The Influence Of Austenite Grain Size On Hot Ductility Of Steel

NOTE This online version of the thesis may have different page formatting and pagination from the paper copy held in the University of Wollongong Library.

UNIVERSITY OF WOLLONGONG COPYRIGHT WARNING You may print or download ONE copy of this document for the purpose of your own research or study. The University does not authorise you to copy, communicate or otherwise make available electronically to any other person any copyright material contained on this site. You are reminded of the following: Copyright owners are entitled to take legal action against persons who infringe their copyright. A reproduction of material that is protected by copyright may be a copyright infringement. A court may impose penalties and award damages in relation to offences and infringements relating to copyright material. Higher penalties may apply, and higher damages may be awarded, for offences and infringements involving the conversion of material into digital or electronic form.

The Influence Of Austenite Grain Size On Hot Ductility Of Steels

A thesis submitted in fulfilment of the requirements for the award of the degree

MASTER OF ENGINEERING BY RESEARCH

From

UNIVERSITY OF WOLLONGONG

By

Suk-Chun Moon, B.Eng (Met.)

DEPARTMENT OF MATERIALS ENGINEERING 2003

CERTIFICATION

I, Suk-Chun Moon, declare that this thesis, submitted in fulfilment of the requirements for the award of Master of Engineering by Research, in the Department of Materials Engineering, University of Wollongong, is wholly my own work unless otherwise referenced or acknowledged. The document has not been submitted for qualifications at any other academic institution.

Suk-Chun Moon December 2003

1

ACKNOWLEDGEMENTS

I wish to express my gratitude to Prof. Rian Dippenaar who has supervised my work. I deeply appreciate his prudent regard on all things. He always encouraged me saying “Excellent”. His words always made me confident. I express my gratitude to POSCO and its staff for giving me this opportunity to study abroad and supporting unsparingly. I also acknowledge Mr. Shin-Eon Kang (POSCO Technical Research Laboratories) for providing the experimental materials. I am deeply indebted to Mr. Bob DeJong for his support with the hot tensile tests on the GLEEBLE 3500. I also thank Mr. Ron Marshall for preparing my specimens and Mr. Greg Tillman for assistance in metallographic preparation. Many thanks to Dr. Dominic Phelan and Mr. Mark Reid for their assistance with my experiments. I thank Rev. Tae-Joo Lee who helped my family to adapt easily when we arrived in Australia in cold windy winter. I am most grateful to my parents for their deep concern for me and my family and for their support over such a long distance. Finally, my most severe thanks to my wife, Kwang-Young and my son, Joon-Young, for their patience, loving support and encouragement to finish my task and giving me good heart and love. Also, thanks to my unborn baby who will bring ‘big surprise’ sooner or later.

2

ABSTRACT

The high temperature ductility of steel increases as the grain size is decreased. However, in microalloyed steels this interdependence of grain size and hot-ductility is generally less pronounced because of the overriding effect of precipitation of carbo-nitrides of alloying elements at grain boundaries. Nevertheless, many types of crack have been shown to be associated with coarse austenite grains, and the tendency to crack is reduced when the formation of coarse austenite grains is prevented by the use of suitable secondary cooling. To our knowledge, a systematic study, to isolate the contribution of grain size on hot-ductility has not been reported and hence, under these circumstances, it is important to define clearly the contributing role of grains size to hot-ductility and therefore it is necessary to separate the effect on hot ductility of microalloying precipitates from that of grain size. The materials used for study were Fe-0.05%C, Fe-0.18%C and Fe-0.45%C alloys, prepared in an experimental facility. Carbon was the only alloying element deliberately added to the melt and tramp elements and impurities were kept to as low a value as experimentally possible. A GLEEBLE 3500 thermomechanical simulator was used to conduct hot-tensile tests and from these results the relationship between austenite grain size and hot-ductility could be determined. The hot-ductility tensile tests in this study were performed in the temperature range 1100 to 700 . Hot-tensile test specimens were either solution treated or melted in-situ (direct cast) and cooled to the test temperature in order to simulate the microstructural characteristics of commercially ascast slab which contains coarse austenite grains at the base of oscillation marks. In order to obtain different grain sizes, the specimens were solution treated at temperatures between 1100 1350

and

for 10minutes. The specimens were then rapidly cooled to the test temperature, in the

range 1100

to 700

at a rate of 200 min-1 and then pulled to fracture at a low strain rate of

7.5×10-4s-1. Specimens that were cast in-situ (referred to as the ‘direct casting condition’) were

3

cooled at rates of 100 min-1 and 200 min-1 respectively in order to study the effect of cooling rate on hot ductility. It was found that grain size increased almost linearly with increasing solution treatment temperature in all the Fe-C alloys and in case of specimens solution treated at 1350

an average

austenite grain size of ~4mm in diameter was obtained. Increasing the grain size resulted in ductility loss under all testing conditions. The existence of a ductility trough between Ar3 and Ae3 temperature was considered to be due to the formation of deformation induced ferrite as evidenced by a constant peak stress region between these two temperatures. However, convincing experimental evidence of the ductility trough extending beyond the Ae3 temperature well into the austenite was found. The mechanism of this interesting, and important observation for low carbon alloy below 0.3%C has not been explored as yet and at present it has to be assumed that it is related to the occurrence of grain boundary sliding because in the high austenite temperature region the grain boundary sliding is favored in a coarse grained structure. For specimens cast in-situ, the largest grains were found in the Fe-0.18%C alloy for both cooling rates. This important observation is attributed to the higher austenitizing temperature of a Fe-0.18%C alloy compared to that of the other Fe-C alloys studied. Moreover, the formation of columnar austenite grains were observed in this alloy on cooling, whereas equi-axed grains were formed in the other two Fe-C alloys. By such columnarization, the surface cracking susceptibility of the peritectic grade steel will be accelerated. The hot-ductility of the Fe-C alloy of nearperitectic composition was the lowest of the alloys studied and the inferior ductility in this alloy is attributed to the coarse grain size and columnar shape of the grains. The Fe-0.45%C alloy had the smallest grain size at any specimens cast in-situ but the hot-ductility was much lower than would have been expected in an alloy of such small grain size. This much reduced ductility may be due to increased grain boundary sliding. At a higher cooling rate of the in-situ melted specimens, smaller grains were produced in all the

4

Fe-C alloys resulting in ductility improvement. This grain refinement obtained at higher cooling rates have important implications for near-net shape casting operations such as thin-slab casting or strip-casting where much higher cooling rates are realized than in the conventional casting process, if the factors related with precipitation are removed. An important insight derived at through this study was that enlarged grain contributed more to reduced ductility at high temperature than did cast structure, at least under the pertaining experimental conditions. This observation has important practical implications because it means that efforts in industry could be concentrated on reducing the chances of forming large austenite grains, such as at the roots of oscillation marks due to a decreased cooling rate, without undue regard to the effect of these measures on cast structure. This study has provided convincing new experimental evidence of the extremely detrimental effect of large austenite grains on hot-ductility in plain carbon steels. The very large columnar shaped grains that can form in alloys of near-peritectic composition is particularly disconcerting. However, the influence of AlN precipitation on austenite grain boundaries on hot-ductility was not studied and it is recommended that this important topic should be included in subsequent investigation. The experimental data on these Fe-C alloys, provided in this study, may now be used to benchmark and further analyse the much larger body of information on low-alloyed steels available in the literature.

5

TABLE OF CONTENTS CHAPTER 1 - INTRODUCTION

1

CHAPTER 2 - LITERATURE REVIEW

4

2.1 Continuous casting process

4

2.1.1 Outline of continuous caster

4

2.1.2 Near net shape casting technology

5

2.2 Slab cracking

9

2.2.1 Features of Cracking

9

2.2.2 Coarse prior austenite grains

11

2.3 Hot ductility test

14

2.3.1 Simulation of straightening operation during continuous casting

14

2.3.2 Suitability of the hot tensile test to the problem of transverse cracking

15

2.4 Fracture mechanisms

16

2.4.1 Region of embrittlement

17

2.4.1.1 Embrittlement by strain concentration and microvoid coalescence at grain boundaries 2.4.1.2 Grain boundary sliding

17 21

2.4.2 High ductility, low temperature region

23

2.4.3 High ductility, high temperature region

23

2.4.4 Hot ductility behaviour of steels

24

2.4.4.1 Plain C-Mn and C-Mn-Al steels

24

2.4.4.2 C-Mn-Al steels with high Al and N levels

25

2.4.4.3. Microalloyed steels

26

2.5 Factors influencing hot ductility

27

2.5.1 Grain Size

27

2.5.2 Precipitation

29

2.5.3 Composition

32

2.5.3.1 Sulphur

32

2.5.3.2 Carbon

35

2.5.4 Cooling rate

38

2.5.5 Strain rate

38

i

2.5.6 Thermal history

39

CHAPTER 3 - EXPERIMENTAL

41

3.1 Preparation of specimens

41

3.2 Hot ductility test

42

3.2.1 GLEEBLE3500

42

3.2.2 Thermomechanical cycles

43

3.2.3 Measurement of hot ductility

45

3.2.4 Correction of stress-strain curve

45

3.2.5 Determination of GLEEBLE setting temperature for melting

47

3.3 Metallography

49

CHAPTER 4 - RESULTS

50

4.1 Hot ductility curves

50

4.1.1 Fe-0.05%C alloys

50

4.1.2 Fe-0.18%C alloys

52

4.1.3 Fe-0.45%C alloys

53

4.2 Stress strain curves

55

4.2.1 Fe-0.05%C alloys

55

4.2.2 Fe-0.18%C alloys

57

4.2.3 Fe-0.45%C alloys

58

4.3 Austenite grain size

60

CHAPTER 5 - DISCUSSION

67

5.1 Grain growth

67

5.2 Ductility troughs

69

5.3 Influence of grain size on hot ductility

74

5.3.1 Reduction in area

74

5.3.2 Position and width of ductility trough

76

5.4 Influence of carbon content on hot ductility

77

5.5 Influence of cooling rate following direct-casting on hot ductility

81

5.6 Comparison between as-cast condition and solution treatment condition

82

5.7 Practical implications of the experimental findings

ii

83

CHAPTER 6 - CONCLUSIONS

85

BIBLIOGRAPHY

88

iii

LIST OF FIGURES Fig.2.1 Schematic diagram of typical continuous casting machine [1]

4

Fig.2.2 Schematic diagram of (a) conventional continuous casting and hot-rolling (b) Thin slab-hot rolling (TSHR) (c) Strip-casting

6

Fig.2.3 Practical examples of thin-slab casting techniques

9

Fig.2.4 Widespread crazing and fine transverse cracks at oscillation marks on the as-cast surface of a line pipe steel slab (top side). Etched in hot HCL. Magnification not specified. CD : Casting Direction [4]

10

Fig.2.5 Crazing around a transverse crack at the base of an oscillation mark on the as-cast top surface of a 0.20%C steel slab. Etched in hot HCL [4].

10

Fig.2.6 Section through the surface across a system of transverse cracks through the coarse-grained zone. Prior austenite grain boundaries white (ferrite-decorated). Etched in nital. Magnification not specified [6].

11

Fig.2.7 Grain growth of austenite during continuous cooling. The specimens were remelted at 1580 , cooled to a given temperature at a rate of 0.28 s-1, and then quenched into water [7].

12

Fig.2.8 Formation of surface cracks due to blown grain during casting [4]

14

Fig.2.9 Schematic diagram of a ductility curve defining the three characteristic regions of hotductility [1]

17

Fig.2.10 Schematic diagram showing mechanism for transformation induced intergranular failure [1]

18

Fig.2.11 Microstructure at room temperature of steel tested at 800 , showing intergranular failure associated with a thin layer of grain boundary ferrite [17]

19

Fig.2.13 Schematic illustration of intergranular microvoid coalescence of Nb-bearing steels. a-c depict deformation in austenite above the Ar3 temperature, d-f depict deformation in the (austenite + ferrite) region

21

Fig.2.14 Schematic models showing the formation of wedge cracks by grain boundary sliding: The arrows indicate the sliding boundaries and the sense of translation [23]

22

Fig.2.15 Intergranular microvoid coalescence type fracture in C-Mn-Al steel, solution treated at 1350 , tested at 850

at strain rate of 10-3s-1 [16]

24

Fig.2.16 (a) C-Mn-Al steel showing flat facets on fracture surface, (b) enlarged view of a showing lack of voiding around MnS inclusions [16]

25

Fig.2.17 Hot ductility curves for C-Mn steels at various grain sizes (a) 0.19wt%C steel (b)

0.65wt%C steel [32]

28

Fig.2.18 Influence of (a) particle size and (b) interparticle spacing on hot ductility of Nb-containing steels, solution treated at 1330 , cooled to a test temperature of 850 , and fractured at a strain rate of 3×10-3s-1 [1]

30 iv

Fig.2.19 Effect of sulphur content on the minimum reduction of area for two cooling rates of 1°Cs-1 and 4°Cs-1 [51]

33

Fig.2.20 Schematic illustration showing the effect of C on surface cracking of continuous cast slabs [7]

35

Fig.2.21 Effect of C content on austenite grain size and calculated value of RA (a) Relationship between austenite grain size and C content. Specimens (0.35%Si-1.5%Mn-0.05%Nb) were remelted at 1580 , cooled to 900

at a rate of 5 s-1 and then quenched. (b)

Calculated values RA which were deduced from the relationship between grain size and RA assuming that the specimens were deformed at 800

at a strain rate of 0.83×10-3s-1 [7]

36

Fig.2.22 Influence of C and Mn on the width of the ductility trough [39]

37

Fig.3.1 Geometry of GLEEBLE specimen

41

Fig.3.2 Schematic diagram of the GLEEBLE testing arrangement

43

Fig.3.3 Schematic diagram of thermomechanical cycles for hot ductility tests under the conditions of (a) Solution treatment and (b) Direct casting

44

Fig.3.4 Geometry of sample after fracture

45

Fig.3.5 Measured tensile force during tensile deformation

46

Fig.3.6 Stress-strain curve. (a) curves obtained from GLEEBLE (b) modified curves

46

Fig.3.7 The results of preliminary experiments for the determination of the apparent melting point, (a) Fe-0.05%C alloy (b) Fe-0.18%C alloy (c) Fe-0.45%C alloy

48

Fig.4.1 Hot ductility curves for Fe-0.05%C alloys under (a) solution treatment condition (b) direct casting condition

51

Fig.4.2 Sample geometry after fracture tested at 850

under solution treatment condition

52

Fig.4.3 Hot ductility curves for Fe-0.18%C alloys under (a) solution treatment condition (b) direct casting condition

