Aleaciones 7000 Al
Aleaciones 7000 Al
Aleaciones 7000 Al
1. Introduction
7000 alloys are used above all in automotive industry and architectural applications. These
materials exhibit medium strength and ductility at room temperature and can be strength
ened by aging treatment. Moreover they are characterized by low quench sensitivity, good
corrosion resistance (due to the absence of Cu addition) and good extrudability (higher than
6061 alloy) [1-5].
Because of their commercial importance, much effort has been spent on investigation of the
precipitation process in AlZnMg alloys [6-10]. The high strength exhibited in the hard
ened state is due to a fine distributions of precipitates, notably of the metastable phase
MgZn2, produced by artificial aging from a supersaturated solid solution. The temperature
of artificial aging influences the kinetics and the sequence of precipitation and if heterogene
ous nucleation of the equilibrium phase appears, a less efficient hardening is obtained. In
this study the response to artificial aging with and without previous solution treatment has
been analyzed in the range of 130C-210C in order to evaluate which effect on hardening is
due to the absence of supersaturation of vacancy rich cluster (VRC) and alloying elements
coming from a solution heat treatment and rapid quenching.
There is strong academic and industrial interest in recrystallization driven by the need to
understand and control this phenomenon in order to optimize properties through the care
ful control of thermomechanical processing schedules [11]. In this paper, the effects of differ
ent heat treatments and Zr content on rate of recrystallization induced by annealing heat
treatment after RT deformation and on further deformation in terms of strain hardening rate
(SHR= d/d), have been analyzed. Recrystallization due to hot deformation by torsion and
tension test at 200C-500C and 10-5s-1-10-3s-1 has been investigated too. During hot working
the Al-Zn-Mg alloys exhibit lower flow stress and higher ductility than Al-Mg alloy (for ex
2013 Leo and Cerri; licensee InTech. This is an open access article distributed under the terms of the
Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits
unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
256
ample 5182 and 5083) [12,13]. Generally dynamic recovery (DRV) is the sole restoration
mechanism in Al alloys [14-17] developing a subgrain structure inside elongated grain. As a
consequence, the flow curves (stress vs strain; vs ) exhibit SH to a steady state regime,
although adiabatic heating may cause a peak and a gradual softening particularly at high
strain rate () and low temperature (T). Ductility is usually high because DRV softened
grains allow accommodation of differential grain boundary (GB) sliding, slowing crack for
mation [18-22]. Solutes, in the form of atmospheres hinder dislocation glide reducing DRV
and ductility and raising the flow stress [21-23]. Moreover fine dispersoids pin dislocations
and reduce DRV [24,25]. Precipitation hardening alloys may present varied behaviours as a
result of changes in precipitate morphology. Growth of precipitates during hot working
leads to good ductility and lower stress as shown for Al-Mg-Si [26,27], Al-Cu-Mg [28,29] and
Al-Zn-Mg-Cu [30,31]. Solution treated alloy can exhibit high peak stress and dislocation
density due to dynamic precipitation (DPN), followed by rapid softening as particles coa
lesce [29-33]. In this paper the microstructure of hot deformed Al-Zn-Mg samples (even
modified with Zr) both by torsion and tension test have been analyzed by SEM and optical
microscopy in order to justify the stress-strain curve shape.
Hot working of many engineering alloys is often accompanied by the formation of internal
cavities [34-38]. The cavitation process depends strongly on alloy composition and micro
structure as well as on the imposed processing condition [3 4 36 37 39]. Particularly large
particles and inclusions, notably on GB, introduce new sources of fissure nucleation lower
ing ductility; solidification segregation and low melting constituents, especially if they
spread along the GB, create severe problems [39]. Such cavitation may lead to premature
failure (i.e. failure at strains lower than those expected based on material properties such as
the strain rate sensitivity index and the strain hardening exponent) or result in a finished
part with degraded mechanical properties. The cavitation process comprises three distinct
stages, which in most cases occur simultaneously, i.e., (i) cavity nucleation, (ii) cavity
growth, and (iii) cavity coalescence. Cavities, which usually nucleate preferentially at GB,
triple points, or second-phase particles, grow by either plasticity- or diffusion-controlled
mechanisms, or a combination of the two [35,37,39]. For a given material, the particular
mechanism varies with the imposed deformation conditions.
