Dynamic tear tests at the Naval Research Laboratory (NRL), using a series of beam specimens with increasing depth, have shown that the toughness of 3-in.-thick 5083 plate varied from surface to mid-plane. To explore the possible extent of such variation in still thicker plate and provide an economical test for assessing the effect of process variables on toughness, similar specimens were tested under static loading. The data obtained could be interpreted with the aid of the plastic resistance concept used by NRL. We found that they would also yield approximate values of strain energy release rate at crack initiation (GIc), thus, linking the dynamic tear test and ratio analysis diagram directly to elastic-plastic fracture mechanics. The test method is much more economical than full plate thickness tests, and the results are believed to be more indicative of the materials' resistance to fracture than small specimen (invalid) KIc values, which are more indicative of material yield strength and the crack length dictated by specimen size, than they are of toughness. Surface layers of one 5083-0 plate, 7.75 in. thick (195 mm), were found to be 20 to 30 percent tougher than center layers at room temperature. Toughness of the surface layers improved somewhat as the temperature was lowered to -260°F (-162°C). However, the improvement of the center layers was more dramatic. At the lower test temperature toughness was essentially uniform throughout the plate thickness.

Size effects in the KIc test were investigated by testing duplicate compact specimens in ten different combinations of five thicknesses and five widths. Curves of crack resistance as a function of crack extension were calculated in terms of strain energy release rate, G, rather than stress intensity factor, K. It was found that the entire set of data could be represented by a single linear equation of the form, R = Ro + ρΔa. The coefficient, ρ, was found to be an inverse function of specimen thickness. When this family of R-curves is compared with the driving force, G, curves for the various specimen sizes, it can be seen that the ratio of maximum load to the measuring point load (Pmax/PQ) in the record of a plane strain toughness test, depends primarily on the width/thickness ratio of the specimen. This ratio appears to bear no relation to the contribution of crack tip plasticity to the apparent crack extension at the measuring point. Accordingly, inclusion of a maximum value of this ratio in the validity requirements of ASTM Method E 399 for the purpose of restricting such a contribution is questioned.

No abstract