2016 Volume 56 Issue 11 Pages 2047-2056
A 0.4C-2Cr-1Mo-2Ni steel (in mass%) was austenitized at 1123 K, followed by ausforming (AF) using multi-pass caliber rolling with a rolling reduction of 74%. The AF samples were subsequently tempered at 773 K and deformed by multi-pass caliber rolling (i.e. warm tempforming, WTF) with a rolling reduction of 46%. Their microstructures and mechanical properties were investigated and compared to those of the quenched and tempered samples (non-AF samples) which were subjected to the sequent WTF with rolling reductions of 46% and 74%. The WTF with rolling reductions ranging 46 to 74% resulted in the strengthening and toughening of the AF and non-AF samples through the evolution of anisotropic and ultrafine grain structures with strong <110>//rolling direction (RD) fiber textures. The AF sample demonstrated the faster kinetics of the microstructural changes, i.e. refinement in the transverse grain size and development of highly elongated grains than the non-AF sample. As a result, the AF sample exhibited a better combination of ultra-high strength and toughness than the non-AF sample, when compared at the rolling reduction of 46%. The combined effect of AF and WTF was especially pronounced in the enhancement of toughness resulting from delamination.
Grain refinement is a promising way for lowering the ductile-to-brittle transition temperature (DBTT) as well as strengthening of steels.1,2) Variety types of thermomechanical treatments have been developed.3) However, ultra grain refinement to 1 μm or less was reported to result in a decrease in Charpy absorbed energy at upper-shelf region.1,2)
Recently, we have reported that ultrafine elongated grain (UFEG) structures with strong <110>//rolling direction (RD) fiber textures, which are similar to those in cold drawn steel wires,4,5,6) are particularly effective in toughening of medium-carbon low-alloy steel bars.7,8,9,10,11) A 0.4C-2Si-1Cr-1Mo steel (mass%) with an UFEG structure had an average yield strength (σy) of 1.84 GPa and an average Charpy V-notch absorbed energy (vE) of 226 J at room temperature.7,9,11) Furthermore, the UFEG structure steels were observed to exhibit an inverse temperature dependence of toughness at subzero temperatures, as a consequence of delamination, in which cracks branched and propagated in the longitudinal direction of impact bars.7,8,9,10,11) The vE of the UFEG structure steels increased as the temperature decreased, in contrast to a ductile-to-brittle transition of conventional ultra-high strength steels. It has also been clarified that the microstructural factors controlling the occurrence of delaminations are the transverse grain size, grain shape, <110>//RD fiber texture, and carbide particle distribution in the UFEG structure.10)
The UFEG structures have been evolved through warm deformation of tempered martensite structures (hereinafter referred to as warm tempforming, WTF), using multi-pass caliber rolling with a rolling reduction of about 80% at 773 K.7,8,9,10,11) The microstructure evolution during the WTF was observed to depend on variations in the crystallographic orientation and the geometric arrangement of blocks and packets in the tempered martensite.9,11) As a result, the WTF samples exhibited hierarchical microstructures consisting of “packet bands” and “ultrafine elongated grains (UFEGs)”; the packet bands were evolved through the extension of the packets in the RD and the UFEG structure mainly evolved through the deformation of fine blocks.11) Furthermore, for the 0.4C-2Si-1Cr-1Mo steel, the width of the packet bands was observed to reduce with decreasing the prior-austenite grain (PAG) size (the PAG size range of 50–600 μm), and the occurrence of the delamination became more pronounced at lower temperatures, leading to the enhancement of vE.11)
On the other hand, ausforming (AF) is a thermomechanical treatment in which austenite is deformed below the austenite recrystallization temperature and then quenched.12) When austenite is deformed at temperatures where austenite is stable, this treatment is referred to as modified AF, to distinguish it from conventional AF which involves the deformation of metastable austenite.13) McEvily Jr and Bush14) observed an inverse temperature dependence of toughness in the vicinity of 473 K for an ausformed 0.2C-3Mo-3Ni steel. In this case, delamination cracks propagated along the boundaries of elongated PAGs. However, the delamination did not occur at room temperature and the vE was reduced to 33 J. Ohmori and Yamazaki15) investigated the combined effect of modified AF and WTF on the mechanical properties of JIS-SNC631 and SCM435 steel plates. They reported that when the modified AF steels were subjected to the WTF at 873 K with a rolling reduction of 50% using plate rolling, the WTF resulted in higher strength but lower ductility and toughness in the yield strength range of 0.8–1.1 GPa. However, there is still a lack of data on this subject.
Since the WTF at 773 K with a rolling reduction of about 80% is a severe plastic deformation, lower rolling reduction is preferable. Morito et al.16) observed in a 18Ni maraging steel, that the packet and block widths decreased by the AF with a rolling reduction of 60% at 773 K, whereas the packets, in addition to PAGs, were elongated along the RD. Hence, a faster kinetics of microstructure evolution from an ausformed martensite into an UFEG structure will be expected in the WTF process. The aim of this study is to determine whether the AF with a rolling reduction of 70%, which may lead to the formation of elongated PAG structure, is effective to produce an excellent combination of ultra-high strength and toughness in the following WTF process at 773 K with a rolling reduction of 50%. A 0.4C-2Cr-1Mo-2Ni steel was subjected to multi-pass caliber rolling, and the influence of the AF on the microstructure, the tensile and Charpy V-notch impact properties of the WTF steels was investigated.