2. Experimental procedures
The compositions of the alloys studied in this investigation are reported in Table1.
In order to distinguish easily the two materials with regard to Zr content they have been
designated respectively as 7000 and 7000Zr. The materials were supplied in the form of
DC cast billet of 20 cm in diameter and 40 cm in length. Cylindrical samples with gage
length of 13 mm and 5mm diameter were cut parallel to the longitudinal axis of the billet
for tensile and torsion tests. From the same billet, cube samples of 10 mm edges were cut
for heat treatments. Artificial aging has been carried out at 130C, 160C, 190C and 210C
up to 432h on the as-cast samples and 48h and solutionized ones (2h at 490C). The ef
fects of heat treatments were analyzed by hardness (HRF) and electrical conductivity
curves.
Zn
Mg
Fe
Si
Ti
7000
5.5
1.2
0.07
0.03
0.01
7000Zr
5.6
1.2
0.07
0.03
0.01
Zr
0.16
The microstructure of as-received alloys has been investigated by optical microscope (Ni
kon Epiphot 200) and Scanning Electron Microscope - Focused Ion Beam (SEM FIB) ZEISS
1540. The chemical composition of the matrix and particles was investigated by Energydispersive X-ray spectroscopy (EDS) analysis. For polarized light observation the samples
were ground according to standard methods, electropolished (80ml perchloric acid, 120ml
distilled water, 800ml ethanol, 20V) and anodized (Barkers reagent). The average grain
size has been evaluated on a population of at least 200 grains by using the NIS software
for imaging analysis.
RT tensile tests were performed on as-cast, solutionized (490C-2h) and peak aged
(490C-2h + 160C-24h) samples and SHR plotted versus (-y). One half of each frac
tured sample coming from an RT tensile test was ground parallel to longitudinal axis up
to the middle plane and annealed at 500C for 3h. After 1,5 h of annealing, the sample
was water quenched and the average grain size calculated in a fixed area close to the frac
ture by using the LUCIA G software. Then a second step of annealing at the same temper
ature and time (total 3h) was applied to each sample in order to follow the
recrystallization behaviour.
Hot tensile tests have been performed on as-cast alloys in the range 250C-400C and
10-5to 10-3s-1. The temperature was measured by two independent thermocouples placed
close to the sample. The true stress-true strain curves were calculated from recorded loaddisplacement data according to the usual formula. Hot torsion tests have been performed
in the range 250C-500C and 10-2to 5s-1. The torque and surface strain have been trans
formed into equivalent stress and strain by the traditional means. One half of each frac
tured sample coming from hot deformation tests were ground parallel to longitudinal axis
up to the mid- plane in order to investigate on both recrystallization (RX) and cavitation
phenomena by optical and SEM analysis. Particularly, for cavitation analysis, micrographs
at 20 have been taken along the length of each metallographic section ( up to 4mm by
fracture mid line) and collected into a montage [Fig.15 a,b,c]. The area of cavities inside
the area of metallographic section (4mm by fracture mid line) and the same area of metal
lographic section has been evaluated using NIS software. Cavity area fractions Cs (%)
(area of cavities divided by the area of metallographic section) were determined. Assum
257
258
ing that cavities are randomly distributed inside the specimen, it has been shown that the
area fraction is equal to volume fraction Cv (%) [30].
Figure 1. Optical micrographs of 7000 (a) and 7000Zr (b) alloys (5X) showing that the microstructure of both alloys is
characterized by dendritic microsegregation. Different grain size is evident comparing (a) and (b)
The average values of hardness in the as-cast and solutionized state (490C-2h) are slightly
higher for 7000Zr alloy (Table 2) despite its grain size being higher. In contrast, the as-cast
7000Zr electrical conductivity is lower (22Ms/m vs 23,5Ms/m) for the higher amount of al
loying. As shown in Fig.3 solution heat treatment (490C-2h) reduce microsegregation and
through dissolving hardening particles the hardness is reduced too (Table 2). EDS analysis
did not find Al-Zn-Mg particles in solutionized alloys while some undissolved FeAl3/FeAl6
type particles were found.
