-
in
Eh
e me ine frougnisondconnsvtate
alumipertie
s, espehe eighe addduluser cono ductatures
s
science community on all aspects of primary and secondary
ke T2 and TB candaries [27]. Thecipitates of the S
d since the earlyto intergranular2.2 wt% Li (e.g.ion is
commonly
ductile intergranular (with shallow dimples) and brittle
intergra-
Contents lists available at ScienceDirect
.el
Materials Science
Materials Science & Engineering A 586 (2013)
418427boundaries. Other precipitates include S and T1 phases within
theE-mail address: [email protected] (A.
Deschamps).nular (featureless except some trace of intergranular
precipitates)fracture is observed. In these alloys the common
feature is thepresence of a high volume fraction of -Al3Li ordered
precipitatesin the matrix, except in a precipitate-free zone close
to the grain
0921-5093/$ - see front matter & 2013 Elsevier B.V. All
rights reserved.http://dx.doi.org/10.1016/j.msea.2013.06.075
n Corresponding author. Tel.: +33 4 76 82 66 07; fax: +33 4 76
82 66 44.programs, for instance under the commercial name
AIRWARE[12]. They attract currently a strong interest in the
materials
observed [2835] in various microstructure conditions, from
veryunderaged to later stages of ageing. Classically a mixture
betweenthe competition between Al alloys and composite materials
for air-craft structures, and durably high fuel prices, new alloy
develop-ment has led to alloys with optimised compositions where
most ofthe issues of former generation alloys have been overcome
[47].With the right combination of copper and lithium contents a
moreefcient precipitation strengthening can be reached,
particularlywith the T1 phase that forms very efcient obstacles to
dislocationmotion because of its thin platelet shape of high aspect
ratio [811]. These alloys nd a number of applications in new
airplane
(nominally Al2CuLi). However, depending on the aand processing
conditions other minor phases libe precipitated, particularly at
the grain bounpresence of Mg can lead to the formation of
pre(Al2CuMg) sequence.
One of the key issues that has been recognisedevelopment of
AlLiCu alloys is their propensityfracture. In alloys containing of
the order of 1.8AA8090, AA2090, AA2091), intergranular
decohestolerance or high costs of processing. However, in the
context of phases, the one offering the highest strength being the
T1 phaselloy composition1. Introduction
Among precipitation hardeningalloys possess a combination of
proattractive for structural applicationsector. Developments
conducted in twere based on the main factor that talloy density and
increases the mohigh lithium content relative to coppapplications,
due to issues related tterm ageing at relatively low tempernium
alloys, AlCuLis that has made themcially in the aerospacehties and
early ninetiesition of Li decreases the[1,2]. Early alloys withtent
have found limitedility losses during long[3], insufcient
damage
processing, microstructure development and various
properties[1325].
The ternary AlCuLi system experiences a complex precipita-tion
sequence, exhibiting aspects of both binary AlCu and AlLisystems
(see [26] for a more complete description). The (standard)addition
of minor alloying elements such as Mg and Ag providesadditional
complexity. The binary sequence of AlCu leads tosingle atomic layer
GPI zones as well as GPII and precipitates,whereas the AlLi
sequence gives the (Al3Li) phase, the forma-tion of which is known
to depend strongly on the Li content. Heattreating the ternary
AlCuLi system can generate a number ofOn the role of microstructure
in governof an aluminumcopperlithium alloy
B. Decreus a,b, A. Deschamps a,n, P. Donnadieu a, J.C.a SIMAP,
INP Grenoble CNRS UJF, BP 75, 38402 St Martin d'Hres Cedex, Franceb
Constellium CRV, Voreppe Research Center, BP 27, 38341 Voreppe
Cedex, France
a r t i c l e i n f o
Article history:Received 26 April 2013Received in revised form24
May 2013Accepted 18 June 2013Available online 6 July 2013
Keywords:AlCuLiPrecipitationIntergranular fracture
a b s t r a c t
The inuence of precipitatAlCuLi alloy, AA2198. Thby changing the
quench rattreatment time. Fracture tevidence clearly the
mechasignicantly occurs in all cThis mechanism is mainlydetermine
the value of tramainly controlled by the s
journal homepage: wwwg fracture behavior
rstrmb
icrostructure on fracture mechanisms is studied in a recently
developedtra-granular and inter-granular microstructures are varied
independentlyom the solution heat treatment, the amount of
pre-stretching and the heathness is evaluated by short bar chevron
tear tests that make possible toms of inter-granular fracture. It
is shown that intergranular ductile fractureitions of heat
treatment where substantial precipitation has taken place.trolled
by the state of inter-granular precipitation and plays a major role
toerse fracture toughness, while the strength and ductility of the
alloy areof intra-granular precipitation.
