Materials Science and Engineering A356 (2003) 326
/336

www.elsevier.com/locate/msea

On the relationship between microstructure, strength and toughness

in AA7050 aluminum alloy
D. Dumont a,b, A. Deschamps a,*, Y. Brechet a
a
LTPCM, CNRS UMR 5614, Domaine Universitaire de Grenoble, INPG, BP 75, 38 402 Saint Martin d’Hères Cedex, France
b
Pechiney Centre de Recherches de Voreppe, 725 rue Aristide Bergès, BP 27, 38 341 Voreppe Cedex, France

Received 25 October 2002; received in revised form 19 February 2003

Abstract

The effect of process parameters such as quench rate and precipitation heat treatment on the compromise between the toughness
and the yield strength of AA7050 aluminum alloy (AlZnMgCu) are investigated, as well as the anisotropy of this compromise in the
rolling plane. Fracture toughness is experimentally approached by the Kahn tear test. The microstructure is studied quantitatively in
detail by a combination of scanning electron microscopy, transmission electron microscopy and small-angle X-ray scattering, and
the relative fractions of the various fracture modes as a function of microstructural state are quantitatively determined on scanning
electron microscopy images. Toughness is confirmed to be minimum at peak strength, and lower for an overaged material than for
an underaged material of the same yield strength. A lower quench rate is shown to result in an overall reduction of toughness, and in
a reduced evolution of this toughness during the aging heat treatment. The overall toughness is also lowered when the main crack
propagation direction is parallel to the preferential elongation direction of the coarse constituent particles (rolling direction). The
competition between intergranular and transgranular fracture is explained in terms of the modifications of the work hardening rate,
and of grain boundary precipitation. The evolution of fracture toughness is qualitatively explained in terms of evolution of yield
stress, strain hardening rate, grain boundary precipitation and intragranular quench-induced precipitates.
# 2003 Elsevier Science B.V. All rights reserved.

Keywords: Aluminum alloy; Al
/Zn
/Mg
/Cu; Fracture toughness; Precipitation; Yield strength; Strain hardening; Kahn tear test

1. Introduction strain hardening exponent [15,16]. The evolution of

fracture toughness in complex situations involving
The design of alloy composition and heat treatments multiple fracture modes have been modeled [7,12,17
/
in aluminum alloys of the 7000 series for aerospace 19], and its evolution as a function of aging treatments
applications aims at a good compromise between and quenching rates has been rationalized [8,20]. How-
strength and damage tolerance [1,2]. The requirements ever, very little data is available, which contains
for high yield stress and good fracture toughness are simultaneously a detailed quantitative analysis of the
known to be contradictory [3,4]. The triangular relation- microstructure and the associated mechanical tests in a
ship between microstructure and these two properties wide range of process parameter situations. The aim of
has been the focus of much research effort [5
/8]. A this paper is to contribute to a better understanding of
number of models are now available which describe the the toughness/strength relationship in the AA7050 alloy,
evolution of toughness as a function of microstructural via (i) a quantitative analysis of the relevant micro-
parameters, such as the volume fraction of constituent structural features controlling the yield strength and
particles [9,10], the fraction of precipitates on grain fracture; and (ii) mechanical tests, including the mea-
boundaries [11
/14], the value of yield strength and the surement of strength, strain hardening and notch
resistance, and a quantitative analysis of the fracture
mechanisms involved.
* Corresponding author. Tel.: /33-476-826607; fax: /33-476-
826644. The individual features controlling fracture of pre-
E-mail address: [email protected] (A. Deschamps). cipitation hardening aluminum alloys are well identified:
0921-5093/03/$ - see front matter # 2003 Elsevier Science B.V. All rights reserved.
doi:10.1016/S0921-5093(03)00145-X
 D. Dumont et al. / Materials Science and Engineering A356 (2003) 326
/336 327

/ The coarse constituent particles act as initiation sites structure are the quench rate and the duration of the
for damage [21
/23], and the toughness is experimen- subsequent aging treatment. The microstructural fea-
tally observed to decrease when the volume fraction tures investigated are the constituent particles, the grain
of constituent particles increases [24
/26]. structure, and the state of precipitation. The mechanical

/ The dispersoids used for avoiding recrystallisation properties studied are the tensile behavior, and the
modify the granular structure and the plastic flow fracture toughness, measured by the Kahn tear test.
[15,16,25], as well as they provide sites for ductile The relationships between the microstructural features
cavity growth [9,17,18]. As a consequence, their direct and the fracture behavior are established from fracto-
influence on toughness is unclear, and contradictory graphy observations (allowing to identify the major
statements exist in the literature: they are reported fracture modes and their evolution with varying micro-
either to enhance toughness [13,27,28] or to be structure). The evolution of fracture modes dominance
detrimental [9,21,29,30]. with respect to microstructural features, and their
relation with the value of fracture toughness are then
The fracture path goes either in an intragranular interpreted qualitatively in terms of competition be-
manner or through the soft zone around grain bound- tween energies dissipated inside the grains and in grain
aries [31,32]. This duality of fracture modes implies an boundary fracture.
influence of the plastic behavior (yield stress and work
hardening capacity) of the grains, as well as of the
intrinsic strength of the grain boundaries. These two 2. Materials and heat treatments
aspects of the materials behavior are in turn influenced
by the characteristics of the heat treatment. The alloy studied in this work is AA7050, whose
composition is (in wt.%) Al
/6.33Zn
/2.46Mg
/2.2Cu
/

