Preventing Abnormal Grain Growth of Austenite in Low Alloy Steels

Abstract

Abnormal grain growth (AGG) in austenitic state is studied in low alloy steel in relation with precipitation state. It is observed that initial austenite grain size and precipitation state plays more important role in defining the normal or abnormal grain growth condition than the final one obtained after a heat treatment. Precipitate volume fraction evolution with time-temperature having similar quantity at the end but different initial grain sizes, showed different grain growth phenomenon. Arguments are presented to rationalize the presented experience. A simplified AGG model is applied to understand the effects of initial mean austenite grain size and precipitate size distribution on the subsequent AGG occurrence.

1. Introduction

Grain size control is an important aspect of materials science, since the grain size distribution strongly affects mechanical properties of many alloys. Grain growth in materials is driven by the decrease in interfacial energy.^1,2,3) One can describe two basic types of grain growth in materials; normal and abnormal. Normal Grain Growth (NGG) is characterized by a self-similar coarsening process in the microstructure, where larger grains grow at the expense of smaller ones.^2,3,4) On the other hand, Abnormal Grain Growth (AGG) represents non stationary and “explosive” growth of a few number of grains. It is often attributed to the presence of second phase particles.^3,5) The presence of abnormal grains in a microstructure is known to deteriorate fatigue life of a component.⁶⁾

The direct impact of grain size control on mechanical properties has driven extensive experimental and numerical modelling studies. There are number of authors who have deduced the parabolic grain growth relationship for mean grain size in the case of normal grain growth (see the pioneer work of Hillert³⁾ and the review of Gladman¹⁾). Subsequently these models were extended to predict and explain AGG. M. Hillert addressed the “abnormal” or “discontinuous” grain growth phenomenon as a defect model³⁾ where it is assumed that precursor for AGG already existed. Both Gladman and Hillert models can predict the propagation of an abnormal grain but failed to give realistic prediction of AGG initiation taking into account precipitation state (size distribution and volume fraction).

Zener⁷⁾ first proposed that during grain growth, second phase particles present in the matrix, exert back pressure during the boundary movement. This back pressure is commonly known as the Zener Pinning. Most of the previous works explained AGG initiation with local unpinning of the larger grains.^1,2,3) Higher growth pressure of the larger grains overcome the pinning and subsequently grow in an abnormal manner. Bréchet et al.⁸⁾ introduced corner pinning that lead to grain size dependent pinning, which may initiate AGG. According to their hypothesis, corner pinning is more efficient in pinning the grain boundary than conventional Zener pinning, leading to a net pinning force that decreases with grain size, which is itself the cause of AGG instability. This model has been recently extended to account for the whole precipitate size distribution and give an explanation for AGG initiation⁹⁾ and propagation.¹⁰⁾

Besides precipitate pinning, AGG is also associated with solute drag effect.¹¹⁾ It is reported that grain-growth mode can be changed depending on the solute diffusivity where initial average grain size play an important role. From the modelling point of view, Monte Carlo and phase field approaches considered the commonly assumed anisotropic grain boundary energy.^12,13) Anisotropic grain boundary energy allows particular grain in the heterogeneous microstructure to grow preferably faster than the others. Thus the comparatively rapid growth of some grains attain abnormality in the microstructure.

All the above mentioned studies are performed to explain the AGG phenomenon. Note that there is no widely accepted mechanism for AGG initiation and propagation. In terms of AGG prevention, proposition of precipitate free alloy composition has been made, but the advantage of precipitates as grain size controller is then lost! To the authors knowledge, there are very few published studies dealing with prevention of AGG without changing the materials chemistry, i.e. by optimizing thermo-mechanical treatment.

Among the few published propositions, Bruno et al.¹⁴⁾ and Murakami et al.¹⁵⁾ propositions to prevent AGG by optimizing precipitation state can be mentioned. According to Bruno et al.,¹⁴⁾ AGG initiation temperature can be increased by optimizing heat treatment to produce small initial precipitate mean size with homogeneous distribution. They reported that AGG initiation temperature can be increased from 1373 K to 1473 K in an alloy containing niobium carbonitride precipitate by having a fine precipitation state. Quite similar claim is also made by Murakami et al.¹⁵⁾ In the present work, AGG is studied in relation with precipitation state and mean grain size evolution in a low alloy steel. Efforts are made to explain in detail the AGG dependency on precipitation state evolution with time and temperature. Experimental work along with numerical modeling shed light on a particular precipitation state evolution which can prevent AGG. A method for preventing AGG is proposed, where a particular materials state can better resist AGG in subsequent austenitizing treatment.

