Materials Science and Engineering, A 147 ( 1991 ) 143-154 143

Plastic behaviour of copper sheets during sequential tension tests

J. H. Schmitt*, J. V. Fernandes and J. J. Gracio

l)epartamento de Engenharia Mec~nica, Faculdade de Ci~ncias e Tecnologia, Universidade de Coimbra, P-3(XX) Coimbra
(Portugal)

M. F. Vieira
Departamento de Engenharia Metal6rgica, Faculdade de Engenharia, Universidade do Porto, P-4099 Porto Codex (Portugal)

(Received January 3, 1991 ; in revised form May 17, 1991 )

Abstract

Sequences of two uniaxial tension tests along different axes were performed on a polycrystalline copper
sheet. In order to reach a better understanding of the physical mechanisms occurring during the second
deformation, the effect of the strain path change on subsequent yield and flow behaviour has been inves-
tigated using optical and transmission electron microscopy. The value of the reloading yield stress is a
function of the angle q~ between the two tensile axes. For • > 15° the requirement of glide of disloca-
tions with a new Burgers vector implies that a very low density of potentially mobile dislocations is
available at the beginning of the reloading deformation. This effect is moderated at • values of about
90 ° by the inverse activity on some slip systems during reloading. The transient observed in the work-
hardening behaviour after the path change is concomitant with the disappearance of some dislocation
walls developed during the prestrain. A more homogeneous dislocation substructure appears for
q~ > 15° as a result of the interactions between the mobile dislocations on the new active slip systems and
the previous dislocation walls; the dissolution of the previous dislocation arrangement is also promoted
by the inversion of the slip direction in some systems, depending on the value of the angle q~. This
results in an increase in dynamic recovery rate during the early stage of reloading, particularly when
approaches 90 °.

1. Introduction In a recent study [2] concerning the plastic

Previous work [1] has shown that during uni- deformation of copper sheets under sequential
axial tension of polycrystalline copper sheets, dis- strain paths, the combined effects of the total
location cell walls develop in the grains after a prestrain, the nature of the preloading mode and
few per cent of deformation. The shape of the the magnitude of the strain path reorientation on
cells is related to the number of active slip planes, the subsequent mechanical behaviour in tension
For small grain sizes (less than 65 ~m) two inter- have been analysed. The mechanical behaviour
secting families of parallel dislocation walls, deli- during subsequent loading appears to be only
miting parallelogram cells, are generally observed weakly affected by the first loading mode. It is
in the grains. In these cases the dislocation walls mainly dependent on the magnitude of the strain
are straight and their orientations are closely path change as defined, for example, by the
parallel to the most active slip planes when esti- cosine of the angle between the two vectors which
mated through a static model (Schmid factor) or a represent the successive strain tensors [3]. It is
Taylor analysis. The deviation from this orienta- therefore of major importance to understand the
tion increases as the predicted number of active effect of the magnitude of the path change on the
slip systems rises, evolution of the dislocation structure.
The present work deals with sequences of two
*On leave from Gdnie Physique et M~canique des Mat~ri- uniaxial tension tests performed along different
aux, (URA CNRS No. 793), Institut National Polytechnique
de Grenoble, ENSPG BP 46, F-38402 Saint Martin d'Heres a x e s and covering a wide range of strain path
C~dex, France. changes. The microstructural evolution during