& 2013 Elsevier B.V. All rights reserved.
sevier.com/locate/msea
& Engineering A
-
grains and either phase [28] or Li and Cu containing phases(T2
and TB) [36] at the grain boundaries. Different causes havebeen
invoked to explain this extensive occurrence of
intergranularfracture and have received extensive attention
particularly in the1990s, and have been reviewed in several papers
([29,30,32,34,35]and references therein): presence of intergranular
precipitatesassociated to the presence of precipitate free zones
(particularlyin terms of -Al3Li precipitation), planar slip (due to
the collectiveshearing of precipitates), embrittlement of grain
boundaries byliquid metal impurities, or a loss of grain boundary
coherency dueto Li segregation. The rst mechanism was deemed more
impor-tant in cases where alloys exhibit ductile intergranular
fracture, onthe basis of the experimental evidence that toughness
andoccurrence of intergranular fracture were well correlated to
thegrain boundary precipitate microstructure [28]. Conversely,
thelast hypothesis seems nowadays to be favoured in the cases
wherebrittle intergranular fracture is observed [34,35].
However, most recently developed alloys such as AA2198 orAA2050
have a Li content limited to lower concentrations (of theorder of 1
wt%), in order to minimise or even suppress the
conditions has proven to be a very efcient way to identify
clearlythe intergranular fracture mode.
Following a detailed study of the microstructure
developmentinside the grains and related mechanical properties
during heattreatment of the AA2198 alloy that has been published
recently[26,38], the aim of the present paper is to evaluate the
evolutionof fracture mechanisms and related fracture toughness in a
largevariety of heat treatments where different parameters are
variedacting on the microstructure at the grain boundaries (quench
ratefrom the solution treatment temperature, subsequent ageing
time)and on the intra-granular microstructure (pre-stretch and
ageingtime). The mechanical properties are evaluated by simple
tensiletests and by short transverse chevron notch (short bars)
samples,giving access to short transverse fracture toughness and
relatedfracture surfaces. The microstructure at the grain
boundariesis characterised by electron microscopy and the
intra-granularmicrostructure has been characterised in the formerly
publishedwork [26].
BW2H in Fig. 2. The tests were carried out at a traverse speedof
0.5 mmmin1 with a pre-load of 5 N, according to ASTME1304-97 [39].
The conditional plane strain toughness K was
B. Decreus et al. / Materials Science & Engineering A 586
(2013) 418427 419formation of the phase. These alloys are
completely free of and their intragranular microstructure in the T8
aged condition isdominated by the T1 phase and, to a lesser extent,
the phase[26]. In such alloys, pronounced intergranuar fracture is
stillobserved in the aged T8 temper [23,24]. Similarly, other
alloyswith relatively low Li concentration such as AA2020, which
areexempt of precipitation have been also shown to be subject
tointergranular fracture and delamination [37]. However, the
rela-tionship between the microstructure of these low-Li
containingalloys and their fracture mode has not been studied in
great detail.The limited number of studies concerning the fracture
behaviourof these recent alloys [23,24] have reported mechanical
testing inthe rolling plane and no detailed microstructural
analysis. Slantedfracture is promoted by the intergranular fracture
mode, and abrous fracture surface is observed, related to the
unrecrystallizedgrain structure, and related very high grain aspect
ratio. The shearcomponent of deformation makes it difcult to
determine pre-cisely the microstructural features linked with the
intergranularfracture mechanism. Although it has been proposed that
reducingthe Li content may be an efcient way to inhibit Li
segregation tothe grain boundaries [35], it is not clear if these
alloys are stillsubjected to this mechanism. In the high Li
containing alloys, mostauthors have used short transverse testing
in order to evidence themechanisms of intergranular fracture
[3134]. Using such testingFig. 1. Optical micrograph of the grain
structure.Qv2. Material and experimental methods
Alloy AA2198 has the composition range (all in wt%)
[2.93.5]Cu[0.81.1]Li[0.250.8]Mg[0.10.5]Ag[0.040.18]Zr. It
wasprovided by Constellium Voreppe Research Centre, France,
asrolled sheet of 12 mm thickness with a fully brous grain
struc-ture. The sheets showed a strong Brass texture (112 {110}).
Fig. 1shows an optical micrograph of the grain microstructure.
The heat treatments consisted rst in a solution treatment anda
quench. Unless stated otherwise, the samples were quenchedinto cold
water. Some samples were also quenched into hot water(80 1C) or air
cooled, in order to vary the grain boundary micro-structure prior
to the articial ageing treatment. The samples werethen stretched at
least 2% plastic strain and kept several weeks atambient
temperature (T351 temper). The articial ageing treat-ment consisted
in a heating ramp to 155 1C at 20 K h1, followedby an isothermal
hold at 155 1C.
