Acta Materialia 52 (2004) 3657–3664

www.actamat-journals.com

Recovery during annealing in a cold rolled low carbon steel.

Part I: Kinetics and microstructural characterization
a,*
A. Mart
ınez-de-Guerenu , F. Arizti a, M. D
ıaz-Fuentes b, I. Guti
errez b

a
Department of Electronics and Communications, Centro de Estudios e Investigaciones T
ecnicas de Gipuzkoa (CEIT)
and Universidad de Navarra (TECNUN), Paseo de Manuel Lardiz
abal 15, 20018 San Sebasti
an, Basque Country, Spain
b
Department of Materials, Centro de Estudios e Investigaciones T
ecnicas de Gipuzkoa (CEIT) and Universidad de Navarra (TECNUN),
Paseo de Manuel Lardiz
abal 15, 20018 San Sebasti
an, Basque Country, Spain
Received 9 February 2004; received in revised form 19 April 2004; accepted 21 April 2004
Available online 18 May 2004

Abstract

A cold rolled low carbon steel has been annealed at low temperatures (300–500
C) in order to promote recovery without in-
teraction with recrystallization. It has been shown that the recovery process can be monitored by means of non-destructive magnetic
techniques using coercive field (Hc ) measurements. Transmission electron microscopy and electron backscattering diffraction
(EBSD) observations were carried out in order to investigate the microstructural changes associated with the measured decrease in
Hc produced by the recovery. The EBSD image quality maps show that on a microstructural scale, the cold work stored energy is
distributed heterogeneously among different texture components, in agreement with previous studies. Recovery mainly affects the
{1 1 1} c-fibre orientations, leading to the formation of well defined subgrains.
 2004 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

Keywords: Low carbon steel; Annealing; Recovery; Coercivity; Subgrain growth

1. Introduction loop. In this relationship dislocations play an important

role via magnetostrictive coupling and the stress field of
It has been firmly established that certain mechanical defects [4].
and magnetic properties of ferritic steels are closely re- Several theoretical statements have been made [5–7]
lated. Mechanical properties such as hardness, strength, relating the effect of dislocation density on coercive field
ductility and strain hardening are structure sensitive and quantitatively. Reported experimental metallographic
are directly related to the microstructure, composition results, supported by transmission electron microscopy,
and processing history. These same parameters also confirm some of the theoretical statements [8,9]. Theo-
control the magnetic characteristics of ferromagnetic retical considerations indicate that the coercive field Hc is
materials [1–3]. Nowadays, there is a growing interest in proportional to the square root of the dislocation density
the characterization of microstructures by magnetic pffiffiffi
Hc / k1 q; ð1Þ
non-destructive techniques. Structure sensitive magnetic
macroscopic properties include coercive field, perme- where k1 is a constant. It has also been shown that the
ability, residual induction and power loss derived from a coercive field is further affected by the distribution or
hysteresis loop. The microstructure of polycrystalline arrangement of dislocations [4,10].
ferromagnetic materials determines the magnetic do- The Taylor relation defining the contribution of dis-
main patterns and the magnetization processes and locations to the flow stress is given by
thereby the characteristic parameters of the hysteresis pffiffiffi
r
ry ¼ aMGb q; ð2Þ
*
Corresponding author. Tel.: +34-943-212800; fax: +34-943-213076. where a depends on the material, M is the Taylor factor,
E-mail address: [email protected] (A. Mart
ınez-de-Guerenu). G is the elastic shear modulus, b is the Burger’s vector

1359-6454/$30.00  2004 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.
doi:10.1016/j.actamat.2004.04.019
3658 A. Mart
ınez-de-Guerenu et al. / Acta Materialia 52 (2004) 3657–3664

and ry is the yield stress of the fully-recrystallized ma- The early work by Michalak and Paxton [15] analyzes
terial. Therefore, accepting this empirical relation and recovery taking place in zone-refined iron after strains of
the theoretical Eq. (1), the following linear relationship 5% and 15%. It shows that recovery is a logarithmic
is verified between the coercive field and the flow stress function of the annealing time up to the maximum
k1 fraction of recovery achievable at each temperature, and
Hc / ðr
ry Þ / k2 ðr
ry Þ; ð3Þ that increasing the amount of pre-strain increases the
aMGb
rate of recovery but not the final fraction of recovery.
where k2 is another constant. A number of authors have Another important result in this same study is that
reported this linear relationship for samples plastically carbon in solution segregating to the dislocations lowers
deformed [7,10]. recovery. This is in agreement with the results obtained
On the other hand, the coercive field is also strongly by other authors [16]. In commercial low carbon steels,
affected by the grain size of the material. Both variables it is expected that enough carbon remains in solution
are related by the following relationship [11,12]: and consequently, inhibits to a certain extent recovery.
const However, most research has concentrated on aluminium
Hc / ; ð4Þ
dk [17] or on iron [17,18], but information is lacking con-
where dk is the grain diameter. This equation is under- cerning recovery in commercial grades.
stood theoretically if the grains are regarded as carriers In the present study, the evolution of the coercive
of magnetization nuclei, the diameter of which is de- field, Hc , during the annealing of a cold rolled extra low
termined by the grain diameter. Changes being carbon steel has been studied. The annealing tempera-
produced on the grain size also have an effect on the tures have been mostly selected in order to avoid
yield stress of the material, through a Hall–Petch type interaction with recrystallization. Microstructural char-
relation [13] acterization carried out through the use of transmission
electron microscopy (TEM) and electron backscattering
ky diffraction (EBSD) techniques has allowed the effect of
ry ¼ ri þ pffiffiffiffiffi ð5Þ
dk the recovery on the microstructure to be studied and to
distinguish between different orientations. The recovery
with ri the friction stress and ky a constant.
kinetics, monitored by the progress of the coercive field,
Low carbon steel sheets for packaging or other
have been modelled using several formulas reported in
applications, are usually cold rolled and subsequently
the literature. This analysis and the equations deduced
annealed. During the annealing, the cold rolled sub-
describing the recovery kinetics for ELC-steels are
structure experiences recovery and recrystallization
reported in part II of this work [19].
processes. Recovery involves both the rearrangement of
dislocations into low energy configurations and the an-
nihilation of dislocations [14]. Recrystallization leads to
2. Experimental procedure
the suppression of dislocations through the nucleation
of defect free volumes and the migration of the inter-
An extra low carbon steel, with the composition
faces produced. The deformation dislocation substruc-
shown in Table 1, was used in this study. The steel was
ture provides the driving pressure for recrystallization,
industrially produced and cold rolled to a final thickness
FR , which can be expressed as
of 0.3 mm through a reduction of 84%. The samples
FR ¼ 12Gb2 q ð6Þ were isothermally annealed in the laboratory, following
annealing cycles performed using a device specially de-
with q the dislocation density, driving the migration of signed for continuous annealing simulation. The sam-
the recrystallization front into the deformed areas. ples were heated up in an argon atmosphere, at a rate of
The mechanisms behind the formation of a recrys- 20
C/s to a temperature within the range of 300–600
C,
tallized nucleus within the deformed structure are not and were hold at this temperature for annealing times in
even now fully understood, but there is some general the range of 1–48,880 s (13.6 h). The annealing cycles
agreement concerning the necessity of some degree of were interrupted and the samples quenched in helium to
recovery which allows the substructural changes re- room temperature at a rate close to )60
C/s.
quired to activate viable nuclei for recrystallization. A system designed and constructed at the authors’
However, extended recovery taking place within the laboratory was used for magnetic measurements of the
deformation structure decreases the driving pressure
for nucleation and for grain boundary migration and
Table 1
consequently retards recrystallization. Both recovery Chemical composition of the steel
and recrystallization are governed by different kinetics
C (%) Mn (%) Al (%) N (%) P (%) Si (%)
and consequently the extent of overlapping depends
0.03 0.19 0.13 0.0035 0.012 0.01
on the annealing temperature and on the material.
 A. Mart
ınez-de-Guerenu et al. / Acta Materialia 52 (2004) 3657–3664 3659

ELC steel. The magnetic measurements were made at 80

room temperature once the samples were removed from 75
the annealing simulation device. Near saturation AC 70
major hysteresis loops were recorded for each annealed

HR (30T)
65
sample. The hysteresis curves were acquired at 1 Hz 60
using an encircling coil wound around the samples. 55
A magnetic field strength of about 4,100 A/m, which was 50 Cold rolled 300 ºC
sufficiently high to saturate the samples, was applied. 400 ºC 450 ºC
45
The tangential magnetic field strength, H t , was deter- 500 ºC 600 ºC
40
mined by a small Hall probe placed at the surface of the 10 1000 100000
steel samples and the induction BðtÞ was derived by the Annealing time (s)
integration of the induced voltage on the encircling coil.
The coercive field value, Hc , was extracted from each Fig. 1. Evolution of the hardness as a function of the annealing time
hysteresis loop as the value of the tangential field where and temperature.
the induction crosses zero. It was measured with an
accuracy of 5 A/m. equal to or lower than 500
C, even after long soaking
HR 30T hardness measurements were perfomed on times (13.6 h), the hardness of the steel is almost in-
the sheet plane of the annealed samples, while Vickers sensitive to the annealing time. Vickers hardness tests
hardness measurements were carried out on the trans- carried out on the transverse section of the sheet give the
verse section. Some samples were cut, polished and same type of result. As an example, Hv for the as-cold
etched in nital 2% for observation by optical micros- rolled material is 214 and after a 4.8 h holding time at
copy. Samples for the cold rolled and the annealed 500
C it has only decreased to 204 Hv. For these low
conditions, were additionally polished in a colloidal temperature annealings, the microstructural observa-
silica solution for EBSD analysis. EBSD scans were tion, Fig. 2(a), reveals the initial deformed grains elon-
carried out on a Philips XL30 SEM equipped with a gated in the rolling direction without any signs of
TSL MSC 2200 EBSD system. All the observations were recrystallization. However, for the annealing carried out
performed on the sample-rolling plane. at 600
C, after about a 10 s soaking, the hardness ex-
Since the image quality indices determined by EBSD periences an important drop, which is the result of re-
analysis can be affected by changing conditions in the crystallization taking place, as can be seen in Figs. 2(b)
microscope or video processing (such as simply chang- and (c). The full softening of the steel is completed for a
ing the contrast and brightness), a normalized image soaking time of about 400 s at this same temperature.
quality index, QNi , has been calculated for each scanned As the hardness tests in the present steel do not give
area as any information about a possible recovery taking place
Qi
Qmin during the annealing carried out at the lowest temper-
QNi ¼ ; ð7Þ atures, magnetic methods were applied in order to get a
Qmax
Qmin
higher degree of resolution to investigate this phenom-
where Qi is the image quality index for the texture enon. In Fig. 3, the evolution of the coercive field has
component i and Qmin and Qmax are the minimum and been plotted as a function of the annealing time in linear
maximum image quality indices of each scanned area. In and logarithmic scales, for the different annealing tem-
order to get a unique value for QNi for each sample, peratures. The results are expressed as the quotient be-
the QNi -s calculated for each of the scans have been tween the instantaneous value of the coercive field, Hc ,
averaged. and that of the cold rolled specimen prior to the an-
The thin-foils for TEM observation were prepared by nealing treatment, Hc def . These graphs show that for the
electropolishing in a 5 pct perchloric acid, 25 pct glyc- annealing temperatures in the range 300–400
C, the
erol and 70 pct ethanol solution, at an applied current of coercive field decreases progressively with the soaking
80 mA. The study of the thin foils was carried out on time. For higher temperatures, in the range 450–500
C,
a PHILIPS CM12 electron microscope operating at an important drop in Hc is observed for short annealing
100 kV. times, followed by a continuous decrease over time to
reach saturation after long times. The microstructural
observations carried out on these samples, see example
3. Results in Fig. 2(a), showed no sign of recrystallization after
these low temperature annealing treatments (between
The graph in Fig. 1 shows the evolution of the 300 and 500
C). These results clearly reveal that only
hardness (HR 30T) during the relatively low tempera- recovery contributes to the variations observed at these
ture annealing treatments carried out in the present temperatures in the magnetic measurements. It should
work. It can be seen that for annealing temperatures be noted that, the decrease in Hc is not accompanied by
3660 A. Mart
ınez-de-Guerenu et al. / Acta Materialia 52 (2004) 3657–3664

Fig. 2. Optical micrographs showing the microstructure of the sample after different annealing treatments: (a) 500
C–4.8 h; (b) 600
C–11 s;
(c) 600
C–31 s; (d) 600
C–1.5 h.

a significant modification of the hardness in the samples, Microbands can be distinguished on the rolling plane to
as can be seen when comparing the graphs in Figs. 1 and 3. run nearly perpendicular to the rolling direction. Some
For the annealing carried out at 600
C, Hc experiences transition bands also become evident, such as those
the same type of initial drop, described for the rest of the identified (TB) in the figure.
tests and subsequently exhibits a continuous decrease On the contrary, it is quite difficult to get a precise
for annealing times in the range 4–30 s. It should be picture within the c-fibre grains. The best picture of the
noted that Hc saturates and remains nearly constant in substructure being present within these areas is obtained
the range 30–420 s, slightly increases afterwards and fi- from the image quality maps, but care is required when
nally decreases to a level of about 0.33 Hc def for an processing the results. Any cleaning process requires
annealing time of 1.5 h. It should be mentioned that in criteria to be applied and these significantly alter the
the range of the times leading to the recrystallization of final result. This is why the image in Fig. 4 corresponds
the cold rolled steel sheet, see Figs. 1 and 2, Hc remains to that obtained directly from the equipment without
almost constant. any type of cleaning. Within these areas, small crystal-
TEM and EBSD observation techniques were applied lites with a diameter of around 1 lm can be distin-
in order to investigate the microstructural changes as- guished. These have a relatively high image quality
sociated with the observed variation in the coercive field. index inside, but are surrounded by dark lines in which
The EBSD image quality map obtained on the rolling the Kikuchi patterns are not properly defined, leading to
plane for the as-cold rolled material is shown in Fig. 4. low image quality indices.
The rolling direction has been placed parallel to the TEM observations carried out on the cold rolled
vertical axis. In the scanned area different regions can material show the complex dislocation microstructure
clearly be distinguished according to the image quality produced by the cold reduction applied to the present
indices. The dark areas (low confidence index) in the steel. The area shown in Fig. 5 corresponds to a c-fibre
image quality map correspond to c-fibre {1 1 1} grains. orientation. Sub-micron cells (around 0.3 lm in size)
On the other hand, lighter regions mainly correspond to surrounded by dislocation walls can be observed on its
a-fibre h1 1 0i grains. All the orientations belonging to interior.
the a-fibre have relatively high confidence indices and, as The EBSD image quality picture in Fig. 6 corre-
a consequence, some microstructural features are clearly sponds to the sample annealed for 4.8 h at 500
C. As
evident within the grains having these orientations. before, by simple observation, the a-fibre regions (in
 A. Mart
ınez-de-Guerenu et al. / Acta Materialia 52 (2004) 3657–3664 3661

1.1
1.0
0.9
0.8
Hc / Hc_def

300 ºC
0.7
0.6 400 ºC
0.5 450 ºC
0.4 500 ºC
0.3 600 ºC
0.2
0 20000 40000 60000
(a) Annealing time (s)

1.0
0.9
Fig. 5. TEM image showing the cell substructure of a region with a
0.8 {1 1 1} orientation belonging to the as-cold rolled material.
300 ºC
Hc / Hc_def

0.7
0.6 400 ºC within the {1 1 1} regions. The dislocations arrange into
more perfect walls and some coarsening of the sub-
0.5 450 ºC
structure seems to have taken place, compared to the as-
0.4 500 ºC cold rolled sample, leading to subgrain sizes in the range
600 ºC
0.3–0.6 lm.
0.3
0.2
10 1000 100000
(b) 4. Discussion
Annealing time (s)

Fig. 3. Evolution of the relative coercive field, as a function of the The results obtained in the present work indicate that
annealing time and temperature on: (a) a linear; (b) a logarithmic time the non-destructive testing method used here, based on a
scale.
particular magnetic characterization technique, is espe-
cially useful to investigate the recovery kinetics. Hard-
light) can clearly be distinguished from c-fibre zones ness measurements, which were used to investigate
(dark zones). The former are quite similar to those ob- recovery in 80% cold rolled iron [16] are in the present
served for the as-cold rolled material. The main changes case unable to detect any significant change in the ma-
produced by the annealing seem to concentrate on the terial during the low temperature annealing, see Fig. 1.
c-fibre regions that appear increasingly subdivided into This is a general trend in ELC steels for which, most of
small crystallites about 1–1.5 lm in diameter. The TEM the times, no significant recovery of the hardness is ap-
observations reveal that during recovery, the complexity parent prior to recrystallization [20]. The presence of
of the microstructure still remains, even if a transition to carbon in solution segregating to the dislocations and
a subgrain structure is clearly observed, see Fig. 7, pinning them may be a factor contributing to such

