Effect of Austenitizing Temperature and Prior Martensite on Ultra-Fine Bainite Transformation Kinetics

Abstract

An evaluation method for bainite transformation kinetics was established by theoretical derivation, dilatometric curve analysis, and microstructure observation. The isothermal transformation of ultra-fine bainite under different austenitizing temperatures and contents of prior martensite was studied using a DIL805L dilatometer. The kinetic parameters (activation energy Q*, autocatalytic factor λ, temperature rate constant κ, unit volume transformation rate, and the number density of nucleation sites N_i) of ultra-fine bainite transformation under different austenitizing temperatures and contents of prior martensite were calculated based on the displacement growth bainite dynamics model. It was found that the autocatalytic factor λ is linear with the austenite grain size d, and the number density of nucleation sites N_i is closely related to the average volume of the bainite subunit V_b. Moreover, the formation of prior martensite and its increase can increase the number of nucleation sites and the nucleation rate of the ultra-fine bainite; thus, the ultra-fine bainite transformation can be accelerated.

1. Introduction

Ultra-high-strength steel is widely used in lightweight automotive parts, armor steel plates, and some wear-resistant materials. Caballero and Bhadeshia et al. [1,2,3] developed a high-carbon, silicon-rich, ultra-fine bainitic steel which exhibits excellent strength and tensile ductility. This kind of steel, containing nanoscale lath-like carbide-free bainite, has become a focus for many researchers. However, the long time taken by bainite transformation has restricted its industrial development. The factors affecting the bainite transformation rate include the optimization of the alloy composition system [2,3,4,5,6,7,8,9], the size of the prior austenite grains [10,11,12,13,14], and the heat treatment parameters. In recent years, researchers have paid attention to the pretransformation of the tissue interface, which can promote the transformation rate of ultra-fine bainite. The mechanism for accelerating the bainite transformation using this parameter has the following four main parts: (1) the phase transformation strain introduced by the preformed martensite [15]; (2) preferred nucleation sites provided by the increased martensite phase interface [16,17]; (3) reduced interfacial energy introduced by martensite [18]; and (4) dislocations introduced by martensitic transformation promoting bainite transformation [19]. There are many opinions on the mechanism of the acceleration of ultra-fine bainite transformation induced by the introduction of prior martensite.

Recently, several kinetic models of bainite transformation have been established, mainly based on two bainitic nucleation models: One was presented by Quidort and Brechet [20]; it is similar to the classical nucleation theory and is combined with the kinetics of bainite formation calculated by Johnson Mehl Avrami (JMA). However, the pre-exponential factor of this model is assumed to be constant, not specified, and the model does not consider autocatalysis. The other model was presented by Bhadeshia et al. [21]; it assumes that the bainite subunit undergoes displacement growth and is combined with the Olson–Cohen martensite nucleation model to describe the bainite nucleation. Also, the overall transformation kinetics are determined by the rate of isothermal bainite; the activation energy is calculated based on a specific linear relationship specified by two parameters. In addition, the consistency of experimental data was optimized using autocatalytic parameters and pre-exponential factors. Van Bohemen [22,23] modelled the concepts recognized in bainite and martensitic theory. Their measurement of potential nucleation sites was more rigorous than previous models. Moreover, the current model only contains two valuables, autocatalytic parameters λ and a constant rate κ, which can better describe isothermal transformation data.

In this paper, the phase transformation kinetics model based on the bainite displacement growth mechanism was used to calculate the activation energy, nucleation rate, and nucleation density of ultra-fine Bainite transformation, which can provide a theoretical basis for the austenite grain size, primary martensite, and its content in accelerating the mechanism of ultra-fine bainite transformation.

2. Material and Experimental Methods

2.1. Material Preparation

The experimental steel was melted at 1450–1550 °C in vacuum to obtain the cast slab. Homogenization heat treatment was performed at 1200 °C for 10 h. After homogenization, hot forging was performed at 1100 °C. The composition of the steel is shown in

Table 1. Composition of experimental steel (wt %).

Elements	C	Si	Mn	Co	Cr	Mo	Ni	Nb	Fe
Composition	0.51	1.72	0.83	0.56	0.98	0.25	0.60	0.04	Bal.

.