2. Method and Materials

In this study, a low alloy steel ‘steel-A’ is chosen, in which the major precipitate phase is AlN. Its composition is given in Table 1. Two initial states of steel-A are studied for the austenite grain growth behavior. Heat treatment cycles performed on the alloys are shown in Fig. 1. The as rolled (AR) state is obtained by rolling cast section from the billet. The rolling cycle was terminated at ~800°C and then slowly air cooled. A Fully Precipitated (FP) state is obtained through the two stages isothermal heat treatment of the AR state: (i) 1 h at 875°C (above Ac₃) and (ii) 2 h at 680°C (below Ac₁) (see Fig. 1).

Table 1. Composition of industrial steel-A alloy in atom%.

C	Si	Mn	Ni	Cr
0.702	0.407	0.455	1.22	1.47
V	Ti	Al	Nb	N
0.00696	0.00194	0.0549	0.00171	0.0598

Fig. 1.

Heat treatment cycles of steel-A alloy to obtain As Rolled (AR) and Fully Precipitated (FP) state and subsequent heat treatment of both the states to obtain different heat treatment states for study. (Online version in color.)

The prior austenite grain boundaries were revealed using a super saturated solution of hot picric acid modified with a surface reactant. Optical microscopy was performed and images were taken in order to measure grain size distribution. In order to perform image analysis it was necessary to hand draw the grain boundary using tracing. The traced images were treated using a image processing software (ImageJ) in order to measure the grain size distribution. Precipitate volume fraction was measured using Electrolytic dissolution of matrix using a potentiostat. As the precipitating phase did not dissolve in the solution, filtration provided separation of second phase particles. Filtered particles were then dissolved in an acidic solution (HNO₃ + HCL + HF + H₂SO₄) and analyzed in an Induction Coupled Plasma (ICP) analyzer. Precipitates quantity were measured in terms of substitutional element (e.g. Al, Nb, V) in Parts Per Million (PPM). In the studied alloy, it is found that the major precipitate is AlN. This is confirmed both by Transmission Electron Microscope (TEM) and ICP analysis. Substitutional element’s mass fraction to constituting precipitate’s volume fraction conversion is shown in expression 2.1. Here, $f_{\upsilon }^{AlN}(T,t)$ is volume fraction of AlN precipitate, $f_m^{Al}$ is the mass fraction of element Al measured using ICP, ρ_Fe and ρ_AlN are molecular density of the matrix and AlN precipitate and M_Al and M_N are molar mass of the substitutional (Al) and interstitial (N) elements.   

$$f_{\upsilon }^{AlN}(T,t)=f_m^{Al}\frac{{\rho }_{Fe}}{{\rho }_{AlN}}\left(\frac{M_N}{M_{Al}}+1\right)$$(2.1)

In this study, precipitate size distribution is measured using carbon extraction replica prepared from different heat treated states. The replicas were analyzed in the TEM using High Angle Annular Diffraction (HAADF) mode (more details on experimental procedures are given in^16,17)). The images that were taken, analyzed using image processing software to measure the equivalent sphere radius. In this regard, it was assumed that the precipitates were geometrically uniform in all directions.

3. Results and Discussion

The studied steel-A alloy was heat treated at different temperatures: 875°C, 900°C, 1050°C and 1200°C. For all the temperatures, except for 1050°C, AR and FP states showed normal grain growth. However, at 1050°C, AR sample clearly exhibited NGG whereas FP samples exhibited AGG. In Figs. 2 and 3 precipitate size distributions and grain size distributions of AR and FP states heat treated at 1050°C for 1 h are presented. Grain size distribution is presented in terms of grain class volume fraction $V_i^f=\left(N_i{D}_i^3\right)/\left({\sum }_i^N{N}_i{D}_i^3\right)$ , where i is the class index, N_i is the number of grains in that class and D_i is the diameter of that class. Precipitate size distributions are presented in terms of distribution density (m^–4) where number of precipitate in a given class (m^–3) is divided by the class width (ΔR). From Fig. 2(d) it can be seen that in AR state heat treated at 1050°C, the grain size distribution remained quasi-stationary. Normal grain growth can also be confirmed from the adjacent micrograph (see Fig. 2(b)). On the other hand, FP state in Fig. 3(d) has a bimodal grain size distribution. Here the larger grain classes are occupying higher volume fraction in comparison with the smaller one, clearly showing AGG. The adjacent micrograph confirms the abnormal nature of grain growth (see Fig. 3(b)).

