Вы находитесь на странице: 1из 5

MATERIALS FORUM VOLUME 31 - 2007 Edited by J.M. Cairney and S.P.

Ringer Institute of Materials Engineering Australasia Ltd

EFFECT OF CARBON ON THE HOT DEFORMATION BEHAVIOUR IN NB-STEELS


H. Beladi and P. D. Hodgson Centre for Material and Fibre Innovation, Deakin University, Geelong, Australia, VIC 3217 ABSTRACT A rapid hot torsion method was used to follow the recrystallization behaviour of Nb microallyed steels with different carbon contents over a wide range of process conditions. The hot deformation behaviour of austenite was not influenced by carbon. However, the post deformation softening behaviour of austenite behaved differently in respect to carbon content in different temperature regimes due to Nb(C,N) precipitates. Keywords: Nb-steel, recrystallization kinetics, carbon, thermomechanical processing In spite of extensive studies performed over the past decades, many details regarding the interaction of recrystallization and solute drag/precipitation behaviour of Nb steels remain unclear [3,4]. One such example is how the addition of an element, such as carbon influences the recrystallization kinetics of austenite in Nb bearing steels during hot working. The current research examines the interaction between deformation, recrystallization and solute drag/strain induced precipitation in Nb-steels. Therefore, the recrystallization kinetics of Nb-steels at different thermomechanical conditions and steel composition (i.e. carbon effect) are investigated.

1. INTRODUCTION Microstructural control during hot working is of considerable importance in the thermomechanical processing of steels because it affects the process capability, dimensional accuracy and the final properties of products. The high temperature deformation behaviour of steel is mainly controlled by recrystallization of austenite, which plays a major role in determining the final microstructure in thermomechanically processed steels. Three recrystallization processes may occur during hot working of austenite: dynamic, static and metadynamic recrystallization [1,3] all of which can alter the austenite grain characteristics (i.e. size and distribution) and affect the rolling loads. Dynamic recrystallization (DRX) takes place during deformation, while static recrystallization (SRX) occurs when there is an appropriate time to nucleate and grow new strain free grains after deformation [1]. Metadynamic recrystallization (MDRX) takes place under certain conditions in which dynamic nuclei formed during deformation grow after deformation [2,3]. Each of these can affect the microstructural evolution under different processing conditions. For most commercial hot deformation processes static recrystallization is the dominant mechanism. This is particularly important in microalloyed steels, where small levels of Nb, or other elements, are added to deliberately retard or stop static recrystallization. The addition of Nb to steel retards recrystallization by the solute drag effect and strain induced precipitation [3]. The latter has a much more profound influence than solute drag and effectively stops recrystallization for all practical inter-deformation times [4]. This leads to the accumulation of strain in the austenite, resulting in a finer ferrite grain size structure on transformation.

2. EXPERIMENTAL PROCEDURE Four Nb microalloyed steels were used in this study (Table 1). Apart from Steel D, all other steels were prepared in a vacuum furnace using electrolytically pure iron with alloying additions. The alloying additions were kept constant apart from the carbon, which varied from 0.04% to 0.16%C (Table 1). Steel D was, however, a commercial Nb microalloyed steel supplied by BlueScope Steel. This steel was used to clarify some aspects of post softening behaviour observed in Steels AC. Table 1. Chemical composition of steels, wt-% Steel A B C D %C 0.042 0.105 0.16 0.07 %Mn 0.83 0.86 0.86 1.5 %Si 0.28 0.29 0.29 0.2 %Nb 0.031 0.037 0.034 0.03 %N 0.0039 0.0041 0.0033 0.0035

11

A hot torsion deformation simulator was employed to deform the samples. The specimens had a gauge length of 20 mm and a gauge diameter of 6.7 mm. The entire length of the specimen was enclosed in a quartz tube, where a positive pressure of argon gas was maintained to prevent oxidation of the specimen during testing. The maximum temperature difference between the two ends of the gauge length was 10C. The reheating temperature was chosen based on the Nb(C,N) solution temperature (Ts), which was estimated by a modified Irvine et al. [5] equation to take into account the Mn and Si levels [6] (Table 2). Therefore, the specimen was reheated to 1250C for 5 min, which is well above Ts for all steels. The specimen was then cooled to 1100C and held for 80 s, followed by two pass strains of 0.4 with an interpass time of 20 s at a strain rate of 1 s-1. The deformed structure was fully recrystallized within the 20 s interpass time after each deformation. This led to a homogeneous austenite grain distribution after the roughing condition. Afterwards, three different deformation schedules were performed to determine different thermomechanical processing characteristics. Table 2. The temperature and deformation characteristics of steels Steel A B C D d (m) 41 33 30 36 Ar3 (C) 826 793 786 820 Tnr (C) 932 970 987 953 Ts (C) 1051 1181 1225 1195