B
A
B1
A1
C
Counts
a
10000
7500
5000
2500
0
Zn
Mg
Matrix
Al
Fe
Counts
0,5 1,0 1,5 2,0 2,5 3,0 3,5 4,0 4,5 5,0 5,5 6,0
10000
7500
5000
2500
0
KeV
Al/Fe particles
( point "B1" in Fig.2a)
Counts
0,5 1,0 1,5 2,0 2,5 3,0 3,5 4,0 4,5 5,0 5,5 6,0
10000
7500
5000
2500
0
KeV
0,5 1,0 1,5 2,0 2,5 3,0 3,5 4,0 4,5 5,0 5,5 6,0
KeV
Figure 2:SEM micrograph showing Al/Fe particles (named B, B1) and particles containing Zn and
Mg (named A,A1) (a);EDS spectrum of elements content (b) into the matrix (named C in Fig.2a), and
Figure 2. SEM
micrograph
showing Al/Fe particles (named B, B1) and particles containing Zn and Mg (named A,A1)
B1 and
A1 particles.
(a); EDS spectrum of elements content (b) into the matrix (named C in Fig.2a), and B1 and A1 particles.
The average values of hardness in the as-cast and solutionized state (490C-2h) are slightly higher for
7000Zr alloy (Table 2) despite its grain size being higher. In contrast, the as-cast 7000Zr electrical
The hardness
andiselectrical
conductivity
curves
performed
solution
treated sam
conductivity
lower (22Ms/m
vs 23,5Ms/m)ageing
for the higher
amount
of alloying.on
As shown
in Fig.3
heat on
treatment
reduce at
microsegregation
and shown
through in
dissolving
hardening
ples (T6 solution
type) and
the as(490C-2h)
cast samples
130-210C are
Fig. 4 and
Fig. 5 respec
particles the hardness is reduced too (Table 2). EDS analysis did not find Al-Zn-Mg particles in
tively. The
aging alloys
treatments
(Fig.
4) of solution
(490C-2h) samples lead to 15 and 13
solutionized
while some
undissolved
FeAl3/FeAltreated
6type particles were found.
The hardness of
andhardness
electrical conductivity
ageingfor
curves
performed
on solution
samples
(T6 tempera
point increments
respectively
7000
and 7000Zr
at treated
the two
lowest
type) and on the as cast samples at 130-210C are shown in Fig. 4 and Fig. 5 respectively. The aging
tures of treatments
aging. At
the
higher
temperatures,
the
precipitation
kinetics
are
faster
but
(Fig. 4) of solution treated (490C-2h) samples lead to 15 and 13 point increments of the hard
respectively
and 7000Zr at the
two lowest temperatures
of aging.starts
At the before
higher the peak
ening is hardness
less efficient
dueforto7000
heterogeneous
nucleations
and overaging
temperatures, the precipitation kinetics are faster but the hardening is less efficient due to
is reached.
Moreover
as
the
VRC
are
not
retained
at
high
temperature
of
aging,
a lower den
heterogeneous nucleations and overaging starts before the peak is reached. Moreover as the VRC are
not retained atprecipitates
high temperature
of aging, a lower
density of the
hardening
precipitates
expected.
sity of hardening
is expected.
Comparing
behaviour
of isthe
two alloys, the
Comparing the behaviour of the two alloys, the response to T6 heat treatment is better at the higher
responsetemperature
to T6 heat
treatment
better
at containing
the higher
temperature
(190C
and at220C)
for the
(190C
and 220C)isfor
the alloy
Zr while
it is similar for
both alloys
the
lower
temperature
of
treatment.
This
behaviour
could
be
due
to
Al
Zr
compounds
that
dont
dissolve
3
alloy containing Zr while it is similar for both alloys at the
lower temperature of treatment.