Fig. 2.

As Rolled (AR) state heat treated at 1050°C for 1 h: (a) AlN precipitate’s HAADF TEM image, (b) micrograph representing austenite (γ) grains, (c) AlN precipitate size distribution (here m^–4 obtained by dividing the precipitate number density m^–3 by the class size ΔR) and (d) grain size distribution in terms of volume fraction. (Online version in color.)

Fig. 3.

Fully Precipitated (FP) state heat treated at 1050°C for 1 h: (a) AlN precipitate’s HAADF TEM image, (b) micrograph representing austenite (γ) grains, (c) AlN precipitate size distribution (here m^–4 obtained by dividing the precipitate number density m^–3 by the class size ΔR) and (d) grain size distribution in terms of volume fraction. (Online version in color.)

An important parameter controlling the grain growth type is the initial mean grain size D_n because it strongly influences the driving force for grain growth (AGG and/or NGG) (i.e. the smaller the grain size, the more interfacial energy stored in the system). An analysis of approximatively 2000 grains lead to a precise determination of the initial D_n and D_n after a heat treatment for both FP and AR states. Here, the initial austenite mean grain size is denoted as the mean grain size obtained just after the ferrite/bainite to austenite transformation. In this study, mean austenite grain sizes obtained for AR and FP states heat treated at 875°C for 5 min are considered as the initial austenite grain sizes. Figure 4(a) shows the values of initial D_n for AR and FP state, which are 16 and 10 μm respectively. These values are coherent with the heat treatment performed to get the FP state. This treatment was indeed designed to maximize the precipitate volume fraction and density, leading thus to a higher pinning pressure and a smaller grain size.

Fig. 4.

(a) Prior austenite mean grain sizes of AR and FP state heat treated at different temperatures for 1 h and (b) Maximum (by diameter) 10% (by number) grain’s total volume fraction. (Online version in color.)

The mean grain size (D_n) evolution in both AR and FP states is presented in Fig. 4(a) and cumulative volume fraction of the largest classes of grains are shown in Fig. 4(b). From Fig. 4(a) it can be seen that the mean grain size evolution trend in both AR and FP states are quite similar. At all the heat treated temperature FP state showed relatively smaller mean grain size in comparison with the AR state. It quite obvious as the FP state has a fully precipitated condition at the beginning of heat treatment. It can be observed that the mean grain size evolution with temperature does not clearly show a transition from normal to abnormal grain growth but in the literature a relatively large increase in mean grain size is always associated with initiation of AGG.^3,13,18) In the Fig. 4(b), the total volume fraction of the 10% largest grains are denoted by V_f(max10%). For AR state, the V_f(max10%) remains quite constant at all the heat treatment temperatures and as stated earlier, the grain growth remains quasi stationary (normal). On the other hand, in FP state shows a large increase in volume fraction occupied by the larger grains. In fact, at 1050°C the larger grains are occupying ~90% of total volume, which is an indication of AGG. So, it is quite apparent that presentation of cumulative volume fraction of largest class grains can be an efficient alternative for presenting the occurrence of AGG at a particular heat treatment.

A comparison between precipitate size distribution of AR and FP states and also heat treated at 1050°C is presented in Fig. 5. AR state showed comparatively smaller precipitates radius classes (see Fig. 5(a)) and lower number density because AR state has low precipitate volume fraction (0.015%). On the other hand, FP state showed a wider size distribution, this is expected as the FP state was heat treated to maximize precipitate volume fraction. After heat treatment at 1050°C for 1 h, AR state (see Fig. 5(c)) showed a wider distribution than FP state (see Fig. 5(d)). In the previous section, it is already shown in Fig. 4(b) that FP state heat treated at 1050°C for 1 h showed a pronounce presence of AGG. So, it might be reasonable to assume that the AGG condition still persisted in the FP state even after the heat treatment. A comparison between grain growth condition and precipitate size distribution shows that comparatively wider precipitate distribution in AR-1050°C-1 h state (max precipitate radius 245 nm) showed NGG condition while comparatively narrow precipitate size distribution in FP-1050°C-1 h state showed AGG condition. This observations are rather contradictory with observations reported by Bruno et al.¹⁴⁾ According Bruno et al., a narrow size distribution is favorable to prevent AGG in the austenitic grains but present observation showed that AGG can be initiated in presence of a comparatively narrow (FP) precipitate size distribution. On the other hand, AR state having wider size distribution showed NGG microstructure after the heat treatment. This indicate that final precipitate size distribution (obtained after a heat treatment) has little impact on the grain growth condition (AGG or NGG). The occurrence of AGG or NGG is indirectly dependent on the precipitate size distribution evolution path which influences the mean austenite grain size.

Fig. 5.

Precipitate size distributions of (a) As rolled state (AR), (b) Fully precipitated (FP), (c) AR state heat treated at 1050°C for 1 h and (d) FP state heat treated at 1050°C for 1 h. (here, distribution density m^–4 obtained by dividing the precipitate number density m^–3 by the class size ΔR). (Online version in color.)

Figure 6(b) showing precipitate volume fraction after one hour of heat treatment at different temperatures. It can be seen that the FP state exhibits a decrease in AlN volume fraction with increasing temperature until complete dissolution up to 1200°C. On the other hand, for the AR state, AlN volume fraction starts from a lower value, then reaches a maximum at 1050°C and finally shows quite similar value as FP state for higher temperatures.

Fig. 6.

(a) AlN volume fraction evolution at 1050°C at different holding times and (b) AlN volume fraction evolution at different temperatures after 1 h holding time. (Online version in color.)

Figure 6(a) represents the time evolution of precipitate volume fraction at 1050°C. It confirms the previous observation that the FP state starts from a higher precipitate volume fraction than AR state until both states reach the same equilibrium at 1050°C.

Numerous authors^2,4,19) have previously mentioned abnormal grain growth association with precipitate dissolution. In the present study, the same argument is verified. It is observed that a fully precipitated state (FP) is subjected to AGG at a particular temperature while a partially precipitated state (AR) did not show any AGG. This particular observation leads to a rather contradictory understanding with previous knowledge about grain growth control. Many previous studies referred to a fully precipitated state being better for austenite grain growth control. The present study showed that fully precipitated state might be susceptible to AGG in subsequent heat treatments where precipitate dissolution is more likely to occur. So it can be stated from the presented results, that a better grain growth control can be achieved if the alloy is partially precipitated prior to subsequent heat treatment and have the possibility to increase in precipitate volume fraction (free solute atoms).

To explain this behavior, it can be noticed that AR and FP state do not exhibit the same initial austenitic grain size D_n (the grain size obtained just after austenite transformation). Moreover, the final precipitate size distribution may also influences the grain grown condition of the alloy.

In order to confirm the importance of both (i) the initial austenitic grain size D_n (the grain size obtained just after austenite transformation) and, (ii) the precipitate size distribution, we use the criterion for AGG propagation introduced in a previous article.¹⁰⁾ Let us consider a situation where a large grain of diameter D_ab is surrounded by smaller normal grains of diameter D_n. AGG can occur if (i) the larger grain grows (dD_ab / dt > 0) and, (ii) the larger grain grows comparatively faster than the others (d(D_ab / D_n) / dt > 0). The abbreviated form of the second condition for AGG can be expressed as follows:   

$$\frac{d}{dt}\left(\frac{D_{ab}}{D_n}\right)=\frac{1}{D_n}\frac{d{D}_{ab}}{dt}-\frac{D_{ab}}{D_n^2}\frac{d{D}_n}{dt}>0$$(3.1)

It can be seen from the Eq. (3.1) that in order for the second condition to be true, the first condition has to be always true. So, the condition for AGG can be only expressed by the Eq. (3.1). Further simplifications of the AGG condition can be achieved as following:   

$$\text{or},\hspace{1em}\left(\frac{d{D}_{ab}}{dt}\right)/\left(\frac{d{D}_n}{dt}\right)>\frac{D_{ab}}{D_n}$$(3.2)

So, AGG can occur in a system where the larger grains growth rate ratio with normal/mean grain growth rate is larger than the corresponding larger grain size ration with normal/mean grain size.

