Abstract
Hot ductility of three grades of C–Mn steel and four grades of microalloyed steel have been evaluated simulating the bending or straightening operation in thin slab continuous casting process. The minimum bending or straightening temperature has been determined from the hot ductility curves generated by plotting per cent reduction in cross-sectional area of specimens against different test temperatures. The results have been discussed by analysing the force elongation curves and the fractographs of all the steels, and possible explanations for the results obtained have been presented. In microalloyed steels containing Nb, multiple troughs are observed in the hot ductility curves instead of single trough observed for remaining steel grades. The ductility is not recovered at 700°C in steels containing Nb, unlike the remaining steel grades. Among the C–Mn steels, steels with higher S content show less ductility. Intergranular fracture is observed in specimens in which ductility obtained is low.
Keywords
Introduction
Transverse cracks are one kind of defects that may appear at the surface or at corners of the continuously cast slabs. These cracks in the slabs turn into rolling defects in coils causing yield loss. In a conventional thick slab casting machine, the slabs are inspected before rolling in order to detect these cracks. The cracks found in the slabs are then removed by scarfing. However, the design of thin slab casting and rolling plant is such that there is no scope to repair the slabs. Hence, it is more important to minimise such transverse cracks in thin slabs so that the yield loss and interruption of operation can be decreased.
The propensity for transverse cracks formation is usually related to the high temperature ductility of the material. These cracks are formed when tensile strains exceed fracture strain of the steel at the corresponding temperature.1,2 In other words, when the applied strain exceeds the ductility of the material, these cracks are formed. The tensile strains get developed on the slab surface or corners due to processes like oscillation of mould, misalignment in the casting machines, bending or straightening of slabs and due to thermal stresses generated because of the thermal gradients along the thickness of the slab. Hence, in order to sustain the tensile strains developed, the material should have sufficient ductility especially at the positions of straightener or bender. However, in steel, ductility loss is observed in certain range of temperature, which is specific to the kind of materials. This temperature range is required to be avoided at the stage of straightening and bending.
The high temperature ductility is evaluated by hot ductility tests. These tests are basically high temperature tensile tests conducted in a laboratory testing machine in which a thermal condition similar to the one in casting machine is maintained. In literature, there are mainly two methods employed to simulate thermal condition observed in the caster machine: reheating the specimen to a solutionising temperature followed by cooling at a cooling rate similar to the one in the caster, and one portion of the specimen is melted and cast in situ followed by cooling at a cooling rate observed in the casting machine. Then, the tensile testing is carried out when a test temperature is achieved. Both of these methods have their advantages and disadvantages. While in the first method, the effects of segregation cannot be studied, in the second method, as the casting is carried out in a thermomechanical simulator, there is a possibility of hole formation at the centre of the specimen. In order to get rid of this hole, a prior compressive strain is applied, so that the hole gets eliminated. Then, the final deformation is given. In this process, there is always a possibility of applying inadequate compression strain or overstraining the samples before final testing. In addition, exact segregation patterns or microstructure can never be replicated in this in situ casting method of testing because of the difference in volume of the material being cast. In the present work, it was decided to carry out the tests using the first method due to ease of testing by a Gleeble-1500D machine.
The ductility is measured by calculating per cent reduction in area (RA) of cross-section of the specimen at the position of fracture after the tensile test. If the cross-sectional area at fracture is small, RA will be more. This means ductility is better and the material can sustain more amount of deformation. However, if the decrease in cross sectional area is small, then RA becomes less and hence the ductility is low. So, the material can sustain less deformation before the formation of cracks. The ductility of steel generally varies with temperature, and there is a temperature range (relevant to casting temperatures) in which the ductility becomes very low. Thus, if the steel is deformed at a temperature in this range, there is a possibility of crack formation. In a thin slab caster, if the temperature of the slab, as it reaches the bender or straightener, falls in the range of low ductility, transverse cracks are propagated due to bending or straightening operation owing to the low ductility of the material. As per Mintz et al, 3 if the RA is >40, then transverse cracks will not be formed in the slab. However, as per Suzuki et al., 4 the criteria are more conservative, and as per this group of researchers, the RA should be >60 to avoid forming transverse cracks in the slab. In the present work, the second criterion will be used to report the minimum straightening or bending temperature.
