High Temperature Deformation and Thermomechanical Treatment

of Low Carbon Steel and Vanadium-Niobium Microalloyed Steel*

By Masanori UEKI,** Shiro HORIE*** and Tadahisa NAKAMURA****

Synopsis

A low carbon steel and a vanadium and niobium microalloyed steel were deformed by torsion with the true strain rates from about 10_3 to 10 s-1 in the temperature range from 900 to 1200°C. Through the determina-tion o f their flow behavior and the observation of the deformed microstruc-ture, dynamic recrystallization behavior was analyzed. Austenite grain refinement by dynamic recrystallization and final ferrite

grain refinement due to microalloying elements in the course of thermo-mechanical treatments were simulated by using a torsion machine. The

final grain size in air-cooled specimen was much reduced in the schedule including only one stage deformation at 800°C particularly in the micro-alloyed steel. The microstructures in the thermomechanical processes revealed the characteristic feature of dynamic restoration process in the two phases region of steels. Especially in the microalloyed steel, austenite recrystallized partly at grain boundaries, whereas grain fragmentation like subgrain formation proceeded in ferrite.

Key words: hot deformation; low carbon steel; microalloyed steel; vanadium and niobium; thermomechanical treatment; grain refinement; dynamic recrystallization.

I. Introduction

Hot working of steels is a complicated process of mechanical working with interplay of various metal-lurgical processes such as phase transformation, and dynamic recovery and recrystallization. Grain refine-ment in hot working is achieved through microalloying combined with controlled rolling. The major role of microalloying in the controlled rolling is to retard both static and dynamic recrystallization in the course of hot rolling. Transformation from unrecrystallized elongated austenite with a high dislocation density effectively reduces grain size of ferrite. The micro-alloying elements, Al, Ti or V, inhibit grain growth through formation of carbides or nitrides before or during rolling, tt whereas Nb retards recrystallization and inhibits the subsequent grain growth by stabilizing the substructure of deformed austenite through solid-solution effect2~ or strain-induced precipitation. 3)

In the present study on the fundamental processes in hot working of steels, high temperature torsional deformation behavior is analyzed in a low carbon

steel and a V-Nb microalloyed steel from the mechan-ical and microstructural points of view. Controlled rolling of these steels is also simulated by hot torsion

test to clarify the effects of microalloying elements and deformation in a+r range on dynamic restoration

and grain refinement.

II. Experimental Procedure

The test materials were a low carbon steel and a V-Nb microalloyed steel. The billets were hot forged to solid bars of 21 mm~b at 1 100 ~ 1000°C, followed by homogenizing at 800°C for 2 h. Tubular torsion test pieces with 10 mm gauge length, 14 mm outer diameter and 10 mm inner diameter were machined from the as homogenized materials. The chemical analysis of the test materials is given in Table 1. The torsion machine was equipped with a quench-ing device4~ and a high frequency induction heating apparatus. Quenching of the test pieces for micro-structural observations was accomplished by syn-chronized supply of He-gas jet from the inner tube and water spray onto the outer surface of the test

piece. The rate of quenching was high enough to suppress static restoration. The test steels were austenitized at 950°C for 15 min and brought to a test temperature of 900, 1 000, 1 100 and 1 200°C. After holding 10 min at the test temperature, torsion test was performed at five true strain rates ranging from about 10_3 to 10 s-1. In order to check the initial austenite grain size, the metallographic examinations were also done on typical samples quenched just before testing, that is after 10 min holding at the test temperature. The

grain sizes measured by the line intercept method are shown in Table 2. Shear stress-shear strain relation was calculated from the torque-twist data by using the method described by Nakamura and Ueki5~ and then con-verted into true stress-true strain relation on the basis of the von Mises criterion.

Table 1. Chemical composition of test steels. (Wt /o )

*

**

***

****

Manuscript received on August 20, 1986; accepted in the final form on February 5, 1987. © Formerly Department of Mechanical System Engineering, Kanazawa Institute of Technology. Nippon Steel Corporation, Ida, Nakahara-ku, Kawasaki 211. Department of Materials Science and Engineering, Tokyo Institute of Technology, Nagatsuta, The Technological University of Nagaoka, Kamitomioka-cho, Nagaoka 940-21.

1987 ISIJ Now atR&D

Midori-ku,

Laboratories-I,

Yokohama 227.

Research Article (453)

( 454 ) Transactions ISIJ, Vol. 27, 1987

III. Results and Discussion

1. Hot Deformation Behavior 1. Stress-Strain Relation

Figures 1 and 2 show the typical true stress-true

strain curves in hot torsion tests of the low carbon steel and the V-Nb microalloyed steel: variations of the flow curves with strain rates at 1 000°C in the

upper half and with temperatures at a strain rate, 9.19 X 10-2 s-l, in the lower half. The flow curves for both steels exhibited a single peak, or multiple

peaks after initial work hardening with the typical flow of dynamic recrystallization. Transition from

a single peak to multiple peaks in the flow curves occurred with decreasing strain rate and increasing

temperature, as discussed in detail in previous

papers.4,6-8) 2. Strain Rate and Temperature Dependence of Flow Stress

The activation energy, Q, for deformation was determined from the temperature dependence of the

peak flow stress, o, in each flow curve at a constant strain rate. The values were 74.8 kcal/mol for the low carbon steel and 71.0 kcal/mol for the micro-alloyed steel. From these values, the o, was cor-related with deformation temperature, T, and strain rate, ~, in terms of the following power-law rela-tionship :

where, A, n : the material constants R: the gas constant.

Rewriting of Eq. (1) in,

A r = exp (QIRTT) - Z ...............(2)

results in the correlation of the peak flow stress, a, with the Zener-Hollomon parameter, Z, as shown in Figs. 3 and 4 for the low carbon and the microalloyed steels. The relationship between a and Z for each steel is expressed as a nearly straight line within the range applicable the power-law stress function. In these figures, the transition of the flow behavior from a single to multiple peaks is distinguished with open and solid marks, respectively. 3. Grain Size of Dynamically Recrystallized Austenite As apparent from the flow behavior, dynamic recrystallization takes place as a restoration process

Table 2. Initial austenite grain sizes for temperature. (µm)

each deformation

Fig. 1. Variations of torsional true stress-true strain curves

of low carbon steel.

Fig. 2. Variations of torsional true stress-true strain curves

of V-Nb microalloyed steel.

Transactions ISIJ, Vol. 27, 1987 (455)

during high temperature deformation of these steels.

The sizes of dynamically recrystallized austenite grain, DS, measured by the line intercept method on the

micrographs taken from the deformed and quenched

specimens were plotted against the Zener-Hollomon

parameter as shown in Fig. 5. The values of DS for both steels were a unique function of Z indepen-

dent of the initial grain size. Dependence of the DS on Z, i.e., the slope of DS res. Z relation, was nearly identical for both steels, but, at a certain value of

z, a finer DS was produced in the microalloyed steel than the low carbon steel. This suggests the effective-

ness of microalloying for grain refinement, as will be mentioned later. The transition of flow behavior

is also indicated in Fig. 5.

2. Simulation of Thermomechanical Treatment (TMT) and Grain Refinement of Ferrite

1. Schedule of TMT Based on the conditions of austenite grain refine-ment through dynamic recrystallization shown in Fig. 5, two types of thermomechanical treatment, schedules I and II shown in Fig. 6, were applied to both steels. The schedule I includes only one stage of deformation in a+' range, whereas II includes the two stages of deformation in r and a+r ranges. The resultant grain sizes for both ferrite and prior austenite in the quenched or air-cooled conditions were measured at the stages (0'), (1), (2) and (3) in the figure. 2. Dynamic Restoration in the Course of TMT The microstructure development during high tem-perature deformation was observed in the specimen quenched by using the device described above. Initially, the observations of the structures were

Fig. 3. Correlation of the peak flow stress with strain rate

and deformation temperature through the Zener-

Hollomon parameter, Z for low carbon steel, indi-

cating also the transition of flow stress behavior

from a single peak to multiple peaks with decreasing

z.

Fig. 4. Correlation of the peak flow stress with strain rate

and deformation temperature through the Zener-

Hollomon parameter, Z for V-Nb microalloyed

steel, indicating also the transition of flow stress

behavior from a single peak to multiple peaks with

decreasing Z.

Fig. 5. Size of dynamically recrystallized grain, Ds, as

a function of the Zener-Hollomon parameter.

Fig .6. Schematic representation of

treatments in schedules I and

thermomechanical

II.

(456) Transactions ISIJ, Vol. 27, 1987

carried out for those revealing the dynamic recrystal-lization type flow in the first stage deformation at the

point (0') in schedule II (Fig. 6), where the deforma-tion was applied by a true strain of 0.75 at 900°C in

r range; the flow curves for both steels are shown in Figs. 1(b) and 2(b). From the shape of the curves, the dynamic recrystallization seems to progress in the test. The microstructures for both steels are compared in Fig. 7, where the specimens were etched by a saturated picric acid-ethanol solution with a few drops of hydrochloric acid. Equiaxed prior austenite grains are surrounded by a small amount of ferrite phase in both steels and are much smaller in grain size in the microalloyed steel than in the low carbon steel. Further, the microstructures in the second stage deformation at the point (2) in schedule II were studied, where the deformation was applied at 800°C in a+r range by the same amount of true strain as the first stage. Figure 8 indicates the difference in the shape of flow curves between the low carbon steel and microalloyed steel, together with the stage of interrupted straining. Although the strain at peak stress, sp, was hardly changed by microalloying, the strain at a steady state of flow stress, increased markedly in the microalloyed steel. From the mechanical behavior and the amount of strain at interruption shown in the figure, the microstructure is thought to evolve in both steels in the following way: the interruption strain is high enough to complete dynamic recrystallization of auste-nite in the low carbon steel, whereas it is insufficient for the microalloyed steel because the interruption strain is much smaller than the Es.

The characteristic features of dynamic restoration above described are well represented in Fig. 9,

comparing the microstructures quenched at the stage

(2) in the schedule II for both steels. The structure is composed of ferrite (a) and prior austenite (r); the light area in micrographs is a phase and the dark one either bainite or martensite transformed from r, where the specimens were etched by the same solution as described above. The r phase has completely recrystallized dynamically in the low carbon steel as shown in Fig. 9(a), but only a small fraction recrys-

Fig. 7. Deformed and quenched microstructures in first 'stage of deformation in TMT schedule

(corresponding to (0') in Fig. 6).

the

II

Fig. 8. Flow curves in

TMT schedule

the second stage

II.

of deformation in

Fig. 9. Deformed and quenched microstructures in second stage of deformation in TMT schedule

(corresponding to (2) in Fig. 6).

the

II

Transactions ISIJ, Vol. 27, 1987 (457)

tallized in the microalloyed steel as shown in Fig. 9(b). This indicates the retardation effect of Nb and V on dynamic recrystallization in the austenite. The effect was reflected in the increased ~S in the flow curve due to microalloying elements. The com-

pounds presumably formed in the microalloyed steel are shown in Table 3 with their calculated dissolving temperatures ;9) all the precipitates remain at 800°C without dissolving in the matrix. These compounds will inhibit grain boundary migration, leading to the retardation of dynamic recrystallization in the micro-alloyed steel. Furthermore, although the a phase shows no evidence of dynamic restoration in the low carbon

steel, some of ferrite grains reveal subboundaries

(or newly formed grain boundaries) within grain interior in the microalloyed steel. It seems that the strengthening of austenite grain due to these com-pounds activated the dynamic restoration of ferrite. Such a fragmentation of grain has also been observed in the microstructure of a 17 % Cr ferritic stainless steel10~ deformed at a high temperature, characterizing the microstructural evolution of the ferritic alloys similar to the higher stacking fault energy fcc metals as reported by the authors." 3. Microstructure in the Course of TMT and Final Grain Refinement

The microstructures quenched or air-cooled in the course of TMT schedules were compared in Fig. 10 for the schedule I and Fig. 11 for II. The measured grain sizes of both ferrite and prior austenite corresponding to the above micrographs are sum-marized in Table 4, including the volume fraction of each phase.

The grain refinement attained by the TMT can

Table 3. Compounds formed in the microalloyed steel and

their dissolving temperatures.

Fig . 10. Comparison of microstructures of TMT schedule I : (1), (2)

for low and (3)

carbon steel

indicate the

and microalloyed steel

stages in Fig. 6.

in the processing

Fig . 11. Comparison of microstructures of TMT schedule II: (1), (2)

for low and (3)

carbon steel and microalloyed steel

indicate the stages in Fig. 6.

~~

in the processing

(458) Transactions ISIJ, Vol. 27, 1987

be evaluated by comparing the final grain size at stage (3) in each schedule; it was smaller in the schedule I compared to that in schedule II as shown in Table 4. The comparison can also be made among the steels; the finer grains were produced in the microalloyed steel through both schedules.

In the TMT schedule II, a phase appears already in the stage (1) to a higher amount for both steels whereas it is slightly observed in the schedule I. This means that deformation in the first stage may promote the r -~ a transformation at a temperature above Ar3 temperature. The amount of a phase apparently increases by the deformation as shown in Table 4 and the microstructure at the processing stage (2), which may be caused by elevation of the Ar3 tempera-ture by deformation as reported by Tanaka and Tabata12~ and Ouchi et a1.13~

The initial austenite grain sizes were larger at processing stage (1) in schedule II than in schedule I; it is approximately twice as large, in spite of rela-tively small DS (Fig. 7) in the point (0'). Air-cooling and the subsequent annealing (at 800°C for 5 min) followed by deformation at the first stage seemed to

promote grain growth of the austenite. The initial austenite grain sizes may affect largely to the kinetics of the dynamic recrystallization of r in the subsequent final deformation stage; a smaller austenite grain size promotes recrystallization by decreasing both 6p and ES as mentioned by Sakai and Jonas.l4) How-ever in the case of low carbon steel, since the ss is much lower than the interruption strain of 0.75, there is no difference in the microstructure in the processing stage (2) for both schedules as seen in the micrographs of Figs. 10 and 11. Significant difference in the microstructures between the schedules is shown in the microalloyed steel. The progress of dynamic recrystallization up to the interruption strain is much faster for the schedule I than for the schedule II, as seen in the micrographs (2) in Figs. 10 and 11. The difference in the fraction recrystallized for both schedules are also shown in Table 4; 51 % for schedule I and 15 % for schedule II.