Fig. 4. Image quality and orientation map corresponding to an area of the as-cold rolled sample on the rolling plane.
3662 A. Mart
ınez-de-Guerenu et al. / Acta Materialia 52 (2004) 3657–3664

Fig. 6. Image quality map obtained on the rolling plane for the sample annealed for 4.8 h at 500
C.

there appears to be a ‘‘saturation’’ value of recovery, at

least for the highest temperature, beyond which the
coercive field does not change significantly with time.
These observations are coherent with reported annealing
studies on polycrystalline iron and aluminium, in which
the recovery kinetics were quantified using the flow
stress [15,28], on silicon–iron single crystals [21] and on
a cold-rolled Ti–Nb stabilized interstitial-free steel [22]
by X-ray peak resolution, on oligocrystalline tantalum
by microhardness measurements [26] and on a Fe-3.25%
Si alloy by coercive field and initial susceptibility mea-
surements [27]. The observed decrease in Hc may be
directly related, through Eq. (1), to the annihilation of
dislocations taking place during the annealing. The de-
crease in the saturation level of Hc with temperature can
Fig. 7. TEM image of the subgrains developed within a {1 1 1} region be understood as an increasing level of recovery being
during the isothermal annealing at 500
C for a holding time of 4.8 h. achievable when the annealing temperature is increased,
which seems reasonable.
behaviour [15]. According to the results in [16], in carbu- As soon as the recrystallization activates at the an-
rized iron the maximum hardness recovery at 400
C is nealing temperature of 600
C, see Figs. 1 and 2, the
about 22%, as compared to 46% in pure iron. In hardness curve reflects the consequent softening of the
agreement with this, in stabilized IF steels hardness re- material. However this mechanical softening is not ac-
covery has a higher contribution during annealing than companied by a decrease in Hc . According to Eq. (1), it
in ELC steels [20]. In addition to carbon, the presence of would be expected that the decrease in the dislocation
other alloying elements, able to segregate to the dislo- density produced by the movement of the recrystalliza-
cations or to interact with carbon may also have an tion front through the material should lead to decreasing
effect on recovery. As a consequence, in the ELC steels Hc values. But instead, Hc is observed to stagnate during
the hardness recovery is probably under the resolution recrystallization. This apparent discrepancy can be ex-
limit of the hardness measurements, but can be investi- plained by taking into account the fact that the coercive
gated by more sensitive techniques. field is highly dependent on the grain size, see Eq. (4). The
Tensile tests [15], X-ray techniques [21,22], thermo- initial grain size of the present material is quite large
electric power measurements [23], microhardness mea- (60 lm) and remains constant during recovery. However,
surements [24–26] and magnetic methods [27] have been during recrystallization the nucleation of new strain free
employed to investigate recovery. In the present case, grains leads to an effective grain refinement (see Fig. 2).
the recovery process has been monitored using coercive Consequently, as a result of the type of dependence of Hc
field measurements. According to Eqs. (1)–(3), Hc is on both the dislocation density and the grain size, the
directly related to the dislocation density and accord- decrease in the mean grain size, will produce an increase
ingly is a proper variable to carry out this type of in Hc , which seems to compensate for the expected de-
analysis. crease produced by the dislocation annihilation taking
The curves in Fig. 3 are typical in a recovery process place during recrystallization.
as they show a high initial rate of change of the mea- According to these results, it seems clear that the
sured property (Hc in this case), which progressively falls technique employed can be applied with great success to
as the isothermal annealing time increases. Moreover, investigate recovery, during which the grain structure
 A. Mart
ınez-de-Guerenu et al. / Acta Materialia 52 (2004) 3657–3664 3663

