Microstructural evolution during thermomechanical processing of a Ti ...

Microstructural Evolution during Thermomechanical Processing of a Ti-Nb Interstitial-Free Steel Just below the Ar3 Temperature I.A. RAUF and J.D. BOYD Laboratory thermomechanical processing (TMP) experiments have been carried out to study the austenite transformation characteristics, precipitation behavior, and recrystallization of deformed ferrite for an interstitial-free (IF) steel in the temperature range just below Ar3. For cooling rates in the range 0.1 7C s21 to 130 7C s21, austenite transforms to either polygonal ferrite (PF) or massive ferrite (MF). The transformation temperatures vary systematically with cooling rate and austenite condition. There is indirect evidence that the transformation rates for both PF and MF are decreased by the presence of substitutional solute atoms and precipitate particles. When unstable austenite is deformed at 850 7C, it transforms to an extremely fine strain-induced MF. Under conditions of high supersaturation of Ti, Nb, and S, (Ti,Nb)xSy precipitates form at 850 7C as coprecipitates on pre-existing (Ti,Nb)N particles and as discrete precipitates within PF grains. Pre-existing intragranular (Ti,Nb)xSy precipitates retard recrystallization and grain coarsening of PF deformed at 850 7C and result in a stable, recovered subgrain structure. The results are relevant to the design of TMP schedules for warm rolling of IF steels.

I.

INTRODUCTION

INTERSTITIAL-FREE (IF) steels are state-of-the-art sheet steels, which have exceptionally good deep drawing properties.[1,2] These steels typically have a total C 1 N content of less than 50 ppmw and contain microalloy additions of Nb and/or Ti. Processing is normally carried out in two stages: (1) production of ‘‘hot band’’ by hot rolling, accelerated cooling, and coiling; and (2) cold rolling, annealing, and coating. Warm rolling in the ferrite range is an alternative processing route for producing IF steels having strength and formability intermediate to hot-rolled sheet and cold-rolled/annealed sheet.[3,4] Most of the previous studies of microstructural evolution during thermomechanical processing (TMP) of IF steels pertain to the production of hot band for subsequent cold rolling and annealing. For example, Najafi-Zadeh et al.[5] studied the hot deformation of IF steels, and austenite transformation temperatures have been reported for IF steels,[5] ultra low carbon steels,[6] and dilute Fe alloys.[7] For the range of cooling rates employed in accelerated cooling of sheet, the transformation products are either polygonal ferrite (PF) or massive ferrite (MF).[7] Studies of precipitation in normally processed hot band report large (100 to 1000 nm) TiN, TiS, and (Ti,Nb)4C2S2 and small (10 to 100 nm) TiC and (Ti,Nb)C precipitates.[8,9,10] The MC precipitates often form as coprecipitates on pre-existing nitride, sulfide, or carbosulfide particles. The few studies of ferrite recrystallization in the warm rolling range[3,4] show that the precipitate distribution prior to deformation is an important I.A. RAUF, formerly with the Department of Materials and Metallurgical Engineering, Queen’s University, is Research Associate, Department of Physics, University of Alberta, Edmonton, AB, Canada T6G 2J1. J.D. BOYD, Professor, is with the Department of Materials and Metallurgical Engineering, Queen’s University, Kingston, ON, Canada K7L 3N6. Manuscript submitted August 24, 1995. METALLURGICAL AND MATERIALS TRANSACTIONS A

factor, and the temperature range for complete recrystallization is very limited. In the present study, austenite transformation characteristics, precipitation behavior, and recrystallization of deformed ferrite were investigated for an IF steel in the temperature range just below Ar3. The objective was to determine the details of microstructural evolution under processing conditions relevant to warm rolling. II.