This behaviour could be due to Al3Zr compounds that dont dissolve at high temperature of
treatment and act as nucleation sites for hardening precipitate phase [45-49].The electrical
conductivity increases with temperature of aging and time because of the draining of solute
from the matrix as the precipitation process proceeds. The aging curves of the as cast sam
ples (Fig. 5) do not show an increment of hardness with respect to the starting state. The ab
sence of super saturation of VRC and alloying elements coming from a solution heat
259
260
treatment and rapid quenching, substantially reduces the nucleation of hardening precipi
tates. Even for this heat treatment the response of 7000Zr alloy is better at the higher temper
ature (190C and 210C) compared with the behavior of the alloy without Zr while it is
similar for both alloys at the lower temperature of treatment. The electrical conductivity in
creases with temperature of aging and time but it is always lower that in the case of T6 heat
treatment because the lower supersaturation of the matrix.
HRF
HRF
As received
Solutionized
(490C-2h)
7000
93
86
7000Zr
95
88
b
Figure 3. 7000 (a) and 7000Zr(b) alloys after solution heat treatment (490C-2h). The dissolution of interdendritic seg
regation is evident compared to Fig.1
RT tensile tests on as-cast and solutionized samples indicate that solution heat treatment
lead to low peak stress and high ductility (Fig.6) due to dissolution of both brittle phases
and hardening particles. Moreover, the RT ductility is always higher for 7000Zr alloy. In
term of SHR (Fig.7), it is always higher for the alloy in the as-cast state compared to solu
tionized because of large uncuttable particles cause Orowan hardening. Moreover 7000Zr al
loy exhibits higher SHR probably due to the interaction of Al3Zr particles with dislocations.
One half of each fractured samples was ground parallel to the longitudinal axis down to the
mid plane and annealed at 500C. After 1,5 h of annealing, the samples were water
quenched and analyzed by optical microscopy for checking any recrystallization phenom
ena. Then, a second step of annealing of 1,5h was applied to each sample (total 3h). Fig. 8 an
Fig.9 illustrate the anodized microstructure of respectively 7000 and 7000Zr specimen as ten
sile tested (first line) and 3 hours annealed (second line). The first column presents pictures
from the as cast sample, the second from the solution treated. Annealing treatment lead to
recrystallization rate that is faster on as- cast alloys compared to the solutionized. This result
can be clarified by considering that dissolved atoms and fine precipitates formed in the ma
trix limit the movement of dislocation during annealing and delay the nucleation and
growth of new grains. The as cast alloy exhibits both the highest strain hardening rate and
low dissolved atoms; both these aspects lead to a shorter recrystallization time. However as
shown in Fig.9 the recrystallization of 7000Zr alloy is incomplete and not homogeneous
even after three hours of annealing because of additions of Zr.
Figure 4. Hardness and electrical conductivity of 7000 and 7000Zr alloys during aging at 130C, 160C, 190C and
210C after solution treatment at 490C-2h (initial value at 0,1min).
The peak stress p decreases with increasing T at constant ; moreover, it decreases with de
creasing at a fixed temperature [Fig. 10]. For each fixed temperature T, the ductility decreas
es as increases for 7000 alloy deformed by torsion and tension test while rises with for 7000
Zr tensile samples. This behavior being more evident with increasing temperature is common
in creep [51,52]. In contrast for hot working as decreases, ductilities increases since the im
proved DRV mitigates stress concentration and nucleation of voids (usually at triple junc
261
262
Figure 5. Hardness and electrical conductivity of as-cast 7000 and 7000Zr alloys during aging at 130C, 160C, 190C
and 210C (initial value at 0,1min).
Figure 6. Room temperature tensile curves for 7000 (a) and 7000Zr samples (b) in the following conditions: as cast
and solution treated 490C-2h
263
264
Figure 7. Strain hardening rate curves for the RT tensile curves of 7000 (a) and 7000Zr (b) alloy in the as-cast and solu
tionized state.
