ultrafine grained steel by innovative deformation cycle

ULTRA FINE GRAINED STEEL BY INNOVATIVE DEFORMATION CYCLES

ECSC Contract No. 7210.PR/288

SUMMARY

Recently, much attention has been devoted world wide in search of materials with superior strength

and ductility combination. Although many methods are available to improve strength, grain size is

perhaps the only method, as of known today, to improve the conflicting demands in properties as, for

example, strength and toughness.

This project was aimed to establish the best processing routes to promote the formation of ultrafine

ferrite (preferably 2-3 µm) and mixed (e.g. ferrite-cementite, ferrite-martensite) microstructures either

in the bulk or the external layers of strips/plates and rods for a wide range of chemical compositions

(0.05-1.6 %C). The experimentation was mainly focused on processing routes that do not require

extreme strains and significant plant changes, making use of microalloying additions (e.g. niobium)

when necessary, for producing these novel microstructures.

In addition, two severe plastic deformation techniques to produce ultrafine grains (Accumulative Roll

Bonding and Equal Channel Angular Pressing) were tested.

Of course different experiments were performed depending on material carbon content and for this

reason different work packages were created.

For low C steels (0.05-0.1%C), experiments were mainly carried out to assess the hot deformation

conditions for producing fine grain sizes through Strain Induced Dynamic Transformation mechanism

(SIDT) and to establish the influence of chemical composition, prior austenitic grain size, strain and

deformation temperature on grain refinement.

SIDT is a dynamic phase transformation occurring during deformation at temperature slightly higher

than Ar3 due to the strain energy accumulated in austenite phase that induces an early γ−α 伊phase

transformation.

Deformation-dilatometry tests confirmed that increasing the strain the amount of strain-induced

transformed ferrite and grain size refinement increase. The critical strain, required for SIDT to occur

is related to the deformation conditions and chemical composition. The increase of carbon content in

solution retards the SIDT and increases the critical strain. Small prior austenite grain size leads to an

increase of strain induced ferrite volume fraction and ferrite grain refinement.

After the determination of SIDT parameters, laboratory rolling tests were carried out to reproduce as

closely as possible the deformation schedules used in dilatometer experiments in order to get ultra fine

grains on sheets surface.

For microalloyed low C steels the static and dynamic precipitation effect of Nb on grain refinement

was studied.

These experiments were aimed at defining the processing window for hot rolling of Nb steels by

adjusting the steel composition and the process parameters (in particular the time between the last

rolling pass and the on-set of accelerated water cooling in the run-out table) in order to obtain the

finest ferrite grains.

The effects of soluble B and P additions were also investigated.

From the results it was concluded that to obtain fine grains in hot rolled Nb steel, all the Nb must be

dissolved at 1250 °C, strain accumulation must occur below Tnr and above Ar3 to induce dynamic

recrystallisation of deformed austenite grains. To reduce the NbC precipitation during cooling, and

also the austenite static recrystallisation, the last pass must be followed by fast cooling to the

γ−α transformation temperature.

Ultrafine ferrite (sub)grains (0.60 - 0.95µm) and uniformly distributed cementite particles were

produced in two plain C - Mn steels with 0.15% - 0.3% carbon and different Mn contents (0.7% and

1.5%). The experiments were carried out by using heavy warm deformation with a subsequently

3

coiling treatment, a process, which can generally be applied in the industry. The fine partially aligned

globular cementite particles (20 - 350nm) of two different size groups were detected in the final

microstructure. A higher Mn partitioning in the cementite of the Mn richer steel is useful to decrease

the pearlite growth rate. Moreover, apart of the decrease in the average ferrite grain size with increase

in the Mn content, it also leads to a change in the grain shape from an elongated to a more equiaxed

morphology. The fraction of high-angle grain boundaries rises from 64% to 74%. As a consequence of

these changes in microstructure, conditioned by the different Mn content, the ultimate tensile strength

increased from 595MPa to 678MPa, without significant difference in the total elongation of 22.7%

and 21.4%.

Other processing schedules were investigated aimed at producing a homogeneous ultra-fine ferrite and

spheroidised carbide aggregate microstructure structure in high carbon (CMn) steels (0.6-1.6wt%C),

via conventional ‘warm’ rolling and innovative Equal Channel Angular Extrusion (ECAE). Suitable

deformation schedules were determined from dilatometry and thermo-mechanical Gleeble simulations.

Evidence of an ultra-fine ferrite and carbide aggregate microstructure following ‘warm’ rolling was

given. A significant improvement in tensile strength, particularly proof stress (0.2%) was also noted in

comparison to material deformed at higher temperatures. Concurrent Equal Channel Angular

Extrusion (ECAE) experiments investigated microstructural evolution with incremental strain.

Extensive analysis was carried out using various techniques, including high resolution Electron Back

Scattered Diffraction (EBSD). Evidence of ferrite grain refinement was noted in an eutectoid

composition steel. A sub-micron ferrite structure was observed following high strains (ε~3.3) which

induced a marked increase in tensile strength.

Accumulative Roll Bonding experiments were conducted using a 0.15%C-1.37Mn steel strip. The

sheets (3.7 mm thick) were cut into two halves, ground/polished, overlapped (stacked), edge welded

and warm rolled at 615 °C. The process was repeated 5 times for a total strain of about 4 and then

annealed at different temperatures. After annealing at different temperatures final mean grain size

ranged between 0.8µm and 6 µm.

The experimentation showed that this severe plastic deformation method gives the possibility to

obtain customized microstructures with improved combination of strength and ductility/toughness

through adequate grain refinement (1-3 µm).

Results showed that the grain size of ferrite can be refined to 1-3 µm by different methods using

advanced thermomechanical processes that can be applied on already existent pilot mills or that can be

applied off line (Pony Mill). Innovative techniques like Accumulative Roll Bonding and Equal

Channel Angular Extrusion showed to be very effective in grain refinement, because they can easily

impose high deformations giving ultrafine grains but that, at this stage, are still in laboratory scale.

Further, the mechanisms of ultrafine grain refinement were clarified.

As expected, large enhancements in strength were observed in the ultrafine grained steels after tensile

tests at room temperature. However a reduced work hardening in the ultrafine grained steels was

detected.

The presence of a second phase such as globular cementite is beneficial to improve both strength and

work hardening rate of steels. Therefore, it is very attractive to produce a relatively coarser

microstructure (i. e. grain size of 2&3 µm) with a second phase (i. e. finer precipitates) in the matrix.

The ductility of steels is significantly improved due to a relative coarse microstructure and an increase

in work hardening rate by the second phase. On the other hand, in addition to grain refinement,

precipitation hardening can also contribute to strengthening.

4

CONTENTS

SUMMARY 3

CONTENTS 5

LIST OF FIGURES 7

LIST OF TABLES 14

KEYWORDS 14

1. INTRODUCTION 15

3.DESCRIPTION OF THE WORK PACKAGES

4. ULTRAFINE SURFACE GRAINED LOW C STEEL 19

4.1 Introduction 19

4.2 Strain Induced Dynamic Transformation 19

4.2.1 Experimental 19

4.2.2 Results 20

4.2.3 Discussion 32

4.3 Deformed undercooled austenite 34

4.4 Laboratory Rolling 36

4.5 Conclusions 43

5. ULTRAFINE LOW and MEDIUM C STEEL 45

5.1 Introduction 45

5.2 Experimental approach 45

5.2.1 Continuous-cooling-transformation diagrams

after deformation of C-Mn steels 46

5.2.2 Ultrafine grained steels produced by warm deformation 46

5.2.3 Microstructure and mechanical properties of ultrafine grained

steels 47

5.3 Results 49

5.3.1 Optimization of austenite deformation temperature 49

5.3.2 Continuous-cooling-transformation diagrams after deformation 49

5.3.3 Effects on microstructure and hardness of the ultrafine grained

steels after compression tests 51

5.3.4 Microstructure and mechanical properties of the conventional

steels 57

5.3.5 Microstructure and mechanical properties of the ultrafine

grained steels 59

5.3.6 Comparison of properties between the conventional and

ultrafine grained steels 72

5.3.7 Microstructure and texture evolution during warm deformation

and annealing 79

5.4 Discussion 86

5.4.1 Ultrafine grained steels produced by warm deformation 86

5.4.2 Formation of the ultrafine grained steels during warm

deformation and annealing 86

5

2. PROGRAM OBJECTIVES 17

17

5.4.3 Texture evolution during annealing of a 0.2%C steel 89

5.4.4 Effect of carbon and manganese on grain refinement 90

5.4.5 Mechanical properties of the ultrafine grained steels 92

5.4.6 Effect of carbon content on work hardening rate 96

5.5 Conclusions 97

6. ACCUMULATIVE ROLL BONDING 101

6.1 Introduction 101

6.2 Experimental 102

6.3 Results and discussion 102

6.3.1 Microstructure 102

6.3.2 Textures 105

6.3.3 Mechanical properties 107

6.3.4 Bonding 107

6.4 Conclusions 108

7. ULTRAFINE MICROALLOYED LOW C STEELS 109




7.3.1 Dilatometry 111

7.3.2 Torsion tests 117

7.3.3 Hot rolling simulations 121

7.3.4 Hot rolling process parameters affecting the ferrite grain size 130

7.3.5 Further Processing: cold rolling and annealing 133

7.4 Conclusions 137

8. ULTRAFINE HIGH C STEEL 139




8.3.1 Dilatometry 141

8.3.2 Gleeble tests 145

8.3.3 Rolling Tests 148

8.3.4 Mechanical properties 154

8.3.5 Laser Welding 157

8.4 Equal Channel Angular Extrusion 158

8.5 Conclusions 166

9. GENERAL CONCLUSIONS 167

10. RECOMMENDATIONS FOR FUTURE WORK 169

11. REFERENCES 170

6

LIST OF FIGURES

Fig. 1 Thermomechanical treatments carried out by dilatometer: , (a) verification of SIDT occurrence

(b) prior austenite grain size determination.

Fig. 2 Microstructure of CSM1 steel quenched after holding at Ar3 +25°C x 5s.

Fig. 3 Ferrite nucleation by deformation induced transformation at austenite grain boundaries in

CSM1 steel at: (a) 40% deformation; (b) 60% deformation.

Fig. 4 Microstructure of sample air cooled after 40% deformation (CSM1).

Fig. 5 Microstructure of CSM05 steel quenched after holding at Ar3 +25°C.

Fig. 6 Ferrite nucleation in CSM05 steel at (a) 40% deformation; (a) 60% deformation.

Fig. 7 Microstructure of CSM05 steel air cooled after 40% deformation

Fig. 8 Thermal cycles used to get different prior austenitic grain sizes (PAGS).

Fig. 9 Hot deformation thermomechanical cycles.

Fig.10 Microstructures of CSM1 steel after quenching, for different prior austenitic grain size and

reductions.

Fig. 11 Microstructure of CSM05 steel after quenching for different prior austenitic grain size and

reductions.

Fig. 12 Macrostructure of CSM1 steel after quenching deformed at 60%.

Fig. 13 Macrostructure of CSM05 steel after quenching deformed at 60%.

Fig. 14 Simulated equivalent strain distribution for samples deformed at 60%.

Fig. 15 Simulated equivalent strain distribution for samples deformed at 60%.

Fig. 16 Macrostructure of CSM1 steel after quenching deformed at 40% at : (a) Ar3+15 °C, (b) Ar3+35

°C.

Fig. 17 OIM of CSM1 sample deformed at 60% with wide prior austenite grain size. (a) Grains map, (b)

orientation map.

Fig. 18 Orientation maps.

Fig. 19 Thermal cycles used to get different prior austenitic grain sizes (PAGS) for samples 988 and

1022.

Fig. 20 Hot deformation thermomechanical cycles for samples 988 and 1022.

Fig. 21 Microstructure of 988 and 1022 steels after quenching for different deformations. (a)

deformation 40% steel 988, (b) deformation 60% steel 988, (c) deformation 40% steel 1022, (d)

deformation 60% steel 1022.

Fig. 22 Schematic illustration of temperature dependencies of the critical strain for SIDT and DRX [47].

Fig. 23 Free energy increase after deformation of austenite.

Fig. 24 Thermomechanical cycle for undercooled austenite.

Fig. 25 Microstructure of CSM1 steel having PAGS =133 µm, deformed of 60% at (a) 816°C, (b)

700°C.

Fig. 26 Microstructure of CSM05 steel having PAGS =160 µm, deformed of 60% at (a) 837°C, (b)

700°C.

Fig. 27 Macrostructure of CSM1 steel after deformation at 700°C.

Fig. 28 Hot rolling thermomechanical cycle carried out on laboratory mill.

Fig. 29 Microstructures at different through positions thickness of samples hot rolled at pilot Mill (1st

test).

Fig. 30 Microstructures at different through positions thickness of samples CSM1 And CSM05 hot

rolled at pilot mill.

Fig. 31 Hardness through the thickness for plain low-C steels.

Fig. 32 Microstructures at different through positions thickness of samples CSM1 And CSM05 hot


Fig. 33 Hardness through the thickness for microalloyed low-C steels.

Fig. 34 Optimization of austenite deformation temperature (experiment I).

Fig. 35 Determination of continuous-cooling-transformation diagrams after deformation (experiment

II).

Fig. 36 Processing schedules for the compression tests. LSW: large strain warm deformation; Trh:

reheating temperature; Ar3: austenite to ferrite transformation temperature; Pf: pearlite finish

temperature; Bf: bainite finish temperature.

7

Fig. 37 Processing schedules for the plane strain compression tests. Trh: reheating temperature; Ar3:

austenite to ferrite transformation temperature; Pf: pearlite finish temperature (a) conventional

route (CR) for getting coarser microstructure, (b) ultrafine grain route (UFR) for producing

ultrafine microstructure, (c) investigation of microstructure evolution during warm deformation

(ε = 0.4, 0.8, 1.2, 1.6), (d) high temperature annealing after large strain warm (LSW)

deformation

Fig. 38 Water quenched microstructure after deformation at 1133 K for the 15C steel (a), and at 1143 K

for the 2C (b), 2CMn (c) and 3C (d) steel (dashed line shows an example of an equiaxed prior

austenite grain). CD: compression direction; TD: transverse direction.

Fig. 39 Continuous-cooling-transformation diagrams of the experimental steels after deformation (ε

=0.3, ε& =10s-1

) at the optimum temperatures (1133 K for the 15C steel, and 1143 K for the

other steels) and continuous cooling. The cooling rates varied from 2 to 64 K/s. F: ferrite; P:

pearlite; B: bainite; Ms: martensite start temperature. (a) 15C steel; (b) 2C steel; (c) 2CMn steel;

and (d) 3C steel.

Fig. 40 Comparison of the results from Fig.39 for all the steels. The cooling curves are omitted for

clarity. Fs: ferrite transformation temperature; Pf: pearlite finish temperature; F: ferrite; P:

pearlite.

Fig. 41 Microstructure changes with increasing cooling rate in the 3C steel. CD: compression direction;

TD: transverse direction. (a) 2 K/s, (b) 16 K/s and (c) 64 K/s.

Fig. 42 Optical micrographs of the 3C steel after the experiments in Fig. 36. CD: compression direction;

TD: transverse direction; Temp.: multi-step deformation and subsequent annealing

temperatures in the experiment III and experiment V, annealing temperature in the experiment

IV.

Fig. 43 SEM images (a) and EBSD maps (b) of the 3C steel after the large-strain deformation (ε = 1.6)

and subsequent 2 h annealing at different temperatures (experiment V). Details of the

processing are given in Fig. 36c. CD: compression direction; TD: transverse direction. The

black lines indicate grain boundary misorientations between 15° and 63°. The white lines

indicate grain boundary misorientations between 2° and 15°.

Fig. 44 Fraction of high-angle grain boundaries (HAGBs) of the 3C steel obtained for the three

different experiments in Fig. 36.

Fig. 45 Average ferrite grain and subgrain sizes in the CD (compression direction) and TD (transverse

direction), respectively, for the 3C steel. The right hand ordinate indicates the corresponding

values for the grain shape aspect ratio (GSAR), data are taken from the experiment V (Fig. 36c):

(a) only high-angle grain boundaries are considered in the measurement (15° ≤ θ < 63°), (b)

both high- and low-angle grain boundaries are considered in the measurement (2° ≤ θ < 63°)

Fig. 46 Average values for the grain shape aspect ratio (GSAR) of the 3C steel after the three different

processes in Fig. 36. (a) only high-angle grain boundaries are considered in the measurement

(15° ≤ θ < 63°), (b) both high- and low-angle grain boundaries are considered in the

measurement (2° ≤ θ < 63°)

Fig. 47 Effect of carbon content on average ferrite (sub)grain size after the experiments in Fig. 36. (a)

high-angle grain boundaries are considered in the measurements (b) high- and low-angle grain

boundaries are considered in the measurements

Fig. 48 Effect of the experimental routes in Fig. 36 on average ferrite (sub)grain size for the different

steels. (a) high-angle grain boundaries are considered in the measurement, (b) high- and low-

angle grain boundaries are considered in the measurement

Fig. 49 Macrohardness changes with temperature after the different experiments in Fig. 36 (Experiment

III: warm deformation of a ferrite-pearlite microstructure; Experiment IV and V: warm

deformation of a bainite microstructure). (a) 15C; (b) 2C; (c) 2CMn; (d) 3C.

Fig. 50 Optical micrographs of the experimental steels after the conventional route. Details of the

processing are given in Fig. 37a: (a) 15C steel; (b) 2C steel; (c) 2CMn steel; (d) 3C steel.

Fig. 51 Grain size of the experimental steels after the conventional route. Details of the processing are

given in Fig. 37a.

Fig. 52 Comparison of stress-strain curves of the experimental steels with different ferrite grain sizes.

The steels were processed by the conventional route (CR), as shown in Fig. 37a.

Fig. 53 Charpy impact properties for subsize specimens of the experimental steels after the

conventional route (CR). Details of the processing are given in Fig. 37a.

8

Fig. 54 SEM images of the experimental steels after large strain deformation and 2 h annealing at 823

K (Fig. 37b: ultrafine grain route):(a) 15C steel; (b) 2C steel; (c) 2CMn steel; (d) 3C steel.

Fig. 55 Average ferrite (sub)grain size (a) and grain shape aspect ratio (GSAR) (b) of the experimental

steels after large strain deformation and 2 h annealing at 823 K (Fig. 37b: ultrafine grain route).

Fig. 56 Fraction of high-angle grain boundaries (HAGBs) of the experimental steels after large strain

deformation and 2 h annealing at 823 K (Fig. 37b: ultrafine grain route).

Fig. 57 Grain boundary characterization of the 2CMn steel after large strain deformation and 2 h

annealing at 823 K (Fig. 37b: ultrafine grain route). (a) EBSD image: the red lines indicate

grain boundary misorientations between 15° and 63°; the green lines indicate grain boundary

misorientations between 2° and 15° (b) grain boundary misorientation distribution.

Fig. 58 Image quality map of the 2C steel after large strain deformation at 823 K and subsequent 2 h

annealing at 973 K (Fig. 37d).

Fig. 59 EBSD maps of the 2C steel processed by large strain deformation at 823 K and subsequent 2 h

annealing at 973 K (Fig. 37d). (a) image quality map; (b) carbon distribution; (c) manganese

distribution.

Fig. 60 Manganese content (mass%) in the matrix and in the cementite particles determined by the

STEM-EDS technique for the 2C (0.22 mass%C-0.74 mass%Mn) and 2CMn (0.23 mass%C-

1.52 mass%Mn) steels after warm deformation (ε = 1.6) and subsequent annealing.

Fig. 61 TEM micrographs of the 2C (0.22 mass%C-0.74 mass%Mn) and 2CMn (0.23 mass%C-1.52

mass%Mn) steels after large strain warm deformation (ε = 1.6) and annealing, respectively. The

arrows “1” point out the fine cementite particles inside the ferrite grains. The arrows “2” point

out the coarse cementite particles at the ferrite grain boundaries. (a) and (b) deformed

microstructure (c) and (d) annealed microstructure

Fig. 62 TEM micrographs of the dislocation structures in the 2C steel after large strain warm

deformation (ε = 1.6) and annealing, respectively. (a) dislocation structure after large strain

warm deformation (b) dislocation structure after annealing

Fig. 63 TEM micrographs of the 2C steel after large strain warm deformation and annealing. (a) mutual

crystallographic orientation of the local areas (adjacent grains and subgrains numbered with 1

to 6). The white arrows point out the cementite particles pinning the subgrain boundary

(misorientation θ = 7.5°),(b) dislocation structures and individual cementite particles inside the

ferrite matrix. The black arrows point out the cementite particles pinning the dislocations

Fig. 64 Comparison of stress-strain curves of the experimental steels with different ferrite grain sizes.

The steels were processed by the large strain deformation and subsequent 2 h annealing at 823

K (Fig. 37b: ultrafine grain route).

Fig. 65 Mechanical properties of the experimental steels after large strain deformation and subsequent

2 h annealing at 823 K (Fig. 37b: ultrafine grain route). (a) Strength; (b) ductility

Fig. 66 Lüders strain of the experimental steels after large strain deformation and subsequent 2 h

annealing at 823 K (Fig. 37b: ultrafine grain route)

Fig. 67 Yield to tensile ratio (LYS/UTS) of the experimental steels after large strain deformation and

subsequent 2 h annealing at 823 K (Fig. 37b: ultrafine grain route). LYS: lower yield stress;

UTS: ultimate tensile stress.

Fig. 68 True stress-strain curves of the ultrafine grained steels with different carbon contents.

Fig. 69 Work hardening rate θw versus true stress σt, both normalized by the temperature dependent

shear modulus µ = 80 GPa. dα: average ferrite grain size.

Fig. 70 Charpy impact properties for subsize specimens of the experimental steels after large strain

deformation and subsequent 2 h annealing at 823 K (Fig. 37b: ultrafine grain route).

Fig. 71 Specific upper shelf energy (related to the ligament area) (a) and ductile-to-brittle transition

temperature (b) of the experimental steels after large strain deformation and subsequent 2 h

annealing at 823 K (Fig. 37b: ultrafine grain route). All values were converted to those for full

size Charpy test specimens according to equation (1) and (2).

Fig. 72 Observation area of Charpy impact fracture surface – in the middle of the thickness (along ND)

and 1 mm away from the V-notch.

Fig. 73 Fracture surfaces of the Charpy impact specimens tested at room temperature (293 K) and -40°

(233 K). The impact specimens of the experimental steels were produced by the ultrafine grain

route (Fig. 37b). (a) 15C steel; (b) 2C steel; (c) 2CMn steel; (d) 3C steel.

9

Fig. 74 Comparison of average ferrite grain size of the experimental steels produced by the

conventional route (CR) (Fig. 37a), and the ultrafine grain route (UFR) (Fig. 37b), respectively.

Open symbols: results after the CR; filled symbols: results after the UFR.

Fig. 75 Comparison of stress-strain curves of the experimental steels with different grain sizes. The

different grain sizes were produced by the conventional route (CR) (Fig. 37a), and the ultrafine

grain route (UFR) (Fig. 37b), respectively. (a) 15C steel; (b) 2C steel; (c) 2CMn steel; (d) 3C

steel.

Fig. 76 Comparison of tensile properties of the experimental steels with different grain sizes. The

different grain sizes were produced by the conventional route (CR) (Fig. 37a), and the ultrafine

grain route (UFR) (Fig. 37b), respectively. Open symbol: results after the CR; filled symbol:

results after the UFR. (a) yield stress, (b) ultimate tensile stress, (c) ratio of lower yield stress to

ultimate tensile stress (LYS/UTS), (d) Lüders strain.

Fig. 77 Change of Charpy impact properties for subsize specimens of the experimental steels with

different grain sizes. The different grain sizes were produced by the conventional route (CR)

(Fig. 37a), and the ultrafine grain route (UFR) (Fig. 37b), respectively. DBTTsmall: ductile-to-

brittle transition temperature of subsize specimen. (a) 15C steel; (b) 2C steel; (c) 2CMn steel; (d)

3C steel.

Fig. 78 Fracture surfaces, taken by SEM, of the ultrafine and conventional 2C steels after Charpy V-

notch impact tests. A fracture surface of the ultrafine 2C steel after impact testing at 103 K (-

170°) (a) total view of the fracture surface; (b) brittle behavior; (c) border area between

cleavage fracture and ductile fracture; (d) ductile behavior, observation area is shown in (a) B

fracture surface of the conventional 2C steel after impact testing at 143 K (-130°) (e) total view

of the fracture surface; (f) appearance of the cleavage fracture, observation area is shown in (e).

Fig. 79 ND orientation maps (taken by EBSD measurement) of the ultrafine grained 2C steel after

Charpy impact testing at 103 K, the same specimen as shown in Fig. 45a but measured in the

transverse direction of the sample. (a) front of a crack: the circles with number “1” and “2”

point out two elongated grains with high-angle grain boundaries in between (b) aligned damage

below a ductile fracture: the ellipse highlights the alignment of microvoids along the grain

boundaries

Fig. 80 Converted full size Charpy impact properties. (a) Specific upper shelf energy (related to the

ligament area) and (b) ductile-to-brittle transition temperature of the experimental steels with

different grain sizes. The different grain sizes were produced by the conventional route (CR)

(Fig. 37a), and the ultrafine grain route (UFR) (Fig. 37b), respectively.

Fig. 81 Lower yield stress and ductile-to-brittle transition temperature of the experimental steels. The

filled symbol represents the results from the steels after ultrafine grain route (UFR) (Fig. 37b).

The open symbol represents the results from the steels processed by the conventional route (CR)

(Fig. 37a). The dot of number “1” indicates the best combination of strength and toughness.

Fig. 82 Optical microstructures of the 2C steel during warm deformation (Fig. 37c) and annealing (Fig.

37b) at 823 K. (a) initial microstructure before large strain warm deformation, (b)

microstructure after one warm deformation step (ε = 0.4), (c) microstructure after two warm

deformation steps (ε = 0.8), (d) microstructure after three warm deformation steps (ε = 1.2), (e)

microstructure after steps warm deformation four (ε = 1.6), (f) microstructure after four

deformation steps and 2 h annealing at 823 K

Fig. 83 Microstructure evolution of the 2C steel during warm deformation (Fig. 37c) and annealing

(Fig. 37b) at 823 K. Arrow “1” points out the large cementite particle at the ferrite grain

boundary. Arrow “2” points out the small cementite particle in the initial pearlite colony region.

Arrow “3” points out the cementite lamella in the initial pearlite colony region. (a) initial

microstructure before large strain warm deformation, (b) microstructure after two warm

deformation steps (ε = 0.8), (c) microstructure after four warm deformation steps (ε = 1.6),(d)

microstructure after four deformation steps and 2 h annealing at 823 K .

Fig. 84 Evolution of grain characteristic of the 2C steel during warm deformation (Fig. 37c) and

annealing (Fig. 37b) at 823 K, each deformation step with ε = 0.4 and ε& = 10-1

. 1.6-Ann.:

processed by four deformation steps and 2 h annealing at 823 K. (a) change of average ferrite

grain size and grain shape aspect ratio (b) change of fraction of high-angle grain boundaries

(HAGBs)

Fig. 85 Grain boundary misorientation distribution for the 2C steel during warm deformation (Fig. 37c)

and annealing (Fig. 37b) at 823 K, each deformation step with ε = 0.4 and ε& = 10-1

. (a)

10

misorientation distribution after two deformation steps (ε = 0.8), (b) misorientation distribution

after two deformation steps (ε = 1.2), (c) misorientation distribution after four deformation

steps (ε = 1.6), (d) misorientation distribution after four deformation steps and 2 h annealing at

823 K

Fig. 86 Total fraction of α-(<110> || RD) fiber and γ-(<111> || ND) fiber for the 2C steel during warm

deformation (Fig. 37c) and annealing (Fig. 37b) at 823 K, each deformation step with ε = 0.4

and ε& = 10-1

. 1.6-Ann.: processed by four deformation steps and 2 h annealing at 823 K.

Fig. 87 Texture evolution for the 2C steel during warm deformation (Fig. 37c) and annealing (Fig. 37b)

at 823 K, each deformation step with ε = 0.4 and ε& = 10-1

. (a) orientation image: α-(<110> ||

RD) fiber in blue and γ-(<111> || ND) fiber in red (b) observed texture in ϕ2 = 45° ODF section

Fig. 88 Sketch of the microstructure evolution during warm deformation.

Fig. 89 Influence of grain size on the cleavage fracture stress *

fσ and the yield stress σy of notched

impact test specimens. The subscript of number “1” indicates the results from the steel with fine

grains (solid lines). The subscript of number “2” indicates the results from the steel with coarse

grains (dashed lines). After [75]

Fig. 90 Schematic drawing of Accumulative Roll Bonding Process.

Fig. 91 Hot rolling conditions and microstructures for each pass of ARB.

Fig. 92 SEM image of the sample after the 5th pass before annealing.

Fig. 93 Microstructures of ARB samples annealed at 600°C x 3 min at different magnifications.

Fig. 94 Microstructures of ARB samples annealed at (a)650°C x 3 min (b) 700°C x 3 min.

Fig. 95 (a) Unique grain colour image of the sample annealed at 650°C x 3 min, (b) Misorientation

angle distribution.

Fig. 96 Orientation Distribution Function of the sample.

Fig. 97 γ - Fibre of the sample.

Fig. 98 Inverse polar figure maps according the colour code reported: (a) {hkl}, (b) <uvw>

Fig. 99 σ−ε curve of ARB samples.

Fig. 100 Bonding situations in (a) et mid-thickness after 5th pass and annealing (b) defect.

Fig. 101 Dilatometric cycles.

Fig. 102 Thermomechanical (torsion) schedules.

Fig. 103 Length change versus time during the γ−α transformation at 650 °C after different holding

times at 900 °C.

Fig. 104 Results of NbC precipitates quantification in dilatometer samples hold at 900 °C and

quenched.

Fig. 105 Kinetics of γ−α phase transformation at 650 °C in steel 419.

Fig. 106 Microstructures after holding at 900 °C for tp= 6, 200, 5000 and 10000 s (from top to bottom)

and for different initial Nb content: 0.041, 0.053 and 0.083 wt% (from left to right).

Isothermal transformation at 650 °C.

Fig. 107 Determination of n and k parameters of the JMA equation from the curves of Fig.105.

Fig. 108 Transformed fraction to ferrite after different times at 650 °C versus the Nb content in NbC

precipitates formed at 900 °C.

Fig. 109 Magnetometer signal recorded during cooling the steel without Nb from 900 °C at cooling

rates 1, 6, 20 and 60 °C/s, respectively.

Fig. 110 DCCT diagrams for steels 430 (top) and 431 (bottom) with and without holding 100 s at 900

°C. The DCCT diagram for steel 464 without Nb is also indicated.

Fig. 111 NbC quantification results in torsion specimens quenched after different steps.

Fig. 112 Measured ferrite grain diameter for different continuous cooling rates in steel 431 deformed

at 900 °C with and without 100 s holding at 900 °C.

Fig. 113 Microstructures after continuous cooling the steel 430 from 900 °C without (left) and with

(right) 100s delay after deformation at a) CR = 60 °C/s, b) CR = 20 °C/s and c) CR = 6 °C/s.

Fig. 114 Effect of NbC precipitation on the final ferrite grains size and hardness in Nb steels.

Fig. 115 Rolling forces and temperatures measured during the 5 rolling passes.

Fig. 116 Yield Stress (YS) and Tensile Strength (TS) measured after coiling the low C low Mn Nb

steel 806 at 500 °C, 550 °C, 600 °C and 650 °C, respectively.

Fig. 117 Microstructure and measured ferrite grain size in steel 806 coiled at 500 °C, 550 °C, 600 °C

and 650 °C, respectively.

11

Fig. 118 Rolling forces and temperatures measured during the 5 rolling passes (new rolling scheme

with higher reduction in the last passes).

Fig. 119 Microstructure and measured ferrite grain size in steel 1134 coiled at 500 °C, 550 °C and 600

°C, respectively.

Fig. 120 Microstructure and measured ferrite grain size in steel 1135 coiled at 500 °C, 550 °C and 600

°C, respectively.

Fig. 121 Microstructure and measured ferrite grain size in the B-bearing steel 1185 coiled at 500 °C,

550 °C and 600 °C, respectively.

Fig. 122 Microstructure and measured ferrite grain size in steel B-bearing steel 1186 microalloyed

with P, coiled at 500 °C, 550 °C and 600 °C, respectively.

Fig. 123 Rolling forces and temperatures measured during the 3 rolling passes (new rolling scheme

with the highest possible reduction in the last pass).

Fig. 124 Microstructure and measured ferrite grain size in steel 1135 coiled at 500 °C, after being

processed using 2 different rolling schemes.

Fig. 125 Microstructure and measured ferrite grain size in steels 1185 and 1186 hot rolled in 3 passes

with 70 % reduction in the last pass and coiled at 500 °C.

Fig. 126 Ferrite grain size measured after rolling and coiling the different low C low Mn Nb steels at

different CT.

Fig. 127 Effects of B and P additions on the ferrite grain refinement.

Fig. 128 Tensile Strength (TS) measured after rolling and coiling the different low C low Mn Nb steels

at different CT. Effects of Ti additions.

Fig. 129 Tensile Strength (TS) measured after rolling and coiling the different low C low Mn Nb-Ti

steels at different CT. Effects of B and P additions.

Fig. 130 Total elongation (A50) versus Tensile Strength (TS) plot for the 5 studied steel compositions

hot rolled following different rolling schemes.

Fig. 131 Hall-Petch relation from industrially produced steel (Carlam HSM) and laboratory rolled Nb-

steel.

Fig. 132 Tensile property improvement using ultra fast cooling between finishing and coiling

compared to conventional (laminar cooling) process.

Fig. 133 Temperature and rolling forces measured during hot rolling steels 1134 and 1135 with 3

different finishing rolling temperatures.

Fig. 134 Microstructures and measured ferrite grain size in steels 1134 and 1135 hot rolled with 3

different FRT.

Fig. 135 Total Elongation A50 versus TS for hot rolled, hot rolled and annealed, and cold rolled and

annealed specimens.

Fig. 136 Uniform elongation Au versus YS for hot rolled, hot rolled and annealed, and cold rolled and

annealed specimens.

Fig. 137 YS/TS ratio versus Au plot of hot rolled, hot rolled and annealed, and cold rolled and

annealed specimens.

Fig. 138 TS and YS variations after annealing and after cold rolling and annealing hot rolled steels 806,

1134 and 1135.

Fig. 139 Au versus YS plot for steels 1134 and 1135 cold rolled and annealed at 720 °C for different

times.

Fig. 140 A50 versus TS plot for steels 1134 and 1135 cold rolled and annealed at 720 °C for different

times.

Fig. 141 Thermomechanical processing cycle used in the rolling trials

Fig. 142 CCT diagram for 0.6wt.%C steel



Fig. 145 CCT diagrams for 1.2wt%C (a) and 1.6wt%C steels (b)

Fig. 146 Influence of strain on the microstructure of 0.8wt.% carbon steel deformed at 550°C (a) 0

strain, (b) 0.36 strain, (c) 0.72 strain and (d) 1.2 strain.

Fig. 147 Influence of cooling rate, (a) 200°C/s, (b) 1°C/s, on microstructure of 0.8wt% carbon steels

after deformation to a strain of ~1.2 at 650°C.

Fig. 148 Influence of cooling rate, (a) 200°C/s, (b) 10°C/s, (c) 1°C/s on microstructure of 0.8wt%

carbon steels after deformation to a strain of ~1.2 at 500°C.

12

Fig. 149 (a-c) Comparison of single and multiple deformation for 0.6wt.% carbon steel deformed at

650°C

Fig. 150 Through thickness hardness profile for 0.8wt.%C steel following rolling trials.

Fig. 151 (a-d) SEM micrographs illustrating the carbide structure present in the 0.6 wt.% C steel

following the rolling trials. Deformation temperature and finishing strain are given

Fig. 152 (a-d) SEM micrographs illustrating the carbide structure present in the 0.8 wt.% C steel


Fig. 153 (a-c) SEM micrographs illustrating the carbide structure present in the 1.2 wt.% C steel


Fig. 154 (a-b) SEM micrographs illustrating the grain structure present in the 0.6 wt. % C steel

following deformation at 600°C, to strains of ε ~ 1.2 and 2




following deformation at 600°C, to strains of ε ~ 1.2.

Fig. 157 EBSD band contrast (a) and orientation (b) maps of 0.6wt.%C steel following rolling at 600°C,

ε~2.

Fig. 158 Tensile properties for 0.6wt.%C steel following rolling trials



Fig. 161 Weld side of steel plate following laser welding

Fig. 162 Penetration side of steel plate following laser welding

Fig. 163 (a and b). Image illustrating the ECAE apparatus (a) and half of the 120° channel die (b)

Fig. 164 Schematic representation of the ECAE process, showing a cross section of a deformed billet,

the ram, the die set and the shear plane of the apparatus [131]

Fig. 165 SEM micrographs illustrating the microstructural evolution during ECAE processing (1 (i), 2

(ii), 3 (iii), 4 (iv) and 5 (v) passes) of 0.6, 0.8 and 1.2 wt. % C steels.

Fig. 166 (a+b). SEM micrographs etched to highlight the subgrain/grain boundaries of 0.8 wt. % C and

1.2 wt. % C (c+d) steel samples following 4 ECAE passes

Fig. 167 Sample B1, 0.8wt%C steel after 1 ECAE pass. Suspect points have been removed and the

data cleaned to remove non-indexed points. LAGB (grey) and HAGBs (black) also shown.

Fig. 168 Sample B3, 0.8wt%C steel after 3 ECAE passes. Suspect points have been removed and the

data cleaned to remove non-indexed points. LAGB (grey) and HAGBs (black) also shown

Fig. 169 Sample B5, 0.8wt%C steel after 5 ECAE passes. Suspect points have been removed and the


Fig. 170 Plot of EBSD statistical data showing the average transverse high angle grain boundary

spacing and high angle grain boundary area fraction (%HAGBs) in the 0.8wt.%C steel, as a

function of strain during ECAE processing.

Fig. 171 Sample A5, 0.6wt%C steel after 1 ECAE pass. Suspect points have been removed and the


Fig. 172 Sample C5, 1.2wt%C steel after 5 ECAE passes. Suspect points have been removed and the


Fig. 173 Tensile properties of 0.6wt.%C steel with increase in ECAE deformation



13

LIST OF TABLES

Table 1 Chemical compositions and critical temperatures measured by dilatometric tests for materials

cast at CSM.

Table 2 Chemical Composition of microalloyed steels received by CRM.

Table 3 Temperatures measured during pilot mill rolling.

Table 4 Hot rolling conditions during the 2nd

test.

Table 5 Grain size measurements at different positions through the thickness.

Table 6 Grain size measurements at different positions through the thickness.

