Recrystallization during. Interstitial Free Steels. A. NAJAFI-ZADEH, J.J. JONAS, and S. YUE. Multipass torsion tests were used to investigate the warm-rolling ...
Grain Refinement by Dynamic Recrystallization during the Simulated Warm-Rolling of Interstitial Free Steels A. NAJAFI-ZADEH, J.J. JONAS, and S. YUE Multipass torsion tests were used to investigate the warm-rolling behavior of three interstitial free (IF) steels containing Ti and/or Nb. All the tests were carried out at a strain rate of 2 s -l, and the samples were water-quenched immediately after particular passes. Whereas the finishing passes were always executed in the single-phase ferrite region, the roughing passes were applied either in the austenite phase (hot/warm-rolling) or in the ferrite region (warm/warm-rolling). In the former case, the temperature of the first roughing pass was 1260 ~ while it was 850 ~ for the warm/warm method. It is shown that dynamic recrystallization occurs to a degree that depends on the composition of the steel and the finishing temperature. Although the finishrolling loads for warm-rolling (in the ferrite region) are no higher than for conventional striprolling in the austenite region, the ultrafine ferrite grain sizes of 1 to 2 / z m that are produced by warm-rolling are an order of magnitude finer than those resulting from conventional processing.
I.
INTRODUCTION
IT is generally accepted that the hot deformation of ferrite is controlled by dynamic recovery processes alone. The reports of dynamic recrystallization in the ferrite range have been restricted to its occurrence in zone-refined or vacuum-melted iron. l~'21 However, Maki et al. t31 have shown that dynamic recrystallization can take place in 430 ferritic stainless when it is deformed at temperatures above 1000 ~ With regard to the warm-rolling of conventional steels, once again the absence of significant work hardening is generally attributed to the ease of dynamic recovery. In the present study, the microstructural changes taking place during simulated warm-rolling were studied on three commercial interstitial free (IF) steels. It was found that grain refinement occurred by means of dynamic recrystallization which led to ferrite grain sizes as low as 1 to 2 pm. Such ultrafine grain sizes can raise the low yield strength of commercial IF steels and improve the dent resistance 141 in this way. It is of interest that the present procedure for refining the ferrite grain size differs from that of Yada et al. of Nippon Steel Corp., Japan, who patented a technique for the production of ultrafine grain sizes in plain C-Mn steels, t51 According to their method, ultrafine grained ferrite is produced when such steels are rolled in the intercritical region (i.e., in the austenite-plus-ferrite, two-phase region). They attributed the grain refinement to the dynamic transformation of austenite to ferrite as well as to the dynamic recrystallization of ferrite. The former mechanism, in which the ultrafine grained ferrite is produced as a result of the repeated nucleation of ferrite at austenite grain boundaries, was considered to dominate, with the dynamic recrystallization of ferrite playing a minor role. They also showed that such dynamic recrystallization only takes place in the intercritical (t~ + 3') region in their materials. A. NAJAFI-ZADEH, Professor, is with the Department of Materials, Isfahan University of Technology, Isfahan, Iran. J.J. JONAS, CSIRANSERC Professor of Steel Processing, and S. YUE, Assistant Professor, are with the Department of Metallurgical Engineering, McGill University, Montreal, PZ, Canada H3A 2A7. Manuscript submitted September 24, 1991. METALLURGICAL TRANSACTIONS A
II.
EXPERIMENTAL PROCEDURE
A. Materials and Techniques
The chemical compositions of the three IF steels studied are given in Table I. These have the same base composition of approximately 0.004 wt pct carbon, 0.03 wt pct silicon, 0.15 wt pct manganese, and 0.003 wt pct nitrogen. The major difference in composition involved the type of carbon stabilizing element employed, i.e., titanium and/or niobium. As can be seen in Table I, the first grade is Ti stabilized; in this material, Ti combines with N and S prior to scavenging C, so that the minimum Ti level required to fully stabilize the steel is given by the following relationship: 48 48 Ti (wt pct) - 7-7. N (wt pct) + ~ S (wt pct) 32 14 48 + - - C (wt pct) 12
[ 1]
In the Ti-Nb IF steels, the usual practice is to add sufficient Ti to react with N to form TiN, leaving Nb to scavenge C as NbC. The third steel was a niobium stabilized IF grade; in such materials, carbon is scavenged by Nb, whereas N reacts with A1 to form AIN. Sulfur in such steels combines with Mn to form MnS. To define the amounts of Ti and/or Nb available to combine with C as well as the quantities of excess Ti and/or Nb likely to remain in solution, the following relationships were employed: 48 48 Ti + = Ti (wt pet) - - - N (wt pct) S (wt pct) 14 3-2
[2]
Nb § = Nb (wt pct)
I3]
Ti* = Ti § - 4C (wt pct)
[4]
93 Nb* = Nb § (wt pct) - --2 C (wt pct) 12
[5]
where Ti and Nb refer to the total titanium and niobium VOLUME 23A, SEPTEMBER 1992--2607
Table I.