point of inflection in the strain hardening rate as a function of the flow stress plot. Finally, interrupted torsion tests were used to study the kinetics of post-deformation softening. After the roughing condition, the specimen was cooled at 1C/s to a given deformation temperature and held for 80 s before a double hit deformation was applied. Each experiment involved deforming to the required strain followed by holding for a given time and then re-deforming to a strain of 0.2. The deformation temperature was varied between 900-1100C. The specimen was isothermally deformed at strain rates that varied between 0.1-3 s-1. Then, the softening fraction, s = 3 1 100 , at a given 3 2 thermomechanical condition was estimated. yield stress of the first deformation,

2 is the yield stress of the second deformation and 3 is the stress at


the end of first deformation [9].

is the

3. RESULTS AND DISCUSSION

3.1 Continuous Stress-Srain Crves Figure 1 shows some of the flow curves for Steels A-C. The curves mostly exhibit work hardening to a peak stress followed by softening to a steady state, typical for the occurrence of DRX. The peak stresses and strains increased with a decrease in temperature and increasing strain rate for all alloys. Carbon appears to not have any specific effect on the stress-strain behaviour (Fig. 1). This is not consistent with early reports that showed a reduction in work hardening rate and flow stress with carbon [10]. However, Collinson et al. [11] also reported very little effect of carbon on the stress-strain behaviour in plain carbon steels for Zener-Hollomon (Z) conditions similar to the current study. They have shown that the carbon effect only appears in steels with >0.4 carbon and under high Z conditions. As noted below one interesting change in behaviour was for the 0.1C steel at 0.1 s-1 and 900C. Here there was no clear peak in the stress-strain curve.

A multipass torsion test was firstly performed during cooling (1C/s) from 1100C to determine the characteristic temperatures for all alloys: the empirical austenite to ferrite transformation (Ar3) and nonrecrystallization temperature (Tnr). This follows the method developed by Bai and co-workers [7]. A pass strain of 0.15 and a strain rate of 1 s-1 were applied. Fourteen passes were performed and the temperature decreased from pass to pass (~20C). The Tnr and Ar3 temperatures for all alloys were estimated (Table 2). The second hot torsion test was performed after the roughing process to establish the continuous flow curve for full dynamic recrystallization (DRX) at different thermomechanical conditions. This was used to identify the characteristic strains at a given thermomechanical condition: the strain to the initiation of DRX (C) and peak stress (P). The characteristic strains, C and P, were identified along the continuous stress-strain curves and C was estimated using the approach developed by Poliak and Jonas [8]. In this method, C is associated with a

3.2 Softening Behaviour During Unloading A rapid method of determining post-deformation softening kinetics through the study of the time to 50% softening (t50) was used for the current analysis (the details of which are described elsewhere [12]). The graph consisted of two distinct regions: strain dependent softening and strain independent softening that changed at the transition strain, * (Fig. 2). 12

180 o -1 160 900 C, 1s 140 Stress (MPa) 120 100 o -1 1100C, 1s 80 A B C 60 A C 40 B o -1 1100C, 0.1s 20 0 0.0 0.5 1.0 Strain