0921-5093/91/$3.50 © Elsevier Sequoia/Printed in The Netherlands

144

the transient period which follows the reloading were performed at room temperature. The initial
was carefully studied. More specifically, an analy- strain rate was 5 x 10- 3 s- ~ and the strain
sis of the active slip systems after the path change, amounts were deduced from extensometer data.
with respect to those previously active, is carried Microstructural observations were performed
out. This leads to an understanding of the using both optical and transmission electron
observed differences in the latent hardening microscopy. The observation plane was the sheet
phenomenon and the Bauschinger effect at var- plane.
ious orientations. The main object of the study is Slip lines were observed through an optical
to clarify the correlation between these micro- microscope by the differential interference
structural effects and the yield stress and flow contrast technique. Prior to testing, the specimens
behaviour after the path change, were polished using a solution containing 250 ml
orthophosphoric acid, 250 ml ethanol, 500 ml
2.. Experimental details distilled water and 3 g urea and using a current
density of 80 A dm-2.
Oxygen-free high purity copper (99.95% Cu) The observation of the dislocation microstruc-
sheet 1 mm thick, cold rolled and annealed, was tures was performed using a 100 kV transmission
used in this investigation. The mean grain size is electron microscope; the samples, taken from the
20/~m, the grains being rather equiaxed, middle area of the tested samples, were mechani-
X-ray analysis revealed the existence of a weak cally ground on both surfaces, then electropol-
rolling texture [4] as is usually observed in ished using a double-jet thinner and a dilute
annealed industrial copper sheets. The copper solution of orthophosphoric acid (2:1) under
component is clearly noted while other minor 10 V tension at 20°C until perforation
texture components such as {110}(112) (brass occurred. Most of the grain s observed by trans-
component), {146}(211) and {123}(412) have a mission electron microscopy (TEM) had an
slightly higher intensity than for a random distri- orientation such that {110} or {112} planes were
bution, parallel to the sheet plane, in agreement with the
In order to measure the influence of the initial texture measurement.
work-hardening anisotropy on the plastic
response of the copper sheet, uniaxial tension
3. Results
tests were performed on ISO 50 samples (75
m m × 12.5 mm) cut in the rolling plane of the 3.1. Macroscopic plastic behaviour
sheet at different angles • with respect to a refer- The stress-strain curves along the reference
ence direction normal to the rolling direction; direction (normal to the rolling direction) have
seven orientations ( ~ = 0 °, 15 °, 30 °, 45 °, 60 °, 75 ° been drawn for three prestrain amounts (namely
and 90 °) were considered. The sequential tension ep = 0.06, 0.12 and 0.18) along various tensile
tests were performed by cutting ISO 50 samples directions • with respect to the reference direc-
from large tensile specimens (500 mm x 220 mm) tion. The prestrain amounts are such that the
which had been previously strained up to various macroscopic deformation may be assumed homo-
strain v a l u e s (ep). The prestrain was obtained by geneous in the whole sample. In Fig. 1 are pre-
pulling the samples along the same directions sented three examples of stress-strain curves for
(same • values) as used for the plastic anisotropy • = 15 °, 45 ° and 90 °. On the same graphs is plot-
study. For any of these • values a set of three ted the tensile curve along the reference direction
prestrained samples was pulled for ep = 0 . 0 6 , (termed the reference curve). For comparison the
0.12 and 0.18. The strain was checked to be amplitude of the plastic anisotropy during uniax-
almost homogeneous over a large region near the ial tension tests along various directions with
middle of the large specimens and it has been respect to the reference direction is shown in Fig.
observed that the strain path followed in this case 2 for the different angles ~ which have been
is very close to that followed in the ISO 50 sam- studied.
pies. After each prestrain, i.e. a • value and a As a general result the macroscopic reloading
prestrain ep, an ISO 50 sample was carefully cut yield stress (which may be defined as the back-
from the middle of the large strained specimen extrapolated stress abe, for instance) is higher
along the transverse direction used as a reference than the stress reached on the reference curve at a
direction. In all cases the uniaxial tension tests strain which is equal to the prestrain amount. The
 145

! i i , i , i , i

(a) @:15° ~ 4

300200 ~ f - 300
200 ~ 5 _

~[ [ 0"- I (RC)
o. ,', 15"-2
~ p" q~= 14~"-4
100 100 60"
75°--65
90°- 7

t I e I , I 0 , I I I , I
O0 0.1 0.2 0.3 0 0.1 0.2 0.3
E E
f ~ ~ ~ Fig. 2. True stress-true strain curves for seven orientations
(b) @:45° ~ (I) between the tensile axis direction and the reference direc-
30 0 - t]on.

200
In order to analyse the previous results in a
y more quantitative way, it is possible to present the
variations of a "normalized" reloading stress with
100 respect to the angle ~ . T h e analysis is performed
according to the discussion given by Schmitt et al.
[7].
, D , i , ~ (1) A reference curve is defined as the
0 0.1 0.2 0.3 stress-strain curve obtained from a non-deform-
E ed sample pulled under the same loading condi-
i i !
(c) @:90 ° tions as for the second deformation in the
300 ~ - sequential test; in the present work the reference
curve is the stress-strain curve recorded during a
tension along the transverse direction.
(2) A reference stress ar is then estimated as
200 jJ the stress value in the reference curve for which
¢
~ ~/ the plastic work is equal to the plastic work done
during the prestraining up to %.
loo (3) After reloading, a macroscopic yield stress
abe is measured using a back-extrapolation
method.
In Fig. 3 the evolution of the normalized
, 0.2 I0.1 , O . 3t reloading stress (OhJ a r ) vs. the angle • is plotted.
E T h e global effect of the prestrain is an increase in
Fig. 1. Examples of typical true stress-true strain curves of reloading yield stress by a factor ranging between
prestrained samples in uniaxial tension (ep=0.06, 0.12 and 1.0 and 1.15. T h e effect is greater the larger the
0.18); the reference curve (RC) is also shown. The orienta-
tion of the first tension is such that (a) • = 15°, (b) q~= 45 ° prestrain amount. For • > 15 ° the normalized
and(c) ~ = 9 0 °. reloading stress increases and reaches its maxi-
mum for • between 45 ° and 60 °. A slight drop is
reloading yield stress is followed by a transient noted for • around 90 °, though the value of the
region with a low strain-hardening rate (Fig. 1). normalized stress is still above unity. This evolu-
T h e amplitude of these effects is dependent upon tion is in qualitative agreement with the results
the angle • as commonly observed for other which were obtained for mild steel during similar
metals (see e.g. refs. 5 and 6). sequential experiments [5], though the crystallo-
146