EXPERIMENTAL

A. Thermomechanical Treatments The material studied was a commercial Ti-Nb IF steel, having the composition given in Table I. The starting material was 25-mm-thick transfer bar, which had been hot rolled and air cooled. Cylindrical dilatometer samples 4mm diameter and 8-mm long were machined directly from the transfer bar with the cylinder axis parallel to the transverse direction of the rolled bar. Laboratory thermomechanical treatments were carried out in a MMC quench-deformation dilatometer. Sample temperature was controlled to 55 7C by a Type S thermocouple spot welded to the sample surface and connected via a computer control system to the induction furnace of the dilatometer. All treatments were carried out in a vacuum of 1025 torr, and the cooling rate was controlled by a flow of He gas on the sample surface. Uniaxial compressive deformation was achieved by a gas-driven platen, which produced a mean strain rate of approximately 1 s21. The following three groups of experiments were carried out. 1. Continuous cooling transformation Two pretreatments were employed. In the first (1A), samples were solution treated for 20 minutes at 1290 7C. In the second (1B), samples were solution treated 20 minutes at 1290 7C, cooled to 1000 7C at 30 7C s21, held 20 s at 1000 VOLUME 28A, JULY 1997—1437

Table I. C

N

Ti

Chemical Composition of IF Steel (Weight Percent) Nb

Si

Mn

P

S

ASA*

0.0038 0.003 0.018 0.019 0.008 0.170 0.006 0.006 0.043

further 100 seconds at 850 7C. Samples are quenched at different stages of these treatments and the microstructures characterized. The details of all the thermomechanical treatments are summarized in Table II.

*Acid-soluble aluminum

B. Microstructural Characterization Table II.

Thermomechanical Treatments

1A

20 min at 1290 7C, continuously cooled to room temperature (RT) at rates 0.1 7C/s to 130 7C/s

1B

20 min at 1290 7C, cool to 1000 7C, deform to ε 5 0.3, 100 s at 1000 7C, continuously cooled to RT at rates 0.1 7C/s to 130 7C/s

2

20 min at 1290 7C, cool to 820 7C at 30 7C/s, 20 to 60 s at 820 7C, quench to RT

3A

20 min at 1290 7C, cool to 850 7C at 30 7C/s, 20 s at 850 7C, deform to ε 5 0.3, 100 s at 850 7C, quench to RT

3B

20 min at 1290 7C, cool to 850 7C at 30 7C/s, 600 s at 850 7C, deform to ε 5 0.3, 100 s at 850 7C, quench to RT

Samples for microstructural characterization were sectioned parallel to the long axis, mounted, mechanically polished, electropolished (75 pct acetic acid-15 pct methanol-10 pct perchloric acid) at 25 7C and 25 V, and lightly etched in 2 pct Nital. Microstructures were characterized by scanning electron microscopy using secondary electron (SE) and electron channeling contrast (ECC) images. The accelerating voltage was 10 kV and the working distance was 10 mm for ECC imaging. Carbon extraction replicas were prepared from selected samples by carbon coating in a vacuum evaporator, etching in 10 pct Nital, floating replicas off in methanol, and collecting them on Cu grids. The replicas were examined in a PHILIPS* CM 20 scanning transmission electron micro*PHILIPS is a trademark of Philips Electronic Instruments Corp., Mahwah, NJ.

scope at 200 kV, and qualitative EDS analyses were made using a Si(Li) light element detector at a specimen tilt of 30 deg and a 10-nm probe diameter. III.

RESULTS

A. Continuous-Cooling-Transformation

Fig. 1—CCT diagram for samples cooled from 1290 7C: pretreatment 1A. (Time scale is cooling time from 1000 7C).

7C, deformed to a true strain of 0.3, and held a further 100 seconds at 1000 7C. After the pretreatment, samples were cooled to ambient temperature at a selected rate in the range 0.1 7C s21 to 130 7C s21, and the transformation temperatures were determined for each cooling path. The final microstructure was characterized for the highest (130 7C s21) and the lowest (0.1 7C s21) cooling rates. 2. Isothermal transformation Samples were solution treated 20 minutes at 1290 7C, cooled at 30 7C s21 to 820 7C, held at 820 7C for times ranging from 20 to 60 seconds, then quenched to ambient temperature. The mean cooling rate during the quench was 130 7C s21. The dilation of the sample was monitored during the 820 7C isothermal treatment, and the microstructures of quenched samples were characterized. 3. Isothermal transformation with deformation Samples were solution treated 20 minutes at 1290 7C, cooled at 30 7C s21 to 850 7C, held either 20 (3A) or 600 seconds (3B) at 850 7C, deformed to ε 5 0.3, and held a 1438—VOLUME 28A, JULY 1997

The initial austenite grain sizes for the 2 pretreatments were estimated from optical micrographs of quenched samples. Assuming the ferrite grain diameter is 1/2 the prior austenite grain diameter, it was determined that after 20 minutes at 1290 7C, the austenite grains are equiaxed and about 400 mm in diameter. After deformation and 100 seconds hold at 1000 7C, the austenite appears to be completely recrystallized equiaxed grains about 160 mm in diameter. The continuous-cooling-transformation (CCT) diagram obtained for the 1290 7C pretreatment (1A) is shown in Figure 1. At the slowest cooling rates, only a single transformation is observed, starting at 880 7C and finishing at 810 7C. With increasing cooling rate, more transformation points appear on the dilation curves, until at cooling rates greater than 20 7C s21, two distinct transformations are indicated. The corresponding microstructures are shown in Figure 2. The 0.1 7C s21 cooling rate results in large (500 to 600 mm) polygonal ferrite (PF) grains (Figure 2(a)). The matching ECC micrograph from the same area (Figure 2(b)) shows weak contrast changes in the PF grains, which indicate a subgrain structure. For the 130 7C s21 cooling rate (Figures 2(c) and (d)), the microstructure is a mixture of areas that have the serrated grain boundaries and fine substructure characteristic of massive ferrite (MF), plus areas of PF similar to that observed at the slowest cooling rate (i.e., equiaxed grains showing distinct grain boundaries in SE, with faint substructure in ECC). However, the PF grain diameter is much smaller at the higher cooling rate (200 to 300 mm). Based on the microstructures, the two transforMETALLURGICAL AND MATERIALS TRANSACTIONS A

Fig. 2—Microstructures after continuous cooling from 1290 7C: pretreatment 1A. (a ) 0.1 7C s21, SE; (b ) 0.1 7C s21, ECC; (c ) 130 7C s21, SE; and (d ) 130 7C s21, ECC.

Fig. 4—Microstructures after deformation and continuous cooling from 1000 7C: pretreatment 1B. (a ) 0.1 7C s21, SE; (b ) 0.1 7C s21, ECC; (c ) 130 7C s21, SE; and (d ) 130 7C s21, ECC.

Fig. 3—CCT diagram for samples deformed and cooled from 1000 7C: pretreatment 1B.

mations indicated by the CCT diagram for 130 7C s cooling rate are taken to be PF at the higher temperatures (770 7C to 840 7C) and MF at the lower temperatures (700 7C to 730 7C). The CCT diagram obtained for the 1000 7C pretreatment (1B) is shown in Figure 3. Only a single transformation is observed over the entire range of cooling rates. At 0.1 7C s21 to 5 7C s21, the transformation start and finish temperatures are essentially constant at 850 7C and 830 7C, respectively. The transformation temperature range decreases with further increase in cooling rate, until at 130 7C s21, the transformation starts at 620 7C and finishes at 540 7C. The microstructures in Figure 4 show 300- to 400-mm-diameter PF with faint substructure at 0.1 7C s21 and 100 pct MF at 130 7C s21. Note, the characteristic feature of MF in all cases is the distinct fine (10 to 15 mm) subgrains, which are generally equiaxed, but with irregular boundaries (Figures 4(d) and 2(d)). By comparison, the subgrains in the PF grains appear as weak contrast changes in the ECC micrographs and are larger (50 to 100 mm) (Figures 2(b), 2(d), and 4(b)). 21

METALLURGICAL AND MATERIALS TRANSACTIONS A

Fig. 5—Microstructures of samples quenched after various holding times at 820 7C: treatment 2. (a ) 20 s, SE; (b ) 20 s, ECC; (c ) 45 s, SE; (d ) 45 s, ECC.

B. Isothermal Transformation The sample dilation during isothermal treatment at 820 7C was close to the limit of measurement of the system, but it was consistent with the microstructural observations, which show that the transformation to PF is complete before 45 seconds. Figures 5(a) and (b) show a partially transformed structure after 20 seconds, where the austenite has formed MF during the quench. Figures 5(c) and (d) show, after 45 seconds, a fully transformed structure of 100- to 200mm PF grains, again with the faint substructure. With further holding at 820 7C, the PF grains coarsen rapidly to ;500 mm and the faint substructure persists. C. Isothermal Transformation with Deformation The quenched structure after 20 seconds at 850 7C (Figures 6(a) and (b)) is a mixture of PF and MF (similar to Figures 2(c) and (d) and 5(a) and (b)), which is indicative VOLUME 28A, JULY 1997—1439

Fig. 6—Microstructures of samples quenched at different stages of treatment 3A at 850 7C. (a ) 20 s, SE; (b ) 20 s, ECC; (c ) ε 5 0.3, SE; (d ) ε 5 0.3, ECC; (e ) 100 s hold, SE; and (f ) 100 s hold, ECC (deformation axis is shown by → ←).

Fig. 7—Microstructures of samples quenched at different stages of treatment 3B at 850 7C. (a ) 600 s, SE; (b ) 600 s, ECC; (c ) ε 5 0.3, SE; (d ) ε 5 0.3, ECC; (e ) 100 s hold, SE; (f ) 100 s hold, ECC (deformation axis is shown by → ←).

of incomplete transformation. After 600 seconds at 850 7C (Figures 7(a) and (b)), the transformation is long complete and the quenched structure is coarse (200 to 300 mm) PF grains (similar to Figures 2(a) and (b)). The quenched microstructure from the deformed, partially transformed sample (Figures 6(c) and (d)) contains areas of small (10 to 50 mm) equiaxed grains (marked PF) and areas of extremely fine elongated grains ;5-mm wide (marked MF). These are taken to represent areas that were PF and austenite, respectively, prior to deformation. It is clear in Figure 6(c) that a deformed coarse PF grain has recrystallized into the small (10 to 50 mm) equiaxed PF grains. The adjacent areas are deformed austenite grains that have transformed into the fine (5-mm-wide) elongated grains reported elsewhere to be strain-induced MF.[6] (Note, this interpretation infers that in this case, the MF forms during deformation rather than during the quench, as in all other examples of MF.) The quenched microstructure from the deformed, completely transformed sample (Figures 7(c) and (d)) shows the original coarse PF grains elongated perpendicular to the deformation axis, containing an incomplete network of smaller (25 to 50-mm) PF grains and 10 to 20-mm diameter subgrains; i.e., the deformed PF grains are partially recrystallized. The two as-deformed structures exhibit very different behavior when held a further 100 seconds at 850 7C. In samples that contained recrystallized PF and strain-induced MF, both phases coarsened rapidly to form 200- to 300-mm PF grains containing a persisting subgrain structure (Figures 6(e) and (f)). There are areas of fine (10 to 50 mm) and coarser (25 to 75 mm) elongated subgrains (Figures 6(f) at

F and C, respectively), suggesting one originated with the recrystallized PF and the other with the strain-induced MF. However, it is not possible to positively determine the origin of these two size groups of subgrains. In samples that contained partially recrystallized PF, the grain boundaries became more clearly defined, but there was little change in the grain size or the subgrain structure (Figures 7(e) and (f)). Thus, although recrystallization of PF is complete after 100 seconds at 850 7C, this structure is resistant to further grain coarsening. Precipitate distributions were studied in samples quenched at various stages of treatments 3A and 3B. The main types of precipitates observed are illustrated in Figure 8, which represents a sample after 600 seconds at 850 7C. The changes in precipitate distributions during treatments 3A and 3B can be summarized as follows: (1) At the beginning of the treatments at 850 7C, there are only large (;1 mm) cuboidal (Ti,Nb)N particles (Figure 8 at A), often with coprecipitated TixSy. During the isothermal treatment of 600 seconds at 850 7C, the TixSy transforms to (Ti,Nb)xSy (Figure 8 at B), and the coprecipitates can become an indistinguishable single (Ti,Nb)x(S,N)y phase. The (Ti 1 Nb)/S ratio varies widely, but it is initially low and achieves values in the range 1 to 2 after 600 seconds at 850 7C. (2) Samples deformed after 20 seconds at 850 7C contain very few small (30 to 100 nm) TixSy precipitates on grain boundaries of the recrystallized ferrite. (3) After 600 seconds at 850 7C, there are many 30- to 100-nm spheroidal precipitates of (Ti,Nb)xSy and (Ti,Nb,Mn)xSy within PF grains (Figure 8 at C).

