Materials Science & Engineering A 615 (2014) 331–339

Contents lists available at ScienceDirect

Materials Science & Engineering A

journal homepage: www.elsevier.com/locate/msea

High temperature processing of brass: Constitutive analysis

of hot working of Cu–Zn alloys
S. Spigarelli n, M. El Mehtedi, M. Cabibbo, F. Gabrielli, D. Ciccarelli
DIISM, Università Politecnica delle Marche, via Brecce Bianche, 60131 Ancona, Italy

art ic l e i nf o a b s t r a c t

Article history: The hot workability of CW602N brass, containing 36.5% Zn and 2% Pb, was investigated in the present study
Received 20 June 2014 by hot torsion tests carried out between 550 and 700 1C. A detailed analysis of the literature data was
Received in revised form preliminarily aimed at developing a constitutive model able to describe the hot deformation of single-phase
25 July 2014
α-brasses. The model, based on the Garofalo sinh equation, gave excellent descriptions of the mechanical
Accepted 26 July 2014
Available online 1 August 2014
response of copper alloys containing Zn amounts ranging from 0% to 30%, but overestimated the strength of
the CW602N at the highest temperature. The model was then combined with the law of mixtures, to take
Keywords: into account the presence of the soft β-phase, that appears when the Zn content increases above 35%, as was
Non-ferrous alloys indeed observed in the CW602N. Both the flow curves and the peak flow stress dependence on strain rate
Mechanical-characterization
were successfully modeled by assuming that the microstructure was composed of hard α-grains and soft
Light microscopy
β-grains, containing Zn amounts close to 36–37% and 42–44% respectively.
Strain measurements
Bulk deformation & 2014 Elsevier B.V. All rights reserved.
Plasticity

1. Introduction as regards the fitness of tensile creep experiments in order to

attain a true stationary microstructural condition [5]. Since fewer
A knowledge of the complex interaction between several recent studies [6–8] have considered the hot workability of brass,
concurring mechanisms that intervene during hot working pro- the aim of the present study was to investigate the hot working
cesses is a fundamental prerequisite for the correct design of the behavior of the CW602N (Cu–36.5% Zn–2% Pb) alloy, and to obtain
production cycle and the optimization of product properties [1]. a self-consistent constitutive model suitable for describing the
The recent development of Finite Element Modeling (FEM), a material behavior of brasses of different compositions.
method that is now universally used to simulate metalworking
processes, puts an increased emphasis on the accuracy that should
characterize the input data, i.e. the physical and mechanical 2. Experimental procedure
properties of the considered material. Even when referring to
thoroughly investigated and “traditional” materials, such as cop- The alloy considered in the present study was CW602N brass
per and Cu–Zn brasses, the amount of recent, i.e. easily accessible, (Cu–36.5% Zn–2% Pb–As, composition in wt%, corresponding to a
data in the form of reliable mechanical datasets, could be insuffi- Cu/Zn ratio close to 1.68; this ratio, translated into a binary alloy
cient to properly simulate simple operations such as forging. In composition, corresponds to Cu–37.2% Zn). Specimens for torsion
fact, the effects of strain, strain rate and temperature in copper tests, 10 mm in diameter with a gage length of 15 mm, were
have received considerable attention, as attested by the early work strained by a computer-controlled hot torsion machine at 550,
of Bonesteel and Sherby [2], who analyzed the effect of chemical 600, 650 and 700 1C. The samples were heated by a high frequency
composition on the creep response of alpha brasses, not to induction coil at 1 1C/s from room temperature to the testing
mention the comprehensive studies on pure Cu and Cu–30% Zn temperature and maintained at this temperature for 300 s. Sub-
carried out by Raj and Langon [3] and Raj [4] respectively. All these sequently the samples were strained up to rupture. Ruptured
studies, and the previous ones reviewed in [3], explicitly consid- samples were quenched using a water-jet.
ered hot deformation under constant stress/constant load, i.e. The torque M and the number of revolutions N were recorded
creep. Nevertheless, this subject is still controversial, in particular and converted to von Mises equivalent stress σ and equivalent
strain ε at the surface
pffiffiffi
3M
ð3 þ n0 þ m0 Þ
n
Corresponding author. Tel.: þ 39 0712204746; fax: þ39 0712204770. σ¼ ð1aÞ
E-mail address: [email protected] (S. Spigarelli). 2πR3

