Abstract
Controlled amounts of cold work are shown to cause a minimum in the ductile to brittle transition temperature (DBTT) in a ferritic steel at a critical level of ∼1·5. Mechanical property assessments show that the hardness values exhibit the same trend. A theory is advanced for explanation of these effects, based on work hardening and Cottrell–Bilby locking models. Consideration is given to an alternative Ashby–Embury model, but it is concluded that the former approach is most successful in predicting the observed DBTT shift behaviour. Although independent of fracture surface type, the degree of plastic deformation shows some dependency on the grain boundary character. This leads to the conclusion that the matrix yield strength is the primary factor in determining the DBTT in these steels. Discussion focuses on methods for exploiting the effect to give higher toughness steels utilising knowledge of how to control matrix hardening and cleavage fracture strength.
List of symbols
Burger's vector shear modulus Boltzmann's constant plane strain fracture toughness temperature dependant exponent absolute temperature critical transition temperature lattice constant activation energy for lattice resistance controlled glide dislocation density lattice resistance for given temperature velocity of dislocation velocity constant
Introduction
It is apparent that it should be possible to control the ductile to brittle transition temperature (DBTT) in ferritic steels by compositional control or thermomechanical processing. Compositional control by manganese additions 1 and additions to reduce the grain size 2 have shown to be instrumental in reducing the DBTT in plain carbon steels. Thermal annealing procedures can soften the matrix to produce a lower DBTT. 3 More recently, the mechanical part of this treatment has become more important in reducing the DBTT, 4 and it is the purpose of this paper to highlight the importance of a critical amount of cold work, which seems to be required to reduce the DBTT to a minimum value. Conventionally, studies show that an amount of cold work will result in a material with a higher yield stress and therefore DBTT5,6; however, these studies typically focus on a substantial degree of deformation (>10). This study focused on much smaller amount of deformation in the region of 0·5–5, offering explanation as to why the findings differ from the sources mentioned above.
Ashby and Embury 7 developed a theory concerning the crack propagation rate in impact-tested material. The model relies on an appreciation of the dislocation activity ahead of a crack tip under high strain rate conditions. The model concludes that there is a critical amount of cold work required, to produce a sufficient dislocation density for plastic deformation to occur around the advancing crack tip. Below this critical value, the constraint of plastic work will not hinder the crack growth and will behave more like a classic Griffith's crack. An interesting implication of this model is the prediction of an intermediate amount of cold work, which gives rise to a minimum in the DBTT. Below this critical cold work level, the Ashby and Embury mechanism applies, and the DBTT will increase with decreasing amounts of cold work. Beyond the critical deformation level, the more traditional arguments for ductile to brittle transition become more important, and increasing cold work results in higher dislocation density. This, in turn, leads to higher matrix yield strength, which means that the DBTT will increase with increasing amounts of cold work. This leaves us with the situation that there is a minimum in the DBTT at some intermediate level of cold work.
This paper describes experiments that confirm this effect in C–Mn pressure vessel steel, with the critical cold work strain being in the region of 2. The addition of manganese to plain carbon steels is known to have a beneficial effect on the fracture properties. Various microstructural and microhardness observations are made, and other models are applied, so that a coherent, fully understandable mechanism can be proposed for the phenomenon. It will be shown that instead of the Ashby and Embury model, a strain aging Cottrell–Bilby locking approach can be used with much greater effect to predict the experimental results. The implications of the model for future thermomechanical processing approaches to DBTT control in ferritic steels will be described.