2.00 . . . . . . . . . . . . . . . . . . graphic nature (and thus the dislocation motions

[] °'°61 I and interactions) is different.
1.16 IA~ ,.o._....,o.~
~.i ~2,E
i ~,,i
o, During reloading, the general trend is a lower
1.12 1 ~ strain-hardening rate than for a monotonic test
t~ ~ ~ along the reference direction at a given stress
to6 (Fig. 4). The decrease is larger the greater the
prestrain amount. In all cases the strain-harden-
1.04 ing rate v s . stress curves tend asymptotically
1.00 [] towards the monotonic curve. Conversely to the
reloading stress results, the evolution of the
o.960 . . . . .15. . . . . . 3. 0. . . . . . 4 5 60 75 90
strain-hardening rate with respect to the value of
(/) (o) • is monotonic: the initial decrease is larger and
the transient longer the larger the • value. For a
Fig. 3. Evolution of the normalized reloading stress ob~/o~ prestrain amount equal to 0.18 and for • > 45 °
with the angle qb between the two successive tension axes.
(See Fig. l(b) for the definition of ob~.) the drop in strain-hardening rate is so important

1000 . . . . i . . . . I . . . . I ' ' 1000 . . . . I ' ' ' I . . . . i ' "

,, ,,,, ca) Ep=O.06 , (C) Ep=0.18
•, IL ~1

800 800 -

2_ 3

45 I~

600 ~ 600 7t~SLL~-

I I i I I

400 400 - "ii ,, 'I

II
,,,,
. . . .~.~'~'~ . ..~A'~. ~ ~ "
O~a- 07 ~....
, ~ , . L.-r%~, ~ , I . . . . I 200 i i , L....r'q~ , I , , , I , ,
20050 200 250 300 150 200 250' 300
(MPa) O" ( M P a )

1 0 0 0 . . . . [ . . . . | . . . . | , ,

', (b) Ep= 0.12

800 -

i"4,'~ r',k F z

e ~o. ~. ....... " Evolution of the work-hardening rate 0 = do/de with

F i g . 4.
.~..-. the tensile stress o." curve 1, sample without prestrain along
200 , , , . L.~.~', ~ . , I . . . . I . , the reference direction; curves 2-7, prestralned samples such
150 200 250 300 t h a t q ) = 1 5 °, 3 0 °, 4 5 °, 6 0 °, 7 5 ° a n d 90 ° respectively; (a)
O" ( M P a ) ep=0.06; (b) ep=0.12;(c) ep=0.18.
 147

that the instability point (maximum load) is tural changes obviously depends on the angle
reached at the beginning of the transient [2]. This between the two successives tensile axes. The sub-
phenomenon is a limiting feature of the sequential structural evolution takes place throughout a
strain path studies when the second loading is a wide range of strain values along the second
uniaxial tension test. strain path (up to at least 0.10 for a prestrain
value of 0.12). At a given macroscopic strain the
3.2. Microscopic observations state of evolution in each individual grain is dif-
Microscopic analyses have been performed at ferent depending on its orientation.
two different scales: dislocation cell structure TEM observations are presented for three
observations in bulk material by TEM; surface values of the angle qb which seem to correspond
characterization of slip lines developed during to some characteristic evolutions of the disloca-
the plastic deformation, tion structure, namely q) = 15 °, 45 ° and 90 °. In the
case of a small amplitude of the strain path
3.2.1. Dislocation substructure change (~ = 15 °) no noticeable change of the dis-
As has been commonly reported in various location arrangement is observed. The substruc-
works on copper (see e.g. refs. 8-10), a disloca- ture which develops during the prestrain evolves
tion cell substructure develops in grains during continuously, the same as for monotonic tension.
uniaxial tension. For a strain of 0.06 a well-devel- For larger amplitudes of strain path change
oped cell structure is observed in most of the (~ = 4 5 ° and 90 °) the dislocation walls created
grains and dislocation walls are well defined in during the prestrain tend to disappear. The dislo-
samples deformed up to strains of 0.12 and 0.25. cation wall thickness increases and the disloca-
In most of the grains two intersecting families of tion density seems to decrease in the walls as
straight dislocation walls are present (Fig. 5), revealed by the change in TEM contrast. At low
forming a closed-cell parallelogram structure as strain, after reloading (about 0.025), the disloca-
already reported [1]. In a few grains equiaxed tion structure is unorganized in most of the grains
cells are observed. The above analysis is generally (Fig. 6): the distribution of the dislocations is
independent of the direction of the tensile axis more homogeneous inside the grain and the dis-
with respect to the rolling direction, location tangles do not present any preferential
After a change in the tensile axis direction, dis- orientation. For a • value of 90 °, in a few grains
location arrangements evolve; at large strains (at the partial disappearance of the dislocation walls
and above 0.26) during the second tension the looks like a dissolution of the pre-existing walls
dislocation cell structure tends to be similar to the (Fig. 7): straight dislocation segments seem to
substructure which develops during a monotonic "move" from the wall vicinity towards the cell
tension from a virgin sample (similar cell shape
and dislocation wall orientation with respect to
the tensile axis). The magnitude of the substruc- . ~.~