Table 7 Chemical compositions (mass%) of the steels, calculated Ae3 (K) [26]

Table 8 Chemical composition of the material used for ARB tests.

Table 9 Steel compositions (10-3

wt%).

Table 10 Compositions (wt. %) of high carbon steels

Table 11 Hardness and microstructure of dilatometer samples

Table 12 EBSD statistical data for 0.6wt%C steel rolled at 600°C to a strain of ~2

Table 13 Tensile properties following rolling trials with finishing rolling temperatures of 600°C and

700°C and strains of ~1.2 and 2.

Table 14 Comparison of the average statistical boundary data for the 0.6, 0.8 and 1.2wt.%C steel

obtained from EBSD maps at different levels of ECAE deformation.

KEYWORDS

Ultrafine grain, Grain refinement, Severe Plastic Deformation, Warm Rolling, SIDT, DIFT,

Accumulative Roll Bonding, Equal Channel Angular Extrusion

14

1. INTRODUCTION

Steel is the most important structural material with a variety of applications across a wide range of

industries. Moreover, steel will continue to be the main structural material in the 21st century by

fulfilling requirements of: (1) higher strength and safety for high buildings, long bridges and high-speed

traffic, (2) longer and maintenance-free life and (3) environment friendly and recyclable. Social

demands such as energy saving, resource saving and environment conservation are increasing, therefore,

it is desirable to develop materials with high strength, long life and which are easy to recycle.

One solution is to produce and use ultrafine grained steel. Among different strengthening mechanisms,

grain refinement is the only method to improve both strength and toughness simultaneously. Therefore,

ultrafine grained steels with relatively simple chemical compositions, strengthened primarily by grain

refinement, show great potential for replacing alloyed high strength steels, thereby, avoiding additional

alloying. Meanwhile, complicated additional heat treatments like soft annealing, quenching and

tempering are not required, and the weldability of the steel is also improved due to a low carbon and

alloying element content compared to high strength quenched and tempered steels. Ultrafine grained

steels also offer other property advantages, such as high strain rate superplasticity at relatively high

temperatures.

For the past 50 years, the drive to produce finer grained steel has continued to obtain better mechanical

properties. So far, grain refining to the level of 5&10 µm has been realized in industrial production by

TMCP (Thermo-Mechanical Controlled Process). The further challenge of ultra grain refinement in

steels to 1 µm or below has been confronted in the recent several years.

Currently, there are two main groups of laboratory methods to refine ferrite grains down to the ultrafine

grain range in bulk steels. These are severe plastic deformation techniques and the advanced

thermomechanical process. Until recently the developed and applied severe plastic deformation

techniques were equal channel angular pressing (ECAP) [1,2], accumulative roll bonding (ARB) [3,4]

and bi-directional deformation [5]. The advanced thermomechanical processes include strain-induced

ferrite transformation (i.e. transformation during rather than after deformation) [6], dynamic

recrystallization of austenite during hot deformation with subsequent γsα (austenite to ferrite)

transformation [7], hot rolling in the intercritical region (i.e. in the austenite + ferrite two-phase region)

[8], and dynamic recrystallization of ferrite during warm rolling [9].

It appears that both severe plastic deformation and the advanced thermomechanical processing are

capable of producing ultrafine ferrite grain sizes, and that an ultrafine grained microstructure may, if

used wisely, be beneficial. However, the deformation and cooling conditions in these routes are often

extreme and outside the capabilities of large scale industrial processing lines.

Giving severe plastic deformation for instance, to produce an ultrafine grained steel with grain size of

0.3 µm, the essential strain is about 4.0 for equal channel angular pressing, 5.6 for accumulative roll

bonding process and 6.4 for bi-directional compression, respectively. This is a high requirement for

deformation tools, force and power. On an actual mass production scale, however, there are limitations

to the application of low temperature large scale deformation. Therefore, it is necessary to develop a

technology to achieve commercially meaningful grain refinement using existing rolling facilities.

The objective of this project is to produce ultrafine grains using processing routes that do not require

extreme strains and significant plant changes, making use of microalloying additions (e.g. niobium)

when necessary, for producing these novel microstructures.

15

2.PROGRAM OBJECTIVES

The objectives of this project are:

1. To establish the best processing routes to promote the formation of UF (preferably 2-3 µm) and

mixed (e.g. ferrite-cementite, ferrite-martensite) microstructures either in the bulk or the

external layers of strips/plates or rods of low and medium carbon steels.

2. To asses the benefits that can be achieved in terms of strength, ductility and toughness by the

formation of UF microstructures

3. To produce very fine bulky microstructure using two innovative deformation methods:

o Accumulative Roll Bonding to be applied for low-C steels

o Equal Channel Angular Pressing to be applied to high C steels

3. DESCRIPTION OF THE WORK PACKAGES

The program was addressed to assess the advantages of ultrafine microstructures for the following types

of steels:

• Low carbon: 0.04-0.16C; 0.15-0.3Si; 0.5-1.6Mn

• Medium carbon: 0.25-0.45C; 0.15-0.3Si; 0.8-1.3Mn.

• High carbon: from 0.8% up to 1.6 %.

The benefits of micro-additions of Nb and/or other elements (e.g. P) were also examined.

The project was focused on two different types of experimentation:

1. to develop advanced thermomechanical processes for producing ultrafine microstructures, that

do not require extreme strains and significant plant changes and that can be applied on existing

hot rolling mills.

2. To test two innovative severe plastic deformation methods to verify their potential in refining

the grains: Accumulative Roll Bonding and Equal Channel Angular Extrusion.

To carry out this experimentation, the following Work Packages were identified:

WP1. UF Surface Grained Steel

Steel grades with a carbon content ranging between 0.05 – 0.1 % C were studied in order to assess the

parameters for Strain Induced Dynamic Transformation occurrence and to study the feasibility of

ultrafine grains at the surface of steel products.

In particular two approaches, either individually or combined, were investigated:

a) intragranular nucleation in the deformed austenite through minimisation of grain boundary

nucleation during the strain induced transformation of austenite after hot/warm rolling by

single/multiple pass;

b) local warm rolling due to supercooling of austenite induced by intermediate water cooling and

roll chilling; combined with localisation of shear strain in a layer close to the surface;

In addition to the influence of C and Mn, the presence of microalloying elements, such as Nb, has

expected to modify the microstructure.

WP2. UF Low C Steel

Steel grades with a carbon content between 0.04 – 0.1 % C containing Mn (0.8 – 1.5 %) and Nb (0.04 –

0.1 %) were laboratory cast and hot rolled on the CRM pilot rolling mill. The rolling schedule together

with the Nb-content were adapted in order to reach an optimal amount of Nb in solution which retards

the ferrite transformation. A high coiling temperature (580 – 650°C) was applied for obtaining a fully

ferritic microstructure, ensuring additional precipitation hardening.

In a second step, the effect of Si (1.5 %) and P (0.1 %) was studied. Torsion and compression tests were

carried out to study the effect of temperature and reduction on the grain refinement.

17

MPI will concentrated their experiments on 0.15 C steel, exploring conditions to develop a uniform

distribution of second phases such as cementite and/or very small islands of martensite, in addition to 2-

3 µm ferrite grains.

Accumulative Roll Bonding (ARB) process was applied on 0.15 C steel by CSM.

WP3. UF Medium C Steel

The following aspects were investigated on medium C steels: optimum range of ferrite/pearlite fraction,

effect of cooling rate over the transformation region, necessary amount of deformation just below the

pearlite finish temperature PF to achieve the spheroidisation of the lamellar cementite (pearlite).

The main contribution of MPI was the simulation of the processing routes by the plane strain

compression test by WUMSI.

Laboratory rolling trials on the basis of the results were carried out.

Based on the PonyMILL concept, the fabrication of ultrafine grained steels is introduced in the present

work. In particular attention was paid to the possibility of its industrial application. For plain C-Mn

steels a rather moderate logarithmic strain of ε = 0.8 & 1.6 and cooling rates which are quite common in

an industrial production were used. At the same time, a simulated final coiling was employed to imitate

a complete industrial processing.

WP4. UF High C Steel

A range of carbon steels (with Nb to aid strain accumulation) were processed to ferrite/carbide

aggregate microstructures (i.e. non-pearlitic). A series of laboratory casts was prepared, of 0.4%, 0.8%,

1.2% and 1.6% C. A range of deformation schedules were applied by Gleeble thermomechanical

simulator and/or deformation dilatometry, from which microstructure and hardness were determined. A

selection of conditions were made on this basis for processing of larger volumes of material suitable for

obtaining tensile test data. Example material was also processed to very high strains by ECAE

processing at the Manchester Materials Science Centre, from which tensile test data were obtained.

The program was a strict co-operation between CSM (coordinator), CRM, Corus and MPI.

18

4.ULTRAFINE SURFACE GRAINED LOW C STEEL

4.1 Introduction

The possibility of making microstructures in which bands of Ultra Fine Ferrite (UFF) grains act as

strengthening component to more normal microstructural constituents was investigated in thin strips

characterised by very large austenite grains by Hodgson et al. [10,11].

Their method of producing a surface layer of UFF in thin strip required austenitisation at a high

temperature to form a coarse austenite grain. The strip was then cooled and hot rolled at a temperature

close to the Ar3 value (700-800 °C), using a modest reduction (30-40%). The supercooling of austenite

owing to roll chilling, combined with the strong shear strain at the strip surface, lead to a very high

nucleation density of ferrite on dislocations within the coarse austenite grains. It seemed that the UFF

grains resulted from strain-induced transformation of intragranular ferrite with nucleation occurring on

the austenite substructure. The reduction or minimisation of grain boundary nucleation by austenite

grain enlargement seemed to promote a substantially instantaneous transformation homogeneously over

the austenite. This process is called Strain Induced Dynamic Transformation (SIDT).

The layered microstructure consisting of ultra fine grained surface layers (generally penetrating to one

quarter of the strip thickness) in the plates were found to be beneficial for improving the YS/ UTS ratio.

In fact, Hodgson [12] used a modified version called the post roll quenching, specifically to produce a

steel with layered microstructures, with the surface having ultra fine grains less than 2 µm and the

centre showing martensite or bainite microstructure. They argued that the quenching produced enough

under-cooling of austenite at the surface by increasing the driving force for nucleation of ferrite.

However, other authors [13-16] showed completely different results showing difficulties to generate

this kind of UFF grains at surface using the methods of Hodgson, but they gave different indications

like high rolling strains, fast cooling and fine austenite grain size to create this type of structure.

It is clear that the key of the process is the understanding of SIDT mechanism, that is still unclear and

presently studied worldwide. Therefore main part of this work package was concerned to the study and

characterization of SIDT mechanism.

Finally, pilot mill rolling tests were carried out to obtain ultrafine surface grains.

4.2 Strain Induced Dynamic Transformation

Driving force for phase transformation is dependent on chemical composition and temperature change.

The typical phase transformations in steels are austenite to ferrite formation, austenite to pearlite,

austenite to bainite, austenite to martensite and precipitation of nitrides and carbides. If an external load

is applied to the system above the static transformation temperature, the free energy of the system will

be certainly raised. If the load (e.g. deformation) is applied to the austenite above transformation

temperature and the free energy of austenite is raised to the extent over the free energy of ferrite at the

same temperature, the transformation from austenite to ferrite will occur, which could not be found to

happen without deformation.

This kind of ferrite transformation induced by deformation has been called as Deformation Induced

Ferrite Transformation (DIFT) or Strain Induced Dynamic Transformation (SIDT).

The aim of this experimentation was to assess the hot deformation conditions in the austenite region for

producing fine grain sizes through SIDT mechanism, and to establish the influence of prior austenitic

grain size, strain, deformation temperature and chemical composition on grain refinement.

4.2.1 Experimental

Dilatometric experiments were carried out on samples cast at CSM and machined from laboratory hot

rolled plates, Moreover, other dilatometric tests were carried out in order to compare the refinement

ability of SIDT and deformation of undercooled austenite and to characterize the SIDT mechanism in

Nb microalloyed steels.

19

The chemical compositions of the steels used for these tests and the measured critical temperatures are

given in 1.

C Si Mn P S Al N Ar1 Ar3

(%) (%) (%) (%) (%) (%) (%) °C °C

CSM05 0.047 0.2 1.44 0.007 0.0023 0.0259 0.0028 625 823

CSM1 0.107 0.2 1.45 0.007 0.0021 0.0258 0.0032 610 790

Table 1- Chemical compositions and critical temperatures measured by dilatometric tests for materials

cast at CSM.

The hot deformation experiments were performed by a Theta Ind. Dilatronic III dilatometer. Samples

had 3.25 mm diameter and 6 mm length. The experiments were carried out in vacuum (2x10-5

mbar)

and quenching was done using He. The cooling rate was 130°C/s.

Before deformation, the samples were homogenised at Ar3+50°C for 10 min, then cooled at a rate of

2°C/s to the deformation temperature (Ar3+25°C). Samples were deformed at 40% and 60%. After

deformation, the samples were either helium quenched or cooled at the rate of 10°C/s.

Samples for metallographic examination were prepared from the deformed specimens by polishing the

longitudinal section down to centre of the sample.

4.2.2 Results

Verification of Deformation Induced Ferrite Transformation occurrence

In order to verify the occurrence of SIDT and to study the occurrence of ferrite formation (before

deformation, during deformation or in the later stages of cooling), the samples after the homogenization

Ar3+50°C for 10 min, were held at the respective Ar3+25°C without deformation for 5 s and then

quenched (Fig. 1a). They were then compared with samples transformed according the

thermomechanical cycle shown in Fig. 1b that were deformed at 40% and 60%.

Microstructure of CSM1 steel (Fig. 2), treated according the cycle shown in Fig. 1a, showed fully

martensitic structure and absence of ferrite nucleation.

T= Ar3+50°C

10 min

WQ

T= Ar3+25°C

Def=

40% or 60%

T= Ar3+50°C

10 min

WQ

T= Ar3+25°C

Def=

40% or 60%

(a) (b)

Fig. 1- Thermomechanical treatments carried out by dilatometer: (a) verification of SIDT occurrence (b)

prior austenite grain size determination.

t = 5s

T= Ar3+50°C

10 min

WQ

T= Ar3+25°C

T= Ar3+50°C

10 min

WQ

T=

20

The microstructure of CSM1 sample deformed to 40% at Ar3+25°C and then quenched is shown in Fig

3a. Ferrite that nucleated at the austenite grain boundaries can be clearly seen. Apart from the ferrite at

the austenite boundaries, the structure consisted of Widmanstatten ferrite (WF), allotromorphic ferrite

and a dark constituent interpreted as bainite. WF was present at the grain boundaries as well as at the

deformation bands. Volume fraction of deformation induced transformed ferrite was less than 30% at

the centre of the sample.

When the deformation was increased to 60% the microstructure consisted of predominantly fine ferritic

grains and some amount of bainite and martensite. The average grain size in 60% deformed samples

was 3 µm (Fig. 3b) and the volume fraction of strain induced ferrite was about 80%.

Samples cooled at 10°C/s after deformation showed typical transformation structure consisting of

homogeneous largely equiaxed polygonal ferrite (Fig. 4). The approximate grain size of these ferrite

grains was 7µm.

Fig. 2- Microstructure of CSM1 steel quenched after holding at Ar3 +25°C x 5s.

(a) (b)

Fig. 3– Ferrite nucleation by deformation induced transformation at austenite grain boundaries in

CSM1 steel at: (a) 40% deformation; (b) 60% deformation.

21

Fig. 4– Microstructure of sample air cooled after 40% deformation (CSM1).

In Fig. 5 the microstructure of CSM05 steel quenched after holding at Ar3 +25°C for 5s without

deformation is shown. It exhibited bainitic structure with some amount of pro-eutectoid ferrite

probably formed during cooling. However, it was difficult to distinguish the bainitic ferrite and

polygonal ferrite by optical metallography.

The microstructure of samples 40% deformed, showed predominantly ferritic structure with some

amount of bainite (Fig. 6a). No WF or degenerate ferrite was observed in these samples.

Fig. 5- Microstructure of CSM05 steel quenched after holding at Ar3 +25°C.

An inhomogeneous grain size distribution was observed and the mean value was around 4-5 µm. The

volume fraction of strain induced transformed ferrite was about 70%.

At 60% deformation the microstructure was again ferrite and some amount of bainite. The ferrite

volume fraction was more than 90%. The average grain size in these samples was in the range of 4-5

µm. (Fig. 6b).

In Fig. 7the microstructure of CSM05 air cooled after 40% deformation is shown. Average grain size is

9 µm.

Observations confirmed that the sample CSM1 was deformed in the austenite phase and that the

deformation caused the formation of ferrite. In sample CSM05 in addition to strain induced ferrite, there

was some pro-eutectoid ferrite observed in the samples quenched from temperatures above Ar3+25°C

without deformation.

This situation was confirmed by the two sets of grain sizes present in the deformed samples. One in the

range 1-3 µm and the other in the range 5-6 µm. The microstructures revealed that, in general,

increasing the deformation the amount of strain-induced transformed ferrite increase, while no clear

trend can be established at this stage about grain refinement.

In CSM1 steel, after 40 % reduction, the SIDT ferrite was formed mainly at the austenite boundaries,

whereas in CSM05 steel, the strain induced ferrite nucleated both at grain boundaries and deformation

bands.

22

(a) (b)

Fig. 6- Ferrite nucleation in CSM05 steel at (a) 40% deformation; (b) 60% deformation.

Fig. 7- Microstructure of CSM05 steel air cooled after 40% deformation

In addition to the strain induced ferrite, Widmanstatten ferrite was present in some regions in CSM1

steel. Because of increased hardenability and lower Ms temperature of this steel, WF formed at low

undercoolings close to the Ae3 temperature. On the other hand, in CSM05 steel, because of lower

carbon content and finer austenite grain size, the hardenability was lower compared to CSM1 steel.

Hence most of the austenite was transformed to ferrite during deformation and cooling and the

remaining austenite enriched in carbon transformed to bainite.

Determination of SIDT Parameters

After the preliminary results described above, the study of the effect of different parameters (prior

austenite grain size, strain, deformation temperature, carbon content) on SIDT mechanism were carried

out.

At first, various tests were performed in order to determine the thermal cycles to get very different prior

austenitic grain sizes for both materials, having a homogeneous grain size distribution as shown in Fig.

8.

In order to get different prior austenite grain sizes before deformation the samples were treated

according to the cycles showed in Fig. 8, then cooled at a rate of 2°C/s to the deformation temperature

of about Ar3+25°C and deformed giving 40% and 60% reductions (Fig. 9). After deformation, the

samples were helium-quenched.

Samples for metallographic examination were prepared from the deformed specimens by polishing the

longitudinal section and observed at the centre of the sample.

In Fig. 10 the microstructures of CSM1 sample are shown.

23

In Fig. 10a, the microstructure of the sample having a small prior austenitic grain size (13µm), 40%

deformed at 816 °C is reported. Ferrite that nucleated at the austenite grain boundaries can be clearly

seen.

When the reduction was increased to 60% (Fig. 10c) the microstructure consisted of predominantly fine

ferritic grains and small amount of bainite and martensite. The average ferrite grain size (FGS) was 3

µm.

In the case of larger prior austenitic grain size (133 µm) and 40% reduction (Fig. 10b), no polygonal

ferrite was detected. If the reduction is increased to 60% the microstructure consisted of some fine

ferritic grains at the austenite boundaries, Widmanstatten ferrite (WF) and bainite. The average ferritic

grain size in this case was 3.6 µm (Fig. 10d).

Fig. 8– Thermal cycles used to get different prior austenitic grain sizes (PAGS).

Fig. 9 – Thermomechanical cycles.

The very fine polygonal ferrite grains observed in quenched specimens after deformation were formed

during deformation. In fact, results from preliminary experimentation showed that the ferrite could not

be formed without deformation in the later stages of cooling. The presence of ferrite in the samples

showed in Fig. 10 means that the γ−α transformation occurred dynamically during deformation.

However, CSM05 steel showed some pro-eutectoid ferrite in the quenched samples due to its lower

hardenability.

In Fig. 11 the microstructures of CSM05 steel deformed at 837°C are shown.

HQ HR

20°C

t = 10min

T = 900¡

HR

20°CHQ

t = 5 min

T = 900¡

HR

20°CHQ

t = 10 min

T = 1150¡C

HQHR

20°C

t = 30 min

T =1150¡

160 µm

133 µm

14.3 µm

13.0 µm

0.05%C 0.1%C

Heating rate: 20°C/s

Helium Quenching

t = 10 min, 30 min

T = 900, 1150 °C

T = Ar3 + 25 °C

Red. = 40% , 60%

cooling rate: 2°C/s

24

The microstructure of sample strained to 40% reduction with small prior austenite grain size (14µm)

showed predominantly polygonal ferrite with a small amount of bainite. No WF or degenerate ferrite

was observed. In this sample, in addition to strain induced ferrite, there was some pro-eutectoid ferrite.

This was evident from the two sets of grain sizes present in the deformed samples. The average grain

size was 4.2 µm.

At 60 % deformation the microstructure was fully ferritic and the average grain size was 3.2 µm. In this

case higher SIDT ferrite volume fraction was detected.

In the case of the sample with coarse prior austenite grain size (160 µm) deformed 40% the

microstructure consisted of some fine ferritic grains at the austenite boundaries, Widmanstatten ferrite

(WF) and bainite. At higher reduction (60%) most of the structure consisted of polygonal ferrite but

some bainite and WF ferrite were still present. The average ferritic grain size in this case was 4 µm.

PAGS = 13 µm PAGS= 133 µm

40%

60%

FGS=3.6 µmFGS=3 µm

Fig. 10-Microstructures of CSM1 steel after quenching, for different prior austenitic grain size and

reductions.

25

PAGS = 14 µm PAGS = 160 µm

40%

60%

FGS=4.2 µm

FGS=3.2 µm FGS=4 µm

FGS=4.9 µm

Fig. 11– Microstructure of CSM05 steel after quenching for different prior austenitic grain size and

reductions.

It is evident that a critical strain for SIDT occurrence exists. It depends on carbon content and prior

austenite grain size.

In Fig. 12 and in Fig. 13 the macrostructure of samples after the thermomechanical cycles showed in

Fig. 9, with the highest prior austenite grain size, are reported. It is evident that the microstructure

inside the sample is quite inhomogeneous. For both samples, nearby the surface no ferrite is present and

it is true until a certain depth after that ferrite is present. It means that a critical strain for SIDT

occurrence exists.

Fig. 12– Macrostructure of CSM1 steel after 60% reduction and quenching.

26

Fig. 13– Macrostructure of CSM05 steel after 60% reduction and quenching.

Fig. 14– Simulated equivalent strain distribution for samples after 60% reduction.

In order to understand the strain distribution inside the sample and the correspondent microstructural

behaviour, a FEM simulation was carried out using MARC code. Simulated equivalent strain

distribution inside the sample is shown in Fig. 14 and quantitative distribution in the central line of the

sample is reported in Fig. 15.

Comparing the model results and the microstructure, it is clear that the microstructure follows the same

trend of the strain. In particular, measuring the deepness at which SIDT begins to occur, results say that

the critical strain for sample CSM1 is about ε=1.5, while for CSM05 it is about ε=1. Applying the same

method to samples having a smaller prior austenite grain size, critical strain resulted to be ε=0.8 for

CSM1 steel and ε=0.6 for CSM05 steel. It means that decreasing the carbon content and the prior

austenite grain size, the critical strain decreases.

27

0

0.5

1

1.5

2

2.5

0 0.5 1 1.5 2 2.5

Sample thickness (mm)

Eq

uiv

ale

nt

str

ain

Fig. 15– Simulated equivalent strain distribution for samples after 60% reduction (average strain ≈0.8).

(a) (b)

Fig. 16- Macrostructure of CSM1 steel after quenching deformed at 40% at : (a) Ar3+15 °C, (b) Ar3+35

°C.

Further tests regarding the effect of deformation temperature on the occurrence of SIDT were carried

out.

Two tests were performed giving 40% reduction at two different deformation temperatures: the first

specimen (Fig. 16a) was deformed at Ar3+15 °C and the other one at Ar3+35 °C (Fig. 16b) in order to

verify some differences in grain size and ferrite volume fraction.

Results showt that grain size decreases with the decrease of deformation temperature, and ferrite

volume fraction increases. It means that the critical strain for SIDT is decreased with the reduction of

temperature.

Textures in SIDT

In order to verify the textural properties of strain induced dynamic transformed ferrite, orientation

measurements on sample CSM1 after the thermomechanical cycle whose microstructure is shown in Fig.

10d were carried out.

28

(a) (b)

Fig.17– OIM of CSM1 sample deformed at 60% with wide prior austenite grain size. (a) Grains map,

(b) orientation map.

In Fig. 17a the grain map reconstructed by Orientation Imaging Microscopy is shown. The prior

austenite grain size that after quenching was transformed in martensite-bainite is dark and the

recrystallized ferritic grains are coloured. It can be noted that most of the ferritic grains were generated

on prior austenite grain boundaries. In Fig. 17b the orientations of the single grains are mapped. The

grains around the prior austenite grain boundaries, generated by SIDT show a homogeneous orientation

distribution. Most of the grains show typical recrystallization texture. It means that these grains are fully

recrystallized and not recovered and they were formed during deformation and not after, otherwise they

had to show deformation texture. Further they show high boundary angles.

29

Fig. 18– Orientation maps.

To confirm these results in Fig. 18 orientation map for planes //ND and the orientation map for

directions //RD, are shown. Both from Fig. 17 and Fig. 18 it is clear that the texture is made mainly by γ

fibres with a strong component at {111}<011>, the typical recrystallization textures of bcc crystals.

SIDT in microalloyed steels

Dilatometric experiments were carried out on samples received by CRM and machined from laboratory

hot rolled plates in order to asses the influence of Nb content on SIDT mechanism occurrence. The

chemical compositions, reported in Table 2, show two different carbon contents, comparable to the

previous plain C-Mn samples for determination of SIDT parameters.

In order to compare the microstructures with the C-Mn steel studied before, preliminary tests to have

the same prior austenite grain size were carried out. The final conditions to have different prior

austenite grain sizes are reported in Fig. 19. However, it was very difficult to find the conditions to have

a very large prior austenite grain therefore the following thermomechanical cycles were applied only on

samples with a fine prior austenite grain size.

Then thermomechanical cycles similar to the previous ones were carried out (Fig. 20).

Steel Code C Si Mn P Nb Al Ti AR3

988 0.121 0.02 1.61 0.012 0.03 0.055 0.007 755

1022 0.064 0.01 1.63 0.011 0.033 0.042 0.001 784

Table 2 – Chemical Composition of microalloyed steels received by CRM.

30

In Fig. 21 the microstructures of samples 988 and 1022, austenitized, deformed under different

conditions and quenched, are shown. Initial prior austenite grain sizes were 15 and 19 µm respectively.

In Fig. 21a, the microstructure of sample 988, 40% deformed at 780 °C is reported. Some ferrite that

nucleated at the austenite grain boundaries can be seen but its volume fraction is less than that of

sample CSM1 (Fig. 10a) with similar carbon content and similar prior austenite grain size.

When the deformation was increased to 60% reduction (Fig. 21b) the volume fraction of deformation

induced transformed ferrite increased but it remained less than 30%.

Fig. 19– Thermal cycles used to get different prior austenitic grain sizes (PAGS) for samples 988 and

1022.

Fig. 20– Hot deformation thermomechanical cycles for samples 988 and 1022.

Heating rate: 20°C/s

Helium Quenching

t = 10 min

T = 1020 °C

T = Ar3+25 °C

Red. = 40% , 60%

cooling rate: 2°C/s

HQ HR

20°C

t = 10min

T = 1020¡

HR

20°CHQ

t = 10 min

T = 1020¡

HR

20°CHQ

t = 10 min

T = 1150¡C

HQHR

20°C

t = 30 min

T =1150¡

106 µm

75 µm

15 µm

19.0 µm

0.06%C 0.12%C

31

40% 60%

0.12

%C

(a) (b)

0.06

%C

(c) (d)

Fig. 21– Microstructure of steels 988 and 1022 after quenching for different deformations: (a) steel 988

40% reduction, (b) steel 988 60% reduction, (c) steel 1022 40% reduction, (d) steel 1022 60%

reduction.

In Fig. 21c the microstructure of the sample 1022, 40% deformed at Ar3 +25°C (T=809°C) is reported.

In this case the SIDT volume fraction is higher (Fig. 21a) as expected from the influence of carbon

content. However if compared to Fig. 11a, the SIDT ferrite is quite less (about 50%). The same trend

can be detected in Fig. 21d, where the sample was subjected to 60% reduction.

4.2.3 Discussion

Before the discussion of the results it is necessary to point out that during the austenite deformation,

SIDT and dynamic recrystallization (DRX) compete with each other and one of them, the one that has

the smaller critical strain, will take place ahead. When the deformation temperature is below Ae3 the

critical strain for SIDT is smaller than for DRX and SIDT becomes the predominant softening

mechanism (Fig. 22).

Influencing factors in SIDT mechanism

From results it is evident that the processing parameters and initial microstructure can influence the

occurrence of SIDT, the final grain size of ferrite and the SIDT ferrite volume fraction. They will be

discussed hereinafter:

(a) – Critical strain

As explained before, deformation accelerates the phase transformation kinetics (Fig. 23). However if

the imposed strain is too small the dynamic transformation is not activated.

32

Ae3Ar3

SIDT of DRX of

Cri

tic

al

Str

ain

Temperature

Fig. 22– Schematic illustration of temperature dependencies of the critical strain for SIDT and DRX

[17].

Therefore, it is clear that there is a critical strain required for SIDT to occur and it is related to the

deformation conditions and chemical composition.

From the above results the increase in the content of carbon in solution increases the critical strain

retarding the SIDT mechanism, while smaller prior austenite grain sizes shows the smaller critical strain

value.

In addition, even the deformation temperature has an important influence on critical strain. It was found

that the critical strain required for SIDT to occur decreases with the deformation temperature, due to the

increase of the transformation driving force.

Gγ

γdeformed

α

Chemical Composition

Fig. 23– Free energy increase after deformation of austenite.

(b) - Strain

It can be confirmed that increasing strain, more and finer SIDT grains can be obtained. When the strain

is small, strain induced transformed ferrite starts to form along the prior austenite grain boundaries. As

the strain is increased, the interiors of austenite grains are activated as nucleation sites for SIDT ferrite,

which are probably deformation bands or twin boundaries.

Results found in literature show that the relationship between SIDT final grain size and strain shows an

asymptotic trend increasing the strain. In fact it can be observed from the above results that increasing

reduction from 40% to 60% the ferrite grain size decreases only of 20% (e.g. 4.2 µm to 3.2 µm), that is

in fully agreement with the other results [17-20].

It means that increasing and increasing the strain, grain size does not decrease significantly.

33

(c) - Prior austenite grain size

From Figs. 10 and 11, it is evident that the smaller prior austenitic grain size, the higher is the volume

fraction of SIDT ferrite under a given deformation condition. This is mainly because the prior austenite

grain boundary is a preferential nucleation site for SIDT ferrite.

A smaller prior austenite grain size has the longer grain boundary area in a given volume of material,

and this will result in the larger volume fraction of ferrite. Moreover, a finer austenite grain reduces the

critical strain for SIDT and therefore results in a finer ferrite grain size at a given strain.

(d) - Deformation Temperature

From our results deformation temperature influences the critical strain, grain size and ferrite volume

fraction. Lower the deformation temperature, lower is the critical strain. If the critical strain for SIDT is

reduced, a finer ferrite grain size and more volume fraction at given strain are observed.

The effect of low deformation temperature is important when the strain is relatively small, because

decreasing the deformation temperature is very effective in increasing the amount of SIDT, but at

relatively high strains decreasing the deformation temperature will affect only slightly the amount of

SIDT.

(e) Steel chemistry

The carbon content has an important role on ultrafine grains formation. The level of ferrite refinement

increases slightly when the carbon content increases, however even the critical strain increases.

Moreover, the presence of Nb retards the SIDT occurrence in terms of higher critical strain needed to

trigger the SIDT mechanism.

Therefore, results show that SIDT was inhibited in the presence of niobium in solid solution compared

to a non-niobium containing steel. However in the case the precipitates of Nb are not dissolved during

heating preceding deformation, SIDT occurs in the Nb steel in a very similar manner as in the non Nb

containing steel, except that the final structure is finer [21].

These elements show that the use of Nb could potentially favour a final refinement through SIDT in

industrial conditions, even if higher strain for SIDT occurrence is required.

(f) - Strain Rate

In this experimentation strain rate was not considered. It is generally thought that the rate of

deformation has a small or even negligible effect on refining grain size during SIDT. A strain rate in

the range of 0.1-10 s-1 was not found to affect significantly the grain size [22]

However Mintz et al. [18] showed that the strain rate can have a certain influence on the fraction of

SIDT ferrite especially when it is formed from a coarse austenite grain. If the strain rate is very low

then extensive dynamic recovery occurs, but if the strain rate is higher then the work hardening takes

place and the conditions for SIDT occurrence are verified; for a finer austenite grain size the strain rate

has less influence on the grain size and the fraction of SIDT.

For a given strain a higher strain rate and a lower deformation temperature are beneficial for the

formation of SIDT ferrite [19].

4.3 Deformed undercooled austenite

Another method to get ultrafine grains increasing the driving force of γ−α transformation is the

deformation of undercooled austenite.

Deformation in undercooled austenite increases the driving force for ferrite formation, contributing to

ultra refinement of ferritic grains.

Adachi et al. [23-25] intensively studied the effect of heavy deformation in undercooled γ region on the

transformation behaviour and revealed that deformation of undercooled austenite enhanced microband

formation rather than dislocation cells with boundaries of larger orientation difference and this is the

main cause for ultra grain refinement.

34

Helium

Quenching

Heating rate:

20 C/sT =700 C

T = 1150 C

t Def. 60%

Quenching to 700 C

Fig. 24–Thermomechanical cycle for undercooled austenite.

In order to compare the refinement ability between SIDT and deformation of undercooled austenite,

other dilatometric tests (Fig. 24) were carried out. In this case samples were heated in order to have the

large prior austenite grain size, rapidly cooled at the deformation temperature of 700°C (ferritic region),

60% deformed and quenched.

Microstructure of CSM1 sample after the thermomechanical cycle shown in Fig. 24 is reported inFig.

25. Fig. 25a shows the microstructure of the sample austenitized at 1150°C x 10 min, cooled at 816°C

and deformed at 60% (SIDT occurrence). Fig. 25b shows the microstructure of the sample austenitized

with the same conditions, quenched to T=700°C and reduced to 60%.

Fig. 26a shows the microstructure of the sample CSM05 austenitized at 1150°C x 30 min cooled at

837°C and 60% deformed. Fig. 26b shows the microstructure of the sample austenitized with the same

conditions, rapidly cooled to T=700°C and reduced of 60%.

Results show that the transformed ferrite volume fraction in the case of deformation of undercooled

austenite is higher, however the morphology and the location of ferrite are different.

Ferrite from deformed undercooled austenite is not polygonal but irregular. Average ferritic grain size is

in both cases around 3-4 µm, but the grain size distribution is different.

Grains starts from less than 1 µm size to arrive to 7-8 µm, while the grain size distribution for the SIDT

ferrite is much more homogenous. Moreover, the samples show a different topological distribution of

ferrite grains. SIDT is fully ferritic in the central part of the sample (Fig. 12) where the local strain is

higher (critical strain for SIDT to occur) while the other sample shows a more homogeneously

distributed ferrite on the sample surface (Fig. 27).

(a) (b)

Fig. 25– Microstructure of CSM1 steel having PAGS =133 µm, deformed of 60% at (a) 816°C, (b)

700°C.

35

(a) (b)

Fig. 26– Microstructure of CSM05 steel having PAGS =160 µm, deformed of 60% at (a) 837°C, (b)

700°C.

Fig. 27 - Macrostructure of CSM1 steel after deformation at 700°C.

Textural analyses were carried out to compare the samples. OIM analyses showed that a larger density

of grains having typical deformation texture is present in the undercooled sample.

4.4 Laboratory Rolling

Laboratory rolling was carried out to reproduce as closely as possible the schedules used in dilatometer

experiments in order to get ultra fine grains on surface.

A first hot rolling tests started from 20 mm initial thickness to 10 mm of final thickness according to the

rolling schedule shown inFig. 28. Samples were heated in a furnace at 1100°C for 40 min and then their

two surfaces were submitted to different cooling methods in order to produce different local

thermomechanical conditions and to understand the best cooling conditions to be used to get ultrafine

grains: one surface was cooled by compressed air and the other was left in still air. When the forced air

cooled surface achieved a temperature of about Ar3+25°C, the piece was rolled in one pass giving 50%

reduction.

Thermocouples were embedded at mid-thickness and close to the surface of the samples. Measured hot

rolling temperatures are shown in Table 3.

36

T = Ar3 + 25 C

Def. = 50%

T = 1100 C

Water Quenching

t = 40 min

Forced Air

cooling

T = Ar3 + 25 C

Def. = 50%

T = 1100 C

Water Quenching

t = 40 min

Forced Air

cooling

Fig. 28 – Hot rolling thermomechanical cycle carried out on CSMlaboratory mill.

Trolling (°C)

Heat Ar3

(°C)

Ar3+25

(°C) Mid-thickness Surface

CSM05 823 848 845 795

CSM1 790 815 800 750

Table 3 – Temperatures measured during pilot mill rolling.

Unfortunately, due to problems during the tests, the samples were hot rolled at too low surface

temperature going into the ferrite region at the surface (Table 4). At the centre of the sample the rolling

temperature was slightly greater than Ar3 and SIDT mechanism was detected. Further it is necessary to

notice that the real deformation was 40%.

In Fig. 29 microstructures of samples hot rolled at CSM pilot mill are shown.

37

CSM1 CSM05

Surface A

Mid-th.

Surface B

Fig. 29– Microstructures at different through positions thickness of samples hot rolled at pilot

Mill (1st test).

Surface “A” is the one cooled by forced air, surface “B” is the one cooled in still air. It can be noted that

the surface “A” shows a thicker layer of ferritic grains compared with the other surface. Grains have a

size around 10 µm and are elongated along the rolling direction due to the deformation in the ferritic

range. Going down to the mid-thickness, the hot rolling temperature increased greater than Ar3.

Therefore, strain induced transformation occurred at mid-thickness even if the volume fraction was low.

Ferrite grains formed at the prior austenite grain boundaries. Like outlined before about SIDT

mechanism, the SIDT ferrite volume fraction is higher when carbon content decreases.

After this first test, the rolling trials were repeated and the rolling conditions were modified.

From the results obtained from SIDT mechanism, it is clear that in order to get ultrafine grains

important parameters are the deformation temperature and the prior austenite grain size.

38

Lower the deformation temperature and finer prior austenite grain size, finer is the ferrite grain size and

more SIDT ferrite volume fraction is present.