The smaller grains are subjected to a classical growth pressure due to grain boundary interface diminution:   

$$P_{D_n}={\lambda }_n\gamma /D_n$$(3.3)

Here γ is the interfacial energy of grain boundary and λ_n is a geometrical factor. However, the larger grain is subjected to a specific growth pressure P_Dab (see Ref. 10 for more details):   

$$P_{D_{ab}}=\frac{2\gamma }{D_n}\left[\mu -(1-\mu ){\left(\frac{D_n}{D_{ab}}\right)}^2+(2\mu -1)\frac{D_n}{D_{ab}}\right]$$(3.4)

Here μ is a compact efficiency constant. All grains are subjected to two kinds of pinning pressures: (i) classical Zener pinning (Eq. (3.5)) and, (ii) corner point pinning (introduced by Bréchet and Militzer⁸⁾ - Eq. (3.6)):   

$$P_Z=2k_s\gamma N_r\mathrm{\hspace{0.17em} }\mathrm{\hspace{0.17em} }\text{with}\mathrm{\hspace{0.17em} }\mathrm{\hspace{0.17em} }{N}_r=\sum _i^n{n}_i{r}_i^2$$(3.5)

  

$$P_C=\frac{K_V\alpha k_s\gamma }{K_A}{D}_n\sum _{i=1}^n{n}_i{r}_i$$(3.6)

Here, n_i and r_i are the number and associated radius of precipitates in each class i, k_s and K_A are geometrical constants and α is the amplification factor of corner point pinning relatively to Zener pinning. In this model, the whole precipitate size distribution is taken from TEM characterization (see Figs. 5(c) and 5(d)). The growth rates of larger and smaller normal grains are calculated using Eqs. (3.7) and (3.8).   

$$\frac{d{D}_{ab}}{dt}=M(P_{D_{ab}}-P_Z-P_C)$$(3.7)

  

$$\frac{d{D}_n}{dt}=M(P_{D_n}-P_Z-P_C)$$(3.8)

Here, M is the grain boundary mobility.

The model parameters are listed in Table 2. All model parameters have the same value as in reference,¹⁰⁾ except for μ and λ_n, that are slightly modified according to the alloy. From these two equations, it is possible to estimate the value of the AGG criterion: d(D_ab / D_n) / dt.

Table 2. List of all model parameters.

ks8)	kv8)	λn	γ	α	KV8)	KA8)	μ
π	4π/3	0.6	1	2	(1.1)3π/6	(1.1)2π/2	0.53

The initial mean austenitic grain size D_n and the final precipitate size distribution (n_i(r_i)) obtained at the end of the heat treatment are taken as input parameters in the model. Figures 2 and 3 show that FP state showed AGG condition and AR state showed NGG after 1 h of heat treatment at 1050°C. Presence of AGG in FP state after the heat treatment indicates that the pinning condition is still suitable for AGG. Even though the continuous evolution of precipitation state occurred during the heat treatment which experimentally difficult to measure for all the intermediate states. In order to keep a simplistic approach of predicting physical parameters impact on grain growth condition, precipitation states obtained at the end of heat treatments are chosen. In the case of mean austenite grain sizes, from Fig. 4(a), it can be observed that in both AR and FP states the increase in mean austenite grain sizes are quite small and can be considered as negligible. So, the choice of initial mean austenite grain size should provide an acceptable approximation of the real situation. This approach also provides the opportunity to have reduced number of experimental data requirement for model application. A full coupled model between precipitation and grain growth may provide more realistic predictions, but it would require the prediction of the precipitation state for all time and temperature, which goes far beyond the scope of this paper.

Influences of mean austenite grain size and precipitation state are separated by the application of cross experimental data in the AGG model. In the case of AR state (see Fig. 7(a-1)), first the D_n and precipitate size distribution (n_ir_i) of AR is applied in the model and then model prediction is evaluated using D_n of FP state with AR state’s precipitate size distribution (see Fig. 7(a-1)). Similar calculations are done also done for FP state (see Figs. 7(b-1) and 7(b-2)). This approach showed the affects of D_n on the grain growth condition having same precipitate size distribution.