53

Fig.4.4 Hot ductility curves for Fe-0.45%C alloys under (a) solution treatment condition (b) direct casting condition

54

Fig.4.5 Stress-strain curves at different test temperatures for Fe-0.05%C alloys solution treated at (a) 1100

(b) 1200

(c) 1350

55

Fig.4.6 Stress-strain curves at different test temperatures for Fe-0.05%C Alloys under direct casting condition at cooling rate (a) 100 min-1 (b) 200 min-1

56

Fig.4.7. Peak stress as a function of test temperature for Fe-0.05%C alloys under (a) solution treatment condition (b) direct casting condition

56

Fig.4.8 Stress-strain curves at different test temperatures for Fe-0.18%C alloys solution treated at temperature (a) 1100

(b) 1200

(c) 1350

57

Fig.4.9 Stress-strain curves at different test temperatures for Fe-0.18%C alloys under direct casting condition at cooling rate (a) 100 min-1 (b) 200 min-1

57

Fig.4.10. Peak stress as a function of test temperature for Fe-0.18%C alloys under (a) solution treatment condition (b) direct casting condition v

58

Fig.4.11 Stress-strain curves at different test temperatures for Fe-0.45%C alloys solution treated at temperature (a) 1100

(b) 1200

(c) 1350

58

Fig.4.12 Stress-strain curves at different test temperatures for Fe-0.45%C alloys under direct casting condition at cooling rate (a) 100 min-1 (b) 200 min-1

59

Fig.4.13 Peak stress as a function of test temperature for Fe-0.45%C alloys under (a) solution treatment condition (b) direct casting condition

59

Fig.4.14 Cross-section of a GLEEBLE specimen of a Fe-0.05%C alloy. Etched in nital (diameter of specimen = 10mm)

61

Fig.4.15 Cross-section of a GLEEBLE specimen of a Fe-0.18%C alloy. Etched in saturated picric acid based etchant (diameter of specimen = 10mm)

62

Fig.4.16 Cross-section of a GLEEBLE specimen of a Fe-0.45%C alloy. Etched in saturated picric acid based etchant (diameter of specimen = 10mm)

63

Fig.4.17 The distribution of austenite grain size of a Fe-0.05%C alloy under (a) solution treatment condition (b) direct casting condition

64

Fig.4.18 The distribution of austenite grain size of a Fe-0.18%C alloy under (a) solution treatment condition (b) direct casting condition

65

Fig.4.19 The distribution of austenite grain size of a Fe-0.45%C alloy under (a) solution treatment condition (b) direct casting condition

66

Fig.5.1 Grain size as a function of solution treatment temperature

68

Fig.5.2 Grain size as a function of cooling rate under direct casting condition

69

Fig.5.3 Hot-ductility curves for Fe-0.05%C alloys for solution treated specimens having various grain sizes

70

Fig.5.4 Hot-ductility curves for Fe-0.18%C alloys for solution treated specimens having various grain sizes

70

Fig.5.5 Hot-ductility curves for Fe-0.45%C alloys for solution treated specimens having various grain sizes

73

Fig.5.6 Relationship between tensile properties and reciprocal of austenite grain size (D) at different test temperatures in solution treated specimen (a) Fe-0.05%C (b) Fe-0.18%C (c) Fe-0.45%c alloy

75

Fig.5.7 Relationship between tensile properties and reciprocal of austenite grain size (D) at different test temperatures under direct casting condition for (a) Fe-0.05%C (b) Fe-0.18%C (c) Fe-0.45%c alloy

75

Fig.5.8 Relationship between minimum RA value and reciprocal of austenite grain size (D) for different Fe-C alloys in the solution treatment condition

76

Fig.5.9 (a) Position of ductility trough, (b) width of trough as a function of grain size under solution treatment condition

77

Fig.5.10 Hot ductility curves from solution treatments at (a) 1100

(b) 1200

(c) 1350

78

Fig.5.11 Effect of C content on austenite grain size and RA value under direct casting condition for cooling rate (a) 100 min-1 (b) 200 min-1

79

Fig.5.12 Hot ductility curves for Fe-0.18%C alloy from two different thermal conditions vi

82

CHAPTER 1 - INTRODUCTION

The continuous casting of steel was introduced commercially around 1960. Through numerous efforts to solve problems related to solidification phenomena, many developments and improvement in technology were made resulting in large-scale production and high yield ratios. However, one of the problems still pertaining to this process has been the occurrence of cracks on the slab surface such as transverse cracking, crazing, corner cracking and star cracking. In many cases these cracks remain even after rolling with high reduction in strip thickness. The introduction of thin-slab casting and hot-direct-rolling (HDR), where surface inspection prior to rolling is not possible, elevated the need to ensure the elimination of these cracks and the improvement of surface quality if defect-free rolling is to be achieved. Maintaining a defect free slab surface has now become a prime requirement for the economic production of HDR of steel sheet. The straightening operation of the continuously cast strand is carried out when the slab surface temperature is in the range 1000

to 700 . These surface temperatures unfortunately coincide

with the temperature range in which steel exhibits a ductility minimum as measured in laboratory hot-tensile tests [1]. Although numerous studies have been devoted to establishing a relationship between transverse cracking and the hot-ductility trough as well as the factors that effect this relationship, only a few researchers [7, 32, 39] have reported the hot-ductility behavior of the steel in relation to the characteristics of austenite grain growth. Generally, the high temperature ductility of steel increases as the grain size is decreased. Refining the grain size leads to a reduction in both the depth and width of the ductility trough [32]. On the other hand, microalloyed steel generally shows little indication that grain size has a significant influence on hot ductility. This observation is mainly due to the overriding effect of the precipitation of AlN or Nb(C,N) at grain boundaries. Nevertheless, in the straightening operation

1

of continuous casting strands, many types of cracks are well associated with coarse austenite grains, and convincing evidence has been provided that the tendency to crack is reduced when the formation of coarse austenite grains is prevented by the use of suitable secondary cooling [6]. Under these circumstances, it is important to define clearly the role of grains size in controlling hot ductility. In order to do this, it is necessary to separate the effect on hot ductility of microalloying precipitates from that of grain size. It is generally conceded that the early stages of solidification and subsequent high temperature phase transformations profoundly effect cast structure, and therefore also hot-ductility during the straightening operation, but conclusions have mostly been drawn from indirect evidence because of the difficulty of conducting reliable experiments at these high temperatures. In the present study, a GLEEBLE3500 thermo-mechanical simulator has been used to evaluate the ductility of three different Fe-C alloys. The relationship between austenite grain size and hotductility of specimens in the temperature range 1100

to 700 , was obtained through hot tensile

tests under conditions similar to that of as-cast slab which contain coarse austenite grains at the base of oscillation marks. For the purposes of this investigation, plain carbon steels have been prepared in a laboratory furnace. Carbon was the only alloying element deliberately added to the melt and tramp elements and impurities were kept to as low a value as was experimentally possible. The rationale behind this approach is to attempt to isolate the effect of austenite grain size on hot-ductility in the absence of complicating factors such as the precipitation of alloy carbides and the presence of low melting point impurities such as iron sulphides. An attempt was made to prepare pure Fe-C alloys but aluminum had to be added as deoxidizer and some ingress of nitrogen could not be avoided. Hence, as will be shown in Table 3.1, there is a distinct possibility that AlN can form in the alloys prepared for this investigation. The influence of AlN grain boundary precipitates on hot-ductility is well documented and no attempt was made in

2

this study to isolate the contribution of AlN grain boundary precipitation to hot-ductility.

3

CHAPTER 2 - LITERATURE REVIEW 2.1 Continuous casting process 2.1.1 Outline of continuous caster A typical continuous caster is shown schematically in Fig.2.1 [1]. A tundish which is supplied with molten steel from a ladle feeds it into the mold, which is oscillating and water-cooled. Mold oscillation is carried by a hydraulically driven mold oscillation system to ensure surface quality in the cast slab and to prevent sticking of the solidifying strand to the mould. This oscillation is responsible for the formation of oscillation marks on the surface of the strand. From mold to secondary cooling zone, molten steel is solidified into slab and cooled. There are many stresses involved in this process ; Friction between solidified shell and mold wall, ferrostatic pressure inside the shell, thermal stresses on the strand surface and bending stresses at the straightening point. Whenever a tensile stress is present when the steel shell is in a region of low ductility, there is the likelihood that crack can form.

Please see print copy for Figure 2.1

Fig.2.1 Schematic diagram of typical continuous casting machine [1]

2.1.2 Near net shape casting technology

4

Near net shape casting is advanced process technology that has been a major focus of recent research and development efforts. The technology includes the methods by which liquid steel is continuously cast into shapes more nearly approximating the final dimensions to be achieved in the hot finishing stage of production. By doing so, the construction cost of plant and the production cost of product can be reduced significantly [2]. These features are illustrated in Fig.2.2. In this figure, the process routes of conventional slab continuous casting, thin-slab casting and continuous strip-casting are compared. Direct linkage or merger of two or more consecutive process steps is a powerful means by which production cost in the steel manufacturing process can be reduced. Thin-slab hot rolling (TSHR) and strip-casting are the latest examples of process developments in which these principles are applied. In the TSHR process, continuous casting and hot-strip rolling are integrated so that the roughing mill is eliminated, whereas in the strip-casting process, the hot-strip rolling process is eliminated completely. The thickness of thin slabs ranges between 50 and 150mm, whereas the thickness of strips produced by strip-casting is about 2-6mm. Efforts are still being made to develop a technology known as DSC for casting of slabs of plain carbon steels with thickness between thin-slab casting and strip-casting by utilizing the horizontal twin belt concept.

S la b (2 0 0 ~ 2 5 0 mm)

R e h e a tin g F u rn a c e

R o u g h in g M ill

F in is h in g M ill

600m

(a) Conventional slab continuous casting and hot-rolling

5

C o ilin g

T h in S la b (5 0 ~ 8 5 m m )

C o il B o x

C o ilin g

200~300m

(b) Thin-slab hot rolling (TSHR)

F in a l S trip (0 .7 ~ 3 .0 m m )

C o ilin g

C a s t S trip (1 .4 ~ 6 m m )

100m

(c) Strip-casting Fig.2.2 Schematic diagram of (a) conventional continuous casting and hot-rolling (b) Thin slab-hot rolling (TSHR) (c) Strip-casting

Recently many steel making industrials have invested in commercial mini mills that employ thin-slab casting technology, thereby deriving technical and economic benefit from this lower cost, streamlined alternative for producing many types of hot-rolled sheets. Also, this process can be rendered environmentally friendly by selecting the electric-arc furnace route of steelmaking which uses recycled scrap and directly reduced iron as a raw material. There are a variety of thin slab-casting techniques currently being used in the world. Fig.2.3 shows several examples of this technology.

6

Please see print copy for Figures 2.3

(a) Thyssen Krupp Stahl AG, Germany (CSP : Compact Strip Production) [3]

Please see print copy for Figures 2.3

(b) Aceria Compacta, Spain (CSP) [3]

Please see print copy for Figures 2.3

(c) AST-Terni, Italy (CSP) [3]

7

(d) Cremona, Italy (ISP) [3]

Please see print copy for Figures 2.3

(e) POSCO #1, South Korea (ISP : In-line Strip Production) [47]

Please see print copy for Figures 2.3

(f) Algoma, Canada (FTSC : Flexible Thin Slab Casting)

Please see print copy for Figures 2.3

(g) CORUS Ijmuiden, Netherlands (DSP : Direct Sheet Plant) [3]

Fig.2.3 Practical examples of thin-slab casting techniques

8

2.2 Slab cracking 2.2.1 Features of Cracking Transverse cracking, crazing and star cracking are linked by an important common feature. Szekeres [4] suggested that in all these cases the cracks follow the boundaries of exceptionally large prior-austenite grains. On the surface of the as-cast product, the diameters of these grains may vary between 1 and 4mm, and in some instances they are even larger, An example is shown in Fig.2.4 [4]. At some locations the large grains may extend to a depth beyond 6mm. Turkdogan [5] referred to these abnormally large grains as “blown grains”. The diameter of blown grains tends to be greater at the base of oscillation marks as shown in Fig.2.5 [4]. Schmidt and Josefsson [6] also provided evidence of blown grains and found that transverse cracks occurred only in areas where blown grains are present, Fig.2.6 [6]. They suggested that the large grains may be caused by secondary recrystallization. Similarly, Maehara et al. [7] maintained that control of the austenite grain size should be the first priority in preventing surface cracking.

Please see print copy for Figure 2.4

Fig.2.4 Widespread crazing and fine transverse cracks at oscillation marks on the as-cast surface of a line pipe steel slab (top side). Etched in hot HCL. Magnification not specified. CD : Casting Direction [4]

9

Please see print copy for Figure 2.5

Fig.2.5 Crazing around a transverse crack at the base of an oscillation mark on the as-cast top surface of a 0.20% C steel slab. Etched in hot HCL [4].

Please see print copy for Figure 2.6

Fig.2.6 Section through the surface across a system of transverse cracks through the coarse-grained zone. Prior austenite grain boundaries white (ferrite-decorated). Etched in nital. Magnification not specified [6].

2.2.2 Coarse prior austenite grains Maehara et al. [7] provided convincing evidence that austenite grain size increases very rapidly in the temperature range of 1450 to 1350

when specimens are re-melted in-situ and

continuously cooled, Fig.2.7 [7]. Moreover, steel containing 0.16% C had significantly larger grains than other plain carbon steels under similar heat treatments. Because the surface temperature of most strands are below 1300

at the point of mold exit, blown grains must

develop while the surface region is still within the mold.

10

Molten steel level fluctuations in the mold give rise to deep transverse depressions and oscillation marks. Wolf [8] suggested that large prior austenite grains form in such depressions because of locally reduced cooling rates as a result of lack of contact between the solidified shell and mold wall.

Please see print copy for Figure 2.7

Fig.2.7 Grain growth of austenite during continuous cooling. The specimens were remelted at 1580 , cooled to a given temperature at a rate of 0.28 s-1, and then quenched into water [7].

Surface cracking mechanism based on the existence of blown grains during casting is schematically illustrated in Fig.2.8 [4]. Stage

represents newly solidified grains on the mold

wall with very small grain size. The grain diameter at the surface is 500 um or less. Stage

represents blown austenite grains that have grown to a size several times larger than

those shown in stage . In this case, the surface temperature is probably above 1350 . In stage , liquid copper (or a Cu alloy), if present at the scale-matrix interface, penetrates the blown grain boundaries and allows microcracks to initiate. However, if liquid copper is not present, solid state sulfides may precipitate on the blown grain boundaries and be the weakening agent responsible for microcrack formation.

11

In stage , the temperature is low enough to allow precipitation of nitrides, e.g., AlN, Nb(C,N), or V(C,N). Subsequently, or simultaneously, the nucleation and growth of proeutectoid ferrite can weaken the grain boundaries and cause a significant loss of ductility. The size of the virgin austenite grains has a profound effect on the nature of proeutectoid ferrite that precipitates and grows along the grain boundaries. With blown grains, the proeutectiod ferrite forms in a continuous, film-like fashion. When the austenite grain size is small, the ferrite grains are more equi-axed and discontinuous. It is far easier for a crack to propagate along a continuous ferrite film as opposed to a discontinuous one. Because AlN precipitates far more rapidly in ferrite than in austenite, the film can be weakened further by AlN precipitation. In stage , the top side of the strand experiences tension during the straightening operation. Thus, any microcracks that are aligned with decorated blown grain boundaries are easily extended, or new ones develop. If the strand surface temperature at the straightener is above the temperature at which proeutectoid ferrite forms, this stage may actually precede stage .

Please see print copy for Figure 2.8

Fig.2.8 Formation of surface cracks due to blown grain during casting [4]

12

2.3 Hot ductility test 2.3.1 Simulation of straightening operation during continuous casting There are several test methods by which the unbending operation during continuous casting may be simulated. These techniques include: hot-bend testing, hot-compression testing, torsion testing and hot-tensile testing. The hot-bend test most closely simulates the unbending operation but it is difficult to quantify the severity of surface cracks that form during such a test [9,10]. The hot-compression test is carried out on a flanged sample where the hoop strain corresponding to the first appearance of a crack is taken as a measure of hot-ductility [11]. Torsion testing produces large strains and the difficulties in interpreting the fracture appearance after failure makes this test unsuitable [12]. The most popular test for studying the problems of transverse cracking is the simple hot-ductility tensile test. 2.3.2 Suitability of the hot tensile test to the problem of transverse cracking It is clear that laboratory hot-ductility tensile tests do not precisely simulate the straightening operation in continuous casting, A major disparity is the degree of straining involved. In the straightening operation, it is at most only 1~2% [9], whereas, in a hot-ductility test, the fracture strains are in the range 5~100%. Thus, the mechanisms which pertain to tensile tests do not necessarily coincide with those associated with transverse cracking. Notwithstanding these objections, the simple hot tensile test has proven to be the most popular test for the study of transverse cracking on laboratory scale. Generally, tests are carried out in a protective atmosphere using a servo-hydraulic load frame equipped with either a furnace, or an induction heater. The GLEEBLE machine has also proven to be popular for the investigation of hot-ductility because of its ability to melt samples and its versatility in simulating thermal cycles. In this instance, heating is effected by electrical resistance, which has the advantage that there is no practical limit to the rate of heating, and temperature gradients can be kept to a minimum [1].

13

Samples are usually heated to a temperature above the solution temperature of the microalloy precipitates, both to dissolve all these particles and to produce a coarse grain size reminiscent of the continuously cast microstructure before the unbending operation. The sample is cooled at the rate experienced by the surface of the strand during the continuous casting operation and is strained at rates between 10-3 and 10-4s-1 [9,13]. These strain rates are specifically chosen to simulate the strain rates associated with straightening operation. More sophisticated simulations of the continuous casting operation involve actual melting the samples, either by induction or electrical resistance, in a quartz tube placed over the mid span region to retain the liquid. In-situ melting can be combined with the complex cooling patterns that are experienced in the secondary cooling zone before straightening, although unfortunately, this technique has rarely been used. Some investigators [14,15] have used total elongation to fracture as a measure of the ductility, and it can provide useful information regarding the role played by dynamic recrystallisation in influencing the ductility. The majority of researchers have used reduction in area (RA) at fracture to provide quantitative information on the fracture strain. Although the RA at the onset of fracture is probably a better measure of ductility in relation to transverse cracking, the total RA has been generally used because it is easier to measure. This measurement has the advantage that it is independent of the fracture geometry of the sample.

2.4 Fracture mechanisms The general characteristics of ductility as a function of temperature are shown in Fig. 2.9 [1]. Three distinct regions may be (i)

a high ductility, low temperature (HDL) region

(ii)

a deep ductility trough, indicating a region of embrittlement

(iii)

a high ductility, high temperature (HDH) region.

14

It is instructive to briefly discuss the deformation and fracture characteristics of material subjected to tensile stress in each of these regions.

Please see print copy for Figure 2.9

Fig.2.9 Schematic diagram of a ductility curve defining the three characteristic regions of hot-ductility [1]

2.4.1 Region of embrittlement The ductility region is invariably associated with intergranular fracture, the fracture facets on microscale, being either covered with fine dimples or microvoids, or they are smooth. Two distinct fracture mechanisms may be deduced from this difference in appearance of the fracture surface. Preferential deformation in regions close to grain boundaries initiates voids at grain boundary inclusions or precipitates, which leads to intergranular failure via microvoid coalescence and in this instance fine dimples and microvoids are expected to be seen on a fracture surface. However, grain boundary sliding in the single phase austenite region followed by wedge cracking would result in a smooth fracture surface [1].

2.4.1.1 Embrittlement by strain concentration and microvoid coalescence at grain boundaries Two distinctly different microstructural features may lead to strain concentrations at austenite grain boundaries. ▪ thin ferrite films 15

▪ precipitate free zones Plain carbon as well as low-alloyed steels have been shown to be susceptible to intergranular fracture in the temperature region in which unbending of the continuously cast strand is performed [1]. For this reason, it is pertinent to further discuss the likely microstructural features that may lead to a mechanism of intergranular fracture. Ferrite films Intergranular failure may occur when the austenite to ferrite transformation has partially occurred and a thin film of ferrite (~5-20

thick) has formed around austenite grains.

Such a situation is depicted in Fig.2.10 [1]. The comparative ease of dynamic recovery in ferrite translates into a low flow stress compared to austenite, and therefore to strain concentration in the ferrite film. This strain concentration leads to ductile voiding, generally at MnS inclusions located at austenite grain boundaries.

Please see print copy for Figure 2.10

Fig.2.10 Schematic diagram showing mechanism for transformation induced intergranular failure [1]

Strain induced ferrite can be formed at temperatures above the Ar3 temperature (the austenite/ferrite transformation start temperature at a constant cooling rate), and often as high as the Ae3 temperature (the austenite/ferrite transformation start temperature under equilibrium condition) when the tensile test is conducted at these temperatures [16,17], Fig.2.11 [17]. Between

16

the Ar3 and Ae3 temperatures, the thickness of the ferrite film forming around austenite grains does not change significantly with temperature. Hence, also in this case a thin ferrite layer will form around austenite grains. At test temperatures below Ar3 the ferrite film thickens rapidly and the ductility is fully recovered when approximately 50% ferrite is present before the tensile test is conducted.

Please see print copy for Figure 2.11

Fig.2.11 Microstructure at room temperature of steel tested at 800 , showing intergranular failure associated with a thin layer of grain boundary ferrite [17]

Although deformation-induced ferrite can be formed quite readily during straining in the course of hot tensile testing, there is as yet, no convincing evidence of the presence of strain-induced ferrite in coarse grained steels at the low strains (~2%) applied during straightening. Essadiqi and Jonas [18] provided limited evidence that strain induced ferrite can be produced in a fine grained (~25 ) low C, Mo containing steel, deformed to a true strain of 0.016 at a low strain rate of 7.4×10-4s-1. This strain and strain rate are similar to those undergone during the straightening operation, but no such evidence has been provided for coarse grained steels.

Precipitate free zones on grain boundaries In Nb-containing steels that have been solution treated prior to cooling to the test temperature, precipitation of Nb(C,N) occur in austenite during deformation. These carbon-nitrides usually precipitate on austenite grain boundaries and such precipitation is frequently accompanied by the formation of relatively weak precipitate free (and 17

carbon depleted) zones (PFZs) on both sides of the austenite grain boundaries (500nm wide) [19]. Fine matrix precipitation can also take place, leading to significant matrix strengthening. The microstructural situation is then similar to the presence of soft films of deformation-induced ferrite on austenite grain boundaries, and microvoid coalescence fractures are frequently observed. In this case however, void formation takes place at the microalloy precipitates (Nb(C,N) and AlN when Nb and Al are both present). This fracture process is schematically shown in Fig.2.13 [20]. When deformation is induced in the single phase austenite region, grain boundary sliding is also likely to be instrumental in the embrittling process, most probably contributing to crack propagation.

Please see print copy Figure 2.13

Fig.2.13 Schematic illustration of intergranular microvoid coalescence of Nb-bearing steels. a-c depict deformation in austenite above the Ar3 temperature, d-f depict deformation in the (austenite + ferrite) region

2.4.1.2 Grain boundary sliding Grain boundary sliding followed by crack propagation occurs in austenite because it displays limited dynamic recovery [1] giving rise to high flow stresses and work hardening rates. These high stresses, in turn, prevent the accommodation, by lattice deformation, of the stresses built up at triple points or grain boundary particles, and this series of events may lead to intergranular failure by the nucleation of grain boundary cracks. This rupture mechanism is usually associated with 18

creep, the latter occurring at strain rates typically below 10-4s-1. However, fractures characteristic of failure initiated by grain boundary sliding are frequently found at the strain rate generally used in hot tensile testing (10-3s-1). Furthermore, Ouchi and Matsumoto [21] have observed grain boundary sliding at strain rates as high as 10-1s-1 in a 0.054%Nb steel strained in tension at 900 . Therefore much insight may be gained from a study of the creep literature about the factors that influence intergranular crack nucleation and growth at high temperatures [22]. Traditionally, intergranular creep defects have been classified as either ‘grain edge’ (r-type, r for rounded), or ‘grain corner’ (w-type, w for wedge) cavities. Both types of cracks are observed in samples tensile tested at strain rates in the range 10-3s-1 to 10-4s-1, and both require grain boundary sliding to nucleate cracks. The models proposed for the formation of w-type cracks are illustrated in Fig.2.14 [23]. These models are important at high temperatures, where other embrittling mechanisms, associated with precipitates and the thin ferritic films, are not viable.

Please see print copy for Figure 2.14

Fig.2.14 Schematic models showing the formation of wedge cracks by grain boundary sliding: The arrows indicate the sliding boundaries and the sense of translation [23]

19

2.4.2 High ductility, low temperature region In the high ductility, low temperature (HDL) region, which coincides with a relatively high volume fraction of ferrite, the strain is no longer concentrated in a thin ferrite film at austenite grain boundaries. Furthermore, the difference in strength between austenite and ferrite decreases with decreasing temperature, thus increasing plastic strain in the austenite and, more importantly, decreasing the strain in the ferrite [24]. The concentration of strain at the grain boundaries is thus minimized, and high ductilities are observed. In Ferrite, dynamic recovery, which is a softening process that operates at all strains, readily takes place [25]. Generally, the ductility is very good when high percentages of ferrite are present in the microstructure, in the vicinity of 700

[16,17]. At this temperature, recovery in the ferrite takes

place with ease, the subgrain size is large, and the flow stress is low. Thus, ferrite flows readily at triple points to relieve stress concentrations, therefore discouraging the initiation of cracks.

2.4.3 High ductility, high temperature region One obvious reason for the improvement in ductility in the high temperature region is the absence of the thin ferrite films. In this region failure occurs either by grain boundary sliding or through strain concentration in the PFZ. However, higher temperatures also lead to less precipitation in the matrix and at the grain boundaries. Finally, increased temperatures lead to lower flow stresses via increased dynamic recovery so that stress concentrations at the crack nucleation sites are reduced [1]. At higher temperature such as in the HDH region, grain boundary migration may occur, leading to increased ductility. Cracks can grow intergranularly during the early stages of deformation, and become isolated within the grains as a result of grain boundary migration. The original cracks are then distorted into elongated voids, until final failure occurs by necking between these voids. One way to achieve a high driving force for grain boundary migration is by dynamic 20

recrystallisation. It is therefore not surprising that the HDH region has been observed to coincide with the onset of dynamic recrystallisation [26,27].

2.4.4 Hot ductility behaviour of steels 2.4.4.1 Plain C-Mn and C-Mn-Al steels For steel with carbon contents below 0.3% and low in Al and N, fractography of samples which have failed intergranulary reveals that the coarse grain surfaces (~30

grain diameter) are covered

by cavities [16], Fig.2.15. These cavities are caused by microvoid coalescence within the thin ferrite films. The increased ductility at lower temperatures in the HDL region therefore corresponds to an increased volume fraction of ferrite. On the other hand, the increased ductility at higher temperature in the HDH region is attributed to the absence of ferrite films on austenite grain boundaries as well as the possibility that grain boundary

Please see print copy for Figure 2.15

Fig.2.15 Intergranular microvoid coalescence type fracture in C-Mn-Al steel, solution treated at 1350 , tested at 850

at strain rate of 10-3s-1 [16]

migration may isolate small cracks and hence prevent the formation of cracks exceeding the critical crack length [1].

2.4.4.2 C-Mn-Al steels with high Al and N levels At the high temperature end of the trough, fractured samples display intergranular failure (Fig.2.16) with flat facets and no evidence of microvoid coalescence or ferrite formation [16,17]. 21

Embrittlement is therefore by grain boundary sliding in austenite, accompanied by crack nucleation at triple points. In case of the high Al-N steels, it is likely that AlN precipitates are formed on the austenite grain boundaries, pinning the boundaries and allowing the cracks formed by grain boundary sliding to join up, as well as encouraging void formation [1].

Please see print copy for Figure 2.16

Fig.2.16 (a) C-Mn-Al steel showing flat facets on fracture surface, (b) enlarged view of a showing lack of voiding around MnS inclusions [16]

Lowering the tensile test temperature to below Ae3 will introduce deformation induced ferrite, giving rise to embrittlement by microvoid coalescence. Thus, in these steels, the ductility trough is a result of both grain boundary sliding and microvoid coalescence in the ferrite. A further reduction in temperature increases the volume fraction of ferrite and the recovery of ductility. Since the high temperature side of the trough occurs in the single phase austenite region, ductility recovery in the HDH region is by grain boundary migration, aided by the dissolution of AlN precipitates.