remains constant and microstructural changes only af- important local lattice misorientations, define small
fect the cold rolling dislocation substructure produced crystallites exhibiting high image quality indices, Fig. 4.
inside them. The TEM image shows a cellular type of substructure
The recovery of the cold-worked microstructure pro- within the {1 1 1} regions, see Fig. 5, but the size of the
duced by plastic deformation in iron-based alloys has cells is significantly lower than the size of the crystallites
been reported to occur by annihilation of point defects, observed by EBSD. The mean diameters obtained from
annihilation, migration, and rearrangement of disloca- EBSD are about three times larger than the cell size
tions and the formation and growth of subgrains [29]. determined by TEM. This is probably related to the
Nes [17] has recently revisited several theories concerning resolution of the EBSD, which is not able to reveal the
recovery and has proposed different models to describe lowest boundaries forming the cellular substructure.
such a process. For zone-refined iron the author pro- The graph in Fig. 8 shows that the normalized image
poses that recovery can be treated as the annealing out of quality indices are higher in the recovered material than
statistically stored dislocations and the subgrain growth. in the cold rolled sample. This is another consequence of
All these treatments of recovery are statistical in na- the effect of recovery. It is observed that the main in-
ture and do not take into account the changes being crease in the normalized image quality affects the ori-
produced on a local scale. It is well known that the cold entations belonging to the c-fibre, while the a-fibre is
work energy is distributed heterogeneously among the much less affected by recovery. This is in agreement with
grains, depending on their crystallographic orientation. a rate of recovery being higher for the regions with a
The EBSD results in Fig. 4 are a clear example of this higher stored energy, which seems reasonable.
effect. In fact, the lowest possible stored energy is lo- The TEM images show experimentally, that recovery
cated near to the Goss {1 1 0}h0 0 1i and the second leads within the {1 1 1} orientations to the progressive
lowest at the rotated cube {0 0 1}h1 1 0i [30]. As sum- sharpening of the cell boundaries and that the latter
marized by Raabe [31], experimental work on both evolve into a well defined subgrain substructure, Fig. 7.
single and polycrystalline specimens shows that On the EBSD images, the crystallites become more
{0 0 1}h1 1 0i grains rarely develop texture gradients or clearly defined, leading progressively to the formation of
develop deformation bands. In the present work, Fig. 8, a network of relatively high misoriented boundaries [20].
the normalized image quality index of these grains is Again, the subgrain size determined by TEM is signifi-
significantly higher than that of the rest of the grains, cantly lower than the approximate mean diameter of the
which is an indication of a lower degree of stored energy crystallites revealed by EBSD. The coarsening of the
associated with these orientations. The graph in Fig. 8 substructure seems to have taken place during recovery.
shows that the normalized image quality decreases along The subgrain growth has previously been reported for
the a-fibre, as the orientation moves from the rotated low carbon steels to activate as a recovery mechanism
cube to the c-fibre. The poorer images obtained from the [32]. However, only abnormal subgrain growth seems to
EBSD scans are due to a high degree of distortion in the produce recrystallization nuclei. These latter have been
Kikuchi patterns produced for orientations belonging to observed by EBSD to form within the network of
the c-fibre. As a result, the interior of the grains cannot crystallites being present inside the {1 1 1} regions [20].
be properly resolved using the EBSD technique. Regions
of very poor image quality, indicating the presence of
5. Conclusions
Normalized image quality index, QNi

0.7
Cold rolled steel • The coercive field measurements have demonstrated
0.6 that they satisfactorily monitor the recovery during
Annealed steel
0.5 low temperature annealing of a cold rolled low car-
bon steel. This technique has the advantage of being
0.4
non-destructive in nature and of giving quantitative
0.3 information about a phenomenon even when
0.2 other techniques like hardness do not have enough
sensitivity.
0.1 • The grain refinement taking place during recrystalli-
0 zation masks the influence on the coercive field mea-
surement of the decrease in the dislocation density
]

0]

1]

]
0]

3]

]
11
10

10

10

31

32

12
1
1

2
)[0
)[1
)[1

)[1

)[1

)[1

)[2

)[1

)[1

)[1

)[1

produced by the movement through the sample of

01

13

12

23

11

11

11

11

11

11

11
(0

(1

(1

(1
(1

(1

(2

(1

(1

(1

(1

the recrystallization front. This effect allows the

α-fiber γ-fiber
recovery and the recrystallization processes taking
Fig. 8. Normalized image quality indices corresponding to the cold place during annealing to be distinguished by using
rolled steel and to the sample annealed for 4.8 h at 500
C. coercive field measurements.
3664 A. Mart
ınez-de-Guerenu et al. / Acta Materialia 52 (2004) 3657–3664