2.2. Determination of the Dilatometric Curve

The experiment material was machined into a Φ 4 mm × 10 mm cylindrical specimen. Then, the specimens were grinded and polished. After that, the martensitic transformation start temperature (Ms) and the bainite isothermal transformation curve were determined using a high-precision Bahr DIL805L dilatometer. The parameters are shown in Figure 1. Heat Treatment Parameter 1 (DIT) was designed to study the effect of the austenitizing temperature on the isothermal transformation of ultra-fine bainite. First, the specimens were heated from room temperature at a heating rate of 5 °C/s to four different austenitizing temperatures of 900 °C, 950 °C, 1000 °C, and 1050 °C, and were held for 180 s. Then, the specimens were cooled to Ms + 20 °C at a cooling rate of 20 °C/s and held for 1 h at the temperature of bainitic transformation (Figure 1a). Heat Treatment Parameter 2 (QBT) was designed to study the effect of prior martensite and its content on the isothermal transformation of ultra-fine Bainite. First, the specimens were heated from room temperature at a heating rate of 5 °C/s to the austenitizing temperature of 1050 °C and were held for 180 s. Then, the specimens were cooled to four different temperatures below Ms at a cooling rate of 20 °C/s and held for 30 s; at last, the specimens were heated at a rate of 20 °C/s to Ms + 20 ° C and held for 1 h at the temperature of bainitic transformation (Figure 1b).

2.3. Microstructure Observation

The surfaces of the samples were grinded and polished and then placed in a constant-temperature etching solution (4 g picric acid + 4 g sodium dodecylbenzene sulfonate + 100 mL distilled water) with a temperature of 70 °C for 5 min. After that, the microstructure of the prior austenite grains was observed using a ZEISS AX10 (Carl Zeiss AG, Jena, Germany) metallographic microscope. The SEM samples were grinded and mechanically polished, then etched with a 2% to 4% (volume fraction) nitric acid solution, and then subjected to a ZEISS ULTRA 55 (Carl Zeiss AG, Jena, Germany) thermal field emission scanning electron microscope for microstructure observation. The TEM samples were polished to 50 μm manually and then punched into a Φ 3 mm discs. After thinning in a Gatan691 (FEI Company, Hillsboro, OR, USA) ion thinning machine, samples were analyzed using a Tecnai G2 F30 S-TWIN (FEI Company, Hillsboro, OR, USA) transmission electron microscope at an acceleration voltage of 200 kV.

3. Results

3.1. Phase Transformation

Figure 2 shows the dilatometric curves obtained under different experimental conditions. Figure 2a shows the dilatometric curve of the Ms transformation temperature. The Ms of the experimental steel was about 260 °C. Figure 2b shows the dilatometric curve of the sample for Heat Treatment Parameter 1. Bainite transformation occurred between the isothermal curves, resulting in nonlinear expansion. Figure 2c,d shows the dilatometric curves of Heat Treatment Parameter 2, where two-stage nonlinear expansion occurred at two isothermal temperatures and physical expansion occurred at elevated temperatures. Early studies on similar alloy systems have shown that when the sample is maintained at a temperature below Ms for 10 s, the larger hump on the dilatometric curve observed for the quenching parameter is due to athermal martensite transformation occurring below the Ms temperature [24]. When the sample was kept at a constant temperature of 280 °C, the formation of athermal martensite was stopped, and the volume expansion isothermal transformation caused by bainite transformation began (Figure 2d). There are no other discontinuities on the dilatometric curve above Ms in Figure 2c, which confirms that the cooling rate during quenching was sufficiently high to prevent any other high-temperature transformation (to ferrite or pearlite) above Ms. Therefore, no transformation was recorded in the dilatation data during quenching to room temperature after isothermal treatment at 280 °C (Figure 2b,c).

3.2. Microstructure Observation

The effect of the prior austenite grain size on the transformation kinetics of steel was investigated in the range of 900–1050 °C. The austenite grain sizes at different austenitizing temperatures are shown in Figure 3. Image-Pro Plus software (Version 6.0, Media Cybernetics, Rockville, MD, USA) was used to measure the size of the prior austenitic grains at different austenitizing temperatures. The intercept method was also used to measure 10 metallographs under each temperature state. As the austenitizing temperature increased, the prior austenite grain size increased from 30 (±3.82) μm to 92 (±2.99) μm.

