle) and the intrinsic mode II threshold is the lowest one of all tested materials. In the hcp lattice of titanium there are higher angles between the slip planes than in the bcc structure and some of the slip systems are harder to activate which results in a bigger deflection angle and a higher intrinsic mode II threshold. The fcc lattice of nickel has even higher angles between the (111) slip planes, which is in agreement with a big measured mean deflection angle of the remote mode II loaded cracks. The mode II intrinsic threshold was also high here, which can be related to a low Schmid factor in the highly deflected planes. This good agreement between the deflection angles and the values of thresholds (see also Tab. 1) is reflected by a successful use of a simple physical-based relationship for intrinsic mode II threshold, Eq. (1), for single- phase metals. This formula takes into account only the strength of interatomic bonds, the dislocation mobility and the crystallography:  K IIeff,th = Gb 1/2 / n α , (1) where   IIm IIm 1 cos 3cos 1 2 2 n      , G is the shear modulus, b is the magnitude of Burgers vector and α IIm is the mean deflection angle of the crack front from the plane of the maximum shear stress [3]. Let us also mention that the effective mode I threshold can be predicted using the formula  K Ieff,th = 0.75 E b 1/2 . (2) The other studied material was the stainless steel. The stainless steel (X5CrNi18-10) had the fcc lattice but, apart from nickel, it has a low stacking-fault energy. The presence of stacking faults prevents dislocations from cross slip and limits their manoeuvrability significantly. Therefore, not only the deflection angle is bigger than that in nickel but also the fracture surface morphology was completely different. Both mode II and mode III cracks propagated in pure mode I and the fracture surfaces contained no crystallographic facets [1]. Therefore, the effective mode II threshold was controlled by O