Chemical Compositions of the Three IF Steels (Weight Percent)
IF Steel
C
Si
Mn
S
N
Ti
Nb
Ti + + Nb +
Ti* + Nb*
Ti Ti-Nb Nb
0.0035 0.0035 0.0036
0.022 0.021 0.024
0.15 0.15 0.16
0.012 0.011 0.013
0.003 0.003 0.003
0.065 0.035 --
-0.028 0.056
0.037 0.036 0.056
0.023 0.016 0.028
Ti§ or Nb§ refers to the Ti or Nb available for combination with C. Ti* or Nb*: refers to the excess of Ti or Nb above the amount required to stabilize C, N, and S.
contents, Ti § or Nb § to the Ti or Nb available to combine with C, and Ti* or Nb* to the excess of titanium or niobium over the amount required to stabilize C, N, and S. The present materials were taken from as-cast slabs supplied by Stelco Steel, Hamilton, PO. Torsion test specimens, with gage lengths of 23 mm and diameters of 6.2 mm, were machined from the as-cast slabs. The torsion experiments were performed on a servohydraulic, computer controlled MTS machine equipped with a Research Incorporated radiant furnace controlled by a Leeds and Northrup temperature programming system. All the torsion tests were performed in a high-purity argon atmosphere to minimize oxidation. Further details of this apparatus are presented elsewhere. 16] Stress-strain curves were determined from the torque-twist behavior in the usual manner, t7~ Prior to testing, all specimens were austenitized at 1200 ~ for 10 minutes in a high-purity argon atmosphere and then normalized by cooling at 5 ~ to 10 ~ to room temperature. The purpose of this treatment was to standardize the thermal and mechanical histories of the specimens. After torsion testing, the samples were mounted and polished so that the subsurface of the gage length was just revealed. Specimens were quenched after testing and subsequently etched in 2 pet nital and Marshall's reagent in order to reveal the final ferrite microstructure. Because of the ultrafine scale of the microstructure, the surfaces of the specimens must be well prepared (i.e., polished down to 0.05/zm) before etching. In most cases, it was necessary to polish and re-etch the specimens several times until the grain structure was satisfactorily revealed over the entire surface. The grain sizes were then measured using standard linear intercept techniques.
Table II. Simulated Strip-Rolling Schedules Used in the Present Work* Pass Number
Equivalent Strain per Pass
Delay Time between Passes (s)
R1 R2 R3 R4 R5 R6 R7 F1 F2 F3 F4 F5
0.23 0.25 0.23 0.29 0.39 0.76 0.56 0.41 0.53 0.55 0.55 0.55
3.5 8 10 12 13 18 150 or 300 3.5 2.5 1.7 0.8 --
*The strain rate was 2 s -~ for each pass. R: refers to roughing. F: refers to finishing.
1260 ~
/
15 rain. ] ,
The pass strains and interpass times used are presented in Table II. This schedule involves seven roughing and five finishing passes and only differs from the strip-rolling schedules employed in the Lake Erie Works of Stelco Steel with regard to the strains associated with the last three finishing passes. These are 0.42, 0.4, and 0.3 for F3, F4, and F5, respectively, at the Lake Erie strip mill, whereas the strain in the last three finishing passes in the present simulations is 0.55.18'9] The total strain employed in the present simulations is thus approximately equal to that which can be applied in a 6 or 7 stand finishing mill. All the simulated passes were executed at an equivalent strain rate of 2 s- , and the specimens were quenched into water about 2 s after deformation.
2608--VOLUME 23A, SEPTEMBER 1992
passes
300s delay time
15 rain
::Ar3 ::
::890 ~ i
i Art
860 ~
Temp.
B. Laboratory Test Schedules
Roughing
/
i ..............
I
passes
150s
-
[First~ ] temp. is well I
[
theArt.
10rain
Water quenching
Time
Fig. I--Coolingcyclesfollowedin the two methods of rollingthe present IF steels.