Figure 4 shows microstructure diagrams of ultra-fine bainite under different heat treatment parameters after austenitization at 1050 °C. Figure 4a,4c shows the ultra-fine bainite structure directly isothermally transformed after austenitizing at 1050 °C. It consists mainly of lamellar bainitic ferrite and retained austenite. There are two main forms of retained austenite: one is a film of retained austenite (film RA) distributed between bainitic ferrite laths; the other is blocks of retained austenite (block RA). The bainite laths in bainitic lath packet have the same crystallographic orientation relationship, while the bainitic lath packets have nonuniform orientation. The blocks of retained austenite are surrounded by bainite lath packets of different orientational relationship and appear as a polygonal shape on a two-dimensional scale, while the film of retained austenite is distributed between the bainite substructures. In addition, besides bainite ferrite, film RA, and blocky RA, there was a small amount of lenticular martensite (PM) in the ultra-fine Bainite structure treated according to Heat Treatment Parameter 2. Further, the bainitic ferrite lath adjacent to the martensite lath had the same growth direction as the martensite (Figure 4b,d). Chester and Bhadeshia [25] estimated the width μ_w of the bainite plate as a function of the transformation temperature by fitting test results:

$${\mu }_w=0.001077\left(T+273\right)-0.2618.$$

(1)

The average volume of the bainite subunit was determined as [26]:

$$V_b={\mu }_w·{\mu }_l·{\mu }_t$$

(2)

where ${\mu }_t$ and ${\mu }_l$ represent the thickness and length of the bainite lath. Combining Equations (1) and (2), it can be found that when the isothermal temperature of bainite is constant, the strip width is fixed. The true thickness of bainitic ferrite (${\mu }_t$) can be calculated using the formula ${\mu }_t=2L_T/\pi$, $E=\pm 2{\sigma }_L/\left(\pi \sqrt{N}\right)$ [27,28], where L_T represents the linear intercept length measured in the TEM image by Image-Pro Plus software (Figure 4c,d), E represents the 95% confidence interval, σ_L represents the standard deviation of the true thickness, and N represents the counting plate quantity. The statistical results showed that for Heat Treatment Parameter 1, the thickness of the bainitic ferrite plates was 297 ± 3 nm, and the thickness was 200 ± 2 nm for Heat Treatment Parameter 2. However, the introduction of martensite can refine the thickness of the bainite lath [29], and since martensite occupies a part of the austenite space, bainite growth is limited and the length is reduced (Figure 4b). As the martensite content increases, the austenite residual space shrinks and bainitic subunit decreases.

When the QBT-25 samples were held for 2 min (Figure 5a), a small amount of bainitic ferrite (BF) was nucleated and grew at the martensite lath interface and the austenite grain boundaries; when the holding time was 5 min (Figure 5b), a large number of bainitic ferrite laths grew along the martensite grain boundaries into austenite grains; when the holding time was 10 min (Figure 5c), the bainite firstly nucleated, also becoming the nucleation site of the new bainite. The sample formed more bainitic ferrite to form the bainite lath packets. As the isothermal time was prolonged, the carbon atoms continued to diffuse into the austenite, causing an increase in the carbon concentration in the austenite. After holding for 20 min (Figure 5d), the carbon concentration in the austenite increased sharply. When the carbon concentration reaches a certain value, the free energy of the austenite (γ) is equal to the new phase bainitic ferrite (α). The transformation of austenite to bainitic ferrite then stops, forming multiphase steel with ultra-fine bainite, martensite, and retained austenite.

3.3. Model Establishment

According to the bainite displacement growth mechanism theory, the bainite nucleation rate equation at temperature T can be expressed using a model established by Bohemen et al. [22]:

$$\frac{dN}{dt}=\frac{kT}{h}(1-f)N_i(1+\lambda f)exp(-\frac{Q*}{RT})$$

(3)

where k, h, and R are the Boltzmann constant k = 1.380649 × 10⁻²³ J/K, Bronck constant h = 6.62607015 × 10⁻³⁴ J·s, and the gas constant R = 8.314 J/(mol·K); $\frac{dN}{dt}$ is the bainite nucleation rate per unit volume; f is the volume fraction of bainite; λ is the bainite nucleation autocatalytic coefficient; and $N_i$ is the number density of nucleation sites.

