ISIJ International, Vol. 44 (2004), No. 3, pp.

618–622

Decarburization Kinetics during Annealing of a Semi-processed

Electrical Steel

Kleiner Marques MARRA, Evandro de Azevedo ALVARENGA and Vicente Tadeu Lopes BUONO1)

Usiminas Research and Development Center, Rod. BR 381 s/n, 35160-900 Ipatinga, MG-Brazil.
1) Federal University of Minas Gerais, Department of Metallurgical and Materials Engineering, Rua Espirito Santo, 35, 30160-
030 Belo Horizonte, MG-Brazil.
(Received on July 15, 2003; accepted in final form on October 28, 2003 )

Semi-processed, non-oriented electrical steels used in motor and transformer cores have their magnetic
properties optimized by means of a heat treatment performed under a moistened nitrogen atmosphere con-
taining small amounts of hydrogen. The oxidation of carbon on the steel surface by the oxidizing gases pre-
sent in this atmosphere decreases the carbon content of the steel. The consequent reduction in the amount
of soluble carbon and in the volume fraction of carbides reduces magnetic losses and the susceptibility of
the steel to magnetic aging. In this work, a cold-rolled, semi-processed electrical steel was submitted to six
thermal treatments covering the most usual conditions employed in industry, combining two types of at-
mosphere (mixtures of nitrogen with 5 % and 10 % H2) and three dew points (10, 20 and 30°C). The treat-
ments were carried out in a pilot electrical oven especially adapted to simulate the thermal cycle and the
moistened atmospheres employed in industry. The results obtained show that, at the beginning of the
process, the decarburization rate is controlled by the chemical reaction of carbon on the steel surface. Later,
an oxide layer may build up, depending on the humidity level of the atmosphere, reducing the rate of decar-
burization.
KEY WORDS: semi-processed electrical steels; decarburization rate; carbon oxidation; decarburizing atmo-
sphere.

due to the higher iron concentration on the steel surface,

1. Introduction
both elements react with the humidity. The hydrogen, pre-
Semi-processed, non-oriented electrical steels are cold sent in the atmosphere has an important role in reducing the
rolled products used mainly as core materials of motors and iron oxides formed during the treatment and allows a pro-
small size transformers for home appliances. A low core longed contact between the atmosphere and the steel, thus
loss and a high permeability are the magnetic properties re- extending the decarburization process.
quired to improve the efficiency of this electric machinery. The knowledge of the parameters controlling the kinetics
In order to present suitable magnetic properties, these steels of decarburization is thus of fundamental importance for
must be submitted to a heat treatment, during which the fol- the producers an the users of this class of steel, so that ap-
lowing events take place: carbon content reduction, stress propriate heat treatment schedules can be established to ob-
relief, ferritic grain growth and surface oxidation. The tain a final product with optimized magnetic properties.
treatment involves two important stages. In the first one, the Analysis of the phenomena taking place during decarbur-
steel is heat treated at high soaking temperatures, between ization of steel leads to the conclusion that the kinetics of
700 and 850°C, under a moistened atmosphere. Such con- the process depends fundamentally on four reaction rates,
ditions promote decarburization, stress relief and grain associated with: (i) diffusion of carbon atoms from bulk to
growth. The second stage is accomplished under an oxidiz- the steel surface; (ii) gas adsorption (water vapor and hy-
ing atmosphere, at temperatures between 500 and 600°C, drogen) on the steel surface; (iii) oxidation of carbon; and
which results in the build up of an iron oxide layer on the (iv) formation of iron oxides. It is important to note that all
steel surface. This heat treatment is usually performed by these processes are thermally activated and should be accel-
the semi-processed steel users. erated by an increase in temperature.
During the heat treatment, decarburization occurs mainly The carbon diffusion rate in thin steel strips during de-
through the reaction of the humidity (H2O), in the furnace carburization was studied by Tomilin et al.1) Considering
atmosphere, with carbon atoms on the steel surface. The the carbon content in the steel surface equals to zero, these
oxidation reaction involves both the iron and carbon atoms, authors developed Eq. (1), giving the average carbon con-
the latter being diffused from the bulk to the steel surface. tent C of the steel at a time t,
The carbon oxidation should prevail, since it involves larger
changes in the free energy than iron oxidation. However,

© 2004 ISIJ 618

 ISIJ International, Vol. 44 (2004), No. 3

∞
8Co
C ∑ (2n1) π
0
2 2
exp{
Dt[(2n1)π /e]2} .......(1)

in terms of the initial carbon content Co, the diffusion coef-

ficient of carbon D at the soaking temperature and the strip
thickness e. Equation (1) is the solution of the differential
equation expressing Fick’s second law for diffusion, in this
particular case.
Regarding the adsorption of water vapor and hydrogen
on the steel surface, it is well known that this reaction is
very fast2) and certainly does not control the decarburiza-
tion process. The third reaction, carbon oxidation at the
steel surface, is a thermally activated process obeying
Arrhenius law,2) that is, the reaction rate increases when the Fig. 1. Schematic representation of the decarburization heat
temperature and/or concentration of C and H2O are in- treatments employed.
creased. Finally, the water vapor reacts also with iron, lead-
ing to scale formation, which may isolate the steel from the
Table 1. Volume concentrations of water vapor and hydrogen
oxidizing atmosphere and thus inhibit further decarburiza- in the furnace atmosphere and the PH2O/PH2 ratios
tion. The slowest reaction among those previously men- used in the decarburization treatments.
tioned determines the overall decarburization rate.
The aim of this work was to identify the phenomenon
controlling the kinetics of decarburization in a typical semi-
processed electrical steel, in order to provide relevant infor-
mation for the users of this type of steel regarding the opti-
mization of magnetic properties during the final annealing
cycle applied to the punched lamellae used as core materi-
als.

2. Experimental
The semi-processed, cold-rolled steel sheet studied,
0.50 mm in thickness, had the analyzed chemical compo-
sition (in mass percent) Fe–0.0052C–0.47Mn–0.73Si–
0.015P–0.007S–0.25Al–0.03Sb–0.0035N. Six thermal a hot water column. A dew point measurer probe (Omega
treatments covering the most usual industrial conditions RH70) was placed in the entrance of the chamber, to con-
have been employed, at the soaking temperature of 760°C, trol the concentration of H2O in the atmosphere. Three val-
combining three dew points (10, 20 and 30°C) and two ues of dew point, 10°C, 20°C and 30°C, were aimed, cover-
types of atmosphere (mixtures of nitrogen with 5% and ing a wide range of heat treating conditions and corre-
10% H2). The decarburization times employed varied from sponding to volume concentrations of, respectively, 1.2%,
5 min to 3 h. The simulated annealing cycle consisted only 2.3% and 4.2% H2O. These values were kept constant dur-
of the decarburization stage, as shown in Fig. 1. ing the treatments by continuous monitoring the dew point
It is important to mention that the conditions above cover and correcting it by changing the height of the water col-
the majority of the decarburization step of the heat treat- umn or the water temperature in the bubbling system. The
ment procedures applied to semi-processed electrical steels. humidity degrees of the atmospheres employed, that is, the
Only the soaking temperature was kept constant at 760°C, ratios between the partial pressures of H2O and H2, which
although the producers of core materials for motors and are shown in Table 1, were evaluated by the ratios between
transformers employ a broad range of soaking tempera- the volume concentrations of H2O and H2 in the treatment
tures, usually between 700°C and 850°C. The influence of atmospheres.
the soaking temperature on the other processes taking place Test specimens for decarburization consisted of rectan-
during the decarburization treatment, that is, stress relief gular strips 280 mm in length, 30 mm wide and 0.5 mm in
and grain growth, is not relevant in the present work. thickness. They were placed in the decarburization chamber
Nevertheless, all phenomena involved in the process are as packages of seven strips, in an arrangement similar to
thermally activated, meaning that decarburization, stress re- that employed in the industry. The strips at the top and the
lief and grain growth occur faster when the soaking temper- bottom of the package were discarded. Carbon concentra-
ature is increased. tion was determined on the remainder strips using an inte-
The treatments were carried out in a pilot electrical fur- grated C and S determination system (Leco 444LS), cali-
nace especially adapted to simulate the thermal cycle and brated for low carbon steels, whose detection limit is
the moistened atmospheres. Pure and dried N2 and H2 gases 10
1 ppm in weight of C. The carbon contents obtained cor-
were used to produce mixtures with volume concentrations respond to the average of six individual measurements per-
of 5% and 10% H2 which, prior to injection into the fur- formed on different regions of the analyzed strips.
nace chamber, were humidified in an bubbling system with

619 © 2004 ISIJ

 ISIJ International, Vol. 44 (2004), No. 3

3. Results and Discussion

The evolution of the carbon content of the steel during
the decarburization treatments performed at 760°C is
shown in Fig. 2, for the six different atmospheres em-
ployed. The curve identified as “diffusion” in Fig. 2 shows
the expected reduction in the carbon content calculated
using Eq. (1), with D1.11710
4 m2 s
1. The Fourier se-
ries was truncated after the 25th term (n25), resulting in a
negligible error. The value of the diffusion coefficient at the
soaking temperature of 760°C was obtained using the clas-
sical expression for D, with a pre-exponential factor of
2.010
6 m2 s
1 and an activation energy of 84.2 kJ/mol.3)
According to results presented in Fig. 2, the decarburiza-
Fig. 2. Carbon content evolution as a function of the decarbur-
tion process happens in a much slower way than that pre-
ization time at 760°C for the ratios PH2O/PH2 indicated.
dicted by Eq. (1), thus indicating that carbon diffusion is
not the controlling mechanism for decarburization of semi-
processed electrical steels. These results also show that the
use of highly moistened atmospheres does not guarantee an Vfk[aC][PH2O](1
exp DG/RT ) ...............(4)
optimum decarburization condition. In Eq. (4), k is the reaction constant of (2) and the value of
The results obtained can be discussed by considering the (1
exp DG/RT) tends to 1 when the free energy tends to
decarburization chemical reaction, which is described by minus infinite.
Eq. (2). Two observations can be made about the decarburization
CH2O↔ COH2 .........................(2) of semi-processed steels. First, the partial pressure of water
vapor is constant in industrial furnaces, because its concen-
Two factors can slow down the decarburization chemical re- tration practically does not change during the decarburiza-
action rate: the build up of an oxide layer on the steel sur- tion treatment. Second, the carbon content of modern steels
face during the heat treatment; and the increase in the con- is very low, usually below 50 ppm, and in this case the car-
centration of CO close to the steel surface, due to the use of bon activity (aC) can be considered directly proportional to
stacked steel strips. The latter is known as the “gap effect”, the carbon content of the steel, according to Henry’s law.2)
and has been studied before by Inokuti,4) whose main con- Thus, the reaction rate depends directly on the carbon con-
clusion was that the decrease in the gap distance between tent and Eq. (4) can be written as:
two successive coils increases the partial pressure of CO in
the spacing between them. The effect of the stacking condi- Vk1[C] ..................................(5)
tion was not explicitly considered in the present work, but In this equation, k1 is a constant given by the following ex-
the observations of Inokuti4) are in agreement both with the pression:
theoretical expectation and the experience in industry: due
to the gap effect, decarburization of individual steel strips k1kC[PH2O](1
exp DG/RT) ..................(6)
occurs much faster than in stacked strips. However, the gap effect must be taken into account in the
Usually, the partial pressure of CO in electrical steel de- decarburization of semi-processed steels, since stacked
carburization furnaces is very low, whereas the partial pres- steel strips are used in the majority of commercial heat
sures of N2, H2O and H2 are kept higher and practically treatments. When stacked steel strips are heat treated, the
constant during the heat treatment. Therefore, considering CO concentration inside the steel strip packages is not neg-
the decarburization reaction of an individual strip, the ligible and reaches a pseudo-equilibrium state with H2O
change in the Gibbs free energy associated to the reaction is and H2. Considering Eq. (4) once again, DG must be taken
given by Eq. (3), as finite, since the CO pressure is no longer zero. As a con-
sequence, the reaction rate is smaller than the rate calculat-
 1 PCO PH2  ed by Eq. (5), because the factor (1
exp DG/RT ) becomes
∆G  RT ln   .....................(3)
 k e PH2O aC  smaller than 1. So, the reaction rate V can also be written as
a function of the carbon content, as given by Eq. (7), where
where R is the universal gas constant, T the absolute tem- the reaction constant k2 is smaller than its corresponding
perature, ke the kinetic constant of the reaction at equilibri- value in the case of an individual strip, k1.
um, aC is the carbon activity and PCO, PH2 and PH2O are the
partial pressures of CO, H2 and H2O, respectively. In the Vk2[C] ..................................(7)
particular case of decarburization of an individual strip, This reduction of the reaction constant actually takes place
PCO, tends to zero and consequently the associated change and is so significant that the chemical reaction may become
in free energy tends to minus infinite. slower than the carbon diffusion, as shown in Fig. 2. In this
According to the literature,2) the decarburization reaction case, the decarburization process is probably controlled by
is reversible. So the reaction rate should be given by Eq. the chemical reaction expressed in Eq. (2). In order to eval-
(4), where Vf is the speed of the reaction from left to right uate this possibility, one should recall that the reaction
in Eq. (2). speed, V, can be represented by the differential equation