7•~o/'t~,~,
. ~ ~ Fig. 6. T E M observation of unorganized dislocation struc-
re~.~t--c.v.~._~._~._~ ~ Iure developed during the second deformation (strain
approximately 0.025) in a grain with (110) axis normal to the
Fig. 5. T E M observation of closed cells developed in a grain sheet plane. The prestrain value is Ep = 0.12 and the angle
with (110) axis normal to the sheet plane after a strain ~etween the two tensile axes is q~ = 4 5 °. (T.A.1, axis of first
amount equal to 0.15. (T.A., tension axis, parallel to the refer- Lension; T.A.2, tensile axis of the subsequent deformation,
ence direction.) parallel to the reference direction.)
148

Fig. 8. TEM observation of closed parallelogram cells devel-

oped during the second%= 0.def°rmati°nl
2 up to betweena
strain value of
0.26 in a grain with (110) axis normal to the sheet plane. The
prestrain value is and the angle the two
tensile axes is • = 90 °.

lowed during a monotonic tension test for which

the same area was studied at different strain

I
amounts (Fig. 9). Up to relatively high strain
,~ ,,.v'/~'i~ (greater than 0.30)one slip line direction is pre-
•:--~ ~ o dominant in most of the grains. The slip traces
are inclined at 400-60 ° to the tensile axis (about
52% of the observed slip lines in a total of 2500
Fig. 7. TEM observation of dislocation cell structure devel-
oped during the second deformation (strain approximately measurements--Fig. 10). In some grains other
0.025) in a grain with (110) axis normal to the sheet plane, families of slip lines are noted, but frequently i n a
The prestrain value is ep = 0.12 and the angle between the restricted a r e a n e a r the grain boundary. For strain
two tensile axes is qb = 90 °. The traces of the slip planes of the
inverted systems are indicated in (b), the enlargement of a values equal to or greater than 0.20 a new slip
detail of (a). line direction is sometimes observed (e.g. see Fig.
9(d)). Some bending of slip lines is apparent at a
strain of 0.30 which may be connected with a
interior in the same way as reported during a certain amount of rotation in some areas of the
Bauschinger experiment [11]. grain and/or with roughness development. The
For these two • values (45 ° and 90°)new dislo- observations made for tensions along various
cation walls appear progressively in grains as the directions • with respect to the reference direc-
total strain increases during the second tensile tion point out a weak dependence of the slip line
loading. This wall development is similar to that patterns on ~. This is mainly related to the weak-
during the first step of the deformation sequence ness of the initial and induced crystallographic
but with a smaller cell size, in agreement with the texture as also shown by the stress-strain curves.
larger stress level during reloading. As observed During sequential tensile tests, observations
during a linear monotonic tension, the new cell have been made on the same intially polished area
structure is mainly formed by two families of after the prestrain and at different strain amounts
rather straight dislocation walls which define during the second tension. This allows new slip
closed parallelogramcells(Fig. 8). lines which develop during reloading to be
viewed. A few examples are given in Figs. 11 and
3.2.2. Slip line analyses 12 for • = 45 ° and 90 °. A statistical analysis may
Surface observations have been made during be done which exhibits the frequency of activa-
sequential tensions in order to check the develop- tion of new slip systems (or at least new slip
ment of new slip systems during reloading and the planes) during reloading as a function of the angle
eventual existence of strain localization at the ~ between the two successive tensions. The dia-
macroscopic scale. To make possible a compar- grams in Fig. 13 show the frequency of orienta-
ison, the evolution of slip lines has first been fol- tion of the new slip lines compared with the
 149

25
2O

Z
10

0
0 10 20 30 40 50 60 70 80 90
(o)

Fig. 10. Diagram showing the orientation distribution of slip

line directions on the sheet plane with respect to the tensile
"-- axis direction during monotonic uniaxial tension. ~ is the
angle between the slip line direction and the tensile axis
direction and N (%) is the percentage of total slip line direc-
tions having an orientation ~2 + 5 ° with the tensile axis.

orientation distribution of the slip lines due to the

! prestrain (this last information corresponds in
fact to the diagram of Fig. 10 where the slip line
orientations are expressed with respect to the
i second tensile axis direction).
The main results may be summarized as fol-
lows.
( 1 ) For ~ = 15°, even after a large strain along
the second tension (greater than 0.15), only for a
very small number of grains are new slip lines
seen.
(2) For ~ = 4 5 ° and 90 ° new slip lines are
clearly visible soon after reloading; they develop
with an average orientation which is identical to
the results obtained in simple tension.
! (3) For the last two examples no trace of
coarse slip or grain boundary shearing has been
noted; in the same way no evidence of strain
localization has been observed on the sample
surface as reported during the deformation of
prestrained steel sheets [12].