2.4.4.3. Microalloyed steels 22

In microalloyed steels precipitation occurs at austenite grain boundaries, often generating precipitation free zones and thus leading to embrittlement in the single phase austenite region by a combination of microvoid coalescence and grain boundary sliding or shear. Fine precipitates may pin the grain boundaries, allowing the cracks to join up. Thus, intergranular fractures at the high temperature end of the trough are of mixed character, containing flat facets as well as coalesced microvoids. Lowering the temperature increases the amount of intergranular microvoid coalescence, until the fractures are entirely due to microvoid coalescence in deformation-induced ferrite. The mechanisms of fracture in the HDH and HDL regions are essentially those pertaining to high Al-N steels, described above [1].

23

2.5 Factors influencing hot ductility 2.5.1 Grain Size The high temperature ductility increases as the grain size is decreased. When the failure is intergranular, refining the grain size affects crack growth via : A.

The decrease in the crack aspect ratio, which controls stress concentration at the crack tip [30].

B.

The difficulty in propagating smaller cracks formed by sliding through triple points [30].

C.

The increase in the specific grain boundary area (for a given volume fraction of precipitate), which reduces the precipitate density on the grain boundaries [11, 21].

D.

The reduction in the critical strain for dynamic recrystallisation by increasing the number of grain boundary nucleation sites [31].

The influence of grain size on the hot ductility of C-Mn steel having 0.19 and 0.65wt%C is illustrated in Fig.2.17 [32]. The grain sizes before deformation were varied by heating cylindrical samples to temperatures of 925-1330

and holding for 15minutes using Instron tensile tester and

GLEEBLE machine. And then the samples were tested to failure at a strain rate of 3×10-3s-1 after cooled to the test temperature of 550-950 . For the 0.19wt%C steel, in the finer grained steels with grain size below 180

, the ductility trough starts close to the Ar3 temperature when films of

ferrite form around the stronger austenite grains. It is difficult to observe deformation induced ferrite because of the much higher austenite to ferrite transformation rate resulting from the fine austenite grain size [17]. The narrow trough in a fine grained steel is then a consequence of the rapid increase in volume fraction of the ferrite which forms when the temperature is lowered to below the Ae3. In the coarse grained steels, on the other hand, deformation induced ferrite can have a pronounced influence on hot ductility over a wide range of temperatures leading to a wide and deep ductility trough. In this case deformation can raise the Ar3 temperature to almost the Ae3 temperature. The volume fraction of deformation induced ferrite is always small in the coarse 24

grained steels, often over a wide temperature range from the Ae3 to the Ar3 (undeformed) because there is insufficient deformation away from the boundary regions to increase the volume fraction of ferrite significantly.

Please see print copy for Figure 2.17

Fig.2.17 Hot ductility curves for C-Mn steels at various grain sizes [32]

Refining the grain size has an even greater influence on the hot ductility of the 0.65wt%C steel as shown in Fig.2.17 (b). Intergranular fracture in coarse grained steel occurs by grain boundary sliding in the austenite resulting in a very wide ductility trough. Raising the carbon level increases the activation energy for dynamic recrystallization, and hence, more grain boundary sliding than grain boundary migration is expected to account for deformation. It is probable that intergranular failure in these steels now only occur below the Ar3 temperature. 25

2.5.2 Precipitation The effect of precipitates on hot ductility depends strongly on their size distribution and interparticle spacing. These characteristics are controlled by composition and the thermomechanical history of the cast strand. Fig.2.18 [1] shows the influence of particle size and interparticle spacing of Nb(C,N) precipitates at the austenite grain boundaries on the hot ductility of C-Mn-Nb-Al steel. The results of the surface examination of plates commercially produced from continuously cast slabs are also shown in the figure. Plates that have been rejected because of the presence of surface cracks contained Nb(C,N) precipitates with mean particle sizes of less than 14nm and interparticle spacings of less than 60nm. Mintz et al. [13] proposed that the excessive surface cracks observed in slabs that contained fine precipitates less than 14nm, pins the boundaries and when deformation occurs by grain boundary sliding in the austenite, cracks are allowed to join up. Microvoid coalescence failures are also encouraged by an increase in the precipitate or inclusion density at the austenite grain boundaries, these being preferential sites for void initiation, and hence when the interparticle spacing is low, typically less than 64nm, excessive cracks can occur.

Please see print copy for Figure 2.18

26

Fig.2.18 Influence of (a) particle size and (b) interparticle spacing on hot ductility of Nb-containing steels, solution treated at 1330 , cooled to a test temperature of 850 , and fractured at a strain rate of 3×10-3s-1 [1]

The precipitation of AlN, Nb(C,N), and V(C,N) in austenite is accelerated by deformation. This acceleration of precipitation in deformed austenite compared the rate of precipitation in undeformed austenite at the same temperature results from the favorable nucleation sites, such as dislocation networks and vacancy clusters, provided by deformation [1]. Both Nb(C,N) and VN can precipitate rapidly during testing at a strain rate <10-1s-1, and hence the precipitation of these carbides occurring during the test can have a very important influence on the experimentally observed hot ductility and transverse cracking. For steels which are solution treated and cooled to the test temperature, Nb extends the ductility trough to higher temperatures than V because the maximum rate of Nb(C,N) precipitation in undeformed austenite occurs at 950 , whereas it is about 885

for VN. Al steels have even narrower ductility troughs because

the maximum rate of AlN precipitation occurs at 815

[33,34].

Under cooling rates typically experienced in continuous casting, AlN will not precipitate in austenite either statically or dynamically, because nucleation is inhibited. However AlN can easily precipitates in ferrite, and thermal cycling through the ferrite/austenite transition encourages such precipitation. Although the bulk cooling rate of a continuously cast slab is roughly constant, temperature cycling that occurs at the strand surface can accelerate AlN precipitation. Depending on the secondary cooling conditions, it is possible for the temperature to fall periodically below the Ar3, particularly at slab corners. This can be very detrimental to the ductility, because AlN precipitates mainly at austenite grain boundaries [35]. In steels containing aluminum and vanadium, the precipitation of carbo-nitrides that occur before deformation (the static precipitate) is more detrimental than those formed during testing, although they are coarser, because they form only at grain boundaries. By contrast, in Nbcontaining steels, dynamic precipitation occurs more readily and reduces hot ductility significantly 27

[15,27]. Moreover, if such steels are solution treated before deformation so as to take carbides and nitrides into solution, very fine Nb(C,N) is formed on the austenite boundaries and within the matrix. Conversely, if the test is done under direct casting conditions, the amount of Nb available to form fine deformation-induced Nb(C,N) precipitation will be reduced because coarse niobium carbo-nitrides precipitate interdendriticly during solidification [36]. Under solution treated testing conditions, Ti additions to steel are most effective in maintaining hot-ductility and reducing or eliminating the ductility trough because TiN and TiN-rich precipitates form at high temperatures close to the solidus and tend to be coarse and randomly distributed having a high enough volume fraction to restrain grain growth at high temperatures (~1350 ) [1]. The benefits of Ti addition therefore result from the refinement of grain size and the ability to combine preferentially with N, thereby preventing the formation of AlN or Nb(C,N) precipitates. In industrial continuous casting, however, Ti addition does not significantly influence the grain size obtained during cooling after solidification [37]. However, the knowledge about the influence of Ti appears to be incomplete as laboratory investigations do not always agree with operational experiences.

2.5.3 Composition 2.5.3.1 Sulphur The effect of sulphur on hot ductility depends largely on the test conditions. For steels solution treated at 1330 , it is the amount of S which can be redissolved at 1330

and subsequently

reprecipitated as sulphides in a fine form at grain boundaries that is most detrimental to ductility [1, 38, 39, 40]. Dissolution of sulphides allows S to segregate to grain boundary precipitate/matrix interfaces as well as austenite grain boundaries. This leads to enhanced nucleation of microvoids and a loss in ductility. The amount of sulphur that redissolves depends on the manganese content of the specimen [39]. Using the solubility data of Turkdogan [41], for a steel containing 1.4% Mn, 28

the amount of sulphur redissolved at 1330°C is ~0.001%. Accordingly, once the sulphur content becomes high enough to reach the maximum dissolvable amount at the solution treatment temperature, the hot ductility behaviour will be independent of the sulphur level. Under direct casting conditions, it is the total sulphur level that is important for controlling hot ductility [1, 38, 39]. Fig.2.19 [51] illustrates the effect of sulphur on RA at two cooling rates, 1°Cs1

and 4°Cs-1, where the ductility decreased as the sulphur content is increased. Abushosha [42]

found that increasing the sulphur level caused lower ductility for C-Mn-Al steels and C-Mn-Al-Nb steels under direct casting conditions. Increasing the sulphur level increases the volume fraction of sulphides formed on solidification, reducing ductility. The reduction in ductility is related to the increase in the volume fraction of sulphides at the interdendritic boundaries. The interdendritic boundaries later form the austenite grain boundaries; as a consequence sulphide particles are located at austenite grain boundaries. These sulphide inclusions enhance intergranular failure in austenite by preventing either austenite grain boundaries from migrating or encourage voiding during grain boundary sliding [1, 39]. Sulphide inclusions are also instrumental in accelerating intergranular failure in thin, deformation induced, ferrite films formed at lower temperatures.

Please see print copy for Figure 2.19

Fig.2.19 Effect of sulphur content on the minimum reduction of area for two cooling rates of 1°Cs-1 and 4°Cs-1 [51]

In direct cast steels that have been reheated, segregation that occured during solidification will 29

influence the amount of sulphur that can redissolve at the holding temperature. The regions surrounding the sulphides are depleted in Mn, so a higher volume fraction of sulphur can be redissolved on reheating [39]. Hence, if specimens are in the as-cast condition and reheated, they may have a lower ductility than if the specimen were only reheated to the solution treatment temperature. Sulphur can impair hot ductility by forming precipitate free zones (PFZs), allowing strain to concentrate in the weaker regions (where there are no precipitates) in the vicinity of austenite boundaries. In specimens exhibiting poor ductility, PFZs have been found responsible for encouraging intergranular failure. Dense precipitation at the prior austenite grain boundaries and an extremely fine dispersion of sulphides in the matrix are observed in these steels [1, 38]. It is generally recommended that sulphur levels be kept to a minimum to avoid transverse cracking. Reducing sulphur levels reduces the volume fraction of sulphides available for precipitation on solidification. Calcium treatment of liquid steel has been shown to be beneficial for improving hot ductility by modifying the sulphides. Sulphur is bound in these modified sulphides and hence, they show little ability to redissolve at 1330°C, thus, reducing the amount of free sulphur available for precipitation. Calcium additions to liquid steel will also reduce the total amount of sulphur in the steel by removal of sulphur in the slag (important for direct cast specimens) and reduce precipitation in this way [1, 42].

30

2.5.3.2 Carbon The carbon content largely determines hot cracking susceptibility of low alloy steels during continuous casting, as shown in Fig.2.20 [7]. Maximum cracking susceptibility is found in the peritectic composition region, 0.10 to 0.16%C. Hot-tensile tests on reheated specimens fail to show this carbon dependence, and the cause of the ductility loss in the medium C steels cannot be ascribed merely to the uneven solidification on the surface of slabs due to the peritectic reaction. The carbon content dependence can be ascribed to microstructural changes during solidification. If intergranular failure in austenite is the dominant failure mode, then hot-ductility of direct-cast specimens will depend largely on the austenite grain size. This is described in Fig.2.21 [7]. Because austenite forms at high temperatures in the peritectic region, large grains develop due to rapid grain growth. For this reason the largest austenite grain sizes are generally found in steels with carbon contents close to the peritectic composition [38].

Please see print copy for Figure 2.20

Fig.2.20 Schematic illustration showing the effect of C on surface cracking of continuous cast slabs [7]

31

Please see print copy for Figure 2.21

(a) Relationship between austenite grain size and C content. Specimens (0.35%Si-1.5%Mn-0.05%Nb) were remelted at 1580 , cooled to 900

at a rate of 5 s-1 and then quenched.

(b) Calculated values RA which were deduced from the relationship between grain size and RA assuming that the specimens were deformed at 800

at a strain rate of 0.83×10-3s-1

Fig.2.21 Effect of C content on austenite grain size and calculated value of RA [7]

Austenite grain growth is largely retarded by the presence of a small amount of a second phase and especially by the presence of fine grain boundary precipitates. Thus the grain size will mainly be determined by the austenitizing temperature, and will be a maximum in steels with the peritectic composition. In steels of hypo-peritectic composition, the second phase is delta-ferrite. In steels of hyper-peritectic composition, however, a liquid phase will be present up to much lower temperatures [7]. Although microalloyed steels exhibit poor ductility in the temperature range 1000-700 , a ductility trough is present even in plain C-Mn steels. The hot-ductility behaviour observed by Crowther and Mintz [6] was a strong function of the carbon content. For steels containing more than 0.3%C, troughs up to ~200

wide were obtained having minimum RA values close to 30%. 32

In steels with carbon contents lower than 0.3%, minimum RA values were found but the ductility troughs were only about 50-100

wide. It has been shown that the width of the trough decreases

with decreasing carbon and manganese contents, Fig.2.22 [39]. The trough in the hot-ductility curve of low-carbon C-Mn steels is mainly a result of the presence of thin films of deformationinduced ferrite which form around the austenite grain boundaries. In higher carbon steels, grain boundary sliding in the austenite can be responsible for the loss in ductility at temperatures above Ae3.

Please see print copy for Figure 2.22

Fig.2.22 Influence of C and Mn on the width of the ductility trough [39]

33

2.5.4 Cooling rate Generally, increasing the cooling rate in the range 25 to 240 min-1, results in lower ductility for most types of steel. In most cases, the decrease in ductility with increasing cooling rate is ascribed to either the formation of finer precipitates or finer inclusions [40]. For C-Mn steels, a finer MnS distribution in the ferrite film surrounding the austenite grains, as well as a reduction in thickness of the ferrite film, can lead to the deterioration in ductility at increased cooling rates [40]. In the case of C-Mn-Al steels, the deterioration in ductility is due to finer AlN precipitation and/or a finer sulphide inclusion distribution [40, 43]. For C-Mn-Al-Nb steels, an increase in cooling rate can lead to a larger amount of Nb being held in solution, resulting in an increase in finer, more detrimental, strain-induced Nb(C,N) precipitation [43]. It is pertinent however, to point out that although a reduction of the cooling rate of the strand may increase the resistance to transverse cracking through coarsening of the precipitates and a reduction in the thermal stresses in the strand, an increased cooling rate may result in ductility improvements through grain refinement [1].