1440—VOLUME 28A, JULY 1997


Fig. 8—(a ) Transmission electron micrograph of extraction replica from sample held 600 s at 850 7C: treatment 3B, showing large (Ti,Nb)N precipitate (A), coprecipitated (Ti,Nb)xSy (B), and small intragranular (Ti,Nb)xSy precipitates (C). (b ) through (d ) EDS spectra from precipitates A, B, and C, respectively.

IV.

DISCUSSION

A. Austenite Transformation Behavior The CCT experiments determined the transformation temperatures and as-transformed microstructures for two different prior austenite conditions and a range of cooling rates. Two distinct transformation products are observed. (1) Polygonal ferrite—equiaxed grains with diameters ranging from 100 to 300 mm for as-transformed PF to 300 to 600 mm for PF coarsened after transformation; a 50- to 100-mm subgrain structure indicated by weak contrast changes in ECC micrographs; transformation temperatures in the range 770 7C to 880 7C. (2) Massive ferrite—generally equiaxed grains, ;100-mm diameter, with serrated grain boundaries; subgrains 10to 15-mm diameter, with irregular boundaries; transformation temperatures in the range 540 to 730 7C. The austenite transformation temperatures agree well with results reported previously for Fe and IF steels.[5,6,7] Of particular note is the work of Wilson,[7] who has reviewed the several distinct austenite transformations in Fe and dilute Fe alloys. He reported transformation temperatures of 840 7C to 860 7C for PF and 735 7C to 760 7C for MF in Fe with varying C contents, and a continuous decrease in these transformation temperature ranges with increasing METALLURGICAL AND MATERIALS TRANSACTIONS A

substitutional solute content. The variation in transformation temperature with solute concentration is larger for MF than PF. This suggests that the effect of solute on the MF transformation is not by the usual solute drag mechanism, whereby the migration rate of the transformation interface is limited by the ability of segregated solute atoms to move with the interface by diffusion.[11] Rather, it could be dispersed solute atoms acting as obstacles to the movement of the interface. Thus, with increasing cooling rate, much lower transformation temperatures are required for the free energy of the MF transformation to be sufficient to overcome the solute obstacles. This increased separation of the temperature ranges for PF and MF with increasing cooling rate is shown by the data in Figure 1. The deformed/recrystallized austenite condition (pretreatment lB) seems to suppress both the PF and MF transformations. Thus, at low cooling rates, the PF transformation start temperature (850 7C) is slightly lower than that (880 7C) for samples cooled directly from 1290 7C (pretreatment 1A). At high cooling rates, the PF transformation is not observed and the MF transformation temperature range decreases rapidly with increasing cooling rate (Figure 3). If nucleation were the rate-controlling mechanism, the kinetics of both PF and MF transformations would be increased by decreasing the austenite grain size since both phases VOLUME 28A, JULY 1997—1441