http://dx.doi.org/10.1016/j.msea.2014.07.091
0921-5093/& 2014 Elsevier B.V. All rights reserved.
332 S. Spigarelli et al. / Materials Science & Engineering A 615 (2014) 331–339

2πNR
ε ¼ pffiffiffi ð1bÞ
3L
where R and L are respectively the radius and the length of the
gage, m0 ¼ ð∂ log M=∂ log NÞ_ is determined at constant strain, and
0
n ¼ ð∂ log M=∂ log NÞ at constant strain rate. The surface equiva-
lent strain-rates were 10  2, 10  1, 1 and 10 s  1. For the peak stress
condition, clearly n0 ¼ 0, while m0 was estimated by plotting the
torque as a function of N_ at a strain of 1, i.e. close to the peak flow
stress.

3. Results

Fig. 1 shows the microstructure of the alloy in the as-received

condition. Although, according to the Cu–Zn phase diagram (Fig. 2)
[9], the alloy should exhibit a single-phase microstructure at room
temperature, Fig. 1 clearly shows the presence of 3% in volume of
the β or β0 phase, presumably the effect of a non-equilibrium
cooling after the production of the billet.
Fig. 3 shows representative equivalent stress vs equivalent strain
curves. In general, the flow curves, after an initial work-hardening
stage, present a peak followed by a moderate softening. Fig. 4 shows
the variation in the peak flow stress as a function of the experimental
strain rate. For the purpose of comparison, the figure also presents
the peak stresses obtained in compression by testing a single-phase
α-brass (Cu–30% Zn) [6], a bi-phase α/β (Cu–40% Zn–3% Pb) brass [7]
and a single phase β-alloy (Cu–44% Zn) [7]. The analysis of Fig. 4

Fig. 3. Representative equivalent stress vs equivalent strain curves.

clearly shows that an increase in the Zn content causes an incre-

Fig. 1. Microstructure of the as received alloy (optical microscopy). mental softening of the alloy.
Fig. 5 shows the microstructure of the alloy after hot-deforma-
tion; in all the investigated samples, the microstructure consists of
refined equiaxed grains. Copper, a low stacking fault energy fcc
metal, exhibits low degrees of dynamic recovery (DRV), and the
substructure formation is sluggish, with smaller and more ragged
subgrains and cells, in comparison with high stacking fault energy
metals such as Al. As a result of the low level of DRV, discountin-
uous dynamic recrystallization (dDRX), i.e. the transformation of
some subgrain boundaries into high-angle boundaries which are
able to migrate, takes place if the temperature is sufficiently
high [1]. The addition of Zn causes a further important decrease
in stacking fault energy (γSF), from 78 (pure copper) down
to 14 mJ/m2 (Cu–30% Zn) [10,11]. Thus, as attested by Fig. 5, hot
working results in a fully equiaxed microstructure of dynamically
recrystallized grains in all the investigated conditions. If on
one hand the presence of large amounts of solute in Cu–30%
Zn facilitates DRX by reducing the efficiency of DRV by lowering
γSF, on the other it also reduces grain boundary mobility, retarding
the progress of dynamic recrystallization. This is the reason for the
Fig. 2. Copper–zinc phase diagram [9]. observed increase with Zn content of the critical strain for the
 S. Spigarelli et al. / Materials Science & Engineering A 615 (2014) 331–339 333

Fig. 4. Peak flow stress as a function of the strain rate for the alloy investigated in the present study, and for Cu–30% Zn [6], Cu–40% Zn–3% Pb [7] and Cu–44% Zn [7].