Compared to the convention thick slab casting, the cooling rate experienced by the slab in thin slab caster is higher. This increase in cooling rate decreases the hot ductility of most of the steels. In C–Mn or in C–Mn–Al steels, with increase in cooling rate, finer MnS inclusions, finer AlN form, which leads to decrease in ductility. 5 In Nb based microalloyed steel, increase in cooling rate suppresses the formation of bigger size Nb based precipitate at higher temperature. In other words, more amount of Nb is held in solution, which at lower temperature on straightening or bending operation leads to formation of more amount of finer strain induced Nb(C,N) precipitates. These finer precipitates drastically reduce the hot ductility. 6
In literature, much of hot ductility data have been presented, but most of them are for conventional casting route. In the present work, the objective is to systematically evaluate hot ductility of both C–Mn steels and microalloyed steels (Nb and V based) at temperatures in the range of 700–1100°C simulating the thermal profile of the caster in a thin slab casting and rolling plant and then identify the range of temperature at which ductility of the materials become unacceptably low. This is followed by characterising fracture surfaces of specimens tested to compare the crack propagation mechanisms at different temperatures for the materials studied. In the present work, the load elongation curves are also generated, which can be useful to calculate the bending/straightening force required to bend or straighten the slabs in the casting machine.
Experimental
Steel slab samples of length 100 mm with width same as that of the slab were cut from the conventionally cast slab of 210 mm thickness. Three of the samples collected are C–Mn grades (A–C), and the rest are microalloyed grades. The compositions of these samples are shown in Table 1a and b .
Chemical composition (wt-) of C–Mn slab samples and corresponding Ae3 and Ar3 (°C)
Chemical composition (wt-) of microalloyed slab samples and corresponding Ae3 and Ar3 (°C)
The samples were then cut into smaller pieces, whose lengths were aligned perpendicular to the direction along which the columnar grains grow as shown in Fig. 1. Tensile specimens of cylindrical shape were machined from these pieces of steel slab samples. The dimensions of the specimens have been mentioned in Fig. 2. In order to measure the temperature during the hot tensile tests, Pt−Pt/Rh thermocouples were spot welded at the centre of the specimens as shown in Fig. 3. The tensile testing at different temperatures in the range of 700–1100°C at every 50°C interval was carried out for each of the grades in a Gleeble (1500D) machine. The schematic diagram of the testing set-up is shown in Fig. 3. The specimens were held in copper grips, and C shaped clamp was used to firmly hold the grips as well as to keep the specimen in position. The thermal cycle followed in these experiments is shown in Fig. 4. The specimens were heated to 1350°C at a heating rate of 10°C s−1 followed by heating to 1440°C at a relatively lower heating rate, i.e. 3°C s−1 and then held at this temperature for 30 s followed by cooling at a cooling rate of 200°C min−1 or 3·3°C s−1. This is the average cooling rate, close to surface of the strand over a temperature range that is observed in thin slab caster for a 50 mm thick slab. 3 The maximum temperature of 1440°C is chosen in order to ensure larger grain size and dissolution of all precipitates. The cooling was continued to test temperature at which the tensile tests of the specimens were carried out. The crosshead speed maintained was 5 mm min−1, which is equivalent to strain rate of 8·33×10−3 approximately assuming the deformation in the tensile test is concentrated over a length of 10 mm around the centre of the specimen. The strain rate experienced by surface of the slab in thin slab caster is calculated using the following mathematical relationships as mentioned by Lankford. 1
Schematic diagram of slab showing positions of slab sample and specimen orientation
Schematic diagram of specimen; all dimensions in millimetres
Schematic diagram of hot ductility experimental set-up in Gleeble 1500D machine
Thermal cycle employed during hot ductility experiment
ϵ = t/2R, where ϵ is surface strain due to bending of slab, t is thickness of the slab and R is bending radius of the slab.
, where ˙ϵ is the strain rate, V is the casting speed and L is the gauge length, which is assumed to be equal to thickness of the slab. Assuming a casting speed of 3·25 m min−1, the strain rate, close to surface of 50 mm thick slab produced in a thin slab caster having a bending radius of 3250 mm, is found to be 8·33×10−3 s−1, which is equal to the approximate strain rate applied in the tensile tests.
Tests were carried out after vacuum of ∼10−3 torr is achieved, and the chamber was then flushed with argon. In these experiments, the diameters of the specimens before and after the test (at the position of the fracture) were measured. The ductility is measured in terms of reduction in area (RA). In order to study the fractographs, samples were cut from the fractured specimens and then the fracture surfaces were observed in a scanning electron microscope.