/ The quenching characteristics influence the propor- 0.11Zr
/0.1Fe
/0.08Si. It was received as 150 mm thick
tion of coarse precipitates able to embrittle the grain plates; all the samples for this investigation were
boundaries, (intergranular precipitation and precipi- sampled at 1/8th of the thickness below the surface, in
tate-free zone (PFZ) close to the grain boundaries) order to ensure a maximum homogeneity of composi-
and to deplete the solid solution (precipitation on tion and grain structure.
dispersoids). The samples, 3 mm thick, were first solutionised for

/ The aging treatment influences the plastic behavior of 80 min, with a temperature ramp from 470 to 483 8C.
the grains. The effects of strain localization asso- They were then quenched with three different proce-
ciated with precipitate shearing is well documented in dures: cold water quench (Fast quench rate F /850 K
the literature [33
/36], and is known to be detrimental s 1), quench in boiling water (Intermediate quench rate
to the fracture toughness and to promote intergra- I /19 K s 1) and quench in boiling water, the sample
nular fracture by stress concentrations at the bound- being held between two 5 mm plates (Slow quench rate
aries ahead of localized shear bands [37,38]. In S /7 K s 1). The quench rates given here are average
addition the intrinsic influence of the yield stress values, in the critical temperature range 450
/250 8C.
and the work hardening behavior on dissipation After quenching, the samples were held for 84 h at
ahead of the crack tip is also to be considered [39,40]. room temperature, and then subjected to a heating ramp

/ In addition, the crystallographic and morphological at 30 K h1 up to 120 8C, left at 120 8C for 6 h,
texture produced by the rolling procedure [32], subjected to a heating ramp at 30 K h1 up to 160 8C,
associated with the presence of the PFZs [41], as and held for various times at 160 8C. This two stage
well as the spatial distribution of constituent particles heat treatment enables a good control of the hardening
[10], give preferential directions for intragranular precipitate distribution, and is coherent with industrial
crack propagation and an intrinsic plastic anisotropy practice. The microstructural and toughness investiga-
to the material, so that the fracture toughness as well tions were carried out notably on three reference states
is expected to be anisotropic within the plate. The of aging: underaged (UA), peakaged (PA) and overaged
influence of the state of recrystallisation is unclear (OA). The UA and OA heat treatments were determined
[42
/44] and will not be considered in the present in order to achieve a constant Vickers microhardness of
study where the granular structure is not influenced 170 HV. The corresponding heat treatment times for the
by the various heat treatments (quenching and different quench rates are indicated in Table 1.
aging).

This evaluation of the microstructural actors of the 3. Experimental methods

problem, as well as the ambiguities existing in the
literature, set up the strategy adopted in the present Tensile tests were performed at a strain rate of 5/
study. The process parameters used to vary the micro- 103 s 1. The strain hardening rate u was calculated by
328 D. Dumont et al. / Materials Science and Engineering A356 (2003) 326
/336

Table 1
Summary of heat treatment times for the three main aging states and
the three different quenching conditions

Quench rate Underaged Peak aged Overaged

Fast 1 h 40 min at 5 h at 60 h at
120 8C 160 8C 160 8C
Intermediate 1 h 40 min at 7 h at 40 h 30 min at
120 8C 160 8C 160 8C
Slow 1 h 30 min at 6 h at 41 h at
120 8C 160 8C 160 8C