METALLURGICAL TRANSACTIONS A
250
250
EST' l~t~F
2OO
200
150
150
_
Stress
Stress (MPa)
(M:Pa) ii!! ':_ .,,+ ! 100
100
+
:
$
.
:
50
50 i
2.5
3
~i
i
I
3.5
i
4
I
i
|
i
4.5
*
l
0 5.5
5
2.5
3
~,
i
a
,,
3.5
4
Strain
Strain
(a)
(b)
i
I
l
4.5
i
i
i
i
i I
5
5.5
250 qEST,
200
150
[.
Stress (MPa) 100
**
+
§
9
@
t
!
"t *+
t"
50
:~. , ,
I |
2.5
3
a |
I
|
*
I
3.5
4
4.5
5
5.5
Strain
(c) Fig. 2 - - F l o w c u r v e s for the Ti steel finish-rolled a c c o r d i n g to the h o t / w a r m - r o l l i n g m e t h o d with TR, = (b) T ~ = 6 7 0 ~ and (c) TF~ = 590 ~
300
977 ~
838 ,
727 ,
636 ~ Temp. ~
...............
250
Stress
. . . . . . . ;"...... ! r~ ] L~0"O~ ', ",..r.": .... ~ + ,
150
s~.
', i n t h e l a s t t h r e e : sJ : finishing passes ',"~"~ s S
]
(MPa)
50
,
i
is~ ~
, I
.8
. ~
',
~_~]
"" ~ ~C FI =710 ~
*~ f f M ' ~ ] ~ s ~
...
." . v~.'~'FI=680 !~
496
~s S~
~ AverageMFS ~
j""
Mean Flow
560 t
]
~
..... i-n. . . . . . . . . t h e first ' finishingpass
'MFS'
. ................
F1 =790~
i i
J
i
a
.9
1
1.1
1.2
1.3
1000fr ( K I) Fig. 3--Influence of finishing pass temperature on mean flow stress of the Ti stecl-rolled according to the hot/warm method for a corrected strain rate of ~ = lO0/s.
METALLURGICAL TRANSACTIONS A
1260 ~
and (a) Tr~ = 7 1 0 ~
Two different roughing procedures were employed, according to which the first seven passes were executed either above or below the Ar3/Arl temperature range. In both cases, the finishing passes were applied in the warmrolling (ferrite) range. These methods are designated here as hot/warm- and warm/warm-rolling (Figure 1). In an industrial scenario, the higher roughing temperature of the hot/warm method requires a longer delay time between roughing and finishing which allows the temperature of the steel to decrease below the At, to enable finish-rolling to be performed in the single-phase ferrite region. Thus, a delay time between roughing and finishing of 300 s was employed for the hot/warm method compared to 150 s for warm/warm-rolling. Over the finish-rolling temperature range, an average cooling rate of 3 ~ was used. After each simulation, the specimens were retested in torsion at room temperature in order to determine the yield strengths associated with each as-hot-rolled microstructure. For this purpose, the method developed by Yue e t a l . I~~ was employed.
VOLUME 23A, SEPTEMBER 1 9 9 2 - - 2 6 0 9
Fig. 4 - - D e v e l o p m e n t of ferrite dynamic recrystallization in the Ti steel when hot/warm-rolled with TR, = 1260 ~ and TF, = 770 ~ quenched after (a) first finishing pass, (b) second finishing pass, (c) third finishing pass, and (d) last finishing pass.
