Springer Science and Business Media LLC

A series of Al-Cu ingots ranging in volume from approximately 50 to 500 and to 5000 cc has been conventionally prepared in two ways, by normal chill casting or by using a magnetic field to control the convective currents. Comparisons allow us to differentiate among the various mechanisms proposed for the origin of equiaxed grains and for the columnar-to-equiaxed transition in castings: i) In normal casting, with the natural convection associated with standard superheats, the majority of grains is provided by the mechanism suggested by Chalmers (and also known as the Big Bang), or by crystal settling from a free surface as suggested by Southin. ii) If new grains, no matter how they may form, are inhibited from mixing with the rest of the liquid, no significant equiaxed structure appears. iii) Unless some special or preselected conditions prevail, other mechanisms such as those involving constitutional supercooling or dendrite remelting, do not seem operative in the formation of a central equiaxed zone.

The mechanical behavior of polycrystalline nickel specimens that were deformed in tension and cathodically charged with hydrogen simultaneously was investigated with particular emphasis on the fracture of such electrodes. This procedure leads to definite, if, however, weak serrated yielding and also markedly reduces the elongation at fracture compared to polycrystals unexposed to hydrogen. Moreover, in contrast to hydrogenated nickel monocrystals which neck down to give a chisel-edge fracture typical of ductile metals, hydrogenated polycrystal fractures are brittle and intergranular. The embrittlement of nickel by hydrogen is shown by means of Auger electron spectroscopy to be associated with the segregation of hydrogen recombination poisons to the grain boundaries. In essence, it is suggested that the entry of hydrogen into the nickel specimens occurs preferentially in the proximity of grain boundary intersections with the free surface, due to the presence therein of Sb and Sn which act as hydrogen recombination poisons and stimulate the absorption of hydrogen by the metal. The presence of such impurities in the grain boundaries suggests that a pressure mechanism is not involved in the intergranular cracking.

The Co-Pt system is well suited for studies of order-disorder by field-ion microscopy. Because of this, and the fact that detailed information on ordering is available for only one Ll2 alloy (Cu3Au), an extensive X-ray study was made of CoPt3. Long-range order falls to lower values (S ≈ 0.64) in this alloy than in Cu3Au. There is a two-phase field nearT c (685°C) extending for about 20°C. The ordered phase has a smaller lattice parameter than the disordered phase close toT c, but at room temperature the reverse is true. Ordering (when the domain size is large) follows Rothstein’s kinetic theory, with an activation energy of 74.0(9) kcal per g-atom. Domain growth is similar to grain growth, with a time exponent of about 0.45 and an average activation energy of about 62 kcal per g-atom. The antiphase domain structure is isotropic, in contrast to Cu3Au. There is little difference in atomic volume of cobalt and platinum in the ordered state as compared to the pure elements.

Correlation functions of random process theory are used to intercompare available fracture toughness (K Ic) vs loading rate $$(\dot K)$$ data of three different mild steels. The analyses confirm that large fluctuations in fracture strength are dependent in an extremely sensitive yet statistically reproducible way upon $$\dot K$$ , and apparently insensitive to temperature and neutron irradiation. The emergence of a common pattern of rate sensitivityK Ic $$K_{Ic} (\dot K)$$ in these ferritic steels provides a discriminating index of strain rate in the initiation stages of the crack growth process, helpful in aligning fracture models.