numerically differentiating the stress strain curve, and

plotted versus the increase in flow stress (s/s0.2) in a
Kocks
/Mecking plot [45,46]. Notch resistance was Fig. 2. Some selected tests have been carried out both by the Kahn
characterized using the Kahn tear test [47]. This test Tear Test geometry and by CT toughness tests. A good correlation is
consists in pulling at constant speed a notched thin found between KQ2 from the toughness tests and the UIE from the
specimen (3 mm thick), with no prior fatigue crack Kahn tests.
initiation, but a tip radius lower than 0.025 mm. The
notch resistance is then characterized by the unit The microstructure evolution has been investigated
initiation energy (UIE) and the unit propagation energy with a range of complementary experimental techniques.
(UPE). The geometry of the sample and the method of Constituent particles have been studied by SEM, both
analysis are shown in Fig. 1. The Kahn tear test has on polished and fracture surfaces. In-situ deformation of
been extensively used to characterize the notch resis- notched samples in the SEM has been used to char-
acterize the sequence of damage initiation. Transmission
tance of precipitation hardening aluminum alloys
electron microscopy (TEM) has been used to character-
[15,47,48]. Despite the fact that the deformation mode
ize the particles present in the various microstructural
is quite different from the classical CT test (closer to
states, namely hardening precipitates, coarse intragra-
plane stress for KTT as compared to plane strain for
nular particles and intergranular precipitates. These
CT), this test has been proven to be remarkably
different particles were also characterized by small-angle
representative of toughness [49,50]. In the present study, X-ray scattering (SAXS). The relative importance of the
the UIE has been used to investigate systematically the different fracture modes has been quantitatively assessed
effect of process parameters on notch resistance. The by image analysis of SEM micrographs.
proportionality between UIE and the square of KQ has
been checked by selected tests on CT specimens, as
shown in Fig. 2. Two orientations of straining in the
plate plane have been investigated: longitudinal direc- 4. Microstructures
tion (L for tensile tests, L
/T for Kahn tests), and
longitudinal
/transverse direction (LT for tensile tests, In this section we will present the microstructural
T
/L for Kahn tests). features found in the 7050 alloy that are most important
for controlling its yield stress and toughness.

4.1. Grain structure and coarse constituent particles

The grain structure of the 7050 alloy, after solution

treatment and quenching, is mostly fibrous, with a
recrystallised fraction of approximately 25% (see Fig. 3).
Constituent particles are present in the form of
Al7Cu2Fe and Al2CuMg (total area fraction on polished
surface of 0.7%) and Mg2Si (area fraction of 0.2%).
These constituent particles are found mostly in conjunc-
tion with the recrystallised areas, suggesting that recrys-
tallisation has occurred mainly with the particle
stimulated nucleation (PSN) mechanism [15]. These
Fig. 1. Load
/displacement curve for a Kahn Tear Test on the fast particles are weakly aligned in the rolling direction, as
quench, under aged state. The UIE is calculated from the area under compared to the long transverse direction, due the
the curve before maximum load, and the UPE after maximum load. rolling mode used in this case (broadsiding).
 D. Dumont et al. / Materials Science and Engineering A356 (2003) 326
/336 329

Fig. 3. Grain structure of the 7050 material in the (L,LT) plane. The
unrecrystallized regions appear in dark.

4.2. Quench-induced precipitates nucleated on dispersoids

When a slow or intermediate quench is applied to the

material after the solution treatment, coarse precipita-
tion occurs on the Al3Zr dispersoids, see Fig. 4(a). These
precipitates are of the equilibrium h phase and are
distributed in bands, which correspond to the maximum
concentration of Zr in the former dendrite interiors of
the cast structure, see Fig. 4(b). The band-like precipi-
tate distribution corresponding to the former casting
Fig. 4. (a) TEM micrograph showing an h particle nucleated on an
dendrites can be evidenced by optical microscopy after Al3Zr dispersoid during a slow quench. (b) TEM micrograph showing
etching of the h particles, see Fig. 5. a band of h particles nucleated on Al3Zr precipitates during an
The effect of the quench rate on this precipitation intermediate quench
process is of course dramatic. No precipitate can be seen
on dispersoids after the fast quench. After the inter-
mediate quench, isolated bands can be observed. The
average diameter of these quench-induced precipitates
was measured both from TEM and field-emission gun
(FEG) SEM observations to 2009/40 nm, and the
volume fraction was estimated from FEG-SEM obser-
vations to 0.59/0.1%. After a slow quench, the bands
cover a large part of the microstructure. The average
size of the particles is still 2009/40 nm, but their volume
fraction is much higher, of the order of 1.59/0.5%.

4.3. Quench-induced precipitates nucleated on grain

boundaries

The second type of quench induced precipitation

occurs on grain and subgrain boundaries. Again, just
after a fast quench, no such precipitates can be
observed. However, they appear subsequently during
Fig. 5. Optical micrograph showing the distribution of etched h
the aging treatment, as can be seen in Fig. 6, and reach
particles nucleated on Al3Zr particles during a slow quench. The
an average size of approximately 409/5 nm in the over- spatial distribution of these quenched-induced precipitates is governed
aged condition. After a slower quench, precipitates are by the peritectic segregation of Zr during solidification, which is
largely present on the grain boundaries, and their size conserved during rolling.
330 D. Dumont et al. / Materials Science and Engineering A356 (2003) 326
/336

size leading to the required hardness is closer to the size

of the peak hardening condition (namely higher in the
underaged condition and lower in the overaged condi-
tion), due to the higher amount of solute trapped in
quench-induced precipitates, not available for harden-
ing. This is reflected in the volume fraction of pre-
cipitates: it is quite similar for a given state for aging for
the fast and intermediate quench rates, and significantly
lower for the slowest rate.