According to the model established by Bohemen et al., the time t required to form the bainite volume fraction f at a given temperature can be calculated as [30]

$$t=\frac{-\ln \left(\frac{1-f}{1+\lambda f}\right)}{\kappa \left(1+\lambda \right)}$$

(4)

where (1 − f) is the volume fraction of untransformed austenite and κ represents the temperature-dependent rate constant, according to the literature [22]:

$$\kappa =\frac{kT}{h}\frac{Z\delta }{d}{\alpha }^m\left(T_h-T\right)exp\left(-\frac{Q*}{RT}\right).$$

(5)

Here, Z represents the austenite geometry factor, Z = 6; δ is the effective thickness of the austenite grain boundary, δ ≈ 1 nm [22,31]; d is the average austenite grain size; and ${\alpha }_m$ represents the composition-related parameter, according to the linear Equation (4) calculation [32]:

$${\alpha }_m=0.0224-0.0107x_c-0.0007x_{\mathrm{Mn}}-0.00005x_{Ni}-0.00012x_{\mathrm{Cr}}-0.0001x_{\mathrm{Mo}}.$$

(6)

T_h represents the highest transformation temperature of bainite. By studying the formation of bainite in different types of steel, Bhadeshia et al. found that the bainite transformation is divided into two processes: nucleation and growth. According to the displacement mechanism of bainite transformation, the growth of the bainite plate is a constant plane strain shape deformation, and even if the semi-equilibrium distribution of carbon atoms occurs in the nucleation process, it is completely diffusionless. Therefore, the bainite phase transformation temperature cannot be above the highest temperature corresponding to the T’₀ curve (the relationship between the C content in austenite and the phase transition temperature), and another two conditions must be met:

ΔG^γ→α < −G_SB

(7)

and

$$∆G_m<G_N$$

(8)

where G_SB, the energy storage capacity of bainite, is about 400 J/mol; ∆G_m is the driving force of bainitic ferrite nucleation for a quasi-equilibrium carbon distribution; and G_N is a common nucleation function. In this paper, ∆G_m was calculated by the method described by Bhadeshia and colleagues, using the computer program “MUCG83”, and G_N is expressed as

$$G_N=3.637 {\mathrm{Jmol}}^{-1}{K}^{-1}(T-273.18)-2540 {\mathrm{Jmol}}^{-1}.$$

(9)

The result is shown in Figure 6. When the above conditions are met, T_h is 500 °C.

4. Analysis and Discussion

4.1. Effect of the Prior Austenite Grain Size on the Transformation Kinetics of Ultra-Fine Bainite

According to the rate equation proposed by Bohemen and Sietsma, the volume fraction of bainite f can be expressed as [22]

$$f=\frac{1-\exp \left\{-\kappa \left(1+\lambda \right)t\right\}}{1+\lambda \exp \left\{-\kappa \left(1+\lambda \right)t\right\}}.$$

(10)

The volume fraction of bainite f can be estimated by experiments [22] using

$$f=\frac{L_t-L_0}{L_{\infty }-L_0}$$

(11)

where L₀ and L_∞ are the initial and final lengths of the isothermal transformation of the sample, and L_t is the length of the sample corresponding to the transformation time t. The L₀ and L_∞ of Heat Treatment Parameter 1 can be obtained from Figure 7.

The bainite transformation kinetics curves of different austenite temperatures were fitted using MATLAB2016a (mathworks, Natick, Massachusetts, MA USA). The results are shown in Figure 8. The experimental results are similar to those of many previous researchers [22,23,31,33]. The autocatalytic factor λ is related to the austenite grain size. The larger the λ, the more obvious the S-shaped curve; the kinetics exhibited a strong catalytic effect and a long incubation period was shown in the experiment (Figure 7). As the austenite grain boundary density increases, the nucleation effect of the grain boundary increases and the autocatalytic nucleation effect decreases. Therefore, in bainite transformation, bainite growth is a displacement process which causes plastic deformation of the surrounding austenite matrix [34,35,36]. The dislocation density near the bainite/austenite interface may change as the bainite growth continues to form a lath bundle. The grain boundary nucleation effect gradually becomes greater than the autocatalytic nucleation. According to the experimental conditions, the swelling characteristics of the steel are mainly affected by the phase transformation, because the isothermal temperature of the bainite is the same, and the thermal expansion caused by the physical thermal effect of the temperature did not have any influence on the relevant performance analysis of the experimental material [37]. It can be seen from Figure 7 that there was a period of incubation before the phase transformation of bainite, and the expansion amount began to increase after a certain period of isothermal temperature; bainite transformation began to occur, and the expansion amount became smaller with time. As the austenitizing temperature increased, the final amount of expansion increased, but the end of the transition was delayed.