However, the application of these rules must match with the possibility of the laboratory rolling mill.

Decreasing the hot rolling temperature causes an increase of the rolling force needed to give the same

reduction. Moreover, finer grain size is related to low temperature and short time thermal treatment

before deformation, that is in contrast with the temperature homogenisation of the sample in the furnace

(for thick samples).

In this test all the analysed grades (C-Mn and C-Mn-Nb) were warm rolled as shown in Table 4.

Rolling conditions

Heat Ar3

(°C)

Ar3+25

(°C)

T Surface

(°C)

Measured

Reduction (%)

CSM05 823 848 846 34

CSM1 790 815 828 34

988 760 785 790 30

1022 785 810 817 30

Table 4 – Hot rolling conditions during the 2nd

test.

The initial thickness was 10 mm and the heating time before deformation was 950°C x 30 min.

After rolling the samples were quenched in order to study the microstructural situation and to verify if

SIDT occurred at surface.

Final microstructures and hardness measures through the thickness are reported in Figs. 30 and 31 for

plain low C steels, and in Figs. 32 and 33 for Nb microalloyed steels respectively.

It can be seen that fine ferritic grains were found at the surfaces for all the materials, while at the centre

presence of quenched microstructure or coarse grains were found.

In Fig. 30 microstructures at different through thickness positions for CSM1 and CSM05 steels are

reported. It can be noted that for CSM1 at surface fine ferritic polygonal grains are detected until a

depth of around 400 µm, after that, mixed ferrite+bainite structure is revealed. Sample CSM05 shows a

ferrite+pearlite structure homogeneously distributed through the thickness showing grain size ranging

from about 4.5 µm at surfaces to 8-9 µm at centre, while CSM1 sample showed a slightly coarser grain

size at surface (around 6 µm) (Table 5).

Hardness measurements (Fig. 31) reflects the situation of grain sizes and microstructure measured at

surface and centre.

For microalloyed Nb steels (Fig. 32) microstructures at different through thickness positions for 988

and 1022 steels are reported. In this case sample 1022, having the lowest C content showed at surface

fine ferritic grains whose dimensions are reported in Table 6, which can be detected until a depth of

around 300 µm, after that, mainly bainitic structure is revealed. The grains at surface were around 2-3

µm but their shape is not polygonal. Sample 988 shows at very thin surface layer with fine ferritic

grains and their shape is not polygonal. Hardness measurements are reported in (Fig. 33).

From these results it can be concluded that SIDT was activated only in the case of plain low C steels

(CSM05 and CSM1), where the strain imposed at surfaces, given during hot rolling, was over the

critical strain for SIDT.

39

CSM1 CSM05

Surface.A Surface A

Center Center

Surface B Surface B

Fig. 30– Microstructures at different through positions thickness of samples CSM1 And CSM05 hot


40

Ferritic Grain Size (µm)

CSM05 CSM1

Surface A 4.5 6

Centre 8-9 -

Surface B 5 7

Table 5 – Grain size measurements at different positions through the thickness.

Sample CSM05

100

150

200

250

300

350

0 2 4 6 8

HV

1

Distance traversed through thickness(mm)

Sample CSM1

100

150

200

250

300

350

0 2 4 6 8

Distance traversed through thickness (mm)

HV

1

Fig.31– Hardness through the thickness for plain low-C steels.

41

988 1022

Surface A Surface A

Center Center

Surface B Surface B

Fig. 32– Microstructures at different through positions thickness of samples CSM1 And CSM05 hot


42

Sample 988

100

150

200

250

300

350

0 2 4 6 8


HV

1

Sample 1022

100

150

200

250

300

350

0 2 4 6 8Distance traversed through thickness (mm)

HV

1

Fig. 33– Hardness through the thickness for microalloyed low-C steels.

Ferritic Grain Size (µm)

988 1022

Surface A 2 3

Centre - -

Surface B 2.5 3

Table 6 – Grain size measurements at different positions through the thickness.

4.5 Conclusions

SIDT route was confirmed to be a method to produce ultrafine grains.

Ultrafine grained ferrite, ranging between 3 and 4 µm, was obtained in all of the four tested low-carbon

steels: two plain low C-Mn, steel (0.05 and 0.1%C) and two microalloyed low C-Mn-Nb steels (0.06

and 0.12 %C and 0.03Nb).

In this work the factors influencing SIDT occurrence and grain size refinement were investigated. The

experimental results showed that the controlling parameters of SIDT route are: refined prior austenite

grain size, low deformation temperature and high strain.

43

In particular it was confirmed that:

(1) There is a critical strain required for SIDT to occur and it is related to the deformation conditions,

prior austenite grain size and chemical composition. Minimum strains required for ultrafine ferrite for

this experimentation were about 0.6, in the case of lowest carbon content (plain low C steel with

0.05%C) and finest prior austenite grain size (14 µm).

(2) When the strain is small, strain induced transformed ferrite starts to form along the prior austenite

grain boundaries. As the strain is increased, the interiors of austenite grains become activated as

nucleation sites for SIDT ferrite increasing the strain SIDT ferrite volume fraction increases.

(3) Higher strain applied results in a finer ferrite grain size and in higher deformation induced ferrite

fractions.

(4) Deformation at a low temperature, just above Ar3, results in finest grain sizes.

(5) Refining the prior austenite grain size highly enhances the ferrite grain size refinement, and it is

quite crucial in practice.

(6) The presence of Nb increases the critical strain compared to a niobium free steel.

The transformed ferrite volume fraction in the case of deformation of undercooled austenite is higher

than in SIDT samples, however the morphology and the location of ferrite are different. Ferrite from

deformed undercooled austenite is not polygonal but irregular and the grain size distribution is wider.

Grains ranges from less than 1 µm size to 8 µm, while the grain size distribution for the SIDT ferrite is

much more homogenous.

The results by deformation dilatometry were applied to pilot hot rolling mill in order to develop

ultrafine surface grained steels.

SIDT route was applied to plain and microalloyed low carbon steels using laboratory-scale pilot mill

and in plain low C steels surface grains whose size was around 4-5 µm were detected.

High shear strain at the surface and low deformation temperature are the main factors which promote

the formation of UFF surface layers, but differently from Hodgson et. al [10,11] we found that starting

from very coarse austenite grains is not beneficial for grain refining. Smaller the prior austenite, finer is

the SIDT ferrite grain. Exceeding the critical strain increased the volume fraction of ferrite.

It is evident that chemical composition of the steel influences the morphology and the volume fraction

of ultrafine grains formed at the surface layers of a strip and it significantly changes the microstructure

formed in the core.

A special note must be done regarding the influence of cooling path following deformation.

Cooling applied just after rolling is of common practice in order to limit the growth of ferrite and so

refine the final structure. This should be also true in the case of strain induced ferrite, and the role of the

cooling path following deformation is of major interest to be investigated.

Besides, cooling also controls the type and morphology of second phases, which has a direct influence

on the final strain hardenability of the steels.

However, in the industrial plants there exists a delay between rolling and cooling, that will also affect

the microstructure, since it enhances the growth of existing ferritic nuclei. Therefore the effect of static

recrystallization must be considered.

44

5. ULTRAFINE LOW AND MEDIUM PLAIN C STEELS

5.1 Introduction

Since on an actual mass production scale there are limitations to the application of low temperature

large scale deformation, it is necessary to develop a technology to achieve commercially meaningful

grain refinement using existing rolling facilities. If it is insufficient, additional special rolling mills (e.g.

Pony Mill) should be added to the already existing rolling lines.

The PonyMILL is a new technology which enables a conventional mill operator to produce a

percentage of the hot mill throughput as cold rolled substitute, economically and with low investment

costs. The physical process route briefly consists of a close coupled coiler after the tandem finishing

mill exit with rapid coil transfer to an un-coiler. The strip is then fed into a single high reduction stand

where the strip is reduced in a fully or partially ferritic state to the desired gage; a coiler then recoils the

strip before it is passed back into the existing coil handling system.

In the present investigation the mechanisms of producing ultrafine grained C-Mn steels under the

consideration of industrial application are described. In particular, the evolution of microstructure and

texture during warm deformation based on PonyMILL technique, the tensile properties at room

temperature and the Charpy impact properties of the ultrafine grained steels, the effect of carbon

contents on improving the work hardening rate (uniform elongation) of ultrafine grained steels, and the

effect of carbon and manganese contents on ultra-grain refinement were studied.

5.2 Experimental

Four plain C-Mn steels with different carbon and manganese contents were melted in a laboratory

vacuum induction furnace. Table 7 shows the chemical compositions of the steels. Compositions are

given in mass%. The study was concentrated on the steels with carbon contents of ≥ 0.15%. The effect

of carbon content (15C, 2C and 3C steel) and the effect of manganese content (2C, 2CMn steel) were

investigated.

Steel Fe C Si Mn P S Al N Ae3

15C bal. 0.17 0.22 0.76 0.004 0.004 0.031 0.001 1107

2C bal. 0.22 0.21 0.74 0.004 0.003 0.029 0.001 1093

2CMn bal. 0.23 0.22 1.52 0.004 0.004 0.03 0.001 1070

3C bal. 0.31 0.22 0.76 0.003 0.003 0.030 0.001 1071

Table 7 Chemical compositions (mass%) of the steels, calculated Ae3 (K) [26]

Each cast ingots had a weight of 70 kg and a size of 140 × 140 × 300 (width × length × height) mm3.

The laboratory samples were machined directly from the ingot into cylindrical specimens (∅5 mm × 10

mm (diameter × length)) for deformation dilatometry, and into rectangular parallelepiped samples, 18 ×

18 × 30 and 50 × 40 × 60 (width × length × height) mm3, for the plane strain compression tests which

were conducted in a large scale set-up with a maximum force of 2.5 MN [27]. This servohydraulic press

is capable of conducting large scale thermomechanical processes by performing multi-step plane strain

compression tests as a realistic approximation of industry-scale rolling operations.

Different experimental routes were tested in order to optimize the process conditions before during and

after deformation in order to get ultrafine structures.

In particular:

• Continuous-cooling-transformation diagrams after deformation of C-Mn steels

• Ultrafine grained steels produced by warm deformation

• Microstructure and mechanical properties of ultrafine grained steels

were studied.

45

5.2.1 Continuous-cooling-transformation diagrams after deformation of C-Mn steels

Optimization of austenite deformation temperature (experiment I)

In order to get a fine grained and homogeneous austenite, the optimum austenitization and deformation

temperatures prior to the γ → α (austenite to ferrite) transformation were determined by pre-tests.

According to this, the samples were austenitized at Ae3 + 100 K for 3 minutes, air cooling to 1013&1153

K, then deformed with a logarithmic strain of ε = 0.3 and a strain rate of ε& = 10 s-1

. The aim of using

different deformation temperatures is to optimize a fully recrystallized austenite as would be expected

after a conventional hot rolling. Water quenching was used immediately after the deformation in order

to study the resulting austenite microstructure. The experiment I is shown is Fig. 34.

Continuous-cooling-transformation diagrams after deformation (experiment II)

To establish a continuous cooling transformation diagram for each of steel, cooling rates from 2 to 64

K/s were used after deformation at the optimum temperature, i.e. 1133 K for the 15C steel, and 1143 K

for the other steels, which was determined by experiment I. Fig. 35 shows the experiment II.

Fig. 34. Optimization of austenite deformation temperature (experiment I).

Fig. 35. Determination of continuous-cooling-transformation diagrams after deformation (experiment

II).

5.2.2 Ultrafine grained steels produced by warm deformation

Warm deformation of a ferrite-pearlite microstructure (experiment III)

A large strain warm deformation was applied after the γ → α transformation in order to refine the ferrite

and to spheroidize the pearlitic cementite lamellae. The optimum cooling rates after the first austenite

deformation step were determined by the continuous-cooling-transformation diagrams after deformation

for each of steel in experiment II. They are 12 K/s, 6.5 K/s, 5 K/s and 5 K/s for the 15C, 2C, 2CMn and

3C steel, respectively. These values were applied in the controlled cooling experiment III (Fig. 36a), to

46

get a fine ferrite-pearlite microstructure (without bainite) prior to the large-strain multi-step warm

deformation. After holding the samples for 2 min, four deformation steps were exerted in the

temperature range between 773 K and 973 K. Each of the four subsequent steps imposed a logarithmic

strain of ε = 0.4 at a strain rate of ε& = 10 s-1

accumulating to a total strain of ε = 1.6, and an inter-step

time of 0.5 s. Subsequently an annealing treatment of 2 h was imposed at the respective deformation

temperature in order to mimic coiling conditions.

Warm deformation of a bainitic microstructure (experiments IV and V)

In experiments IV and V, a bainitic microstructure was produced by using a cooling rate of 50 K/s after

the deformation in the austenite regime. This treatment provided the starting microstructure for the

subsequent large strain deformation. In experiment IV (Fig. 36b), the large strain deformation was

performed below the bainite finish temperature. In order to obtain a recrystallized ferrite containing

homogeneously distributed cementite particles, the material was reheated after the large strain

deformation to different temperature levels in the range between 773 K and 1003 K, using a holding

time of 2 h to approximate coiling conditions. In experiment V (Fig. 36c), after the complete

transformation the material was first reheated to the respective deformation temperatures in the range

between 823 K and 1003 K. Coiling was simulated by holding the samples for 2 h at the deformation

temperature.

(a) (b) (c)

Fig. 36. Processing schedules for the compression tests. LSW: large strain warm deformation; Trh:

reheating temperature; Ar3: austenite to ferrite transformation temperature; Pf: pearlite finish

temperature; Bf: bainite finish temperature.

5.2.3 Microstructure and mechanical properties of ultrafine grained steels

Initial ferrite-pearlite microstructure before large strain warm deformation (experiment VI)

Since austenite refinement is essential for obtaining small grains after transformation the experiments

were in all cases carried out with a low austenitization temperature (experiment I). After reheating with

a heating rate of 10 K/s, the samples were austenitized at Ae3 + 100 K for 3 minutes to obtain a fine

austenite grain size. After air cooling to 1133 K (15C steel) or 1143 K (2C, 2CMn, 3C steel), a one-step

deformation was used for all materials with a true logarithmic strain of ε = 0.3 and a strain rate of ε& =

10 s-1

, in order to obtain a fully recrystallized austenite, as would be expected after a conventional

recrystallization controlled hot rolling. This was followed by the controlled cooling down to pearlite

finish temperature with cooling rates of 12 K/s (15C steel), 6.5 K/s (2C steel) and 5 K/s (2CMn, 3C

steel), which were determined by pre-tests of deformation dilatometry as maximum cooling rates to get

a bainite free ferrite-pearlite microstructure in each steel (experiment II).

In the following, different experimental routes were carried out after the same treatment (experiment

VI).

Conventional route (CR)

In order to study the initial pearlite-ferrite microstructure before large strain warm deformation and the

effect of grain size on mechanical properties, a conventional route was carried out as shown in Fig. 37a.

That is, the samples were water quenched at 823 K after the controlled cooling in order to get a bainite

free ferrite-pearlite microstructure.

47

Ultrafine grain route (UFR)

After 2 minutes holding at 823 K, which corresponds to the time necessary for some handling in the

industrial processing, the large strain warm deformation was performed by a four-step plane strain

compression with short inter-step times of 0.5 s. Each of the four subsequent steps imposed a

logarithmic strain of ε = 0.4 at a strain rate of ε& = 10 s-1

accumulating to a total strain of ε = 1.6.

Subsequently an annealing treatment of 2 h at 823 K was imposed in order to mimic coiling conditions.

The processing route is shown in Fig. 37b. It should be noted that for the 2CMn steel the sample was

first cooled to 673 K, then, reheated to the large strain warm deformation temperature. This is due to the

pearlite finish temperature which is below 823 K for the 2CMn steel.

Microstructure evolution during warm deformation

In order to study the evolution of microstructure and the formation of the ultrafine structure during the

large strain warm deformation, samples were water quenched after different strains, as shown in Fig.

37c. In this way, the microstructure variations with increasing strain (ε = 0.4, 0.8, 1.2, 1.6) were studied.

High temperature annealing

The 2C steel was annealed for 2 h at an elevated temperature of 973 K after the large strain warm

deformation at 823 K, Fig. 37d. The reason for the annealing at high temperature is to obtain a coarser

size of cementite particles. The coarser size of particles is required for the qualitative measurement of

carbon/manganese distribution in particles by energy-dispersive spectrometry (EDS). The particles with

a size of less than 100 nm can not be represented in an electron backscatter diffraction (EBSD) map

since the step size used in the present study is 50 nm. A smaller step size will lead to bad pattern quality

and lower resolution for the EBSD and EDS measurement.

Fig. 37. Processing schedules for the plane strain compression tests. Trh: reheating temperature; Ar3:

austenite to ferrite transformation temperature; Pf: pearlite finish temperature

(a) conventional route (CR) for getting coarser microstructure, (b) ultrafine grain route (UFR) for

producing ultrafine microstructure, (c) investigation of microstructure evolution during warm

deformation (ε = 0.4, 0.8, 1.2, 1.6), (d) high temperature annealing after large strain warm (LSW)

deformation

(a) (b)

(c) (d)

48

5.3. Results

5.3.1. Optimization of austenite deformation temperature

The optimum deformation temperatures to get a fine and recrystallized austenite are 1133 K for the 15C

steel, and 1143 K for the 2C, 2CMn and 3C steels. In all materials higher deformation temperatures lead

to coarse microstructures, whereas at lower temperatures, some pancaking phenomena indicate an

incomplete recrystallization. The water quenched microstructures of the steels after deformation at the

optimum temperatures are shown in Fig. 38. As shown by the dashed lines in the figure, the initial

austenite grain boundaries decorated by ferrite are equiaxed, which demonstrates that austenite

recrystallization occurred before transformation.

Fig. 38. Water quenched microstructure after deformation at 1133 K for the 15C steel (a), and at 1143

K for the 2C (b), 2CMn (c) and 3C (d) steel (dashed line shows an example of an equiaxed prior

austenite grain). CD: compression direction; TD: transverse direction.

5.3.2. Continuous-cooling-transformation diagrams after deformation

Fig. 39 shows the continuous-cooling-transformation diagrams of the steels after austenite deformation

at the optimum temperatures. If the cooling rate is smaller than 12 K/s, a fully ferrite and pearlite

microstructure can be obtained in the 15C steel. Similarly, the critical cooling rate to get a fully ferrite

and pearlite microstructure for the 2C, 2CMn and 3C steel is about 6.5 K/s, 5 K/s and 5 K/s,

respectively.

(a) (b)

(c) (d)

49

(a) (b)

(c) (d)

Fig. 39. Continuous-cooling-transformation diagrams of the experimental steels after deformation (ε

=0.3, ε& =10s-1

) at the optimum temperatures (1133 K for the 15C steel, and 1143 K for the other steels)

and continuous cooling. The cooling rates varied from 2 to 64 K/s. F: ferrite; P: pearlite; B: bainite; Ms:

martensite start temperature. (a) 15C steel; (b) 2C steel; (c) 2CMn steel; and (d) 3C steel.

Effect of chemical composition on phase transformation

Fig. 40 shows the ferrite and pearlite transformation diagrams of the experimental steels at the cooling

rates of 2&64 K/s. The illustration confirms that higher carbon content retards the diffusion controlled

transformation (austenite to ferrite), which is reflected in a shift of the beginning transformation curve

(ferrite start temperature Fs) to the lower temperatures and longer times with increasing carbon content.

On the other hand, the carbon content has little effect on the pearlite finish temperature curve (pearlite

finish temperature Pf, which is similar for the 15C, 2C and 3C steels.

Manganese is an element that enlarges the austenite region. It obviously retards ferrite formation and

the following pearlite and bainite formation, which can be seen from the curves (the lines with triangle)

of the 2CMn steel in Fig. 40.

50

Fig. 40. Comparison of the results from Fig.39 for all the steels. The cooling curves are omitted for

clarity. Fs: ferrite transformation temperature; Pf: pearlite finish temperature; F: ferrite; P: pearlite.

Effect of cooling rate on microstructure

As shown in Fig. 39 the microstructure changes with increasing cooling rate. As an example, Fig. 41

shows the microstructures of the 3C steel with increasing cooling rate. The main differences between

the microstructures produced by different cooling rates were the amount and the size of pro-eutectoid

ferrite, the thickness of pearlitic cementite lamellae and the presence of bainite and martensite. After a

slow cooling rate of 2 K/s, the microstructure contains coarse polygonal ferrite and thick lamellar

pearlite, Fig. 41a. After a faster cooling rate, the size of polygonal ferrite and the thickness of pearlitic

cementite lamellae are smaller. Increasing the cooling rates above 5 K/s, bainite (Fig. 41b) and later

martensite (Fig. 41c) are present in the microstructure. Therefore, the cooling rate of 5 K/s was used in

the experiment to get an initial fine and bainite free ferrite and pearlite microstructure before large

strain warm deformation.

Fig. 41. Microstructure changes with increasing cooling rate in the 3C steel. CD: compression direction;

TD: transverse direction. (a) 2 K/s, (b) 16 K/s and (c) 64 K/s.

5.3.3 Effects on microstructure and hardness of the ultrafine grained steels after compression tests

Effect of temperature on microstructure

In order to study systematically the microstructure of the steels produced by the different experiments

(Fig. 36), the 3C steel was selected as an example.

Fig. 42 shows the optical micrographs of the 3C steel processed by the experiments in Fig.36 at the

elevated temperatures. The temperature has different meanings for the different processes. It represents

the multi-step deformation and annealing temperature in the experiment III and experiment V, while it

represents only the annealing temperature in the experiment IV. The microstructures obtained for the

three processing schemes are similar, i.e. they consist of ferrite and globular cementite particles. The

microstructure after the experiment V is also shown as a SEM image (Fig. 43a) and EBSD map (Fig.

51

43b). The microstructure obtained for a heat treatment above 913 K reveals a more equiaxed grain

shape. Lower temperatures lead to finer ferrite grains and cementite particles (5&350 nm), but the grain

shape is more elongated than that after 973 K. An increase in the deformation temperature results in

coarser and more equiaxed ferrite grains. The distribution of cementite particles tends to be more

homogeneous and the particle size becomes larger (250&950 nm) at temperatures of 973 K.

Fig. 42. Optical micrographs of the 3C steel after the experiments in Fig. 36. CD: compression

direction; TD: transverse direction; Temp.: multi-step deformation and subsequent annealing

temperatures in the experiment III and experiment V, annealing temperature in the experiment IV.

52

Fig. 43b represents the EBSD maps from the same samples as shown in Fig. 43a. The maps show the

image quality index (IQ) of the pseudo-Kikuchi lines together with the grain boundaries. Black lines

indicate misorientations above 15° between adjacent points, i.e. they represent high-angle grain

boundary segments. The white lines indicate misorientations between 2° and 15°. The white spots at the

grain boundary triple points visible in the map taken on the sample treated at 973 K are cementite

particles with sizes in the range between 300 nm and 950 nm, i.e. they are larger than the step size (100

nm) of the EBSD measurement.

(a)

(b)

Fig. 43. SEM images (a) and EBSD maps (b) of the 3C steel after the large-strain deformation (ε = 1.6)

and subsequent 2 h annealing at different temperatures (experiment V). Details of the processing are

given in Fig. 36c. CD: compression direction; TD: transverse direction. The black lines indicate grain

boundary misorientations between 15° and 63°. The white lines indicate grain boundary misorientations

between 2° and 15°.

Effect of temperature on grain boundary orientation distribution

Fig. 44 shows the fraction of high-angle grain boundaries of the 3C steel for the three experiments in

Fig. 36. For the different deformations (experiment III and V) and annealing temperatures (experiment

IV) ranging from 773 K to 973 K, the fraction of high-angle grain boundaries increases with

temperature but remains between 55% and 70%. A detailed analysis of the grain boundary maps,

though, reveals that many grains are bordered by both, high- and low-angle grain boundaries.

Fig. 45 shows the EBSD analysis of the grain sizes together with the grain shape aspect ratios as

obtained by experiment V. The data reveal true grain sizes (only high-angle grain boundaries are

considered) in the range between 1 µm and 2 µm (Fig. 45a). The aspect ratio for these grains amounts

to 2 for the low temperature regime and drops to 1.5 for higher temperatures. Fig. 45b, the (sub)grains

(all the grains with both high- and low-angle grain boundaries) reveal a decreased aspect ratio of about

1.2 when the temperature increased to 913 K.

Fig. 46 shows the corresponding average data for the grain shape aspect ratio for all the processes in Fig.

36. All grains and subgrains are elongated perpendicular to the compression direction at low

temperatures, but more equiaxed at high temperatures. For the experiment IV the aspect ratio of the

grains and subgrains almost drops by a factor of two within the investigated temperature range.

53

Fig. 44. Fraction of high-angle grain boundaries (HAGBs) of the 3C steel obtained for the three

different experiments in Fig. 36.

Fig. 45. Average ferrite grain and subgrain sizes in the CD (compression direction) and TD (transverse

direction), respectively, for the 3C steel. The right hand ordinate indicates the corresponding values for

the grain shape aspect ratio (GSAR), data are taken from the experiment V (Fig. 36c): (a) only high-

angle grain boundaries are considered in the measurement (15° ≤ θ < 63°), (b) both high- and low-

angle grain boundaries are considered in the measurement (2° ≤ θ < 63°)

Fig. 46. Average values for the grain shape aspect ratio (GSAR) of the 3C steel after the three different

processes in Fig. 36. (a) only high-angle grain boundaries are considered in the measurement (15° ≤ θ

< 63°), (b) both high- and low-angle grain boundaries are considered in the measurement (2° ≤ θ < 63°)

54

Effect of carbon content on grain size

The effect of carbon content on the average ferrite (sub)grain size after the different experiments (Fig.

36) is shown in Fig. 47. With increasing carbon content the average ferrite (sub)grain size decreases in

the experiment III and IV (Fig. 47A and B). It is interesting to note that this phenomenon is not valid in

the experiment IV, Fig. 47C, when the temperature is between 823 K and 843 K. That means the

average grain size is similar for all the materials in the experiment IV at this temperature range.

(a)

(b)

Fig. 47. Effect of carbon content on average ferrite (sub)grain size after the experiments in Fig. 36. (a)

high-angle grain boundaries are considered in the measurements

(b) high- and low-angle grain boundaries are considered in the measurements

55

Effect of experimental procedures on grain size

The effect of the experimental routes in Fig. 36 on the average ferrite (sub)grain size of the steels is

shown in Fig. 48. It can be seen that the average grain size of each steel is similar after the different

routes.

Effect of temperature on hardness

The effect of temperature on macrohardness (HV5) is shown in Fig. 49 for the steels processed by the

different experiments in Fig. 36. The increase in temperature leads to a decrease in hardness. For each

of steel processed by the different routes, the macrohardness is quite similar. However, at the same

temperature, the increase in carbon and especially manganese content is effective in increasing hardness.

Fig. 48. Effect of the experimental routes in Fig. 36 on average ferrite (sub)grain size for the different

steels. (a) high-angle grain boundaries are considered in the measurement, (b) high- and low-angle grain

boundaries are considered in the measurement

56

Fig. 49. Macrohardness changes with temperature after the different experiments in Fig. 36

(Experiment III: warm deformation of a ferrite-pearlite microstructure; Experiment IV and V: warm

deformation of a bainite microstructure).

(a) 15C; (b) 2C; (c) 2CMn; (d) 3C.

5.3.4 Microstructure and mechanical properties of the conventional steels

Microstructure

The microstructures of the steels after the conventional route (Fig. 37a) consist of ferrite and pearlite

(Fig. 50). Higher carbon content leads to finer ferrite grains and a higher volume fraction of pearlite.

Compared to the lower manganese steel of 2C, a much finer ferrite microstructure can be observed in

the higher manganese steel 2CMn (Fig. 50c).

Fig. 51 shows the average ferrite grain size after the conventional route. A higher carbon content and

especially a higher manganese content results in finer ferrite

(a) (b)

(c) (d)

57

Fig. 50. Optical micrographs of the experimental steels after the conventional route. Details of the

processing are given in Fig. 37a: (a) 15C steel; (b) 2C steel; (c) 2CMn steel; (d) 3C steel.

Fig. 51. Grain size of the experimental steels after the conventional route. Details of the processing are

given in Fig. 37a.

Mechanical properties

The engineering stress-strain curves of the steels after the conventional route are shown in Fig. 52. The

tensile tests were carried out at room temperature. The average ferrite grain size of each of the steels is

also shown in the figure. It can be seen that higher manganese and carbon contents lead to smaller grain

sizes and subsequently higher strength. However, the total elongation hardly changes especially when

the carbon content increases from 0.15% to 0.3%.

(c) (d)

(a) (b)

58

Fig. 53 shows the results of Charpy impact testing on subsize specimens of the experimental steels after

the conventional route. With increasing carbon content the upper shelf energy decreases obviously in

the ferrite-pearlite steels.

Fig. 52. Comparison of stress-strain curves of the experimental steels with different ferrite grain sizes.

The steels were processed by the conventional route (CR), as shown in Fig. 37a.

Fig. 53. Charpy impact properties for subsize specimens of the experimental steels after the

conventional route (CR). Details of the processing are given in Fig. 37a.

5.3.5 Microstructure and mechanical properties of the ultrafine grained steels

Microstructure

Fig. 54 shows the SEM micrographs of the experimental steels after the large strain warm deformation

(ε = 1.6) and annealing at 823 K for 2 h (Fig. 37b). In all steels the microstructure consists of ferrite

with a small grain size and globular cementite. The finer cementite particles (5&90 nm) are distributed

inside the ferrite grains. The coarser ones (90&350 nm) are located at the ferrite grain boundaries, acting

as obstacles for the migration of ferrite grain boundaries in ND. Increasing carbon content from 0.15%

to 0.3% the chains of cementite particles locating at ferrite grain boundaries disappear. The cementite

59

particles are distributed more homogenously at both grain boundaries and inside the ferrite grains.

Meanwhile the ferrite grain size becomes smaller and the grain shape more equiaxed. Compared to the

2C steel, much finer and more equiaxed ferrite grains can be observed in the 2CMn steel.

The effects of carbon and manganese contents on the average ferrite (sub)grain (grain and subgrain)

size and grain shape aspect ratio (GSAR) are shown in Fig. 55. It can be seen that the increase in carbon

and manganese contents leads to a decrease in the average ferrite grain size and grain shape aspect ratio.

Especially the manganese shows a significant effect, a finer and nearly perfect equiaxed grain structure

is represented in the 2CMn steel. The fraction of high-angle grain boundaries is shown in Fig. 56. The

increase in carbon content leads to a slightly increase of the fraction of high-angle grain boundaries.

Comparably, the increase in manganese content is more effective to enhance the fraction of high-angle

grain boundaries. As an example, Fig. 57a shows the grain boundary misorientation map of the 2CMn

steel after the ultrafine grain route. The red lines represent high-angle grain boundaries and the green

lines low-angle grain boundaries. It can be seen that most of the grain boundaries are decorated by the

red line segments. The grain boundary misorientations with bimodal distribution characteristic can be

observed in Fig. 57b. The two peak values were represented at both low- and high-angle grain boundary

misorientation regions, i.e. one is in the range of 2&12° and the other is between 50° and 62°.

(a) (b)

(c) (d)

Fig. 54. SEM images of the experimental steels after large strain deformation and 2 h annealing at 823

K (Fig. 37b: ultrafine grain route):(a) 15C steel; (b) 2C steel; (c) 2CMn steel; (d) 3C steel.

60

Fig. 55. Average ferrite (sub)grain size (a) and grain shape aspect ratio (GSAR) (b) of the experimental

steels after large strain deformation and 2 h annealing at 823 K (Fig. 37b: ultrafine grain route).

Fig. 56. Fraction of high-angle grain boundaries (HAGBs) of the experimental steels after large strain

deformation and 2 h annealing at 823 K (Fig. 37b: ultrafine grain route).

(a)

(b)

61

Fig. 57. Grain boundary characterization of the 2CMn steel after large strain deformation and 2 h

annealing at 823 K (Fig. 37b: ultrafine grain route).

(a) EBSD image: the red lines indicate grain boundary misorientations between 15° and 63°; the green

lines indicate grain boundary misorientations between 2° and 15°

(b) grain boundary misorientation distribution

Effect of manganese on grain refinement

(a) EDS analysis

As shown in Fig. 54 and 55, very fine and more equiaxed grains were observed in the higher manganese

steel 2CMn. Due to the same experimental processing applied for all the steels, therefore, special

attention was paid on the effect of manganese on grain refinement, especially the effect of manganese

during large strain warm deformation and subsequent annealing.

Fig. 58 shows the image quality map of the 2C steel after large strain warm deformation (ε = 1.6) and

subsequent 2 h annealing at 973 K. The reason to use this high annealing temperature after warm

deformation at 823 K is to get a large cementite size. The cementite particles with a size of less than

100 nm can not be represented in an EBSD map (step size: 50 nm). A smaller step size will lead to bad

pattern quality and lower the resolution for EBSD and EDS technique to measure the distribution of

carbon and manganese. The image quality map, Fig. 58, represents the same microstructure components

as in the case of the annealing temperature of 823 K: ferrite and globular cementite. Nevertheless,

compared to the SEM micrograph of the 2C steel in Fig. 54b, some coarse ferrite grains appear in the

microstructure. A large number of cementite particles can be observed inside the coarser grains.

(a)

(b)

62

Fig. 58. Image quality map of the 2C steel after large strain deformation at 823 K and subsequent 2 h


Fig. 59. EBSD maps of the 2C steel processed by large strain deformation at 823 K and subsequent 2 h


(a) image quality map; (b) carbon distribution; (c) manganese distribution.

(b)

(a)

(c)

63

In order to study the distributions of carbon and manganese, a small area in Fig. 58 was selected for the

EBSD and EDS measurements. Fig. 59a shows the image quality map of the measured area. Fig. 59b

and c represent the distribution of carbon and manganese, respectively. The color band from blue to red

displays the increase of local carbon and manganese concentrations. In Fig. 59b and c, the arrows point

out red areas which indicate high carbon and manganese concentrations, respectively. It shows that

besides carbon, manganese is also preferred to locate at the same positions as the cementite particles,

shown in Fig. 59a.

(b) TEM analysis

Fig. 60. Manganese content (mass%) in the matrix and in the cementite particles determined by the

STEM-EDS technique for the 2C (0.22 mass%C-0.74 mass%Mn) and 2CMn (0.23 mass%C-1.52

mass%Mn) steels after warm deformation (ε = 1.6) and subsequent annealing.

The distributions of carbon and manganese can be qualitatively analyzed by the EBSD and EDS

measurements. Using TEM and scanning electron microscopy (STEM) with EDS technique, the amount

of manganese in solid solution in both the ferrite matrix and cementite particles can be quantitatively

determined. Fig. 60 shows the participation of Mn in both matrix and particles for the 2C and 2CMn

steels processed by the large strain warm deformation and subsequent annealing. Each value represents

an average value from between five to ten measurements. As expected, the manganese content in the

ferrite matrix is higher in the 2CMn steel if compared to that in the 2C steel, which is related to the

chemical compositions of each steel (Table 7). However, it is important to note that the manganese

content in the cementite particles of the 2CMn steel (~7.5%) is significantly higher than that in the 2C

steel (~3.1%).

The TEM micrographs of the 2C and 2CMn steels after large strain warm deformation and direct water

cooling are shown in Fig. 61a and b, and after large strain warm deformation and a subsequent

annealing are represented in Fig. 61c and d. All microstructures are characterized by ultrafine ferrite

grains and globular cementite particles. The ferrite grain size hardly changed after the post deformation

annealing (Fig. 61c and d), if compared with those in the microstructures directly after large strain

warm deformation (Fig. 61a and b). Only a slight coarsening of the cementite particles was observed

after annealing. However, much finer cementite particles were found in the 2CMn steel, especially after

annealing, if compared to those of the 2C steel (compare Fig. 61d with c).

64

Fig. 61. TEM micrographs of the 2C (0.22 mass%C-0.74 mass%Mn) and 2CMn (0.23 mass%C-1.52

mass%Mn) steels after large strain warm deformation (ε = 1.6) and annealing, respectively. The arrows

“1” point out the fine cementite particles inside the ferrite grains. The arrows “2” point out the coarse

cementite particles at the ferrite grain boundaries.

(a) and (b) deformed microstructure

(c) and (d) annealed microstructure

Fig. 62. TEM micrographs of the dislocation structures in the 2C steel after large strain warm

deformation (ε = 1.6) and annealing, respectively.

(a) dislocation structure after large strain warm deformation

(b) dislocation structure after annealing

(a) (b)

(c) (d)

(a) (b)

65

Fig. 63. TEM micrographs of the 2C steel after large strain warm deformation and annealing.

(a) mutual crystallographic orientation of the local areas (adjacent grains and subgrains numbered with

1 to 6). The white arrows point out the cementite particles pinning the subgrain boundary

(misorientation θ = 7.5°),(b) dislocation structures and individual cementite particles inside the ferrite

matrix. The black arrows point out the cementite particles pinning the dislocations

Two different size groups of cementite particles are presented in the 2C steel (Fig. 61a and c). The finer

cementite particles (5&90 nm) are distributed inside the ferrite grains (see arrows 1). The planar arrays

of larger cementite particles (90&350 nm) are located at the ferrite grain boundaries (see arrows 2),

acting as obstacles for the migration of ferrite grain boundaries. A lower number of small size particles

were observed in this steel after annealing (Fig. 61c). In contrast to this, in the 2CMn steel,

homogeneously dispersed fine cementite particles (5&120 nm) (not clearly dividable in two size groups)

could be observed (Fig. 61b and d).

Fig. 62a and b show the dislocation structures in the 2C steel after the large strain warm deformation

and annealing, respectively.

In Fig. 63a, the mutual crystallographic misorientations of local areas (adjacent grains or subgrains

numbered with 1 to 6) are determined by means of Kikuchi-line analysis in the TEM. The curvature of

the subgrain boundary (misorientation of 7.5°) demonstrates a pinning effect of the particles on the

ferrite grain boundary. Fig. 63b shows the dislocation structures and individual cementite particles

inside the ferrite matrix. The black arrows point out the cementite particles pinning the dislocations.


Fig. 64 shows the stress-strain curves of the ultrafine grained steels after tensile tests at room

temperature. The details of the tensile properties are shown in Fig. 65. Each dot represents an average

value from three tests. The increase of carbon and manganese contents leads to an increase in strength,

but less difference in ductility. Taking the manganese as an example, with increasing the manganese

content from 0.7% to 1.5%, the lower yield stress increases from 532 MPa to 622 MPa and the ultimate

tensile strength from 595 MPa to 678 MPa, while the total and uniform elongations hardly change, such

as the total elongation is 22.7% and 21.4%, the uniform elongation is 10.2% and 10.0% for the 2C and

2CMn steel, respectively. The Lüders strain decreases obviously with the increase of carbon content, as

shown in Fig. 66. The yield ratio (LYS/UTS) of the experimental steels is shown in Fig. 67. The ratio

firstly increases with increasing carbon content from 0.15% to 0.2%. However, it obviously decreases

when the carbon content reaches 0.3%.