5. Mechanical properties

A necessary step before understanding the evolution

of notch resistance is to characterize in detail the tensile
properties, namely yield stress and strain hardening. Fig.
7 shows the evolution of yield stress during aging as a
function of the different quench rates. As can be
expected from the loss of solute to quench-induced
Fig. 6. TEM micrograph of intergranular precipitates after a fast precipitates available for fine scale hardening precipita-
quench and in the overaged state.
tion, the level of yield stress is lowered in the case of a
slow quench. This decrease of yield stress is at maximum
vary from 559/5 nm after an intermediate quench to about 10% of the nominal value.
759/5 nm after a slow quench. These precipitates do not The strain hardening behavior plays also a very
evolve significantly during the aging treatment. important role for the fracture properties, and is well
known to evolve with the state of aging. Fig. 8 shows the
4.4. Hardening precipitates Kocks
/Mecking plots, representing the strain harden-
ing rate u /ds/do versus the incremental flow stress s/
The hardening precipitates are formed during the s0.2, for the different states of aging in the cases of the
aging treatment. This aging treatment occurs in three fast and slow quench. The evolution of the strain
steps: natural aging, where Guinier
/Preston (GP) zones hardening rate is classically described as follows: the
of diameter less than 1.5 nm develop, aging at 120 8C, underaged material, containing shearable GP zones
where h? precipitates nucleate on the dissolving GP along with a still high solute supersaturation, shows a
zones, and aging at 160 8C, where h? precipitates obtain high strain hardening rate, comparable to that of the
their optimal size (diameter of approximately 6 nm) and material in full solid solution. This high work hardening
are then gradually replaced by the equilibrium h rate is possibly related to dynamic precipitation during
precipitates. This precipitation occurs homogeneously the tensile test [51]. The lowest strain hardening rate is
in the grain and subgrain interiors. It was characterized experienced in the peak aging condition, due to the
quantitatively for the main states of aging and the combination of a low solute content and the presence of
different quench rates by SAXS. The results are
summarized in Table 2.
It appears that the precipitate size leading to a given
hardness level (170 HV for the underaged and overaged
material) is not very sensitive to the quench rate, except
for the slowest quench rate. In this case the precipitate

Table 2
Main characteristics of the hardening precipitates (average radius and
volume fraction) as a function of quench rate and aging state

Quench rate Underaged Peak aged Overaged

R (Å) fv (%) R (Å) fv (%) R (Å) fv (%)

Fast 12.8 1.81 29 4.32 67 5.26

Intermediate 12.4 1.81 35.4 4.00 70.6 4.87
Slow 14.5 3.03 32.9 4.09 54.1 4.21 Fig. 7. Evolution of the yield stress with aging time, as a function of
the quench rate.
 D. Dumont et al. / Materials Science and Engineering A356 (2003) 326
/336 331

Fig. 8. Strain hardening rate vs. incremental stress showing the Fig. 9. Evolution of the relationship between UIE and yield stress as a
evolution of the strain hardening capability as a function of aging function of quench rate. Loading direction is L
/T.
state and quench rate (F, fast quench; S, slow quench). The quench
rate appears to be of second order as compared to the aging state.
resistance. This effect is to be compared to the 10%
difference in yield stress: the latter appears to be much
shearable precipitates. In the overaged condition, the less quench sensitive than the notch resistance. Addi-
strain hardening rate increases again due to the storage tionally, lowering the quench rate also minimizes the
of geometrically necessary dislocations on non-shear- evolution of notch resistance during the aging treatment.
able precipitates, but this high initial strain hardening Finally, the anisotropy of the notch resistance is
rate is not sustained for very long because of the high shown in Fig. 10. The testing direction has little
rate of dynamic recovery due to the low solute content. influence on the yield stress (i.e. the yield stress
Compared to the strong influence on the yield anisotropy is very small), but changes the overall level
strength, the influence of quench rate on strain hard- of notch resistance, which is much lower when pulling in
ening rate appears to be of secondary importance. Only the LT direction (i.e. the crack propagating along the
in the underaged condition is the strain hardening rate rolling direction). It can be noticed as well that the notch
somewhat lowered in the case of a slow quench, resistance is much less dependent upon the state of aging
probably due to the lower solute content resulting in the latter case.
from the presence of the coarse quench induced In order to interpret these clear qualitative features,
precipitates. In the following we will consider the strain we need to investigate, via analysis of fracture surfaces,
hardening properties to be unchanged by changes in the occurrence of the different fracture modes.
quench rate.
A convenient way to present the evolution of the
compromise yield strength/toughness with the different 6. Elementary fracture mechanisms
aging and quench parameters is to draw graphs UIE
versus s0.2, in which each point corresponds to a given The fracture surfaces of the Kahn tear test specimens
heat treatment. show four basic fracture mechanisms, which are dis-
Fig. 9 shows the evolution of this compromise during
aging for the three quench rates. We can first concen-
trate on the behavior after a fast quench. In this classical
situation, two main features are observed:

i) The minimum notch resistance is associated with

maximum yield strength;
ii) For a given yield strength, the overaged material
shows a lower notch resistance as compared to the
underaged material.