Tensile tests were carried out in the rolling (L) direction of
theplates using at tensile tests of section 53 mm2 and gaugelength
60 mm, at a constant strain rate of 1.5104 s1.
Short bar chevron tear tests were carried out with a
loadingdirection in the short transverse (ST) direction of the
plates withsample dimensions (in mm) of 1912.511 corresponding
toFig. 2. Geometry of the short bar chevron notched sample.
-
calculated as follows:
KQvM YnmFmB
Wp
where Q stands for the conditional nature of the
measuredtoughness, and M states for the determination from the
maximumload. Ynm is the tabulated minimum stress intensity factor
(depend-ing on the sample geometry), and Fm is the maximum
loadmeasured during the test. Some conditions must be met
tocalculate this toughness value [39]. All the results presented
here,except for the observation of fracture surfaces, were obtained
invalid conditions; other results are omitted.
Fracture surfaces were observed using a standard
scanningelectron microscope LEO Stereoscan 440 at 20 kV. High
resolutionimages were obtained on a Zeiss Ultra-55 FEG-SEM at 4 kV
within-lens secondary and back scattered detectors.
Conventional transmission electron micrographs were obtainedon
electropolished samples with a Jeol 3010 microscope operatingat 300
kV. Complementary observations for the grain boundarymicrostructure
were performed in scanning transmission electronmicroscopy (STEM)
mode on a FEI Titan 80300 equipped a highangle annular dark eld
(HAADF) detector.
3. Tensile tests
Fig. 3 shows the tensile curves obtained for samples heattreated
from 1 h to 500 h at 155 1C. The microstructure evolutionduring
this ageing treatment has been described in detail in [26].
0
100
200
300
400
500
600
0 2 4 6 8 10 12 14 16
True
stre
ss (M
Pa)
True strain (%)
1h
4h8h
16h100h
500h
Fig. 3. Tensile curves as a function of ageing time at 155
1C.
esol
B. Decreus et al. / Materials Science & Engineering A 586
(2013) 418427420Fig. 4. Fracture surfaces of tensile test samples
(SEM micrographs) at low and high r
155 1C.ution for two different states of ageing: (a) and (b) 4 h
at 155 1C; (c) and (d) 16 h at
-
-1/2
), 4*
U.A
. (%
)
B. Decreus et al. / Materials Science & Engineering A 586
(2013) 418427 421300
400
500
d (N
)
1h
8h
16h
100h
4hAfter 1 or 4 h at 155 1C, the microstructure consists mainly
of ahomogeneous solid solution, and no signicant fraction of
thestrengthening T1 precipitates is present. After 8 h at 155 1C,
theformation of the T1 phase is really signicant and the
microstruc-ture between 16 and 500 h is mostly stable, consisting
on afraction close to equilibrium of very ne T1 platelets resulting
inan almost constant yield strength. Fig. 4 shows the
fracturesurfaces of the tensile specimens tested in two ageing
conditions,namely in the absence of T1 precipitates (4 h (a, b))
and near peak
0
100
200Loa
Displacement (mm)
500h
0 0.5 1 1.5 2
Kqv
M (M
Pa.
m
Fig. 5. (a) Loaddisplacement curves for the short bar tests on
samples aged for differenwith corresponding yield stress and
uniform elongation of tensile tests in the same age
Fig. 6. Low resolution fracture surfaces (SEM micrographs) of
the notch tip of the short bAn increase of the fraction of at
intergranular fracture is observed when the ageing ti40
50
60
70
300
400
500
600
Yield stress
Yield stress
Toughnessstrength (16 h (c,d)). In the rst case, fracture
appears to becompletely ductile trans-granular. The dimples are
oriented dueto the slanted geometry of the fracture surface. In the
second case,the fracture surface appears as multiple ne steps of a
few mm,with smooth walls between them. These fractographs
resemblethat obtained in similar conditions by Chen et al. [23] and
Steglichet al. [24] and are characteristic of a fracture controlled
byintergranular decohesion with a heavily elongated grain
structurein the loading direction. Due to shearing and friction
during the
0
10
20
30
0
100
200
1 10 100 1000
(MP
a)
Time at 155C
Elongation
t times at 155 1C. (b) Values of toughness measured from the
short bar tests alonging conditions.
ar tests for the following ageing times at 155 1C: (a) 1 h; (b)
8 h; (c) 16 h; (d) 100 h.me proceeds.
-
B. Decreus et al. / Materials Science & Engineering A 586
(2013) 418427422last stages of the fracture process, it is
impossible to describe themicroscopic mechanisms that explain this
occurrence of intergra-nular fracture from these observations.