The activation energy Q*, the nucleation rate per unit volume dN/dt, and the number of nucleation sites N_i of the different austenite grain sizes when the bainite transformation was 1% were determined using Equations (3) and (5), wherein the average volume of the bainite subunit V_b was determined by Equation (2); the results are shown in

Table 2. Calculation results of austenite grain size, autocatalytic factor λ, rate parameter κ, activation energy Q*, nucleation rate dN/dt, and nucleation density number N_i at different austenitizing temperatures.

Taus/ °C	Grain Size d/μm	λ	Vb m3	κ	Q* KJ/mol	Ni m−3	dN/dt m−3s−1
1050	92.82 ± 2.99	78	1.36 × 10−18	5.04 × 10−5	146.14	1.67 × 1014	7.22 × 1013
1000	72.21 ± 1.99	57	1.36 × 10−18	7.19 × 10−5	145.66	2.15 × 1014	8.37 × 1013
950	49.8 ± 2.34	24	1.36 × 10−18	1.94 × 10−4	142.81	3.11 × 1014	1.75 × 1014
900	32.74 ± 3.82	17	1.36 × 10−18	4.60 × 10−4	140.77	4.74 × 1014	3.91 × 1014

. The activation energy of isothermal bainite increased with increasing austenite grain size, but the overall change is not obvious. The nucleation rate dN/dt and the number of nucleation sites N_i increased rapidly with decreasing austenite grain size. The number of nucleation sites (N_i) of small grain sizes at the beginning of bainite transformation was an order of magnitude larger than that of large-sized grains. This indicates that the low austenite temperature is more prone to bainite transformation, while the effect of the activation energy is very small; this is mainly because the smaller the austenite grain size, the larger the grain boundary area, which is more conducive to providing more nucleation sites. This is consistent with the results of many previous studies [10,11,38]. Bainite transformation involves nucleation and growth. In terms of the nucleation sites, the austenite grain size is small, and the grain boundary area is larger, which can provide more nucleation sites for bainite transformation. It is well known that a smaller austenite grain size limits bainite lath growth, which is not of benefit to bainite transformation.

4.2. Effect of Martensite on the Transformation Kinetics of Ultra-Fine Bainite

It is assumed that the temperature was raised to the bainite phase transformation temperature after being kept warm below temperature Ms, and the prior martensite content f_M = ab/ac [39] was calculated according to Figure 9 using the lever theorem.

Assuming that the remaining austenite after martensite transformation can convert 100% to bainite, the time t required to form the bainite volume fraction f can be calculated as

$$t=\frac{-\ln \left(\frac{\left(1-f_{\mathrm{M}}\right)-f}{1+\lambda f}\right)}{\kappa \left(1+\lambda \right)}.$$

(12)

Then, the volume fraction of bainite $f$ can be expressed as

$$f=\frac{\left(1-f_{\mathrm{M}}\right)-\exp \left(-\kappa \left(1+\lambda \right)t\right)}{1+\lambda \exp (-\kappa \left(1+\lambda \right)\mathrm{t}}$$

(13)

and the volume fraction of bainite $f$ estimated experimentally is

$$f=\frac{L_t-L_b}{L_{\infty }-L_0}.$$

(14)

Here, L₀ and L_∞ are the initial and final lengths of the isothermal transformation of the sample, L_t is the length of the sample corresponding to the transformation time t, and L_b is the length of the bainite which begins to transform; L₀, L_b, and L_∞ of Heat Treatment Parameter 2 can be obtained from Figure 10.

The same method was used for fitting, and the fitting results are shown in Figure 11 and in

Table 3. Calculation results of rate parameter κ, activation energy Q*, nucleation rate dN/dt, and the number density of nucleation sites N_i under different heat treatment parameters.