4. Discussion
The above results show the now well-estab-
lished effects of a prestrain on the subsequent
plastic behaviour of polycrystalline metals,
namely an increase in macroscopic reloading
yield stress and a drop in work-hardening rate
during the early deformation along the second
strain path. The various strain path changes which
have been investigated in the present paper allow
the influence of the amplitude of the strain path
Fig. 9. Optical obervations of slip lines developed during change (the angle between the axes of the two
monotonic uniaxial tension along the reference direction
(horizontal in the figure) at increase strain values (a) 0.05, (b) successive tensions) on these two mechanical
0.10, (c) 0.20 and (d) 0.30. parameters to be studied.
150

Fig. 12. Optical observations of slip lines (a) after a prestrain

of ep = 0.12 and (b) during the subsequent tension along the
reference direction (horizontal in the figure) at a strain value
of 0.15. The angle between the two tensile axes is ¢ = 90*.

l valid for the low strain amounts which have been

imposed as a prestrain (not greater than 0.18). In
fact, the flow stress during the monotonic tension
may be expressed as the product of the average
Taylor factor ( M r ) and the effective resolved
shear stress f [13, 14]:
w,,,\ Or = (Mr r) = ( m r ) ~
Fig. 11. Optical observations of slip lines (a) after a prestrai 0
of e_ = 0.12, and during the subsequent tension parallel to the Since multiple slip occurs in most of the grains
refe~'ence direction at strain values of (b) 0.05 and (c) 0.15. during tension [1], in agreement w i t h t h e Taylor
The angle between the two tensile axes is ¢ = 45 °. assumption, the intragranular behaviour can be
considered isotropic throughout a monotonic
4.1. Reloading yieM stress loading [15-17]. The evolution of i may then be
expressed as a function of a single parameter
At first sight the evolution of the reloading whatever the direction of the tensile axis with
yield stress with respect to the angle ¢P between respect to the rolling direction. A usual proce-
the two successive tension axes (Fig. 3) may dure is to relate the effective resolved shear stress
appear to be due to a crystallographic texture to the plastic work [7]. The differences between
effect and/or to intragranular anisotropic harden- the various stress-strain curves, corresponding to
ing. In fact, the use of the ratio Obe/Or allows the the prestrain paths, are thus connected with the
effect of the intragranular hardening alone to be values of the Taylor factor. The reloading yield
examined, assuming that the texture evolution is stress abe is expressed as
weakly dependent on the tensile direction during
the prestrain [7]. This assumption is essentially Obe=(mberbe)=(Mbe)(f+A'~)
 151

25 material axis. The above equation thus reduces to

20 @=15°
abe Af
-- -]+--
Z 1510 Or 7"

