Aluminum Alloy 7075: The Relation between Microstructure and Toughness

The application of high strength aluminum alloys, particularly the Al-Zn-Mg-Cu-based 7000-series alloys, could be extended if it were not for their relatively low fracture toughness, and considerable research has been conducted on the factors affecting toughness of these alloys.

It is well established that one major variable is strength, since as strength is increased there is an accompanying decrease in toughness. There is also evidence that fracture is sensitive to heat treatment, and particularly that the toughness associated with an underaged structure is superior to that of an overaged structure having the same yield stress.

However, considerable use is being made commercially of over-aged tempers, particularly two-step overaged tempers such as T73; although these were developed primarily to improve stress-corrosion resistance, alloys thus aged have significantly greater toughness than those aged to maximum strength. The published information is hence somewhat confusing and little attempt has been made to relate these effects to microstructural differences.

Where microstructure and fracture modes have been studied, it has been shown that fracture in these alloys may be complex. Fracture by transgranular dimple rupture, by brittle or ductile intergranular failure, and by mixtures of these modes has been observed. Several authors have suggested that transgranular fracture is facilitated by inclusion particles. Precipitation reactions in copper-free Al-Zn-Mg alloys have been extensively studied and several attempts have been made to relate ductility or toughness to the precipitate-free zones that are commonly observed adjacent to grain boundaries. They showed that toughness was strongly dependent on fracture mode; the toughness associated with intergranular fracture was considerably lower than that observed for transgranular fracture.

This paper reports the variation of both toughness, as measured by tear resistance, and fracture path with aging treatment for a commercial 7075 aluminum alloy, with the aim of clarifying and explaining the dependence of toughness on microstructure.

The inverse relation between propagation energy and yield stress is to be expected; as the yield stress is increased, the extent to which the stress concentration ahead of a crack or notch may be relaxed by plastic deformation is decreased. It is, therefore, one of these processes which is affecting the observed toughness.

It has been postulated that the mechanical properties of Al-Zn-Mg based alloys should be sensitive to matrix microstructures. G.P. zones, being coherent, are sheared by dislocations and this process favors the formation of coplanar dislocation arrays and pile-ups at boundaries. Although η’ plates are thought to be partially coherent and hence should permit considerable shearing, the deformation of η’ dispersions has been shown to occur in bands rather than by coplanar flow. Stress concentrations at boundaries should hence be greater for G.P. zone dispersions than for η’ dispersions.

The close similarity between the propagation energies for 120° and 150°C underaged structures indicates that, in the present case, matrix microstructure does not have a major effect on toughness. This is supported by the two-stage η’ structure (4 hr at 120° + 1 hr at 177°C) having similar propagation energy to the G.P. zone structure (24 hr at 120°C). The apparent insensitivity to microstructural variations is most probably due to the dispersoid particles. These, acting as barriers to dislocation motion, break up the coplanar flow and hence greatly decrease the local stress concentrations in the structure. In 7075, deformation, even in the structures aged at 120°C, occurs in bands rather than along single planes.

Similarly there is no first order effect of precipitate-free zone width. The zone at 120°C is much smaller than that at 150°, but no significant effect on toughness is observed. Also the precipitate-free zone width is constant with aging time and cannot be responsible for the underaged/overaged differences.

Fractography indicates that the decrease in toughness at a given strength level due to overaging is associated with a transition from transgranular to intergranular fracture. The transgranular dimples have the same characteristics in all conditions. Their size and distribution fluctuate considerably across the fracture surface as does the spacing of the dispersoid particles in the infrastructure. This together with the shape similarity between the dispersoid particles and the particles observed in the dimples supports the suggestions that transgranular fracture is nucleated by the dispersoid particles, most probably by decohesion of the particle/matrix interface. As further support, it has been shown that removing the chromium from 7075 raises the transgranular propagation energy considerably.

The intergranular fracture surfaces vary considerably with aging treatments. Their dimple size increases with increasing aging time and temperature and the dimple depth appears to increase with increasing temperature. It is suggested that fracture proceeds by the nucleation of voids at the grain boundary MgZn₂-particle/matrix interface and the subsequent coalescence of these voids. Hence the dimple size increases as the MgZn₂ particles coarsen; the dimple depth may be related to the precipitate-free zone width, a wide zone permitting more ductile flow than a narrow zone as suggested by Ryum. The effect on toughness is, however, minor.

The observed variation of propagation energy with heat treatment may be explained if the coarsening of grain boundary precipitation results in a steady decrease in the intergranular fracture stress relative to the transgranular fracture stress. If this relative decrease occurs at a constant yield stress then the fracture stress and fracture mode will vary.

The transgranular stress should be independent of aging time but may decrease slightly if the precipitation of MgZn₂ on the dispersoid interface facilitates fracture. The intergranular fracture stress is presented as a band on graphic, because the fracture stress for a particular boundary will depend on its orientation relative to the applied stress-the band represents the difference between the most favorable orientation for fracture (normal to applied stress) and the least (parallel to applied stress). Varying the yield stress will vary toughness, but, at a given strength, this effect will be the same for underaged and overaged structures.

The only grain boundary microstructural variable which may be correlated with a steady decrease in intergranular fracture stress is the grain boundary precipitate size. Similarly, that the directly quenched specimens have coarser grain boundary precipitates than identically aged specimens which have been quenched to 0°C; the directly quenched specimens have significantly lower toughnesses.

A rationalization of the effect of grain boundary precipitate size may be developed if the particles are considered as cracks. The observation of particle-nucleated dimples is generally held to be due to void initiation ahead of the main crack; if the MgZn₂ particle interfaces fracture ahead of the crack, secondary cracks are produced with the same length as the particles. Crack propagation then proceeds by the ductile fracture of the grain boundary regions between the main crack and the secondary cracks. These regions will be subjected to a stress concentration due to the secondary cracks, in addition to the concentration caused by the main crack.

An array of brittle particles along a grain boundary may be considered as an array of colinear cracks. Paris and Sih have used linear elastic analysis to describe the stress concentration effects of regularly spaced, identical, colinear cracks; they showed that the stress concentration factor of such an array was approximately proportional to the square root of the length of the individual cracks.

While there is no available analysis which describes the effects of a similar array within the plastic zone of a larger crack, it is probable that any additional stress concentration effect due to the array will show a similar dependence upon crack length. From the close analogy between brittle particles and cracks, it is suggested that the coarsening of the grain boundary particles leads to an increase in the stress concentration due to these particles. Hence intergranular void coalescence would occur under an increasing stress concentration as aging proceeds.

Similar considerations do not apply for the transgranular fracture mode as there is no significant coarsening of the dispersoid particles during aging. It is to be expected, therefore, that the effect of increasing the aging time would be to lower the intergranular fracture stress relative to that of the transgranular mode, consistent with the observed transition of fracture mode.

It can be concluded that: