You are on page 1of 11

Journal of

J. Min. Metall. Sect. B-Metall. 47 (2) B (2011) 199 - 209 Mining and
Metallurgy

AustenIte GrAIn Growth CAlCulAtIon


of 0.028% nb steel

D. Priadi*,#, r.A.M. napitupulu*,** and e.s. siradj*

*Metalurgy and Material Engineering Department, Universitas Indonesia, Depok, Indonesia


**Mechanical Engineering Department, Universitas HKBP Nommensen, Medan, Indonesia

(Received 01 September 2010; accepted 26 January 2011)

Abstract

Modeling of microstructural evolution has become a powerful tool for materials and process
design by providing quantitative relationships for microstructure, composition and processing.
Insufficient attention has been paid to predicting the austenite grain growth of microalloyed steel and
the effect of undissolved microalloys. In this research, we attempted to calculate a mathematical
model for austenite grain growth of 0.028% Nb steel, which can account for abnormal grain growth.
The quantitative calculation of austenite grain growth generated from this model fit well with the
experimental grain growth data obtained during reheating of niobium steels. The results of this study
showed that increasing the temperature increases the austenite grain size, with a sharp gradient
observed at higher temperatures.

Keywords: Austenite; Grain growth; Calculation; Niobium steel.

1. Introduction boundary migration, a polycrystalline


material will evolve towards a single crystal.
It is well established that grain growth is In reality, this goal is rarely attained due to
the evolution of a microstructure due to the the unavoidable defects and impurities in
motion of grain boundaries driven by the materials, even for high purity zone refined
reduction in grain boundary interfacial metals.
energy [1]. Given a sufficiently high The grain growth is characteristic of the
temperature and no factors impeding grain growth of larger grain at the expense of
# Corresponding author: dedi@metal.ui.ac.id

DOI:10.2298/JMMB100901001P
200 D. Priadi / JMM 47 (2) B (2011) 199 - 209

smaller ones, leading to an increase in 2. experimental Procedures


average grain size [1]. Grain growth control
in microalloyed steels during hot working Commercial Nb-microalloyed steel was
processing, critically depends on second used for this study; detailed composition of
phase particles [2, 3]. The appropriate the steel is provided in Table 1. Rectangular
employment of microalloying elements in samples in 30 mm width and 60 mm length
high strength low alloy (HSLA) steels, (parallel to the rolling direction) were
coupled with hot working, can provide machined from hot-rolled coil (HRC)
improvements in both strength and material. A hole of 2 mm in diameter was
toughness [4, 5, 6]. This is achieved by drilled in the center of a longitudinal-
suitable manipulation of the recrystallization transverse (L-T) section of the sample for a
and precipitation phenomena that take both thermocouple insertion.
place during deformation. Heat treatment of the sample fitted with
Austenite grain growth in microalloyed the thermocouple was performed by heating
steels is influenced by numerous factors, at a constant rate of 0.3Cs-1 to temperatures
including austenizing time and temperature, of 900, 1000 and 1100 C, followed by
alloy composition, hot working history, initial holding for 10 minutes to permit full
grain size distribution, and rate of heating to austenitization of the material. Then the
the austenitizing temperature [6-8]. By specimens were quenched in water to view
suppressing austenite recrystallization, these the austenite boundary. Light microscopy of
elements act as ferrite grain refiners, thus steel plates was conducted in an Olympus
increasing the yield strength and decreasing metallurgical, microscope on 4%
the impact transition temperature [6]. picraletched specimens.
Manohar et al. [9], predicted the austenite
grain growth for microalloyed steels, 3. empirical Model
especially for Ti steel. They assumed that
fine TiN precipitates control the austenite Empirical equations are widely used to
grain growth behavior at all temperatures. describe grain growth behavior during
The influence of possible mixed precipitates isothermal heating. The power law
such as Nb(C, N) was not accounted for in relationship was first specified by Beck et al.
their mathematical model. Referring to their (1948) for normal grain growth during
empirical predictions, the aim of this study is isothermal conditions [10]:
to observe and recalculate the austenite grain dn don = k t ...(1)
size and the influencing behavior of Nb(C,
N) precipitates during reheating of Nb where n is the grain growth exponent, k is
microalloyed steels. the reaction rate constant, and dn and don are
grain sizes at time t and t=0, respectively. It
Table 1. Steel Composition (Weight Percent)
C Si Mn P S Al Nb Cu N Fe
0.085 0.222 1.45 <0.003 <0.003 0.049 0.028 0.045 <0.003 98.14
D. Priadi / JMM 47 (2) B (2011) 199 - 209 201

has been found by Sellars and Whiteman in diffusion constant at grain boundary for iron
1978 that the k value can be expressed by an (cm2s-1), Dgbo is the diffusion coefficient (=
Arrhenius type equation [11]. They 2.0 cm2s-1) [9], Q is activation energy for
previously published grain growth data on diffusion (= 171544 Jmol-1) [9], R is the
low carbon-manganese steel. They arrived at universal gas constant (= 8.31 Jmol-1K-1), T
the following general expression for is the absolute temperature (K) and is
evaluating the constant k in Eq. (1) as : thickness on grain boundary (= lattice
dn don =[A exp(-Qgg/RT)] t ...(2) parameter of austenite = 3.59 10-10 mm) [9].
Empirical models were compared to
where n and A are constants which depend observe relationships with each other.
on material composition and processing Assuming a hypothetical typical value for
conditions, Qgg is the activation energy for dRX of 80 m and solving all above models
grain growth, R is the universal gas constant by substituting T = 860, 960 and 1060 oC, for
and T is the temperature in Kelvin [10]. holding time t = 10 s and 1000 s, the results
Empirical equations for grain growth of C- obtained are shown in Figs. 1 and 2.
Mn steels are included in Table 2. It can be observed from Figs.1 and 2,
Yoshie and Nishizawa [9] presented and using empirical models for predicted grain
proposed another empirical model to predict growth of austenite in C-Mn steels giving
grain growth of austenite in cast C-Mn steel. different results. In general, if increase the
Their model represented the k constant as : holding time will take effect to don<<dn in C-
k = V Dgb/ R T ...(3) Mn steels, then the Eq. (1) can be simplified
to:
and, Dgb = Dgboexp(-QD/RT) dn = k t ...(4)
where is grain boundary energy so the values of n and k can be obtained
(= 80010-7 Jcm-2) [9], V is the molar volume by regression of experimentally measured
of austenite (= 6.97 cm3mol-1) [9], Dgb is the grain sizes and holding time [1].

Table 2. Summary of empirical models describing austenite grain growth [9,17].


Austenite Qgg
source steel N K
Conditions
3.87 1032 for 400000 for
Statically
Sellars et al., (1979) C Mn 10 T>1000oC T>1000oC 914000
recrystallized
10 for T<1000oC
53 for T<1000oC
Statically
Namba et al., (1979) Low C-Mn 2 4.27 1012 66600
recrystallized
Statically or
Hodgson and Gibbs C-Mn and
Metadynamically 7 1.45 1027 400000
(1992) C-Mn-V
recrystallized
Hodgson and Gibbs Statically or
(1992) Beynon and C-Mn-Nb Metadynamically 4.5 4.1 1023 435000
Sellars (1992) recrystallized
202 D. Priadi / JMM 47 (2) B (2011) 199 - 209

The empirical model of C-Mn-Nb steel neglecting of initial grain size don would give
either underestimates or overestimates grain rise to a large deviation on predicted values
sizes due to its inability to account for from the model. It is obvious that the grain
undissolved microalloyed effects on growth exponent n is dependent on the
austenite grain growth (Fig. 1 and 2). temperature and materials. As a
Actually, in microalloyed steels, e.g., Nb consequence, when Becks formula is
microalloyed steel, the austenite grain employed under continuous heating
growth kinetics are expected to be slower conditions, the n value should correlate with
than in plain C-Mn composition due to solute temperature [1]. So, developing an empirical
drag [12]. The final grain size after a certain model, to predict the undissolved
heat treatment does not satisfy the microalloyed effect on austenite grain
requirement in Eq. (1), and therefore, growth in Nb microalloyed steels is needed.

Fig. 1. Predicted austenite grain growth in C-Mn Steels, t = 10 s

Fig. 2. Predicted austenite grain growth in C-Mn Steels, t = 1000 s.


D. Priadi / JMM 47 (2) B (2011) 199 - 209 203

4. Calculation of a Model where Z is the ratio of growing grains to


matrix grains (rmax/rmean), thus representing
4.1. Background the heterogeneity of grain size distribution in
the sample. Gladman found that Z values
In the presence of second phase particles between 1.41 and 2 correlated closely with
(such as in Nb microalloyed steels), a the experimental results. Thus, the value of
fundamental approach in identifying a the constant kz could be between 0.05 and
theoretical solution for the grain growth 0.26. So the value of kz = 0.17 (Fig.3) close
behavior, centers on finding a hypothetical to the Gladman value of kz, then
grain radius Rcr, such that any grain with Rcr = 0.17 r/fv = d/2 ...(7)
radius equal to Rcr will neither grow nor
shrink [9]. Any grain with a radius equal to where d is the critical grain diameter.
Rcr is considered to be thermodynamically Solving Eq. (7) thus facilitates the
stable because the driving force for grain calculations of stable austenite grain size
growth of such a grain exactly balances the under any combination of volume fraction
pinning force exerted by particles on the and particle radius of precipitates. The
moving grain boundary. dissolution and coarsening of Nb microalloy
Zener proposed that the driving pressure precipitates therefore play a pivotal role in
for grain growth due to the curvature of the deciding the stable austenite grain size. The
grain boundary would be counteracted by a methodology used to estimate these two
pinning (drag) pressure exerted by the parameters is described in the following
particles on the boundary. Zener [10] first sections.
assigned a quantitative value for a critical
maximum grain radius Rcr as:
Rcr = kz r/fv ...(5)
where kz is a constant, r is the mean
particle radius, fv is the volume fraction of
particles. In the original Zener model, kz was
set to 4/3.
Experimental data for the ratio of limiting
grain radius to particle radius (Rcr/r) as a
function of volume fraction of particles (fv)
shows that the value of solubility of
(Ti,Nb)CN in austenite phase is lies directly
on the line for kz = 0.17 [9] (Fig. 3).
Gladman [13] analysed the effect of grain
sized distribution on the driving pressure
Fig.3. Experimental data of ratio of limiting
(and therefore Rcr) and found that :
grain radius to particle radius (Rcr/r), as a
Rcr = [(0.25 0.33/Z)] r/fv ...(6) function of volume fraction of particles (fv) [13].
204 D. Priadi / JMM 47 (2) B (2011) 199 - 209

4.2. Calculating nb(C,n) solubility form (Xp), can now be calculated as =Total
Nb-Xs, where Xs is the percent of niobium in
Choice of the equation describing the solution of austenite for differing values
solubility of Nb(C,N) in austenite is very of T. Total niobium in this case is 0.028%.
important because partitioning of niobium in Results of these calculations are presented in
precipitates and in austenite matrix plays a Table 3.
crucial role in deciding grain growth of
microalloyed austenite. In the current work, 4.3. Calculating Volume fraction of
we have used the Nb(C,N) solubility product Precipitates
equation [10] :
log[Nb][C]0.83[N]0.14 = 4.46 9800/T Volume V of a given mass m of a
material can be given as [9]:
[Nb][C]0.83[N]0.14 =
log1(4.46 9800/T) ...(8) V = m Na Vu / (Nu Ar) ...(11)

Multiplying both sides by [Nb]/[C]0.83 where Na is Avogadros number, Vu is the


[N]0.14 results in: volume of one unit cell of the material, Nu is
the number of atoms per unit cell of the
[Nb]2=[log1(4.46 material, and Ar is the atomic weight of the
9800/T)]{[Nb]/[C]0.83[N]0.14} ...(9) material. Nb(C,N) phase is regarded as a
It is assumed that the formation and complete ideal solid solution of NbC and
dissolution of Nb(C,N) precipitates proceed NbN because the interaction between C and
such that stoichiometric relationships of N in Nb(C,N) can be ignored [14].
Nb:C[=7.74] and Nb:N[=6.63] [10] are Therefore, the volume and the mass of
maintained whenever the overall Nb(C,N) Nb(C,N) can be written as:
ratio in the given composition of steel is VNb(C,N) = VNbC + VNbN
close to the stoichiometric ratio of
0.14[Nb]/[Nb]0.97. Therefore, substituting = 0.128mNbC + 0.119mNbN [15] ...(12)
[Nb]/[C]0.83[N]0.14=0.14[Nb]/[Nb]0.97 in
where
Eq. (9) gives:
mNbC= mNb + mC = mNb + (mNb/7.74) =
[Nb]1.97=[log1(4.46 9800/T)]0.14 ...(10) 1.129mNb
Using this equation, the amount of mNbN = mNb + mN = mNb + (mN/6.63) =
niobium out of solution, i.e., in precipitate 1.151mNb

Table 3. Predicted partitioning of Nb in dissolved and precipitated forms at temperatures of


interest
Temperature [Nb]=Xs=Predicted % Nb Xp = Predicted % Nb
(oC) in solution in austenite in ppt form = 0.028-Xs
860 0.00102 0.02698
960 0.00435 0.02365
1060 0.01489 0.01311
D. Priadi / JMM 47 (2) B (2011) 199 - 209 205

so VNb(C,N) = 0.282mNb and radius and the Ostwald ripening theory.


VFe = 0.125mFe [9] ...(13) According to the Wagners diffusion-
controlled model [16], particle size under
where mFe = 1 mNb(C,N) and Ostwald ripening conditions can be
mNb(C,N) = Xp/100. described by:
Volume fraction of precipitates can now r3 ro3 = k t ...(15)
be calculated from: Wagner, Lifshitz and Slyozov analyzed
fv=VNb(C,N)/(VNb(C,N) + VFe) ...(14) the particle coarsening more rigorously,
maintaining that the growth process begins
For different values of T we can thus
when the grains (of second phase) have
calculate the corresponding volume fraction
reached appreciable size and the degree of
of precipitates (Fig. 4).
supersaturation of the matrix has become
very slight [9]. In this situation, coalescence
4.4. Calculating Particle Coarsening
starts (i.e., growth of larger grains by
incorporation of smaller ones begins). This
The Nb(C,N) particle in Nb-HSLA steels
theory predicts particle grain growth through
plays an important role in retarding the
the equation given below:
recrystallization process. After the formation
of the new phase, the high density of small r3 ro3 = [8 V D t Xs]
precipitation tends to coarsen into a lower / [9 R T] ...(16)
density of larger particles with a smaller total where r is the average final particle radius
interfacial area. The diffusional growth and (cm) at time t, r is the average initial particle
o
diminution of precipitated particles in a solid radius (cm) at time t , is interfacial energy
o
matrix is generally known as Ostwald (= grain boundary energy = 310-7 Jcm-3)
ripening [16]. [14], V is the molar volume of Nb(C,N)
For evaluation purposes, we assessed (=6.72 cm3mol-1) [14], D is the diffusivity of
particle coarsening in terms of the mean cube solute (Nb) in matrix (austenite) (D= cm2s-

Fig. 4. Predicted volume fraction of precipitate at temperatures of interest.


206 D. Priadi / JMM 47 (2) B (2011) 199 - 209

1), t is the time for particle coarsening (s), Xs Eq. (16), we can estimate particle coarsening
is the concentration of the saturated solution at different temperatures (Fig. 6).
(Table. 3), R is the universal gas constant (=
8.31 Jmol-1K-1) and T is the absolute 4.5. Calculation of Austenite Grain size
temperature (K).
Diffusivity of Nb in austenite [14] is The micrographs of the samples after
given by: continuous heating at different temperatures
held for t = 600 s are shown in Fig. 7. The
DNb = 5.6 exp(-286000/RT) cm2s-1 (17)
measured mean linear intercept and the
Niobium diffusivities are given in Fig. 5, calculated real grain size, combined with the
for selected temperatures. substitution of predicted fv values (Fig. 4)
By considering initial mean particle and predicted r values (Fig. 6) in Eq. (7), are
radius, ro = 1.3 10-7 cm [14], t = 600 s, and predictive of stable austenite grain size (Figs.
substituting appropriate numerical values in 8 and 9).

Fig. 5. Diffusivity of Nb in austenite at temperatures of interest

Fig. 6. Predicted particle coarsening with respect to temperature at t = 600 s.


D. Priadi / JMM 47 (2) B (2011) 199 - 209 207

Fig. 7. Micrographs grain size (1000x) at different temperatures in a continous reheating


at various temperatures held for 600 s, a). 860oC, b). 960oC, c). 1060oC.

Fig. 8. Comparison of predicted and experimental grain growth behavior


in 0.028% Nb microalloyed steel.

Fig. 9. Smallest and largest mean grain sizes and predicted grain growth
for 0.028% Nb microalloyed steel at t = 600 s.
208 D. Priadi / JMM 47 (2) B (2011) 199 - 209

5. Discussion considered normal in the presence of


second phase particles.
The evolution of the microstructure At temperatures above 1030 oC, the
during the thermal history was observed by largest grain size increased sharply to
quenching the specimens at specified exceed the predicted grain size. At this
temperatures (Fig. 7). After quenching a stage, we observed mixed grain sizes, with
specimen at 860 oC, the microstructure most grain remaining small, and a small
showed only martensite, with fully number of grains developing into unusually
undissolved precipitates at the austenite large sizes compared with their neighbors.
grain boundary. Fine and uniform A significant ratio difference between the
distributed austenite grain size was largest and smallest observed grain size is
observed due to the presence of precipitates indicated; the ratio of the mean size of the
as the second phase particles. Increasing largest grain is 1.74 times the mean size of
the reheating temperature reduced the the smallest grain (Fig. 8). This behavior is
undissolved precipitates. When the typical of abnormal grain growth as
specimen was quenched at 1060oC, the described by Zener and Gladman [9] and is
microstructure was entirely martensite, due to progressive particle dissolution and
denoting a fully austenitic structure prior to particle coarsening.
quenching, because the reheating
temperature exceeded the solubility 6. Conclusions
temperature equation of Nb(C,N) [Eq. (8)].
Two distinct stages in grain growth 1. Grain growth of 0.028% Nb steels is
behavior of 0.028% Nb steels were influenced by heating temperature, where
observed (Figs. 8 and 9). At lower grain sizes increase with increasing
temperatures (<1030 oC), small and reheated temperature.
uniform grain size was observed which 2. A plot of the size of the largest and
very closely approached the predicted grain smallest grain versus temperature provides
sizes. Low temperatures indicated a an effective and sensitive method for
marginal increase in mean grain size and an detecting the onset and termination of
insignificant differentiation between the abnormal grain growth in 0.028% Nb
smallest and the largest grain sizes. steels.
According to Hillerts size distribution 3. Abnormal grain growth increases
theory [9], a grain radius greater than 1.8 rapidly at the temperature above 1030oC.
times the mean grain radius is unstable and 4. Grain growth behavior and the effect
predictive of abnormally growth. This of Nb microalloy precipitates can be
means that abnormal grain growth due to successfully explained by developing the
the size different effect is either calculated model in this work.
insignificant in niobium steels at lower 5. Grain growth predictions for Nb
temperature. Grain growth in this steels fit well with the experimental grain
temperature region, therefore, can be growth data for these steels.
D. Priadi / JMM 47 (2) B (2011) 199 - 209 209

Acknowledgments Int., 38 (1998) 913.


[14] S. Akamatsu, T. Senuma, M. Hasebe, ISIJ
The authors are grateful for the Grant Int., 32 (1992) 275.
from the Directorate on Research and [15] M. Gomez, S.F. Medina, P. Valles, ISIJ Int.,
Community Service, Universitas Indonesia. 45 (2005) 1711.
[16] E.S. Siradj, Journal of Technology, 4 (1998)
references 348.
[17] J.G. Lenard, M. Pietrzyk, L. Cser,
[1] H.R. Wang, W.Wang, Materials Sciences and
Mathematical and Physical Simulation of the
Technology, 24 (2008) 228.
Properties of Hot Rolled Products, 1st ed., UK:
[2] F. Fang, Y. Qi-long, Y. Cai-fu, S. Hang, Elsevier Science Ltd. , Oxford, Kidlington, 1999,
Journal of Iron and Steel Research International, p.167.
17(2) (2010) 36.
[3] K.A. Alogab, D.K. Matlock, J.G. Speer, H.J.
Kleebe, ISIJ Int., 47 (2007) 1034.
[4] N. Tao, et. all., Journal of Iron and Steel
Research International, 17(11) (2010) 73.
[5] V. Goryany, V Radsinsky, J. Min. Metall.
Sect. B-Metall. 38 (3-4) B (2002) 171.
[6] L. Jiang, A.O. Humphreys, J. J. Jonas, ISIJ
Int., 44 (2004) 381.
[7] S.S. Zhang, et. all., Materials Science and
Engineering A, 528 (2011) 4967.
[8] C. Yun, L. Zhang, S. Liao, H. Gao, Journal of
Materials Engineering and Performance, 19 (1)
(2010) 112.
[9] P.A. Manohar, D.P. Dunne, T. Chandra, C.R.
Killmore, ISIJ Int., 36 (1996) 194.
[10] T. Liu, Modeling Microstructural Evolution
of Microalloyed Forging Steels during Thermo
mechanical Processing, Thesis, Queens
University, Canada, 2001, p.16.
[11] C.M. Sellars, G.J. Davies, The physicals
metallurgy of hot working, Proc. of the
International Conf. on the hot working and
forming process, 1979, eds. TMS. London, p.3.
[12] F. Siciliano, J.J. Jonas, Metallurgical and
Materials Trans. A, 31A (2000) 511.
[13] P.A. Manohar, M. Ferry, T. Chandra, ISIJ