2.5.5 Strain rate With respect to straightening operations where the strain rate varies between ~10-3-10-4s-1, an increase in the strain rate significantly improves hot ductility. The fracture appearance changes from intergranular to ductile by an increase in the strain rate [1, 38, 40]. This improvement of hot ductility with strain rate may be attributed to the following [1, 21, 44]: . Insufficient time to allow for strain induced precipitation, . Less grain boundary sliding, . Insufficient time for the formation of voids near the precipitates or inclusions at grain boundaries . Prevention of the formation of deformation-induced ferrite 34

In commercial operations changes in casting speed can only increase the strain rate by a factor of two and hence, it is not possible to change the cracking behavior by a change in casting steed through its influence on strain rate alone but a change in casting speed can still modify the temperature distribution and thus have an indirect effect on cracking behavior [1, 45].

2.5.6 Thermal history The actual thermal cycle at the surface of a continuously cast steel slab is very complex because of the temperature oscillation introduced by the alternate impingement of water sprays into the slab and contact with the rolls [35]. Computer simulations [46] have shown that there is very large temperature drop just below the mold, where the surface temperature reaches 600-700 . The surface is then reheated by heat transfer from the interior of slab to over 1000 , after which it is cooled more steadily along the rest of the strand. This cyclic behavior of the surface temperature has been incorporated into tensile tests to simulate actual continuous casting. The results showed that, if the temperature drops below the final test temperature during cycling, increased fine precipitation at the test temperature reduces ductility. This has been demonstrated for C-Mn-Al and C-Mn-Nb-Al steels [17, 36, 46]. Furthermore, if the temperature falls below the transformation so that ferrite is present, precipitation is further enhanced due to the lower solubilities of nitrides and carbonitrides in ferrite than in austenite. This is again expected to be detrimental to hot-ductility.

35

CHAPTER 3 - EXPERIMENTAL 3.1 Preparation of specimens The specimens used in this investigation were prepared in a laboratory scale vacuum induction melting furnace. Table 3.1 shows the chemical composition of these specimens. In order to eliminate the overriding effect of precipitation of alloying element at the austenite boundaries masking any influence of grain size, tramp element and impurities were kept as low a level as possible. Nevertheless, Al was used as deoxidizer and some ingress of N occurred. Solidified casting ingots were reheated to 1200

and then hot-rolled into 15mm thick plates using a

laboratory scale rolling mill. Cylindrical GLEEBLE specimens, with 10mm diameter and 115mm in length were machined from these plates and the details of such specimens are shown in Fig.3.1.

Table 3.1 Chemical composition of specimens Specimens

Elements, wt% C

P

Mn

Si

S

A

0.05

0.002

<0.01

<0.005

0.002

B

0.18

0.002

<0.01

<0.005

C

0.45

0.002

<0.01

<0.005

Ni,Cr,Mo

Al

Nb

Ti, V

N

<0.002

0.016

<0.001

<0.003

0.0015

0.002

<0.002

0.034

<0.001

<0.003

0.0016

0.002

<0.002

0.025

<0.001

<0.003

0.0016

,Cu,Sn

Thermocouple Type-R (Pt-Pt13%Rh)

10 mm 115 mm Fig.3.1 Geometry of GLEEBLE specimen

3.2 Hot ductility test

36

3.2.1 GLEEBLE3500 GLEEBLE3500 is a thermomechanical simulator in which the samples are heated by electrical resistance. Sample cooling can be achieved by several methods such as simple cooling by anvil/grips or by any combination of air/inert gas/water quenching. Testing can be done in either ambient or inert atmosphere or alternatively under vacuum. The testing is done fully automatic, or with a limited manual control. It is possible to deform under uniaxial tensile, uniaxial compressive or plane strain compressive conditions at the desired temperature. The time – temperature – deformation program is written on a desk-top computer through the use of table programming. The table program is converted automatically to an executable script program by the machine software. Script programs are executed, not table programs. Machine capabilities, as specified by the manufacturer, are: . Heating rate

up to 3000 s-1.

. Cooling rate

up to 600 s-1.

. Ram speed

up to 1000mms-1 (which affords a very high strain rate).

. Maximum Load

10 tons.

A schematic diagram of the GLEEBLE testing arrangement is shown in Fig.3.2. In order to control the temperature of a specimen, a type-R thermocouple (Pt / Pt-13%Rh) is spot-welded onto the specimen at the middle of the span.

37

Fig.3.2 Schematic diagram of the GLEEBLE testing arrangement

3.2.2 Thermomechanical cycles Fig.3.3 shows the thermomechanical cycles imposed during hot ductility tests. In order to attain different grain sizes, specimens were solution treated at pre-determined temperatures between 1100

and 1350

for 10minutes in order to allow sufficient grain growth, Fig.3.3 (a). The

specimens were then cooled to the test temperature, in the range 1100

to 700

at a rate of

200 min-1. After holding for 1 minute at the test temperature, the specimens were pulled to fracture at a low strain rate of 7.5×10-4s-1, simulating the strain rate obtained during straightening of cast slab. Tensile tests were conducted at temperature intervals of 50 1100

in the temperature range

to 700 . Tests were carried out in a vacuum of approximately 10-3 atm.

In order to simulate direct casting conditions, specimens were melted and then cooled to the test temperature (Preliminarily experiments conducted to determine the melting point will be discussed in Section 3.2.5). In order to prevent the specimen from collapsing during melting, quartz crucibles covering the middle part of GLEEBLE specimen were used to contain the melted zone of the specimen. Two different cooling rates, 100 min-1 to simulate conventional continuous casting and 200 min-1 for the simulation of thin-slab continuous casting, were used to study the effect of cooling rate on hot ductility, Fig.3.3 (b). 38

For the purposes of metallographic examination, specifically to determine grain size, samples were subjected to the same heat treatments and water quenched after cooling to the test temperature but without applying deformation. 1100-1350¡ É 10m in 200¡ É /m in

7.5¡ ¿ 10-4s-1

10¡ É /s

700-1050¡ É 1m in

Tim e

(a) Solution treatment M elting 20sec 100¡ É /m in

200¡ É /m in

7.5¡ ¿ 10-4s-1 700-1050¡ É

1m in

10¡ É /s

Tim e

(b) Direct casting Fig.3.3 Schematic diagram of thermomechanical cycles for hot ductility tests under the conditions of (a) Solution treatment and (b) Direct casting

3.2.3 Measurement of hot ductility Reduction in area at fracture has been a most popular method for assessing hot-ductility. Fig.3.4

39

shows the geometry of sample after fracture. Reduction in area at fracture was calculated as follows:

RA (%) = (A0 A1) / A0 × 100

Where:

(3.1)

A0 = cross-sectional area before test A1 = cross-sectional area after fracture

A1 was calculated by an average of four measurements.

A1

A0

Fig.3.4 Geometry of sample after fracture

3.2.4 Correction of stress-strain curve The load-curve sensed by the load-cell during a tensile test indicates that there is a residual force remaining after the specimen has been fractured as shown in Fig.3.5. Both Friction between the ram and chamber and software resetting errors seemed to cause this effect. It is therefore necessary to correct the load obtained from GLEEBLE load cell every time. The net tensile force was obtained by shifting the load curve down (or up) by the amount of the residual load. After doing this, modified stress-strain curves could be obtained. Fig.3.6 shows an example of this correction showing that exactly the same result was obtained from two different experiments using the same steel grade specimens under the same test conditions.

40

350 Mesured tensile force from load cell

300

Force (kgf)

250 200 150

Net tensile force

100 50 0 1000

1200

1400

1600

1800

Time (sec)

Fig.3.5 Measured tensile force during tensile deformation (b)

(a) 40

40

1

Stress (MPa)

30

30

20

20

1

2 10

10

0

0

-10

-10

0.0

0.1

0.2

0.3 0.4 Strain

0.5

0.6

0.7 0.0

2

0.1

0.2

0.3 0.4 Strain

0.5

0.6

Fig.3.6 Stress-strain curve. (a) curves obtained from GLEEBLE (b) modified curves

41

0.7

3.2.5 Determination of GLEEBLE setting temperature for melting There exists a radial thermal gradient in GLEEBLE specimens due to heat loss at the specimen surface. This heat loss is mainly due to radiative heat loss under vacuum conditions. Heat loss by convection in vacuum is reported to be about 0.1% of the total heat loss [48]. In the case of specimen temperatures above 1400 , this temperature difference between the surface and center of a sample is approximately 100°C [49]. Because only surface temperature of a sample is measured in a GLEEBLE test and this temperature is used to control the heating current, it was necessary to determine the apparent melting point of the three specimen types used before a melting experiment could be conducted. Several preliminary experiments were conducted to determine an accurate estimate of this apparent melting point. Samples within a quartz crucible in its center span was heated up to 1500

a very slow rate of 0.2 sec-1 under a small compressive force of 8kgf. During heating the

stroke (the relative distance between the two rams) as well as the temperature was recorded. This experiment was repeated 2-3 times. Fig.3.7 shows the effects measured. Specimen starts to melt from the center of specimen due to thermal gradient. The stroke increases or may become stable until some outer part of specimen begin to melt. As melting of outer zone is initiated, the specimen seems to collapse resulting in a decrease of the stroke (because there is no resisting power against the applied compressive force). Sometimes the thermocouple detached just before or after the stroke collapse which resulted in the loss of experimental control. The GLEEBLE setting temperatures for the melting experiments were set at 1440 , 1440 0.05%C, Fe-0.18%C and Fe-0.45%C alloys respectively.

42

and 1430

for the Fe-

1460 0.64

1442? ¡ É

1440

1420

1400

0.60

1380 0.58

Temperature (? ¡ É)

Stroke (mm)

0.62

1360 0.56

1340 200

250

300

350

400

450

Time (sec)

(a) Fe-0.05%C alloy 1.1

1460

1444? ¡ É

1440

1420

0.9

1400 0.8 1380

Temperature (? ¡ É)

Stroke (mm)

1.0

0.7 1360 0.6 350

400

450

500

550

600

650

Time (sec)

(b) Fe-0.18%C alloy 1460

0.74

1440

1435? ¡ É

1420

0.70

1400 0.68 1380

Temperature (? ¡ É)

Stroke (mm)

0.72

0.66 1360 0.64 350

400

450

500

550

600

Time (sec)

(c) Fe-0.45%C alloy Fig.3.7 The results of preliminary experiments for the determination of the apparent melting point, (a) Fe0.05%C alloy (b) Fe-0.18%C alloy (c) Fe-0.45%C alloy

43

3.3 Metallography In order to establish the relationship between the heat treatment condition and austenite grain size, specimens were heat-treated within the GLEEBLE. Heat treated specimens were then cut into appropriate sizes using an Accutom precision cutting machine. Samples were polished up to 1

finishing. In order to delineate prior austenite grain boundaries, the following etching technique was used.

Ten drops of hydrochloric acid and ten drops of teepol were added to 70ml of saturated picric acid. The etchant was heated to 67°C and polished samples were immersed for 10-15 minutes in this solution. In some cases, nital (2.5% nitric acid) was used as an etching agent.

44

CHAPTER 4 - RESULTS

4.1 Hot ductility curves In this chapter, the temperature range in which a ductility trough is found, is defined as the range in which the RA value is below 40%.

4.1.1 Fe-0.05%C alloys Fig.4.1 shows the reduction in area (RA) as a function of test temperature for Fe-0.05%C alloys. For the solution treatment condition, at temperatures below 800 , almost the same RA value is found. On the other hand, at temperatures above 900

higher solution treatment temperatures

resulted in lower RA values. In the temperature range 830-1020

a ductility trough was found for

specimens solution treated at 1350 . The minimum RA values are 15.7% at 850 , 40.1% at 900 and 65.6% at 800 , for solution treatment temperatures 1100 , 1200

and 1350

respectively.

For direct cast condition, the RA values at a cooling rate of 200 min-1 were slightly higher than that for 100 min-1 at all test temperatures.

100%

RA (%)

80%

60% 40%

20% 0% 650

1100 1200

, solution , solution

1350

, solution

700

750

800

850

900

950

Test temperature ( )

(a) solution treatment

45

1000

1050

1100

100%

RA (%)

80%

60%

40%

20%

0% 650

700

750

800

850

900

Test temperature (

100

/min, melting

200

/min, melting

950

1000

1050

1100

)

(b) direct casting condition

Fig.4.1 Hot ductility curves for Fe-0.05%C alloys under (a) solution treatment condition (b) direct casting condition

When specimens were solution treated at 1100

or 1200

and tested at 850 , it was not

possible to obtain RA values because neck formation did not always coincide with the center of the specimen in axial direction as shown in Fig.4.2.

Fig.4.2 Sample geometry after fracture tested at 850

under solution treatment condition

4.1.2 Fe-0.18%C alloys Fig.4.3 shows RA curves as a function of test temperature for Fe-0.18%C. When specimens were solution treated, the ductility trough was present at 780-860 , 770-880 46

and 760-960

for

solution treatment condition temperature 1100 , 1200

and 1350

respectively. At test

temperatures below 800 , the curves show similar RA values for all solution treatment temperatures. As in the case of Fe-0.05%C alloy, the higher solution treatment temperatures cause poor ductility when the test temperature is above 850

especially for specimens solution treated

at 1350 . The minimum RA value was 11.0% at 850 , 13.2% at 850 solution treatment temperature 1100 , 1200

and 1350

and 32.3% at 800

respectively.

For the direct casting condition, the curves showed the minimum ductility at 800 cooling rates.

100% 1100 , solution 1200 , solution

80%

RA (%)

1350 , solution

60% 40% 20% 0% 650

700

750

800

850

900

950 1000 1050 1100

Test temperature ( )

(a) solution treatment

47

for

for both

100%

RA (%)

80% 60% 40% 100 /min, melting

20% 0% 650

200 /min, melting

700

750

800

850

900

950

1000 1050 1100

Test temperature ( )

(b) direct casting condition Fig.4.3 Hot ductility curves for Fe-0.18%C alloys under (a) solution treatment condition (b) direct casting condition

4.1.3 Fe-0.45%C alloys The measured reduction in area (RA) as a function of temperature for Fe-0.45%C alloys is shown in Fig.4.4. Unlike the other alloys, the RA value peaked at 700

in both heat treatment

conditions. In case of solution treatment condition, the ductility trough began at 850 , 890 solution treatment temperature 1100 , 1200

and 1350

and 970

for

respectively. Ductility was not

recovered at temperatures below 650 . At all test temperatures, a higher solution treatment temperature showed the lower RA value. For direct cast condition, the RA values at a cooling rate of 200 min-1 were slightly higher than that for 100 min-1 in all test temperatures.