Fig. 68 represents the true stress-strain curves (the part before necking starts) of the ultrafine grained

steels with different carbon contents. A fast increase of the true stress can be observed in the 3C steel.

Fig. 69 shows the data of stress-strain curves (Fig. 64) of the ultrafine grained steels (produced by the

ultrafine grain route in Fig. 37b) in Kocks-Mecking plots. The data were converted to true stress–true

strain and plotted in differential form. The figure displays the relationship between work hardening rate

θw versus true stress σt, both normalized by the shear modulus µ = 80 GPa. The curves shown in Fig. 69

are the part between the Lüders strain and the strain where the necking starts. An obvious feature of

Fig.69 is that the work hardening rate decreases with stress in all curves.

66

This decrease may be approximately described by the straight coarse lines. The value shown in the

figure (i.e. 21.59, 23.25 and 21.69) is the slope of each straight line (〉θw / 〉σt), which displays the

difference in work hardening behavior. On the right up of the figure, the grain sizes of the steels are

represented.

Fig. 64. Comparison of stress-strain curves of the experimental steels with different ferrite grain sizes.

The steels were processed by the large strain deformation and subsequent 2 h annealing at 823 K (Fig.

37b: ultrafine grain route).

Fig. 65. Mechanical properties of the experimental steels after large strain deformation and subsequent

2 h annealing at 823 K (Fig. 37b: ultrafine grain route).

(a) Strength; (b) ductility

(a)

(b)

67

Fig. 66 - Lüders strain of the experimental steels after large strain deformation and subsequent 2 h

annealing at 823 K (Fig. 37b: ultrafine grain route)

Fig. 67. Yield to tensile ratio (LYS/UTS) of the experimental steels after large strain deformation and

subsequent 2 h annealing at 823 K (Fig. 37b: ultrafine grain route). LYS: lower yield stress; UTS:

ultimate tensile stress.

Fig. 68. True stress-strain curves of the ultrafine grained steels with different carbon contents.

68

Fig. 69. Work hardening rate θw versus true stress σt, both normalized by the temperature dependent

shear modulus µ = 80 GPa. dα: average ferrite grain size.

Fig. 70. Charpy impact properties for subsize specimens of the experimental steels after large strain

deformation and subsequent 2 h annealing at 823 K (Fig. 37b: ultrafine grain route).

69

Fig. 71. Specific upper shelf energy (related to the ligament area) (a) and ductile-to-brittle transition

temperature (b) of the experimental steels after large strain deformation and subsequent 2 h annealing at

823 K (Fig. 37b: ultrafine grain route). All values were converted to those for full size Charpy test

specimens according to equation (1) and (2).

Fig. 72. Observation area of Charpy impact fracture surface – in the middle of the thickness (along ND)

and 1 mm away from the V-notch.

(b)

(a)

70

(a)

(b)

(c)

(d)

Fig. 73. Fracture surfaces of the Charpy impact specimens tested at room temperature (293 K) and -40°

(233 K). The impact specimens of the experimental steels were produced by the ultrafine grain route

(Fig. 37b).

(a) 15C steel; (b) 2C steel; (c) 2CMn steel; (d) 3C steel.

Fig. 70 shows the results of Charpy impact testing on subsize specimens of the experimental steels.

With increasing carbon content the Charpy impact curves shift to higher temperatures and the upper

shelf of the notch impact energy decreases. Compared to the results from the conventional route (Fig.

53, the steels with ferrite-pearlite microstructure after the conventional route), the effect of carbon

content in decreasing the upper shelf energy is less pronounced in the ultrafine grained steels. In Fig. 70,

there is only a small difference between the ultrafine grained 2C and 2CMn steels.

71

As described before, it is possible to use the Charpy impact properties of subsize specimen to predict

the ductile-to-brittle transition temperature (DBTT) and the upper shelf energy (USE) of full size

specimen. The correlations of Charpy impact properties between subsize and full size specimens are

shown in equation (1) and (2).

[USE/Bb]Charpy-V = -0.001175 [USE/Bb]2small + 5.029 [USE/Bb]small - 75.07 (1)

DBTTCharpy-V = 1.169 [DBTTsmall] + 61.35 (2)

According to these correlations, the values of the specific upper shelf energy and the ductile-to-brittle

transition temperature calculated for full size specimens are represented in Fig. 71. With increasing

carbon content the specific upper shelf energy decreases and the ductile-to-brittle transition temperature

increases. However, the ductile-to-brittle transition temperature decreases when the carbon content

increased from 0.15% to 0.2%.

The observation area of fracture surface after Charpy impact test was in the center of the sample and 1

mm below the V-notch, as shown in Fig. 72. The reason to select this location as the observation area is

because the stress state during impact testing at this position promotes a more brittle fracture. The

fracture surfaces of the impact specimens tested at 293 K and 233 K are shown in Fig. 73. It can be seen

that the fracture surfaces of all the experimental steels reveal dimples in conformity with a ductile

fracture. No significant difference in the appearance of dimples could be established in the fracture

surfaces of the steels with different compositions.

5.3.6 Comparison of properties between the conventional and ultrafine grained steels

Grain size

Fig. 74 shows the average ferrite grain size produced by the conventional route (CR) (Fig. 37a), and the

ultrafine grain route (UFR) (Fig. 37b), respectively. The UFR is effective to produce a finer

microstructure. The effect is even obvious when the initial microstructure (before large strain warm

deformation) is coarse, such as for the 15C steel.

Tensile properties

The effect of grain refinement on the tensile properties is represented in Fig. 75. It can be seen that,

compared to the steels produced by the conventional route, there are obvious increases in the lower

yield stress, ultimate tensile stress and Lüders strain in the ultrafine grained steels. The ductility of the

ultrafine grained steels decreases but the total elongation is still above 20%.

The details of the tensile properties are shown in Fig. 76. With decreasing average ferrite grain size of

each of the steel, the lower yield stress increases by 40&56% (Fig. 76a), the ultimate tensile stress

increases by 12&19% (Fig. 76b) and the yield to tensile ratio increases by 25&37% (Fig. 76c). The

decrease in grain size has evident effect on increasing the Lüders strain, i.e. 27&73% increment for the

different steels (Fig. 76d).

72

Fig. 74. Comparison of average ferrite grain size of the experimental steels produced by the

conventional route (CR) (Fig. 37a), and the ultrafine grain route (UFR) (Fig. 37b), respectively. Open

symbols: results after the CR; filled symbols: results after the UFR.

Fig. 75. Comparison of stress-strain curves of the experimental steels with different grain sizes. The

different grain sizes were produced by the conventional route (CR) (Fig. 37a), and the ultrafine grain

route (UFR) (Fig. 37b), respectively.


(a) (b)

(c) (d)

73

Fig. 76. Comparison of tensile properties of the experimental steels with different grain sizes. The

different grain sizes were produced by the conventional route (CR) (Fig. 37a), and the ultrafine grain

route (UFR) (Fig. 37b), respectively. Open symbol: results after the CR; filled symbol: results after the

UFR. (a) yield stress, (b) ultimate tensile stress, (c) ratio of lower yield stress to ultimate tensile stress

(LYS/UTS), (d) Lüders strain.

(a) (b)

(c) (d)

74

Fig. 77. Change of Charpy impact properties for subsize specimens of the experimental steels with

different grain sizes. The different grain sizes were produced by the conventional route (CR) (Fig. 37a),

and the ultrafine grain route (UFR) (Fig. 37b), respectively. DBTTsmall: ductile-to-brittle transition

temperature of subsize specimen.


Toughness

The Charpy impact transition curves of the ultrafine grained steel and the coarser grained steel are

represented in Fig. 77. In order to demonstrate the changes of the Charpy impact transition curves with

grain refinement, the 2C steel was chosen as an example since it has the best toughness (the lowest

ductile-to-brittle transition temperature) among the experimental ultrafine grained steels (Fig. 71b).

After impact testing on subsize specimens, the improvement of impact properties in the ultrafine

grained 2C steel is shown in Fig. 77b. A sigmoidal relationship exists between Charpy impact absorbed

energy and temperature. Compared to the coarser microstructure the upper shelf energy is decreased

and the transition curve is more flat in the ultrafine grained steel. In order to determine the ductile-to-

brittle transition temperature, the following criterion was used: the ductile-to-brittle transition

temperature is defined by the temperature at half of the upper shelf energy, Fig. 77b. It can be observed

that the ductile-to-brittle transition temperature of the ultrafine grained steel is lower than that of the

conventional steel. The change of the absorbed energy with temperature close to the ductile-to-brittle

transition temperature is quite small in the steel with finer microstructure.

(d)(c)

(b)(a)

75

Fig. 78. Fracture surfaces, taken by SEM, of the ultrafine and conventional 2C steels after Charpy V-

notch impact tests. A fracture surface of the ultrafine 2C steel after impact testing at 103 K (-170°) (a)

total view of the fracture surface; (b) brittle behavior; (c) border area between cleavage fracture and

ductile fracture; (d) ductile behavior, observation area is shown in (a)

B fracture surface of the conventional 2C steel after impact testing at 143 K (-130°) (e) total view of the

fracture surface; (f) appearance of the cleavage fracture, observation area is shown in (e).

76

Fig. 79. ND orientation maps (taken by EBSD measurement) of the ultrafine grained 2C steel after

Charpy impact testing at 103 K, the same specimen as shown in Fig. 45a but measured in the transverse

direction of the sample. (a) front of a crack: the circles with number “1” and “2” point out two

elongated grains with high-angle grain boundaries in between (b) aligned damage below a ductile

fracture: the ellipse highlights the alignment of microvoids along the grain boundaries

It is important to note that the lower shelf energy of the ultrafine grained steel is much higher than that

of the conventional steel. For example, in Fig. 77b, the absorbed energy is about 2 J for the ultrafine

grained 2C steel when the impact test temperature was 103 K, in contrast, it is only about 0.5 J for the

conventional steel tested at even higher temperature of 143 K. Fig. 78 A and B show the fracture

surfaces of the ultrafine grained and conventional 2C steel tested at 103 K and 143 K, respectively. It

can be seen from Fig. 78a that the fracture surface of the ultrafine grained steel consists of both ductile

and brittle areas and these two areas appear alternatively. Nearly 50% of the surface displays dimpled

fracture. If the ductile-to-brittle transition temperature is defined by the temperature at which 50% of

the fracture surface is covered with a brittle area, the ductile-to-brittle transition temperature of the

ultrafine grained 2C steel becomes lower than 103 K for the subsize specimen. The details of the brittle

and ductile behaviors are shown in Fig. 78b&d. The observation area is in the middle of the thickness

(along ND) and 1 mm away from the V-notch, as shown in Fig. 78a. In contrast to the ultrafine grained

microstructure, in the coarser microstructure of the 2C steel a fully cleavage fracture is shown already at

143 K (Fig. 78e). A large magnification of the cleavage fracture is displayed in Fig. 78f.

(b)

(a)

77

Fig. 80. Converted full size Charpy impact properties. (a) Specific upper shelf energy (related to the

ligament area) and (b) ductile-to-brittle transition temperature of the experimental steels with different

grain sizes. The different grain sizes were produced by the conventional route (CR) (Fig. 37a), and the

ultrafine grain route (UFR) (Fig. 37b), respectively.

Fig. 79 shows the orientation maps of the ultrafine 2C steel after Charpy impact testing at 103 K, the

same specimen as in Fig. 78a, but measured in the transverse direction of the sample. The maps were

taken by EBSD measurements and show orientation in ND. The correspondence between the colors and

the crystal orientations is represented in the stereographic triangle in Fig. 79b. At the tip of the crack,

Fig. 79a, there are two main colors, i.e. red and blue. This indicates that in Fig. 79a (the front of a crack)

the microtexture consists of ND || <111> (in blue) and ND || <001> (in red) texture components. It can

be seen that the crack separates the blue and red grains, which shows that the crack spreads along the

grain boundary of the grains numbered with “1” and “2”. Fig. 79b shows the ND orientation map below

the ductile fracture. Compared to Fig. 79a, however, colors close to green dominate in Fig. 46b due to

the bending during the Charpy test. The aligned damage (highlighted by the ellipse) below the ductile

fracture can be seen clearly.

The values of the Charpy impact properties, calculated for full size specimens, are summarized in Fig.

80 for both the ultrafine grained and conventional steels. It can be seen that the specific upper shelf

energy and the ductile-to-brittle transition temperature decrease in the ultrafine grained steels.

(b)

(a)

78

Summary of mechanical properties

In Fig. 81, the variations in yield stress and ductile-to-brittle transition temperature of all the

experimental steels produced by the ultrafine grain route (Fig. 37b) and the conventional route (Fig. 37a)

are represented together. The symbols in the upper left area of the figure show an optimum combination

of strength and toughness, i.e. higher lower yield stress and lower ductile-to-brittle transition

temperature. Thus, it can be seen that the ultrafine grained steels with dispersed cementite particles have

an improved strength and toughness combination.

Among the experimental steels, the steel with the higher manganese content has the best combination of

strength and toughness, as shown by the dot of number “1” in Fig. 81.

Fig. 81. Lower yield stress and ductile-to-brittle transition temperature of the experimental steels. The

filled symbol represents the results from the steels after ultrafine grain route (UFR) (Fig. 37b). The

open symbol represents the results from the steels processed by the conventional route (CR) (Fig. 37a).

The dot of number “1” indicates the best combination of strength and toughness.

5.3.7 Microstructure and texture evolution during warm deformation and annealing

In order to study the microstructure evolution and the formation of the ultrafine grained microstructure

during the warm deformation, samples were water quenched after different strains, Fig. 37c. Fig. 82

shows the microstructural evolution of the 2C steel during the deformation. The initial ferrite-pearlite

microstructure before the warm deformation is represented in Fig. 82a. It can be seen from Fig. 82 and

Fig. 84a that after the first deformation step (ε = 0.4) the ferrite-pearlite microstructures were finer than

that before the warm deformation and elongated in the rolling direction. Ferrite grain boundaries were

visible, but there existed many faint contours inside ferrite grains after the first deformation step.

79

Fig. 82. Optical microstructures of the 2C steel during warm deformation (Fig. 37c) and annealing (Fig.

37b) at 823 K.

(a) initial microstructure before large strain warm deformation

(b) microstructure after one warm deformation step (ε = 0.4)

(c) microstructure after two warm deformation steps (ε = 0.8)

(d) microstructure after three warm deformation steps (ε = 1.2)

(e) microstructure after steps warm deformation four (ε = 1.6)

(f) microstructure after four deformation steps and 2 h annealing at 823 K

(a) (b)

(c) (d)

(e) (f)

80

Fig. 83. Microstructure evolution of the 2C steel during warm deformation (Fig. 37c) and annealing

(Fig. 37b) at 823 K. Arrow “1” points out the large cementite particle at the ferrite grain boundary.

Arrow “2” points out the small cementite particle in the initial pearlite colony region. Arrow “3” points

out the cementite lamella in the initial pearlite colony region. (a) initial microstructure before large

strain warm deformation, (b) microstructure after two warm deformation steps (ε = 0.8), (c)

microstructure after four warm deformation steps (ε = 1.6),(d) microstructure after four deformation

steps and 2 h annealing at 823 K .

(c) (d)

(a) (b)

81

Fig. 84. Evolution of grain characteristic of the 2C steel during warm deformation (Fig. 37c) and

annealing (Fig. 37b) at 823 K, each deformation step with ε = 0.4 and ε& = 10-1

.

1.6-Ann.: processed by four deformation steps and 2 h annealing at 823 K.

(a) change of average ferrite grain size and grain shape aspect ratio

(b) change of fraction of high-angle grain boundaries (HAGBs)

(a)

(b)

82

Fig. 85. Grain boundary misorientation distribution for the 2C steel during warm deformation (Fig. 37c)

and annealing (Fig. 37b) at 823 K, each deformation step with ε = 0.4 and ε& = 10-1

.

(a) misorientation distribution after two deformation steps (ε = 0.8), (b) misorientation distribution after

two deformation steps (ε = 1.2), (c) misorientation distribution after four deformation steps (ε = 1.6), (d)

misorientation distribution after four deformation steps and 2 h annealing at 823 K

Fig. 86. Total fraction of α-(<110> || RD) fiber and γ-(<111> || ND) fiber for the 2C steel during warm

deformation (Fig. 37c) and annealing (Fig. 37b) at 823 K, each deformation step with ε = 0.4 and ε& =

10-1

. 1.6-Ann.: processed by four deformation steps and 2 h annealing at 823 K.

(a) (b)

(c) (d)

83

(a)

(b)

Fig. 87. Texture evolution for the 2C steel during warm deformation (Fig. 37c) and annealing (Fig. 37b)

at 823 K, each deformation step with ε = 0.4 and ε& = 10-1

.

(a) orientation image: α-(<110> || RD) fiber in blue and γ-(<111> || ND) fiber in red

(b) observed texture in ϕ2 = 45° ODF section

Fig. 88. Sketch of the microstructure evolution during warm deformation.

84

After two deformation steps (ε = 0.8), the microstructure is too fine to be resolved by the optical

micrograph, Fig. 49. Using a higher magnification, as shown by the SEM image in Fig. 83b, it can be

observed that the microstructure consists of ferrite and partially spheroidized cementite. The former

pearlite colonies are elongated and still can be distinguished clearly. The fraction of high-angle grain

boundaries decreases obviously, Fig. 84b.

After three deformation steps (ε = 1.2), the average ferrite grain size decreases slightly and the grain

shape aspect ratio hardly changes, Fig. 84a. On the other hand, the fraction of high-angle grain

boundaries increases, Fig. 84b. The details of the distributions of grain boundary misorientations are

shown in Fig. 85b. It can be seen that the number fraction of the misorientations of less than about 8°

decrease prominently if compared to that after two steps deformation, as shown in Fig. 85a. Meanwhile,

a strong α-(<110> || RD) fiber was formed, Fig. 86.

Most of the pearlite lamellae were spheroidized into cementite particles after four deformation steps (ε

= 1.6), Fig. 83c. The clear alignments of the cementite particles which decorate the ferrite grain

boundaries still can be seen (arrow “1”). Different sizes of cementite particles (arrows “1” and “2”) and

cementite fragments (arrow “3”) were inhomogeneously distributed in the ferrite matrix. Ferrite grain

size decreases and the grain shape becomes more equiaxed with increasing strain, Fig. 84a. The fraction

of high-angle grain boundaries hardly changes, Fig. 84b. The average misorientation angle increases to

27.2°, Fig. 85c, even the fraction of low-angle misorientation (i.e. less than 8°) increases simultaneously.

As shown in Fig. 86, the fraction of the γ-(<111> || ND) fiber increases.

After annealing the samples (processed by four deformation steps ε = 1.6) for 2 h, there is nearly no

change in the grain size. However, the grain shape becomes more equiaxed after annealing, Fig. 84a.

Nearly all cementite fragments are spheroidized into discrete particles and homogeneously distributed

in the ferrite matrix, Fig. 83d. The fraction of high-angle grain boundaries increases after annealing, Fig.

84b and Fig. 85d. After annealing, however, in contrast to interstitial free steels which show a strong γ-

fiber, a strong α-fiber was found, Fig. 86.

Fig. 87a shows the distributions of α- and γ-fiber during warm deformation and subsequent annealing.

The blue color indicates the α-(<110> || RD) fiber. The red color indicates the γ-(<111> || ND) fiber.

The gradient of color represents the angular deviation from a perfect α- or γ-fiber. Taking the α-fiber

for instance, when the color of blue changes to the light blue the orientation direction of <110> changes

from exactly parallel to [100] to with an angle of 15° to [100]. Black lines indicate misorientations θ ≥

15° between adjacent grains. The gray lines indicate misorientations between 2° and 15°. It can be seen

that the area fraction of blue color decreases when the logarithmic strain increases from 1.2 to 1.6.

However, it increases after annealing. This shows the same tendency as shown in Fig. 86, i.e. the α-

fiber first decreases when the strain increases from 1.2 to 1.6 then increases after annealing. The

corresponding texture evolution (ϕ2 = 45° section of ODF (orientation distribution function)) during

warm deformation and subsequent annealing is shown in Fig. 87b.

As shown above, during warm deformation the increase in strain affects the alignment of pearlitic

cementite lamellae or later the alignment of cementite stringers and the anisotropic ferrite grain growth.

The effect of deformation on the microstructure evolution during warm deformation can be simply

displayed by use of Fig. 88. The figure shows schematically the microstructure evolution during the

deformation of the 2C steel processed by the route in Fig. 37c. The initial ferrite-pearlite microstructure

is shown in Fig. 88a. After some deformation, the pro-eutectoid ferrite grains become elongated. The

pearlitic cementite lamellae rotate toward an alignment perpendicular to the compression direction. At

the same time, some pearlitic cementite lamellae disintegrate into short fragments, which decorate the

grain boundaries of pearlitic ferrite, seen in Fig. 88b. After large strain deformation, these fragments

spheroidize into discrete cementite particles. Smaller cementite particles are also observed inside ferrite

grains. Larger cementite particles are distributed at the ferrite grain boundaries and act as obstacles to

grain boundary migration. After a large strain deformation the cementite particles are aligned in rows

perpendicular to the compression direction. This makes grain boundary migration more difficult in the

compression direction, which leads to a more elongated grain shape. Most of the large flat pro-eutectoid

ferrite grains are divided into several subgrains (Fig. 88c).

85

5.4. Discussion

5.4.1. Ultrafine grained steels produced by warm deformation

The material contained a uniform dispersion of fine cementite particles (with an average size below 250

nm) which may contribute to a strong inhibition of grain boundary migration in terms of Zener pining,

thereby effectively preserving a deformation grain shape at temperatures lower than 873 K (Fig. 36).

Higher temperatures above 873 K lead to faster diffusion and Ostwald ripening of the cementite

particles. The classical Zener equation indicates that for a given volume fraction of particles the Zener

back driving force on the grain boundary is increasing proportional to the inverse diameter of particles.

If the cementite particles reach a diameter above 300 nm by ripening, they can be considered as weak

obstacles to the capillary driven motion of the ferrite grain boundaries so that higher temperatures entail

more equiaxed grain shapes of the ferrite (Fig. 36).

In comparison with alternative routes producing an ultrafine single phase ferrite microstructure, the

present results which show an uniform dispersion of cementite particles in a fine ferrite matrix are

advantageous. This is because such microstructures increase the work hardening rate and reduce the

tendency for plastic instabilities. This is important in particular if the grain size of the ferrite is very

small [28,29]. The increase in work hardening can be attributed to an increase in the rate of dislocation

multiplication at particles and to the decrease of the mean free path of dislocations.

Conventional approaches use a quenching and tempering heat treatment to obtain fine carbide particles

distributed homogeneously in the ferritic matrix [30]. This requires, especially for large scale samples, a

certain amount of expensive alloying elements to get the required hardenability. The present study

pursued a large strain warm deformation strategy below the γ → α transformation temperature of a

ferrite plus pearlite or a bainitic microstructure, respectively, to reach the same aim.

The dispersed globular cementite particles appear in all the microstructures due to the effect of the large

strain deformation. Most of the cementite particles are distributed on the ferrite grain boundaries and act

as obstacles to grain boundary migration (Fig. 43). The pronounced alignment of cementite particles

can be explained in terms of their position on the rims of the elongated ferrite grains. This effect also

explains the elongation of the ferrite grains, especially at lower deformation temperatures (Fig. 45 and

47). At higher deformation/annealing temperatures the cementite particles may coarsen significantly,

especially the particles located at grain boundary triple points (Fig. 43).

Quite similar results, such as the fraction of high-angle grain boundaries and grain size, were obtained

from three experimental routes at different cooling rates and temperatures. This makes the experimental

parameters flexible for industrial application. For example, experiment III with its moderate cooling

rate is favorable for rolling mills with a limited cooling rate on the run out table, experiment V with its

relatively high deformation and annealing temperatures is more interesting in cases where the rolling

forces may be the limiting factor. On the other hand, the average ferrite grain size maintained about 1

µm even after 2 h isothermal treatment at 823 K, which implies the obtained ultrafine ferrite and

cementite microstructures are relatively stable under these conditions.

The increase in macrohardness with decreasing temperature (Fig. 49) can be mainly attributed to the

effect of grain refinement. It is known that an addition of carbon and manganese results in a higher

hardness due to solid solution and/or precipitation hardening. On the other hand, this addition can also

refine the microstructure (Fig. 47), which increases hardness according to the Hall-Petch relationship.

The effect of carbon and manganese on grain refinement will be discussed in detail in section 5.4.4.

5.4.2. Formation of the ultrafine grained steels during warm deformation and annealing

Microstructural evolution of ferrite

The microstructural evolution of ferrite in Fig. 82 reveals several details regarding the role of strain

during multi-step warm deformation. After a slight deformation (ε = 0.4), the ferrite grain size

decreases and the grains become elongated along the rolling direction, Fig. 82b and Fig. 84a. The pro-

eutectoid ferrite grain boundaries are visible, but there exists many faint contours inside ferrite grains

which probably imply the fragmentation of ferrite grains due to the deformation.

86

With increasing the logarithmic strain to 0.8 the ferrite grain size decreases rapidly and the grain shape

is more elongate, Fig. 82c and Fig. 84a. Meanwhile, the decrease in the fraction of high-angle grain

boundaries suggests grain subdivision occurs during the deformation (Fig. 84b). The pearlite colonies

stretching in the rolling direction (Fig. 83b) demonstrates the grains rotated during deformation.

A further increase of the strain up to 1.6 leads to only a less refinement of the ferrite grains (Fig. 84a).

There seems to be an effective strain of ~0.8 for refining the microstructure effectively in the present

study. This phenomenon was also observed by Shin and Prangnell [31&33] in a low carbon and an Al-

Mg alloy, respectively, processed by severe plastic deformation. As reported by Tsuji [34] grain

subdivision is of essential importance in ultra-grain refinement especially for an initial coarse

microstructure. Large strain deformation could result in finer grain subdivision due to a large amount of

dislocations produced. The appearances of the more equiaxed grain structure and the increase in the

fraction of high-angle grain boundaries after the effective strain (Fig. 84 and 85b&d) can be attributed to

recovery (especially polygonization), called also continuous recrystallization or recrystallization in situ.

It is different from grain structures formed by primary recrystallization because the ferrite grains were

elongated (i.e. the grain shape aspect ratio is about 2.8 after the deformation of ε = 1.6) and dislocations

could be observed inside the grains (Fig. 65a). The reasons why only recovery but not primary

recrystallization takes place will be described in section 4.2.4.3. Recovery transforms the crystals

subdivided by deformation-induced boundaries into ultrafine grains surrounded by clear grain

boundaries (polycrystalline “grains”). Therefore, further increasing strain over the effective value is

beneficial for grain shape adjustment and high-angle grain boundaries formation. It can be concluded

that the ultra-grain subdivision is the essential process for ultrafine grained microstructure formation

and recovery is also necessary to form clear ultrafine grain boundaries.

After 2 h annealing at 823 K, the fraction of high-angle grain boundaries slightly increases, Fig. 84b

and 86d. As shown by other investigations [35,36], recovery clearly contributes to the formation of new

high-angle grain boundaries. The possible mechanisms are the accumulation of geometrically necessary

dislocations in the subgrain boundaries [37], the increase of misorientation angle by the merging of

lower-angle grain boundaries during subgrain coalescence [35,36] and the subgrain growth with

migration of low-angle grain boundaries via dislocation motion. From the results in this study it is not

possible to prove exactly which mechanism is really active or relevant for the microstructure

development in the present investigation. Nevertheless, some of the interesting features found in the

microstructures will be discussed to illustrate the possible mechanisms of recovery in the steels. As

shown in Fig. 84a, the ferrite grain size hardly changes but the grain shape is more equiaxed (i.e. the

grain shape aspect ratio is about 1.6) after annealing. The mechanism of this phenomenon can be

attributed to the accumulation of dislocations into subgrain boundaries in RD and/or subgrain growth in

ND.

The possibility of subgrain growth is shown in Fig. 63a. The curvature of the subgrain boundaries

(misorientation θ = 7.5°) indicates the tendency to grow of the subgrain “1”. As reported in [38], low-

angle grain boundaries in general have a low mobility. The reason is the difference in structure between

high- and low-angle grain boundaries. For high-angle grain boundaries, migration can occur easily by

single atom jumping across the rather open grain boundary structure. Low-angle grain boundaries

composed of dislocation arrays migrate by the motion of dislocations, i.e. a coordinated movement of

atoms. While the motion of an individual dislocation by glide is easy, the motion of an array might be

more difficult, since the motion of the arrays causes a change in the shape of two subgrains, which is

resisted by surrounding subgrains. As reported in [38], the limitative motion of the array is the climb of

edge dislocations and the activation energy of the process has the order of magnitude of the activation

energy for self-diffusion. On the other hand, even a rather low strain leads to a decrease of the self-

diffusion activation energy. In the present work, the large strain deformation (accumulative strain of 1.6)

facilitates the subgrain growth during annealing.

Spheroidization of lamellar pearlite

According to the work of Chattopadhyay and Sellars an excess of vacancy which formed during the

deformation promotes carbon diffusion [39]. After large strain deformation lamellae kinks are

characteristic for pearlite [40,41]. An important factor for the acceleration of the spheroidization

process can be a local difference between the equilibrium carbon concentrations in ferrite near the

surface of a deformed lamella with different curvature radii. The equilibrium carbon concentration in

87

ferrite in the vicinity of the lamella with a small curvature radius is higher than that of a larger one.

After a large strain deformation of pearlite the numerous kinks of the lamellae occur with small radii,

therefore, the carbon concentration in ferrite near kinks is essential higher than that close to the flat

parts of the lamellae. According to the Gibbs-Thompson equation and together with the high defect

density, the carbon diffusion leads to a rapid dissolution of lamellae kinks and a simultaneous

deposition of carbon on the flat cementite lamella.

It was reported that the fracture of lamellae in the pearlite colonies and lamellae oriented perpendicular

to the rolling direction can also accelerate the spheroidization. Fragments of the former lamellae located

at a prior austenite grain boundary can easily be formed because of an accelerated diffusion along the

boundary, which leads to a faster coarsening of these cementite fragments, which can be seen from

arrow “1” in Fig. 83c.

For the case of a large strain deformation the substructure of the pearlite lamellae can also exhibit an

essential effect on the spheroidization process. The interface adjacent to the sub-boundary in the

cementite lamella with a large local curvature and the surrounding ferrite provokes a quick carbon

dissolution that leads to a local lamella division.

The appearance of a typical former pearlitic cementite lamellae is shown by arrow “3” in Fig. 83c after

a large strain deformation (ε = 1.6). It indicates that the spheroidization is not fully finished after the

large strain warm deformation. The small spheroidized cementite particles are distributed in the initial

pearlite colony (arrow “2” in Fig. 83c) and the clear alignments of the cementite particles which

decorate the ferrite grain boundaries (arrow “1”) indicate that the large strain warm deformation without

the subsequent annealing is insufficient for a homogeneous distribution of cementite particles.

Distribution of cementite particles

Apart from the processes of spheroidization and coarsening of cementite, which are typical for eutectoid

steels [42], the process of a homogeneous distribution of cementite after the spheroidization can be

observed in the present ferrite-pearlite steel after annealing, Fig. 54 and 50d. This means that after a

large strain warm deformation with a subsequent annealing cementite particles can also be found even

within the former pro-eutectoid ferrite regions. During deformation/annealing, the fine cementite

particles may dissolve and some carbon atoms are assumed to diffuse from the areas of the former

pearlite colonies to the cementite free areas of the former pro-eutectoid ferrite followed by a subsequent

reprecipitation and coarsening. As a result the cementite particles are distributed rather homogeneously

in the ferritic matrix, Fig. 54 and 83d. Thus, it is not possible any more to distinguish between the initial

pearlite and pro-eutectoid ferrite regions, as shown in Fig. 88a.

The main condition for the redistribution of cementite, as shown in [43], seems to be a high dislocation

density in the heavily deformed initial pearlite colonies, which was estimated to be the order of 1016

m-2

for a low carbon steel. In case of a plain low carbon steel [43] with low recrystallization temperature,

recrystallization would slow down or even stop the redistribution of cementite by significantly reducing

the dislocation density. As a result, some pearlite colonies may still exist even after annealing [43].

However, a microalloyed low carbon steel with higher recrystallization temperature offers the

possibility that only recovery takes place even at fairly high temperatures. In this situation the

dislocation density will not be reduced very much, so that the redistribution of cementite can take place

rapidly assisted by a fast pipe diffusion along the dislocation cores.

The driving force for the cementite redistribution can be a gradient of the solute carbon. Inside the

initial pearlite colony around the cementite particles the carbon concentration is essentially higher

compared to that in the pro-eutectoid ferrite. The high density of dislocations and vacancies during or

after large strain warm deformation facilitates the solute carbon diffusion to the areas with lower carbon

concentration, i.e. pro-eutectoid ferrite, and subsequent reprecipitation in the most energetically

favorable places such as triple points of ferrite grain boundaries or dislocation nodes. Additionally, due

to a fast grain boundary diffusion, the particles located at the grain boundaries and triple junctions will

have a size advantage in the later Ostwald-ripening process.

88

Effect of cementite particles

(a) Grain size

Just after spheroidization the size of cementite particles can be below 100 nm. On one hand, these fine

cementite particles can increase the rate of dislocation generation and develop surrounding deformation

zones in the matrix that contain large local misorientation gradients [44]. New high-angle grain

boundaries can form within such deformation zones, therefore, a finer ferrite microstructure can be

produced. On the other hand, the presence of these fine cementite particles retards the softening of

ferrite in the areas of the former pearlite colonies (arrow “2” in Fig. 83c). The formation and growth of

ferrite subgrains are controlled by solute carbon and cementite particles which grow by diffusion of

carbon. The spacing of particles determines the size of the subgrains [43].

(b) High temperature stability

Once an ultrafine microstructure is formed, an average grain size of 1.3 µm was found to be stable in

the 2C steel after 2 h annealing at 823 K (Fig. 84a). This is due to the homogeneously distributed

cementite particles which are fairly stable even after the deformation/annealing at this temperature. The

fine particles lead to a high dragging force for the migration of grain boundaries due to Zener pinning.

(c) Inhibition of primary recrystallization

As shown in Fig. 65 and 63b, a large number of dislocations exist in both deformed and annealed

specimens for the 2C and 2CMn steels. This demonstrates recrystallization does not occur after the

deformation/annealing processing. The reasons that only recovery and not primary recrystallization

takes place here may be described as follows. Due to the spheroidization of pearlitic cementite lamellae

during deformation, a fine dispersion of cementite particles is present in the microstructure after the

large strain deformation. The fine particles lead to a high dragging force for the migration of grain

boundaries due to Zener pinning. On the other hand, it is known that due to the high stacking fault

energy the recovery in ferrite can proceed very fast. In this case the dislocation-rearrangement to form

energetically more favorable configurations starts everywhere, but the subsequent migration of high-

angle grain boundaries may be stopped very soon by the fine particles. The recovery process decreases

both the stored energy and the local stored energy gradient, which decreases the probability of the

successful nucleation of a recrystallized grain or slows down its growth.

The fine particles stabilize the ultrafine microstructure when it has formed. Further increasing the

annealing temperature to even higher temperature (i.e. 973 K) will speed up the diffusion of carbon

atoms. As shown in [36], for two-phase alloys the decomposition before recrystallization is controlled

by the lower activation energy for diffusion of interstitial atoms (carbon). The processes of

spheroidization and distribution of cementite might be assumed to be decomposition, since they can be

attributed to a local change of chemical composition, i.e. the change of carbon content in solid solution

during lamellae dissolution, cementite re-precipitation, as well as dissolution of the fine particles with a

coarsening of larger ones. The coarsening of the cementite particles leads to a reduction of the Zener

drag effect, so that primary recrystallization and grain growth can occur at high annealing temperatures

(Fig. 58). But the homogeneously distributed relatively fine cementite particles can still produce a

stabilizing effect on the fine grained ferrite matrix (Fig. 58). This confirms that the fine cementite

particles can stabilize the ultrafine microstructure as well as inhibit primary recrystallization.

5.4.3. Texture evolution during annealing of a 0.2%C steel

After large strain deformation the texture of the samples consists of the α-(<110> || RD) and γ-(<111> ||

ND) fibers typical for rolled ferritic steels. Annealing the samples after large strain deformation,

however, in contrast to interstitial free steels which show a strong γ-fiber, a strong α-fiber was found,

Fig. 86 and 87. The microstructure of the annealed material consists of still elongated ferrite grains with

high dislocation densities indicating that no recrystallization had occurred. The texture and

microstructure observations may be explained as follows: (i) The occurrence of recrystallization is

inhibited by cementite particles inhibiting either the nucleation or the growth of the recrystallized grains.

(ii) Due to the higher stored energy of γ-fiber grains as compared to α-fiber grains the latter slowly

consume their γ-fiber neighbors, thereby keeping their internal subgrain structure. The process

89

occurring during annealing may be addressed as continuous recrystallization, proceeding without

nucleation of new grains.

5.4.4. Effect of carbon and manganese on grain refinement

As shown above, a finer ferrite microstructure may be observed in the high carbon especially high

manganese steel after the same processing. The different size of ferrite grains and, as described below,

mechanical properties in the steels investigated can be related to the difference in the chemical

compositions. Therefore, the effect of carbon and manganese on the microstructure refinement will be

discussed.

Carbon

In Fig. 51 and 50, the increase in carbon content leads to a finer ferrite-pearlite microstructure after the

conventional route in Fig. 37a. This is due to the formation of ferrite and pearlite is retarded by

increasing carbon content, as shown by the DCCT diagrams in Fig. 39. The lower transformation

temperature leads to a finer transformation microstructure. Meanwhile, higher carbon content also

results in a higher volume fraction of pearlite in the transformed ferrite-pearlite microstructure.

When the experimental steels are processed by the ultrafine grain route (Fig. 37b), the increase in

carbon content also leads to a finer ferrite microstructure, a more equiaxed grain shape and a more

homogeneously distribution of cementite particles, Fig. 54 and 55.