onset of DRX, of the peak strain (εp) and of the strain-hardening A final aspect of the microstructure which must be considered
rate [1], thus explaining the high values of εp which are well is the possible effect of Pb on hot workability; this issue has
evident in Fig. 3 (0.8–1). already been investigated by Vaandrager and Pharr [12], who
In addition, non-negligible amounts of β or β0 phase can be demonstrated that Pb does not play any significant role during hot
observed in Fig. 5. This clearly indicates that in all the investigated deformation, thus suggesting neglecting this feature in the sub-
testing conditions, during hot working, the microstructure was sequent analysis.
composed of a matrix of a harder fcc α-phase, and by low fractions
of a softer bcc β-phase. The higher amount of the latter phase
observed in the samples tested above 600 1C is in substantial 4. Discussion
agreement with the phase diagram, which predicts, for a Cu–37.2%
Zn binary alloy, 10% and 28% of β at 650 and 700 1C respectively. 4.1. A universal equation for creep and hot working
The substantial presence of this phase in the hot-deformed
samples is also attested by the SEM micrographs presented in The previous section showed that the investigated alloy is
Fig. 6. characterized by a bi-phase microstructure, with a dominant
334 S. Spigarelli et al. / Materials Science & Engineering A 615 (2014) 331–339

Fig. 6. Microstructure of the sample torsioned at 650 1C and 0.01 s  1 (scanning

electron microscopy).

α-brasses. The present constitutive analysis represents an attempt

to obtain a model able to describe the hot-working response of
single phase α-brasses. The model was based on the conventional
Garofalo equation, rewritten in the following form [13,14]:

ε_ ¼ AD½ sinhðα σ=GÞn ð2Þ

where ε_ and σ are the testing strain rate and the peak flow stress
under constant strain rate, or the minimum creep rate and the
stress in creep. The other parameters which appear in Eq. (2) were
the stress exponent n, the diffusivity D, the shear modulus G, and
two material constants (A and α). Eq. (2) reduces to

ε_ ¼ AD αn ðσ=GÞn ð3Þ

for ασ/Go0.6, i.e. in typical creep conditions. The stress exponent

which appears in Eq. (3) has been estimated to range between
3.4 and 5.0 for brass [4,15]. Bonesteel and Sherby [2] did not observe
any significant variations in the magnitude of the stress exponent
with Zn-content, since they estimated n¼5 for Zn ranging from 10%
to 30%. Based on these premises, it can be reasonably supposed that
n ranges between 4 and 5 also in Eq. (2) are in excellent agreement
with the findings of Pernis et al. [8] who found n¼ 4.3 for a 30% Zn
brass. A value of the stress exponent of 4.5 for both Eqs. (2) and (3)
can thus be safely assumed.
Fig. 5. Microstructure of the sample torsioned at 0.01 s  1: (a) 650 1C; (b) 700 1C The identification of the magnitude of the α parameter
(optical microscopy).
deserves additional comments. In general terms, the increase in
the content of elements in solid solution leads to significant
presence of the harder α-phase, and increasing volume fractions of decreases in the creep rate even in the climb-controlled creep
the softer β-phase as the temperature rises above 600 1C. regime [16]. This effect has been successfully described in single-
The predominance of the α-phase suggests that in principle the phase Mg–Al and Mg–Zn alloys by supposing that α decreases with
alloy could have a behavior similar to that of single-phase increasing Al-content [14,17,18]. Following the same approach, in
 S. Spigarelli et al. / Materials Science & Engineering A 615 (2014) 331–339 335

the case of the present study, α was again assumed to be a function

of the amount of Zn in solid solution.
While the quantification of the stress exponent is a relatively
simple procedure, the situation is far more complex in the case of
diffusivity, which can be written as
D ¼ D0 expð  Q =RTÞ ð4Þ