Short bar chevron tests
Fig. 7. Detail of ductile fracture area of the short bar sample
aged 1 h at 155 1C: (a) in seelectron mode showing the presence of
large intermetallics associated with the dimple
Fig. 8. Detail of ductile fracture area of the short bar sample
aged 100 h at 155 1C: (aintermetallics on the intergranular
surface; (b) medium resolution secondary electrointergranular
fracture surface; (c) high resolution secondary electrons image
showing tpresented in the next paragraph allow for loading in the
directionnormal to the grain boundaries and observing the details
of thefracture mechanisms.
condary electron mode showing the presence of dimples and (b) in
back-scattereds.
) low resolution image in back-scattered electrons mode showing
no large scalens image showing the presence of a high density of
small dimples on the athe presence of small particles in these
dimples.
-
B. Decreus et al. / Materials Science & Engineering A 586
(2013) 418427 4234. Chevron notch short bar tests
Fig. 5(a) shows the loaddisplacement curves for the chevronnotch
short bar tests after different heat treatment times. For
thesamples heat treated 1 h and 4 h at 155 1C where the material
isstill very soft, generalised plasticity prevented the measurement
offracture toughness from these tests. However, it appears
clearlythat the behaviour changes dramatically between 4 and 8 h,
thenbetween 8 h and 16 h, and again between 16 h and 100 h,
afterwhich it remains stable. Fig. 5(b) shows the evolution with
heattreatment time of three parameters of the mechanical
tests,namely the yield strength and uniform elongation obtained
fromthe tensile tests, and the fracture toughness obtained from
theshort bar tests. As evidenced in [26], the evolution of yield
stress isessentially representative of the evolution of the volume
fractionof T1 precipitates inside the grains. When T1 precipitates
rstappear the yield stress starts to increase, and when this
fractionsaturates, the yield stress stabilises around 500 MPa. The
evolutionof elongation is quite well correlated to this evolution
of yieldstress, as discussed in detail in [38]. The conjunction of
theincrease in yield strength and the corresponding reduction of
thestrain hardening rate capability results in a reduction in
uniformelongation. Interestingly, the evolution of fracture
toughness isquite different from that of the uniform elongation.
The toughnessalso decreases with heat treatment time, but this
decrease occurslater. In the rst stages of the increase of yield
strength, thestrength increase compensates the reduction in
elongation to keepthe toughness roughly constant. It is only after
50 h at 155 1C that
Fig. 9. Transmission electron micrographs showing the
microstructure at the grain and susample aged for 16 h at 155 1C;
(b) dark-eld micrograph showing a grain boundary inboundaries in
the sample aged for 100 h at 155 1C.the toughness decreases
signicantly, while the yield strength andelongation (albeit tested
with a different loading direction) arenow constant with ageing
time.
In order to understand the origin of this toughness drop
duringthe peak strength plateau, fractographic observation of the
shortbar samples will now be presented.
Fig. 6 presents low resolution images of the fracture
surfacesclose to the notch tip, for different heat treatment times
(1 h, beforethe increase in strength; 8 h, during the strength
increase; 16 hclose to peak strength before the drop in toughness;
and 100 h atpeak strength after the drop in toughness). In the rst
sample thefracture is entirely ductile transgranular. No at areas
are recorded.After 8 h and 16 h at 155 1C, a large majority of the
fracture surfaceis still ductile transgranular; however a few at
and smooth (at thisscale) surfaces can be observed that correspond
to interganularfracture. After 100 h at 155 1C, most of the
fracture surface consistsof at and smooth intergranular zones. It
is therefore tempting toassociate the decrease in fracture
toughness to the occurrence of afracture mode of low energy
dissipation [40], namely inter-granulardecohesion along grains that
have a large extension normal to theshort transverse loading
direction.
The details of the fracture surface are shown in Fig. 7 for
thesample aged 1 h at 155 1C, and in Fig. 8 for the sample aged 100
hat 155 1C. In the rst case, the ductile transgranular fracture
isclassically characterised by large dimples, which include
brokenintermetallic particles. These particles have a bright
contrast(Fig. 7b) indicating a larger average atomic number than
thematrix that is conrmed by the composition close to Al7Cu2Fe
b-grain boundaries. (a) Dark-eld micrograph showing a sub-grain
boundary in thethe sample aged for 16 h at 155 1C; (c) and (d)
STEM-HAADF micrographs of grain
-
ing treatment, the intragranular and intergranular
microstructures
rt bf qu
B. Decreus et al. / Materials Science & Engineering A 586
(2013) 418427424given by EDS analysis. In the second case, the
situation is entirelydifferent. The at areas of the fracture
surface do not present anychemical contrast at low resolution (Fig.