The evolution of the reloading stress (as

5 expressed by the ratio A~?/f) with respect to the
o 0 10 20 30 40 50 60 70 80 90 amplitude of the change in strain path has to be
(o) related to the physical mechanisms which occur
25 during the subsequent deformation. The
observed increase in yield stress could be due
20 d)=45°
either (i) to an interaction between the newly
~,s ~ ~ ~ i ~ ~ i ~ active slip systems and the dislocation substruc-
ture developed during the prestrain, i.e. a disloca-
lO tion wall effect, or (ii) to a low number of possible
5 active sources during reloading.
0 (i) With the hypothesis of a dislocation wall
0 10 20 30 no so so 70 80 90 effect the ratio O'be/Or would be mainly related to
(o) the new active slip planes and be a function of the
25 mutual misorientation between these slip planes
20 0=90° and the walls. The dislocation substructure devel-
j~ 7 - ~ ~ oped during the prestrain is composed of closed
15 cells accounting for an intense multiple slip in the
z lo whole grain [1]. The main influence of the dislo-
cation walls is expected to induce a fairly iso-
s tropic intragranular hardening. It thus cannot
o account for the dependence of the reloading
0 10 20 30 40 50 60 70 80 90
~(o) stress evolution on the angle between the two
Fig. 13. Diagrams showing the orientation distribution of the successive tensions.
slip line directions on the sheet plane with respect to the (ii) The second effect may be analysed by
second tensile axis direction. The hatched bars indicate the investigating the nature of the active slip systems
slip line directions just after prestrain; the solid bars repre- during the two successive strain paths. Assuming
sent the ratio, as a percentage, between the number of new
slip line directions (developed during reloading) and the total that the intragranular behaviour can be analysed
number of slip line directions after prestrain, through the Taylor model [1], for different grain
orientations which are representative of the main
where (Mbe) is the average Taylor factor along the components of the texture in the sheets, the sets
reference direction (the reloading direction) of possible active slip systems have been esti-
accounting for the texture present during the mated during the two successive tensions for var-
prestrain and f"+ A f- is the effective reloading ious values of ~. Table 1 gives partial results for
shear stress including the intragranular effects the copper and brass components. The possible
due to the change in strain path. Since abe and o r slip systems (labelled using the Schmid and Boas
are defined at the same "equivalent strain" (i.e. notation) are listed in Table 1 when they account
for the same plastic work), the value of ~ is the for more than 20% of the total shear in at least
same for abe as for o r. The ratio Obe/O r is one set of Taylor solutions; this explains the fact
that in some cases less than five slip systems are
O'be- (mbe) q (mbe) A f listed. The line for ~ = 0 ° corresponds to the
or (Mr) (Mr) f active slip systems along the second tension,
while the others, for • values between 15 ° and
With the assumption of a weak dependence of the 90 °, correspond to the different prestrain condi-
textural evolution on the tensile direction during tions. The proportion of new active slip systems
the prestrain, (Mbe) is nearly equal to (Mr) since or of inverted active slip systems (local Bausch-
the two tensions (i.e. the reference and subse- inger conditions), in the various cases of change
quent tensions) are performed along the same in strain path which have been investigated, is
152

TABLE 1
Active slip systems during the two successive tensions for the copper and brass texture components according to the Taylor
model

la,I Copper texture Brass texture

(deg) {112/(111) {110/(112)

0 B4 B5 C3 D----6 ]~ ~ B2 B5" c a
~ B--5'c3 L"5
15 7~k3A6 B4 B5 C3 C-~ D4 D---6 or

~ ~ B'-~B--5'C 3 L"-5D 4 I3-"6

30 ] ~ ] ~ C 3 ~'513-1 D--6 or
~ ~ B2 B-5 b--4 D6
~ ] ~ 1~-4] ~ C3 L-'n5D4 13--6
45 ] ~ C3 13i- or
~ i~-3])-g D6
~ 7~ B-~ B-5 C 3 L"-5'D4 D"6
60 ] ~ B4 ]5-i-13-4 or
]~ ~ B--g/g~ L'~JC5 l)-g D6

75 B4 ]3-i-~ or
~ B-4~-J c5
90 B4 B5 C1 C5 D-i"I)'g A2 ] ~ ] ~ I ~

For the brass component the results are different for tension along d~ and for tension along -d9 ; the two sets of solutions which are
listed in the table correspond to positive and negative • values. For the copper texture only one set of solutions is given because
of the crystallographic symmetry.

thus easily seen. It must be first noted that a larger the dislocation substructure during reloading, and
dependence of the slip system changes on the with the surface analyses, for which no new slip
amplitude of the strain path change exists for the trace develops for this low amplitude of strain
brass component; the net macroscopic effect is, path change. For higher values of the angle • the
however, given by the weighted results for the dif- activity of new slip systems may be clearly con-
ferent texture components present in the sheet. A nected with the appearance of new slip traces on
statistical study indicates that on average the the samplesurface.
number of new active slip systems increases for • For • > 15 ° the requirement of glide of dislo-
increasing between 15 ° and 60 °, then is constant cations with a new Burgers vector implies that a
for higher • values. It is noticeable that the very low density of potentially mobile disloca-
occurrence of a new slip system corresponds to tions is available at the beginning of the reloading
the slip of dislocations with a Burgers vector stage. A high reloading stress is then needed to
which has not been activated during the prestrain, initiate the multiplication process, as also
The fact that O'be/Or appears to be related to the observed for the latent hardening experiments on
number of new active Burgers vectors rather than copper single crystals (see e.g. refs. 17-19). This
new active planes reinforces the assumption of a effect is moderated at • values of about 90 ° by
lack of mobile dislocations during reloading and the inverse activity on some slip systems during
weakens the hypothesis (i) of a dislocation wall reloading. The internal stresses due to the dislo-
effect. For • >160 ° a few slip systems exist for cation walls tend to decrease the critical shear
which the direction of slip is inverted between the stress on these systems. This effect leads to an
prestrain and the subsequent deformation. The overall decrease in the yield stress required for
number of such slip systems increases slightly for plastic flow of the grains and tends to counteract
increasing from 60 ° to 90 °. At • = 90 ° the the latent hardening effect on the other slip sys-
number of inverted slip systems is approximately tems. It explains why the maximum of the abe~Or
equal to the number of new active slip systems, vs. • curve occurs between 45 ° and 60 °.
The weak evolution of the stress for q3 = 15 ° is
understood by the fact that in most of the grains 4.2. Work hardening
the same set of slip systems is active for both The observed decrease in strain-hardening
strain paths. This agrees with the T E M observa- rate during reloading at a given stress level (Fig.
tions, for which almost no change is observed in 4) is concomitant with a rearrangement of the
 153