where D0 is the frequency factor (a constant), and Q is the activation

energy for the diffusional mechanism that controls deformation. For
pure metals and class M alloys [3,19], n¼4–5, and Q is equivalent to
the activation energy for self-diffusion of the considered metal. Thus,
in the case of pure copper, the activation energy for creep should
correspond to the activation energy for self-diffusion for temperatures
above 723 K (450 1C). This value is substantially higher than the
activation energy for creep estimated in [2] (161, 159 and 130 kJ mol  1
for Zn¼10%, 20% and 30% respectively, at temperatures ranging
between 548 and 700 1C), but closer to the value calculated by tensile
creep [4] and torsion testing [8] for a 30% Zn brass (180 kJ mol  1).
An increase in Zn content to 36% [15] led to a further decrease in the Fig. 7. Activation energy QZn and frequency factor Dn0Zn for diffusion of Zn in brass
activation energy for creep (118–153 kJ mol  1), although this set [4,29,31–34].
of experimental data was obtained at relatively low temperatures,
250–550 1C (523–823 K), i.e. in a regime where alternative mechan- relevant frequency factors. In this regard, the above-mentioned diffu-
isms, such as obstacle-controlled glide or pipe diffusion, could sion studies demonstrated that
result in activation energies close to 135 kJ mol  1 even in pure
copper [3,20]. i. in brass, the diffusivity of Zn is always higher than that of Cu;
It is therefore evident that a significant decrease in the ii. an increase in the Zn content leads to an increase in the
activation energy for high temperature deformation is observed diffusivity of both Cu and Zn in brass, i.e. to a decrease in the
when the Zn-content of the alloy increases and/or when the activation energy for diffusion of the relevant species.
testing temperature falls below 450 1C. On the other hand, even
in the case of high-temperature testing, in order to obtain a On the basis of these findings, it could be concluded that the
reliable constitutive equation valid for single-phase α-brasses of quantification of the effect of Zn on diffusivity is a straightforward
different chemical compositions, a detailed analysis of the varia- task, due to the large amounts of experimental data available in
tion in diffusivity with Zn content is a necessary step. the literature. In fact, a comparison of the available results in a
single graph, presented in Fig. 7, shows that there is a considerable
scatter of the literature data. In broad terms, in the case of
4.2. Diffusivity in Cu–Zn brass
diffusion of Zn in brass, two self-consistent sets of literature data
can be identified, namely
As demonstrated above, a major difficulty in defining a
universal constitutive equation for alloys containing different
i. the data reported in [4,29,31,33] (set I), which collapse on two
amounts of Zn is that this element significantly affects diffusivity.
broad scatter-bands, where QZn and Dn0Zn decrease and remain
In particular, the complex diffusion coefficient for dislocation
roughly constant with increasing Zn content, respectively;
climb (Dc), the mechanism that is thought to control deformation
ii. the data illustrated in [32,34] (set II); in this case, both QZn and
in pure metals and class M alloys, can be expressed by the
Dn0Zn strongly decrease with an increase in Zn content.
following equation [21,22]:

Dc ¼ D0c expðQ c =RTÞ ¼ ðN Cu DnCu þ N Zn DnZn Þ=0:78 ð5Þ The calculation of the effect of Zn content on the diffusivity of
n n
this element in brass was based on set I, since it included data
where DZn and DCu are the diffusivities for tracer diffusion of Zn from four different studies. Fig. 8 again plots these data, together
and Cu in brass, respectively. Thus, in order to properly estimate with two curves which were thought to describe the observed
the effects of the chemical composition on diffusion-controlled trends sufficiently well; thus, while QZn monotonically decreases
deformation mechanisms, the influence of the Zn content in brass with increasing Zn content in brass, Dn0Zn increases, reaching an
on diffusivity must be closely evaluated. almost constant value for Zn 425%.
Following the initially controversial findings of Kirkendall in A similar procedure was followed to quantify the effect of Zn
1939 [23–26], many different studies have analyzed the diffusion of content on the diffusion of Cu in brass [4,29,31] (Eq. (6); Fig. 9).
Zn (and Cu) in α-brass, in order to confirm or refute the so-called Once again, it can easily be observed that a variation in the Zn
Kirkendall effect (see, for example, [27–32]). One of the most content causes a marked decrease in QCu and a moderate increase
significant implications of Kerkendall's work is that diffusion and in Dn0Cu .
related phenomena (such as creep and hot deformation) involve an The combination of the curves presented in Figs. 8 and 9, with
unequal interchange of solute and solvent atoms during diffusion. Eqs. (5)–(7), allows the complex diffusion coefficient for climb,
The tracer diffusivities of Cu and Zn in brass, in particular, can be presented in Fig. 10, to be estimated. Fig. 11 thus reports the
expressed by the usual Arrhenius form, i.e. variation in Qc and D0c, as given by the interpolation of the data
DnCu ¼ Dn0Cu expð  Q Cu =RTÞ ð6Þ presented in Fig. 10 using Eq. (5). The graph also shows the values
of the activation energy and the frequency factor for self-diffusion
in pure Cu [31]. The figure shows that, due to the combined effect
DnZn ¼ Dn0Zn expð  Q Zn =RTÞ ð7Þ
of the variations in both DnZn and DnCu , Qc monotonically decreases
where QCu and QZn are the activation energies for the diffusion of with increasing Zn content, while D0c presents a sigmoidal
copper and zinc in brass, respectively, and Dn0Cu and Dn0Zn are the behavior, which again increases with the amount of Zn.
336 S. Spigarelli et al. / Materials Science & Engineering A 615 (2014) 331–339

Fig. 8. Activation energy QZn and frequency factor Dn0Zn for diffusion of Zn in brass Fig. 11. Variation of Qc and D0c, as given by interpolation of the data presented in
data of set I [4,29,31,33]. Fig. 10 by Eq. (5).

Fig. 9. Activation energy QCu and frequency factor Dn0Cu for diffusion of Cu in brass Fig. 12. Variation of α with Zn-content.
[4,29,31].

properly calculated, requires the determination of only two para-

meters, i.e. A and α. As mentioned above, in this study it was
assumed that A is a constant, while α varies with the chemical
composition. The best fitting of the data obtained by Padmavard-
hani and Prasad [6] who tested a number of different alloys under
constant strain rate compression (CSR/Co), by Raj [4] who tested a
Cu–30% Zn brass under constant tensile stress (CS/Te), and by Raj
and Langdon [20] who analyzed the tensile creep (CS/Te) behavior
of pure copper, led to the variation in α with Zn content presented
in Fig. 12. Fig. 13 shows the variation in the diffusivity-normalized
strain rate as a function of sinhðασ=GÞ for different alloys, tested in
compression, tension and torsion, under constant strain rate or
constant stress. The plots show that all the data collapse on a
straight line of slope n ¼4.5. This analysis demonstrates that
Eq. (2) provides a reliable description of the material behavior in
the considered temperature range, i.e. above 500 1C. Even in the
Fig. 10. Complex diffusion coefficient for climb as a function of temperature for case of CW602N, the description given by the model is quite good
alloys with increasing Zn-contents. (see also Fig. 14) although an important deviation is observed at
700 1C and 550 1C. As mentioned above, this alloy contains a non-
4.3. Application of the model to single-phase Cu–Zn α-brasses negligible amount of soft β-phase, which should play a significant
role in high-temperature deformation. This conclusion suggested
The description of the experimental data available by Eq. (2), refining the model taking into account the bi-phase nature of the
once the stress exponent is fixed (n¼ 4.5) and the diffusivity Dc is microstructure in alloys containing Zn amounts in excess of 36%.
 S. Spigarelli et al. / Materials Science & Engineering A 615 (2014) 331–339 337

Fig. 13. Variation of the diffusivity-normalized strain rate as a function of sinhðασ=GÞ for different alloys, tested in compression (CSR/Co) [6], tension (CS/Te) [4,20] and
torsion (CSR/To) [this issue,8], under constant strain rate (CSR) or constant stress (CS). For reference purpose, the dataset obtained by testing under constant stress the Cu–
30% Zn have been reported in each plot.