III. BEHAVIOR OF IF STEELS UNDER HOT/WARM-ROLLING CONDITIONS
A. Effect of Finish Start Temperature on the Stress-Strain Curves The influence of first finishing pass temperature on the stress-strain curves associated with the five finishing passes is illustrated in Figure 2 for the Ti steel. As can be seen in the diagram, there is an accumulation of strain, i.e., of work hardening, from the first to the second pass. There is no further increase in flow stress beyond the second pass, despite the approximately 15 ~ decrease in temperature associated with the final three passes. This lack of increase in flow stress suggests that some form of dynamic recrystallization was taking place during simulated rolling, as in the case of austenite, IH,~21leading to a decrease in the isothermal flow stress and offsetting the effect of the decrease in temperature in this way. It should be noted that although the reduction per pass is less than that required to initiate dynamic recrystallization, the relatively low temperatures and short interpass times (1 to 2 s) allow the dislocation density to accumulate until it reaches and exceeds the critical level required for the propagation of this type of recrystallization. 2610--VOLUME 23A, SEPTEMBER 1992
Water-
B. Effect of Pass Temperature on the Mean Flow Stress under Strip-Rolling Conditions For the prediction of industrial rolling loads, the mean flow stress pertaining to each pass must be corrected for the actual strain rate experienced in the mill. For this purpose, the general relationship between strain rate ~, temperature T, and flow stress 0" can be expressed by the following equation: [6] = A[sinh (a0")]" exp ( - Q / R T ) Here A, a, and n are constants, Q is the activation energy for hot deformation, and R is the universal gas constant. At high stresses and relatively low temperatures, Eq. [6] reduces to an exponential relation: = B exp (/30") exp ( - Q / R T )
[7]
where/3 = an. If oq and 0"2 are the mean flow stresses at strain rates ~, and ~2, where ~ is the simulation and e2 the hot strip mill strain rate, Eq. [7] leads to the relation: e2 - - -- exp [/3(o'2 - 0-0]
at T = constant
or ~2
0"2 = 0"1 + C l o g _
[8]
METALLURGICAL TRANSACTIONS A
Fig. 5 - - D e p e n d e n c e of ferrite grain size of the Ti steel on first finishing pass temperature when rolled according to the h o t / w a r m method with TR~ = 1260 ~ and (a) T~ = 710 ~ (b) T~ = 670 ~ and (c) TF~ = 590 ~
where C = 2.303//3. In the present work, the constant C took the values 35 and 24 for the austenite and ferrite regions, respectively, when the stress units were MPa. It is possible to predict the mean flow stresses associated with a particular mill with the aid of Eq. [8]. The effect of finish pass temperature on the expected mean flow stress under mill conditions is depicted in Figure 3, where the finishing strain rate has been set at 100 s -~ for simplicity. This figure shows that the mean flow stress for the first finishing pass increases approximately linearly with l/T, as does the average for the last three passes. Thus, the following relationship can be used to link the mean flow stress and finishing temperature: or=A+-
B T
where A and B are constants and the stress is in MPa. Here, A = - 2 2 0 and B = 3.6 x 105 (with a correlation coefficient r = 0.995) for the first finishing pass and A = - 3 5 5 and B = 5.2 • 105 (with a correlation coefficient r = 0.997) for the average flow stress of the last three finishing passes. It can be seen from Figure 3 that the difference between the mean flow stress of the first and later finishing METALLURGICAL TRANSACTIONS A
passes diminishes as the temperature is increased. This indicates that the amount of work hardening retained during the interpass time interval decreases at the higher temperatures.
C. Microstructural Observations 1. Progress of dynamic recrystallization in the ferrite To confirm that dynamic recrystallization occurred in the ferrite during warm-rolling, several tests were carried out in which the specimens were quenched after specific passes. The results obtained on the Ti steel when rolled with a first finishing pass temperature of 770 ~ are depicted in Figure 4. As can be seen from Figure 4(a), the ferrite grain structure after the first finishing pass is elongated and some bulging of the grain boundaries can be observed. After the second finishing pass, numerous small grains have formed along the original boundaries, as shown in Figure 4(b). The density of new dynamic grains increases after the third finishing pass (Figure 4(c)), and after the fifth finishing pass, the original structure is almost totally refined by the occurrence of dynamic and metadynamic recrystallization (Figure 4(d)). VOLUME 23A, SEPTEMBER 1992--2611
Fig. 6 - - M i c r o s t m c t u r e of the Nb steel when rolled according to the h o t / w a r m method with TR~ = 1260 ~
(a)
and TF~ = 820 ~
(b)
(c) Fig. 7 - - ( a ) Stress-strain curves for the Ti steel processed according to the w a r m / w a r m method with TRj = 850 ~ and TF~ = 700 ~ (b) Stressstrain curves for the Ti-Nb steel processed according to the w a r m / w a r m method with TRt = 850 ~ and TrL = 700 ~ (c) Stress-strain curves for the Nb steel processed according to the w a r m / w a r m method with TR~ = 850 ~ and TF~ = 700 ~ 2612--VOLUME 23A, SEPTEMBER 1992
METALLURGICAL TRANSACTIONS A
z7o
.
150
+
12o
Ti-Nb
689
653
Temp.~
i
steel ~
"
',
150 Mean Flow Stress I00 (MPa)
§
t'
"
727
--- .Nbsteel
§
130
....
200
§
140 Stress (MPa)
769
Ti steel
/ ( "r
160
814
863 250
4r §
i
50
t'
i
:
i Finlshing i ,,~ passes 9
Roughing passes
+
110
+
100
~
.~,
2.5
Is
I |
3
-.*
L I
11
3.5
I i
4
I~
,f
I i
4.5
i i
.......