dislocation substructure as seen by the T E M (no obvious evidence of coarse slip) or TEM
observations. Most of the dislocation walls gener- observations (no microbands or trace of localized
ated during the prestrain tend to disappear and shearing inside the grains).
the dislocation distribution is frequently more (iii) The disappearance of some dislocation
homogeneous inside the grain during the first walls and the development of a more homogene-
steps of the subsequent deformation. A new kind ous distribution of dislocations result certainly in
of dislocation substructure gradually develops an increase in dynamic recovery rate during the
whose orientation is in agreement with the new early stage of reloading. The fact that even at rela-
loading conditions. Three main causes may then tively low • values the slip direction is inverted
account for the transient decrease in strain-hard- on some slip systems favours the dissolution of
ening rate: dislocation walls, as seen in the extreme case of
(i) the disappearance of dislocation walls; the Bauschinger experiments [11], and the annihi-
(ii) the rapid multiplication of mobile disloca- lation of a large number of dislocations which
tions on the new active slip systems; were tangled in the walls. An example is clearly
(iii) the decrease in total forest dislocation seen in Fig. 7 for a 90 ° change in strain path. For
density due to an increase in dynamic recovery, this grain orientation (Fig. 7) the Taylor analysis
(i) The current intragranular stress can be shows that multiple slip on four slip planes takes
expressed as place during the prestrain, in agreement with the
r = aGbp~/2 rather equiaxed dislocation structure which de-
veloped. The new tensile direction induces an
where a is a parameter which depends on the dis- inversion of the slip direction on two of these slip
location interactions [20] and the dislocation planes, while the other two slip planes are still
arrangement [21], G is the shear modulus, b is the active but along another slip direction. The
Burgers vector amplitude and p is the dislocation straight dislocation segments which are present in
density. From recent computer simulations the the cell interior during reloading are well aligned
value of a has been shown to be lower for a horn- with the traces of the slip planes of the two
ogeneous than a non-homogeneous distribution inverted systems (Fig. 7). This example, sup-
of dislocations at a constant dislocation density ported by some other T E M observations, tends
[21]; it also corresponds to a lower strain-harden- to confirm that a Bauschinger-like substructural
ing rate. The partial "dissolution" of the disloca- evolution occurs in some grains during the stud-
tion walls induces a decrease in a with a ied change in strain path, in agreement with the
consequent negative contribution to the evolution prediction of the Taylor analysis. This partial
of the strain-hardening rate. The difference in the Bauschinger effect is thus an explanation of the
range of a values (a between 0.30 and 0.24 for a disappearance of some dislocation walls beyond
well-defined dislocation cell structure and the destabilization of some others by interaction
between 0.20 and 0.16 for a random distribution with the newly active dislocations. The dynamic
of dislocations [21])seems, however, too weak to recovery is thus more important the closer the
explain by itself the observed large drop in strain- change in strain path is to the Bauschinger orien-
hardening rate. tation. This is in agreement with the larger drop
(ii) The activation of dislocations with a new in strain-hardening rate for • values near 90 °, the
Burgers vector has been seen to be one of the opposite of the effect on the reloading stress level.
main factors which influences the stress during In any case the drop which has been observed in
reloading. The increase in stress for the activation the present work is less than for pseudo-Bausch-
of new sources leads to a rapid multiplication of inger (as during a sequence of shearing followed
the mobile dislocations on the new active slip by tension [25])or Bauschinger experiments [11]
systems, which usually implies at least a stagna- on similar material.
tion of the work-hardening rate. This could be an The gradual evolution of the dislocation sub-
important factor contributing to the observed structure cannot be associated with any specific
behaviour during reloading, though this multipli- observation of localization of the macroscopic
cation process has been frequently connected deformation as was reported, for instance, for
with a localization of the deformation due to mild steel during rolling-tension sequences [12].
"avalanche" glide [22-24]. Such a mechanism has The early plastic instability process described
not been observed either by surface observation elsewhere [2] then appears as a mechanical conse-
154