Fig. 7.

(a-1) Model prediction for As Rolled (AR) state heat treated at 1050°C for 1 h and D_n = 16 μm, (b-1) Fully Precipitated (FP) State heat treated at 1050°C for 1 h and D_n = 10 μm, (a-2) model prediction for AR state heat treated at 1050°C for 1 h and D_n = 10 μm and (b-2) Fully Precipitated (FP) State heat treated at 1050°C for 1 h and D_n = 16 μm. (Online version in color.)

From Fig. 7, it can be observed that the model prediction is in agreement with the experimental observations. The AR state heat treated at 1050°C for 1 h showed d(D_ab / D_n) / dt = 0 (see Fig. 7(a-1)). In this particular case, although the larger grains can grow, they can not grow comparatively faster than the smaller ones. So, the microstructure can not attain AGG conditions. Same heat treatment for the FP state showed that the condition for AGG (d(D_ab / D_n) / dt > 0) is fulfilled (see Fig. 7(b-1)), which is again in agreement with the experimental observations. In FP state, the model predicts that grain larger than 2 times of the mean grain size D_n in the system can grow comparatively faster. As the largest grains in the system can outgrow the smaller ones it is highly probable that AGG microstructure will be induced.

Model prediction for AR state with FP state mean grain size D_n = 10 μm (see Fig. 7(a-2)) shows AGG condition. AGG model prediction of AGG with smaller mean grain size refers to the fact that comparatively larger initial mean austenite grain size can prevent AGG onset. In this particular case, importance of initial mean austenite grain size is the dominant factor to determine AGG/NGG condition. Model prediction for FP with AR state mean grain size D_n = 16 μm (see Fig. 7(b-2)) shows NGG condition which is opposite of previous prediction of AGG. This observation again shows that AGG condition is largely influenced by the initial mean austenite grain size. An increase austenite grain size can effectively reduce the probability of AGG occurrence while smaller mean austenite grain size can increase the AGG probability. All the model results are listed in Table 3.

Table 3. AGG model results (grain boundary mobility, M = 1).

State	Dn (μm)	Pinning (Nm–2)	$\frac{1}{M}\frac{d{D}_n}{dt}$ (m/s)	${\left(\frac{1}{M}\frac{d{D}_{ab}}{dt}\right)}_{D_{ab}=2D_n}$ (m/s)	$\left(\frac{d{D}_{ab}}{dt}\right)/\left(\frac{d{D}_n}{dt}\right)$	$\frac{D_{ab}}{D_n}$	AGG/NGG
AR	16	18532	25823	36767	1.42	2.0	NGG
FP	10	23889	36111	73000	2.02	2.0	AGG
AR	10	36549	23450	51930	2.21	2.0	AGG
FP	16	14966	22534	43147	1.91	2.0	NGG

As the precipitate volume factions are quite equal in both states (AR and FP) after heat treatment at 1050°C for 1 h, it is apparent that the initial mean austenite grain size and precipitate size distribution characteristics determined the onset of AGG. It should be noted that in the FP state a comparatively smaller initial mean austenite grain size (10 μm), increases the probability of pre-existing larger grains to be surrounded by higher number (compare to the AR state) of smaller grains. Higher number of surrounding smaller grains contributes to the increase in the larger grains growth pressure. As a result, driving pressure of the relatively larger grains increases and consequently causes AGG initiation. On the other hand, in the AR state precipitate volume fraction at the beginning of 1050°C heat treatment was 0.017%, which was 1.6 times less than the FP state. Lower initial precipitate volume fraction results in lower pinning pressure, allowed the initial austenite grain size (16 μm) to be larger than the FP state (10 μm). Following the same analogy, larger mean austenite grain size allowed pre-existing larger grains to be surrounded by comparatively (compared to the FP state) lower number of smaller grains. Thus, the driving pressure of the pre-existing larger grains are decreased and consequently decreasing the probability of AGG initiation. From model predictions and also experimental results it can be said that AGG in materials is a phenomenon induced by interplay of microstructural properties such as: precipitate volume fraction, size distribution and initial mean austenite grain size where initial mean austenite grain size playing the dominant role. In order to avoid AGG, heat treatment cycles can be optimized to have a partially precipitates states with possibility of precipitate growth in the subsequent scheduled heat treatment. This should allow the larger precipitates to grow or coarsen leading to a wider precipitate size distribution. As the materials is partially precipitated, lower pinning pressure at the beginning of heat treatment will also allow the initial mean austenite grain size to be large enough to reduce the growth pressure of larger grains in the microstructure. Thus precipitation state evolution indirectly influencing the grain growth condition by allowing initial mean austenite grain size to be large. Lower driving pressure of the larger grains combined with favorable pinning condition provided by wide precipitate size distribution are observed to be ideal to reduce the chances of AGG onset.