nucleate at austenite grain boundaries. However, the difference in austenite grain size for the two pretreatments (400 vs 160 mm) is apparently not enough to produce a significant increase in the nucleation rate, and some other factor associated with the 1000 7C pretreatment must be responsible for the observed decrease in the PF and MF transformation rates. It is well known that precipitates form in deformed austenite of Ti- and Ti-Nb IF steels,[8,9,10] and it appears that the inhibiting effect of these precipitates on the growth of PF and MF outweighs the effects of either substitutional solute atoms or austenite grain size. The isothermal transformation experiments at 820 7C give additional information on the PF transformation kinetics. Under the conditions of low undercooling at 820 7C, the austenite-to-PF transformation is complete in less than 45 seconds. Taking a PF grain diameter of 150 mm and assuming a constant growth rate gives 0.003 mm s21 for PF growth rate. This is lower than the value of ;0.1 mm s21 reported for pure Fe and is closer to the value of 0.02 mm s21 reported for binary substitutional alloys of Fe.[7] Since there should be no precipitates present during isothermal PF formation at 820 7C (treatment 2), it is concluded that solute atoms limit the PF growth rate in this case. The as-transformed microstructures resulting from the continuous-cooling and isothermal treatments agree well with the accepted definitions of PF and MF.[7,12,13] The PF forms at low undercooling by nucleation at austenite grain boundaries and migration of an incoherent g-a interface. Long-range diffusion occurs and there is complete alloy partitioning. Note that in very dilute Fe alloys, the PF transformation can still occur at cooling rates of 102 7C to 103 7C to s21.[7] Transformation strains are relaxed by diffusion, and the resulting microstructure is equiaxed a grains containing a very low dislocation density. The ECC micrographs of PF show weak contrast changes, which indicate that large PF grains can be subdivided into three or four subgrains (e.g., Figure 4(b)). There are also more diffuse contrast changes within PF grains, which are not understood and may arise from the topography of the electropolished surface. The MF also forms at low undercooling, nucleated at austenite grain boundaries and by migration of an incoherent boundary. However, the MF transformation is partitionless, involving short-range atom transfer across the transformation interface, and transformation strains are relaxed by slip. The resulting microstructure is a grains with serrated boundaries (sometimes described as ‘‘ragged’’[7]) containing equiaxed subgrains with irregular boundaries (e.g., Figures 4(d) and 2(d)). It has been shown that MF subgrains have an intermediate dislocation density of the order of 109 cm22.[13] When unstable austenite is deformed at 850 7C, it transforms to a strain-induced MF, as described by Yada et al.[6] This microstructure is characterized by fine elongated grains containing ;5-mm subgrains (Figures 7(c) and (d)); i.e., it is much finer than the thermal MF formed by quenching. B. Precipitation, Recrystallization, and Grain Coarsening in PF In the present experiments, the processing schedule for the isothermal treatments at 850 7C was different from the usual

1442—VOLUME 28A, JULY 1997

processing of IF steels, in that samples were cooled directly from 1290 7C to 850 7C. This would produce supersaturation of Ti, Nb, and S and result in the observed precipitation during the hold at 850 7C of (Ti,Nb)xSy on pre-existing (Ti,Nb)N particles and as discrete precipitates within PF grains. This is a lower temperature for sulfide formation than that found for normally processed hot band (.950 7C).[8,9,10] Since the precipitates were characterized in the present study by EDS analysis only, it was not possible to determine their C contents. It is to be expected that C in such an IF steel should be stabilized in precipitate phases. The recrystallization and grain coarsening behavior of the deformed PF is quite different depending on whether samples are held at 850 7C for 20 or 600 seconds prior to deformation (treatments 3A and 3B, respectively). This is attributed to the different precipitate distributions present at deformation in these two cases. There are no small intragranular (Ti,Nb)xSy precipitates present after 20 seconds at 850 7C. The deformed PF in these samples recrystallizes essentially instantaneously, and both the recrystallized PF and strain-induced MF grains coarsen rapidly during the subsequent 100 seconds hold at 850 7C (Figures 6(e) and (f)). The distribution of (Ti,Nb)xSy precipitates present after 600 seconds at 850 7C retards recrystallization and grain coarsening of the deformed PF during the subsequent 100 seconds hold at 850 7C and produces a stable recovered subgrain structure (Figures 7(e) and (f)). These results emphasize the importance of controlling the precipitate distribution when processing IF steels by warm rolling. To achieve a strong ND//^111& recrystallization texture and good deep drawability, Senuna et al.[3] suggest removing solute C by precipitating Ti4C2S2 and (Ti,Nb)C prior to warm rolling. The present results show that precipitates present prior to deformation in the warmrolling range can also inhibit subsequent recrystallization, even at 850 7C. Another processing approach is to produce high strength sheet by warm rolling, recrystallization, and subsequent precipitation strengthening.[4] In this case, the processing schedule is designed to delay precipitation until recrystallization is complete, but this is often defeated by precipitation on the deformed PF substructure. Clearly, the relationship between precipitation, deformation, and recrystallization is complex, and this is the reason for the narrow processing window for warm rolling of IF steels.