48

100% 1100 , solution 1200 , solution 1350 , solution

RA (%)

80% 60% 40% 20% 0% 600

650

700

750

800

850

900

950

1000 1050

950

1000

Test temperature ( )

(a) solution treatment 100% 100 /min, melting

RA (%)

80%

200 /min, melting

60% 40% 20% 0% 600

650

700

750

800

850

900

1050

Test temperature ( )

(b) direct casting condition Fig.4.4 Hot ductility curves for Fe-0.45%C alloys under (a) solution treatment condition (b) direct casting condition

49

4.2 Stress-strain curves 4.2.1 Fe-0.05%C alloys The stress-strain tensile test curves for Fe-0.05%C alloys are shown in Fig.4.5 and Fig.4.6 under solution treatment condition and direct casting condition respectively. At lower test temperatures, once the maximum stress was reached, the stress dropped rapidly to failure. But, at higher test temperatures, more ductility was retained. Fig.4.7 shows peak stress as a function of test temperature. By and large, the peak stress decreases with increasing test temperature as expected, except in the region 800-950

for solution treatment condition and 800-900

for direct

casting condition. 50

50

50 700¡ É

700¡ É

700¡ É

40

40

800¡ É

750¡ É

Stress (MPa)

830¡ É

900¡ É

750¡ É

30

40

800¡ É

750¡ É

950¡ É

30

870¡ É

30

800¡ É

20

20

1000¡ É

20 1000¡ É

10

10

10

1050¡ É

1000¡ É 1100¡ É 950¡ É 1050¡ É

0 0.0

0.1

0.2

0.3

0.4

0.5

0 0.6 0.0

0.1

0.2

0.3

Strain

(a) 1100

0.4

0.5

0.6

0 0.7 0.0

0.1

0.2

Strain

(b) 1200

0.3

0.4

0.5

Strain

(c) 1350

Fig.4.5 Stress-strain curves at different test temperatures for Fe-0.05%C alloys solution treated at (a) 1100 (b) 1200

(c) 1350

50

0.6

50

50

700¡ É

40

700¡ É

40

Stress (MPa)

900¡ É 800¡ É

900¡ É

30

30

800¡ É

1000¡ É

20

1000¡ É

20

10

10

0

0 0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7 0.0

0.1

0.2

Strain

0.3

0.4

0.5

0.6

0.7

Strain

(a) 100 min-1

(b) 200 min-1

Fig.4.6 Stress-strain curves at different test temperatures for Fe-0.05%C alloys under direct casting condition at cooling rate (a) 100 min-1 (b) 200 min-1

60

60

100 /min, melting

1100¡ É , solution

200 /min, melting

1350¡ É , solution

40

Peak stress (MPa)

Peak stress (MPa)

1200¡ É , solution

20

40

20

Ae3

0 650

700

750

800

850

900

950

1000

1050 1100

1150

0 650

700

750

800

850

900

950

Test temperature(

) Test temperature(¡ É

(a) solution treatment condition

1000 1050 1100 1150 )

(b) direct casting condition

Fig.4.7. Peak stress as a function of test temperature for Fe-0.05%C alloys under (a) solution treatment condition (b) direct casting condition

51

4.2.2 Fe-0.18%C alloys The stress-strain curves for tensile tests of Fe-0.18%C alloys are shown in Fig.4.8 and Fig.4.9 under solution treatment conditions and direct casting conditions respectively. Fig.4.10 shows peak stress as a function of test temperature. All these curves show the same trend as Fe-0.05%C alloys. The flattened region in peak stress value are at 800-900 and 700-800

for solution treatment condition

for direct casting condition.

60

60

700¡ É

60 700¡ É

700¡ É

50

850¡ É

800¡ É

50

800¡ É

Stress (MPa)

900¡ É

40

50 750¡ É

850¡ É 750¡ É

40

800¡ É

40

850¡ É

750¡ É

30

950¡ É

1000¡ É

20

950¡ É

30

900¡ É

900¡ É

30

20

20 1000¡ É

10

0 0.0

0.1

0.2

0.3

0.4

0 0.6 0.0

0.5

950¡ É

1050¡ É

10

0.1

10

0.2

0.3

Strain

0.4

0.5

1000¡ É

0 0.6 0.0

0.1

0.2

0.3

Strain

(a) 1100

0.4

0.5

Strain

(b) 1200

(c) 1350

Fig.4.8 Stress-strain curves at different test temperatures for Fe-0.18%C alloys solution treated at temperature (a) 1100 60

(b) 1200 60

700¡ É

700¡ É

800¡ É

Stress (MPa)

50

800¡ É

50

40

(c) 1350

40

900¡ É

30

30

20

20

10

10

0

900¡ É

1000¡ É

0 0.0

0.1

0.2

0.3

0.4

Strain

0.0

0.1

0.2

0.3

0.4

Strain

(a) 100 min-1

(b) 200 min-1

Fig.4.9 Stress-strain curves at different test temperatures for Fe-0.18%C alloys under direct casting condition at cooling rate (a) 100 min-1 (b) 200 min-1

52

0.6

80

80

1100¡ É , solution 1350¡ É , solution

Peak stress (MPa)

Peak stress (MPa)

100 /min, melting

1200¡ ,Ésolution

60

40

Ae3

20

0 650

700

750

800

200 /min, melting

60

40

20

850

900

950

1000

1050

0 650

1100

700

750

800

Test temperature(¡ É )

850

900

950

1000

1050

1100

Test temperature( )

(a) solution treatment condition

(b) direct casting condition

Fig.4.10. Peak stress as a function of test temperature for Fe-0.18%C alloys under (a) solution treatment condition (b) direct casting condition

4.2.3 Fe-0.45%C alloys The stress-strain curves of tensile tests of Fe-0.45%C alloys are shown in Fig.4.11 and Fig.4.12 under solution treatment condition and direct casting condition respectively. Fig.4.13 shows peak stress as a function of test temperature. As the carbon content increases, peak stress values increase. No significant flattened regions in peak stress were observed, unlike the other alloys. 90

700¡ É

90

80

80

700¡ É

70

700¡ É

90 80

750¡ É

70

70

Stress (MPa)

750¡ É 800¡ É

60

800¡ É

60

50

60

850¡ É

800¡ É

50

40

50

40

900¡ É

850¡ É

40

900¡ É

900¡ É

30

30

30

20

20

20

10

10

10

0

0 0.0

0.1

0.2

0.3

0.4

0.5

0 0.6 0.0

0.1

0.2

0.3

Strain

0.4

0.5

0.0 0.6

0.1

Strain

(a) 1100

0.2

0.3

0.4

Strain

(b) 1200

(c) 1350

Fig.4.11 Stress-strain curves at different test temperatures for Fe-0.45%C alloys solution treated at temperature (a) 1100

(b) 1200

53

(c) 1350

0.5

0.6

100

100

700¡ É

700¡ É

90

90

80

80 750¡ É

Stress (MPa)

70

750¡ É

70

800¡ É

60

800¡ É

60

50 40

850¡ É

50

850¡ É

900¡ É

40

900¡ É

30

30 1000¡ É

20

20

10

10

0

0 0.0

0.1

0.2

0.3

0.4

0.5 0.0

0.1

0.2

0.3

0.4

0.5

Strain

Strain

(a) 100 min-1

(b) 200 min-1

Fig.4.12 Stress-strain curves at different test temperatures for Fe-0.45%C alloys under direct casting condition at cooling rate (a) 100 min-1 (b) 200 min-1 100

120

1100¡ É , solution 1200¡ É , solution

100 /min, melting

100

1350¡ É , solution

Peak stress (MPa)

Peak stress (MPa)

80 60 40 Ae3

200 /min, melting 80 60 40

20 20

0 650

700

750

800

850

900

950

1000

1050

0 650

700

750

) Test temperature(¡ É

800

850

900

950

1000

1050

Test temperature( )

(a) solution treatment condition

(b) direct casting condition

Fig.4.13 Peak stress as a function of test temperature for Fe-0.45%C alloys under (a) solution treatment condition (b) direct casting condition

54

4.3 Austenite grain size In order to determine the grain size, samples were subjected to the same heat treatments without applying deformation. Specimens were then water quenched following cooling into the two-phase region so that ferrite could form on the austenite grain boundaries. Specimens were alternatively etched in a picric acid based solution or nital. Fig.4.14, Fig.4.15 and Fig.4.16 show the cross-section of GLEEBLE heat-treated specimen for Fe-0.05%C, Fe-0.18%C and Fe-0.45%C alloys respectively. The diameter of each cross-sectional specimen is 10mm. In determining the mean grain size, generally known measuring method such as linear intercept technique could not be used because the grain sizes were too large. Individual grain areas were measured using software UTHSCS IMAGE TOOL and the grain size was calculated using the equivalent diameter to account for different morphologies. The distributions of austenite grain size are shown in Fig.4.17, Fig.4.18 and Fig.4.19. In these graphs each measured grain sizes is plotted as a function of the distance between the specimen center and the grain center. For specimens having large grains, in order to get more data two or three cross-sections of specimen were used in obtaining the graph. Due to the radial thermal gradient present in GLEEBLE specimens, there is grain size differences between the center and surface regions as shown in these graphs. This difference becomes larger as the solution treatment temperature is increased. However, in the case of direct casting, this difference is smaller than in the case where specimens were solution treated. For the direct casting condition, when the specimen is melted, the radial thermal gradient seems to decrease, resulting in a more homogeneous grain size.

55

Solution

1100

1200

treatment

1350

Direct casting

100 min-1

200 min-1

Fig.4.14 Cross-section of a GLEEBLE specimen of a Fe-0.05%C alloy. Etched in nital. For the sake of clarity the position of grain boundaries were photo-enhanced in the photographs ‘Direct casting’ (diameter of specimen = 10mm)

56

Solution

1100

1200

1280

1350

treatment

Direct casting

100 min-1

200 min-1

Fig.4.15 Cross-section of a GLEEBLE specimen of a Fe-0.18%C alloy. Etched in saturated picric acid based etchant. For the sake of clarity the position of grain boundaries were photo-enhanced in the photographs ‘200 min-1-Direct casting’ (diameter of specimen = 10mm)

57

Solution

1200

1100

treatment

1350

Direct casting

100 min-1

200 min-1

Fig.4.16 Cross-section of a GLEEBLE specimen of a Fe-0.45%C alloy. Etched in saturated picric acid based etchant (diameter of specimen = 10mm)

58

6

Grain size (mm)

1100 , solution 1200 , solution 1350 , solution 4

2

0 0

1

2

3

4

5

Distance from center (mm)

(a) solution treatment

6 100 /min, melting

Grain size (mm)

200 /min, melting 4

2

0 0

1

2

3

4

5

Distance from center (mm)

(b) direct casting Fig.4.17 The distribution of austenite grain size of a Fe-0.05%C alloy under (a) solution treatment condition (b) direct casting condition

59

6

Grain size (mm)

1100 1200 1280 1350

, , , ,

solution solution solution solution

4

2

0 0

1

2

3

4

5

Distance from center (mm)

(a) solution treatment

6 100 /min, melting

Grain size (mm)

200 /min, melting 4

2

0 0

1

2

3

4

5

Distance from center (mm)

(b) direct casting Fig.4.18 The distribution of austenite grain size of a Fe-0.18%C alloy under (a) solution treatment condition (b) direct casting condition

60

6

Grain size (mm)

1100 , solution 1200 , solution 1350 , solution 4

2

0 0

1

2

3

4

5

Distance from center (mm)

(a) solution treatment

6

100 /min, melting

Grain size (mm)

200 /min, melting 4

2

0 0

1

2

3

4

5

Distance from center (mm)

(b) direct casting Fig.4.19 The distribution of austenite grain size of a Fe-0.45%C alloy under (a) solution treatment condition (b) direct casting condition

61

CHAPTER 5 - DISCUSSION 5.1 Grain growth In order to study the influence of grain size on hot ductility, it is necessary to determine the representative grain size resulting from the thermal history of each specimens used in this study. The cross-sections shown in the previous chapter reveal a single plane, possibly masking the cross-section that actually contains the largest grain. For this reason, and also because of the grain size difference between center and surface, it would be incorrect to determine the average grain size by merely averaging the grain sizes in the plane of the polished surface. In an attempt to overcome this problem, only selected grains which occupied 40% of the cross-sectional area of the specimen were used in determining the ‘average’ grain size. Although this is an arbitrary choice, it gives a better measure of the effective or true grain size. Fig.5.1 shows grain size as a function of solution treatment temperature. In this graph, the maximum grain sizes as well as the mean size of the selected grains are included. The grain size increases almost linearly with increasing solution treatment temperature. It is surprising that in the case of a solution treatment at a temperature of 1350

the average grain size reached 4mm in

diameter in all steel grades. In the Fe-0.05%C alloy, a maximum grain size of almost 6mm was obtained. Other researches [7, 32] using Nb containing steel and plain C-Mn steel under similar thermal conditions obtained grain sizes of 0.4-1.5mm in diameter. This observation implies that alloying elements or alloying compounds pin the grain boundaries and prevent grain boundary migration at these high solution treatment temperatures. Carbon content seems to have little effect on grain size at any given solution treatment temperature. On the other hand, when specimens are melted in-situ, there is a significant influence of carbon on grain size as shown in Fig.5.2. In the in-situ melted specimens the largest grains were found in the Fe-0.18%C alloy. Inspection of the Fe-C phase diagram reveals that in the steel of peritectic composition, Fe-0.18%C, 62

austenite forms on cooling at a much higher temperature than either the Fe-0.05%C alloy or the Fe0.45%C alloy (almost 100

higher). Since grain growth is a sharp function of temperature, austenite

grains will grow much larger in any given time in a steel of peritectic composition. Therefore the grain size in the steel of peritectic composition is expected to be largest at any given cooling rate. That this argument holds, is bourne out by the grain sizes obtained at different cooling rates. At a cooling rate of 100 min-1, the specimen spends more time in the high temperature region than at a cooling rate of 200 min-1 and hence the grain size is larger. The grain size of the Fe-0.05%C alloy is also relatively large because delta-ferrite grains grow very large in the single delta-ferrite phase on cooling following solidification. The austenite grain size is determined, in large measure by the delta-ferrite size and hence, large austenite grains form [50]. 6

Average(selected)

M axim um

Grain size (mm)

0.05%C 0.18%C 0.45%C

0.05%C 0.18%C 0.45%C

4

2

0 1050

1100

1150

1200

1250

1300

1350

1400

) Solution treatment temperature ( ¡ É

Fig.5.1 Grain size as a function of solution treatment temperature

63

5

Grain size (mm)

