Spatial distribution of defects in ultra fine grained copper prepared by high pressure torsion

Abstract

Bulk materials with ultra fine grain structure can be fabricated by severe plastic deformation. Among variety of techniques based on severe plastic deformation high pressure torsion is the most efficient method for grain refinement down to nano-scale. In torsion deformation the strain distribution across the sample is non-uniform and increases with increasing radial distance from the centre of the sample corresponding to the axis of torsional straining. Due to this reason it is very important to examine homogeneity of ultra fine grained structure of samples prepared by high pressure torsion. In the present work positron annihilation spectroscopy was employed for mapping of spatial distribution of defects in ultra fine grained copper prepared by high pressure torsion. Spatial distribution of defects was examined by means of (i) Doppler broadening using S parameter for mapping of defect density and (ii) positron lifetime spectroscopy. Spatially resolved positron annihilation studies were combined with mapping by microhardness testing. Hardness is sensitive to dislocation density due to work hardening but is practically not affected by vacancies while positron annihilation is sensitive both to dislocations and vacancies. Our investigations revealed that ultra fine grained copper contains dislocations and vacancy clusters created by agglomeration of deformation-induced vacancies. Average size of vacancy clusters increases with increasing radial distance from the centre of the sample due to higher production rate of vacancies resulting in larger clusters. During high pressure torsion deformation microhardness increases firstly at the periphery of the sample due to the highest imposed strain. With increasing number of high pressure torsion revolutions the hardness increases also in the centre and finally becomes practically uniform across the whole sample indicating the homogeneous distribution of dislocations. Doppler broadening mapping revealed a remarkable increase of S parameter at the sample periphery due to larger size of vacancy clusters. The S parameter remains significantly enhanced at the periphery even after 25 revolutions. Hence, contrary to dislocation density spatial distribution of vacancy clusters is far from being uniform even after prolonged high pressure torsion deformation.