Figure 8. Optical micrographs of 7000 tensioned specimens before (a,b) and after 3h of annealing at 500C (c,d)
where (a,c) are as cast, (b,d) are solution treated
Figure 9. Optical micrographs of 7000Zr tensioned specimens before (a,b) and after 3h of annealing at 500C (c,d)
where (a,c) are as-cast, (b,d) are solution treated.
1,0
160
140
0,8
Failure strainf
Stress (MPa)
120
100
0,6
80
-3 -1
10 s
-4 -1
10 s
-5 -1
10 s
40
20
0
250
300
350
Temperature (C)
400
0,0
250
Failure strainf
Stress (MPa)
150
10-2s-1
10-1s-1
1s-1
100
50
0
150
200
250
500
10-2s-1
10-1s-1
1s-1
5s-1
0
150
550
200
250
500
550
1,0
120
Stress (MPa)
400
5s-1
140
0,8
Failure strainf
100
0,6
80
60
300
350
Temperature [C]
10
200
20
10-5 s-1
0,2
250
40
10-3 s-1
10-4 s-1
0,4
60
160
0,4
10-3s-1
10-4s-1
10-5s-1
250
10-3 s-1
10-4 s-1
10-5 s-1
0,2
0,0
300
350
Temperature(C)
400
250
300
350
Temperature(C)
400
Figure 10. Peak stress variations (a,b,c) and failure stress (d,e,f) as a function of T for the as-cast 7000 alloy hot de
formed by tension (a,d) and by torsion (b,e) and for 7000 Zr alloy hot deformed by tension (c,f).
Figure 11. Precipitated particles in hot torsioned samples at 300C-0,1s-1(a), 300C-0,01s-1 (b) and 400C-0,1s-1(c): at
fixed T (a,b) the number of particles increases with strain rate due to enhanced precipitation on dislocations, while as
T increases the size of particles increase due to overaging (a,c).
265
266
Figure 12. Optical micrographs (50X) of torsioned samples on longitudinal planes close to fracture surface after Keller
etchant showing that only the samples deformed at the highest T exhibit SRX.
Figure 13. Optical micrographs (20X) of torsioned samples on longitudinal planes close to fracture surface after anod
izing showing that the microstructure at 400C and is characterized mainly by subgrains. Subgrains have been ob
served in some grains at 300C and 0.1 s-1-0.01s-1
Figure 14. SRX close to fracture surface of hot deformed tensile test 7000 alloy at 400C-10-4s-1 with Zr (a) and with
out (b)
Figure 15. Optical micrographs of cavitation phenomena in as-cast 7000 (a), 7000Zr (b), solutionized 7000 alloys (c)
deformed by tensile test at 350C 10-4 s-1. SEM micrographs of 7000 as-cast alloy (d) showing that cavities originate at
brittle particles.
When cavity growth is controlled by plastic deformation, the simplest model for cavity
growth assumes the following form [38]:
Cv = C0 exp (h )
(1)
267
268
where C0 is the initial volume fraction of cavities and is the fracture strain. Following Lee
and Huang [40], who based their analysis on the Stowell et al. relationship [41], one can ex
press the cavity growth exponent as follows [42]:
( )
( )
2 2-m
3 m +1
h =
sinh
m
3 2+m
2
(2)
Figure 16. Cavitation (Cs%) versus log strain rate showing that the area of fissures increases with T and, at the highest
T, with decreasing strain rate
Figure 17. Cavitation (Cs%) versus distance from fracture at 250C (a) and 400C (b) showing that at the lowest T the
area of fissures decrease as distance from fracture increases while at 400C cavities are randomly distributed along the
longitudinal area of the samples.
When the fracture strain is substituted into Eqn.1, an obvious approximation, since the frac
ture strain is affected by necking, the plot presented (with C0 arbitrarily assumed to be
0.03%) in Fig. 18 are obtained. Although this calculation is based on very rough assump
tions, the model of cavity growth controlled by plastic strain describes very well the ob
served cavitation trend. It can be thus reasonably concluded that, in the investigated range
of experimental conditions, the cavity growth is mainly driven by plastic straining. GBS sub
stantially contributes to enhance the fraction of cavities at the highest T and justifies the ob
served difference between the calculated value and experimental one (Fig.18).