V.

CONCLUSIONS

The laboratory TMP experiments have elucidated the following important features of the austenite transformation, precipitation, and ferrite recrystallization in a Ti-Nb IF steel. 1. For cooling rates in the range 0.1 7C s21 to 130 7C s21, austenite transforms to either PF or MF. The transformation temperatures vary systematically with cooling rate and austenite condition. There is indirect evidence that the transformation rates for both PF and MF are decreased by the presence of substitutional solute atoms and precipitate particles. 2. When unstable austenite is deformed at 850 7C, it transforms to an extremely fine strain-induced MF. 3. Under conditions of high supersaturation of Ti, Nb, and


S, (Ti,Nb)xSy precipitates form at 850 7C as coprecipitates on pre-existing (Ti,Nb)N particles and as discrete precipitates within PF grains. 4. Pre-existing intragranular (Ti,Nb)xSy precipitates retard recrystallization and grain coarsening of PF deformed at 850 7C and result in a stable, recovered subgrain structure. 5. These results demonstrate the importance of controlling the precipitate distribution when processing IF steels by warm rolling.

2. 3. 4. 5. 6.

ACKNOWLEDGMENTS The authors are grateful to L.V. Whitlock, for assistance with the dilatometer experiments, and to D.E. Overby, Stelco Steel, for supplying the IF steel. This research was funded by the Ontario Centre for Materials Research.

REFERENCES 1. G. Krauss, D.O. Wilshynsky, and D.K. Matlock: in Interstitial Free Steel Sheet, Processing Fabrication and Properties, L.E. Collins and


7. 8. 9. 10.

11. 12. 13.

D.L. Baragar, eds., Canadian Institute of Mining, Metallurgy and Petroleum, Montreal, 1991, pp. 1-14. International Forum for Physical Metallurgy of IF Steels, Iron and Steel Institute of Japan, Tokyo, 1994. T. Senuma, H. Yada, R. Shimizu, and J. Harase: Acta Metall. Mater., 1990, vol. 38, pp. 2673-81. L.M Perera, I.A. Rauf, J.D. Boyd, and S. Saimoto: in Advances in Hot Deformation Texture and Microstructures, J.J. Jonas, T.R. Bieler, and K.J. Bowman, eds., TMS, Warrendale, PA, 1993, pp. 27-39. A. Najafi-Zadeh, S. Yue, and J.J. Jonas: Iron Steel Inst. Jpn. Int., 1992, vol. 32, pp. 213-21. H. Yada, Y. Matsumura, and T. Senuma: Proc. Int. Conf. on Martensite Transformations, Japan Institute of Metals, Tokyo, 1986, pp. 515-20. E.A. Wilson: Met. Sci., 1984, vol. 18, pp. 471-84. M. Prikryl, Y.P. Lin, and S.V. Subramanian: Scripta Metall. Mater., 1990, vol. 24, pp. 375-80. N. Yoshinaga, K. Ushioda, S. Akamatsu, and O. Akisue: Iron Steel Inst. Jpn. Int., 1994, vol. 34, pp. 24-32. M. Hua, C.I. Garcia, and A.J. De Ardo: in Phase Transformations during the Thermal/Mechanical Processing of Steel, E.B. Hawbolt and S. Yue, eds., Canadian Institute of Mining, Metallurgy and Petroleum, Montreal, 1995, pp. 285-90. M. Hillert and B. Sundman: Acta Metall., 1977, vol. 25, pp. 11-18. B. Jonson and J. Agren: Acta Metall. Mater., 1990, vol. 38, pp. 433-38. M.J. Roberts: Metall. Trans., 1970, vol. 1, pp. 3287-94.

VOLUME 28A, JULY 1997—1443