One reason for the finer ferrite microstructure in the higher carbon steel can be the finer initial

transformed ferrite-pearlite microstructure. The other reason can be attributed to the higher volume

fraction of spheroidized cementite particles in the higher carbon steel. Since there exists an effective

strain to refine the initial especially coarse microstructure effectively, the second case of a large number

of particles seems to be more important in the study. During the deformation, pearlite lamellae

spheroidized into cementite particles which located at the most energetically favorable places such as

grain boundaries. As mentioned above, fine particles are effective to act as obstacles for grain boundary

migration during the large strain deformation and the subsequent annealing. This can, on one hand,

refine the microstructure by slowing down the grain growth rate. On the other hand, the primary

recrystallization process can be inhibited by particles, which leads to an increase in recrystallization

temperature. A higher dislocation density and a finer substructure can be saved in the higher carbon

steel which can facilitate the formation of subgrains or even polycrystalline “grains” by the

accumulation of dislocations at the subgrain boundaries during deformation/annealing.

A higher carbon concentration in steels led to a higher volume fraction of pearlite after phase

transformation. As a result, a homogeneous distribution of cementite particles was observed after the

whole processing (the ultrafine grain route, Fig. 37b) in the higher carbon steel. Meanwhile, the

homogeneous distribution means the cementite particles are located not only at grain boundaries but

also in the grain interior. During the large strain deformation, the ferrite grains were elongated along the

rolling direction. A large quantity of dislocations and substructures may exist in one elongated ferrite

grain. The cementite particles in the grain interior develop surrounding deformation zones in the matrix

that contain large local misorientation gradients. New high-angle grain boundaries can form within such

deformation zones, therefore, the grain shape is more equiaxed in the high carbon steel due to the effect

of particles (inside the grain) on the subdivision of an elongated grain.

Manganese

It is known that the transformation processes and the resulting microstructures in steel are sensitive to

the presence of further substitutional alloy elements. The addition of manganese retards the formation

of ferrite and pearlite significantly (Fig. 39). Thus, a finer transformed microstructure was found in the

higher manganese steel (Fig. 50 and 51). After the large strain deformation and subsequent annealing,

the higher manganese steel shows a prominent microstructure characteristic, i.e. ultrafine ferrite with an

average grain size of 0.8 µm (Fig. 55a), nearly equiaxed grain shape (grain shape aspect ratio of 1.1)

(Fig. 55b), a large fraction of high-angle grain boundaries (Fig. 56 and 57), and a homogeneous

distribution of cementite particles (Fig. 54c). Therefore, the effect of manganese on grain refinement

will be discussed in detail by comparison the results from different manganese contents of the 2C and

2CMn steels.

90

The TEM images, Fig. 61, represent the same microstructural components in the steels with different

manganese contents after large strain deformation and annealing processing: namely ultrafine grained

ferrite and globular cementite. Compared to the 2C steel, Fig. 61a and c, the ferrite grain size is smaller

and the grain shape is more equiaxed in the higher manganese 2CMn steel, Fig. 61b and d. The effect of

manganese on these microstructural characteristics is supposed to be manifold during the processing. At

equilibrium, manganese will partition differently between cementite and ferrite. As a carbide-forming

element, it will be more concentrated in the cementite [45]. However, during the formation of pearlite,

manganese will not participate in cementite very much since the transformation time is too short for a

significant diffusion of the substitutional solution of Mn. After the formation of pearlite, manganese

diffusion occurs during the warm deformation which will be enhanced by the deformation because of a

high dislocation density as well as an excess vacancy concentration produced by dislocation interaction

like cutting. During and after the restoration process, which may partially proceed dynamically, the

introduction of high-angle grain boundaries into the microstructure additionally facilitates the

manganese diffusion. Therefore, a cohenite Fe3C type phase (Fe3-XMnXC) [46] can easily be stabilized

by further substituting Fe by Mn in its lattice.

During the large strain warm deformation itself and in the course of the subsequent annealing, the

smaller particles of Fe3-XMnXC may go into solution and the larger particles grow at the expense of the

smaller ones. The Gibbs-Thomson equation [47] shows that smaller particles have a higher

concentration of solute atoms in the matrix at the interface boundary than bigger ones, so that a

redistribution of manganese occurs between smaller and larger particles. The finer particles, detected in

the 2CMn steel at the end of the processing, may be attributed to a sluggish dissolution [48] as well as

to a slow Ostwald ripening of the manganese richer, more stable cohenite particles (Fe3-XMnXC) in this

steel. Both processes are controlled by the substitutional diffusion of manganese. In the 2C steel, with a

lower manganese content, larger cementite particles were observed, Fig. 61a and c. Additional to this,

the coarse particles are mainly located in a few layers. Such “banding” phenomenon is supported by

higher diffusibility [49], which is more adequate in the 2C steel but not sufficient in the 2CMn steel.

However, the particles still seem to be fine enough to pin the ferrite grain boundaries during the long

time annealing even in the lower Mn steel (compare a and c in Fig. 61). Therefore, the ultrafine grained

ferrite microstructure produced is relatively stable against a rapid grain coarsening in the condition of

annealing.

In Fig. 63a, the curvature of the subgrain boundary (misorientation 7.5°) demonstrates a substantial

evidence that fine particles (pointed out by white arrows) may exert a strong Zener drag, even on

subgrain boundaries. This leads to the assumption that only recovery can be considered as the decisive

process responsible for the restoration of the severely deformed ferrite-pearlite structure in the steels

studied. A further subgrain growth process or even a recrystallization seems to be inhibited by a high

pinning force due to Zener drag.

The formation of ultrafine grained structure that occurs under large strain warm deformation is

frequently accompanied by the development of high internal stresses [50-52]. In addition to the high

dislocation density evolved, the grain boundaries due to their non-equilibrium character can

supplementary produce long-range stress fields inside the ultrafine grains. The TEM examination

indicated that the dislocation density evolved inside the ultrafine grained 2CMn steel is higher than that

in the 2C steel (Fig. 63b and Fig. 62b). The higher dislocation density in the manganese richer steel is

considered to be the result of the interaction of dislocations with the finer and more numerous cementite

particles observed (black arrows in Fig. 63b).

The large amount of dislocations inside the grains (Fig. 63b and 62b) demonstrates an incomplete

recovery process even after the long time annealing. Compared to the microstructure found just after

large strain warm deformation, the annealing did not result in a remarkable grain growth (Fig. 61), or a

significant dislocation density decrease inside the fine grains in the 2C and 2CMn steels (Fig. 63b and

63). This large amount of dislocations evolved in the fine grain interiors at the end of the processing is

rather contradictory to the state after a classical recrystallization mechanism. Obviously, another

restoration process must proceed to produce a high fraction of high-angle grain boundaries under the

condition of large strain warm deformation. The deformed substructure including subgrains and high

dislocation densities that developed in the fine-grain interiors can be essentially stable against annealing

as compared to the recovery of the fine-grain boundaries. Therefore, the static recovery (as relaxation

process) can be sub-classified into two sub-processes according to their kinetics. The first one leads to a

91

rapid restoration of the non-equilibrium grain boundaries, and the second one results in the recovery of

the deformed substructure. In the present study of ultrafine grained steels, the process develops more

readily at grain boundaries as compared to grain interiors due to a large number of grain boundaries,

leading to a more rapid release of lattice distortions associated with the non-equilibrium state of

ultrafine grain boundaries. However, in coarse grained steel it is more preferred that the recovery occurs

in the grain interiors [53] due to a small volume fraction of areas near the grain boundary.

It can also be referred to a study, carried out on the single phase IF steel [54], which showed that with

increasing strain (up to 5.6) the severe deformation, assisted by recovery, leads to a more and more

refined substructure with increasing misorientations of the developing subgrain boundaries. This results

in the formation of an increasing number of high-angle grain boundaries. In the current investigation a

similar microstructure development was found at a significantly lower strain (ε = 1.6). It seems that the

presence of fine particles supports this process. The reason for a finer microstructure (Fig. 54, 55a, 61b

and d) and a higher number of high-angle grain boundaries in the higher manganese steel (Fig. 56, 57)

can be attributed to the controlling effect of the Fe3-XMnXC particles. During the large strain warm

deformation the evolving substructure will be finer in the 2CMn steel due to the finer dispersion of the

particles and their impending effect on the displacement of subgrain boundaries. In accordance to this,

the finer average ferrite grain size and the higher fraction of high-angle grain boundaries were obtained

after large strain warm deformation/annealing in the manganese richer steel 2CMn with finer dispersed

Fe3-XMnXC particles.

5.4.5. Mechanical properties of the ultrafine grained steels

Tensile properties

In Fig. 75 typical tensile stress--strain curves are shown for the ultrafine grained steels and the

conventional steels with different ferrite grain sizes. A summary of the data obtained from the tensile

tests is shown in Fig. 76 where the lower yield stress, ultimate tensile stress, yield ratio and Lüders

strain are plotted against carbon content. It can be seen that grain refinement has led to a significant

increase in strength, yield ratio and Lüders strain.

(a) Strength

As similar to other ultrafine grained materials, it is evident that large enhancements in strength

especially in yield stress were observed in the ultrafine grained plain C-Mn steels after tensile tests at

room temperature. Numerous investigations have shown that the yield stress of unalloyed low-carbon

steel increases in inverse proportion to the square root of the ferrite grain size, that is the Hall-Petch

relationship.

However, it should be noted that the ultrafine grained steels, processed by the large strain deformation

and annealing (Fig. 37b), exhibit dislocations and substructures inside the ultrafine grains (Fig. 62 and

63) as well as an elongated grain morphology (Fig. 55b). This means that the increase in strength of

these ultrafine grained specimens might be affected not only by grain refinement strengthening but also

by work hardening and texture strengthening.

If only considering the ultrafine grained steels (with the same work hardening and texture strengthening)

processed by the ultrafine grain route (Fig. 37b) the increase in carbon and especially manganese

content results in the increase in strength (Fig. 65a). However, it was reported by Santos and Bruzszek

[55] that the variation in volume fractions of secondary phases does not affect the tensile properties to

the same degree as the variation in ferrite grain size. Thus, ferrite grain refinement is the main factor

increasing the strength in the investigated steels.

(b) Ductility

One of the more unusual aspects of the ultrafine grained steels is the nature of the stress-strain curves.

Typical examples (Fig. 75) indicate a reduced work hardening in the ultrafine grained steels. This is

reflected in the yield ratio (lower yield stress/ultimate tensile stress), which is between 0.88 and 0.90

(Fig. 76c), compared with 0.7 for the conventional steels. The reason for the decrease in tensile ductility

for the ultrafine grained steels can be explained as follows:

92

Firstly, dynamic recovery as a softening mechanism is able to reduce the apparent work hardening rate.

During tensile deformation, dislocations causing intragranular strain are trapped at grain boundaries.

Then, the kinetics of dynamic recovery is associated with that of the spreading of trapped lattice

dislocations into grain boundaries [56&58]. The change of the dislocation density during dynamic

recovery by the trapped lattice dislocations spreading into grain boundary was studied by K.-T Park and

Y.-S. Kim [59]. It showed that dislocations which contribute to the intragranular strain are trapped at

grain boundaries have enough time to spread into grain boundaries during deformation. This leads to no

significant accumulation of dislocations and, consequently, less work hardening. As mentioned before,

it is assumed that there are two kinds of recovery mechanisms: a slow recovery in the grain interior and

a much fast recovery in the vicinity of grain boundaries. In coarse grained materials the later one was

not clearly observed due to a lower volume fraction of area near grain boundaries. However, in ultrafine

grained materials the faster recovery near the grain boundaries seems to be significant.

Secondly, the decrease in tensile ductility can be simply explained in terms of plastic instability, which

leads to necking due to excessive localized deformation. The condition of necking in tensile test is

predicted by the Considère's criterion [60], σt = dσt/dεt. When the slope of the true-stress true-strain

curve (work hardening rate), dσt/dεt, is equal to the true stress σt, necking begins. As mentioned above,

ultra grain refinement greatly increases the flow stress of the materials, especially at the early stage of

plastic deformation. On the other hand, the present work also clarified that work hardening after

yielding becomes difficult to occur with decreasing grain size. As a result, a plastic instability (necking)

occurs at a very early stage during the tensile test in ultrafine grained materials, as can be easily

understood from the Considère's criterion, which results in a limited uniform elongation.

The yield ratio is high in the ultrafine grained steels. However, it is interesting to note that apart from

the high strength an acceptable ductility can still be found in the ultrafine grained steels investigated,

which is reflected in the total elongation of over 20% and the uniform elongation of about 10% (Fig.

65b). This is quite different from the results reported by other investigators, i.e. the total elongation is

no more than 10%. This can be attributed to the additional second phase – cementite particles offering

the opportunity to improve the work hardening capacity. The reason for the improvement in the work

hardening rate in the present ultrafine grained steels will be explained in section 5.4.6.

(c) Lüders strain

In Fig. 76d it can be seen that the Lüders strain increases dramatically in the ultrafine grained steels.

This might be, before tensile testing, the solute atmospheres already formed around dislocations in the

ultrafine grained steels after 2 h annealing (strain aging), as shown in Fig. 37b. Nevertheless, this can

not fully explain why the Lüders strain of the ultrafine grained steels is far in excess of that in the

conventional steels. Previous work shows that grain refinement is effective in increasing Lüders strain.

The Lüders strain can be plotted against the inverse square root of the grain size [61], which is similar

to the Hall-Petch type relation.

Large Lüders strain has been noted previously by Lloyd and Morris [62] in a fine grained (1&3 µm) Al-

6%Ni alloy that contained little magnesium in solution. They observed that ultrafine grain sizes

promoted yield drops and low work hardening. In the work of Hayes and Keyte [63] an inverse

relationship between the Lüders strain and grain size was proposed in an Al-3%Mg alloy. Pronounced

yield drops and very large Lüders strain thus appear to be a characteristic of ultrafine alloys. These

effects can be linked to a low density of mobile dislocations, dislocation sources within the grains, and

the low work hardening rate for ultrafine grained alloys. These result in yielding taking place by a very

rapid increase in dislocation density in a narrow band initiated at some local stress concentration. The

high yield stress combined with the low work hardening means that a high strain level must be reached

in the initial Lüders band before the local flow stress increases sufficiently for it to start to propagate

along the gage length, and this leads to a large Lüders strain in fine grained steels.

93

(d) Toughness

The microstructure evaluation and the tensile testing show that the increase in carbon and manganese

contents and/or the accompanying decrease in grain size improve the strength properties. Nevertheless,

different effects on toughness can be expected from the grain size and carbon content.

Smaller grain size leads commonly to a lower ductile-to-brittle transition temperature, as shown in Fig.

80b. This can be simply understood from the standpoint of cleavage crack initiation and propagation. It

is generally known that the grain size is one of the major factors determining the cleavage fracture unit

[64,65]. The cleavage fracture unit is reduced as the grain size decreases. When the cleavage crack

propagates across a field of grains, both of the crack-tip dislocation emission and the formation of the

cleavage facets are interrupted by the grain boundaries. If the cleavage cracking across a high-angle

grain boundary the crack front must be geometrically necessarily branched, which, together with the

separation of the grain boundary between the break-through points, results in the additional fracture

work. This toughening effect can lower the ductile-to-brittle transition temperature considerably and has

been noticed for decades in experiments where the behavior of the grain-sized microcracks was

dominant [66&68].

For example, when the cleavage crack paths are carefully observed at high magnifications, it can be

seen that the paths were changed by high-angle grain boundaries. Furthermore, it was observed that

cleavage cracks can also be arrested by high-angle grain boundaries [69]. Thus, the decrease in grain

size (increasing the volume fraction of grain boundaries) can limit the propagation of the initiated

cleavage crack and raise the fracture toughness in the transition region.

Fig. 89 shows schematically why grain refinement is beneficial to decrease the ductile-to-brittle

transition temperature. The ductile-to-brittle transition of a bcc alloy is accompanied by a change in

fracture mechanism from void coalescence to cleavage, and so the ductile-to-brittle transition

temperature can be quantitatively interpreted as a result of the competition between the flow stress and

cleavage fracture stress of the alloy [70]. In a simplified model the temperature, at which the yield stress

σy is equal to the cleavage fracture stress *

fσ , is the ductile-to-brittle transition temperature. According

to the Hall-Petch relation, grain refinement contributes to an increase in the yield stress (σy1 → σy2).

Simultaneously, the grain refinement is more effective in increasing the cleavage fracture stress (*

1fσ →

*

2fσ ). For example, a grain refinement of a steel with 0.1%C can lead to an increase in the cleavage

fracture stress four times more than the increase in the yield stress [71]. Due to a more significant

increase in the *

fσ than the σy, therefore, the ductile-to-brittle transition temperature is lowered by

grain refinement (T1 → T2).

Grain refinement leads to an improvement in toughness. A higher carbon content, however, has an

adverse effect on toughness. For example, according to [72], for medium carbon ferrite-pearlite steel

increasing carbon content from 0.1% to 0.2% leads to an increase of the ductile-to-brittle transition

temperature by about 25 K. However, in the current investigation, the ductile-to-brittle transition

temperature is obviously decreased in the ultrafine grained steels even if the carbon content increases

from 0.15% to 0.2%, Fig. 71b.

As mentioned above, the ductile-to-brittle transition temperature will generally be shifted to a lower

temperature if the cleavage fracture stress is increased or the flow stress is decreased. Since the flow

stress of the 15C steel is lower than that of the 2C steel (Fig. 65a), the decrease in ductile-to-brittle

transition temperature, can only be explained by increasing the cleavage fracture stress in the 2C steel.

Nevertheless, it is known that for steels without changing in grain size, the ductile-to-brittle transition

temperature increases with increasing carbon content. This is attributed to the precipitation of coarse

rod type and spherical type cementite particles, which are considered as initiation sites for damage

[73,74] and are major detrimental microstructural features in steels [70]. This is due to a low adhesion

between the particles and matrix and to differences in their elastic and plastic properties [75].

Therefore, the effect of grain size on the ductile-to-brittle transition temperature must be taken in

consideration. For the occurrence of cleavage fracture, the free distance of two neighbor high-angle

grain boundaries (Griffith type crack length) is assumed to be important. This distance decreases with

decreasing grain size in the ultrafine grained steels with higher carbon content and, therefore, the

cleavage fracture stress will actually rise. Therefore, the decrease in grain size reveals a surprisingly

strong effect on improving toughness even in the case of a higher carbon content.

94

The lower ductile-to-brittle transition temperature observed in the steel with higher carbon content can

be attributed to the grain refinement function by increasing carbon content.

Fig. 89. Influence of grain size on the cleavage fracture stress *

fσ and the yield stress σy of notched

impact test specimens. The subscript of number “1” indicates the results from the steel with fine grains

(solid lines). The subscript of number “2” indicates the results from the steel with coarse grains (dashed

lines). After [75]

(e) Shelf energy

The ductile-to-brittle transition in steels is associated with two different failure mechanisms. At high

temperature in the upper shelf toughness range, fracture occurs by nucleation and coalescence of

microvoids that produces ductile tearing. This process requires extensive plastic deformation and,

therefore, large amounts of energy. At low temperature, fracture occurs by cleavage which is the sudden

separation of atomic planes across the specimen.

In Fig. 77, the lower shelf energy in the ultrafine grained steels is significantly higher than that in the

coarse grained steels. This means the ultrafine grained steels show improved toughness and partially

ductile behavior even at very low temperatures (i.e. at 103 K). Another reason for this phenomenon

might be anisotropy of toughness properties in the ultrafine grained steels due to the large strain

deformation.

After the large strain deformation, a strong alignment of the microstructure was formed along the

rolling direction. Thus, the spread of the crack (below the V-notch) was retarded in the transverse

direction by the microstructure ligaments, which can be seen from Fig. 78a and 79a that the crack

spread also in the rolling direction. In Fig. 79a, the crack path is between two elongated grains with

high-angle grain boundaries in between (i.e. grain “1” and grain “2”), which demonstrates that the grain

boundary is a favorable path for crack propagation. This may be due to the alignment of cementite

particles at the grain boundaries, as shown in Fig. 79b, where the alignment of microvoids along the

grain boundaries is obvious.

On the other hand, the distribution of sulphides, which exist as manganese sulphides in most structural

steels, is different in the rolling, transverse and through thickness directions [76]. After the large strain

deformation, a higher notch impact energy is expected in the rolling direction due to a lower fraction of

interface between matrix and inclusions on the crack path (below V-notch and along TD). Due to the

very low sulphur content (0.004 mass%) in all the steels investigated, there should be no significant

effect of manganese sulphide on the isotropy.

In Fig. 80a, the specific upper shelf energy is higher in the coarse grained 15C, 2C and 2CMn steels.

This might be attributed to a higher ductility in the coarse grained steels, which is also reflected in the

larger area below the stress-strain curve in these steels (Fig. 75).

95

Evidence for optimization of strength and toughness in the ultrafine grained steels

The chemical composition, heat treatment and deformation have a complex effect on the mechanical

properties. Knowledge of the influence of the microstructure on strength-toughness properties makes it

possible to optimize the properties of the steel. In Fig. 81, the variations in yield stress and ductile-to-

brittle transition temperature of all the experimental steels produced by the ultrafine grain route and the

conventional route are represented together. The spots in the upper left area of the figure show an

optimum combination of strength and toughness, i.e. higher lower yield stress and lower ductile-to-

brittle transition temperature. Thus, it can be seen that the ultrafine grained steels with dispersed

cementite particles have a significant improvement in both strength and toughness.

Among the experimental steels, the ultrafine grained steel with the higher manganese content has the

best combination of strength and toughness, as shown by the spot of number “1” in Fig. 81. The

difference in the sizes of ferrite grains and cementite particles, as well as the mechanical properties in

the investigated steels can be related to the difference in the manganese content. In general, manganese

is a strong austenite stabilizer, depressing the A3 temperature. It is essential to cause a microstructure

refinement after the transformation. This is in accordance with the result in Fig. 50c that the finest

microstructure was observed in the 2CMn steel after transformation. As a carbide-forming element,

manganese will be more concentrated in the cementite after the large strain warm deformation. A

cohenite Fe3C type phase (Fe3-XMnXC) can be stabilized by further substituting Fe by Mn in its lattice.

During the large strain deformation itself and in the course of the subsequent annealing, the finer

cohenite particles, due to a sluggish dissolution as well as a slow Ostwald ripening, are more effective

to pin the ferrite grain boundaries against grain growth. On the other hand, the fine particles are also

beneficial to store the geometrically necessary dislocations and substructures inside the grains, which

make grain subdivision easily. Therefore, the finest microstructure (Fig. 54c) was observed in the

higher manganese steel after the ultrafine grain route.

5.4.6. Effect of carbon content on work hardening rate

In Fig. 68, when the true strain εt ≤ 0.09, the stress-strain curves of the ultrafine grained steels exhibit

work hardening. This is readily apparent in the 3C steel in which a high rate of work hardening

increases the flow stress from about 590 MPa to nearly 720 MPa within a strain increment of 0.1.

Fig. 69 shows the Kocks-Mecking plots, representing the work hardening rate θw = dσt / dεt versus the

true stress σt, for the ultrafine grained steels with different carbon contents. The curve presented is the

part between the Lüders strain and the strain where the necking starts. As expected from work

hardening theory, the results can be represented by straight lines. The slope of each straight line

(〉θw/〉σt) displays the rate of work hardening.

For example, a high value of 〉θw/〉σt means work hardening rate decreases rapidly for the same

increment of true stress. Therefore, the values of 21.69 and 21.59 suggest that the 3C and the 15C steel

have quite similar work hardening rate. On the right up of Fig. 69, it shows the average ferrite grain size

of the 3C steel (1.1 µm) is smaller than that of the 15C steel (1.4 µm). As mentioned above, the grain

refinement leads to an obvious decrease in work hardening rate. Thus, the similar work hardening rate

observed in the steels with different grain size can be attributed to the effect of carbon content. In brief,

if keeping steels with the same grain size the increase in carbon content leads to an increase in work

hardening rate. This is due to a large volume fraction of cementite particles which are effective in

increasing work hardening rate by the accumulation of geometrically necessary dislocations near the

particles.

Many two-phase alloys work-harden much faster than do pure single crystals. This is because the two

phases are not equally easy to deform. One component (often dispersed as small particles) deforms less

than the other, or not at all, so that gradients of deformation from with a wavelength equal to the

spacing between the phases or particles. Such alloys are ‘plastically non-homogeneous’, because

gradients of plastic deformation are imposed by the microstructure [77]. Dislocations are stored in them

to accommodate the deformation gradients, and so allow compatible deformation of the two phases.

These are called ‘geometrically-necessary’ dislocations to distinguish them from the ‘statistically-

stored’ dislocations which accumulate in pure crystals during straining.

Geometrically-necessary dislocations control the work hardening of the specimen when their density

exceeds that of the statistically-stored ones.

96

They contribute to hardening in two ways [78]: by acting as individual obstacles to slip, and

(collectively) by creating a long-range back-stress, with wavelength equal to the particle spacing.

In all dislocation models, work hardening is a consequence of the fact that some fraction of the mobile

dislocations that produce the strain do not exit through the surface, are not absorbed in grain boundaries,

and do not annihilate each other, but are stored in the crystals or react with other dislocations to form

new obstacles to continued slip [78].

For metallic crystals containing hard particles (A ‘hard’ particle is one that resists plastic shear, and

consequently can only deform elastically when the matrix shears plastically [79].), it is assumed by

Ashby [79] that work hardening occurs because of the interaction of geometrically necessary

dislocation loops (secondary dislocations) nucleated at the particle-matrix interface and dislocations

moving on the primary slip plane. That is, the interaction of primary, glide, dislocations with the

secondary dislocation loops that intersect the slip plane contribute to work hardening. On the basis of

the Ashby model in which the extra work hardening 〉σ, due to the accumulation of dislocation at the

particles, is expresses by the relationship [80]: 21

23

2ÕÖÔ

ÄÅÃ

=∆r

bfDM tε

µσ (3)

where M is the Taylor factor with an assumed magnitude of 2.75, µ is the shear modulus of iron (81

GPa), b is the magnitude of Burger's vector (0.248 nm), f is the volume fraction of particles, r is the

average particle radius and εt is the true strain. It can be seen that a large amount of fine particles are

beneficial to increase work hardening. D is a constant, from experimental data for polycrystalline b.c.c

alloys, varying from 0.35 to 0.52 [81,82]. Furthermore, the values of D appear to increase with

increasing temperature provided that recovery does not occur to a significant extent [81,82]. If recovery

occurs, then the magnitude of D should decrease as thermally activated plastic relaxation and/or

diffusion-controlled dislocation rearrangement and annihilation increase [83,84]. Thus at high

temperatures and low strain rates, particle hardening should increase the yield stress but have relatively

little effect on work hardening [84].

The rapid accumulation of dislocation substructures by particles, characterized by a high density of

geometrically necessary dislocations and high accompanying flow stresses, triggers the dynamic

recovery [80] during the tensile test. As mentioned before, the finer particles were found to pin the

dislocations. Thus, the 3C steel with a large amount of cementite particles (due to the high carbon

content and large strain deformation) can still maintain pronounced work hardening even with an

ultrafine microstructure. The extra work hardening is due to the accumulation of dislocations at the

particles.

As a result, the ultrafine grained steels investigated exhibit not only high strength but also acceptable

ductility, which can be seen from the total elongation of about 22% and the uniform elongation of about

10% in these steels (Fig. 65b). The strengthening effects of second-phase particles are well understood

in conventional grain size metals with regard to size, distribution, and mechanical properties of the

second phase. Hard dispersion particles increase the yield stress due to the Orowan mechanism and also

increase subsequent work hardening rates due to dislocation storage processes required for compatible

plastic strain [85]. The study of second phase particles with varying sizes is likely to provide additional

optimization of strength and ductility in ultrafine grained steels.

It can be concluded that the additional second phase – cementite particles offer the opportunity to

improve the work hardening capacity. This was reflected by the simultaneous increase of yield and

ultimate tensile stresses with decreasing the average ferrite grain size, while the elongations are nearly

the same in all the ultrafine grained steels.

5.5 Conclusions

Three new thermomechanical treatments were developed to produce ultrafine ferrite grains in plain C-

Mn steels using large strain warm deformation and subsequent annealing (as a laboratory simulation of

industrial coiling). The processing routes for producing ultrafine microstructures were found to be

flexible regarding cooling rates and the initial microstructures. This offers a large processing window

for the industrial application of the thermomechanical treatment suggested in this work.

97

Large scale specimens ~10 mm in thickness were produced in plain C-Mn steels by plane strain

compression. The final microstructure consisted of ultrafine ferrite and homogeneously distributed

cementite particles. The average ferrite grain size changed from 0.8 µm to 1.4 µm for the steels with

different carbon and manganese contents. These ultrafine microstructures were stable against

grain/particle coarsening during 2 h annealing at 823 K.

The mechanical tests showed that the ultrafine grained steels exhibit higher strength and improved

toughness than the conventional steels. A significant work hardening was found in the ultrafine grained

steels. This led to a reasonable ductility, such as 10% uniform elongation and 22% total elongation.

Grain refinement also results in an increase in the Lüders strain.

A second phase such as globular cementite is beneficial to improve both strength and work hardening

rate of steels. Therefore, it is very attractive to produce a relatively coarser microstructure (i. e. grain

size of 2&3 µm) with a second phase (i. e. finer precipitates) in the matrix. The ductility of steels will be

significantly improved due to a relative coarse microstructure and an increase in work hardening rate by

the second phase. On the other hand, in addition to grain refinement strengthening, precipitation

hardening can also contribute to an obvious improvement in strength.

Formation of the ultrafine microstructure

The grain refinement by large strain warm deformation can be attributed to deformation induced grain

subdivision and recovery. The grain subdivision is an essential process for the ultrafine grained

microstructure formation, and recovery is necessary to form high-angle grain boundaries. In order to

produce ultrafine grained steels by large strain deformation, it is important to avoid primary

recrystallization. Primary recrystallization will reduce significantly the dislocation densities and remove

the substructures which are important for the formation of subgrains and later the ultrafine grains.

During a 2 h annealing at 823 K, a further recovery occurred instead of grain growth. This facilitated

the formation of high-angle grain boundaries and the establishment of a more equiaxed grain shape. The

main mechanism of the recovery during annealing can be attributed to the accumulation of

geometrically necessary dislocations into (sub)grain (grain and subgrain) boundaries.

Spheroidization of pearlite during the large strain warm deformation was accelerated by the formation

of cementite lamellae kinks, subboundaries and fracture of lamellae. High equilibrium carbon

concentration in ferrite near these lamellar defects provoked a quick local dissolution of the lamella,

which led to the division of lamella as well as a subsequent or simultaneous spheroidization.

After the large strain warm deformation and annealing a rather homogeneous distribution of cementite

was found in an initial ferrite-pearlite steel, which can be due to several effects. Firstly, it can be

attributed to the spheroidization of cementite lamellae. Secondly, it is due to the locally partial

dissolution of cementite lamellae and cementite particles. This is enhanced by a high dislocation density

inside the ferrite and the dislocation cores increase significantly the carbon solubility in ferrite. Thirdly,

the solute carbon atoms may diffuse into the cementite-free proeutectoid ferrite areas with lower

equilibrium carbon concentration and re-precipitate. All of these processes were supported by a high

density of dislocations and lattice defects due to the large strain warm deformation. Therefore, primary

recrystallization is also not preferred for the homogeneous distribution of cementite particles.

The fine cementite particles are beneficial for the formation of a fine ferrite microstructure. Especially

they are effective in inhibiting the grain boundary migration due to Zener drag. This contributes to

stabilize the ultrafine microstructure against grain coarsening, as well as to inhibit primary

recrystallization which leads to an increase in recrystallization temperature. Therefore, the temperature

window for rolling/annealing is enlarged.

Abnormal texture evolution during annealing of the 0.2%C steel

After annealing the 0.2%C steel processed by the large strain warm deformation, a strong α-fiber was

found. The result in the present study can be attributed to the inhibition of primary recrystallization by

cementite particles. The process occurring during annealing may be addressed as continuous

recrystallization, proceeding without nucleation of new grains.

98

Effect of carbon and manganese

A finer ferrite and a more homogeneous distribution of cementite were found in the higher carbon steel

after the deformation and annealing process. It can be attributed to the initial finer ferrite-pearlite

microstructure before large strain warm deformation and the higher volume fraction of spheroidized

cementite particles during/after deformation in the higher carbon steel. In comparison, the large number

of particles seems to be more important for the formation of the final ultrafine grained microstructure

since the process included a large strain warm deformation. A higher carbon concentration in hypo-

eutectoid steels led to a higher volume fraction of pearlite after phase transformation. This facilitated a

more homogeneous distribution of cementite particles after the deformation and annealing processing in

the higher carbon steel.

Moreover, the increase in carbon content was beneficial to increase the work hardening rate. This is due

to a large volume fraction of cementite particles which are effective in increasing the work hardening

rate by the accumulation of geometrically necessary dislocations near the particles. The increase in

work hardening rate is important for improving the ductility (especially the uniform elongation) of the

ultrafine grained steels.

In contrast to carbon, the increase in manganese content had a more significant contribution to grain

refinement. This is, firstly, due to the initial finer transformed microstructure in the higher manganese

steel because of manganese depressing the transformation temperature obviously. Secondly, it can be

attributed to manganese substitutes iron in cementite and forms a more stable cohenite (Fe3-XMnXC).

During the large strain warm deformation itself and in the course of the subsequent annealing, the finer

cohenite particles, due to a sluggish dissolution as well as a slow Ostwald ripening, are more effective

to pin the ferrite grain boundaries against grain growth. Thirdly, the fine particles are also beneficial to

store the geometrically necessary dislocations and substructures inside the grains, which facilitate grain

subdivision. Therefore, the finest microstucture was observed in the higher manganese steel after the

deformation and annealing process.

The prominent combination of mechanical properties found in the ultrafine grained steel with higher

manganese content can be attributed to the effect of an obvious grain refinement by the extra

manganese.

99

6. ACCUMULATIVE ROLL BONDING OF LOW C STEELS

6.1 Introduction

Ultrafine structures are being studied world-wide on various materials using different types of processes

to give intense deformation like torsion straining, equal channel angular extrusion, mechanical milling

of powder metals, etc., But the majority of these processes also exhibits severe limitations because they

are generally restricted to small samples and are difficult to be implemented on an industrial scale.

A promising technique for industrial implementation is the Accumulative Roll Bonding (ARB) [86-89],

a process able to impose severe plastic deformation on large samples even at a lab scale.

The schematic drawing of Accumulative Roll Bonding (ARB) process is reported in Fig. 90.

In this process a strip is placed on top of another strip whose facing surfaces are cleaned both

mechanically and chemically to ensure good bonding. The two layers of the strip are rolled together and

the material is joined together. The rolled strip is then sectioned into two halves. The cut strips are again

stacked together and rolled. The whole process can be repeated again and again to accumulate strain.

In ARB, rolling is not only a deformation process but also a bonding process carried out to obtain one-

body solid final material. In order to obtain one-body solid material the sheets are heated before rolling

in order to realize good bonding and to reduce the rolling force. In such case the heating temperature

must be below the recrystallization temperature of the material because recrystallisation cancels out the

accumulated strain.

Fig. 90 – Schematic drawing of Accumulative Roll Bonding Process.

The issues concerning the accumulative roll bonding experiments were inadequate bonding at the first

pass, oxidation at the interface and time and rolling temperature optimisation in order to prevent

recrystallisation and to accumulate strain after each pass.

Oxidation of the surface during heat treatment has been minimised by welding the corners and sides

before annealing. Further the use of stainless steel foils to wrap the samples has been tested. Results

show that oxidation improved a little but presence of oxide was detected. The material used for

accumulative roll bonding tests is CSM_ARB whose chemical composition is reported in Table 8.

Steel code C Si Mn P S Al N

CSM_ARB 0.153 0.2 1.37 0.012 0.008 0.034 -

Table 8 - Chemical composition of the material used for ARB tests.

101

6.2 Experimental

Trial experiments have been carried out using a 0.15%C-1.37Mn commercial steel strip. The initial size

of samples was 300 mm in length, 50 mm in width and 3.7 mm in thickness. After degreasing and

wire-brushing the surfaces to remove the oxide scale, two strips were overlapped, edge welded, heated

and roll bonded by 50% reduction in thickness (ε=0.8) in one pass.

Initial technological difficulties were encountered in the production of ARB samples. The main

problems were: inadequate bonding due to oxidation at the interface, soaking time and rolling

temperature optimisation in order to prevent recrystallisation and to accumulate strain after each pass.

At first, oxidation of the surface during heat treatment was minimised by the use of stainless steel foils

to wrap the samples during heating. In this way oxidation between the strips a little decreased, however

technical problems in unwrapping the foils before rolling occurred.

So, welding the corners and sides before annealing was the method that gave the best results in terms of

oxidation and managing times.

Another problem was the perfect planarity of the strips. In our pilot mill planarity cannot be controlled,

so the overlapping of strips sometimes resulted in not good bonding.

In order to check the best condition of roll bonding temperature and soaking time to have deformation

accumulation without recovery or recrystallization, some tests were made at different temperatures in

the range 600-700°C and for different soaking times (1 - 15 min). The heating temperature before

rolling pass, defined as the best compromise among no recrystallization, deformation accumulation,

good bonding and decrease of rolling force was found to be T=615°C.

So, before each rolling pass the sample was heated to 615°C and maintained at this temperature for 2

min in order to homogenise the temperature of the sample. Using this rolling conditions, it was noted

that the minimum reduction needed for bonding was 40%.

The process was repeated 5 times for a total strain of about 4 (the total equivalent strain e after n cycles

of the 50% roll-bonding ARB is Ｅ 0.8n).

Further it is important to notice that the layers included in a sheet treated after n ARB cycles at 50%

deformation becomes 2n. In this case, after 5 passes there will be 32 layers having each one about

120µm thickness.

During rolling the main problem was to obtain for each pass a deformation of 50%.

6.3 Results and discussion

In Fig.91 the applied schedule and the corresponding microstructures are reported.

After the first pass, elongated grains due to deformation can be observed and after the second pass a

well deformed structure can be detected. Increasing the reduction, the structure is more and more

deformed and it is not possible to distinguish each layer by optical microscopy (except the central one

as showed in the 4th pass microstructure that sometime was not uniformly bonded). Each original layer

after the 5th pass is about 120 µm thick. Fig. 92 shows a SEM image of the deformed microstructure

after the 5th pass. From this figure no recrystallised structure can be detected.

6.3.1 Microstructure

Microstructural observations of as-ARB processed materials along the rolling direction (RD) showed

more and more deformed structures as shown in Fig. 91.

After the first pass elongated grains due to deformation can be observed and after the second pass a well

deformed structure can be detected. Increasing the reduction, the structure is more and more deformed

and it is not possible to distinguish each layer by optical microscopy. Fig. 92 shows a SEM image of the

deformed microstructure after the 5th pass. The specimen severely deformed up to 4.0 of total strain

exhibited the pancake shaped UFGs elongated along the rolling direction as shown in Fig.91e, which

have been typically observed in the ARB processed materials [87-89]. The mean thickness and length

of the elongated UFGs were 0.25 µm and 1.5 µm, respectively. Subsequent annealing procedure has

been carried out.