Experimental method
Samples of a C–Mn pressure vessel plate steel were annealed at 980°C for 24 h, followed by slow cooling to form a ferritic/pearlitic microstructure. The elemental composition of the carbon manganese pressure vessel steel is Fe–0·18C–1·30Mn–0·02S–0·02P–0·36Si–0·10Cu (wt-) Sections were cut from the plate of dimension 100×50×25 mm. These were subsequently annealed at 980°C, then cold rolled at the University of Sheffield. The focus of this report will be on specimens with a cold work percentage 0–2·5. The annealing was carried out in an exothermic gas generator, which produces a reducing atmosphere. This prevented scale formation, which could affect the measured thickness values. Table 1 shows the specimen thickness before, and after rolling, along with the actual achieved reduction. The samples were sectioned and examined metallographically in an Olympus light optical microscope, and with the LEO 1530VP field emission gun scanning electron microscope. Hardness tests were carried out on all samples with a Mitutoyo Vickers Hardness testing machine. Charpy impact testing was performed on specimens machined from the rolled plate, with dimensions 10×10×55 mm, following standard BS 131-6:1998. The impact testing was carried out at Loughborough University using a Losenhausen impact test machine at temperatures ranging from −100 to 100°C. Tensile testing was performed on dumbbell shaped samples machined from the rolled plate, using a Lloyds Tensile test machine with a 50 kN load cell.
Thickness of cold rolled specimens before and after deformation, showing achieved thickness reduction
Results
Figure 1 shows a selection of optical micrographs of the C–Mn steel after various degrees of cold work. A clear ferritic pearlitic structure is seen with a grain size of ∼50 μm, measured by line intersection analysis. Subsequent cold rolling does not substantially affect the microstructure.
Optical micrograph showing microstructure of C–Mn steel at varying degrees of cold work
The DBTT was determined by taking the temperature at which the impact energy line reached the midway point between maximum and minimum energy values. Fractographic information is displayed in Fig. 2, showing the fracture surfaces for 0·5 and 2·5 specimens. These clearly show the predominance of transgranular fracture; there are some minor features which may indeed show evidence of intergranular, but these account for <5 of the fracture surface. Values of the DBTT as a function of cold work are given in Fig. 3.
Images (SEM) showing predominantly transgranular fracture surface. Only 0·5 and 2·5 specimens are shown; however, all specimens showed very similar fracture features
DBTT values as a function of cold work
The C–Mn pressure vessel steel tensile test results are shown in Fig. 4a, where the samples were tested at a high strain rate. The results show an absent yield point at above 1 deformation. The corresponding hardness values, which should correspond to the yield point values, are given in Fig. 5. A clear drop in hardness appears at about 1 deformation.
a tensile test results showing effect of prior cold work and b schematic to show relationship between prior cold work and upper yield point
Hardness test results showing minimum hardness around 1 (±2 HV)
Discussion
The metallography demonstrates that the microstructure of the annealed and cold rolled C–Mn steel was ferrite/pearlite, with a mean grain size of 50 μm. This is typical for pressure vessel steel. Figure 4a shows the tensile test results, and a smoothing of the yield point can be observed at >1 cold work. This is thought to be due to the high strain rate used (10 mm min−1). The prominent yield point (ambiguous from Fig. 4a due to high strain test rate), diminishes as higher degrees of cold work are introduced. Plain carbon steels exhibit an upper yield point, which disappears and a critical value of cold working. Figure b shows an idealised case of what normally would be expected from this material under tensile testing. This upper yield point is strain dependant; however, this relationship falls outside the scope of this project. Figure 4a indicates that this process is occurring, even if the detail cannot be observed from the tensile test data.
The mechanism for introducing cleavage failure is imagined to be from one of two sources, each of which can be described by the following two mechanistic models, explaining the effect that cold work has. These are:
Ashby and Embury model, 7 which caters for high strain rate deformation and the ease of fast crack propagation in varying dislocation density situations
Cottrell–Bilby model, 8 in which the material exhibits a yield point at deformations below a critical level, during which the upper yield point reduces with increasing deformation. Thereafter, the hardness increases with increasing cold work due to the enhanced dislocation density effect.