© 2004 ISIJ 620

 ISIJ International, Vol. 44 (2004), No. 3

dC So, it can be inferred that, until the beginning of the iron

V 
 k 2C ............................(8) oxide building up on the steel surface, the kinetics of the
dt carbon reduction process is controlled by the chemical re-
The integration of this equation in time, between 0 and t, action expressed in Eq. (2). In a subsequent stage, after
results in Eq. (9), where Co is the initial carbon content and oxide formation, the decarburization rate is gradually
C is the carbon content at a time t. slowed down, as evidenced by the non-linear segments in
the curves shown in Fig. 3.
C  Table 2 shows the values of the kinetic constant k2 given
ln  o   k 2t ...............................(9) by the slopes of the linear portions of the curves plotted in
 C 
Fig. 3. As expected, the trend is that increasing the humidi-
To verify the validity of Eq. (9), the experimental results ty degree, the rate of the decarburization chemical reaction
shown in Fig. 2 were plotted in a graph of ln(Co /C) against is increased, as illustrated in Fig. 6. However, such a factor
t, shown in Fig. 3. It can be observed that, for low humidity
atmospheres, PH2O /PH20.24, a linear relationship is ob-
tained, demonstrating that, for these atmospheres, the de-
carburization process is controlled by the reaction ex-
pressed in Eq. (2). However, atmospheres of higher mois-
ture contents, PH2O /PH20.24, show, in addition to an initial
linear behavior, a regime in which the reaction rate continu-
ously slows down.
Observation of the surface of specimens submitted to the
decarburization treatments, as exemplified in Fig. 4,
showed that only for high moisture content conditions
(PH2O /PH20.24) there was a significant surface oxidation.
This was not the case for samples treated in atmospheres
where the humidity degree was lower than 0.24. Figure 5
shows the change in mass (average of three strips) inside Fig. 3. Logarithmic variation of the carbon content with the de-
the packages of the samples treated under different atmo- carburization time at 760°C for the ratios PH2O/PH2 indi-
spheres. cated.

Fig. 4. Scanning electron microscopy images of the surface of strips after 3 h of treatment at 760°C at the ratios PH2O/PH2
of (a) 0.12 and (b) 0.84.

Table 2. Values of the kinetic constant k2 obtained from the

linear portions of the curves shown in Fig. 3.

Fig 5. Change in mass of specimens treated for 2 h at 760°C

under the ratios PH2O/PH2 indicated.

621 © 2004 ISIJ

 ISIJ International, Vol. 44 (2004), No. 3

4. Conclusions
(1) It was demonstrated that the chemical reaction
CH2O↔COH2 controls, at least initially, the decarbur-
ization of semi-processed electrical steels. As a conse-
quence, the reduction of carbon content obeys a logarithmic
decay law, ln(Co /C)k2t, where Co is the initial carbon con-
tent and k2 is the reaction constant.
(2) The reaction constant k2 increases with the degree
of humidity, accelerating the initial stages of decarburiza-
tion, and decreases when the CO vapor pressure in the
vicinities of the steel strips is increased, slowing down the
reaction rate.
(3) In a subsequent stage, the iron oxide build up on
Fig. 6. Variation of the kinetic constant k2 with the ratio PH2O/PH2
the steel surface slows down the chemical reaction, which
in the linear portion of the curves shown in Fig. 3.
reduces the decarburization treatment effectiveness. There-
fore the increase of the moisture content does not assure a
may simultaneously advance the formation of the oxide more efficient decarburization, since oxide formation may
layer. Taking as an example the curves relative to the treat- be speeded up.
ments under the humidity degrees 0.42 and 0.84, Fig. 3
shows that, although the higher moisture content atmos- REFERENCES
phere presents initially a steeper curve, the decarburization 1) I. Tomilin, V. G. Shvartsman, A. G. Borisenko, A. G. Dukhynov and
attenuation started first for this condition when compared to A. G. Petrenko: Stahl Eisen (in English), 1 (1969), 586.
the moisture content 0.42. As a consequence, this last con- 2) P. W. Atkins: Physical Chemistry, 5th ed., Oxford University Press,
dition resulted in a more effective decarburization, because Oxford, (1994), 1200.
3) R. W. K. Honeycombe and H. K. D. H. Bhadeshia: Steels Micro-
the iron oxide layer build up was delayed.
structure and Properties, 2nd ed., Arnold, London, (1995), 1.
4) Y. Inokuti: Trans. Iron Steel Inst. Jpn., 15 (1975), 324.

© 2004 ISIJ 622