4.4. Constitutive model for α/β Cu–Zn brasses the form

ε_ ¼ ð1  f s Þ_εh þ f s ε_ s ð10Þ
The microstructure of binary Cu–Zn α/β brass is a typical
example of a composite structure composed of a hard phase (the
α-grains) and a softer constituent (the β-grains) [35,36]. Therefore, Eqs. (9) and (10) could then be used to model the peak flow
the rule of mixtures could be effectively used to model the stress dependence on testing strain rate in binary α/β brass; in
evolution of strain and stress, in the following form [37]: particular, in order to simplify the calculation, it was preliminarily
assumed that no dramatic difference in deformation was experi-
ε ¼ ð1  f s Þεh þ f s εs ð8Þ enced by the two phases. According to the diagram shown in
Fig. 2, the microstructure of binary brasses is composed of α and
and
β constituents which, between 550 and 700 1C, approximately
σ ¼ ð1 f s Þσ h þ f s σ s ð9Þ contain 35.5–37.5% and 41.5–44% Zn respectively. Thus, Eq. (9) can
be directly used to estimate the flow stress, if the constitutive
where the suffixes s and h denote the soft and the hard regions models for both phases are available. While this requirement is
respectively, and fs is the volume fraction of the soft β-phase. At a more easily fulfilled, at least in the case of the peak flow stress of
given temperature, provided that the volume fraction of the soft the α-phase, by the availability of the above-described constitutive
constituent does not change with time, Eq. (8) can be rewritten in model and of numerous sets of literature data, the mechanical
338 S. Spigarelli et al. / Materials Science & Engineering A 615 (2014) 331–339

Fig. 14. Description of the peak flow stress dependence on strain rate for the
CW602N alloy; the broken curves were obtained by considering the material a
single-phase alloy with 37.2% Zn (see Section 4.3). The solid curves were calculated
by the composite model, assuming that the alloy is composed by α and β phases
containing 36% and 44% Zn respectively (Section 4.4).