102

N° Pass Rolling T

(°C) Red (%)

1st pass 615 38.3

2nd

pass 615 47.3

3rd

pass 615 49.5

4th pass 615 54.3

5th pass 615 48.6

Fig. 91– Hot rolling conditions and microstructures for each pass of ARB.

103

Fig. 92– SEM image of the sample after the 5th pass before annealing.

Fig. 93– Microstructures of ARB samples annealed at 600°C x 3 min at different magnifications.

(a) (b)

Fig. 94– Microstructures of ARB samples annealed at (a)650°C x 3 min (b) 700°C x 3 min.

In order to macroscopically check if the material accumulated strain during passes, hardness was

measured after each ARB pass and it was characterized by a continuous increase from HV5= 230 to 270

(5th pass).

104

Samples after the 5th pass were annealed at 3 different temperatures: 600, 650 and 700°C for 3 minutes.

In Fig. 93 the microstructure of the sample annealed at 600°C x 3 minutes at different magnifications is

shown. Microstructure was quite homogeneous and polygonal grains were observed. Mean grain size

was 0.8 µm.

In Fig. 94a microstructure of the sample annealed at 650°C x 3 minutes was quite homogeneous and

polygonal grains were observed. In this case mean grain size was 1.7 µm, while in Fig. 94b the

microstructure of the sample annealed at 700°C x 3 min showed an inhomogeneous microstructure with

some coarse grains, indicating that abnormal grain growth occurred. For this annealing condition the

average grain size was 6 µm.

From these results it is clear that increasing the number of passes the non recrystallization temperature

decreases and during heating and warm rolling some microstructural changes must have occurred. In

fact after annealing at 600°C x 3 min deformed at ε = 4 recovery occurred.

6.3.2 Texture

In order to verify if a real ultrafine structure was obtained, orientation measurements were carried out.

Textural analyses were carried out in the central layers of the sample

In Fig.95(a) orientation map for planes //ND and in Fig.95(b) the orientation map for directions //RD,

are shown. In Fig.96 orientation distribution function of the same sample is reported. Both from Fig.95

and Fig. 96 it is clear that the texture is made mainly by α and γ fibres with a strong component at

{100}<011>, the typical deformation textures of bcc crystals. However, it is interesting to note how the

intensity distribution along the γ fibre (Fig. 97) has not peaks at {111}<110> (at 0° and 60°) as

expected in the case of deformation textures of ferritic steels, but it is exactly in the middle between

{111}<110> and {111}<112> that are the typical recrystallization textures of bcc steels.

In Fig. 98(a) the unique grain colour image of the sample annealed at 650°C x 3 min is shown. Border

lines between different colours represent high angle boundary grains with a misorientation angle

exceeding 15°, whereas the fine lines inside the same colour area delineate low angle grain boundaries.

This image shows that high angle boundary grains are not polygonal but elongated along the rolling

direction. Inside the elongated grains there are a lot of small grains with low boundary angles. This is a

typical situation of a recovered (and not recrystallised) material. It is confirmed in Fig.98(b) where

grain boundaries misorientation distribution shows a high fraction of low angles boundaries (53%

against the 47% of high boundary angle). If we fix the misorientation angle to 10° a higher fraction of

high boundary angles (63%) is detected.

(a) (b)

Fig. 95– (a) Unique grain colour image of the sample annealed at 650°C x 3 min, (b) Misorientation

angle distribution.

105

Fig. 96– Orientation Distribution Function of the sample.

Fig. 97– γ - Fibre of the sample.

(a) (b)

Fig. 98– Inverse polar figure maps according the colour code reported: (a) {hkl}, (b) <uvw>

106

6.3.3 Mechanical properties

Tensile tests showed that decreasing the annealing temperature of ARB specimens (εtot=4) from 700 to

600°C for 3 minutes, allows to increase the YS from 315 MPa (average grain size 6µm) to 600 MPa

(average grain size 0.8µm), without significant effects on ductility (A50%=14-15%). However, the yield

to tensile ratio increased from 0.73 to 0.96. The material annealed at 650°C (average grain size 1.7µm)

exhibited an intermediate behaviour with YS=510 MPa and YS/TS =0.92.

However, these values if compared with standard mechanical properties for this steel grade (TSＥ

420MPa and total elongation Ｅ13%) are quite good. It is interesting to notice that total elongation is

not badly deteriorated decreasing the grain size. Mechanical properties are reported in Fig. 99.

Tensile tests

0

100

200

300

400

500

600

700

0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16

strain

MP

a

650°C x 3 min

700°C x 3 min

600°C x 3 min

anticipated failure

Fig. 99- σ−ε curve of ARB samples.

(a) (b)

Fig. 100– Bonding situations in (a) et mid-thickness after 5th pass and annealing (b) defect.

6.3.4 Bonding

A cleaning procedure was set up in order to have good bonding among layers during rolling. Oxidation

of the surface during heat treatment has been minimised by welding the corners and edges before

annealing.

From microstructural observations interface among layers formed in the previous rolling passes are not

detectable anymore; it means that fully metallurgical bonding occurred. A bonded layer can be observed

107

at the left of Fig.100a. It can be noted that it works like a grain boundary even if in the right part the

bonded layer is not detectable anymore. Further, very seldom some oxide residuals have been detected

as shown in Fig. 100a.

Sometime situations like that shown in Fig. 100b occurred at the central layer of the sample. It can be

noted that in this case bonding was not so good because a lot of oxides were trapped between the layers

during hot rolling. Separation among layers are usually limited only at the last bounded boundary at the

centre. This means that the joining at the bonded boundaries becomes stronger and stronger during

rolling in subsequent cycles.

6.4 Conclusions

ARB process was applied to 0.15%C-1.37%Mn steel in order to obtain ultrafine grains. Different

annealing temperatures originated grain sizes ranging from 0.8µm to 6µm.

The experimentation showed that this severe plastic deformation method, that could be extended to

industrial scale, gives the possibility to obtain customized microstructures with improved combination

of strength and ductility/toughness through adequate grain refinement (1-3 µm).

Microstructure refinement and the relevant increase of tensile strength could be a solution to decrease

alloying elements at a given strength level, allowing the improvement of recyclability, weldability,

coatability and toughness.

The critical point for the production of ultrafine grained steels by ARB is the surface cleaning, which is

essential for a good bonding.

108

7. MICROALLOYED LOW C STEELS

7.1 Introduction

The grain refinement by means of dynamic recrystallisation, combined with the presence of Nb in

solution, has been already exploited by CRM to produce a very fine dual-phase microstructure,

beneficial to ductility, in hot rolled strips continuously cooled (DPCC) [90,91]. However, the ultra fine

dual phase (UFDP) route established during the previous ECSC project was though to be too

complicated and thus expensive for adoption commercially.

The present research will focus on ultra fine ferrite microstructure obtained in low carbon steel

microalloyed with Nb.

The key factor to refine the microstructure is to finish hot rolling below the non recrystallisation

temperature, Tnr, and above the transformation temperature Ar3. The temperature range between Tnr and

Ar3 should be enlarged to make practicable the rolling schedule on an industrial line, with consistent

results. Therefore a sufficient amount of Nb is added to avoid static recrystallisation between the rolling

passes.

During the hot rolling process of medium-low carbon steel containing Nb, several phenomena occur.

Nb retards static recrystallisation at lower rolling temperatures and smaller interpass times. It increases

the non-recrystallisation temperature, Tnr [92]. However, strain accumulation favours dynamic

recrystallisation, beneficial for the austenite grains refinement[93,94]. During cooling from the

austenitic region, NbC static precipitation can occur. Moreover, straining induces dynamic precipitation.

All these precipitation and recrystallisation phenomena take place more or less simultaneously,

depending on the initial steel composition and thermomechanical process. Schmitz et al. have found

that, for the same content of Nb in solution, the grain size of the ferrite which transformed from a

partially dynamically recrystallised austenite or from a pancaked austenite is the same [95]. This result

indicates that it is the Nb in solution rather than the occurrence of dynamic recrystallisation that

influences the ferrite grain refinement.

Dilatometry was combined to ICP (Inductively Coupled Plasma) analyses to study the static

precipitation of NbC at 900 °C in Nb steels. The results of NbC quantification are compared to the

kinetics of ferrite formation at 650 °C in 3 steels with different initial Nb contents.

Torsion tests were performed with Nb steels to study the effects of strain induced NbC precipitation on

the final ferrite grain size. These experiments aimed at reducing the processing window for further hot

rolling of Nb steels by adjusting the steel composition and the process parameters (in particular the time

between the last rolling pass and the cooling to the coiling temperature) in order to obtain the finest

ferrite grains.

Five steel compositions containing Nb and different amounts of Ti have been cast, hot-rolled and

processed by ultra fast cooling to different coiling temperatures. The effects of soluble B and P

additions have also been investigated. The mechanical properties and the final grain sizes have been

analysed.

Finally, several hot rolled plates have been cold rolled and annealed to study the heredity of the

improved tensile properties after further processing.

7.2. Experimental

The steel compositions used for the study of the ferrite grains refinement are given in Table 9. The

theoretical temperature for the end of static recrystallisation, Tnr [96] and the temperature of NbC

precipitation in austenite [97] are indicated and were calculated using the following equations:

Tnr,°C = 887+464wt%C+890wt%Ti-357wt%Si+6445wt%Nb-644wt%Nb1/2+363wt%Al (4)

TNbC, °C = -6770/(LOG((wt%Nb)*(wt%C))-2.26)-273 (5)

109

Cylindrical samples with 10 mm length and 4 mm diameter of the CMnNb steels 416, 419 and 568,

respectively, were reheated in the dilatometer to 1250 °C to dissolve all the Nb and subsequently cooled

to 900 °C, then held at 900 °C during different times tp (6 s, 200 s, 5000 s and 10000 s) in order to

precipitate different amounts of Nb. The samples were then quenched to room temperature or to the

isothermal transformation temperature 600 °C, 650 °C or 700 °C. The dilatometer cycles are

summarised in Fig.101.

Torsion specimens with 6 mm diameter of the CMnNb steels 430 and 431 and of the reference steel 464

without Nb, were subjected to different torsion schedules shown in Fig. 102. Specimens were reheated

to 1250 °C. A roughing pass was simulated at 1100 °C (ε = 0.5). A finishing pass of ε = 2 was given at

900 °C to simulate strain accumulation in the last rolling passes. One series was then immediately

cooled with a cooling rate between 1 °C/s and 60 °C/s. A second series was first held during 100 s after

finishing, before cooling. The phase transformations during cooling after hot torsion were monitored by

a thermomagnometer [98].

Torsion specimens and dilatometer samples quenched to room temperature were dissolved and filtered

in order to determine their amount of Nb precipitated by means of inductively coupled plasma (ICP)

analysis. The details of the procedure are given elsewhere [99].

Five steel compositions were cast to study the effects of Nb precipitation on the formation of ultrafine

ferrite during hot rolling. All steels contained 0.06 wt% C and 0.055 wt% Nb. The effects of Ti, B and P

were studied. Pieces with 150 mm width, 110 mm length and 30 mm thickness were reheated one hour

at 1250 °C. They were covered by stainless steel foils to prevent their oxidation during the

austenitisation treatment. Hot rolling experiments were performed using 5 passes or 3 passes with high

rolling reductions in the last passes above Ar3. The details of the rolling cycles are given in the

following paragraphs. Hot rolled plates with 3 to 4 mm thickness were cooled at 300 °C/s to the coiling

temperature using a mist jet cooling device. Coiling simulations were performed by cooling at 50 °C/h

in a furnace.

Casting nr. C Mn Si P Nb Ti B Al sol. N Tnr TNbC

416 87 1530 130 16 53 22 0 40 3.3 1108 1200

419 83 1410 124 15 83 22 0 40 3.5 1265 1258

568 82 1650 18 14 45 1 1 49 8.3 1091 1170

430 72 1540 23 16 41 2 0 40 3.7 1062 1140

431 34 1510 22 17 40 0 0 40 3.8 1038 1048

464 79 1470 23 15 0 0 0 40 4 930 /

806 62 500 6 11 59 3 0 43 4.8 1156 1168

1134 57 540 8 11 54 25 0 50 6.2 1146 1146

1135 58 520 14 11 54 57 0 52 8.8 1173 1148

1185 62 500 20 8 55 21 3.2 27 5 1141 11591186 65 500 14 91 55 26 4.1 45 5.1 1155 1165

Torsion tests

Rolling trials

Dilatometry

Table 9 - Steel compositions (10-3

wt%).

110

300 s at 1250 °C

60 °C/s

60 °C/s

tp at 900 °C

700 °C

650 °C

600 °C

tp = 6 s, 200 s, 5000 s and

10000 s

solution

treatment

precipitation

γ → α transformation

Fig.101 - Dilatometric cycles.

300 s

1250°C eps=0.5 at

1100°Ceps=2 at

900°C

100 s

delay

Fig. 102 - Thermomechanical (torsion) schedules.

Tensile specimens with 50 mm gauge length (A50) were machined from hot rolled plates and the

microstructures were investigated by light optical microscopy. The grains sizes were measured using

the line intercept method.

Finally, hot rolled plates were cold rolled to 1.5 mm thickness (67 % reduction). Tensile specimens A50

were machined and directly annealed to varied temperatures in salt bath.

7.3. Results And Discussion

7.3.1 Dilatometry

The change in length measured during holding at 650 °C the austenitic samples after different

precipitation times at 900 °C is shown for the 3 studied compositions in Fig. 103. Static precipitation of

NbC occurred at 900 °C. Those precipitates acted as nucleation sites for the γ - α transformation. On the

contrary, the presence of Nb in solution increased the Solute Drag Like Effect (SDLE) [100], thus

decreasing the kinetics of the γ - α transformation. Note that this effect was stronger in the steel 568

containing the lowest amount of Nb and no Ti. In the other steels, Nb was only partially dissolved at

1250 °C, or Nb was partially precipitated during the cooling to 900 °C, leading to the presence of some

111

NbC at 900 °C. In particular, the temperature of NbC dissolution calculated for the steel 419 by means

of the solubility product in Table 9 is above 1250 °C.

NbC quantification was done by means of ICP analysis carried out on dilatometer samples quenched

after holding each steel composition at 900 °C for 6 s, 200 s, 5000 s and 10000 s, respectively. The

results of ICP measurements are shown in Fig. 104. It is very clear that NbC statically precipitated

during holding at 900 °C. The amount of NbC precipitates present before the γ−α transformation acted

as nucleation sites and increased the kinetics of the phase transformation. The microstructures obtained

after the isothermal γ−α transformations at 650 °C are shown in Fig. 106 for the 3 steels with increasing

Nb content and for different holding times. The NbC precipitation at 900 °C clearly influences the

microstructures by providing nucleation sites for the ferrite formation, increasing the number of α

grains, thus reducing their size.

The volume fraction of transformed austenite y(t) corresponds to the fraction of ferrite formed, fα, and

was calculated from the dilatation curves at each time t using the relation:

0

0)(1)(

ll

ltlfty

f −

−=−= γ

(3)

where l(t), l0 and lf correspond to the sample length at time t, at the beginning of the transformation and

at the end of the transformation, respectively. The results of this procedure are shown in Fig. 105 for the

steel composition 419 isothermally transformed at 650 °C.

At each transformation temperature, the Johnson-Mehl-Avrami (JMA) analysis was carried out by

fitting the experimental data to the JMA equation [101]:

[ ]ntkty )(exp1)( −−= (6)

where

ÕÖÔ

ÄÅÃ −

=RT

QATk exp)(

(7)

and y(t) is the fraction of austenite transformed, t, the time, T, the temperature, A, the frequency factor,

Q, the activation energy, k(T), the temperature-dependent rate constant and n, the Avrami exponent. R

is the gas constant, 8.314 JK-1

mol-1

. The constants n and k were determined experimentally by

modifying equation (4) to:

)ln(ln)(1

1lnln tnk

ty+=ÕÕ

Ö

ÔÄÄÅ

ÃÕÕÖ

ÔÄÄÅ

Ã−

(8)

Fig. 107 shows the double logarithmic plot for the steel 419. Using this procedure for each steel

composition, the values of n and k were determined and the experimental dilatometric curves were

fitted to the JMA equation. The fraction of austenite transformed to ferrite was calculated for different

isothermal transformation times and was plotted against the NbC content before the start of the

transformation (NbC precipitated at 900 °C) in Fig. 108. A clear correlation between the Nb content in

the precipitates, i.e. the number of precipitates and the kinetics of phase transformation is observed.

112

419: 0.083 C - 0.083 Nb - 0.012 Si. 650°C

0

5

10

15

20

25

30

35

40

45

0 100 200 300 400 500 600 700 800 900 1000

Time (s)

Ch

an

ge

in

le

ng

th (

µm

)

10000 s

5000 s

200 s

6 s

416: 0.087 C - 0.053 Nb - 0.013 Si. 650°C

0

5

10

15

20

25

30

35

40

45

0 100 200 300 400 500 600 700 800 900 1000

Time (s)

Ch

an

ge i

n l

en

gth

(µ

m)

10000 s

5000 s

200 s

6 s

568: 0.082 C - 0.045 Nb. 650°C

0

5

10

15

20

25

30

0 100 200 300 400 500 600 700 800 900 1000

Time (s)

Ch

an

ge

in

le

ng

th (

µm

)

10000 s

5000 s

200 s

6 s

Fig. 103 - Length change versus time during the γ−α transformation at 650 °C after different holding

times at 900 °C.

113

0

5

10

15

20

25

30

35

40

45

50

55

60

0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000

time at 900 °C, s

Nb

in

pre

cip

itate

s,

10

-3 w

t%

53Nb-130Si-22Ti

83Nb-124Si-22Ti

45Nb-18Si

Fig. 104 Results of NbC precipitates quantification in dilatometer samples hold at 900 °C and quenched.

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

1 10 100 1000

t, s

f

tp = 200 s

tp = 5000 s

tp = 10000 s

tp = 6 s

tp = 6 s

tp = 10000 s

tp = 200 s

tp = 5000 s

Fig.105 - Kinetics of γ−α phase transformation at 650 °C in steel 419.

114

t p a

t 9

00

°C

Initial Nb content

80µm

Fig.106 - Microstructures after holding at 900 °C for tp= 6, 200, 5000 and 10000 s (from top to bottom)

and for different initial Nb content: 0.041, 0.053 and 0.083 wt% (from left to right). Isothermal

transformation at 650 °C.

115

y (6 s) = 1.5192x - 8.5881

R2 = 0.9988

y(200 s) = 1.034x - 5.4737

R2 = 0.9999

y (5000 s) = 1.2912x - 7.1411

R2 = 0.9992

y(10000 s) = 0.6664x - 3.1843

R2 = 0.9984

-4

-3

-2

-1

0

1

2 3 4 5 6

ln t

ln l

n (

1/(

1-f

))

6 s

200 s

5000 s

10000 s

Fig.107 - Determination of n and k parameters of the JMA equation from the curves of Fig.105.

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

1.1

0 100 200 300 400 500 600

ppm Nb

Ferr

ite f

racti

on

f

t at 650 °C = 100 s

t at 650 °C = 500 s

t at 650 °C = 200 s

Fig.108 -. Transformed fraction to ferrite after different times at 650 °C versus the Nb content in NbC

precipitates formed at 900 °C.

Dilatometric measurements showed that increasing the holding time, tp, at 900 °C increases the number

of NbC precipitates that act as nucleation sites for the γ−α transformation, which leads to the formation

of more, thus finer ferrite grains.

116

7.3.2 Torsion tests

The static precipitation of NbC during cooling (effect of cooling rate) and dynamic precipitation of NbC

after hot deformation were studied by means of hot torsion tests combined with ICP analyses for Nb

quantification.

The signal provided by the magnetometer allowed the determination of the start and finish temperatures

of the γ−α transformation during cooling. An example of curves measured experimentally in the Nb-

free steel 464 is shown in Fig. 109. The jump of potential is due to the paramagnetic-ferromagnetic

transformation that begins when ferrite starts to form during cooling from the austenite region. These

signal analyses permitted to plot DCCT (Deformation Continuous Cooling Transformation) diagrams

for the studied compositions with and without the holding time of 100 s at 900 °C. It can be seen in Fig.

110 that only at high cooling rates (> 20 °C/s) the start temperature of austenite decomposition is raised

after holding 100 s at 900 °C, likely due to the precipitation of Nb before cooling. At lower cooling

rates, Nb has enough time to precipitate during the cooling prior to the γ−α transformation.

464: 0.079C, no Nb

-2.5

-2

-1.5

-1

-0.5

0

0.5

1

100 200 300 400 500 600 700 800 900 1000 1100 1200

Temperature

Mag

neto

mete

r (V

)

1°C/s

6°C/s

20°C/s

60°C/s

Fig.109 -. Magnetometer signal recorded during cooling the steel without Nb from 900 °C at cooling

rates 1, 6, 20 and 60 °C/s, respectively.

Increasing the cooling rate after the deformation at 900 °C decreased the number of NbC precipitates.

However, coarser ferrite grains were obtained in the specimens containing higher amount of NbC

precipitates, i.e. those cooled at the slower rates. The coarse ferrite grains obtained at low cooling rates

were due to the fact that the ferrite grains had more time to grow. In specimens cooled at rates higher

than 15 °C/s, few differences in the NbC content were measured and no differences in the grain size

were observed.

To study the strain induced precipitation of NbC, torsion specimens were strained at 1100 °C with

ε=0.5, then cooled in air to 900 °C, then strained with ε=2. Specimens were quenched after the first

deformation, just after the second deformation or 100 s after the second deformation.

117

Steel 430: 0.07 %C - 0.04 %Nb

Ac1 Nb = 625 °C

Ac3 Nb = 762 °C

Ac1 No Nb = 653 °C

Ac3 No Nb = 768 °C

400

450

500

550

600

650

700

750

800

850

900

1 10 100 1000

Cooling time from 900 °C, s

Te

mp

era

ture

, °C

tp = 0 s

tp = 100 s

no Nb

60 °C/s 20 °C/s 1 °C/s6 °C/s

120

149

145

137

187

168

146

189

178

146

180

191

HV 5kg

no Nb

Nb, tp = 0 s

Nb, tp = 100 s

Steel 431: 0.03 %C - 0.04 %Nb

Ac1 = 638 °C

Ac3 = 761 °C

400

450

500

550

600

650

700

750

800

850

900

1 10 100 1000

Cooling time from 900 °C, s

Te

mp

era

ture

, °C

tp = 0 s

tp = 100 s

1 °C/s6 °C/s20 °C/s60 °C/s

HV 5kg

tp = 0 s

tp = 100 s

167

166

171

162

154

150

166

165

Fig. 110 -. DCCT diagrams for steels 430 (top) and 431 (bottom) with and without holding 100 s at 900

°C. The DCCT diagram for steel 464 without Nb is also indicated.

The results of NbC quantification are shown for both steels 430 and 431 in Fig.111. It can be seen that a

large part of NbC already precipitated at 1100 °C in both steels. After the second deformation, at 900

°C, some NbC again precipitated. Finally, the amount of Nb precipitated as NbC nearly doubled after

100 s holding at 900 °C. Considering that these NbC precipitates may act as nucleation sites for the

ferrite formation as observed previously with dilatometer experiments, finer ferrite grains are expected.

118

0

2

4

6

8

10

12

14

16

18

20

Nb

in

Nb

C p

rec

ipit

ate

s, 1

0-3

wt%

after 1st def, 1100 °C after 2nd def, 900 °C after 2nd def + 100s, 900 °C

Step

72C-41Nb

34C-40Nb

Fig. 111 - NbC quantification results in torsion specimens quenched after different steps.

Steel 431: 0.03 %C - 0.04 %Nb

0

1

2

3

4

5

6

0 10 20 30 40 50 60 70

Cooling rate, °C/s

Gra

in d

iam

ete

r, µ

m

tp = 0 s

tp = 100 s

Fig. 112 - Measured ferrite grain diameter for different continuous cooling rates in steel 431 deformed

at 900 °C with and without 100 s holding at 900 °C.

However, it was clearly seen that the holding time of 100 s at 900 °C led to coarser ferrite grains, as

shown in Fig. 112. Microstructure observations (Fig. 113) showed that after 100 s holding at 900 °C,

the ferrite formed from partially recrystallised austenite. Torsion experiments showed, that increasing tp

after the deformation at 900°C led to more recrystallised austenite, responsible for the formation of

coarser ferrite grains.

119

a) CR = 60 °C/s

b) CR = 20 °C/s

c) CR = 6 °C/s

no delay delay 100s

Fig. 113 - Microstructures after continuous cooling the steel 430 from 900 °C without (left) and with

(right) 100s delay after deformation at a) CR = 60 °C/s, b) CR = 20 °C/s and c) CR = 6 °C/s.

Fig. 114 shows the hardness and grain size as functions of the amount of NbC precipitated in steel 431.

The precipitation hardening due to NbC precipitation is not observed. It is very likely that the NbC

quantification technique involving a filtering of precipitates does not account for the finest precipitates

responsible for precipitation hardening. The clear correlation between grains size and hardness is

observed. This graph shows that in Nb steels hot deformed at 900 °C, Nb must be in solution to obtain

the finest ferrite grains.

20 µm

120

Effects of NbC precipitation on GS and hardness in 0.03 wt% C - 0.04 wt% Nb steel

140

145

150

155

160

165

170

175

180

50 100 150 200 250 300 350 400

Nb in (NbC) precipitates, ppm

Ha

rdn

ess, H

v5

3

3.5

4

4.5

5

5.5

6

6.5

7

Gra

ins s

ize, µ

m

Fig. 114 - Effect of NbC precipitation on the final ferrite grains size and hardness in Nb steels.

7.3.3 Hot rolling simulation

To obtain fine grains in Nb microalloyed steels, there is a compromise to find between:

- dynamic recrystallisation during high rolling reductions;

- static recrystallisation retarded by soluble Nb below Tnr, the non recrystallisation temperature;

- dynamic NbC precipitation during hot rolling;

- static NbC precipitation during cooling / coiling.

Fine recrystallised austenite grains are beneficial for the production of fine ferrite grains. Fine NbC

precipitates are beneficial for the production of fine ferrite grains as they act as nucleation cycles.

To obtain fine grains in hot rolled Nb steel, all the Nb must be dissolved at 1250 °C, strain

accumulation must occur below Tnr and above Ar3 to induce dynamic recrystallisation of deformed

austenite grains. To reduce the NbC precipitation during cooling, and also the austenite static

recrystallisation, the last pass must be followed by fast cooling to the γ−α transformation temperature.

First rolling trials were performed using steel 806 containing 0.06 wt% C, 0.5 wt% Mn, 0.06 wt% Nb.

That low C-Nb steel was hot rolled with the maximum reduction below Tnr but above Ar3, followed by

ultra fast cooling to the coiling temperature (CT) in the range 650 – 500 °C. The rolling forces and

temperatures measured during the trials are given in Fig. 115.

Fig. 116 shows the yield stress and tensile strength measured after coiling at temperatures between 650

°C and 500 °C. The maximum precipitation hardening due to fine NbC precipitation in the ferrite is

clearly seen after coiling at 600 °C. In spite of the small rolling reduction, the low finish rolling

temperature of 850 °C (before the last pass) followed by ultra fast cooling led to ferrite grains with 3 to

5.5 µm diameter. The microstructure was similar for all coiling temperatures and consisted in ferrite

and small amounts of pearlite, as seen in Fig. 117. The ferrite grain size was found to decrease with

decreasing CT, likely due to the presence of more nucleation sites provided by NbC precipitated, and

less time for the growth of coarser ferrite grains during the cooling down to CT.

121

600

650

700

750

800

850

900

950

1000

1050

1100

1150

1200

50 60 70 80 90 100 110

Time (sec)

Te

mp

era

ture

(°C

)

0

10

20

30

40

50

60

70

80

90

100

110

120

Ro

llin

g f

orc

e (

t)

Pass 1 Pass 2 Pass 3 Pass 4 Pass 5

Tin, °C 1100 1050 1000 950 900

thickness in, mm 20 16 12 9 6

thickness out, mm 16 12 9 6 4

reduction, % 20 25 25 33 33

Fig. 115 - Rolling forces and temperatures measured during the 5 rolling passes.

420

440

460

480

500

520

540

560

450 500 550 600 650 700

Coiling temperature, °C

YS

, T

S,

MP

a

Fig. 116 - Yield Stress (YS) and Tensile Strength (TS) measured after coiling the low C low Mn Nb

steel 806 at 500 °C, 550 °C, 600 °C and 650 °C, respectively.

122

600 °C

dα = 5.5 µm

650 °C

dα = 5.2 µm

500 °C

dα = 3.0 µm550 °C

dα = 4.3 µm

806: 600 ppm Nb, no Ti; 5 passes above 850 °C

Fig. 117 - Microstructure and measured ferrite grain size in steel 806 coiled at 500 °C, 550 °C, 600 °C

and 650 °C, respectively.

New rolling trials with higher rolling reduction in the last passes and higher finishing rolling

temperature in order to avoid ferrite formation during rolling have been performed on four steel

compositions 1134, 1135, 1185, 1186 containing Nb, Ti, B and P additions. The thermomechanical

schedule is given in Fig. 118 with the measured temperatures and rolling forces. It can be seen that

during the last pass a high rolling force was necessary to achieve the 50 % reduction. The

microstructures obtained with this rolling scheme are shown in Figs. 119 and 120 for the steels 1134

and 1135 with 0.025 and 0.057 wt% Ti, respectively.

In both steels the maximum grain size is measured after coiling at 550 °C, corresponding to the

maximum precipitation of NbC and TiC. The average ferrite grain size was coarser than in steel without

Ti. It is very likely that Ti additions led to TiN precipitates formed at high temperatures. These

precipitates should act as nucleation sites for the ferrite nucleation. However, as seen during torsion

experiments, the static recrystallisation of austenite is very fast at 900 °C and leads to coarser ferrite

grains, even if the number of precipitates is higher.

B additions in steels 1185 and 1186 aimed at reducing the nucleation sites for the ferrite formation by

segregation of soluble B at the austenite grain boundaries [102]. After hot rolling and fast cooling to the

transformation temperature, the sites available for the ferrite nucleation are the defects induced by

plastic deformation during rolling and TiN and NbC precipitates. Figs. 121 and 122 show the

microstructures obtained in steels 1185 and 1186, respectively. Before coiling at 500 °C, plates were

undercooled below 500 °C, and other plates were cooled to 600 °C. These differences in the cooling

cycle are indicated as UFC500 and UFC600, respectively. In both B-bearing steels the ferrite grain size

was finer than in steels 1134 and 1135. Note that the morphology of the ferrite grains is also different

(less isotropic). The finest ferrite grain size is achieved with the P-containing steel. This alloying

element promotes the γ−α transformation and contributes strongly to substitutional hardening.

Comparing the rolling schemes of Fig. 115 and Fig. 118, it can be seen that the finishing rolling

temperature (FRT) was higher in the second schedule. It is very likely that the FRT is responsible for

different recrystallisation rate and may be an important parameter to obtain fine grains. Due to the high

force needed to obtain the 50 % reduction the last rolling pass was given at 900 °C, also reducing the

risk of austenite to ferrite transformation during rolling.

123

600

650

700

750

800

850

900

950

1000

1050

1100

1150

100 110 120 130 140 150

Time (sec)

Te

mp

era

ture

(°C

)

0102030405060708090100110120130140150160170180190200210220

Ro

llin

g f

orc

e (

t)

Pass 1 Pass 2 Pass 3 Pass 4 Pass 5

Tin, °C 1100 1050 1000 950 900

thickness in, mm 30 20 15 10 6

thickness out, mm 20 15 10 6 3

reduction, % 33.33333 25 33.33333 40 50

Fig. 118 - Rolling forces and temperatures measured during the 5 rolling passes (new rolling scheme

with higher reduction in the last passes).

600 °C

dα = 5.8 µm1134:550 ppm Nb, 250 ppm Ti

5 passes above 900 °C

550 °C

dα = 5.9 µm

500 °C

dα = 4.8 µm

Fig. 119 - Microstructure and measured ferrite grain size in steel 1134 coiled at 500 °C, 550 °C and 600

°C, respectively.

124

600 °C

dα = 4.9 µm1135:550 ppm Nb, 570 ppm Ti

5 passes above 900 °C

550 °C

dα = 6.5 µm

500 °C

dα = 5.0 µm

Fig. 120 - Microstructure and measured ferrite grain size in steel 1135 coiled at 500 °C, 550 °C and 600

°C, respectively.

600 °C 550 °C

500 °C

UFC 600 UFC 500

1185

dα = 5.7 µm dα = 4.5 µm

dα = 5.1 µm dα = 4.3 µm

Hv: Hv:

179 167

Hv: Hv:

156 170

Fig. 121 - Microstructure and measured ferrite grain size in the B-bearing steel 1185 coiled at 500 °C,

550 °C and 600 °C, respectively.

125

500 °C

1186

600 °C 550 °C

UFC 600 UFC 500

dα = 6.2 µm dα = 3.8 µm

dα = 3.4 µm dα = 5.1 µm

Hv: Hv:

224 184

Hv: Hv:

172 212

Fig. 122 - Microstructure and measured ferrite grain size in steel B-bearing steel 1186 microalloyed

with P, coiled at 500 °C, 550 °C and 600 °C, respectively.

Pass 1 Pass 2 Pass 3

Tin, °C 1050 1000 950

thickness in, mm 30 20 10

thickness out, mm 20 10 3

reduction, % 33.33333 50 70

600

650

700

750

800

850

900

950

1000

1050

1100

120 130 140 150 160 170

Time (sec)

Te

mp

era

ture

(°C

)

0153045607590105120135150165180195210225240255270285300

Ro

llin

g f

orc

e (

t)

Fig. 123 - Rolling forces and temperatures measured during the 3 rolling passes (new rolling scheme

with the highest possible reduction in the last pass).

New rolling tests were performed on the steel compositions 1135, 1185 and 1186. Only 3 rolling passes

were given above 950 °C to obtain the maximum possible reduction of 70 % (with CRM laboratory mill

126

capacity) in the last pass. It can be seen in Fig. 123 that the limit of the rolling mill power was reached

with 300 t rolling force in the last pass. This high reduction was followed by ultra fast cooling and

coiling simulation from 500 °C. The microstructure is shown in Fig. 124 for the steel composition 1135

and is compared to the microstructure obtained with the 5 passes schedule. Fig. 125 shows the

microstructures obtained in steels 1185 and 1186 processed with the 3 passes rolling schedule. The

ferrite grain size was down to 1 µm finer with this high reduction rolling scheme.

1135:

550 ppm Nb, 570 ppm Ti500 °C

dα = 5.0 µm

5 passes above 900 °C 3 passes above 900 °C

500 °C

dα = 3.9 µm

Fig. 124 - Microstructure and measured ferrite grain size in steel 1135 coiled at 500 °C, after being

processed using 2 different rolling schemes.

500 °C 500 °C

1185 1186

3 passesdα = 4.4 µm

Hv: 167

dα = 2.9 µm

Hv: 188

Fig. 125 - Microstructure and measured ferrite grain size in steels 1185 and 1186

hot rolled in 3 passes with 70 % reduction in the last pass and coiled at 500 °C.

Fig. 126 is a summary of the obtained ferrite grain size in the 3 processed B-free steels in function of

the coiling temperature. It must be kept in mind that for the Ti-free steel 806, the FRT was lower than

for other steels. However, Ti additions to Nb steel seem detrimental for the grain refinement and for

127

NbC precipitation hardening efficiency. Some authors have shown, that during the casting stage, Ti

additions change the precipitation mode of Nb(C,N) in austenite and their composition [103].

In steel containing 0.1 wt%C, 0.005 wt%N, 0.03 wt%Nb and 0.001 wt%Ti TiN and AlN precipitate at

high temperature (T>1130 °C) and NbC0.7N0.3 form at T<980 °C. These precipitates precipitated along

prior austenite grain boundaries.

In steel with 0.018 wt%Ti and 0.005-0.0025 wt%N, Nb0.5Ti0.5(C0.5-0.8N0.5-0.2) precipitates formed at

T<980 °C. They precipitated on TiN nucleation sites inside γ grains. Ouchy et al. summarised the

improvement of high temperature ductility brought by Ti additions and/or reduction of N content in the

steel as follows [104]:

1. Suppression of AlN precipitation;

2. Change in chemistry of carbonitrides from Nb(C0.6N0.26) with molar ratio C/Nb > 2 to NbC0.86;

3. Complete suppression of NbN.

The effects of B and P additions on the ferrite grain size are shown in Fig. 127. Only 30 ppm B are

sufficient to refine the ferrite grain from 5.9 µm to 4.5 µm (coiling at 550 °C).

2

2.5

3

3.5

4

4.5

5

5.5

6

6.5

7

475 500 525 550 575 600 625 650 675


Ferr

ite g

rain

siz

e, µ

m

No Ti

250 ppm Ti

570 ppmTi

570 ppm Ti, 3 passes

Increasing Ti

Increasing

reduction

Fig. 126 - Ferrite grain size measured after rolling and coiling the different low C low Mn Nb steels at

different CT.

128

2

2.5

3

3.5

4

4.5

5

5.5

6

6.5

475 500 525 550 575 600 625


Ferr

ite g

rain

siz

e, µ

m

250 ppm Ti

250 ppm Ti + B

250 ppm Ti + B + 900 ppm P

Fig. 127 - Effects of B and P additions on the ferrite grain refinement.

Fig. 128 shows the tensile strength (TS) measured for the different steels using A50 tensile specimens. In

steels 806 (no Ti) and 1134 (250 ppm Ti), TS increases with decreasing CT, due to ferrite grain

refinement and precipitation hardening. In steel 1135 (570 ppm Ti), precipitation hardening contributes

more to TS than the grain size.

450

475

500

525

550

575

600

475 525 575 625 675


TS

, M

Pa

No Ti

250 ppm Ti

570 ppm Ti

570 ppm Ti ,3 passes

Fig. 128 - Tensile Strength (TS) measured after rolling and coiling the different low C low Mn Nb

steels at different CT. Effects of Ti additions.

Fig. 129 shows the tensile strength (TS) measured for the steels 1134, 1185 and 1186. The TS increase

in B-bearing steel 1185 is mainly due to the lower ferrite grain size. Note that in steel 1186, the tensile

properties remain very stable with the coiling temperature. This would allow flexibility in the cooling

stop temperature. This consistency in mechanical properties is clearly seen in Fig. 130, which

summarises the tensile properties of the 5 steel compositions after different rolling cycles.

129

450

475

500

525

550

575

600

625

475 500 525 550 575 600 625


TS

, M

Pa

250 ppm Ti

250 ppm Ti + B

250 ppm Ti + B + 900 ppm P

Ti-B-P, 3 passes

Fig. 129 - Tensile Strength (TS) measured after rolling and coiling the different low C low Mn Nb-Ti

steels at different CT. Effects of B and P additions.

