Scripta METALLURGICA Vol. 21, pp. 1523-1528, 1987 Pergamon Journals, Ltd Printed in the U.S.A. All rights reserved THE DYNAMICS OF DISLOCATION GENERATION AT CRACK TIPS AND THE DUCTILE-BRITTLE TRANSITION P.B. Hirsch, S.O. Roberts, and J. Samuels Department of Metallurgy and Science of Materials, University of Oxford, Parks Road, Oxford 0Xl 3PH, U.K. (Received July 21, 1987) (Revised August 31, 1987) Introduction During the last few years there has been an upsurge in interest in the phenomena which control whether crystalline solids containing cracks deform in a brittle or ductile manner. The current state of the subject is described in the 1986 Viewpoint Set No. i0 in Scripta Metallurgica (i-8), and in Thomson's review article (9). One of the main conclusions of the Viewpoint Set No. 10 (1-8) is that the role of dislocations generated from external sources, as compared with that of dislocations nucleated at and emitted from the crack tip, in controlling T c is not understood. The purpose of the present paper is to report briefly on some recent experiments on Si and theoretical computations which address this and other problems covering the ductile-brittle transition of intrinsically brittle materials. Further details will be published in the Proceedings of the Aussois Conference on Mechanisms and Mechanics of Plasticity (April 1987) (i0), and elsewhere. Results for Si Fracture experiments were carried out using 4-point bend tests on rectangular bars of float zone Si, with their long axis (25mm) parallel to [iii], their shorter axes along [i~0] (imm) and [i12] (3mm), respectively. The specimens contained thumbnail cracks (radii 13um or 37um) introduced by hardness indentation on the (110) surface (a method known to introduce sharp cracks (11)), and a small plastic zone just below the indentation; internal stresses were reduced by annealing at 800'C. The results were as follows:- l) The ductile-brittle transition is very sharp as found previously by St.John (12) and Brede and Haasen (13) for a different crack geometry; an example is shown in Fig.1. The fracture is brittle below Tc; above T c the whole specimen deforms plastically and the curve shown in Fig.l above T c is the yield stress variation with temperature. The transition region extends only over a few degrees; fracture in the transition region is brittle but occurs at stresses higher than below T c . 2) T e is strain-rate dependent, again as found previously (12,13). Fig.2 shows log (K) against To-I, where K is the stress intensity factor, for the results from the present work, compared with those from St.John's experiments (12). Table I lists the values of activation energies from the lines in Fig.2, for intrinsic and n-type material, and specimens with cracks of radius 13um, together with those for dislocation velocity in Si (14). These results confirm and extend those of St.John (12), and show that T c is controlled by the activation energy for dislocation velocity; similar results have been obtained by Brede and Haasen (13). Writing the dislocation velocity in the form v = Armexp[-U/kT] = rmv o (I) the results show that at T c , K is proportional to (vo). (2) 3) There is a substantial shift of T¢ to higher temperatures for the larger crack size (37~m) for the same intrinsic material. 4) The values of T c in the Oxford experiments are ~IO0'C lower than those in St.John's experiments for the same material and comparable strain-rates; in the latter experiments the cracks were introduced in a different way and fracture toughness tests were carried out using a trapezoid type specimen geometry in which K is approximately independent of crack length. 1523 0036-9748/87 $3.00 + .00 Copyright (c) 1987 Pergamon Journals Ltd.