5
5.5
.88
I
I
l
I
.92
.96
1
1.04
1.08
1000~ ( K 1)
6
(a)
Strain
(a)
1290
250
1198 ,
1043
1116
(.
Ti steel
- - - - i Nb
160
Temp. ~
steel
150 i
Flow Stress I00 (MPa)
~..4
Finishing
50 ,9
120
863
Mean
P (NPa) 130
917
.... !~:N~i'~i!:.
200
i '
170
977
i
'
~
'§
-.
%
§247
0
$
.64
§
3
I
I
I
.68
.72
.76
.8
.84
i
i
.88
+
§
(b)
"t
2.5
~mes
1
lOOOfr ( Ka)
110 : i 100 ;"~". . . . . . . .
Roughing passes
~ ............ 3.5
4
Ill
4.5
5
i
I L l |
5.5
Strain
(b) Fig. 8 - - ( a ) Finishing flow curves for the Ti-Nb steel rolled according to the w a r m / w a r m method with six finishing passes and TR~ = 850 ~ and Tet = 700 ~ (b) Finishing flow curves for the Ti-Nb steel rolled according to the w a r m / w a r m method with six finishing passes and TR~ = 850 ~ In this case, a constant finishing temperature of 700 ~ was employed.
It should be pointed out that although the microstructures of Figure 4 look as if they could have resulted solely from static recrystallization, such an interpass softening process is inconsistent with the shapes of the flow curves in Figure 2 and with the flow stress levels in Figure 3. 2. Effect of pass temperature on the microstructure The influence of first finishing pass temperature on the ferrite grain size of the Ti steel is shown in Figure 5. The stress-strain curves corresponding to these microstructures have been presented in Figure 2. When the first finishing pass temperature was 710 ~ (Figure 5(a)), an average ferrite grain size of 1.8 /zm was produced. When TFI was further lowered to 650 ~ and 590 ~ the grain size decreased to 1.5 /xm (Figure 5(b)) and 1.3/zm (Figure 5(c)), respectively. It is evident from the figure that the ferrite grain boundaries are fairly irregular METALLURGICAL TRANSACTIONS A
Fig. 9 - - ( a ) Influence of inverse pass temperature on MFS for the three IF steels rolled according to the w a r m / w a r m method with TR~ = 850 ~ and Tr~ = 700 ~ Here, the strain rates have been corrected to ~R = 20/S and ~r = 100/s. (b) Influence of inverse pass temperature on MFS for the three IF steels rolled conventionally with TR~ = 1260 ~ and Tr~ = 960 ~ The strain rates have been corrected to eR = 2 0 / s and er = 100/s.
and that the grain size is not uniform. These microstructural features are similar to those associated with the dynamic recrystallization of austenite. The ferrite microstructure of the Nb steel when rolled according to the hot/warm-rolling method with a first finishing pass temperature of 820 ~ is shown in Figure 6. Some coarse ferrite grains can be seen to remain in the microstructure. This appears to be related to the high degree of softening that takes place in the interpass intervals at the higher rolling temperatures (Figure 3). That is, when higher first finishing pass temperatures are used, it becomes more difficult to accumulate sufficient strain to ensure that dynamic recrystallization penetrates entirely through the original microstructure. The microstructures and associated stress-strain curves presented above show that dynamic recrystallization occurs in the ferrite of IF steels. This raises the question of why this mechanism is not observed in the ferrite of plain carbon steels. A possible answer is that the presence of conventional carbon levels leads to sufficient strain-induced carbide precipitation for the inhibition of VOLUME 23A, SEPTEMBER 1992--2613
Fig. 1 0 - - F e r r i t e microstructures of the three IF steels rolled according to the w a r m / w a r m method with TR~ = 850 ~ steel, (b) Ti-Nb steel, and (c) Nb steel.
and Te~ = 700 ~
(a) Ti
dynamic recrystallization. By contrast, when the C and N concentrations reach sufficiently low levels, the relative absence of precipitation permits the dynamic recrystallization of ferrite to occur.