The predictions for the Ashby and Embury model
7
are given in Fig. 6. They show that, at a critical dislocation density, the DBTT is at a minimum. This dislocation density is difficult to relate exactly to the cold work, but the model shows critical values being around 6×1014 m−2. The values obtained from the model cannot be considered valid; however, the predicted trend exhibits good agreement with that observed in the experimental data. Parameters used in the calculation are given in Table 2. The model relies on predicting the degree of interaction between dislocations and a fast propagating crack tip. Below a certain dislocation density, there will be insufficient dislocations to assist in blunting the crack tip. Therefore, the material has a low work of fracture and high DBTT. Above the critical degree of cold work, there are sufficient dislocations to interact, and the crack tip is allowed to blunt, whilst the crack passes through. This results in the work of fracture increasing and a lowering of the DBTT. Finally, the model accounts for very high densities of dislocations being counterproductive through work hardening, so that the yield strength increases, and the DBTT is raised. Equation (1) shows the calculation used to forecast the DBTT shift
Ashby–Embury predictions of relationship between cold work and DBTT
Parameters and values used in Ashby–Embury model calculation
The Cottrell–Bilby 8 solute atmosphere dislocation drag model suggests that there will be a large upper yield point in undeformed material. This is because, in the absence of cold work, there is insufficient strain to separate dislocations from the interstitial carbon atoms in the ferritic matrix. In the critical range of deformation between 0 and 1, the yield point becomes less prominent (see Fig. 4a). This results in a drop in the yield strength as measured by hardness and tensile testing. The hardness also shows a drop with increasing deformation in the 0–1 range (Fig. 5).
The Cottrell–Bilby mechanism seems to be more appropriate to the situation being discussed. The hardness measurements, showing that the hardness drops for strains <1, would not be predicted from the Ashby–Embury model because the hardness should increase monotonically with strain in that model: no special restrictions are put on the dislocations in the 0–1 strain range.
It is important to mention that the Ashbury–Embury model assumes control of the DBTT through changes in the cleavage fracture strength and not through the matrix yield strength as is the case for the Cottrell–Bilby model. Figure 7 looks at the separate effects on DBTT of the Ashby–Embury model (Fig. 7a), the Cottrell–Bilby model (Fig. 7b) and then considers the combined effect (Fig. 7c). The cross-over of the localised yield strength and the cleavage fracture strength curves, which represents the DBTT, moves from A→B→C in the separate models. The combined model shows the additional reduction in DBTT in going from A→B→C before the cold work effect begins to dominate above 1 strain to increase the DBTT to D. In addition, it is suggested that grain size can have an effect even if the cleavage is transgranular; however, this is an area of continuing work.
Effects on the DBTT using a Ashby–Embury model, b Cottrell–Bilby model and c combined model
As shown above, neither one of these models describes the full picture; however, taking aspects of both into the combined model, a fuller explanation can be given. If the Cottrell–Bilby mechanism was the sole explanation for this DBTT reduction trend, then after several months, the interstitial carbon would return at room temperature to the dislocations, and the upper yield point would return (the DBTT would return from B to A in Fig. 7a). If the Ashby–Embury mechanism operates, then there is no condition imposed that lowers the yield strength at small strains; therefore, the DBTT will remain indefinitely at B, independent of time.
Finally, it must be emphasised that the Ashby–Embury model applies to fast crack growth fracture, and therefore, if the DBTT/fracture toughness measurements were performed at slow strain rates, this component of the DBTT shift would be lost.
Conclusions
A series of experiments designed to show the effect of cold work on DBTT in C–Mn pressure vessel steel is described. It is demonstrated that reductions in DBTT of ∼10°C are discovered at a strain of 2, and the DBTT returns to higher temperatures with larger amounts of cold work. Also, a similar minimum in hardness levels of ∼20 HV is observed at 1·5 cold work level. Two models, due to Ashby–Embury and Cottrell–Bilby, are proposed to explain this effect. The Cottrell–Bilby mechanism provides the best explanation for the effect, but a combined model is proposed that takes advantage of the versatility of the Ashby–Embury model. This combined approach could lead to substantial permanent benefits being realised for DBTT reduction in ferritic steels for the future.