Results
Hot ductility curves generated by plotting RA against different test temperatures are shown in Figs. 5 and 6 for both C–Mn grades and microalloyed grades respectively. Ductility trough is observed for each of these steel grades in these curves. The position and the width of the trough are varying for each of these grades. Among the C–Mn grades, grade A, with the highest amount of S, has the maximum width of the trough, and the minimum width is found for grade B, which has a higher Mn/S ratio. The minimum ductility is observed at 900, 800 and 825°C for A–C grades respectively. The minimum straightening temperature for each of these grades is defined as the temperature where the RA equals 60. These temperatures as obtained from each of the curves are 940, 880 and 920°C respectively. Only the higher side of the hot ductility curve is taken into consideration to determine these straightening temperatures. Three kinds of microalloyed steels have also been examined: Nb and V based, Nb based and V based steel. Two grades of Nb–V based steel have been taken for the analysis. The ductilities of these two grades are very similar to each other at low temperature range. There is a small difference in ductility seen in temperature range of 800–1000°C. Grade D has a little bit higher amount of Nb and more amount of V. The difference in ductility between grade D and E is not very large, indicating lesser influence of V on the difference between the hot ductility for these two grades. The minimum ductility is observed at 775°C for both of these grades. The minimum straightening temperatures as obtained from the hot ductility curves are 930 and 925°C for D and E grades respectively. Another point to be noted is that, in comparison to C–Mn grades, these two grades show much lower ductility at low temperature like 700°C. The ductility (RA) at temperature range of 650–750°C is <40, which is much lower in comparison to C–Mn steel grades. In addition, there seems to be two stages of ductility loss for the Nb–V based steel grades. This two-stage troughs are also observed in the hot ductility curve for grade F, which is Nb based grade. Another difference to be noted is that in the higher temperature side of the ductility trough, the ductility loss for C–Mn grades is very sharp. As shown in Fig. 5, for grade A, at temperatures >950°C, the amount of ductility is >80, but as the temperature is decreased <950°C, the ductility sharply decreases to <20. Similar trend is observed for grade B also, but this sudden drop in ductility occurs <900°C. For both Nb and Nb–V based grades, this ductility loss is gradual as evident from the slope of their ductility curves between temperatures 800 and 950°C. The ductility curve for V based steel is similar to the curves for C–Mn grades. There is no two-stage ductility loss observed in the curve.
Hot ductility curves for C–Mn steel grades
a hot ductility curves for Nb+V based microalloyed steel, b hot ductility curves for Nb based microalloyed steel and c hot ductility curves for V based microalloyed steel
The force versus elongation curves are shown in Figs. 7 and 8 for all of the steel grades being studied. For each of the steel grades, the strength of the steel decreases with rise in temperature. The elongations for certain temperatures are low. For grade A, it can be seen from these curves, the elongations are very limited at temperatures of 800–925°C. Beyond the peak load, sudden drop in the load displacement curves are seen at these temperatures, unlike the delay in load drop observed for temperatures of 700, 1000 and 1050°C. Minimum elongation is observed at temperature 850°C. Similar observations are found in other C–Mn grades and V based steel grade. For Nb–V based and Nb based grades, the load displacement curve at temperature ∼700°C shows sudden load drop after the peak load unlike the C–Mn steel grades. But at temperature beyond 1000°C, the shape of the load displacement curve is similar to the C–Mn grades.
a force elongation curves for steel A, b force elongation curves for steel B and c force elongation curves for steel C
a force elongation curves for steel D, b force elongation curve for steel E, c force elongation curve for steel F and d force elongation curve for steel G
Fractographs for both C–Mn and microalloyed steel grades are shown in Figs. 9–12. The features on the fracture surface shown include dimples or microvoids, flat faces, cavities and mixed features with both flat facets and microvoids. The appearance of these features on fracture surface depends on the temperatures at which the specimens have been tested. For steel grade A, the fractograph shows a lot of dimples at 700°C, indicating high ductility. However, at 825 and 925°C, fracture surface contains flat facets representing intergranular failure. In steel grade C, fracture surface of specimen tested at 900°C, contain a lot of cavity-like features apart from very small amount of microvoids. Similarly in the Nb–V based steel grade E, mostly cavities are seen on sample representing test at 900°C. Cavity-like features are also seen in steel samples tested at 900°C for material F, but in material F, the cavities are much more distinctly visible with fewer no of smaller voids. At temperature 800°C, the fracture surface contains only flat features in material E, whereas at 850°C, the fracture surface contains mainly flat features and some amount of dimples. In samples with high ductility such as the one tested at 1100°C, only microvoids are seen.