8(a)), which means thatno coarse intermetallic particle is
associated with this fracturemechanism. However, at higher
resolution a homogeneous dis-tribution of very ne dimples is
observed (Fig. 8(b) and (c)). Theirsize is between 100 and 200 nm,
and very small particles (lessthan 50 nm in diameter) are observed
at the core of the dimples.A quantitative chemical analysis of
these particles is clearly out ofthe range of EDS-SEM, but a
qualitative comparison between theEDS signal on these particles and
on the surrounding matrixclearly shows that they are enriched in Cu
and not in Fe. It is notpossible using this technique to estimate a
potential enrichment inLi. Although not shown here for space
reason, it should bementioned that similar observations were made
on all at areasof short bar samples, with a general tendency of
decreasing thedimple size with increasing the ageing time [41].
These fractographic observations provide good evidence thatthe
intergranular fracture mode responsible for the at areas onthe
short bar test fracture surfaces is actually ductile, with adamage
initiation related to ne Cu-rich particles lying on the
30
35
40
45
50
55
60
Undeformed
T351
Kqv
M (M
Pa.
m-1
/2)
Microhardness
Def 16h
Undef 16h
Undef 100hDef 100h
Increasing ageing time
100 120 140 160 180 200
Fig. 10. (a) Evolution of the compromise between toughness (as
measured by the shocompared to samples aged from an undeformed
condition; (b) inuence of the rate otime at 155 1C.grain
boundaries. These Cu-rich particles, less than 100 nm indiameter,
are likely to be intergranular precipitates formed duringthe
articial ageing treatment. In order to conrm this hypothesis,TEM
observations are reported in Fig. 9. Fig. 9(a) and (b)
showsconventional dark eld micrograph of a sub-grain boundary and
agrain boundary after 16 h at 155 1C (at the beginning of the
peakstrength plateau). In the matrix, a high density of ne T1
plateletsis observed. At sub-grain boundaries, this density is even
higher,but the precipitates have a similar shape and size. Since
T1precipitates are known to nucleate on dislocations,
sub-grainboundaries can be regarded as regions of particularly high
densityof nucleation sites for the precipitates. At the grain
boundaries,however (Fig. 9(b)), hardly any precipitation can be
observed,which is consistent with the fact that fracture is still
mostlytransgranular in the short bar tests. Fig. 9(c) and (d)
showsSTEM-HAADF (Z-contrast) images at grain boundaries of
thesample aged 100 h at 155 1C. Inside the grains the
microstructureis qualitatively identical to that of the sample aged
16 h (it canactually be shown that the two intragranular
microstructures arequantitatively almost identical [26]). However
the grain bound-aries now include a high density of incoherent
particles of sizeapproximately 50 nm. Their bright contrast in the
HAADF micro-graphs tells that they must be Cu-rich, in agreement
with [36] andevolved concurrently, so that unravelling their
respective effectsremains challenging.
In order to provide additional evidence, we have sought tochange
independently the intergranular and intragranular micro-structures
in two ways. First, we have compared samples aged inthe classical
way (namely quenched, stretched and aged) towith similar
observations in [13]. These precipitates may be T2 orTB phases;
however their crystal structure was not determinedhere. Their size
and morphology suit particularly well with that ofthe particles
observed at the core of the dimples on the fracturesurfaces (Fig.
8(c)).
5. Effect of the variation of intergranular and
intragranularmicrostructures
The preceding experiments show that the evolution of trans-verse
toughness is correlated with the state of
intergranularprecipitation, which in turn controls the occurrence
of ductileintergranular fracture. However, during the studied
articial age-
0
10
20
30
40
50
60
70
0 20 40 60 80 100 120
Kqv
M (M
Pa.
m-1
/2)
Time at 155C (h)
Fast quench
Intermediate quench
Slow quench
ar test) and microhardness in the sample aged at 155 1C from the
T351 condition asenching on the toughness (as measured by the short
bar test) evolution with ageingsamples where the stretching step
was omitted. As shown in[42], intragranular precipitation in the
absence of pre-deformationis very sluggish and the precipitation
kinetics is decreased by afactor of 10. However it is likely that
the precipitation kinetics atthe grain boundaries is at rst order
unaffected by the absence ofstretching and thus it becomes possible
to compare samples withdifferent intra-granular precipitates and
similar inter-granularmicrostructure. Secondly, we have compared
samples aged bythe same procedure (including the stretch), but
having experi-enced different quenching conditions from the
solution treatment(intermediate quench, in hot water and slow
quench, air cooled).These samples can thus be expected to have
similar intragranularprecipitates but different inter-granular
microstructures, with ahigher fraction of intergranular
precipitates in the slowly cooledmaterials.
Fig. 10(a) shows a graph representing the different sets
ofmicrohardness/toughness parameters for samples aged at 155 1Cin
the deformed and undeformed materials. In both cases aninverse
correlation exists, however very different in the twosamples. As
seen in the former sections, for the stretchedsamples, at near peak
strength (16 h) the toughness is still veryhigh. At this ageing
time, the unstretched material shows acomparable toughness value;
however the microhardness is
-
B. Decreus et al. / Materials Science & Engineering A 586
(2013) 418427 425much lower due to the sluggish intragranular
precipitationkinetics. After 100 h at 155 1C, the unstretched
material hasreached the hardness value of the stretched material
aged 16 h,but has a much lower toughness. Its toughness is now
similar tothat of the stretched material aged 100 h at 155 1C.