0

5

10

15

20

25

30

35

40

400 450 500 550 600 650

TS, MPa

A5

0,

%

806: no Ti, 50 Nb

1134: 25 Ti, 50 Nb

1135: 50 Ti, 50 Nb

1185: 20 Ti, B, 50 Nb

1186: 20 Ti, B, 50 Nb, 90 P

3 passes

Fig. 130 - Total elongation (A50) versus Tensile Strength (TS) plot for the 5 studied steel compositions

hot rolled following different rolling schemes.

7.3.4 Hot rolling process parameters affecting the ferrite grain size

The main parameters that determine the final ferrite grain size during hot rolling are: the chemical

composition that affects the transformation temperatures, the austenite grain size, the cooling rate

between finishing rolling and coiling, the coiling temperature, the delay time between the end of rolling

and cooling, etc.

The austenite grain size is determined by the rolling conditions (reductions, temperatures, interpass

times, line speed, etc.) and by the chemical composition (Nb additions, etc.).

During this research work some of these parameters were studied, and their main effects on the ferrite

grain size are summarised hereinafter:

130

- Microalloy content:

Nb is a very important microalloying element, widely used for grain refinement. When in solid solution

in austenite, it retards recrystallisation during rolling, and also lowers the kinetics of the γ−α

transformation. Fig. 131 shows a tensile strength versus ferrite grain size plot for industrial coils (data

from Carlam hot strip mill). It is clearly seen that low carbon steels microalloyed with Nb exhibits finer

grains and consequently higher tensile strength.

The other effect of microalloying with Nb is precipitation hardening. The steel composition must be

optimised to benefit both from grain refinement and precipitation hardening with Nb. During this

research work it was shown in particular that Ti in combination with Nb was detrimental because TiN

precipitates formed at high temperatures increase the rate of NbC precipitation in the austenite and

change their composition.

300

325

350

375

400

425

450

475

500

525

550

575

600

625

650

7 8 9 10 11 12 13 14 15 16 17 18 19 20

d-1/2

, mm-1/2

TS

, M

Pa

7 8 9 10 11 12 13 14 15 16 17 18 19 20

d, µm

Industrial LC steel

Industrial LC-Nb steel

Hall-Petch relation

Laboratory Nb steels + UFC

TS = 24.9 d-1/2

+ 116

20.4 15.6 12.3 10 8.3 6.9 5.9 5.1 4.4 3.9 3.5 3.1 2.8 2.5

Fig. 131 - Hall-Petch relation from industrially produced steel (Carlam HSM) and laboratory rolled Nb-

steel.

60C-420Mn-60Nb steel (10-3 wt%)

464

548539

590

400410420430440450460470480490500510520530540550560570580590600

Conventional UFC (Carlam)

Hot rolling process

YS

, T

S, M

Pa

YS, MPa

TS, MPa

ASTM 12:

d = 5.4 µm

ASTM 14:

d = 2.8 µmdα ìσppt ê

Fig. 132 - Tensile property improvement using ultra fast cooling between finishing and coiling

compared to conventional (laminar cooling) process.

131

- Cooling rate:

Fig.131 shows that using accelerated cooling between the end of rolling and coiling increased the

precipitation hardening efficiency with similar grain size.

Fig.132 shows an industrial example of the ferrite grain refinement and tensile property improvement in

steel containing 60.10-3

wt%C, 420.10-3

wt%Mn and 60.10-3

wt%Nb, hot rolled and cooled to the

coiling temperature around 650 °C using the conventional (laminar cooling) process and the ultra fast

cooling pilot installation of Carlam. [105].

Accelerated cooling reduces the growth of ferrite grains during cooling, leading to finer microstructure

and prevents NbC precipitation during cooling. This very important result shows that the microalloy

content can be reduced to obtain similar strength level if fast cooling or ultra fast cooling is applied in

place of the conventional (laminar) cooling.

- Coiling temperature:

In all studied steels, decreasing the coiling temperature led to finer ferrite grains, provided ultra fast

cooling is applied. If slow cooling is applied after rolling, the reverse situation is observed, as ferrite

nucleates and grows during the cooling.

- Rolling reduction (strain accumulation):

Increasing the rolling reduction, especially in the last rolling passes cumulates the strain and reduces the

austenite grain size. To have the maximum efficiency, the last rolling reduction must be immediately

followed by fast cooling to avoid static recrystallisation of the austenite.

- Finishing rolling temperature:

This parameter determines the amount of accumulated strain, especially in microalloyed steel when it is

below Tnr and above Ar3. Steels 1134 and 1135, microalloyed with Ti and Nb were hot rolled in 5

passes following the rolling schedule of Fig.118, with different temperatures before the last pass, 900°C,

875 °c and 840 °C, respectively. The measured forces and temperatures are shown in Fig.133. The

rolling forces during the last pass were very close when finishing at 900 °C or 875 °C, likely due to the

smaller interpass time between the last 2 passes above 900 °C. The accumulated strain was very similar

in both cases and few differences in the final ferrite grain size were expected. The higher rolling force

measured during rolling at 840 °C indicates that more strain was accumulated and slightly finer ferrite

grains were expected.

132

600

650

700

750

800

850

900

950

1000

1050

1100

1150

100 110 120 130 140 150 160 170

Time (sec)

Te

mp

era

ture

(°C

)

0102030405060708090100110120130140150160170180190200210220

Ro

llin

g f

orc

e (

t)

Fig. 133 - Temperature and rolling forces measured during hot rolling steels 1134 and 1135 with 3

different finishing rolling temperatures.

Fig. 134 shows the microstructures obtained with the 3 different finishing rolling temperatures. Finer

grains were obtained in steel 1134 with the lowest FRT. It is however important to take into account the

transfer time between the finishing mill and the cooling device. In the course of the present experiments,

due to high reductions, the mill configuration did not allow short delay time before cooling. Therefore,

the benefit of strain accumulation with high reduction at low temperature may be partially lost due to

austenite recrystallisation and NbC precipitation before coolin

840 °C 875 °C 900 °C

1134

1135

3.9 µm 4.8 µm 5.0 µm 5.0 µm 4.8 µm 5.0 µm

Fig. 134 - Microstructures and measured ferrite grain size in steels 1134 and 1135 hot rolled with 3

different FRT.

- Time between the end of rolling and cooling (delay time):

Hot torsion experiments showed strong difference between cooling immediately after deformation or

100 s later. As mentioned before, during hot rolling experiments the transfer time between the last

rolling pass and the cooling device was quite long (between 5 and 10 s). The maximum efficiency of

accumulated strain and soluble Nb on the ferrite grain size was not achieved. To obtain ultra fine ferrite

grains, Nakayama Steel Works have developed a cooling system for strong cooling just after rolling

133

[106]. The delay time seems to be a major parameter for the control of the grain size and is worth being

studied more deeply.

7.3.5 Further processing: cold rolling and annealing

Steel 806, 1134 and 1135 were hot rolled to strips with 4 mm thickness and fast cooled (300 °C/s) to

different coiling temperatures. Tensile specimens with 50 mm gauge length (A50) were machined from

hot rolled strips. Some of these tensile specimens were directly annealed in salt bath for 100 s at

temperatures of 620 °C, 670 °C, 720 °C and 770 °C, respectively. This procedure was applied for the 3

steel compositions in order to estimate the potential in terms of precipitation hardening due to the Nb

remaining in solution after hot rolling.

Hot rolled strips of steels 1134 and 1135 were also cold rolled to 1.5 mm thickness. A50 tensile

specimens were machined and the same thermal treatments were applied.

Fig. 135 shows the tensile properties measured on hot rolled, hot rolled and annealed, and cold rolled

and annealed specimens. In all cases hot rolled strips contained enough Nb, kept in solution by using

ultra fast cooling, to obtain NbC precipitation hardening after annealing.

0

5

10

15

20

25

30

35

450 500 550 600 650 700 750 800

TS, MPa

A5

0,

%

As hot-rolled

Hot rolled and annealed

Cold-rolled and annealed

Fig. 135 – Total Elongation A50 versus TS for hot rolled, hot rolled and annealed, and cold rolled and

annealed specimens.

134

0

2

4

6

8

10

12

14

16

18

20

300 350 400 450 500 550 600 650 700 750 800

YS, MPa

Au,

%

As hot-rolled

Hot rolled and annealed

Cold-rolled and annealed

Fig. 136 – Uniform elongation Au versus YS for hot rolled, hot rolled and annealed, and cold rolled and

annealed specimens.

The uniform elongation (Au) versus yield stress (YS) for the same tensile specimens is reported in Fig.

136. For some specimens, annealing improved the tensile strength but reduced the yield stress. The

reason is that in specimens annealed above 720 °C, some austenite formed and dual phase ferrite +

martensite microstructures were obtained after subsequent cooling to room temperature. The YS/TS

ratio plotted against uniform elongation in Fig. 137 shows clearly both categories of microstructures

obtained after annealing.

0.4

0.5

0.6

0.7

0.8

0.9

1

0 5 10 15 20

Uniform elongation, %

YS

/TS

As Hot rolled

Hot rolled + annealed

Cold rolled + annealed

FERRITE

FERRITE + MARTENSITE

Fig. 137 - YS/TS ratio versus Au plot of hot rolled, hot rolled and annealed, and cold rolled and

annealed specimens.

135

0

50

100

150

200

250

300

-100 -50 0 50 100 150 200 250 300

YS, MPa

TS

, M

Pa

Cold rolled + annealed

Hot rolled + annealed

Dual Phase

Ferrite

(NbC precipitation hardening)

Fig..138 - TS and YS variations after annealing and after cold rolling and annealing hot rolled steels

806, 1134 and 1135.

Fig. 138 is a summary of the increase of TS after annealing, plotted against the variations of YS.

Distinction is clearly made between dual phase microstructures and ferrite microstructures. Specimens

annealed directly after hot rolling exhibit increased YS and TS up to 100 MPa, only due to NbC

precipitation hardening.

0

2

4

6

8

10

12

14

16

18

20

400 420 440 460 480 500 520 540 560 580 600 620 640 660 680 700 720 740 760 780

YS, MPa

Au,

%

30 s

60 s

100 s

200 s

500 s

Fig. 139 - Au versus YS plot for steels 1134 and 1135 cold rolled and annealed at 720 °C for different

times.

136

0

5

10

15

20

25

30

480 500 520 540 560 580 600 620 640 660 680 700 720 740 760 780

TS, MPa

A5

0,

%

30 s

60 s

100 s

200 s

500 s

Fig. 140 - A50 versus TS plot for steels 1134 and 1135 cold rolled and annealed at 720 °C for different

times.

In cold rolled specimens, more than 250 MPa increase of TS and YS was measured as a consequence of

NbC precipitation hardening and grain refinement by cold deformation. It must be noted that in this last

type of material, the microstructure was not fully recrystallised. As a consequence the uniform

elongation may be lower than 10 %.

Fig. 139 and Fig. 140 show the tensile properties of steels 1134 and 1135 cold rolled and annealed at

720 °C for different times. A large range of tensile properties was obtained by varying the annealing

parameters.

These very promising results obtained after combining fine grains and NbC precipitation hardening are

worth being studied further, with focus on Nb quantification and ferrite recrystallisation.

7.4. CONCLUSIONS

Dilatometry and hot torsion tests were performed on Nb-microalloyed low C steels. The results are

summarised as follows.

In dilatometric samples (no deformation):

- increasing tp at 900 °C increases the rate of γ−α transformation, finer ferrite grains are

obtained due to more nucleation sites;

- increasing the initial Nb content, with the same C content, increases the number of fine NbC

precipitates, leading to finer ferrite grains.

- grain refinement is due to NbC static precipitation in austenite: nucleation sites.

In torsion samples (hot rolling simulation):

- increasing tp after the deformation at 900 °C leads to more recrystallised austenite, producing

coarser ferrite grains;

- increasing the cooling rate reduces the growth of ferrite, producing finer ferrite grains;

- grain refinement is favoured when Nb remains in solution before the γ−α transformation.

In Nb microalloyed steels, there is a compromise to find between:



137



To obtain fine grains in hot rolled Nb steel, all the Nb must be dissolved at 1250 °C, strain

accumulation must occur below Tnr and above Ar3 to induce dynamic recrystallisation of deformed

austenite grains. To reduce the NbC precipitation during cooling, and also the austenite static

recrystallisation, the last pass must be followed by fast cooling to the γ−α transformation temperature.

Hot rolling simulations were performed on steels microalloyed with Nb, Ti, B and P. Different rolling

schedules were applied to increase the rolling reduction in the last passes. Ultra fast cooling was applied

between the end of rolling and the coiling in order to keep Nb in solution to obtain the maximum

precipitation hardening efficiency.

Fully ferritic microstructures with grain size between 3 and 6 µm were obtained. The finest grains were

obtained with the low C steel 806, microalloyed with Nb without very high rolling reduction in the last

pass. Ti was found to reduce the beneficial effect of Nb on both grain refinement and precipitation

hardening.

The main hot rolling parameters that affect the ferrite grain size are listed below:

- Microalloy content: Nb must be in solution to refine efficently the grain;

- Cooling rate: must be high to avoid ferrite grain growth;

- Coiling temperature: must be low enough to avoid ferrite grain growth;

- Rolling reduction (strain accumulation): must be high to cumulate strain;

- Finishing rolling temperature: must be low (above Ar3) to cumulate strain;

- Time between the end of rolling and cooling (delay time): must be as short as possible to avoid

austenite recrystallisation.

Finally, hot rolled strips were cold rolled and annealed to study the benefit of fine grains and NbC

precipitation hardening. The very promising results show that high yield stress (> 600 MPa) and high

tensile strength fully ferritic steels can be obtained after cold rolling and annealing low C microalloyed

steel, provided ultra fast cooling has been applied after hot rolling.

138

8. ULTRAFINE HIGH C STEELS

8.1 Introduction

The objectives of this work package (WP4) within the overall project are to define suitable thermo-

mechanical processing conditions to produce ultra-fine grain microstructures in high carbon steels and

assess the microstructural evolution and mechanical properties of such materials.

The transformation behaviour and initial thermo-mechanical processing conditions to produce ultra-fine

microstructures have been determined using small-scale thermo-mechanical (Gleeble) simulations [113-

115]. This indicated that deformation was required below a temperature of ~650°C. These processing

conditions have been applied to rolling schedules to assess the repeatability on a larger scale and also to

enable further investigation into the structure-property relationships. The results of the application of

the thermo-mechanical processing conditions to the laboratory scale rolling trials are presented. The

influence of deformation temperature and strain on the resultant microstructure and tensile properties

was also assessed.

As a comparison, three further high carbon steels have been cast with a microalloying addition of

vanadium. It is envisaged that this addition will influence the microstructure and properties,

particularly the formation of fine cementite particles [116].

The use of Equal Channel Angular Extrusion (ECAE) has also been investigated within this work

package, as a comparative means to conventional rolling, of imparting severe plastic deformation to

produce an ultra-fine microstructure. Since its inception in the former Soviet Union in 1972 by Segal et

al [111] there has been extensive research into the relatively simple technique, the majority of which

has taken place in the 1990’s [e.g. 112-116], coinciding with an escalation in the interest in ultra-fine

grained materials. Due to the repeatable nature of the extrusion process, the technique is capable of

introducing very large amounts of strain, without fracture and with no overall change in dimensions.

The equipment used for this part of the work package was based at Manchester Materials Science

Centre (MMSC). The collaboration with MMSC enabled access to not only the ECAE equipment but

also world leading expertise [117-123]. This report evaluates the microstructural evolution during

ECAE and presents results obtained using High Resolution Electron Back Scattered Diffraction (EBSD).

8.2 Experimental

In accordance with the objectives, four high carbon (0.6 – 1.6 wt. %) 60 kg vacuum laboratory melts

were produced previously during the project with a base composition of 0.2 wt. % silicon and 0.7 wt. %

manganese. The compositions are given in Table 10. Such compositions were selected because previous

superplasticity work has demonstrated that ultra-fine structures can be achieved in these types of alloys,

when processed appropriately [124, 125].

C Si Mn P S Al N

0.61 0.20 0.69 0.016 0.005 0.031 0.0057

0.81 0.20 0.70 0.015 <0.005 0.031 0.0062

1.22 0.20 0.69 0.016 <0.005 0.030 0.0069

1.61 0.20 0.69 0.015 <0.005 0.029 0.0064

Table 10: Compositions (wt. %) of high carbon steels

The ingots were furnace reheated to 1250°C, held at temperature for four hours and then rolled to

75mm plate primarily to reduce segregation and consolidate porosity.

Samples for dilatometry were machined from each plate and transformation characteristics assessed in a

Bahr 105 dilatometer. The dilatometer samples were heated rapidly to 1150°C, held for 5 minutes and

then cooled to room temperature using cooling rates in the range of 64 to 0.25°C/s (800-500°C).

Transformation start and finish temperatures were determined by measuring the sample dilation.

139

Samples were prepared for metallographic examination and Vickers hardness values (HV30) were

determined according to BS EN ISO 6507-1.

Cylindrical samples (10mm long and 6mm) in diameter were prepared for uni-axial compression testing

on a Gleeble 1500 thermo-mechanical simulator. The ends of each sample were lubricated with 5 layers

of graphite to reduce friction. The samples were rapidly heated to 1150°C, held for 5 minutes and then

cooled to deformation temperatures in the range 700-500°C at a cooling rate of 10°C/s. The temperature

range and cooling rate were considered likely to generate ultra-fine structures, based on the dilatometry

results. After holding at deformation temperature for 15 seconds to equilibrate temperature within the

sample, the cylinder was deformed to a final length, equating to mean strains of 0.36, 0.72 and 1.2 at a

mean strain rate of 21/s. Immediately following deformation, cooling rates of 1°C/s, 10°C/s and

200°C/s were applied. Deformed samples were then sectioned, mechanically polished and etched in

picral and Villella’s reagent for metallographic examination by Scanning Electron Microscopy (SEM).

Cylindrical billets, 100mm x 15mm diameter, were machined from the plates in the rolling direction

and processed by Equal Channel Angular Extrusion (ECAE). The ECAE equipment was designed for

the processing of non-ferrous alloys, predominantly aluminium. Due to the large difference in material

flow stress values between that of steel and aluminium, processing was carried out with the dies heated

to a constant 500°C. Graphite lubrication was used and the billets were heated at 50°C/min and held at

temperature for 15 minutes prior to processing. Up to 10 passes were completed with no billet rotation

between extrusion cycles (Route A). A circular cross-section ECAE die was used with a die angle of

120°, giving a nominal von Mises’ effective strain of εvm ~ 0.67 per pass. Full details of the ECAE

equipment and deformation behaviour of the die can be found elsewhere [126].

The remainder of the rolled plates were cut into small blocks (150x75x75mm) for further rolling

experiments. Feedstock samples of the 0.6, 0.8 and 1.2wt.%C steels were furnace reheated and

homogenised at 1150°C. Samples were then rolled to either a plate thickness of ~62 or 50 mm, with an

aim temperature of 900°C, in preparation for strains of ~1.2 or ~2 (respectively) to be introduced during

finish rolling. The samples were then air-cooled until the temperature reached the final deformation

temperature, i.e. 700°C or 600°C before several finish roll passes were applied. The aim final thickness

of the plate was ~7mm, however the actual thickness was closer to ~9mm. Details of the thermo-

mechanical cycle are given in Fig. 141. It should be noted that it was not possible to complete rolling of

the 1.2wt.%C steel at a deformation temperature of 600°C to a strain of ~2, as the load approached the

maximum capability of the mill.

Fig. 141 - Thermomechanical processing cycle used in the rolling trials

6-12 finishing passes, i ~ 1.2/2

700°C

6-12 finishing passes, i ~ 1.2/2

600°C

Temperature

Time

1150°C

900°C

Ar3

Air cool

Roughing passes

140

The deformation structures of the ECAE processed billets and select rolled plates, were characterised at

the centre of the ND-ED plane (die symmetry plane) and RD-ND plane respectively, using SEM

metallography and high resolution Electron Back Scattered Diffraction (EBSD). SEM samples were

etched in a Picral and Villela’s solution for approximately 90 seconds to reveal the carbide structure

before secondary electron imaging in a Philips XL40 SEM. EBSD samples were mechanically polished

and given a light electropolish in 92% acetic and 8% perchloric acid solution, at ~70 Volts for

approximately 3 seconds. A Philips XL30 FEG-SEM and a Camscan Maxim 2040s FEG-SEM, both

fitted with automated HKL-EBSD pattern collection systems with a spatial resolution of ~50nm and

angular resolution of ~1°, were used for data acquisition. Data analysis was carried out using VMap, a

software package developed at MMSC by Prof. F. John Humphreys. A minimum misorientation cut-off

of 1.5° was used in all cases to eliminate excessive misorientation noise. High angle grain boundaries

(HAGBs) are defined as being equal to, or greater than 15° in misorientation and low angle grain

boundaries (LAGBs), a misorientation less than 15°.

Tensile properties were determined for both rolled and ECAE deformed material. Two tensile test

pieces were machined from the centre of each rolled plate and from each ECAE processed billet

(following longitudinal EDM wire cutting). Testing was carried out in accordance with BS EN 10002.

8.3 Results and discussion

8.3.1 Dilatometry

Continuous cooling transformation (CCT) diagrams were produced for all four experimental steels in

order to determine suitable cooling rates and temperatures for deformation experiments. The CCT

diagrams for the four steels are shown in Figs. 1422 to 145. Hardness measurements and

microstructural evaluation was carried out for each sample given in Table 11.

For the 0.6wt.% carbon steel, metallography indicated that the microstructure consisted of pearlite and a

small amount of ferrite for cooling rates of 10°C/s and below, and pearlite with increasing amounts of

martensite for cooling rates of 30°C/s and above. At a cooling rate of 10C/s, the transformation start

and finish temperatures were 620 and 600°C respectively. Very similar observations for the 0.8wt.%

carbon steel were made.

For the 1.2wt.% carbon steel, the microstructure consisted of pearlite and cementite at cooling rates of

9°C/s and below and pearlite with increasing amounts of martensite for cooling rates in excess of 9°C/s.

The transformation start and finish temperatures were 620 and 615°C respectively. Similar

microstructures were observed in the 1.6wt.% carbon steel, including a small amount of graphite.

Fig. 142 - CCT diagram for 0.6wt.%C steel

141



142

Fig. 145 - CCT diagrams for 1.2wt%C (a) and 1.6wt%C steels (b)

143

M=Martensite, P=Pearlite, F=Ferrite, Fe3C=Cementite, G=Graphite

Table 11 - Hardness and microstructure of dilatometer samples

Cooling rate, °C/s

(800-500°C)

Vickers Hardness

(HV30)

Microstructure

0.6wt.% Carbon steel

63.8 808 M+P

29.7 651 M+P

9.6 304 P+F

6.4 285 P+F

4 284 P+F

2 289 P+F

1 249 P+F

0.5 249 P+F

0.25 247 P+F


63.8 849 M+P

29.4 677 M+P

9.4 368 P+F

6.2 347 P+F

4 366 P+F

2 355 P+F

1 309 P+F

0.5 311 P+F

0.25 307 P+F


63.8 727 M+P

19.7 526 P+M

9.3 458 P+Fe3C

6.3 437 P+Fe3C

4 434 P+Fe3C

2 415 P+Fe3C

1 357 P+Fe3C

0.5 347 P+Fe3C

0.25 312 P+Fe3C


62.5 500 M+P+G

20.4 453 P+Fe3C+M+G

10 415 P+Fe3C+G

8 405 P+Fe3C+G

4 379 P+Fe3C+G

2 358 P+Fe3C+G

1 318 P+Fe3C+G

0.5 305 P+Fe3C+G

0.25 276 P+Fe3C+G

144

8.3.2 Gleeble

Hot compression testing was carried out using a Gleeble testing machine to establish the deformation

conditions, including the minimum strain and temperature, required to generate ultra-fine ferrite and

carbide structures. Fig. 146 illustrates the influence of strain on the microstructure of 0.8wt.%C steel

deformed at 550°C. It was apparent that spheroidisation of the pearlitic cementite was only partially

complete following strains of 0.36 (Fig. 146(b)) and 0.72 (Fig. 146(c)) at 550°C. A similar behaviour

was observed at deformation temperatures of 500 and 600°C in each of the three steels. By comparison,

a relatively homogeneous and spheroidised carbide structure was observed following an applied strain

of 1.2. This suggested that mean strains in excess of 0.72 were required to promote pearlite

spheroidisation and produce a ‘composite’ structure of ultra-fine ferrite and carbide aggregate (UFF-

CA) at deformation temperatures below the transformation temperature (~600°C). Similar values of

required strain were noted by Niikura et al [127] in heavy deformation experiments on grain refinement

in plain C-Mn-Si steel. The strain distribution at various positions in a deformed Gleeble sample was

estimated from a Finite Element Analysis (FEA) using the Abaqus© software package. Analysis

showed that when deforming the 10 x 6mm samples, consistent with inducing a mean average strain of

1.2, the strain is low in areas constrained by friction, i.e. near the sample surface. In contrast, the centre

of the sample experienced a strain approximately 2.3 times greater than the average strain. As a

consequence, rolling schedules to induce finishing strains of ~1.2 and 2 were used to try to induce

similar microstructures.

(a) (b)

(c) (d)

Fig. 146 – Influence of strain on the microstructure of 0.8wt.% carbon steel deformed at 550°C (a) 0

strain, (b) 0.36 strain, (c) 0.72 strain and (d) 1.2 strain.

145

(a) 200°C/s (b) 1°C/s

Fig. 147 - Influence of cooling rate, (a) 200°C/s, (b) 1°C/s, on microstructure of 0.8wt% carbon steels

after deformation to a strain of ~1.2 at 650°C.

(a) 200°C/s (b) 10°C/s

(c) 1°C/s

Fig. 148 - Influence of cooling rate, (a) 200°C/s, (b) 10°C/s, (c) 1°C/s on microstructure of 0.8wt%

carbon steels after deformation to a strain of ~1.2 at 500°C.

The influence of cooling rate following deformation on the microstructure was determined for each

steel. Samples were deformed to a mean true strain of 1.2 at either 650 or 500°C and then immediately

cooled at 200, 10 or 1°C/s. Example microstructures are shown in Figs. 147 (650°C) and 148 (500°C).

Irrespective of carbon content, it was evident that extensive carbide spheroidisation was present even

after very rapid cooling (200°C/s), this suggested that the spheroidisation of pearlitic carbides was an

extremely quick process. Evidence in support of this can be found in the literature [128] which states

that the spheroidisation process can occur in less than 1 second. Changes in cooling rate, following

deformation at 500°C (Fig. 148(a-c)), had no significant effect upon the microstructure. However,

146

following deformation at 650°C (Fig. 147(b)) the slowest cooling rate (1°C/s), led to significantly

coarser carbides.

Influence of multiple deformation on development of microstructure

The hot compression testing described thus far has used a single burst of deformation. To study the

influence of multiple deformation on microstructural development, two bursts of deformation were

introduced, both corresponding to a mean true strain of 0.6, with a 10 second interval between them.

Deformation was carried out at 650 and 550°C and the samples were cooled at 10°C/s following

deformation.

An example of the microstructures produced after a single burst and multiple burst of deformation at

650°C for the 0.6wt.% carbon steel is shown in Fig. 150. It can be seen that pearlite spheroidisation was

only partial after a strain of 0.6 Previous study had revealed that spheroidisation was largely complete

after a strain of 1.2 (e.g. Figs. 146(d), 147 and 148). After multiple deformation a uniform and highly

spheroidised distribution of carbides is observed (Fig. 149 (b)). It is considered that the greater time

involved in multiple deformation (10 second interval between deformation 'hits') has promoted a more

uniform and spheroidised structure, with slightly larger carbides. Fig. 149(c) has been etched to reveal

the boundaries after multiple deformation.

(a) Strain 0.6 (b) Strain 0.6 + 0.6

(c) Strain 0.6 + 0.6, etched to reveal boundaries

Fig. 149– (a-c) Comparison of single and multiple deformation for 0.6wt.% carbon steel deformed at

650°C

147

8.3.3 Rolling Tests

Deformation by warm rolling to conventional strains

Influence of deformation and temperature

Schedules for hot rolling were devised based upon the results of the Gleeble simulations. Deformation

temperatures of 700 and 600°C and finishing strains of ~1.2 and 2 were implemented with a cooling

rate of approximately <5°C/s, using the schedule defined in Fig. 141. The hardness was found to be

approximately 30HV (5kg) higher, for each composition, when the deformation temperature was

reduced from 700°C to 600°C. Consistent material hardness was observed through the thickness of all

the rolled plates, e.g. 0.8wt.%C steel, Fig. 150.

240

250

260

270

280

290

300

0 1 2 3 4 5 6 7 8 9 10


Ha

rdn

ess

HV

(5k

g)

T Defmtn 700°C, i=1.2

T Defmtn 600°C, i=1.2

T Defmtn 700°C, i=2

T Defmtn 600°C, i=2

Fig. 150 - Through thickness hardness profile for 0.8wt.%C steel following rolling trials.

Electron micrographs that illustrate the structures formed in the rolled 0.6, 0.8 and 1.2wt.%C steel are

shown in Figs. 151-153. After deformation at 700°C (above the austenite to pearlite transformation

temperature), a typical pearlite microstructure was produced (e.g. Figs. 151-153(a) and (c)) in all cases.

Substantial areas of grain boundary ferrite were evident in the 0.6wt.%C steel. The 0.8wt.%C steel was

characterised by a fully pearlitic microstructure and increasing the carbon content to 1.2wt.%C led to

the formation of cementite laths at the prior austenite grain boundaries. Figs 151, 152 and 153 (a+c)

show the microstructure present at mid-thickness of the final gauge 0.6, 0.8 and 1.2wt.%C material,

illustrating a pearlitic transformation following finish rolling at 700°C.

By contrast, a partially spheroidised microstructure was often found at sub-surface regions in each of

the steels, e.g. Figs. 151, 152 and 153(b), particularly in the lower strained samples (i.e. ~1.2). This is

attributed to accelerated local cooling, possibly due to roll chill, which has lead to transformation during

deformation, and/or discrete strain localisation. Small ‘pockets’ of lamellar cementite, thought to be due

to localised temperature differences, were also apparent in the mid-thickness of a number of samples

deformed at 600°C.

148

Subsurface Mid-thickness

(a) 700°C, ε ~ 1.2

(b) 600°C, ε ~ 1.2

(c) 700°C, ε ~ 2

(d) 600°C, ε ~ 2

Fig. 151- (a-d) SEM micrographs illustrating the carbide structure present in the 0.6 wt.% C steel

following the rolling trials. Deformation temperature and finishing strain are given.

149


(a) 700°C, ε ~ 1.2

(b) 600°C, ε ~ 1.2

(b) 700°C, ε ~ 2

(d) 600°C, ε ~ 2

Fig. 152 - (a-d) SEM micrographs illustrating the carbide structure present in the 0.8 wt.% C steel


150


(a) 700°C, ε ~ 1.2

(c) 600°C, ε ~ 1.2

(d)

(b) 700°C, ε ~ 2

(c)

Fig. 153 - (a-c) SEM micrographs illustrating the carbide structure present in the 1.2 wt.% C steel


151

In general, an ultra-fine ferrite and spheroidised carbide structure was evident following deformation at

600°C to a strain of ~2 in all compositions (Figs. 151(d), 152(d), 154(b) and 155(b)). The samples

shown in Figs. 154-156 were etched in Marshall’s reagent to reveal the subgrain/grain boundaries. A

‘banded’ structure was observed in the 0.6wt.%C steel (Fig. 154), consisting of large pro-eutectoid

ferrite grains separated by more equiaxed and refined ultra-fine ferrite and carbide aggregate. A more

homogeneous ferrite microstructure was present in the higher carbon steels (0.8 and 1.2wt.%C, Figs.

155 and 156 respectively) after deformation at 600°C. Assessment of samples rolled at 600°C, and

preferentially etched to reveal the boundary structure, suggested that there was little subsequent

refinement in the ferrite grains size as a result of increased deformation strain.


(a) 600°C, ε ~ 1.2


(b) 600°C, ε ~ 2

Fig. 154 - (a-b) SEM micrographs illustrating the grain structure present in the 0.6 wt. % C steel


152


(a) 600°C, ε ~ 1.2


(b) 600°C, ε ~ 2





following deformation at 600°C, to strains of ε ~ 1.2.

High resolution EBSD was carried out to determine the nature of the boundaries observed in the ferrite

matrix in the rolled sample, and corresponding statistics. Initial attempts were made to index both the

ferrite and cementite volumes. Unfortunately this proved ultimately unsuccessful, mainly due to

differential etching of the ‘soft’ ferrite matrix in comparison with the ‘hard’ cementite particles which

resulted in a ‘shadowing’ effect in addition to problems with accurate pattern indexing of the Fe3C. An

example EBSD Euler angle map for the 0.6wt.%C steel rolled at 600°C to a strain of ~2 is given in Fig.

153

154. A fine step size of 0.1µm was used so that accurate subgrain information could be obtained (Fig.

154).

Fig. 157 shows that after deformation by controlled rolling the ferrite did not have what could be

described as an ultra-fine grain structure, and predominantly contained elongated fibrous grains with

subgrains. Some relatively large grains containing low angle grain boundaries (LAGBs) were evident,

which were considered to correspond to areas of pro-eutectoid ferrite (Fig. 157). Some bands of finer,

ferrite grains were observed adjacent to these areas, aligned in the rolling direction (Fig. 157), formed

as a consequence of the pearlitic spheroidisation. EBSD statistical boundary data confirmed that an

ultra-fine structure consisting of both low angle and high angle grain boundaries was achieved

following deformation at 600°C to a strain of ~2. The majority of the boundaries present were low

angle; between 1.5 and 15° misorientation (~60%, see Table 12) and approximately 40% could be

described as high angle grain boundaries (HAGBs) i.e. in excess of 15° misorientation. The average

grain boundary spacing was 2.19µm and subgrain spacing 1.14µm. However, it should be noted that

>75% of the boundaries had a misorientation in excess of 5°, suggesting that the sub-structure was well

developed.

(a) (b)

Fig. 157 - EBSD band contrast (a) and orientation (b) maps of 0.6wt.%C steel following rolling at

600°C, ε~2.

Intercept Length (µm) Boundaries (%) Steel

(C wt%) HAGB HAGB X HAGB Y LAGB HAGB LAGB

0.6 2.19 1.20 3.17 1.14 42 58

Table 12 - EBSD statistical data for 0.6wt%C steel rolled at 600°C to a strain of ~2

8.3.4 Mechanical properties

Following the rolling trials at 600°C and 700°C, tensile properties were determined for each of the steel

compositions. The results are given in Table 13 and Figs.158-160 show plots of the results.

LAGBs

HAGBs

TD

RD

154

Carbon

(wt%)

0.2%

PS (MPa)

0.5%

PS (MPa)

UTS

(MPa)

Total Elong.

(%)

R of A

(%)

0.2%PS:UTS

ratio

Normal rolling 0.6 351 386 743 18 33 0.47

TDef 700, Strain=1.2 0.6 533 527 761 26 55 0.70

TDef 600, Strain=1.2 0.6 670 672 831 24 54 0.81

TDef 700, Strain=2 0.6 488 495 752 17 51 0.65

TDef 600, Strain=2 0.6 636 636 777 21 49 0.82

Normal rolling 0.8 418 475 878 13 11 0.48

TDef 700, Strain=1.2 0.8 472 500 902 14 39 0.52

TDef 600, Strain=1.2 0.8 724 728 911 18 38 0.79

TDef 700, Strain=2 0.8 466 496 897 20 42 0.52

TDef 600, Strain=2 0.8 750 750 891 18 43 0.84

Normal rolling 1.2 385 419 860 10 7 0.45

TDef 700, Strain=1.2 1.2 750 749 886 10 30 0.85

TDef 600, Strain=1.2 1.2 850 858 1045 17 34 0.81

TDef 700, Strain=2 1.2 694 682 857 14 35 0.81

Table 13 – Tensile properties following rolling trials with finishing rolling temperatures of 600°C and

700°C and strains of ~1.2 and 2.

0

200

400

600

800

1000

1200

Normal

rolling

TDef 700,

Strain=1.2

TDef 600,

Strain=1.2

TDef 700,

Strain=2

TDef 600,

Strain=2

Ten

sile

Str

eng

th (

MP

a)

0.2%PS

0.5%PS

UTS

Fig. 158 - Tensile properties for 0.6wt.%C steel following rolling trials

155

0

200

400

600

800

1000

1200

Normal

rolling

TDef 700,

Strain=1.2

TDef 600,

Strain=1.2

TDef 700,

Strain=2

TDef 600,

Strain=2

Ten

sile

Str

eng

th (

MP

a)

0.2%PS

0.5%PS

UTS


0

200

400

600

800

1000

1200

Normal rolling TDef 700,

Strain=1.2

TDef 600,

Strain=1.2

TDef 700,

Strain=2

Ten

sile

Str

eng

th (

MP

a)

0.2%PS

0.5%PS

UTS


In all cases a marked increase in proof stress (PS) was observed, particularly at deformation

temperatures of 600°C (Fig. 145). In general, UTS values were greater after deformation at 600°C in

comparison to the ‘Normal’ material (rolled at 1050°C) and 700°C (Figs. 158-160). No obvious

relationship was present between deformation and elongation in any of the steels. It should be noted that

samples deformed at 600°C exhibited higher PS:UTS ratios, however unstable yielding was not evident

and a substantial increment of work hardening was retained in contrast to many other ultra-fine

‘grained’ materials [129, 130].

156

8.3.5 Laser Welding

Preliminary bead-on-plate laser welding trials were carried out on four plain 0.8wt. %C steel plates,

following warm rolling to conventional strains. Welding was carried out using a Ferranti Photonics

AF5kW CO2 with the following parameters: 4kW laser power; 0.5m/min welding speed; beam focused

4mm into plate; 0.6mm focus spot diameter. The aim of the small weld study was to investigate the

laser weldability of such high carbon steels, assess its repeatability and in particular, evaluate the effect

of the weld process on ultra-fine microstructures. Laser welding was selected as the joining technique to

minimise the incident area, resultant welded volume of material and subsequent Heat Affected Zone

(HAZ). It was considered that laser welding would result in minimum disruption to the ultra-fine

structure developed during rolling. A full evaluation was not possible within the time constraints of the

study but successful welding and full penetration, with minimal HAZ in comparison to conventional arc

welding was achieved in each of the four cases. Examples are given in Figs. 161 and 162. It was

accepted however, that a substantial post weld heat treatment would be required to ensure material

performance.