A B C B A 900 C, 0.1s
o -1

be explained by the solute drag effect of Nb, which retards the mobility of grain boundaries (i.e. recrystallization) [3]. It is interesting to note that the solute drag effect is much more pronounced in the strain dependent than strain independent regimes. At a temperature lower than Tnr, there is a dramatic change in the post-deformation softening with the steel composition (i.e. carbon and Mn), which could be explained by the formation of Nb(C,N). The effect of Nb(C,N) precipitates on grain boundary mobility through pinning is much stronger than Nb will have in solution through solute drag [4]. Carbon significantly increases the driving force for Nb(C,N) formation at a given thermomechanical condition. Therefore, both the rate and the amount of Nb(C,N) precipitation increases with carbon, resulting in a much lower static recrystallization rate. This effect appears to be more pronounced when the Nb(C,N) precipitates are formed during deformation. This will retard DRX and significantly change the state of the austenite at the point of unloading. Therefore, this can explain the significant difference in post deformation behaviour of Steels A and B at strain rate of 0.1 s-1 at 900C. For this condition Steel A showed evidence of DRX, while little or no DRX occurred in Steel B (Figs. 1 and 3b). Dutta and Sellars developed a model for strain induced precipitation as a function of composition and process conditions [13]. This model was slightly modified by one of the current authors [4] based on torsion tests on a range of steels. It is interesting to note that using these two models gives precipitation times (for T=900C, =0.8 and strain rate of 1 s -1) of 50-100 s, 3-6 s and 1-3 s for carbon contents of 0.04, 0.1 and 0.16, respectively. As recrystallization takes around 10 s for 50% by extrapolation of high temperature data it suggests the two higher carbon contents will have precipitation before recrystallization, whereas Steel A will be recrystallized before precipitation, which supports the current results.
Steel A Steel B Steel C 1000 C 1050 C 1100 C
o o o

1.5

2.0

Figure 1. Stress-strain curves of steels at different thermomechanical conditions. A, B and C represent Steel A, Steel B and Steel C, respectively. t50 is a strong function of thermomechanical parameters as well as steel composition. The softening kinetics were accelerated by increasing temperature in all alloys (Fig. 3), although the strain independent softening is significantly less affected by temperature than strain dependent softening (Fig. 3). However, the strain rate showed a strong influence on the strain independent softening and less effect on strain dependent softening. These trends are consistent with the results reported in the literature [12]. The effect of carbon on the kinetics of post-deformation softening differed depending upon the temperature. This effect could be divided into two regions: above and below the Tnr.
10 Time to 50% Softening (s) Steel A Steel B Steel C Steel D C-Mn[12]

strain dependent softening *

strain independent softening 1 Strain


Time to 50% softening (s) t (sec)

0.1 0.1

1000

900 C

& b) =0.1/s

b)

Figure 2. Time to 50% softening as a function of strain at different alloys at strain rate of 1s-1 and deformation temperature of 1000C. The kinetics of post-deformation softening were very similar for all alloys above the Tnr for both the strain dependent and strain independent regions (Fig. 3). In other words, the steel composition (i.e. C and Mn levels) does not have a significant effect on post deformation softening behaviour above the Tnr. However, the kinetics of softening were significantly slower than a C-Mn steel of similar composition but without Nb (Fig. 2). This can 13

100

900 C

50

10

1 0.1 Strain 1

Figure 3. Time for 50% softening for different alloys at different thermomechanical conditions: a) strain rate of 1 s-1 and b) strain rate of 0.1 s-1. (break point P correspond to strain induced precipitation)

For Steel B at 0.1 s-1 the precipitation times are similar to the deformation times, which explain why all results are increased. However for Steel A the precipitation times are much longer (120-250 s for =1) than either the deformation and recrystallization times. Therefore, the models do not completely explain the current results. A break point was observed in the strain dependent softening regime for Steels A and D at deformation temperature of 900C and below 900C, respectively (Figs. 3a and 4) as summarized in Table 3. Interestingly, the sharp change in strain dependent behaviour was not observed for Steels B-D at 900C for similar thermomechanical conditions (Figs. 3a and 4). This also disappeared at a strain rate of less than 1/s for all steels (Fig. 3b). It appears that this behaviour depends on the thermomechanical conditions and steel composition as previously reported by others [14].
Time to 50% softening (s) 1000 100
B C
o Steel D, 900C SteelD, 900 C o Steel D, 875 CC Steel D, 8750 o Steel D, 850 C Steel D, 850C o Steel A, 900 C Steel A, 900C C

Using the modified Dutta and Sellars model, the time for 5% static recrystallization is estimated as 2 s at 850C (point B in Fig. 4), which is close to the time for precipitation (i.e. 3.5-21 s). The same calculation for 875C estimates a time for 5% recrystallization and precipitation of 3.7 s and 2.5-13 s, respectively. These outcomes showed that at the break point, the onset of strain induced precipitation, appears to overlap the time for initiation of static recrystallization. The reason for this observation is not clear yet, but suggests that precipitation before recrystallization can significantly suppress the post deformation softening kinetics. The significant change in the post deformation softening kinetics of Steels A and D at 900C could partly be explained by the difference in the strain rate as the post deformation softening rate increases with the strain. However, no break point was observed at 900C for Steel D. This suggests that Mn retards strain induced precipitation of Nb(C,N) since the Mn level in Steel D is two times higher than Steel A.