Therefore allthese results are consistent with a hardness
controlled by thestate of intragranular precipitation and a
toughness mostlycontrolled by the inter-granular precipitation
(which is at rstorder controlled by the ageing time).
Fig. 10(b) shows the evolution with heat treatment time
oftoughness for samples quenched with different procedures
(fast,intermediate and slow). Clearly, decreasing the quench
ratereduces very signicantly the toughness value in all
ageingconditions. The remaining evolution is more complex, with
aconvergence of the intermediate and fast quench samples at
longageing time, while the slowly quenched material keeps a
lowertoughness at any ageing time. The fracture surfaces of the
shortbar tests after 100 h at 155 1C for the intermediate or
slowquenched materials are shown in Fig. 11. They present a
highfraction of at intergranular areas, with a high density of
smalldimples (albeit larger than that of the fast quenched
material,presumably due to a larger size of intergranular
precipitates), andsome areas of facetted brittle fracture, which
could correspond tothe formation at the grain boundaries of a
continuous layer of abrittle incoherent intermetallic phase after a
sufciently slowquench.
Fig. 11. Fracture surfaces (SEM micrographs) of short bar
samples on slowly quenched m6. Discussion
As summarised in the Section 1, the fracture behaviour of AlLiCu
alloys containing of the order of 2 wt% Li has been exten-sively
investigated in the literature. These alloys have been shownto be
prone to intergranular fracture, often of brittle nature.
Thisbehaviour was shown to be strongly related to the presence of
Liby several mechanisms that depend on the alloy composition
(andrelated precipitates formed) and heat treatment, namely Li
segre-gation, grain boundary precipitates, zones free of
precipitates atthe grain boundaries, and planar slip due to the
ordered phases.In the more recently developed alloys where the Li
content islimited to about 1 wt% in order to suppress the formation
of tothe benet of the T1 phase, the literature on the fracture
mechan-isms is still incomplete, although intergranular fracture
has alsobeen observed in such alloys to play a prominent role.
Themicrostructural observations made in the present paper and inthe
two companion papers published earlier on the same alloy[26,38], as
well as the fractographic observations reported here,show that
these alloys have a very different microstructure ascompared to
their higher Li counterparts. This is especiallyimportant with
respect to the mechanisms that have been invokedto explain the
fracture behaviour. First, they do not show any precipitation, nor
any precipitate-free zone (PFZ). After ageing, theT1 platelets have
an extension of more than 50 nm, which helpsto explain the absence
of PFZ because any precipitate nucleated
aterials: (a) intermediate quench, 100 h ageing; (b)(d) slow
quench, 100 h ageing.
-
grains, so that we believe that it is related to the decohesion
along
[12] T. Warner, J.C. Ehrstrm, B. Chenal, F. Eberl, Light Met.
Age 67 (2009) 3.[13] N. Brodusch, M. Trudeau, P. Michaud, L.
Rodrigue, J. Boselli, R. Gauvin, Microsc.
B. Decreus et al. / Materials Science & Engineering A 586
(2013) 418427426the interface of large facetted precipitates. In
all other cases themacroscopically at intergranular fracture
surfaces included a highdensity of very small dimples (size of
about 100 nm), associatedwith precipitates whose size and
composition were consistentwith that observed at the grain
boundaries in the transmissionelectron microscope, namely Cu-rich,
of average atomic numberhigher than the matrix as proven by the
higher brightness in Z-contrast.
Therefore, it can be expected that the mechanisms
inducingintergranular fracture may be of a different nature in the
high-Liand low-Li containing alloys. Although we cannot rule out
com-pletely a loss of the grain boundary cohesion due to the
presenceof Li in the alloy, a good indication that such decohesion
does notplay a major role in the low-Li containing alloy
investigated here isthat we have not observed brittle fracture. In
addition the values oftoughness are quite high for transverse
toughness tests in pre-cipitation hardening aluminium alloys, which
is a further indica-tion that grain boundary cohesion is not
affected at rst order.Ductile grain boundary fracture is usually
related to three para-meters [28,43]: the presence of grain
boundary precipitates, a softprecipitate free zone, and a high
stress triaxiality. In the presentcase no signicant precipitate
free zone is visible. However thecompetition between intergranular
and transgranular fractureappears to be affected by the trixiality
ratio, since grain boundaryfracture is systematically observed in
the areas of highest triaxi-ality of the chevron notch specimen
(close to the initial notch).This effect of external triaxiality
may actually help the acceleratedvoid growth in such conditions
where no PFZ is present.