Fig. 161 - Weld side of steel plate following laser welding

Fig. 162 - Penetration side of steel plate following laser welding

157

8.4 Equal Channel Angular Extrusion

ECAE processing was used to investigate the advantages of a shear deformation mode and to study

the effect of increasing the deformation strain beyond that achievable by rolling. In the work

presented here, the maximum deformation temperature was limited to 500°C by the design of the

equipment. The MMSC ECAE equipment and channel die are shown in Fig. 163(a and b)

respectively.

(a) (b)

Fig. 163 - (a and b). Image illustrating the ECAE apparatus (a) and half of the 120° channel die (b)

A simple schematic illustrating the ECAE process is shown in Fig. 164. Essentially, a hydraulic ram

forces a billet shaped volume of metal through the angle in the channel die.

Ram

Billet

Die

φ

Fig. 164 - Schematic representation of the ECAE process, showing a cross section of a deformed billet,

the ram, the die set and the shear plane of the apparatus [131]

158

Microstructural Evolution

Representative SEM images showing the microstructures of the three steels (0.6, 0.8 and 1.2 wt.%C),

processed using ECAE at 500°C and air-cooled are shown in Fig. 165. Following one ECAE pass,

evidence of shear banding and deformation of the pearlite lamellar was apparent in all three steel

compositions (Fig. 165(i)). This shear banding was particularly marked in the 0.6 wt.%C steel (Fig.

165(i)) where a large pearlite colony had been sheared into a number of smaller ones. After 2 ECAE

passes (Fig. 165(ii)), the increasing strain led to further break up of large pearlite lamellar colonies and

extensive shear banding is present. This was particularly noticeable in the 0.8wt.%C steel (Fig. 165(ii)).

Evidence of the onset of break-up and spheroidisation of the lamellar cementite is shown for the 0.6

wt.%C steel (Fig. 165(ii)). Spheroidisation was also observed in the 1.2 wt.%C steel following 2 ECAE

passes (equivalent to εvm~2) (Fig. 165(ii)). Spheroidisation continued in the 0.6 and 0.8 wt. % C steels

with increasing strain following 3 ECAE passes (Fig. 165(iii)). Further ECAE deformation to 4 passes

resulted in continued spheroidisation in all steels (Fig. 165 (iv)). Fully spheroidal carbides were clearly

visible in the 0.6 and 0.8wt.%C steels (Fig. 165 (iv)) following deformation. Heterogeneous flow of the

matrix and lamellar cementite around large primary cementite particles was also observed in the 1.2 wt.

% C steel (Fig. 165(iv)). On completion of 5 ECAE passes, an accumulation of deformation artefacts

was evident in the 1.2wt.%C steel (Fig. 165 (v)); i.e. deformed lamellar flow around large, ‘hard’

cementite particles and also shear banding. A mix of deformed pearlite and spheroidised carbides was

observed in the 0.6 and 0.8wt.%C steels (Fig. 165 (v)), suggesting that the ECAE deformation was

relatively inhomogeneous.

Similar microstructures to those shown in this report (Fig. 165) have been observed by other workers

following multi-pass rolling (6, 7). Spheroidisation and partial dissolution of pearlite has also been

observed following wire drawing at low temperatures. It is well known that spheroidisation is

accelerated by the break-up of the lamellar, which increases the surface area and the site density for

Gibbs-Thompson capillary driven diffusion. Dissolution has been attributed to either dislocation cores

being a more energetically favourable site for carbon atoms than cementite, or that the surface energy of

cementite is increased by dislocation interaction such that the cementite becomes thermodynamically

unstable [132, 133]. Shin et al [110] reported similar findings relating to pearlite partial dissolution and

spheroidisation following ECAE of low carbon steel at temperatures of 350°C.

Example SEM images following severe deformation, preferentially etched to illustrate the boundary

structure are shown in Fig. 166. The samples have been pre-etched in 2% Nital then etched in Marshalls

reagent for approximately 2 minutes. Secondary electron imaging is used at a small working distance

and high operating voltage (KeV). The images show that there was a submicron (<1µm) structure

present following severe ECAE deformation (Fig. 149). However, it is unclear whether the observed

boundaries and differences in orientation contrast correspond to discrete changes in misorientation of a

‘true’ grain structure or that of a refined subgrain structure.

Microstructural analysis using EBSD

High resolution (0.1µm step size) EBSD was carried out to determine the extent of grain refinement in

the 0.8wt.%C steel with increasing strain. Fig. 167 shows example Euler orientation maps that illustrate

the microstructural evolution for the 0.8wt.%C steel. Statistical boundary data is given in Table 12. As

the strain increased, a high density of HAGBs developed leading to a submicron HAGB spacing by a

strain of 3.3 (see Figs 150(a) – (c)) and Table 14). Considerable refinement was evident after 3 passes

(Fig. 168), with ‘ribbons’ of refined ferrite grains forming in the extrusion direction, containing well

defined subgrains. With increased deformation, the HA and LAGB spacing continued to reduce and the

structure became more equiaxed in appearance (Fig. 169 and Table 14). However, the EBSD maps

(Figs. 167–169) also indicated that this ‘refinement’ was quite heterogeneous in nature and that the

structure only became relatively homogeneous following large deformation strains (i.e εvm ~ 3.33).

Following a strain of εvm ~ 3.33 (5 passes), a sub-micron grain size was observed (~0.7µm) and the

HAGB content was approximately 60%. A plot illustrating the reduction of HAGB spacing and increase

in HAGB percentage area with increasing deformation is shown in Fig. 170. Apps et al [123] have

observed similar findings during the refinement of a particle containing aluminium alloy.

159

0.6 wt. % C 0.8 wt. % C 1.2 wt. % C

(i) (i)

(ii)

(i)

(ii)(ii)

(iii) (iii) (iii)

(iv) (iv)(iv)

(v) (v) (v)

Fig. 165 - SEM micrographs illustrating the microstructural evolution during ECAE processing (1 (i), 2

(ii), 3 (iii), 4 (iv) and 5 (v) passes) of 0.6, 0.8 and 1.2 wt. % C steels.

160

(a)

(b)

(c) (d)

Fig. 166 - (a+b). SEM micrographs etched to highlight the subgrain/grain boundaries of 0.8 wt. % C

and 1.2 wt. % C (c+d) steel samples following 4 ECAE passes

Steel Intercept Length (µm) Boundaries (%)

(C wt%)

No. of

ECAE

Passes

HAGB

(mean)

HAGB X HAGB Y LAGB HAGB LAGB

0.8 1 2.98 2.53 3.42 0.74 23 77

0.8 3 1.41 1.04 1.89 0.79 40 60

0.8 5 0.66 0.55 0.77 0.86 58 42

0.6 5 1.81 1.57 2.05 0.68 31 69

1.2 5 1.02 0.92 1.11 0.89 48 52

Table 14 – Comparison of the average statistical boundary data for the 0.6, 0.8 and 1.2wt.%C steel

obtained from EBSD maps at different levels of ECAE deformation.

161

Deformed pearlite lamellar

Fig. 167 - Sample B1, 0.8wt%C steel after 1 ECAE pass. Suspect points have been removed and the


Large areas of LAGBs ultra-fine ferrite

Fig. 168 - Sample B3, 0.8wt%C steel after 3 ECAE passes. Suspect points have been removed and the


Similar EBSD analysis was carried out on the 0.6 (Fig. 171) and 1.2wt.%C (Fig. 172) steels following

ECAE deformation to 5 passes (εvm~3.33). Statistical boundary data from the EBSD measurements is

given in Table 14. A refined substructure was evident in the 0.6wt.%C steel after 5 passes, yet there was

little HAGB content ~30%, particularly in comparison to the 0.8wt.%C steel (~60%) following the

same amount of deformation (see Table 14).

It is thought that this lack of HAGB generation and refinement may be due to the majority of the initial

deformation being concentrated in the areas of primary ferrite rather than the pearlite lamellar areas,

leaving a relatively large HAGB spacing or grain size. The boundary statistics for the 1.2wt.%C steel

after 5 ECAE passes (Table 14) also suggest that refinement proceeded at a ‘slower’ pace in

comparison to the 0.8wt%C steel (compare Table 14). No evidence for accelerated, homogeneous

refinement due to large cementite particles was noted, excluding the obvious matrix refinement

observed in the immediate vicinity of said particles.

162

Some large areas of LAGBs remain. Structure is generally ultra-fine

Fig. 169 - Sample B5, 0.8wt%C steel after 5 ECAE passes. Suspect points have been removed and the


0

1

2

3

4

0 1 2 3 4Strain (von Mises)

Boundar

y S

pac

ing (

µm

)

0

20

40

60

80

100

0 1 2 3 4 5 6Number of ECAE passes

Per

centa

ge

HA

GB

(%

)

Grain Width

% HAGB

Fig. 170 - Plot of EBSD statistical data showing the average transverse high angle grain boundary

spacing and high angle grain boundary area fraction (%HAGBs) in the 0.8wt.%C steel, as a function of

strain during ECAE processing.

163

Large areas of LAGBs exhibiting ultra-fine substructure

Fig. 171 - Sample A5, 0.6wt%C steel after 1 ECAE pass. Suspect points have been removed and the


Large cementite particles area of deformed lamellar and

microstructural ‘flow’ Heterogeneous ferrite refinement

Fig. 172 - Sample C5, 1.2wt%C steel after 5 ECAE passes. Suspect points have been removed and the



After ECAE processing, similar tensile behaviour was observed in all the steels as a function of strain

(number of ECAE passes) as illustrated in Figs. 173-175. Values of 0.2% proof stress (PS), UTS and

elongation were recorded for each steel. In each case, a large increase in proof stress and a lesser

increase in UTS was observed following the first ECAE pass (Figs. 173-175). Subsequent passes

resulted in a slower increase in proof stress and plateau in UTS values with the two values tending to

164

converge at high levels of deformation, e.g. proof stress and UTS following five passes for 0.8wt.%C

steel are 1070MPa and 1105MPa respectively (Fig. 174). Further increase in strength was observed

with increasing strain and a second plateau in tensile strength was reached following seven passes.

Proof stress and UTS values were essentially identical after ten passes (εvm~6.67), i.e. 1239MPa and

1242MPa respectively (Fig. 174). Elongation values tended to show more scatter, but in general a

decrease with increasing deformation was observed. The results suggested that although an increase in

tensile strength was clearly apparent, it was somewhat to the detriment of work hardening, ductility and

also the advent of plastic instability and associated phenomena. Such high ratios of PS:UTS are

common in ultra-fine grained materials and occur as the subgrain size approaches that of the grain size.

0

200

400

600

800

1000

1200

1400

0 1 2 3 4 5Number of ECAE passes

Ten

sile

Str

eng

th (

MP

a)

0

5

10

15

20

25

30

35

Elo

ng

atio

n (

%)

0.2%PS (MPa)

UTS

Elongation

Fig. 173 - Tensile properties of 0.6wt.%C steel with increase in ECAE deformation

0

200

400

600

800

1000

1200

1400

0 1 2 3 4 5 6 7 8 9 10Number of ECAE passes

Ten

sile

Str

eng

th (

MP

a)

0

5

10

15

20

25

30

35

Elo

ng

atio

n (

%)

0.2%PS (MPa)

UTS

Elongation


165

0

200

400

600

800

1000

1200

1400

0 1 2 3 4 5Number of ECAE passes

Ten

sile

Str

ength

(M

Pa)

0

5

10

15

20

25

30

35

Elo

ngat

ion (

%)

0.2%PS (MPa)

UTS

Elongation


8.5 Conclusions

Efforts to produce a two-phase ultra-fine grained structure in high carbon (0.6-1.2wt.%) steels through

innovative deformation, i.e. heavy draft ‘warm’ rolling and ECAE, have proved successful. A

comprehensive programme of thermomechanical simulation was carried out to determine suitable

process parameters. It has been shown that deformation strains of ~1.2 induced during warm rolling at

~600°C are sufficient to promote carbide spheroidisation and refinement of the ferrite matrix.

Quantitative EBSD measurements record a grain size in the order of ~2µm and corresponding subgrain

size of ~1µm. A marked increase in strength, particularly proof strength, was noted following warm

rolling and a substantial increment of work hardening is retained with no loss in ductility.

ECAE deformation has illustrated that pearlite spheroidisation occurs in each steel following two ECAE

passes. An ultra-fine ferrite and carbide aggregate structure is observed after four passes. Ferrite grain

refinement occurs more rapidly in a pearlitic eutectoid composition (0.8wt.%C) steel, resulting in a

submicron grain size following moderate deformation strains (~5 passes). The presence of coarse

cementite particles in the 1.2wt.%C steel was observed to promote development of high misorientation

boundaries within their immediate vicinity. However, due to the large and relatively un-deformable

nature of the cementite particles in comparison with the ‘soft’ ferrite matrix, much of the initial grain

refinement was limited to these deformation zones and overall HAGB formation and refinement of

grain size proceeded at a slower rate. Slow grain refinement observed in the 0.6wt.%C steel is

considered a result of the single phase, primary ferrite regions ‘absorbing’ the majority of the imparted

deformation, leaving the pearlite deformed to a lower effective strain. A large increase in tensile

strength was observed following the initial ECAE pass. Subsequent deformation resulted in a lesser

increase in proof stress and UTS values plateau following two passes. In general, elongation values

decreased with increasing deformation for all steels.

The ability to form a relatively homogeneous ultra-fine ferrite and carbide aggregate microstructure

with improved tensile strength yet retain work hardening, in simple high carbon steels using

conventional rolling practices, is particularly attractive for future commercial application. EBSD has

proved invaluable in assessing the evolution of deformed microstructures, by allowing ultra-fine grain

structures to be rigorously distinguished from microstructures predominantly containing subgrains.

166

9 . GENERAL CONCLUSIONS

It has been shown in this work that the ferrite grain size can be refined to 1-3 µm by different methods.

They can be defined as advanced thermomechanical processes that can be applied on already existent

pilot mills or that can be applied off line (Pony Mill), and others that can be defined as new innovative

techniques like ARB and ECAE. The latter methods can impose high deformations giving promising

results but at this stage they still remain to a laboratory scale.

Different parameters were considered in designing a processing route to obtain fine grains and

simultaneously suitable mechanical properties. Mainly the carbon content and presence of

microalloying elements influenced the best processing route to be applied to the chosen material.

A finer ferrite and a more homogeneous distribution of cementite were found in the higher carbon steel

after the deformation and annealing process. It can be attributed to the initial finer ferrite-pearlite

microstructure before large strain warm deformation and the higher volume fraction of spheroidized

cementite particles during/after deformation in the higher carbon steel.

Moreover, the increase in carbon content was beneficial to increase the work hardening rate. This is due

to a large volume fraction of cementite particles which are effective in increasing the work hardening

rate by the accumulation of geometrically necessary dislocations near the particles. The increase in

work hardening rate is important for improving the ductility (especially the uniform elongation) of the

ultrafine grained steels.

Even by SIDT mechanism the level of ferrite refinement increased slightly when the carbon content

increased, but in this case the critical strain for SIDT occurrence increased. Critical strain becomes

higher and higher increasing the C content until it is quite difficult to activate this mechanism.

Results from literature [134] agree that over 0.2-0.3 %C the critical strain necessary to activate SIDT

becomes so high that dynamic recrystallization is more favourable. Therefore this mechanism can be

applied only for low C steels.

In contrast to carbon, the increase in manganese content had a more significant contribution to grain

refinement in the C content range of 0.15-03%C.

This is, firstly, due to the initial finer transformed microstructure in the higher manganese steel because

of manganese depressing the transformation temperature obviously. Secondly, it can be attributed to

manganese substitutes iron in cementite and forms a more stable cohenite (Fe3-XMnXC). During the

large strain warm deformation itself and in the course of the subsequent annealing, the finer cohenite

particles, due to a sluggish dissolution as well as a slow Ostwald ripening, are more effective to pin the

ferrite grain boundaries against grain growth. Thirdly, the fine particles are also beneficial to store the

geometrically necessary dislocations and substructures inside the grains, which facilitate grain

subdivision. Therefore, the finest microstucture was observed in the higher manganese steel after the

deformation and annealing process.

The prominent combination of mechanical properties found in the ultrafine grained steel with higher

manganese content can be attributed to the effect of an obvious grain refinement by the extra

manganese.

The tested processing routes that successfully produced ultrafine ferrite can be summarized as follows:

Nb microalloyed steels

In Nb microalloyed steels, there is a compromise to be found between different mechanisms occurring

during thermomechanical process, that act in different directions for grain refinement occurrence. They

are:





The most important requirement for grain refinement is to keep Nb in solution to obtain the maximum

precipitation hardening efficiency. Ti was found to reduce the beneficial effect of Nb on both grain

refinement and precipitation hardening.

167

The main hot rolling parameters that affect the ferrite grain size are listed below:

- Microalloy content: Nb must be in solution to refine efficently the grain;

- Cooling rate: must be high to avoid ferrite grain growth;

- Coiling temperature: must be low enough to avoid ferrite grain growth;

- Rolling reduction (strain accumulation): must be high to cumulate strain;

- Finishing rolling temperature: must be low (above Ar3) to cumulate strain;

- Time between the end of rolling and cooling (delay time): must be as short as possible to

avoid austenite recrystallisation.

Plain and microalloyed low C steels by SIDT

SIDT mechanism can be exploited for grain refinement in plain or microalloyed low C steels having a

low C content (<0.2%C). The following hot rolling parameters should be controlled:

- Austenitization at low temperature in order to keep a small prior austenite grain size;

- Rolling reduction in the deformation temperature range Ar3-Ar3+30°C

- Minimum required strain of ε=0.6 to activate SIDT (depending on carbon and Nb content)

Plain C steels by large strain warm deformation

For plain C steels the main hot rolling parameters that affect the ferrite grain size are:

- Austenitization at low temperature or small deformation in austenite region in order to keep a

small prior austenite grain size;

- Rolling at deformation temperature under the pearlitic range

- Minimum strain of ε=0.8 to begin the spheroidisation of cementite

Mechanical Properties

As expected, large enhancements in strength and a reduced work hardening were observed in the

ultrafine grained steels. This is reflected in the yield ratio (lower yield stress/ultimate tensile stress),

which is around 0.9, compared with 0.7 for the conventional steels.

However, it is interesting to note that apart from the high strength an acceptable ductility can still be

found in the ultrafine grained steels investigated, which is reflected in the total elongation of over 20%

and the uniform elongation of about 10% for low-medium C steels. This is quite different from the

results reported by other investigators, i.e. the total elongation is no more than 10%. This can be

attributed to the additional second phase – cementite particles offering the opportunity to improve the

work hardening rate.

A second phase such as globular cementite is beneficial to improve both strength and work hardening

rate of steels. Therefore, it is very attractive to produce a relatively coarser microstructure (i. e. grain

size of 2&3 µm) with a second phase (i. e. finer precipitates) in the matrix for the ductility improvement.

On the other hand, in addition to grain refinement strengthening, precipitation hardening can also

contribute to strengthening.

Moreover, grain refinement also results in an increase in the Lüders strain, in a decrease of ductile-to-

brittle transition temperature, and in improvement in toughness.

Severe plastic deformation techniques

Accumulative Roll bonding and Equal Channel Angular Extrusion showed to be very effective to refine

grains imparting high strain.

Their results were in agreement with those on ultrafine grained materials obtained applying advanced

thermomechanical processes, e.g. large warm rolling. Those new deformation methods can help to

better understand the formation and the mechanisms that are involved in highly deformed materials.

168

10. RECOMMENDATIONS FOR FUTURE WORK

The ability to form a relatively homogeneous ultra-fine ferrite and carbide aggregate microstructure

with improved tensile strength yet retain work hardening using conventional rolling practices, is

particularly attractive for future commercial application.

This work was only the first stage for investigating the potential of the processing of ultrafine grained

steels. All the tested routes were shown to be successful to produce ultrafine microstructures and their

influencing factors were assessed.

In order to continue in the development work of ultrafine grained steels, practical steel manufacturing

processing must be performed with industry experts to verify the routes convenient for bulk production

in a steel mill.

Moreover, in this work the mechanical properties of the ultrafine structures were well tested, however

response to various application like welding were hardly begun. More extensive investigation would be

required to assess the properties achievable.

169

11. REFERENCES

[1] Valiev RZ. Mater Sci and Eng A 1997; 234: 59.

[2] Segal VM. Mater Sci and Eng A 1999; 271: 322.

[3] Saito Y, Utsunomiya H, Tsuji N, Sakai T. Acta Mater 1999; 47: 579.

[4] Tsuji N, Ueji R, Minamino Y. Scripta Mater 2002; 47: 69.

[5] Inoue T, Torizuka S, Nagai K. International Symposium on Ultrafine grained Steels

(ISUGS 2001). The Iron and Steel Institute of Japan, Fukuoka, Japan; 2001. p. 88–91.

[6] Hodgson PD, Hickson MR, Gibbs RK. Scripta Mater 1999; 40: 1179.

[7] Kaspar R, Distl JS, Pawelski O. Steel Research 1988; 59: 421.

[8] Matsumura Y, Yada H. ISIJ Int. 1987; 27: 492.

[9] Najafi-Zadeh A, Jonas JJ, Yue S. Metall Trans A 1992; 23: 2607.

[10] P.D. Hodgson et al., Patent Application PCT/AU94/00356, Publication WO 95/01459, 1995

[11] P.J Hurley, B.C. Muddle, P.D. Hodgson, Proc. “Thermomechanical Processing of Steels 2000”,

24-26 May 2000, London, p. 476

[12] M.R. Hickson and Hodgson, Materials Science and Technology, vol.15, (1999), p. 85

[13] Choi J.K., Seo D.H., Umk K. and Choo W.Y, ISIJ International vol. 43, (2003), p. 746

[14] Priestner R., Al-Horr Y.M. and Ibraheem A.K., Material Science and Tech., vol. 18, (2002), p.

973.

[15] Mintz B. Lewis J. and Jonas J.J., Material Science and Tech., voil. 13, (1997), p. 379

[16] Yang Z.M. and Wang R.Z., ISIJ International vol. 43, (2003), p. 761

[17] Choi JK, Seo DH, Lee JS, Umk K & Choo WY (2003) ISIJ International 43: 746-754.

[18] Mintz B, Lewis J & Jonas JJ (1997),Materials Science and Technology 13: 379-385.

[19] Yang ZM & Wang RZ (2003) ISIJ International 43: 761-766.

[20] Hurley PJ & Hodgson PD (2001) Materials Science and Engineering A 302: 206-214.

[21] S.C.Hong, S.H.Lim, H.S.Hong, K.J.Lee, D.H.Shin, K.S.Lee, Materials Science and Technology,

February 2004, vol. 20.

[22] Du LX, Caibei Z, Hua D, Xianghua L & Guodong W., ISIJ International vol. 42, (2002), p.1119.

[23] M.R. Hickson and P.D. Hodgson, Material Science and Technology, Vol. 15, 1999, p. 85

[24] M.Niikura, M. Fujioka,Y.Adachi, A. Matsukara, T.Yokota,Y. Shirota, Y. Hagiwara, Journal of

Materials Processing Technology, 117, 2001, p. 341

[25] P.D.Hodgson, M.R. Hickson and R.K.Gibbs,Proc. Conf. “Microalloying in steels”, 63, 1988,

Switzerland, Trans Tech.

[26] Jansson B, Schalin M, Sundman B. Phase Equilibria 1993;14:557.

[27] Kaspar R, Pawelski O. Materialprüfung 1989;31:14.

[28] Pawelski O, Gopinathan V. Mechanical Working Technology 1981;5:267.

[29] Kaspar R, Faul H. MP Materialprüfung 2001; 43: 18.

[30] Glez JC, Driver JH. Acta Mter 2003; 51: 2989.

[31] Shin DH, Kim BC, Kim YS, Park KT. Acta Mater 2000; 48: 2247.

[32] Apps PJ, Bowen JR, Prangnell PB. Acta Mater 2003; 51: 2811.

[33] Gholinia A, Humphreys FJ, Prangnell PB. Acta Mater 2002; 50: 4461.

[34] Tsuji N, Toyoda T, Minamino Y, Koizumi Y, Yamane T, Komatsu M, Kiritani M. Mater

Sci and Eng A 2003; 350: 108.

[35] Hornbogen E, Köster U. Recrystallization of Two-phase Alloys. In: Haessner F, editor.

Recrystallization of Metallic Materials. Dr. Riederer Verlag GmbH: 1978.

[36] Hornbogen E. Metall Trans A 1979; 10: 947.

[37] Hales SJ, Mcnelley TR, Mcqueen HJ. Metall Trans A 1991; 22: 1037.

[38] Doherty RD. Nucleation. In: Haessner F, editor. Recrystallization of Metallic Materials.

Dr. Riederer Verlag GmbH: 1978. p. 42.

[39] Chattopadhyay S, Sellars CM. Acta Metall 1982; 30: 157.

[40] Watanabe T, Karashima S. Physica Status Solidi 1970; 42: 749.

[41] Robbins JK, Shepard OC, Sherby OD. Iron and Steel Inst 1964; 202: 804.

[42] Kaspar R, Kapellner W, Lang C. Steel Research 1988; 59: 492.

[43] Shin DH, Park KT, Kim YS. Scripta Mater 2003; 48: 469.

[44] Humphreys FJ, Hatherly M. Recrystallization and Related Annealing Phenomen. UK.

Pergamon: 1995. p. 252.

170

[45] Portern DA, Easterling KE. Phase Transformations in Metals and Alloys. UK. Van Nostrand

Reinhold, Workingham, England, 1981.

[46] Zhang YC, Schleich DM. Solid State Chemistry 1994; 110: 270.

[47] Gottstein G. Physical Foundations of Materials Science. Springer-Verlag, Berlin Heidelberg, 2004.

[48] Björklund S, Donaghey LF, Hillert M. Acta Metall 1972; 20: 867.

[49] Kaspar R, Chen Q, Streißelberger A. Arch. Eisenhüttenwes 1983; 54: 273.

[50] Valiev RZ, Islamgaliev RK, Alexandrov IV. Progress in MaterSci 2000; 45: 103.

[51] Belyakov A, Sakai T, Miura H, Kaibyshev R. Scripta Mater 2000; 42: 319.

[52] Gertsman VY, Birringer R, Valiev RZ, Gleiter H. Scripta Metall et Mater 1994; 30: 229.

[53] Belyakov A, Sakai Y, Hara T, Kimura Y, Tsuzaki K. Scripta Mater 2001; 45: 1213.

[54] Tsuji N, Ueji R, Minamino Y. Microstructure, Texture, and Mechanical Properties of Ultrafine

Grained IF Steel. In: Takechi H, editor. IF Steels 2003. 2003, p347.

[55] Santos DB, Bruzszek RK, Rodrigues PC, Pereloma EV. Mater Sci and Eng A 2003; 346:

189.

[56] Valiev RZ, Kozlov EV, Ivanov YF, Lian J, Nazarov AA, Baudelet B. Acta Metall et Mater 1994;

42: 2467.

[57] Lian J, Baudelet B, Nazarov AA. Mater Sci and Eng A 1993; 172: 23.

[58] Lojkowski W. Acta Metall et Mater 1991; 39: 1891.

[59] Park K-T, Kim Y-S, Lee JG, Shin DH. Mater Sci and Eng A 2000; 293: 165.

[60] Reed-Hill RE, Abbaschian R. Physical Metall Principles. PWS Publishing Company:

1994. p. 158.

[61] Wang ZC, Prangnell PB. Mater Sci and Eng A 2002; 328: 87.

[62] Lloyd DJ, Morris LR. Acta Metall 1977; 25: 857.

[63] Hayes JS, Keyte R, Prangnell PB. Mater Sci and Tech 2000; 16: 1259.

[64] Kim S, Im YR, Lee S, Lee HC, Oh YJ, Hong JH. Metall and Mater Trans A: Physical

Metall and Mate Sci 2001; 32: 903.

[65] Brozzo P, Buzzichelli G, Mascanzoni A, Mirabile M. Metal Sci 1977; 11: 123.

[66] Hulla D. Acta Metall 1960; 8: 11.

[67] Ritchie RO, Knott JF, Rice JR. Mechanics and Physics of Solids 1973; 21: 395.

[68] Knott JF. Pressure Vessels and Piping 1995; 64: 225.

[69] Kim S, Lee S, Lee BS. Mater Sci and Eng A 2003; 359: 198.

[70] Im YR, Oh YJ, Lee BJ, Hong JH, Lee HC. Nuclear Mater 2001; 297:138.

[71] Pitsch W, Sauthoff G, Hougardy HP. Steel. In: Verein Deutscher Eisenhüttenleute, editor. A

Handbook for Mater Research and Eng, vol. 1. Springer-Verlag: 1992. p. 83.

[72] Pickerin FB. Physical Metallurgy and the Design of Steels. Applied Science Publishers LTD,

London: 1978.

[73] Broek D. Eng Fracture Mechanics 1973; 5: 55.

[74] Cai H, Evans JT, Boomer D. Eng Fracture Mechanics 1992; 42: 589.

[75] Petch NJ. Iron and Steel Inst 1953; 174: 25.

[76] Dahl W. Steel. In: Verein Deutscher Eisenhüttenleute, editor. A Handbook for Mater

Research and Eng, vol. 1. Springer-Verlag: 1992. p. 283.

[77] Ashby MF. Philo Magazine 1970;21:399.

[78] Kochs UF. Eng Mater and Tech 1976; 98:76.

[79] Ashby MF. Philo Magazine 1966;14:1157.

[80] Kosco JB, Koss DA. Mater Sci and Eng A 1993;169:1.

[81] Jones RH. Metall Trans 1973;4:2799.

[82] Jones RH. Metall Trans; 1974: 5: 970.

[83] Brown LM, Stobbs WM. Philo Magazine 1976;34:351.

[84] Humphreys FJ, Hirsch PB. Philo Magazine 1976;34:373.

[85] Koch CC. Scripta Mater 2003; 49: 657.

[86] Y. Saito, H. Utsunomiya, N. Tsuji, and T. Sakai, Acta mater., vol. 47, No.2, 1999, p579

[87] N. Tsuji, Y. Saito, H. Utsunomiya and S. Tanigawa, Script Materialia, vol.40, No.7, 1999, p795

[88] N. Tsuji, Y. Saito, S.H. Lee and Y. Minamino; Proceedings of Conference NANOSPD2,

December 2002, Vienna, Austria. Ed. M.J.Zehetbauer, R.Z. Valiev.

[89] N. Tsuji, Y. Minamino, Y.Koizumi and Y. Saito;”Int. Symp. On Processing and Fabrication of

Advanced Materials”, Materials Solution 2002, ASM&TMS Columbus (OH).

[90] A. Schmitz et al., Proc. Of 40th MWSP conf., Vol. 36, 1998, p.295

171

[91] J.C. Herman, Proc. Of the 'Thermomechanical processing of steels 2000' conf., 24-26 May 2000,

London, p. 585.

[92] T.M. Maccagno, J.J. Jonas, S. Yue, B.J. McCrady, R. Slobodian, D. Deeks: ISIJ International, 34

(11), 1994, p. 917.

[93] F.H. Samuel, S. Yue, J.J. Jonas. K.R. Barnes: Trans ISIJ, 30, 1990, p. 216.

[94] R. Kaspar, J.S. Distl, O. Pawelski: Steel Research, 59, 1988, p. 421.

[95] A. Schmitz, A. De Ro: Proc. of the 'Thermomechanical processing of steels 2000' conf., 24-26

May 2000, London, p. 702.

[96] F. Boratto, R. Barbosa, S. Yue, J.J. Jonas: Proc. Int. Conf. Physical Metallurgy of Processing of

Steels and Other Metals (THERMEC ’88); ed. I. Tamura, ISIJ, Tokyo, 1988, p. 383.

[97] R.C. Hudd, A. Jones, M.N. Kale: JISI, 209, 1971, p.121.

[98] J. Neutjens, A. Schmitz, Ph. Harlet: 'On the use of Torsion Testing for the study of Phase

Transformation after Austenite Pancaking in a Hot Strip Finishing Mill', Thermec 2000.

[99] Development of Methods for the Characterization and Modelling of Precipitation in Steels, ECSC

Project 7210-PR/160, Technical report N°3.

[100] W.T. Reynolds, Jr., F.Z. Li, C.K. Shui, H.I. Aaronson: Metallurgical Transactions A, vol. 21A,

June 1990, p. 1433.

[101] J. Burke: The Kinetics of Phase Transformation in Metals, Pergamon Press, Elmsford, NY, 1965.

[102] C. Mesplont, S. Vandeputte and B.C. De Cooman: Z. Metallkde 93, 2002, No. 11, pp.1108-1118.

[103] Ken-Ichiro Suzuki, Shoji Miyaga, Yoshiyuki Saito and Kazuhiro Shiotani: ISIJ International,

Vol.35 (1995), No.1, pp.34-41.

[104] C. Ouchy, T. Sanpei and K. Matsumoto: Tetsu-to-Hagané, 69 (1979), A293.

[105] Ultra Fast Cooling in the Hot Strip Mill (Phase I), ECSC contract number 7215-CA/204 (96-

T3.06).

[106] T. Morimoto et al. Proc. 2nd

Int. Conf. On Thermomechanical Processing of Steels, June 15-17,

2004, Liège, Belgium, p.415.

[107] D N Crowther: ‘Ultra-fine grained steel by innovative deformation cycles’, ECSC Agreement

No. 7210.PB/288 (D3.02/01), Report No. M8657-1(ST)012, Corus Research, Development &

Technology, Swinden Technology Centre, 2001.


No. 7210.PB/288 (D3.02/01), Report No. M8657-1(ST)012, Corus Research, Development &

Technology, Swinden Technology Centre, 2002.


No. 7210.PB/288 (D3.02/01), Report code STC/SMT CTE/CR/08657/2003/A, Ref. Source No.

108386, Corus Research, Development & Technology, Swinden Technology Centre, 2003.

[110] D H Shin, Y S Kim and E J Lavernia: ‘Formation of Fine Cementite Precipitates by Static

Annealing of ECAP Low C Steels’, Acta Mat., 2001, Vol. 49, pp2387-2393.

[111] V M Segal, V I Reznikov, A E Drobyshevskiy, V I Kopylov: ‘Plastic working of metals by

simple shear’, Russ. Metal., 1981, 1, pp99-105.

[112] R Z Valiev, N A Krasilnikov and N K Tsenev: ‘Plastic deformation of alloys with submicron-

grained structure’, Mat. Sci. Eng., 1991, A137, pp35-40.

[113] R Z Valiev, A V Korznikov and R R Mulyukov: ‘Structure and properties of metals with

submicrocrystalline structures’, The Physics of Metals and Metallography, 1992, Vol. 4, pp70-

86.

[114] V M Segal: ‘Materials processing by simple shear’, Materials Science and Engineering, 1995,

A197 [2], pp157-164.

[115] V M Segal: ‘Equal channel angular extrusion: from macromechanics to structure formation’,

Materials Science and Engineering, 1999, A271 [1], pp322-333.

[116] K Nakashima, Z Horita, M Nemeto and T G Langdon: ‘Influence of channel die angle on

development of ultra-fine grains in equal channel angular pressing’, Acta Mat., 1998, Vol. 46,

pp1589-1599.

[117] F J Humphreys, P B Prangnell, J R Bowen, A Gholinia and C Harris: ‘Developing stable fine-

grain microstructures by large strain deformation’, Philos. Trans. Royal Soc., 1999, A357,

pp1663-1680.

[118] A Gholinia, J Sarkar, P J Withers and P B Prangnell: ‘Ultra-fine grain structures formed by

thermomechanical processing of spray cast Al-Li alloys’, Mat. Sci. Tech., 1999, 15, pp605-615.

172

[119] A Gholinia, P B Prangnell and M V Markuchev: ‘The effect of strain path on the development of

deformation structures in severely deformed aluminium alloys processed by ECAE’, Acta Mat.,

2000, Vol. 48, pp1115-1130.

[120] A Gholinia, F J Humphreys and P B Prangnell: ‘Processing ultra-fine grain structures by

conventional routes’, Mat. Sci. Tech., 2000, Vol. 16, pp1251-1255.

[121] J S Hayes, R Keyte and P B Prangnell: ‘Effect of grain size on the tensile behaviour of a

submicron grained Al-3%Mg alloy produced by severe deformation’, Mat. Sci. Tech., 2000, Vol.

16, pp1259-1263.

[122] R Priestner and A K Ibraheem: ‘Processing of steel for ultra-fine ferrite grain structures’, Mat.

Sci. Tech., 2000, Vol. 16, pp1267-1272.

[123] P J Apps, J R Bowen, P B Prangnell: ‘The effect of coarse second-phase particles on the rate of

grain refinement during severe deformation processing’, Acta Mat., 2003, Vol. 54, pp2811-2822.

[124] O.D. Sherby, B Walser, C.M. Young and E.M. Cady, “Superplastic Ultra-High-C Steels”, Scripta

Metallurgica, Vol. 9, 1975, 569-573.

[125] B. Walser and O.D. Sherby, “Mechanical Behaviour of Superplastic Ultrahigh Carbon Steels at

Elevated Temperature”, Metallurgical Transactions A, Vol. 10A, 1979, 1461-1471.

[126] J.R. Bowen, A. Gholinia, S.M. Roberts and P.B. Prangnell, “Analysis of the billet deformation

behaviour in equal channel angular extrusion”, Materials Science and Engineering, A287 1, 2000,

87-99.

[127] Niikura et al, “new concepts for ultra refinement of grain size in Super Metal Project”, Journal of

Materials Processing Technology, 117, 2001, 341-346.

[128] S. Chattopadhyay and C.M. Sellars, “Kinetics of Pearlite Spheroidisation During Static

Annealing and During Hot Deformation”, Acta Materialia, 1982, 157-170.

[129] W.B. Morrison and R.L. Miller, “The Ductility of Ultrafine Grained Alloys”, Ultrafine-Grain

Metals, J.J. Burke and V. Weiss, Eds., Syracuse University Press, 1970, 183.

[130] P.D. Hodgson, M.R. Hickson and R.K. Gibbs, “The Production and Mechanical Properties of

Ultrafine Ferrite”, Materials Science Forum, Vols. 284-286, 1998, 63-72.

[131] J R Bowen: ’The formation of ultra-fine grained model aluminium and steel alloys’, PhD Thesis,

UMIST, 2000.

[132] J.Laguillaume, G.Kapelski and B.Baudelet, “Cementite Dissolution in Heavily Cold Drawn

Pearlitic Steel Wires”, Acta Materialia, Vol. 45, 1997,1201-1212.

[133] W.J. Nam et al, “Effect of Interlamellar Spacing on Cementite Dissolution during Wire Drawing

of Pearlitic Steel Wires”, Scripta Materialia, Vol. 42, 2000, 457-463.

[134] Yada H, Matsumura Y & Pnakajima K (1984) Ferrite Steel Having Ultra-fine Grains and a

Method for Producing the Same, United State Patent No. 4, 466842.

173

ultrafine grained steel by innovative deformation cycle

Documents