response of the β-constituent should be properly quantified. To

that end, the data obtained by Padmavardhani and Prasad [7] on a
Cu–44% Zn are extremely useful.
The accuracy of the composite model in describing the beha-
vior of these alloys, can then be tested by considering the
Cu–40% Zn–3% Pb alloy [6]. This material, once the Cu/Zn ratio is
recalculated to take into account the presence of Pb according to
the phase diagram presented in Fig. 2, should contain 55%, 70%,
78% and 95% of β-phases at 550, 600, 650 and 700 1C respectively.
Thus, a rough estimate of the flow stress variation with strain can
be simply obtained by substituting, in Eq. (9), the experimental
stress values at a given strain, obtained in torsion for Cu–36.5%
Zn–2% Pb (no data are available for a single phase Cu–36% Zn) and
in compression for Cu–44% Zn (instead of Cu–42% Zn which is
Fig. 15. Stress vs strain curves for Cu–40% Zn–3% Pb [6] calculated by the composite
closer to the real composition of the β-phase). In particular, the
model; the behavior of the α and β phases is assumed to be equivalent to that of the
flow stress can be calculated by the following equation: Cu–36.5% Zn–2% Pb (data for Cu–36% Zn are not available) and Cu–44% Zn [6]
″ 0 ″ 0 ″ 0 respectively.
σ ¼ σ s ff s  f s ð1  f s Þ=ð1  f s Þg þσ 0 ½ð1  f s Þ=ð1  f s Þ ð11Þ
0 ″
where and are the equilibrium volume fractions of β-phase in
fs fs
Cu–36.5% Zn–2% Pb and Cu–40% Zn–3% Pb respectively, while σ0 same calculation can be carried out for the CW602N alloy, with
and σs are the flow stresses for Cu–36.5% Zn–2% Pb and Cu–44% Zn fs ¼3.0%, 3.3%, 10% and 28% (equilibrium volume fraction of β at
(Fig. 15). The agreement between the calculated curves and the 550, 600, 650 and 700 1C respectively; solid curves in Fig. 14).
experimental data is excellent, thus demonstrating that the Again, although the model is based on rough assumptions
composite model gives a coherent picture of the deformation of (in particular, the real volume fraction of β-phase could differ from
these alloys. the equilibrium one, due to short time at high temperature), an
A subsequent step is the recalculation of the peak flow stress excellent agreement between the experimental data and the
values using the composite model. For Cu–44% Zn alloy, which, as curves is obtained, in particular at the highest temperature, for
mentioned, has a mechanical behavior almost analogous to that of which the single-phase model described in Section 4.3 gave a
the β-phase present in these binary alloys, the peak flow stress relatively poor description.
depends on the testing strain rate with an equation in the form
_ n expð  Q =RTÞ
ε_ ¼ Aβ ½ sinhðαβ σ=GÞ ð12Þ
β
5. Conclusions
1 7 1
where αβ ¼350, n ¼3, Qβ ¼ 92 kJ mol and Aβ ¼1.44  10 s . The
constitutive equation for the α-phase could easily be computed by The present study analyzed the hot workability of Cu–36.5%
the model described in the previous section, with Zn¼ 36%. The Zn–2% Pb (CW602N) brass, investigated by torsion testing at
resulting model curves, which again show an excellent agreement temperatures ranging from 550 to 700 1C. A comparison with
with the experimental data, are presented in Fig. 16. The only literature data obtained in compression, torsion and tensile creep,
significant deviation is observed at the highest temperature, demonstrated that the mechanical behavior of CW602N is inter-
probably as an effect of the overestimation of the β-content. The mediate between those observed in brasses containing 30% and
 S. Spigarelli et al. / Materials Science & Engineering A 615 (2014) 331–339 339

[2] R.M. Bonesteel, O.D. Sherby, Acta Metall. 14 (1966) 385–391.