Although it is proposed15) that finer grains and unrecrystallized r grains are both effective to produce fine ferrite grains at the final stage (3), the dynamically recrystallized fine r structure in the schedule I might be much effective for the final ferrite grain refinement in the present case. Comparing the final micro-structures in the stage (3), even in the low carbon steel, a considerable grain refinement was achieved because of fine austenite grains provided by dynamic recrystallization. Although the final grains appear finer in the low carbon steel, extremely fine grains are mixed in the microalloyed steel, as shown in Fig. 12. Such a mixed grain structure must be formed by the transformation from the austenite grain during

Table 4. Initial

grain

grain size, grain

size of ferrite by

sizes in the processing stage of the the treatment. (µm)

thermomechanical treatment and finally attained

Fig . 12. Comparison of

alloyed steel in

processing stage

final microstructures for the air cooled condition

(3) in Fig. 6.

micro-

at the

Transactions ISIJ, Vol. 27, 1987 (459)

dynamic recrystallization at the final deformation stage.

I V. Conclusions

High temperature deformation behavior was exam-

ined in the low carbon steel and the V-Nb micro-alloyed steel, by using the torsion testing of tubular

test pieces. These steels exhibited a typical flow of

dynamic recrystallization in the present experimental range. The size of dynamically recrystallized grain

could be represented as a function of the Zener-Hol-lomon parameter independent of the initial grain

size. Thermomechanical treatment was simulated by

using the torsion machine for both low carbon and V-Nb microalloyed steel. Two simulation schedules

were employed; deformation in one stage or two stages. The final grain size in air-cooled specimen was extensively reduced in the schedule including

only one stage deformation at 800°C particularly in the microalloyed steel. Dynamically recrystallized

fine austenite acts as the nucleation site of ferrite more effectively for the final grain refinement of

ferrite than the unrecrystallized austenite. The microstructures in the thermomechanical

processes revealed the characteristic feature of dynamic restoration process in the two phases region of steels. Especially in the V-Nb microalloyed steel, recrystal-

lization in austenite was found to be retarded signifi-candy due to the presence of Nb and V, and the

fragmentation of grain (or observed in ferrite phase.

1)

2)

3)

4)

5) 6)

7)

S) 9)

10)

11)

12)

13)

14) 15)

subgrain formation) was

REFERENCES

K.J. Irvine, T. Gladman, J. Orr and F. B. Pickering:

JISI, 208 (1970), 717. A. Le Bon, J. Rofes-Vernis and C. Rossard: Met. Sci., 9

(1975), 36. II. Kubota, I. Kozasu, T. Shimizu, S. Murai and Y. Kaneko: Nippon Kokan Tech. Rep., 46 (1969), 205. M. Ueki, M. FIattori, S. Hone and T. Nakamura : Tetsu-to-Hagane, 72 (1986), 482. T. Nakamura and M. Ueki : Trans. ISIJ, 15 (1975), 185. M. Ueki, S. Hone and T. Nakamura: Script. Met., 19

(1985), 547. M. Ueki, M. Hattori, S. hone and T. Nakamura: Trans. ISIJ, 26 (1986), 906. M. Ueki: J. Mat. Sci. Lett., 5 (1986), 1263. W. C. Leslie: The Physical Metallurgy of Steels, Mc-Graw-PIill, New York, (1981), 135. M. Ueki: Doctoral Thesis to Tokyo Institute of Tech-nology, (1976), 97. M. Ueki, S. Hone and T. Nakamura: Microstructure and Mechanical Behaviour of Materials, ed, by G.1-Iaicheng and I-I. Jiawen, Chameleon Press, London, (1986), 473. T. Tanaka and N. Tabata : Tetsu-to-Hagane, 64 (1978), 1353. C. Ouchi, T. Sampei and I. Kozasu: Tetsu-to-Hagane, 67

(1981), 143. T. Sakai and J.J. Jonas: Acta Met., 32 (1984), 189. M. Umemoto, H. Ohtsuka and I. Tamura: Trans. ISIJ, 23 (1983), 775.