THE STRESS intensity factors associated with a central crack, situated along the material interface between two different elastic media and subjected to the action of opening mode loads, have been determined in the present paper. Unlike previous results (refs [l, 8, 91 cited in the paper), Ku was found to be zero and K, was found to be independent of the material constants of the bimaterial plate. The results in this paper are erroneous and mistakes are committed from the lack of understanding of the author on (1) the definition of the stress intensity factors for an interface crack; and (2) the mathematical nature of the mixed boundary problem. These points are illustrated in the following paragraphs.
(1) To begin with, an interface crack possesses a different order of stress singularity from the eommon inverse square root singularity for a crack in a homogeneous, isotropic elastic medium. This singularity has a rapid oscillatory character of the type I-"'~+~&, where r is the distance measured from the crack tip and E is a material constant given in the paper, and was first obtained by Williams (ref. [9] cited in the paper). The stress intensity factors for interface cracks are still defined as the amplitudes of the stress singularity corresponding to various modes of loading conditions. However, because of this singular characteristic, the physical meaning of these stress intensity factors is different from those for single phase materials. Failing to recognize this point, the author proceeded to extract the stress intensity factor from eq. (3.28) in the paper which enforces the inverse square root singularity as if the problem is concerned with a single phase medium. This leads to the physically unrealistic results that the stress intensity factor for a dissimilar material system being identical to that for a homogeneous, isotropic medium.
(2) In determining the relationships between constants by enforcing the boundary conditions, the author erred in failing to distinguish the region of applicability of these conditions. For example, eqs (3.19) and (3.20) are both in error because these relationships are not valid in general as the author specified. Consequently, the solution obtained has no relevance to the problem defined.
In fact, integral transform techniques have been used to solve mixed boundary problems of similar nature treated in this paper. Abramian et al. [l] solved the contact problem of a circular stamp on an elastic half space in the presence of adhesion using the Fourier transform and Hankel integrals. By properly treating the mixed boundary conditions, they have successfully recovered the oscillating singularity at the edge of the circular stamp.
It is unfortunate that errors of such fundamental nature contained in this paper can escape the review process. One can only hope that discussions of the present nature can help to alert the reviewers and thus preventing this from happening in the future to this fine journal.