4. Conclusion

It is shown that AGG propagation is influenced by the prior thermal history of the materials. This itself influences both the initial grain size and the precipitate size distribution. Difference in initial grain size and precipitate size distribution with similar precipitate volume fraction can induce very different grain growth phenomenon. A proposition is presented to prevent AGG by optimizing precipitation state to a partial one which can effectively prevent AGG in subsequent heat treatment. Simple model gave accurate prediction of AGG taking into account the initial mean grain size and the whole precipitate size distribution determined experimentally. As It is often difficult to separate the influence of precipitation state and initial prior austenite grain size in the onset of AGG, such approach could make it possible to discriminate between these two microstructural parameters. A comparative picture of different microstructural properties affect on grain growth conditions is also presented.

Acknowledgement

The authors would like to extend their sincere gratitude to ‘The Centre LYonnais de Microscopie (CLYM)’ for facilitating the use of their Electron Microscopy facilities and ASCOMETAL for financing this project.

References

• 1)   T.  Gladman: Proc. Royal Soc. London. Ser. A, Math. Phys. Sci., 294 (1966), 298.
• 2)   I.  Andersen,  O.  Grong and  N.  Ryum: Acta Metall. Mater., 43 (1995), 2689.
• 3)   M.  Hillert: Acta Metall., 13 (1965), 227.
• 4)   P. R.  Rios: Acta Mater., 45 (1997), 1785.
• 5)   T.  Gladman: The Physical Metallurgy of Microalloyed Steels, Inst. of Mater., Maney for the Inst. of Materials, London, (2002).
• 6)   S. W.  Thompson and  G.  Krauss: Metall. Trans. A, 20A (1989), 2279.
• 7)   C.  Zener: J. Appl. Phys., 20 (1949), 950.
• 8)   Y.  Bréchet and  M.  Militzer: Scr. Mater., 52 (2005), 1299.
• 9)   C.  Leguen,  M.  Perez,  T.  Epicier,  D.  Acevedo and  T.  Sourmail: Proc. of Euromat 2009, ed. by D. Keogh, Federation of European Materials Societies, Glasgow, United Kingdom, (2009).
• 10)   M. A.  Razzak,  M.  Perez,  T.  Sourmail,  S.  Cazottes and  M.  Frotey: ISIJ Int., 52 (2012), 2278.
• 11)   S. G.  Kim and  Y. B.  Park: Acta Mater., 56 (2008), 3739.
• 12)   A.  Rollett,  D.  Srolovitz and  P.  Anderson: Acta Metall., 37 (1989), 1227.
• 13)   U. P. W. B. J.  Rudnizki and  B.  Zeislmair: Comput. Mater. Sci., 10 (2010), 209.
• 14)   J.  Bruno and  P. R.  Rios: Scr. Metall. Mater., 32 (1995), 601.
• 15)   T.  Murakami,  H.  Hatano and  H.  Yaguchi: Mater. Sci. Forum, 539–543 (2007), 4167.
• 16)   D.  Acevedo-Reyes,  M.  Perez,  S.  Pecoraro,  A.  Vincent,  T.  Epicier and  P.  Dierickx: Mater. Sci. Forum, 500–501 (2005), 695.
• 17)   D.  Acevedo-Reyes,  M.  Perez,  T.  Epicier,  A.  Bogner,  T.  Epicier and  C.  Verdu: J. Micro., 232 (2008), 112.
• 18)   C. V.  Thompson,  H.  Frost and  F.  Spaepen: Acta Metall., 35 (1987), 887.
• 19)   D.  Srolovitz,  G.  Grest and  M.  Anserson: Acta Metall., 33 (1985), 2233.