4

3

2 Average (selected)

1

Maximum

0.05%C 0.18%C 0.45%C

0.05%C 0.18%C 0.45%C

0 100¡ É /min

200¡ É /min

Fig.5.2 Grain size as a function of cooling rate under direct casting condition

5.2 Ductility troughs Fig.5.3 and Fig.5.4 shows the hot-ductility curves for solution treated specimens having various grain sizes for Fe-0.05%C and Fe-0.18%C alloys respectively. The Ae1 temperature (the eutectoid transformation temperature under equilibrium condition) and Ae3 temperature (the austenite/ferrite transformation start temperature under equilibrium condition) which are calculated using software MTDATA are also shown on the graph. As shown in these figures, as the grain sizes increase, the ductility trough extends towards higher temperatures and even beyond the Ae3 temperature. The reason for the existence of a low ductility region just below the Ae3 temperature may be explained as follows. According to Crowther [16] and Cardoso [17], deformation induced ferrite can be formed at temperatures above the Ar3 temperature (the austenite/ferrite transformation start temperature at a constant cooling rate), and often as high as the Ae3 temperature when the tensile test is conducted at these temperatures. The comparative ease of dynamic recovery in ferrite translates into a low flow stress compared to

64

100%

RA (%)

80%

60%

40%

Ae1 0.7mm

20%

1.6mm 4.3mm

Ae3 0% 650

700

750

800

850

900

950

1000

1050

1100

) Test temperature (¡ É

Fig.5.3 Hot-ductility curves for Fe-0.05%C alloys for solution treated specimens having various grain sizes

100%

RA (%)

80%

Ae1 Ae3

60% 40%

0.4mm 20%

1.4mm 3.8mm

0% 650

700

750

800

850

900

950

1000 1050

1100

) Test temperature (¡ É

Fig.5.4 Hot-ductility curves for Fe-0.18%C alloys for solution treated specimens having various grain sizes

austenite, and therefore to strain concentration in the ferrite film. This strain concentration leads to ductile voiding, generally at void nucleation sites located at austenite grain boundaries and this results in a loss of ductility. Coarsening the grain size causes the temperature of the start of the ductility trough to increase up to the Ae3 temperature. However, there is no evidence in the literature that deformation induced ferrite can form beyond the Ae3 temperature for the steels having a carbon content of less than 0.3%. The extension of ductility trough beyond the Ae3 temperature as shown in Fig.5.3 and Fig.5.4, could possibly be ascribed to the occurrence of grain 65

boundary sliding, although evidence in the literature suggests this mechanism of failure is only favored in higher carbon steel containing more than 0.3%C [16]. In coarse grained steels, it is difficult for dynamic recrystallisation to occur due to the decrease in the number of grain boundary nucleation sites where dynamic recrystallisation initiates, thereby decreasing the possibility of ductility improvement via grain boundary migration [1, 45]. On the other hand, AlN precipitates can play a role in widening the ductility trough into the single phase austenite region by encouraging grain boundary sliding. It is likely that AlN precipitates are formed on the austenite grain boundaries, pinning the boundaries and allowing the cracks formed by grain boundary sliding to join up, as well as encouraging void formation. It is recognized that the precipitation of AlN on grain boundaries can have a significant influence on hot-ductility. However, in this study, it was not possible to study this aspect in detail. It is therefore recommended that the effect of AlN precipitation in these alloys be distinguished from the effect of grain size on hot-ductility by fractography and transmission electron microscopy. It is interesting to compare the position of the ductility trough relative to temperature to the temperature dependence of the peak stress obtained in high-temperature tensile tests. The peak stress vs. temperature curves for the Fe-0.05%C and Fe-0.18%C alloys are shown in Fig.4.7 and 4.10 respectively. On lowering the temperature of the tensile test, the strength of austenite increases down to a temperature of about 950

in the case of the Fe-0.05%C alloy and 900

the case of the Fe-0.18%C alloy. At test temperatures between 950

in

and 750 , the peak stress is

lower than what it would have been for pure austenite and higher than for pure ferrite, evidently because the test temperature falls within the two-phase (ferrite + austenite) region and the mixture of ferrite and austenite is softer than austenite. However, this softening occurs at a temperature at least 50

above the Ae3 temperature. The formation of ferrite between the Ar3 and Ae3

temperature during deformation is a well known phenomenon and such ferrite is usually referred to ‘deformation induced ferrite’.

66

It is more difficult to explain the softening behavior above the Ae3 temperature. One possibility is that the temperature, measured in the surface of the tensile specimen, is not fully representative of the bulk temperature of the specimen in the region where deformation occurs. Hence that the temperatures boundaring the two-phase region are displaced. There is clearly a radial temperature distribution in a hot-tensile test specimen in the GLEEBLE arrangement. However, experimental evidence suggests that the temperature measured on the surface of the specimen in the area of neck formation is actually lower than the temperature in the center. This temperature difference also becomes higher at higher test temperature. It therefore seems improbable that the extended ‘softening temperature region’ is merely due to a temperature displacement of the two-phase region. In the Fe-0.45%C alloy, the two phase region is much narrower and little evidence is found of the influence of the two-phase region on the peak stress except in the specimen containing very large grains as shown in Fig.4.13. Fig.5.3 shows a very strong dependence on grain size of ductility as well as the temperature range of the ductility trough for the Fe-0.05%C alloy. The ductility trough becomes deeper and wider with an increase in grain size and the reduction in ductility in the larger-grained specimens at temperatures above the Ae3 temperature can clearly not be attributed to the formation of ferrite. In specimens with a grain size of 4.3mm, a significant reduction in ductility occurs on lowering the test temperature from 1100

to 1000 . For this temperature range the structure is clearly

austenitic and the reduction in ductility must be attributed mainly to a grain size effect. The dependence of ductility on temperature of the Fe-0.18%C and Fe-0.45%C alloys show similar behavior but the ductility troughs become wider with an increase in carbon content. These aspects are analysed more quantitatively below. Fig.5.5 shows hot ductility curves for the Fe-0.45%C alloy with various grain sizes. The low ductility above Ae3 and between Ae3 and Ar3 seems to be due to grain boundary sliding and deformation induced ferrite respectively as in the case of the other alloys. But at 700 , there is a

67

slight increase in ductility. The drop in ductility below a temperature of 700

probably arises from

the fact that a relatively large amount of pearlite forms in this alloy compared to the other alloys at temperatures below Ar1 (the eutectoid transformation temperature at a constant cooling rate). Until the Ar1 temperature is reached, ductility will increase due to an increase in the volume fraction of ferrite. Once the Ar1 temperature is reached, pearlite can form and the strength of the matrix is increased. In addition, the presence of a thin film of ferrite at the grain boundaries, leads to even more strain concentration at the grain boundaries and the ductility decreases even further [40]. 100% 0.7mm 2.6mm 4.1mm

RA (%)

80% Ae3

Ae1

60% 40% 20% 0% 600

650

700

750

800

850

900

950

1000 1050

Test temperature (¡ É )

Fig.5.5 Hot-ductility curves for Fe-0.45%C alloys for solution treated specimens having various grain sizes (The Ar1 and Ar3 temperature are not shown in the figure)

5.3 Influence of grain size on hot ductility 5.3.1 Reduction in area The relationship between tensile properties and the reciprocal of austenite grain size (D) at several test temperatures is shown in Fig.5.6 and Fig 5.7 for the solution treatment conditions and direct casting conditions respectively. The ductility generally decreases with increasing grain size under all testing conditions notwithstanding the fact that the matrix strength was essentially independent of the grain size at any given test temperature. These results are in good agreement with previous studies in which Nb-containing steels [7] and plain C-Mn steels [32] were used. However in these studies, the steels contained relatively small grains 68

(smaller than 1.5mm in diameter). In the case of direct casting, the change in grain size results from the difference of cooling rate. At the higher cooling rate there is less time available for grain growth in the high austenite temperature region. Hence, the grain size decreases with an increase in cooling rate and a concomitant ductility improvement results as shown in Fig.5.7. 60 800¡ É

50

Peak Stress (MPa)

900¡ É

40 900¡ É

900¡ É

30

1000¡ É

1000¡ É

20 10 0 100

1000¡ É

1000¡ É

80

RA (%)

900¡ É

60 900¡ É

900¡ É

40 800¡ É

20

0 0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

-1

1/D (mm )

(a) Fe-0.05%C

0.0

0.5

1.0

1.5

2.0 -1

1/D (mm )

(b) Fe-0.18%C

2.5

3.0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

-1

1/D (mm )

(c) Fe-0.45%C

Fig.5.6 Relationship between tensile properties and reciprocal of austenite grain size (D) at different test temperatures in solution treated specimen (a) Fe-0.05%C (b) Fe-0.18%C (c) Fe-0.45%c alloy

69

Peak Stress (MPa)

70

0

60

0

50

0

800¡ É 850¡ É

0

40 900¡ É

30

0

1000¡ É

20

0

0 100

0 0 1000¡ É

1000¡ É

0

10

80

900¡ É

0

RA (%)

900¡ É

60

0

40

0

20

0

1000¡ É 900¡ É

0

0.30

0.35

0.40

850¡ É 800¡ É

0 0.40.25

0.26

0.27

0.28

0.29

0.30

0.35

-1

-1

(a) Fe-0.05%C

0.40

0.45

0.50

0.55

-1

1/D (mm )

1/D (mm )

1/D (mm )

(b) Fe-0.18%C

(c) Fe-0.45%C

Fig.5.7 Relationship between tensile properties and reciprocal of austenite grain size (D) at different test temperatures under direct casting condition for (a) Fe-0.05%C (b) Fe-0.18%C (c) Fe-0.45%C alloy

Fig.5.8 shows the effect of austenite grain size on the minimum RA value for the solution treatment condition. It is clear that with increasing grain size minimum RA value decrease for all steel grades. Above all, this is obvious in case of the Fe-0.05%C alloys, showing a much higher RA value at the same grain size. It is generally known that the lower carbon contents exhibit the more ductile behavior. 80 70

0.05%C 0.18%C 0.45%C

Minimum RA (%)

60 50 40 30 20 10 0

0.0

0.5

1.0

1.5

2.0

2.5

-1

1/D (mm )

Fig.5.8 Relationship between minimum RA value and reciprocal of austenite grain size (D) for different Fe70

C alloys in the solution treatment condition

5.3.2 Position and width of ductility trough As a measure of the position of a ductility trough, the temperature at minimum RA value was taken except in the case of very wide ductility troughs where the temperature was chosen to be at the center of the trough. This information is summarized in Fig.5.9 as a function of grain size. Also Included is the width of trough taken at 40% of RA value. With increasing grain size, the

P o sitio n o f d u c tility tro u g h (¡ É )

position of the trough as well as the width of the trough are increased.

950

(a)

0.05%C 0.18%C 0.45%C

(b)

900

850

800 300 250

W id th o f tro u g h (¡ É )

0.05%C 0.18%C 0.45%C

200 150 100 50 0

0

1

2

3

4

Grain size (mm)

Fig.5.9 (a) Position of ductility trough, (b) width of trough as a function of grain size under solution treatment condition

71

5.4 Influence of carbon content on hot ductility Fig.5.10 shows hot ductility curves for the alloys of different carbon contents at the same solution treatment temperature. It is clear that at any given solution treatment temperature the position and shape of the ductility curve vary with carbon content. As shown in Fig.5.8 and Fig.5.9 (a), the carbon content affected the position of trough as well as the depth of trough under solution treatment condition. 100%

RA (%)

80%

60%

40%

`

0.05%C 0.18%C 0.45%C

20%

0% 600

650

700

750

800

850

900

950

1000 1050 1100

Test temperature ( )

(a) 1100 100%

RA (%)

80% 60% 40%

`

0.05%C 0.18%C 0.45%C

20% 0% 600

650

700

750

800

850

900

Test temperature ( )

(b) 1200

72

950

1000 1050 1100

100% 0.05%C 0.18%C 0.45%C

RA (%)

80% 60% 40%

`

20% 0% 600

650

700

750

800

850

900

950

1000 1050 1100

Test temperature ( )

(c) 1350 Fig.5.10 Hot ductility curves from solution treatments at (a) 1100

(b) 1200

(c) 1350

In order to investigate the influence of carbon content on hot ductility under direct casting conditions, the austenite grains size and the variation of the corresponding RA value were plotted against % C as shown in Fig.5.11. The Fe-0.18%C alloy exhibits the largest grain size at both cooling rates and the large grains in this Fe-C alloy results from the higher austenitizing temperature compared to that of other Fe-C alloys as discussed earlier. The loss of ductility seems to be directly related to an increase in grain size and the RA value for Fe-0.18%C alloy has the lowest value. Although the Fe-0.45%C alloy had the smallest grain size, it was still very brittle. This observation may be explained

73

4

3

3

Grain size (mm)

Grain size (mm)

4

2

1

800¡ É 900¡ É 1000¡ É

80

800¡ É 900¡ É 1000¡ É

80 60

RA (%)

60

RA (%)

1

0 100

0 100

40

40 20

20 0

2

0.0

0.1

0.2

0.3

0.4

0

0.5

0.0

0.1

0.2

0.3

0.4

0.5

C (%)

C (%)

(a) 100 min-1

(b) 200 min-1

Fig.5.11 Effect of C content on austenite grain size and RA value under direct casting condition for cooling rate (a) 100 min-1 (b) 200 min-1

by grain boundary sliding becoming dominant as the carbon content increases. Crowther and Mintz [16] showed that increasing the carbon content to above 0.3% in a coarse grained steel (~300 ) causes intergranular failure to occur by grain boundary sliding in the austenite, resulting in a very wide ductility trough. Increasing the carbon content was found to increase the activation energy for dynamic recrystallization, and hence to encourage more grain boundary sliding and linkage of cracks. Inspection of the cross-section of heat treated samples in the direct-casting condition shown in Fig.4.14, Fig.4.15 and Fig.4.16 reveals interesting information. Columnar austenite grains were formed in the Fe-0.18%C alloy compared to the equi-axed grains observed in the lower or higher C alloys. The formation of columnar grains in the Fe-C alloy close to the peritectic composition may be related to the larger temperature gradient during cooling of specimens of this composition 74

or to the occurrence of the peritectic reaction followed by the peritectic transformation to austenite [7]. This observation has great practical significance because the susceptibility of slabs to surface cracking can be significantly accelerated since the effective grain size affecting intergranular fracture can be taken as the length of the columnar grains rather than the average grain size [7].