• EBSD scans and TEM observations clearly show that [10] Asti
e B, Degauque J, Porteseil JL, Vergne R. IEEE Trans Magn
the cold work stored energy is distributed heteroge- 1981;MAG-17(6):2929.
[11] Mager A. Ann Phys 6 F Lpz 1952;11:15.
neously among different texture components. The [12] Degauque J, Asti
e B, Porteseil JL, Vergne R. J Magn Magn Mater
c-fibre grains store a relatively high energy and ac- 1982;26:61.
commodate it into a cellular substructure. During re- [13] Gladman T. The physical metallurgy of microalloyed steels.
covery, the latter evolves towards a subgrain London: The Institute of metals; 1997. p. 39.
substructure of increasing size. A network of rela- [14] Humphreys FJ, Hatherly M. Recovery after deformation. In:
Recrystallization and related annealing phenomena. Oxford:
tively high angle boundaries is observed through Pergamon Press; 1996. p. 127 [Chapter 5].
EBSD to form within the c-fibre grains. This network [15] Michalak JT, Paxton HW. Trans Met Soc AIME 1961;221:850.
cannot directly be related, at this stage, with the sub- [16] Venturello G, Antonione C, Bonaccorso F. Trans Met Soc AIME
grains observed by TEM. 1963;227:1433.
• The EBSD results reveal that the rate of recovery into [17] Nes E. Acta Metall Mater 1995;43:2189.
[18] Leslie WC, Michalak JT, Aul FW. The annealing of cold-worked
the c-fibre is higher than that for other a-fibre iron. In: Spencer CW, Werner FE, editors. Iron and its dilute solid
orientations. solutions. New York: Interscience; 1963. p. 119.
[19] Mart
ınez-de-Guerenu A, Arizti F, Guti
errez I. Recovery
during annealing in a cold rolled low carbon steel. Part II:
Modelling the kinetics. Acta Mater 2004, doi:10.1016/j.actamat.
Acknowledgements
2004.04.020.
[20] Bocos JL, Novillo E, Petite MM, Iza-Mendia A, Gutierrez I.
The financial support of the MCYT of Spain un- Metall Mater Trans A 2003;34A:827.
der the project: FIT-020100-2003-3 is gratefully [21] Hu H. Annealing of silicon–iron single crystals. In: Himmel L,
acknowledged. editor. Recovery and recrystallization of metals. New York:
Science publishers; 1963. p. 311.
[22] Mukunthan K, Hawbolt EB. Metall Mater Trans A 1996;27A:
3410.
References [23] Khelfaoui F, Thollet G, Guenin G. Mater Sci Eng A 2002;338:
305.
[1] Jiles DC. NDT Int 1988;21(5):311. [24] Chen SP, Kuijpers NCW, van der Zwaag S. Mater Sci Eng A
[2] Bhattacharya DK, Jayakumar T, Baldev R. In: Murphy CRL, Raj 2003;341:296.
B, Prabhakar O, Srinivasulu A, editors. Non destructive evalua- [25] Ferry M, Munroe PR. Mater Sci Eng A 2003;358:142.
tion and quality assurance. Bangalore, India: Interline Publishing; [26] Hupalo MF, Sandim HRZ. Mater Sci Eng A 2001;318:216.
1993. p. 110. [27] Chaudhary LN, Qureshi AH. IEEE Trans Magn 1971;7(3):
[3] Mikheev MN, Gorkunov ES. Sov J NDT 1981;17(8):579. 560.
[4] Degauque J. Solid State Phenom 1994;35–36:335. [28] Kwiecinski J, Wyrzykowski JW. Acta Metall Mater 1993;41(11):
[5] Vicena F. Czech J Phys 1955;5:480. 3089.
[6] Kersten M. Z Angew Phys 1956;8:496. [29] Byrne JG. Recovery. In: Recovery, recrystallization and grain
[7] Tra€uble H. The influence of crystal defects on magnetization growth. New York: The Macmillan Co.; 1965. p. 37 [Chapter 3].
processes in ferromagnetic single crystals. In: Berkowitz AE, [30] Ray RK, Jonas JJ, Hook RE. Int Mater Rev 1994;39:129.
Kneller E, editors. Magnetism and metallurgy, vol. 2. New York: [31] Raabe D. Scr Metall Mater 1995;33:735.
Academic Press; 1969. p. 621. [32] R
egl
e H. In: Gottstein G, Molodov DA, editors. Proceedings of
[8] Qureshi AH, Chaudhary LN. J Appl Phys 1970;41:1042. the First Joint International Conference on Recrystallization and
[9] Sternberk J, Kratochv
ılov
a E, Hreb
ık J, Gemperle A. Phys Stat Grain Growth, vol. 2. New York: Springer-Verlag; 2001.
Sol (a) 1983;79:523. p. 707.