Changing the quench rate has a dramatic effect on the

value of the notch resistance. Between an ideal quench
and the 7 K s 1 quench, the UIE is reduced by up to a
factor of 4 (which corresponds to a factor of 2 in
fracture toughness). The main effect of reducing the Fig. 10. Evolution of the relationship between UIE and yield stress as
quench efficiency is thus an overall decrease of the notch a function of the loading direction, after a fast quench.
332 D. Dumont et al. / Materials Science and Engineering A356 (2003) 326
/336

tributed in various proportions depending on the constituent particles being very brittle, their fracture or
thermomechanical condition of the alloy: decohesion can occur in the very first stages of plastic
flow, providing a damage initiation. In order to better
i) Fracture or decohesion of the coarse constituent understand the relationship between this particle frac-
particles (essentially Al7Cu2Fe and Mg2Si), see Fig. ture and the macroscopic crack propagation, in-situ
11(a); straining in the SEM was performed on the cold
ii) Ductile transgranular fracture, characterized by
quenched, underaged material (showing maximum
dimples, the center of which contain constituent
notch resistance). The outcome is that most particles
particles, see Fig. 11(b);
in the section in front of the notch are damaged in the
iii) Ductile transgranular shear fracture, showing large
first stages of plasticity, very long before the macro-
planar areas crossed by intense slip lines, see Fig.
11(c); scopic fracture of the material occurs. These observa-
iv) Ductile intergranular (or inter-subgranular) frac- tions suggest that the influence of the constituent
ture, characterized by relatively smooth surfaces, particles on the toughness can be treated independently
revealing the initial grain structure, see Fig. 11(d). from the other metallurgical effects such as plastic
properties of the matrix or weakening of grain bound-
The first fracture event (fracture of constituent aries by intergranular precipitation.
particles) is known to contribute significantly to the This hypothesis is further supported by the measure-
overall toughness level of the material: the coarse ments of the surface area fraction of constituent

Fig. 11. Elementray fracture mechanisms met in the fracture surfaces of Kahn Tear Test specimens. (a) Intermetallic fracture (intermediate quench,
under aged); (b) ductile fracture (fast quench. over aged); (c) ductile shear fracture (fast quench, under aged); (d) intergranular fracture (slow quench,
peak aged).
 D. Dumont et al. / Materials Science and Engineering A356 (2003) 326
/336 333

Table 3 the proportion of this fracture mode. It has to be

Area fractions in % of intermetallic Fe,Cu phases on the fracture emphasized that this fracture mode is ductile and very
surfaces, as a function of loading direction, quench rate and aging state
dissipative, since large local plastic deformations are
Loading direction Underaged Peak aged Oeraged involved. Thus it is not surprising to find the highest
values of toughness when it is present.
Fast Slow Fast Slow Fast Slow
The effect of aging on the distribution of fracture
LT loading 2.2 2.2 2.2 2.3 2.1 modes is twofold. During the aging treatment, the
TL loading 3.4 3.4 3.4 3.4 ductile shear fracture mode is replaced by the ductile
fracture mode, and an increase of the proportion of
These area fractions can be compared with the aeraged area fraction as
measured on a polished surface, which is 0.7%. intergranular fracture is observed. The proportion of
intergranular fracture is maximum at peak aging and
decreases slightly in the over-aging condition. This
particles present on the fracture surfaces. Table 3 shows
apparition of the intergranular fracture mode is related
the area fraction measurements for both pulling direc-
to the precipitation on grain boundaries which reduces
tions and a number of different microstructural states.
their cohesion, and to the development of the soft PFZ
The fraction of constituent particles is remarkably
which promotes strain localization in the grain bound-
constant for a given sample orientation, regardless of
ary vicinity. The decrease of intergranular fracture from
quench rate and aging treatment. In all cases it is much
the peak aging condition to the over aging state can be
higher than the bulk value of 0.7%. As expected, the explained by the lower difference in yield stress between
higher area fraction is encountered in the T
/L Kahn the PFZ and the grain interior in the latter case,
specimens, where the crack propagation in mode I is in promoting a more homogeneous plastic flow [52].
the rolling direction, which is also the direction along Finally, the effect of quench rate on the fracture mode
which the constituent particles are distributed. This is mainly to increase the proportion of intergranular
observation suggests that the main factor determining fracture. Notably, it is observed that the under-aged
the difference in notch resistance between the L
/T and material after a slow quench shows a significant
T
/L directions is the spatial distribution of the con- proportion of intergranular fracture, whereas the under-
stituent particles. aged material after a fast quench did not exhibit this
The other fracture events are clearly interrelated and fracture mode at all. This is undoubtedly due to the
cannot be studied separately. Their relative importance precipitation during the slow quenching of coarse
as a function of the microstructural state was quantified precipitates on the grain boundaries. This conclusion
by studying a large number of fracture surface micro- is further supported by the fact that the proportion of
graphs. This data is summarized in Table 4. intergranular fracture is almost independent on the
It can be first pointed out that the ductile shear mode aging condition in the case of a slow quench: the coarse
is only present in the underaged condition, and mostly quench induced precipitates are not significantly mod-
after a rapid quench. This fracture mode is actually ified during the aging treatment at 160 8C.
related to the presence of intense shear banding, which
originate in these microstructures due to the combina-
tion of shearable precipitates and a very low or negative 7. Discussion: effects of the microstructural features on
strain rate sensitivity, resulting in Portevin
/LeChâtelier notch resistance and yield stress
serrated flow. When the material is slowly quenched,
coarse intragranular particles probably inhibit the for- Having identified the main microstructural features
mation of macroscopic shear bands and thus decrease affecting the mechanical properties and the notch
resistance in our testing conditions, we are now in
Table 4
position to explain, the effect of aging, the effect of
Area fractions (in %) of the three main fracture modes (I, inter- quenching and the effect of pulling direction on the
granular; S, transgranular shear; D, transgranular ductile) general features of the [s0.2, UIE] diagram.
Loading direction Underaged Peak aged Overaged
7.1. Influence of the heat treatment
Fast Slow Fast Slow Fast Slow
We have shown that the evolution of yield stress
L
/T 10% I 45% I
/ 50% I 20% I 30% I
15% S 5% S
during aging is mainly controlled by the hardening
75% D 50% D 50% D 80% D 70% D precipitates. The maximum value of yield stress is
T
/L 20% I 25% I
/ 10% I 20% I
determined by the amount of solute not trapped in the
50% S coarse quench induced precipitates.
50% D 80% D 75% D 90% D 80% D The evolution of notch resistance during aging is
controlled both by the plastic properties of the matrix
334 D. Dumont et al. / Materials Science and Engineering A356 (2003) 326
/336