Taus °C	Heat Treatment Parameters	fM %	λ	Vb m3	κ	Q* KJ/mol	Ni m−3	dN/dt m−3s−1
1050	DIT	0	78	1.36 × 1018	5.40 × 105	146.14	1.67 × 1014	7.22 × 1013
1050	QBT-15	21	670	6.50 × 1019	1.43 × 105	145.86	3.50 × 1014	1.67 × 1014
1050	QBT-25	35	130	5.20 × 1019	5.36 × 105	142.60	4.37 × 1014	2.3 × 1014
1050	QBT-40	49	50	3.90 × 1019	1.09 × 104	139.65	5.83 × 1014	4.15 × 1014
1050	QBT-50	56	24	2.60 × 1019	2.07 × 104	136.84	8.75 × 1014	9.77 × 1014

. The influence of pre-quenching the austenite before isothermal treatment on the transformation kinetics of bainite has been the subject of previous studies. Kawata et al. [16] showed that bainite transformation from the two-phase structure of austenite and martensite was faster than that from single-phase austenite, and the nucleation of bainitic ferrite was accelerated at the martensite/austenite boundary. The martensite/austenite boundary was considered to be the main nucleation site of bainitic ferrite. In this paper, as the martensite content increased, the self-destructive nucleation of the grain boundary was strongly weakened, but the transformation rate of the bainite and the number density of nucleation sites per unit volume were gradually increased (Table 3). This phenomenon is consistent with the findings of Kawata et al. It can be seen from Table 3 that the two-stage isothermal treatment had lower activation energy Q* than the single-stage isothermal, which may be due to the introduction of the initial martensite to produce the distortion strain; this increased the driving force of the bainite transformation so that the activation energy was reduced, but the reduction was not obvious, similar to in the study by Smanio Véronique [40]. However, before the bainite began its transformation, the nucleation rate and the number density of nucleation sites (N_i) in the two-stage isothermal treatment were an order of magnitude larger than those in the single-stage isothermal, which indicated that the number of bainite nucleation sites increased due to the introduction of prior martensite, thereby increasing the transformation rate of the bainite.

Compared with DIT, although the rate parameter κ of QBT-15 was reduced, its number density of nucleation sites N_i was increased by one order of magnitude, mainly because the ultra-fine bainite unit volume fraction was reduced. A schematic diagram of the bainite nucleation and growth stages under different heat treatment schemes is shown in Figure 12. With the single-step isothermal parameter, the completely austenitized sample was cooled to a temperature above M_S. Firstly, bainite started to nucleate at the austenite grain boundary and then grew inside the grain, accompanying the diffusion of carbon from the new phase ultrafine bainite (α) to austenite (γ). When the concentration of carbon reached a certain value, the free energy values of the new phase bainite ferrite (α) and austenite (γ) were equal, and the bainite phase transformation stopped because the driving force was insufficient to cause the FCC to continue to transform into the BCC structure. In the case of the two-stage isothermal, a completely austenitized sample generated a certain amount of martensite after being cooled to a temperature below M_S, and the martensitic transformation produced volume expansion and an increase in the dislocation density in the surrounding austenite. Compared to the single-step isothermal, the growth space of the two-stage isothermal bainite decreased, and the length of the bainite lath decreased, which caused the average volume to decrease and the number of nucleation sites of bainite transformation to be increased. Then, the temperature was raised to above the temperature M_S in the isothermal phase, and the bainite began to grow in the austenite grain boundary and the martensite interface with the austenite grain. Similarly, when the free energy values of the new phase bainite ferrite (α) and austenite (γ) were equal, the bainite phase transformation stopped.

5. Conclusions

The effects of different austenitizing temperature and prior martensite formation on ultra-fine bainite transformation were studied based on phase transformation kinetic model calculations of the bainite displacement growth mechanism combined with microstructural observation. Our conclusions are as follows:

• The decrease of austenite temperature will reduce the size of the prior austenite grains, reduce the activation of bainite transformation, and increase the nucleation rate and the number density of nucleation sites at the grain boundaries, thus accelerating the ultra-fine bainite phase transformation.
• Prior martensite can cause abnormal strain in the austenite, which reduces the activation energy of the ultra-fine bainite transformation, promotes the nucleation rate, increases the nucleation sites of the ultra-fine bainite, and increases the number density of nucleation sites N_i, thus accelerating the ultra-fine bainite phase transformation.
• As the content of prior martensite increases, the average volume of ultra-fine bainite subunit decreases, the nucleation density and the nucleation rate of ultrafine bainite phase transformation will increases.