quence of the decrease in strain-hardening rate at Acknowledgments

a relatively high level of yield stress.
T h e authors are indebted to the CNRS for
financial support during the two months' stay of
5. Conclusions J.-H. Schmitt at Coimbra University. M. F. Vieira
thanks the Instituto Nacional de Investigaqfio
Sequences of uniaxial tension tests along dif- Cientffica (INIC) for a research grant. T h e
ferent axes were p e r f o r m e d on polycrystalline authors express their grateful thanks to Professor
c o p p e r sheets. T h e correlation established Bernard Baudelet for encouraging this work and
between the macroscopic and microscopic for stimulating discussions. T h e y would like to
behaviours after the path change enables the fol- acknowledge the permission given to use the
lowing conclusions to be drawn. T E M of the Faculty of Medicine of the University
After the path change the increase observed in of Coimbra. T h e c o p p e r sheets on which the
the normalized reloading yield stress is linked to experimental work has been performed were pro-
the activation of new slip systems (not active dur- vided by E U R O P A M E T A L L I - L M I (Florence,
ing the prestrain) with a low density of mobile Italy).
dislocations. For the angle q~ between the two
tensile axes equal to 15 °, a weak evolution of the References
reloading yield stress is understood by the fact
1 J.J. Gracio, J. V. Fernandes and J.-H. Schmitt, Mater. Sci.
that in most of the grains the same set of slip sys- Eng. A, 118(1989) 97.
tems is active for both strain paths. T h e n u m b e r 2 M. F. Vieira, J.-H. Schmitt, J. J. Gracio and J. V.
of new slip systems increases as • increases up to Fernandes,J. Mater. Process. Technol., 24 (1990) 313.
45°; for this • value the reloading yield stress 3 J.-H. Schmitt, E. Aernoudt and B. Baudelet, Mater. Sci.
reaches the maximum. W h e n ~ is increased from Eng., 75(1985) 13.
60 ° to 90 °, the n u m b e r of new slip systems 4 J.L. Raphanel, personalcommunication, 1988.
5 J. L. Raphanel, J.-H. Schmitt and B. Baudelet, Int. J.
remains stationary; in some slip systems the Plast., 2(1986) 371.
direction of slip is inverted between the prestrain 6 J. L. Raphanel, E. Rauch, E. L. Shen and J.-H. Schmitt,
and the subsequent deformation. At • = 90 ° the Set. Metall., 21 (1987) 1087.
n u m b e r of inversed slip directions is already of 7 J.-H. Schmitt, E. L. Shen and J. L. Raphanel, submitted to
Int. J. Plast.
the same order as the number of new slip systems. 8 T. Tabata, K. Takagi and H. Fugita, Trans. Jpn. Inst. Met.,
This explains the drop in relative yield stress as • 16 (1975) 569.
approaches 90 °. 9 N. Hansen and B. Ralph, Acta Metall., 30(1982) 417.
T h e transient observed in the work-hardening I0 M. Doner, H. Chang and H. Conrad, Metall. Trans., 5
behaviour after reloading is concomitant with a (1974) 1383.
rearrangement of dislocation substructure at the 11 N. Christodoulou, O. T. Woo and S. R. MacEwen, Acta
Metall., 34(1986) 1553.
beginning of the second tension. For O > 1 5 ° 12 A. KorbelandP. Martin, ActaMetall.,36(1988)2575.
most of the dislocation walls generated during the 13 U. F. Kocks, Metall. Trans., 1 (1970) 1121.
prestrain tend to disappear owing to the rapid 14 c. Tome, G. R. Canova, U. F. Kocks, N. Christodoulou
multiplication of mobile dislocations on the new and J. J. Jonas, Acta Metall., 32 (1984) 1637.
active slip systems which interact with them. T h e 15 P. Franciosi and A. Zaoui, Acta Metall., 30 (1982) 1627.
16 P. Franciosi and A. Zaoui, Acta Metall., 30 (1982) 2141.
consequent appearance of a more homogeneous 17 P. Franciosi, Acta Metall., 33(1985)1601.
distribution of dislocations results in an increase 18 P. J. Jackson and Z. S. Basinski, Can. J. Phys., 45 (1967)
in dynamic recovery rate during the early stage of 707.
reloading; this is mainly verified as • tends to 90 ° 19 P. Franciosi, M. Berveiller and A. Zaoui, Aeta Metall., 28
(1980) 273.
by the inversion of the slip direction in some sys- 20 F. F. Lavrentev, Mater. Sci. Eng., 46(1980) 191.
tems. 21 F. Hernandez Olivares and J. Gil Sevillano, Acta Metall.,
T h e early plastic instability, which takes place 35(1987) 631.
after a prestrain value equal to 0.18 and for 22 Z.S. BasinskiandS. J. Basinski, Phil. Mag., 9(1964)51.
45°--.
< ~ ~<90 °, is only a mechanical consequence 23 z. s. Basinski and P, J. Jackson, Phys. Status Solidi, 9
of the important drop in strain-hardening rate (1965)805.
24 Z. S. Basinski and P. J. Jackson, Phys. Status Solidi, 10
occurring at the beginning of the second tension (1965) 45.
and cannot be associated with any specific evolu- 25 E. L. Shen, Thesis, Institut National Polytechniquede
tion of the dislocation substructure. Grenoble, 1988.