and the grain boundary structure. During the aging quench rate for the present double aging treatment.
treatment three parameters change simultaneously: the However, two phenomena can be observed: the time to
yield stress, which shows a maximum at peak aging, the peak aging is somewhat longer in the slowly quenched
strain hardening rate, which is high in the underaged materials, and the corresponding precipitate size is quite
condition and low in the peak aged and overaged larger. This shows that the slower quench rate results in
condition, and the grain boundary precipitation, which a less efficient GP zone formation and h? nucleation on
develops during aging if the material has been rapidly GP zones, leading to a less efficient hardening. Thus the
quenched, and is approximately independent of the heat effect of quench rate on the yield stress as a function of
treatment if the material has been slowly quenched. aging results from the combination of a less efficient
From the present experiments, we can conclude that hardening and a decrease in the solute available for
the two factors influencing most the evolution of notch hardening precipitation, leading to a lower volume
resistance during aging are grain boundary precipitation fraction of small precipitates [53].
and strain hardening rate. The effect of quench rate on notch resistance is much
more dramatic. Several factors can explain the severe
i) The important influence of grain boundary precipi-
decrease in notch resistance when the efficiency of the
tation is illustrated by the fact that the influence of
quench is decreased: grain boundary precipitation,
the aging condition on the notch resistance is much
coarse intragranular precipitation, yield stress of the
reduced when a slow quench rate is applied (Fig. 9).
matrix and work hardening rate. Given the magnitude
This is a consequence of the constant grain bound-
of the notch resistance decrease, the two last factors are
ary precipitation structure in the latter case, whereas
very weakly dependent on the quench rate and thus can
the yield stress and strain hardening rate evolve
be considered to be of secondary importance.
almost in the same manner as compared to the fast
Grain boundary precipitation is necessarily of major
quenched material.
importance. We have seen that it was controlling for a
ii) The influence of strain hardening rate (through the
significant part the decrease of notch resistance from the
energy dissipation at the crack tip during fracture,
under-aged to the peak-aged condition in the rapidly
see e.g. [15,16]) is evidenced by the difference in
notch resistance between the underaged and over- quenched material. We have also evidenced that a
aged materials in the slowly quenched material (Fig. slower quench resulted in a higher area fraction of
9): whereas it has the same grain boundary pre- intergranular fracture. However, the grain boundary
cipitate structure and almost the same yield precipitation is not sufficient to explain by itself the
strength, the underaged material shows a signifi- decrease in notch resistance, for the two following
cantly higher notch resistance. reasons:

In conclusion, the difference in notch resistance i) The toughest state (underaged) after a slow quench
between an underaged and overaged material of similar has a much lower notch resistance than the weakest
hardness is controlled mainly by the strain hardening state (both peak and overaged) after a fast quench,
properties in the case of a slow quench. However, when despite a much larger strain hardening rate and a
the heat treatment follows a rapid quench, the evolution much lower yield stress. The fast quenched overaged
of the notch resistance results from the combination of material having a large fraction of its grain bound-
strain hardening properties and grain boundary pre- aries covered with coarse precipitates, the contrast
cipitation processes. in notch resistance has to originate from an other
The absolute value of yield stress also influences the source.
notch resistance. The material in the overaged condition ii) As shown in Table 4, even in the slowly quenched
shows a reduced yield stress contrast between the PFZ material, the proportion of intergranular fracture
and the grain interior as compared to the peak aged on the fracture surfaces is never higher than 50%.
material, which promotes a more homogeneous flow. Thus, in order to explain such a high decrease in
This can explain the observed fact that it has a slightly notch resistance as is observed experimentally, it is
higher notch resistance. necessary to take into account a mechanism which
decreases the intrinsic resistance of the grain inter-
iors.
7.2. Influence of the quench rate
It is actually expected that the bands of coarse
We have observed that the evolution of yield stress as precipitates formed in the grain interiors on the dis-
a function of quench rate for a given aging condition is persoids during a slow quench significantly reduce the
relatively weak, namely 10% for the two extreme quench energy required to nucleate and coarsen voids in the
rates used in the present study. The general precipitation grain interior, thus decreasing their intrinsic toughness.
mechanisms are not fundamentally modified by the This mechanism must be responsible for a significant
 D. Dumont et al. / Materials Science and Engineering A356 (2003) 326
/336 335

part of the notch resistance decrease in the case of a slow Acknowledgements

quench.
The authors wish to thank Dr J.C. Ehrström and C.
Sigli, Professors J.D. Embury and A. Pineau for useful
7.3. Anisotropy of the mechanical properties discussions. This work was supported by Pechiney
Centre de Recherches de Voreppe and ANRT through
The loading direction in the (L,LT) plane is irrelevant a PhD grant (DD).
to many features of the microstructure: the grain
structure is rather isotropic, both intragranular and
intergranular precipitates are isotropically distributed as
well, the texture is not very pronounced, as evidenced by References
the very weak dependence of the yield stress on the
[1] T. Megson, Aircraft Structures, Butterworth Heinemann, Wo-
tensile direction. The main microstructural feature
burn, MA, USA, 1999.
which is anisotropically distributed is the constituent [2] J. Fielding, Introduction to Aircraft Design, Cambridge Uni-
particles, which are globally aligned in the rolling versity Press, Cambridge, UK, 1999.
direction. Thus the main effect of the loading direction [3] M. Ashby, D. Jones, Engineering Materials, vol. 1, Butterworth
on the notch resistance is on the damage initiation, Heinemann, Woburn, MA, USA, 1996.
which occurs much more densely in the T
/L specimens [4] I. Polmear, Light Alloys, Arnold, London, UK, 1995.
[5] J.A. Wagner, R.N. Shenoy, Metall. Trans. 22A (11) (1991) 2809.
as compared to the L
/T specimens. This is reflected by [6] R.C. Dorward, C. Bouvier, Mater. Sci. Eng. A254 (1998) 33.
the larger area fraction of constituent particles on the [7] N.U. Deshpande, A.M. Gokhale, D.K. Denzer, J. Liu, Metall.
fracture surface in the T
/L specimens, which conversely Mater. Trans. 29A (1998) 1191.
is at first order independent of all the other parameters [8] N. Kamp, I. Sinclair, M.J. Starink, Metall. Mater. Trans. 33A
(aging and quenching states). As a consequence, the (2002) 1125.
[9] G.T. Hahn, A.R. Rosenfield, Metall. Trans. 6A (1975) 653.
UIE is always smaller in the T
/L specimens, irrespec- [10] P. Achon, J.C. Ehstrom, A. Pineau, J. de Phys. IV 6 (1996) C6-3.
tively of heat treatment. [11] J.D. Embury, E. Nes, Z. Metallk. 65 (1974) 45.
[12] E. Hornbogen, M. Gräf, Acta Metall. 25 (1977) 877.
[13] T. Kawabata, O. Izumi, Acta Metall. 24 (1976) 817.
[14] B.Q. Li, A.P. Reynolds, J. Mater. Sci. 33 (24) (1998) 5849.
[15] G.G. Garret, J.F. Knott, Metall. Trans. 9A (1978) 1187.
8. Conclusions [16] C.Q. Chen, J.F. Knott, Met. Sci. 15 (1981) 357.
[17] T. Kawabata, O. Izumi, Acta Metall. 25 (1977) 505.
A detailed analysis of the microstructure evolution for [18] M. Sugamata, C.P. Blankenship, E.A. Starke, Jr., Mater. Sci.
different quench rates and heat treatments, coupled with Eng. A163 (1993) 1.
a fractography analysis of the elementary failure modes [19] A.M. Gokhale, N.U. Deshpande, D.K. Denzer, J. Liu, Metall.
Mater. Trans. 29A (4) (1998) 1203.
allows to understand qualitatively the main reasons for [20] J.T Staley, R.D. Doherty, A.P. Jaworski, Metall. Trans. 24A
the evolution in the toughness/yield stress compromise (1993) 2417.
in the 7050 alloy. The spatial distribution of constituent [21] D. Broek, Eng. Fract. Mech. 5 (5) (1973) 55.
particles has been identified as the main cause of [22] G. Lapasset, C. Renon, la Recherche Aérospatiale 5 (1985) 313.
toughness anisotropy. Intergranular precipitation as a [23] H. Cai, J.T. Evans, Eng. Fract. Mech. 42 (4) (1992) 589.
[24] C.J. Peel, R.N. Wilson, P.J.E. Forsyth, Met. Sci. J. 6 (1972) 102.
process weakening the strength of the grain boundaries, [25] J.A. Blind, J.W. Martin, Mater. Sci. Eng. 57 (1983) 49.
together with the evolution of the strain hardening [26] T. Ohira, T. Kishi, Mater. Sci. Eng. 78 (1986) 9.
capability of the material depending on the state of [27] B.J. Dunwoody, D.M. Moore, A.T. Thomas, J. Inst. Met. 101
precipitation, account both for the low value of tough- (1973) 172.
[28] D.S. Thompson, B.S. Subramanya, S. Levy, Metall. Trans. 2
ness at the peak aged treatment, and for the difference in
(1971) 1149.
toughness for underaged and overaged samples with [29] D.S. Thompson, R.E. Zinkham, Eng. Fract. Mech. 7 (1975) 389.
similar yield stress. After a slow quench, coarse intra- [30] D.S. Thompson, Metall. Trans. 6A (1975) 677.
granular precipitates appear to play a key role in the [31] A.K. Vasudévan, R.D. Doherty, Acta Metall. 35 (1987) 1193.
decrease of the intrinsic toughness of the grain interiors. [32] R. Lebensohn, G. Canova, D. Solas, Y. Brechet, Acta Mater. 44
The present approach provides some guidelines for alloy (1) (1996) 315.
[33] N. Ryum, B. Haegland, T. Lindtveit, Z. Metallk. 58 (1967) 28.
development when a good compromise between yield [34] N. Ryum, Acta Metall. 16 (1968) 327.
stress and toughness is searched for. The quantitative [35] S.P. Lynch, A.R. Wilson, R.T. Byrnes, Mater. Sci. Eng. A172
interpretation of the various phenomena studied in the (1993) 79.
present paper is in progress, involving both the estima- [36] R. Hamano, Metall. Trans. 24A (1) (1993) 127.
[37] N. Ryum, Acta Metall. 17 (1969) 921.
tion of toughness in the case of competing failure modes
[38] K. Welpmann, A. Gysler, G. Lütjering, Z. Metallk. 71 (1980) 7.
[7], and the quantitative micro-mechanical modeling of [39] A.T. Zehnder, C.Y. Hui, Scripta Mater. 42 (10) (2000) 1001.
the influence of grain boundary precipitation on the [40] T. Pardoen, J.W. Hutchinson, J. Mech. Phys. Solids 48 (2000)
interfacial strength [11,54]. 2467.
336 D. Dumont et al. / Materials Science and Engineering A356 (2003) 326
/336