75

5.5 Influence of cooling rate following direct-casting on hot ductility As discussed in the literature review an increase in the cooling rate generally results in reduced ductility in most types of steel. In most cases, the decrease in ductility with increasing cooling rate is ascribed to either the formation of finer precipitates or finer interdendritic inclusions [40]. The decrease in ductility is the result of the overriding effect of precipitation of alloy carbo-nitrides on the austenite grain boundaries, thereby masking any influence of grain size on ductility. Because plain carbon steels in which tramp elements and impurities were kept very low were used in this study the effect of alloying element precipitation was minimized. When a specimen is cooled from the molten state, the lower cooling rate allows time for grains to grow in the high austenite temperature region. Consequently, an increased cooling rate may result in ductility improvements through grain refinement. For this reason, in terms of grain size refinement, a higher cooling rate may be an advantage in near-net shape casting such as thin-slab casting or strip-casting processes. As shown in Fig.5.2, with an increase in cooling rate from 100 min-1 to 200 min-1, the grain size decreased and accordingly, ductility is improved in all steel grades and at most test temperatures as shown in Fig.4.1 (b), Fig.4.3 (b) and Fig.4.4 (b).

76

5.6 Comparison between as-cast condition and solution treatment condition It is interesting to note that the ductility of the as cast specimens was sometimes better than that of solution treated specimens as shown in Fig.5.12. It is well known that an as-cast structure is inferior to reheated structures with respect to high temperature mechanical properties due to the differences in microstructure. Generally, the austenite grain structure of reheated slabs has a smaller grain size than that of the as-cast structure. However, it follows from the data displayed in Fig.5.12 that the coarse grained structure of the solution treated steel exhibits much poorer hotductility than the as-cast structure in the temperature range 820

to 1000 .

80% 1350 , solution 200 /min, melting

RA (%)

60%

40%

20%

0% 650

700

750

800

850

900

950

1000

1050

Test temperature ( )

Fig.5.12 Hot ductility curves for Fe-0.18%C alloy from two different thermal conditions

77

5.7 Practical implications of the experimental findings This study was undertaken to probe the influence of grain size on hot-ductility primarily because evidence in the literature implicated reduced hot-ductility in continuously cast slabs to the presence of large grains occurring at the roots of oscillation marks [4, 6, 7]. Convincing experimental result was found that an increased austenite grain size results in a deepening as well as a widening of the ductility trough observed in the temperature range 700

to 1100 . The

detrimental effect of grain size was most severe in alloys of near peritectic composition and moreover, in specimens directly cast in a GLEEBLE arrangement, austenite grains were of columnar nature compared to equi-axed structures observed in the other alloys. A further finding of great significance was that the presence of large austenite grains seemed to be more detrimental to hot-ductility in the austenitic region in the solution treated specimens than an as-cast structure, at least under the pertaining experimental conditions. It is pertinent to relate these findings to continuous casting practice. Very large austenite grains have been observed at the roots of oscillation marks [4] in commercially produced continuously cast slab, especially in steels of near-peritectic composition. These large austenite grains form when the peritectic transformation occurs and the thin solidifying shell shrinks from the mould locally. The concomitant lowering of the heat extraction rate results in this thin part of the shell being less efficiently cooled and hence, more time is allowed for austenite grain growth to occur. The observation that columnar grains form in the GLEEBLE samples that have been melted insitu, compounds the problem because in the presence of such columnar grains, it is not the average grain size that will determine the surface crack susceptibility, but the size of the extended columnar grains. Although large austenite grains are formed in the mould, the detrimental effect thereof is mainly realized during unbending where the temperature corresponds to that of the ductility trough. During unbending the surface of the slab is subjected to tensile stresses and if large grains 78

exist at the roots of oscillation marks, the surface crack susceptibility is substantially increased. In order to prevent the formation of large austenite grains, remedial action need to be taken in the mould. Attempts need to be made to ensure that the volume contraction resulting from the peritectic transition does not cause large differentials in the cooling rate of the thin solidifying shell. The judicious selection of mould flux that will equalize, at least in part, the heat extraction along the surface of the strand is of the essence and some success has already been achieved in this regard. A better understanding of the mechanism of the peritectic reaction and quantitative information on the rate of the peritectic transformations are also required if a strategy is to be designed to prevent the formation of these excessively large, or ‘blown-out’ grains. Fortunately great strides are currently being made in pursuit of this goal [52].

79

CHAPTER 6 - CONCLUSIONS

1. The grain size of the three Fe-C alloys studied, increases almost linearly with increasing solution treatment temperature. A solution treatment temperature of 1350

rendered an average

austenite grain size of ~4mm in diameter in all three Fe-C alloys. This coarsening of grain size is attributed to the absence of alloying elements or impurity elements which can restrict grain growth at this high solution treatment temperature.

2. In the case where the specimens were melted in-situ (direct casting condition), the largest grains were found in the Fe-0.18%C alloy at both cooling rates used. This observation can be explained by the higher temperature at which the first austenite forms on cooling in the Fe-0.18%C alloy than in the others, so that the grains have a better chance to grow at high temperature.

3. Columnar austenite grains were formed in the Fe-0.18%C alloy on cooling from the molten state, whereas equi-axed grains were formed in the lower or higher C alloys.

4. Increasing grain size resulted in a loss of ductility, widening and deepening the ductility trough under all test conditions. The existence of a ductility trough after the Ae3 temperature is reached on cooling, may be due to the formation of deformation induced ferrite. The extension of the ductility trough to temperatures higher than the Ae3 may possibly be ascribed to the occurrence of grain boundary sliding. There is a possibility that the precipitation of AlN on austenite grain boundaries can contribute to the loss of ductility but this aspect was not studied in detail.

5. The peak stress as a function of test temperature in hot tensile tests showed a flattening of the peak stress in the ferrite-austenite duplex region. In the Fe-0.05%C and Fe-0.18%C alloys, the

80

Ae3 temperature falls within this flattened region, indicating that deformation induce ferrite can form at least up to the Ae3 temperature. The extension of this flattened region beyond the Ae3 temperature can not be explained as yet.

6. Because there is an almost linear relationship between peak stress and temperature for the Fe0.45%C alloy, it seems that the main embrittling mechanism in this steel is grain boundary sliding.

7. In the case of the Fe-0.45%C alloy, the loss of ductility below 700

is most probably due to the

formation of a relatively large amount of pearlite at temperatures below Ar1 compared to the other alloys. Once the Ar1 temperature is reached, pearlite can form and the strength of the matrix is increased. In addition, the presence of a thin film of ferrite at the grain boundaries, leads to even more strain concentration at the grain boundaries and the ductility decreases even further.

8. With increasing grain size the position of the ductility trough as well as the width of trough is increased.

9. In the case where specimens were cast in-situ (direct casting condition), the higher cooling rate allows less time for grains to grow in the high austenite temperature region and hence, ductility improves.

10. For the specimens solution treated to render different grain sizes, the position of the ductility trough as well as the depth of the trough decreases as the carbon content increases. This observation correlates with the tendency of decreasing the temperature range of the flattened

81

region in the peak stress curve as a function of test temperature with increasing carbon content. This reflects the characteristics of two phase region, i.e., the higher carbon content represent the lower Ar3 temperature and narrowing two phase region.

11. Under direct casting conditions, the Fe-0.18%C alloy showed the largest grain size at both cooling rates and consequently the RA value for this alloy represents the lowest value amongst the three alloys.

82

BIBLIOGRAPHY

1. B. Mintz, S. Yue, J.J. Jonas, 'Hot ductility of steels and its relationship to the problem of transverse cracking during continuous casting,' International Materials Review, 1991. 36(5), 187~217 2. O. J. Kwon, 'Thin slab Hot rolling process in Asia : Installation, Core Technology and Competitiveness,' 35th IISI, 2001 3. D. Ameling, 'Thin slab casting-hot rolling in the EU,' 35th IISI, 2001 4. E. S Szekeres, 'A review of strand casting factors affecting transverse cracking,' Pres'n, 6th International Conference on Clean Steel, Hungary, 2002 5. E. T. Turdogan, 'Fundamentals of Steelmaking,' Institute of Materials, London, 1996, 312 6. L. Schmidt, A. Josefsson, 'On the formation and avoidance of transverse cracks in continuously cast slabs from curved mould machines,' Scandinavian Journal of Metallurgy 3, 1974, 193~199 7. Y. Maehara, K. Yasumoto, Y. Sugitani, 'Effect of carbon on hot ductility of as-cast low alloy steels,' ISIJ Transactions, 1985. 25, 1045~1092 8. M. Wolf, 'Intergranular surface cracks in bloom casting,' Trans. ISIJ, Vol. 24, 1984 9. W. T. Lankford, Metall. Trans., 1972, 3, 1331 10. N. W. Blake, 'Hot ductility testing of simulation castings,' Rep. MRL/AMP/87/5. BHP, Melbourne, July 1987 11. J. Y. Fu, C. I. Garcia, S. Pytel, in 'Processing, microstructure and properties of HSLA steels,' Warrendal, PA, Metallurgical Society of AIME, 1988, 27 12. L. E. Cepeda, J. M. Rodriguez-Ibabe, J. J. Urcola, Mater. Sci. Technol., 1989, 5, 1191 13. B. Mintz, J.M. Arrowsmith, 'Hot-ductility behaviour of C-Mn-Nb-Al steels and its relationship to crack propagation during the straightening of continously cast strand,' Metals Technology,

83

1979. Vo. 6, 24~32 14. J. R. Wilcox, R. W. K. Honeycombe, Met. Technol., 1984, 11, 217 15. J. R. Wilcox, R. W. K. Honeycombe, Mater. Sci. Technol., 1987, 3, 849 16. D. N. Crowther, B. Mintz, Mater. Sci. Technol., 1986, 2, 671 17. G. I. S. L. Cardoso, S. Yue, Proc. AIME Mech. Work. Steel Process. Conf., 1989, 31, 585 18. E. Essadiqi, J. J. Jonas, Metall. Trans., 1988, 19A, 417 19. B. Mintz, J.R. Wilcox, D.N. Crowther, 'Hot ductility of directly cast C-Mn-Nb-Al steel,' Materials Science and Technology, 1986. Vol. 2, 589~594 20. Y. Maehara, Y. Ohmori, Mater. Sci. Eng., 1984, 62, 109 21. C. Ouchi, K. Matsumoto, Trans. Iron Steel Inst. Jpn, 1982, 22, 181 22. H. E. Evans, 'Creep fracture,' Oxford, Pergamon Press, 1984 23. H. C. Chang, N. J. Grant, Trans. AIME, 1956, 206, 544 24. P. J. Wray, Met. Technol., 1981, 8, 466 25. D. M. Keene, C. M. Sellars, W. J. McG. Tegart, in 'Deformation under hot working conditions,' London, The Iron and Steel Institute, 1968, 21 26. G. Bernard, J. P. Birat, B. Conseil, Rev. Metall., 1978, 75, 467 27. D. N. Crowther, Z. Mohamed, B. Mintz, Trans. Iron Steel Inst. Jpn, 1987, 27, 366 28. R. E. Reed-Hill, R. Abbaschian, 'Thermally activated plastic deformation,' Physical Metallurgy Priciples, 3rd edition, PWS-KENT Publishing, 1994 29. B. Mintz, R. Abushosha, M. Shaker, Materials Science and Technology, 1993, 9(10), 907-14 30. V. Kutumba Rao, D. M. R. Taplin, P. Ramo Rao, Metall. Trans., 1975, 6A, 77-86 31. C. M. Sellars, in 'Hot working and forming processes,' London, The Metals Society, 1980, 3 32. D.N. Crowther, B. Mintz, 'Influence of grain size on hot ductility of plain C-Mn steels,' Materials Science and Technology, 1986. 2(9), 951~955 33. I. Weiss, J. J. Jonas, Metall. Trans., 1979, 10A, 831

84

34. W. C. Leslie, R. L. Rickett, C. L. Dotson, Trans. ASM., 1954, 46, 1470 35. T. Nozaki, J. Matsuno, K. Murata, Trans. Iron Steel Inst. Jpn, 1978, 18, 330 36. B. Mintz, J.M Stewart, D.N. Crowther, 'The influence of cyclic temperature oscillations on precipitation and hot ductility of a C-Mn-Nb-Al steel,' Transactions ISIJ, 1987. Vol. 27, 959~964 37. R. Abushosha, R. Vipond, B. Mintz, 'Influence of titanium on hot ductility of as cast steels,' Materials Science and Technology, 1991. Vol. 7, 613~621 38. Y. Maehara, et al., 'Surface cracking mechanism of continuously cast low carbon low alloy steel slabs,' Materials Science and Technology, 1990. 6(9), 793~806 39. B. Mintz, R. Abushosha, 'Effectiveness of hot tensile tests in simulating straightening in continuous casting,' Materials Science and Technology, 1992. 8(2), 171~177 40. R. Abushosha, S. Ayyad, B. Mintz, 'Influence of cooling rate and MnS inclusions on hot ductility of steels,' Materials Science and Technology, 1998. 14(3), 227~235 41. E. T. Turdogan, S. Ignatowicz, J. Pearson, J. Iron Steel Inst., 1955, 180, 349 42. R. Abushosha, R. Vipond, B. Mintz, 'Influence of sulphur and niobium on hot ductility of as cast steels,' Materials Science and Technology, 1991, 7(12), 1101-1107 43. R. Abushosha, S. Ayyad, B. Mintz, 'Influence of cooling rate on hot ductility of C-Mn-Al and C-Mn-Nb-Al steels,' Materials Science and Technology, 1998, 14(4), 346-351 44. T. Maki, T. Nagamichi, N. Abe, Tetsu-to-Hagane (J. Iron Steel Inst. Jpn), 1985, 71, 1367 45. B. Mintz, 'The influence of composition on the hot ductility of steels and to the problem of transverse cracking,' ISIJ International, 1999. 39(9), 833-855 46. T. H. Coleman, J. R. Wilcox, Materials Science and Technology, 1985, 1, 80 47. POSCO products catalogue - Hot rolled steel coil by thin-slab casting and rolling process 48. Y. Bayazitoglu, M. N. Ozisik, 'Elements of heat transfer,' McGraw Hill, New York, 1988 49. H. Suzuki, S. Nishimura, S. Yamaguchi, 'Characteristics of hot ductility in steels subjected to

85

melting and solidification,' Transactions ISIJ, 1982, 22, 48-56 50. R. J. Dippenaar, D. J. Phelan, 'Delta-ferrite recovery structures in low carbon steels,' Proceedings Vol.1, Mills Symposium, Metals, Slags and Glasses: High temperature Properties and Phenomena, London, UK, 2002, 283-292 51. W. R. Irving, 'Continuous casting of steel,' London, The Institute of Materials, 1993 52. D. J. Phelan, M. Reid, R. J. Dippenaar, Private communication

86