[3] S.V. Raj, T.G. Langdon, Acta Metall. 37 (1989) 843–852.
[4] S.V. Raj, J. Mater. Sci. 26 (1991) 3533–3543.
[5] W. Blum, J. Dvořak, P. Král, P. Eisenlohr, V. Sklenička, J. Mater. Sci. 49 (2014)
2987–2997.
[6] D. Padmavardhani, YV.R.K. Prasad, Mater. Sci. Eng. A 157 (1992) 43–51.
[7] D. Padmavardhani, YV.R.K. Prasad, Metall. Trans. 22A (1991) 2993–3001.
[8] R. Pernis, J. Kasala, J. Bořuta, Kov. Mater. 48 (2010) 41–46.
[9] A.P. Miodownik, ASM Handbook – Phase Diagrams, vol. 3, ASM International,
Materials Park, Ohio, USA, 1992.
[10] A. Rohatgi, K.S. Vecchio, G.T. Gray, Metall. Mater. Trans. 32A (2001) 135–145.
[11] Y.H. Zhao, X.Z. Liao, Y.T. Zhu, Z. Horita, T.G. Langdon, Mater. Sci. Eng. A 410–411
(2005) 188–193.
[12] B.L. Vaandrager, G.M. Pharr, Scr. Metall. 18 (1984) 1337–1339.
[13] H.J. McQueen, M.E. Kassner, Mater. Sci. Eng. A 410–411 (2005) 58–61.
[14] S. Spigarelli, M. El Mehtedi, Mater. Sci. Eng. A 527 (2010) 5708–5714.
[15] V. Buršiková, J. Buršik, V. Navrátil, K. Milička, Mater. Sci. Eng. A 324 (2002)
235–238.
[16] Z. Horita, T.G. Langdon, in: H.J. McQueen, J.-P. Bailon, J.I. Dickson, J.J. Jonas,
M.G. Abken (Eds.), Strength of Metals and Alloys, Pergamon Press, Oxford,
1985, pp. 797–802.
[17] S. Spigarelli, O.A. Ruano, M. El Mehtedi, J.A. delValle, Mater. Sci. Eng. A 570
(2013) 135–148.
[18] S. Spigarelli, M. El Mehtedi, Mater. Sci. Eng. A 527 (2009) 126–131.
[19] S.S. Vagarali, T.G. Langdon, Acta Metall. 29 (1981) 1969–1982.
[20] S.V. Raj, T.G. Langdon, Acta Metall. 39 (1991) 1823–1832.
[21] R. Fuentes-Samaniego, W.D. Nix, G.M. Pound, Philos. Mag. A 42 (1980)
591–600.
Fig. 16. Description of the peak flow stress dependence on strain rate for the Cu–
[22] R. Fuentes-Samaniego, W.D. Nix, Scr. Metall. 15 (1981) 15–20.
40% Zn–3% Pb alloy [6]; the solid curves were calculated by the composite model,
[23] H. Nakajima, J. Mater. Res. 49 (1997) 15–19.
assuming that the alloy is composed by α and β phases containing 36% and 44% Zn
[24] E. Kirkendall, L. Thomassen, C. Upthegrove, Trans. AIME 133 (1939) 186–203.
respectively (Section 4.4). [25] E.O. Kirkendall, Trans. AIME 147 (1942) 104–110.
[26] A.D. Smigelskas, E.O. Kirkendall, Trans. AIME 171 (1947) 130–142.
[27] J. Steigman, W. Shockley, F.C. Nix, Phys. Rev. 56 (1939) 13–21.
40% Zn. A constitutive model, based on the Garofalo equation, was [28] J.A. Brinkman, C.E. Dixon, C.J. Meechan, Acta Metall. 2 (1954) 38–48.
developed to obtain a reliable description of the hot strength of [29] J. Hino, C. Tomikuza, C. Wert, Acta Metall. 5 (1957) 41–49.
single-phase α-brass. Although the model gives excellent descrip- [30] N.L. Peterson, S.J. Rohtman, Phys. Rev. B 2 (1970) 1540–1547.
[31] K.J. Anusavice, R.T. DeHoff, Metall. Trans. 3 (1972) 1279–1298.
tions of these materials, the accuracy is considerably lower at the [32] J. Kučera, B. Million, J. Plšková, Phys. Status Solidi A 11 (1972) 361–366.
highest temperature in the case of CW602N. This alloy contains a [33] R. Resnick, R. Balluffi, Trans. Am. Inst. Min. Metall. Eng. 203 (1955) 1004–1010.
non-negligible amount of the soft bcc β-phase, and for this reason, [34] L.C. Correa Da Silva, R.F. Mehl, Trans. Am. Inst. Min. Metall. Eng. 191 (1951)
155–173.
it presents marked similarities with other bi-phase α/β materials
[35] A. Khezri-Yazdan, K.N. Subramanian, J. Mater. Sci. 19 (1984) 3191–3195.
such as the Cu–40% Zn alloy. The composite model can then be [36] D. Hua, W. Qingling, M. Longxiang, J. Mater. Sci. 27 (1992) 607–610.
used to model the hot deformation of these more complex alloys, [37] R. Sedláček, W. Blum, Comput. Mater. Sci. 25 (2002) 200–206.
obtaining an excellent description of the experimental results.

References

[1] H.J. McQueen, S. Spigarelli, M.E. Kassner, E. Evangelista, Hot Deformation and
Processing of Aluminum Alloys, CRC Press, Roca Baton (2011) 267–276.