[41] C. Sauer, F. Busongo, G. Lutjering, Mater. Sci. Forum 396
/402, USA, 1994, Georgia Institute of Technology Atlanta, GA, USA
part 2 (2002) 1115 (Proc. of the 8th ICAA, Cambridge UK). (1994) 443.
[42] O.E. Alarcon, A.M. Nazar, W.A. Monteiro, Mater. Sci. Eng. [49] T. Kobayashi, M. Niinomi, K. Ikeda, J. Jpn. Light Met. 38 (1)
A138 (1991) 275. (1988) 9.
[43] R.C. Dorward, D.J. Beerntsen, Metall. Trans. 26A (1995) 2481. [50] T. Kobayashi, M. Niinomi, Y. Takabayashi, S. Kohmura, J. Jpn.
[44] B. Morere, J.C. Ehrström, P.J. Gregson, Metall. Mater. Trans. Light Met. 38 (11) (1988) 723.
31A (10) (2000) 2503. [51] A. Deschamps, M. Niewczas, F. Bley, Y. Brechet, D. Embury, L.
[45] U.F. Kocks, J. Eng. Mater. Technol. 98 (1976) 76. Lesinq, F. Livet, Phil. Mag. A79 (10) (1999) 2485.
[46] Y. Estrin, in: A.S. Krauz, K. Krauz (Eds.), Unified Constitutive [52] G.M. Ludtka, D.E. Laughlin, Metall. Trans. 13A (1982)
Laws of Plastic Deformation, Academic Press, London, UK, 411.
1996, p. 69. [53] A. Deschamps, Y. Bréchet, Mater. Sci. Eng. A251 (1998)
[47] J.G. Kaufman, A.F. Knoll, Mater. Res. Std. 4 (1964) 151. 200.
[48] H.D. Dudgeon, N.C. Parson, S.A. Court, R.A. Ricks, Proceed- [54] T. Pardoen, D. Dumont, Y. Brechet, A. Deschamps, J. Mechanics
ings of the 4th Int. Conf. on Aluminum Alloys, vol. I, Atlanta, Phys. Solids, in press.