IV. B E H A V I O R OF IF STEELS U N D E R WARM/WARM-ROLLING CONDITIONS A. Flow Stress
The flow curves for the three IF steels rolled totally in the ferrite region are presented in Figures 7(a) through (c). Here, the first roughing and finishing passes were carried out at 850 ~ and 700 ~ respectively. In the roughing passes, the flow stresses of the Ti and Ti-Nb steels are similar, whereas the Nb steel flow stresses are about 10 MPa higher in each pass. When these ferrite roughing stress-strain curves are compared with those of conventional (austenite) hotrolling, t12.~3~ it is evident that the behaviors of the first five roughing passes are similar, whereas those of passes 6 and 7 are different. In conventional hot-rolling, when the strains of passes 6 and 7 are equal to or greater than 0.5, dynamic recrystallization takes place in the 2614--VOLUME 23A, SEPTEMBER 1992
Fig. 11 - - E f f e c t of ferrite grain size on yield stress of IF steels.
austenite. By contrast, under warm/warm-rolling conditions, there is sufficient interpass softening, despite the lower temperatures, to prevent the accumulation of enough work hardening to initiate dynamic recrystallization. In the finishing passes of the three steels (Figures 7(a) through (c)), it is evident that there is an accumulation of work hardening from the first to the second pass, after METALLURGICAL TRANSACTIONS A
3
I0
I
I
I
I
lllt
I
y
I0
I
I
Because of the radial temperature gradients that develop in torsion specimens when these are subjected to continuous cooling, it is difficult to be certain of the mean temperature that applies to each pass and, therefore, to construct accurate mean flow stress vs 1 / T diagrams. Thus, the conclusion that dynamic recrystallization was taking place during the finishing passes of Figure 7 required further confirmation. For this reason, a series of tests was performed in which all the finishing passes were executed isothermally at 700 ~ The flow curves for the Ti-Nb steel finish-rolled under continuous cooling and under isothermal conditions are presented in Figures 8(a) and (b), respectively. As can be seen from Figure 8(b), the isothermal flow curves exhibit a maximum in the second finishing pass followed by flow softening. The shape of the envelope of these curves is characteristic of dynamic recrystallization and thus lends strong support to the interpretation of Figure 7.
I III
region
2 a regaon
Mean Flow Stress (MPa)
! .............. ! i ...............
i test results omA i
o * i Ti steel ::
1
i
/: = 2/s
i .............
J
r ....."; .................7-'I
A 9 iT i - N b steel i !..................
t. . . . . . . . . . . . . . . . . . . . . . .
a . i N b steel
o o A istrain rate correctlon i ! ~ = 1001s ,
lO
I
I
I
I
J
I I llll
0
10
I
10
I
I
I
I
B. Prediction of Mean Flow Stresses under Mill Conditions I
I
2
I
lO
Ferrite Grain Size (pro) Fig. 12--Relationship between the average MFS of the last three finishing passes and the dynamically recrystallized grain size for the aust e n i t e 1~2,~31 a n d f e r r i t e r e g i o n s .
903
800 '
680
I
'
I
596 '
I
527 '
Temp. ~ o 9 T i steel o 9 Ti-Nbsteel ~ 9
Ferrite Grain Size
\
,. o
(pro)
i -with ', ~ = 100Is
0
, .85
l te;t-;esalt---:] ~
1 -
t .......
I
Nbsteel
9
C. Microstructural Observations
I with ~ = 21s I
- "~ ~
_ -~_
J
,
.95
I 1.05
,
I 1.15
As was done for Figure 3, a correction was applied to the experimental mean flow stresses to account for the differences between the mill and laboratory strain rates. The results of these calculations are illustrated in Figure 9 for roughing and finishing strain rates of 20 and 100 s-', respectively. The warm/warm-rolling flow stresses (Figure 9(a)) can be compared with those corresponding to conventional strip-rolling, I121 which are presented in Figure 9(b). It can be seen that the roughing passes of warm/warm-rolling lead to mean flow stresses approximately 25 MPa higher than in conventional rolling, whereas the mean flow stresses are about the same for both methods in the finishing passes. These results indicate that the separation forces in the finishing passes of warm/warm-rolling need not differ significantly from those experienced in conventional hotrolling.
1.25
1000/T ( K-q Fig. 13--Dependence of ferrite grain size on inverse absolute temperature of the first finishing pass.