4. CONCLUSION Continuous and interrupted hot torsion tests were performed to study the influence of carbon content on the hot deformation and recrystallization behaviour of austenite in Nb-steels. Carbon appeared to have no specific influence on hot working of austenite in the range of 900-1100C. However, the post-deformation softening behaviour of austenite was influenced by the carbon with different effects in different temperature regimes. At deformation temperatures above Tnr, the post-deformation softening was slightly influenced by carbon. However, the post-deformation softening was significantly retarded with carbon at the deformation temperature lower than Tnr due to Nb(C,N) precipitates. If these precipitates formed during deformation then the recrystallization rate was markedly reduced, even at very high strains. In addition, there was also a break point in the t50 curves when strain induced precipitation occurred prior to initiation of DRX.

10 1 0.1

Strain

Figure 4. Time for 50 % softening as a function of strain for Steels A and D at strain rate of 1 and 3 s-1, respectively (break points A, B and C correspond to strain induced precipitation) [14]. The strain to the break point was close to the critical strain for the initiation of dynamic recrystallization for both Steels A and D at all deformation conditions (Table 3 and Fig. 4). At a lower strain than the break point, there was an increased strain dependence (Fig. 4). However, this concept that there is another recrystallization process is misleading as holding for longer times at a strain of 0.7 for Steel D did not give a significant increase in softening. Table 3. The break point characteristics of Steels A and D in strain dependent softening regime Steel A D D Td (C) 900 875 850 Strain rate 1/s 3/s 3/s break point 0.8 0.7 1.1 C 0.92 0.73 1

Acknowledgements This research was supported by grants through ARC (H.B.) including an ARC Federation Fellowship (P.D.H.). The authors would like to acknowledged Dr. S. H. Zahiri and Mr. R. Pow for their assistance and useful discussion. The authors also thank BlueScope for the supply of Steel D.

References 1. 14 J. J. Jonas: Mater. Sci. Eng. A, 1984, vol. 5, p. 155.

2. 3. 4. 5. 6. 7. 8.

C. Roucoules, P. D. Hodgson, S. Yue and J. J. Jonas: Metall. Mater. Trans. A, 1994, vol. 25, p. 389. J. J. Jonas and I. Weiss: Met. Sci., 1979, vol. 13, p. 238. P. D. Hodgson and R. K. Gibbs: ISIJ Int., 1992, vol. 32, p. 1329-1338. K. J. Irvine, F. B. Pickering and T. Gladman: J. Iron Steel Inst., 1967, vol. 205, p. 161. F. Siciliano JR and J. J. Jonas: Metall. Mater. Trans. A, 2000, vol. 31, p. 511. D. Q. Bai, S.Yue, W. P. Sun and J. J. Jonas, Metall. Mater. Trans. A, 1980, vol. 11, p. 387. E. I. Poliak and J. J. Jonas: ISIJ Inter., 2003, vol. 43, p. 684-691.

9.

10. 11. 12. 13. 14.

R. K. Gibbs, P. D. Hodgson and B. A. Parker: Morris E. Fine Symposium, eds. P. K. Liaw, J. R. Weertman, H. L. Marcus, J. S. Santner, TMS, USA, 1991, p. 73. H. W. Mead and C. E. Birchenall: Trans. AIME, 1956, vol. 206, p. 1336. D. C. Collinson, P. D. Hodgson and C. H. J. Davies: Thermec97, Wollongong, Australia, 1997, p. 483. M. R. Cartmill, M. R. Barnett, S. H. Zahiri and P. D. Hodgson: ISIJ Int., 2005, vol. 45, p. 1903-1908. B. Dutta and C. M. Sellars: Mater. Sci. Technol., 1987, vol. 3, p. 197. P. D. Hodgson, S. H. Zahiri and J. J. Whale: ISIJ Int., 2004, vol. 44, p. 1224-1229.

15

Вам также может понравиться