Figure 18. Comparison of measured Cs% and calculated Cv% versus log strain rate at 250C and 400C
4. Conclusions
The main conclusions are summarized in the following :
In the as received state the grain size of 7000 Zr alloy is larger than that of 7000 alloy due
to reaction between Zr and Ti that reduce the nucleation power of both elements and the
microstructure of both alloys is characterized by dendritic microsegregation. The solution
heat treatment at 490C-2h leads to a strong reduction of segregation and to a complete
dissolution of hardening Al-Zn-Mg particles as shown by EDS analysis. As consequence
the hardness decreases. For the larger grain size and higher amount of alloying the electri
cal conductivity of 7000Zr is always lower than that of 7000 alloy.
The response to heat treatment for both the as cast and solutionized samples is better at
the higher temperature (190C and 210C) for the alloy containing Zr, while it is similar
for both alloys at the lower temperature of treatment. This behaviour could be due to
Al3Zr compounds that dont solutionize at high temperature of treatment and moreover,
harden the alloy and can act as nucleation sites for hardening precipitate phase. Aging
treatment of the as-cast alloys for the range of imposed times is ineffective in terms of in
creasing hardness.
269
270
Concerning RT tensile tests, the SHR is higher for the as cast sample comparing to the sol
utionized. As consequence the recrystallization rate of as cast sample is faster too. Due to
Zr effect it is incomplete and not homogeneous in the Zr modified alloy.
During hot tensile test the as cast alloys exhibit high flow stress at low temperature due to
reduced DRV. At higher T, both improved DRV and overaging of particles lead to re
duced peak stress. The phenomena is much more evident during torsion because of T in
volved. For each fixed temperature T, the ductility decreases as increases for 7000 alloy
deformed by torsion and tension while rises with for 7000 Zr tensile samples.
The microstructure of both torsion and tensile samples hot deformed at the highest tem
peratures exhibit some SRX, more evident in the alloy without Zr.
Cavitation in 7000 hot tensioned samples increases with T and with decreasing strain rate.
At the lowest T, 7000 as cast alloy cavitation decreases as distance from fracture surface
increases while at the highest T cavities are randomly distributed along longitudinal sur
face of samples, suggesting a more active role of GBS on cavitation nucleation and
growth. Theoretical calculation have shown that cavity growth is mainly driven by plastic
straining but GBS substantially contributes to enhance the fraction of cavities and justify
the observed difference between the calculated and experimental cavitation values at the
highest T and lowest strain rate. Cavitation is reduced if the alloy is solutionized before
deformation.
Author details
P. Leo* and E. Cerri
Universit del Salento, via per Arnesano, Lecce, Italy
References
[1] H.J. McQueen, P. Leo and E. Cerri; Aluminum Alloys: Fabrication, Characterization
and Applications, : Weimin Yin, Subodh Das, Zhengdong Long Eds: TMS USA 2009
pp.37-44
[2] J.E. Hatch,ed. Aluminum, Properties, Physical Metallurgy, ASM, Metals Park, Oh,
1984
[3] H.J.McQueen, O.C. Celliers, Canadian Metal. Quart. 36 (1997), 73-86
[4] K. Laue, H.Stenger, Extrusion: Process Machinery, Tooling, American Society for
Metals, Metals Park, OH (1981), 1-62; 124-152
[5] Y.Baba, H.Yoshido, Proc. 2nd Intnl.Al. Exrusion Tech. Sem., Aluminium Association,
Washington, (1997) Vol. 1, 301-306
[6] T. Engdahl, V. Hansen, P.J. Warren, K. Stiller , Mater. Sci. Eng. A 327, (2002), 59
[7] K. Stiller, P.J.Warren, V. Hansen, J. Angenete, J. Gjonnes, Mater. Sci. Eng. A270,
(1999),55
[8] S.K. Maloney, K. Hono, I.J. Polmear, S.P. Ringer, Scripta Mater. 41, 10, (1999), 1031
[9] H.Z.Li, V. Hansen, J. Jonnes, L. R. Wallenberg, Acta Mater. 47, 9. (1999), 2651
[10] J.C. Werwnskiold, A. Deschemps, Y. Brchet, Mater.Sci. Eng. A293, (2000), 267
[11] M.J. Jones , F.J. Humphreys, Acta Mater. 51 (2003) 21492159
[12] H.J. McQueen, N.Owen, Materials in Automotive Industry E.Essadigi et al., Eds.,
Mat.Soc.Cim, Montreal (2001), pp.189-205
[13] W. Blum, H.J. McQueen, Aluminum Alloys, Physical Mechanical Properties, ICAA5
Mater. Sci. Forum 217-222,(1996), pp. 31-42
[14] H.J. McQueen, W. Blum. Aluminium Alloys, Phisical and Mechanical Properties
ICAA6,T.Sato, ed Japan Inst. Metals (1998), pp. 99-112
[15] R.D.Doherty et al. Mat. Sci. Eng. A 238, 2 (1997) pp.219-274
[16] M. El Mehtedi, S. Spigarelli, E Evangelista, La metallurgia italiana, (September 2007)
pp. 21-28
[17] H.J. McQueen, in Advanced Materials for 21st Century, J.R. Weertman Symp, Y.P.