The ferrite microstructures of the three IF steels when deformed according to the warm/warm-rolling schedule are illustrated in Figure 10. In the Ti steel, the grains are relatively equiaxed, whereas in the Ti-Nb and Nb steels, many of the grains are elongated in the rolling direction. Figure 10 also indicates that the ferrite grain size decreases from the Ti through the Ti-Nb to the Nb steels, the average grain sizes being 2.4, 1.9, and 1.7/xm, respectively. Thus, the effect of Nb on decreasing the ferrite grain size is more pronounced than that of Ti; this is probably related to the difference in their solute drag effects. ItSj
V.
which the flow stress remains at about the same level. As pointed out previously, the lack of increase in flow stress with decreasing temperature indicates that dynamic recrystallization is taking place during deformation in these steels, followed by only minimal amounts of metadynamic recrystallization.tJ4] METALLURGICAL TRANSACTIONS A
FERRITE
GRAIN
SIZE
AND M E C H A N I C A L P R O P E R T I E S The small size of the torsion samples did not permit the machining of standard tensile specimens for subsequent testing. Nevertheless, it was possible to estimate the room-temperature yield strength in tension from an empirical relationship that was developed between the V O L U M E 23A, S E P T E M B E R 1 9 9 2 - - 2615
tensile and torsional yield strengths, tl~ The predicted tensile yield strengths are shown in Figure 11. Here, the curves of Picketing and Gladman t~61 relating the yield stress and impact transition temperature of mild steel vs the ferrite grain size are also displayed. Using regression analysis, the following relationship was deduced from the present data for the IF steels: O'y = 83 + 435dS ~/2 Here O-yis in MPa, d~ is in/zm, and the correlation coefficient r = 0.97. From these data, yield stresses of 430 and 518 MPa with impact transition temperatures of - 1 8 0 ~ and - 2 4 0 ~ can be expected when the ferrite grain sizes are 1.5 and 1 /zm, respectively. It should be noted that the previous relation cannot be compared directly with expressions determined on statically recrystallized or isothermally transformed materials which generally contain lower levels of dislocations. When dynamic recrystallization is taking place, as in the present IF steels, the grains contain a spread of dislocation densities which are expected to contribute a component of substructure strengthening and which may effect the fracture properties as well. A. Effect of Mean Flow Stress on Ferrite Grain Size The relationship between the ferrite grain size and the average mean flow stress of the last three finishing passes for the two rolling methods is displayed in Figure 12. Here, the Ti, Ti-Nb, and Nb steels are represented by full rounds, triangles, and squares, respectively. Using regression analysis, the following relationship was deduced: d~ = 1.54
•
103o'~-1"3
with a correlation coefficient of r = 0.87. Here, d~ is the ferrite grain size (/xm) and o'4 is the average mean flow stress of the last three finishing passes. To predict grain size in a strip mill, the mean flow stresses developed during processing are required; these can be obtained from Eq. [8]. The predictions are also displayed in Figure 12, where the expected strip mill values are represented by open rounds, triangles, and squares for the Ti, Ti-Nb, and Nb steels, respectively. From this figure, it is clear that finer ferrite grain sizes can be expected in strip mills than in torsion tests because of the higher strain rates. In Figure 12, the data of Najafi-Zadeh et al. 112'131for the same IF steels rolled in the austenite range are also presented. The figure indicates that for the same level of mean flow stress, the ferrite grain sizes produced by hot/warm- or warm/warm-rolling are an order of magnitude finer than those resulting from conventional (austenite) processing. B. Effect of First Finishing Pass Temperature The influence of first finishing pass temperature on fertite grain size under both laboratory and mill conditions is displayed in Figure 13 for the two processes. Here, the estimates for the mill were obtained using Eqs. [8] and [10]. As expected, lowering the finishing pass temperature leads to finer ferrite grain sizes. [17] It is of interest, however, that as the finishing temperature 2616--VOLUME 23A, SEPTEMBER 1992
is reduced, the difference in grain size between the test and mill conditions decreases until a ferrite grain size of around 1.2 /zm is reached. Thus, 1 /xm seems to be a lower limit in grain size for the present process and steel chemistries.
IV.