The evolution of fracture mechanism as well as the evolution
oftoughness with the different microstructural states evaluated
hereare consistent with the proposed mechanism controlling
inter-granular fracture, namely nucleation and growth of voids at
thegrain boundary precipitates. The toughness drops
simultaneouslyto the increase in the fraction of intergranular
fracture whenageing proceeds, and this happens independently (at
rst order) ofthe intragranular precipitation kinetics, which was
varied byplaying on the plastic deformation prior to the heat
treatment.Similarly, a slower quench rate from the solution
treatment, whichfavours a high fraction of grain boundary
precipitates, results in alower fracture toughness.
7. Conclusions
In the present work we have attempted to shed light on
thefracture mechanisms that prevail in the recently developed
AlCuLi alloys with moderate Li content where the microstructure
afterarticial ageing is dominated by the precipitation of the T1
phaseand where no phase is formed. In order to study the
microscopicmechanisms related to intergranular fracture, which is
frequentlymet in these alloys, testing has been performed in the
shorttransverse direction. By studying the effect of heat treatment
time,pre-stretch before ageing, and quench rate from the solutionat
a distance smaller than 25 nm from a grain boundary canextend to
the grain boundary provided that some Cu and Lisupersaturation
still remains. Secondly, although they are proneto planar slip in
the naturally aged condition due to the presenceof solute clusters,
when aged, their slip becomes quite homoge-neous [38], because of
the particular shearing mechanism of the T1precipitate, which is
very different from that of . Thirdly, noevidence of brittle
intergranular fracture has been found in thepresent study for all
the heat treatments investigated, except forthe slowly quenched
material. However in the latter case, thisbrittle fracture was
facetted and did not follow the shape of thetreatment, we have been
able to evaluate independently the effectMicroanal. 18 (2012)
1393.[14] N. Ward, A. Tran, A. Abad, E.W. Lee, M. Hahn, E. Fordan,
O.S. Es-Said, J. Mater.
Eng. Perform. 20 (2011) 989.[15] R.J. McDonald, A.J. Beaudoin,
Model. Simul. Mater. Sci. Eng. 18 (2010) 065007.[16] S. Kalyanam,
A.J. Beaudoin, R.H. Dodds Jr., F. Barlat, Eng. Fract. Mech. 76
(2009)
2174.[17] I. Eberl, C. Hantrais, J.-C. Ehrtsrom, C. Nardin, Sci.
Technol. Weld. Join. 15 (2010)
699.[18] S. Richard, C. Sarrazin-Baudoux, J. Petit, Proceedings
of the TMS 2009,
San Francisco, Minerals, Metals & Materials Soc, 2009, pp.
6976.[19] A. Astarita, A. Squillace, A. Scala, A. Prisco, J. Mater.
Eng. Perform. 21 (2012)
1763.[20] C. Bitondo, U. Prisco, A. Squilace, P. Buonadonna, G.
Dionoro, Int. J. Adv. Manuf.
Technol. 53 (2011) 505.[21] P. Cavaliere, J. Light Met. 2 (2002)
247.[22] P. Cavaliere, M. Cabibbo, F. Panella, A. Squillace, Mater.
Des. 30 (2009) 3622.[23] J. Chen, Y. Madi, T.F. Morgeneyer, J.
Besson, Comput. Mater. Sci. 50 (2011) 1365.[24] D. Steglich, H.
Wafai, J. Besson, Eng. Fract. Mech. 77 (2010) 3501.[25] Y.E. Ma, Z.
Zhao, B. Liu, W. Li, Mater. Sci. Eng. A-Struct. Mater. Prop.
Microstruct. Process 569 (2013) 41.[26] B. Decreus, A.
Deschamps, F. De Geuser, P. Donnadieu, C. Sigli, M. Weyland,
Acta Mater. 61 (2013) 2207.[27] S.C. Wang, M.J. Starink, Int.
Mater. Rev. 50 (2005) 193.[28] A.K. Vasudvan, R.D. Doherty, Acta
Metall. 35 (1987) 1193.of intragranular and intergranular
microstructure on the strength,the uniform elongation in tension,
and transverse fracture tough-ness as well as related
fractography.
While the strength and uniform elongation are mainly con-trolled
by the state of precipitation within the grains, we havestrong
indications that the transverse fracture toughness is
largelydetermined by the occurrence of intergranular fracture,
which iscontrolled by the presence of incoherent, Cu rich grain
boundaryprecipitates. These precipitates serve as damage nucleation
sitesand the resulting fracture surfaces present a high density of
sub-micrometric dimples. In practice, the choice of a heat
treatmentat the very beginning of the peak strength plateau, where
theintergranular precipitation microstructure is not yet fully
devel-oped, makes possible to avoid a massive appearance of
intergra-nular fracture and thus keep a high toughness value.