Chung Ed, TMS-AIME, Warrendale, PA, (1999), pp. 159-168
[18] S. Fulop , K.C. Cadien, M.J. Luton, H.J. McQueen, J. Test Eval. 5 (1977) pp.419-426
[19] N.D. Ryan, H.J. McQueen, J. Mech. Working Tech. 12 (1986) pp. 279-296
[20] C.M. Sellars, W.J. McG. Tegart, Int. Metall. Rev. 17 (1972) pp. 1-24
[21] H.J. McQueen in Hot deformation of Aluminum alloys, T.G. Langdon, H.D.Mer
chant, J.G.Morris and M.A.Zaidi Eds, Warrendale, PA, 1990, pp. 31-54.
[22] H.J. McQueen and J.J. Jonas in C.Q. Chen et al (eds.), Aluminum alloys 90 (Icaa2) Bei
jing, 1990, pp.727-742
[23] H.J. McQueen and K.Conrod in Microstructural control in Aluminum alloys, E.H.
Chia and H.J. McQueen (Eds), TMS-AIME, Warrendale, PA, 1986, pp.197-219
[24] H.J. McQueen, E. Evangelista, J. Bowles and G. Crawford, Met Sci. 18, (1984), p.
395-402
[25] E. Evangelista, H.J. McQueen and E. Bonetti in Deformation of Multi-phase and par
ticles containing Materials, J.B. Bildensoren et al (eds), Riso National Laboratory,
Roskilde, 1983, pp.243-250
[26] H.J. McQueen in Hot deformation of Aluminum alloys, T.G. Langdon, H.D.Mer
chant, J.G.Morris and M.A.Zaidi (eds), Warrendale, PA, 1990, pp. 105-120
271
272
[45] P. Leo, E. Cerri, H.J. McQueen and S. Chiozzi, Mater. Sci. Forum 604-605; 2009, pp.
67-76
[46] L.K. Berg, G. Gjonnes,V. Hansen et al. Acta Mater, 49, (2001),3493
[47] H. Loffer, I. Kovacs, J. Lendvai, J. Mater. Sci. 18, (1983), 2215
[48] Z. Katz, N.Ryum, Scripta Met. 15, (1981), 265
[49] J.D.Verhoeven, Fundamental of Physical Metallurgy J. Wiley e Sons, NY (1975),
Chapter 6,
[50] P.Leo, E. Cerri, H.McQueen and A.Taurino, ICA11, in Aluminium Alloys Their
Physical and Mechanical Properties J. Hirsch, B. Skrotzki and G. Gottstein eds,Wi
ley-VCH (DE) 2008, pp. 1868-1874.
[51] M.E. Kassner, M.T. Perez-Prado, Fondamental of creep in metals and alloys, Elsivier
2004, 1-15
[52] R.W. Evans, B. Wilshire, Introduction to creep, Oakdale Printing Company Ltd, 1993,
16-22
273