CONCLUSIONS
The principal conclusions that can be drawn from the present work are the following: 1. When IF steels are finish-rolled below the Ar~, i.e., in the single-phase ferrite region, rough-rolling can be carried out either in the austenite (hot/warm method) or in the ferrite (warm/warm method) temperature range. Using either processing route, dynamic recrystallization (followed by metadynamic recrystallization) occurs in the fertite during finishing and produces ultraflne ferrite grain sizes of 1 to 3 /zm, depending on the conditions of rolling. 2. The main barrier to the occurrence of dynamic recrystallization in ferrite is the presence of interstitials such as C and N. Removing these elements from the matrix reduces the possibility of strain-induced precipitation and increases the likelihood that dynamic recrystallization can take place. 3. The Ti and/or Nb additions retard softening during the interpass intervals by means of solute drag effects and thus promote dynamic recrystallization by enabling strain accumulation to occur. 4. The recrystallized ferrite grain size decreases with decreasing finishing pass temperature. The minimum ferrite grain size expected under mill conditions for the present chemistries is about 1.1 /xm. Such ultrafine grain sizes lead to considerable increases in both yield strength and toughness. 5. At the same levels of rolling load, processing in the ferrite range can produce grain sizes an order of magnitude finer than can be obtained from conventional (austenite rolling) processing routes. ACKNOWLEDGMENTS One of the authors (AN-Z) expresses his thanks to the 1UT for granting a period of sabbatical leave during which this work was carried out. The authors are indebted to K.R. Barnes of Stelco Steel for supplying the material and information about the rolling schedules and to Professor P.R. Cetlin of the Federal University of Minas Gerais, Brazil, for many stimulating discussions. They acknowledge with gratitude the financial support received from the Canadian Steel Industry Research Association, the Natural Sciences and Engineering Research Council of Canada, and the Quebec Ministry of Education (FCAR program). REFERENCES 1. G. Glover and C.M. Sellars: Metall. Trans., 1972, vol. 3, pp. 2271-80. 2. G. Glover and C.M. Sellars: Metall. Trans., 1973, vol. 4, pp. 765-75. 3. T. Maki, S. Okagushi, and I. Tamura: Proc. 7th Int. Conf. on METALLURGICAL TRANSACTIONS A
4.
5. 6.
7. 8.
9.
10.
the Strength of Metals and Alloys (ICSMA 7), 1985, vol. 1, pp. 923-28. Charles G. Brun and Jose L. Pansera: Metallurgy of VacuumDegassed Steel Products, R. Pradhan, ed., TMS, Indianapolis, IN, 1989, pp. 229-46. H. Yada, Y. Matsumura, and K. Naka: U.S. Patent 4,466,842, Nippon Steel Corp., Japan, 1984. T. Chandra, S. Yue, J.J. Jonas, and R.J. Ackert: Proc. 4th Int. Conf., The Science and Technology of Flat Rolling, Deauville, France, June 1987, vol. 2, p. FI8.1. S.L. Semiatin, G. Lahoti, and J.J. Jonas: ASM Metals Handbook, 9th ed., ASM, Metals Park, OH, 1985, p. 154. A. Najafi-Zadeh, J.W. Bowden, F.H. Samuel, and J.J. Jonas: CIMM, 29th Annual Conf. of Metallurgists, Hamilton, ON, Aug. 26-30, 1990, paper no. 40.8. A. Najafi-Zadeh and J.J. Jonas: Int. Conf. on Processing, Microstructure and Properties of Microalloyed and Other Modern LowAlloy Steels, Pittsburgh, PA, 1991, ISS-AIME, Warrendale, PA, pp. 153-63. S. Yue, F. Boratto, and J.J. Jonas: Proc. Conf. on Hot and ColdRolled Sheet Steels, Cincinatti, OH, 1987, R. Pradhan and
METALLURGICAL TRANSACTIONS A
11. 12.
13. 14. 15.
16. 17.
G. Ludkovsky, eds., TMS-AIME, Warrendale, PA, 1988, pp. 349-59. L.N. Pussegoda, S. Yue, and J.J. Jonas: Metall. Trans. A, 1990, vol. 21A, pp. 153-64. A. Najafi-Zadeh, S. Yue, and J.J. Jonas: Int. Symp. on InterstitialFree Steel Sheet: Processing, Fabrication and Properties, 30th Annual Conf. of Metallurgists, Ottawa, Canada, Aug. 19-20, 1991, L.E. Collins and D.L. Baragar, eds., CIM, Montreal, pp. 93-103. A. Najafi-Zadeh, S. Yue, and J.J. Jonas: Iron Steel Inst. Jpn. Int., 1992, vol. 32, pp. 213-21. J.J. Jonas and T. Sakai: ASM Materials Science Seminar, G. Krauss, ed., ASM, Metals Park, OH, 1984, pp. 185-243. K. Tsunoyama, S. Satoh, Y. Yamazaki, and H. Abe: Metallurgy of Vacuum-Degassed Steel Products, R. Pradhan, ed., TMS, Indianapolis, IN, 1989, pp. 127-41. F.B. Pickering and T. Gladman: Physical Metallurgy and the Design of Steels, Applied Science Publishers, London, 1978, p. 16. A. Najafi-Zadeh, J.J. Jonas, and S. Yue: Methods of Processing Interstitial Free (IF) Steels to Produce Ultrafine Ferrite Grain Sizes, U.S. Patent Application 652 872, Feb. 1991.
VOLUME 23A, SEPTEMBER 1992-- 2617