Fractographs of sample A at a 700°C, b 825°C and c 925°C
Fractographs of sample C at 900°C
Fractographs of sample E at a 825°C, b 850°C, c 900°C and d 1100°C
Fractographs of sample F at 900°C
Discussion
The sharp ductility loss in the ductility trough is related to the intergranular fracture occurring in the material. In C–Mn steel, intergranular fracture is mainly attributed to formation of thin film of ferrite along austenite grain boundaries.
7
As the material with such a microstructure is deformed, ferrite being soft as compared to austenite will be preferentially deformed than the later as a result of which strain concentration occurs along this soft phase. This effect in turn leads to microcrack formation along grain boundaries, which subsequently form intergranular fracture.7,10,12 The thin film of ferrite can be induced by deformation and can form between the temperature of Ar3 and Ae3. Thus, at temperatures between Ar3 and Ae3, the specimen may fail with an intergranular failure in a hot ductility test if such a film of ferrite forms. The Ae3 and Ar3 temperatures for the steels studied in the present work have been presented in Tables 1 and 2. While Ae3 temperatures have been calculated using Thermocalc software, Ar3 is calculated using the regression equation developed by Mintz et al.,
12
and the equation is presented as follows
Hot ductility results for both C–Mn and microalloyed steels
Thus, lower ductility in all the steels, observed in the temperature range between their respective Ar3 and Ae3, can be correlated to the formation of this thin film of ferrite. However, for steel A, at 925°C, the failure as observed in the fractograph (Fig. 9) is intergranular. At this temperature, which is above Ae3, only austenite phase will be there in the microstructure. So it is obvious that formation of thin film of ferrite cannot be the cause of intergranular failure. This failure has occurred when grain boundary sliding would have caused decohesion or formation of grain boundary cracks. The decohesion could have been due to segregation of elements like S as described by Tacikowski et al. 8 Even if some S will be in combined form with Mn at this temperature, the amount of S available in the uncombined form in a steel with low Mn/S ratio will be more and on segregation to grain boundary, this S will cause the damage in terms of lowering the ductility. Hence, in steel A, which is having Mn/S ratio of 36, the ductility is found be very low at 925°C. As per Lankford, 1 the grain boundary cracks can form in steel containing S due to formation of FeS along austenite boundaries, which provides easy path for crack propagation because of the low melting point of Fe–FeS eutectic. The melting point of the eutectic is 982°C as mentioned by Wilber et al. 11 The test temperature being discussed is 925°C, which is much below the eutectic melting point. Hence, formation of FeS cannot be the cause of this embrittlement. However, inclusions like MnS may get precipitated along grain boundaries. Because of the low Mn/S ratio, the temperature at which MnS will precipitate out will be lower, and if the temperature of precipitation is close to this test temperature of 925°C, then the precipitates formed will be of smaller size due to reduced diffusivity of element Mn. These small size precipitates will restrict the grain boundary sliding causing strain concentration and then microvoids get formed around these inclusions due to incompatibility in deformation between inclusions and the matrix ferrite phase. These microvoids in turn link up with each other forming microcracks along the grain boundaries. This may also cause intergranular failure resulting in lower ductility. All the C–Mn steels being discussed contain 0·033–0·053 wt-Al. It is very likely that AlN can form at temperatures <950°C as presented by Crowther et al. 9 and then these AlN reduce hot ductility in presence of S along the grain boundaries. For grade C, at 900°C, the fracture surface contains microvoids that surround large cavity-like features. Some of these cavity-like features seem to be a result of more than one flat facets meeting at a point. For this grade, the amount of Al is relatively higher, and hence, the volume fraction of AlN formed at this temperature will also be higher. The microvoids seem to have formed around AlN precipitates, which would have formed along austenite grain boundaries. These voids weaken the grain boundary forming microcracks, and they in turn lead to intergranular fracture forming these flat facets.