Particularcare must yet be taken to ensure a sufciently fast quench
ratefrom the solution treatment so that the grain boundary
micro-structure prior to articial ageing is free of
precipitates.
Acknowledgements
C. Sigli is thanked for fruitful discussions. L. Charpenay
isthanked for helping with the short bar tests. The French
researchagency (ANR) is thanked for nancial support under the
projectALICANTDE. We are grateful to the Canadian Centre for
ElectronMicroscopy, a facility funded by the Canada Foundation
forInnovation and the Ontario Government where the HAADF STEMwork
was carried out.
References
[1] E. Starke Jr., J.T. Staley, Prog. Aerosp. Sci. 32 (1996)
131.[2] F.W. Gayle, F.H. Heubaum, J.R. Pickens, Scr. Metall. Mater.
24 (1990) 79.[3] S. Katsikis, B. Noble, S.J. Harris, Mat. Sci. Eng.
A 485 (2008) 613.[4] T. Warner, Mat. Sci. Forum 519521 (2006)
1271.[5] P. Lequeu, K.P. Smith, A. Danielou, J. Mater. Eng.
Perform. 19 (2010) 841.[6] A. Danielou, J.P. Ronxin, C. Nardin,
J.C. Ehrstrm, Proceedings of the 13th
International Conference on Aluminium Alloys, TMS, Warrendale,
USA,Pittsburgh, PA, 2012, pp. 511516.
[7] J. Boselli, G. Bray, R.J. Rioja, D. Mooy, G. Venema, G.
Feyen, W. Wang,Proceedings of the 13th International Conference on
Aluminium Alloys, TMS,Warrendale, USA, Pittsburgh, PA, 2012, pp.
581586.
[8] J.C. Huang, A.J. Ardell, Mater. Sci. Technol. 3 (1987)
176.[9] J.M. Howe, J. Lee, A.K. Vasudvan, Metall. Trans. A 19A
(1988) 2911.[10] P. Donnadieu, Y. Shao, F. De Geuser, G.A. Botton,
S. Lazar, M. Cheynet, M. de
Boissieu, A. Deschamps, Acta Mater. 59 (2011) 462.[11] C. Dwyer,
M. Weyland, L.Y. Chang, B.C. Muddle, Appl. Phys. Lett. 98
(2011)
201909.[29] S. Lynch, Mater. Sci. Eng. A. 136 (1991) 25.
-
[30] S. Lynch, A. Wilson, R. Byrnes, Mater. Sci. Eng. A. 172
(1993) 79.[31] C.G. Bennet, S.P. Lynch, R.B. Nethercott, M. Kerr,
E.D. Sweet, Mat. Sci. Eng.
(1998) 32.[32] S.P. Lynch, B.C. Muddle, T. Pasang, Acta Mater.
49 (2001) 2863.[33] S.P. Lynch, B.C. Muddle, T. Pasang, Philos.
Mag. A 82 (2002) 3361.[34] A.C. Bregianos, A.G. Crosky, P.R.
Munroe, A.K. Hellier, Int. J. Fract. 161 (2010)
141.[35] T. Pasang, N. Symonds, S. Moutsos, R.J.H. Wanhill, S.P.
Lynch, Eng. Fail. Anal. 22
(2012) 166.[36] N. Eswara Prasad, K.S. Prasad, S.V. Kamat, G.
Malakondaiah, Eng. Fract. Mech.
51 (1995) 87.
[37] K.V. Jata, A.K. Vasudevan, Mater. Sci. Eng. A 241 (1998)
104.[38] A. Deschamps, B. Decreus, F. De Geuser, T. Dorin, M.
Weyland, Acta Mater.
61 (2013) 4010.[39] ASTM-International, Standard Test Method for
Plain-strain (chevron-notch)
Fracture Toughness of Metallic Materials, 2002.[40] D. Dumont,
A. Deschamps, Y. Brechet, Acta Mater. 52 (2004) 2529.[41] B.
Decreus, Ph.D. thesis, Institut Polytechnique de Grenoble, France,
2010.[42] B. Decreus, A. Deschamps, F. De Geuser, C. Sigli,
Advanced Engineering
Materials http://dx.doi.org/10.1002/adem.201300098, in
press.[43] T. Pardoen, D. Dumont, A. Deschamps, Y. Brechet, J.
Mech. Phys. Solids 51
(2003) 637.
B. Decreus et al. / Materials Science & Engineering A 586
(2013) 418427 427
On the role of microstructure in governing fracture behavior of
an aluminumcopperlithium alloyIntroductionMaterial and experimental
methodsTensile testsChevron notch short bar testsEffect of the
variation of intergranular and intragranular
microstructuresDiscussionConclusionsAcknowledgementsReferences