The ductility loss due to intergranular fracture has led to the sudden drop in some force elongation curve shown in Fig. 7. At higher temperatures, fracture occurred by mechanism involving nucleation of voids, void growth and void coalescence. Void growth and coalescence mostly occur beyond peak load. It means the process of void growth and coalescence is not getting interrupted by grain boundary crack propagation and failure is delayed by some mechanisms at higher temperatures >950°C. The mechanism could be grain boundary migration or dynamic recrystallisation as described by Ref. 3.
In microalloyed steel E, thin film of deformation induced ferrite can form at low temperature such as at 800°C, which is between their respective Ar3 and Ae3, and this film can result in intergranular fracture as seen in Fig. 11. The effects of intergranular fracture are reflected on its ductility and on the force elongation curve shown in Figs. 6 and 8. In material E, at 850°C, the fracture surface contains mainly flat features and some amount of dimples. The presence of dimples seems to be along the prior austenite grain boundaries. This indicates that these dimples have formed out of microvoids, and these microvoids are formed due to the grain boundary precipitates. The possible grain boundary precipitates are Nb (C, N) and V (C,N) as mentioned in literatures. 3 As more and more deformation is given, these microvoids link up forming intergranular cracks, and then later, they cause fracture leaving flat features on the fracture surface. In the fractograph of material E tested at 900°C, a few big cavities-like features are seen, and they are elliptical in shape. These features are the traces of intergranular cracks formed due to precipitates at grain boundaries and these cracks would have got isolated by mechanism like grain boundary migration making the intergranular cracks look like cavities. Some big cavities are also formed when more than one flat facets meet. Similar cavity-like features are also seen in fractograph of material F at 900°C.
The ductility loss due to intergranular cracks formation is happening in two temperature ranges. Towards the higher temperature side of the ductility troughs, the loss is due to precipitates along austenite grain boundaries and at lower temperature, the ductility loss, which is of higher amount, is due to formation of thin film of ferrite. These two-stage loss is clearly reflected in the ductility curves for Nb and Nb–V based microalloyed steel. In rest of the alloys studied, this two-stage ductility loss is not evident in the hot ductility curves. The Ae3 and Ar3 temperatures of the alloys without Nb content are higher, and they are closer to the temperatures at which the ductility loss due to precipitate formation or grain boundary segregation of elements in austenite occurs. Thus, the range of temperature in which these two ductility loss occurs is smaller as compared to the microalloyed steel containing Nb. Hence, the transition of the two-stage ductility loss is not clearly visible in the steels without Nb. The same reasons also explain the gradual loss in ductility as shown in Fig. 6 for microalloyed steel and the sharp decrease in ductility observed in C–Mn steels and V based steels.
The ductility of the material at low temperature side of the trough is proportional to the amount of ferrite in the microstructure. As the amount of ferrite increases, the ductility becomes higher. The Ar3 temperatures for steels D, E and F, which contain element Nb, are 700, 736 and 701°C respectively. Since the Ar3 temperature of these materials are very close to 700°C, the amount of ferrite will also be very less in the microstructure at 700°C. It is because of this reason, the ductility of these materials at 700°C is still low as compared to the ones for C–Mn grades and V based steel grade whose Ar3 temperatures are much higher.
Conclusions
In C–Mn steel, decreasing the Mn/S ratio decreases the hot ductility. In other words, the bending or straightening temperature is increased with decrease in Mn/S ratio. Thus, in C–Mn steels, in order to prevent formation of transverse cracks in thin slab, sulphur level in the chemistry needs to be kept low.
Straightening or bending temperature in thin slab caster should never be kept lower than the Ae3 in order to avoid forming thin film of ferrite along the grain boundaries, which is one of the causes for intergranular failure.
Ductility trough is seen in both Nb and V based steel. Multiple troughs are seen in the hot ductility curve for steel containing Nb.
The recovery of ductility in steel containing Nb at lower temperature range (<750°C) is much less as compared to the ones for V based steel and the C–Mn steel grades.
Ductility gets drastically reduced when specimen fails with intergranular fracture. The straightening and bending operation may be performed at such a temperature where this kind of failure can be avoided.
Footnotes
Acknowledgement
The authors would like to thank Ms V. Singh, who has helped them in carrying out a large number of hot ductility experiments in gleeble machine. The authors also thank R&D management of TATA Steel Ltd for their support in carrying out the present work. The authors are also thankful to Mr S. Prakash and Mr P. Patnaik for facilitating slab samples from LD2 & Slab Caster of TATA Steel Ltd.