A steel with a chemical composition of 0.40 C, 0.27 Si, 0.20 Mn, 2.01 Cr, 1.03 Mo, 2.04 Ni, 0.001 P, <0.001 S, 0.020 Al, 0.0015 N, 0.0040 O and the balance Fe (all in mass%) was used in this study. A 100 kg ingot was prepared by vacuum melting and casting. The ingot was subsequently homogenized at 1473 K and hot-rolled into plates with a thickness of 4 cm. Figure 1 presents calculated phase diagram of Fe-2Cr-1Mo-2Ni-C alloy system and the diagrams of thermomechanical treatments. Blocks measuring 12×4×4 cm were cut from the hot-rolled plates and subjected to three different thermomechanical treatments (TMTs) in order to vary the initial microstructure before the WTF as follows.
Fe-2Cr-1Mo-2Ni-C (mass%) phase diagram calculated by Thermo-Calc software (a), thermomechanical treatment diagrams involving ausforming (AF) and warm tempforming (WTF) (b), and schematic drawing of a pair of rolls used for caliber rolling and the relationship between the position of a rolled bar and a Charpy V-notch (CVN) bar (c).
(1) The blocks were solution-treated at 1473 K for 0.5 h to reduce undissolved carbides, and then subjected to multi-pass caliber rolling into squared bars with a cross-sectional area of 15.2 cm2, followed by air cooling. The squared bars were austenitized at 1123 K for 1 h, and subjected to multi-pass caliber rolling into squared bars with a cross-sectional area of 3.9 cm2, followed by air quenching (the AF samples). The rolling reduction was 74% in 8 passes. The finishing rolling temperature was experimentally estimated to be above 950 K; the present steel was mainly deformed in the temperature region where austenite is stable (it was categorized as a Modified AF). In the present AF, the steel bars were air quenched because quench cracking occurred when the steel bars were water quenched. The temperatures of Bs17) and Ms18) were calculated to be 675 K and 569 K, respectively, and the average cooling rate for the AF sample during the air quenching was estimated to be about 30 K/min in the temperature range of 773 to 573 K. Figure 2 shows the optical microstructure of the AF sample in as-air quenched condition. Dark-etching constituents inside elongated PAGs appear to be bainite and/or auto-tempered martensite. The Vickers hardness of the air quenched samples was HV654, and was slightly lower than that of the water quenched samples (=HV690). Concerning the influence of banitic structure on the mechanical properties of UFEG structure steel, Jafari et al.19) reported that the UFEG structures obtained from initial martensitic and bainitic structures had almost identical mechanical properties for medium-carbon low-alloy steels (JIS-SCM440 steel except for ultralow phosphorus and sulfur concentrations).
Microstructure of the AF sample in as-air quenched condition (3% Nital etch).
(2) A part of the AF bars were austenitized at 1123 K for 1 h, followed by air quenching (the Q1 samples).
(3) The blocks were solution-treated at 1473 K for 0.5 h, and then subjected to caliber rolling into squared bars with a cross-sectional area of 8.1 cm2, followed by air cooling. These hot rolled bars were austenitized at 1123 K for 1 h, followed by water quenching (the Q2 samples).
The AF, Q1 and Q2 samples which were processed through treatments TMT 1–3 were tempered at 773 K for 1 h and subsequently subjected to WTF using multi-pass caliber rolling into squared bars with a cross-sectional area of 2.1 cm2, followed by air cooling. The total rolling reduction (r) were 46% in 6 passes for the AF and Q1 samples and 74% in 13 passes for the Q2 samples, which corresponds to the equivalent strain (εeq=2/√3ln (1/(1-r/100)) of 0.7 and 1.5, respectively. During the WTF, the samples were held in a furnace for 5 min following every three or four passes. Some of the AF and Q2 samples were tempered at 773 K for 1 h and subsequently air cooled.
The principal axes of the squared bar were defined (Fig. 1(c)). The axis that was coincident with the rolling direction was defined as the RD, the axis which was coincident with the direction of the main working force at the final pass was defined as the ND, and the axis which was normal to the RD and ND was defined as the TD.
2.2. Microstructural Characterization and Mechanical TestingThe microstructures were observed by optical microscopy, scanning electron microscopy (SEM), and transmission electron microscopy (TEM). The electron backscatter diffraction pattern (EBSP) analysis was performed using a SEM which was equipped with a field emission gun (FE-SEM). The EBSP measurements were done on a hexagonal grid with a step size of 50 nm in the area of 25×25 μm2, or 200 nm in the area of 100×100 μm2. The conditions for the FE-SEM/EBSP analysis are described in detail elsewhere.9,20) The integrated intensities of the (110) XRD peaks (Im) were measured on the RD planes. The ratio of Im to the integrated intensity of the (110) XRD peak (Is) for the standard samples was measured to evaluate the development of the <110>//RD fiber textures.
Tensile tests were conducted on the JIS Z 2201-14A specimens (with a diameter of 6 mm, and machined in the RD (JIS-14A specimens)), at a crosshead speed of 0.85 mm/min. The gauge length of the JIS-14A specimens was 30 mm. A σy of 0.2% offset was reported. Charpy impact tests were performed for the full-size 2 mm V-notch specimens that were machined in the RD. The striking direction (SD) of the impact tests had an angle of ~45° to the TD and ND (Fig. 1(c)).
Figure 3 shows the matrix microstructures before the WTF. The initial microstructure of the non-AF samples is characterized as a tempered martensitic structure where the equiaxed PAGs were subdivided into packets (Fig. 3(a)). These packets were further subdivided into fine blocks. As for the AF sample, the PAGs and the packets were elongated along the RD (Figs. 2 and 3(b)). The elongated packets were further subdivided into fine blocks. The average intercept length (ILAV) for the PAGs along the SD (ILγ) was measured by optical microscopy; they were 16 μm for the Q1 and Q2 samples (non-AF samples), and 11 μm for the AF sample, respectively. The ILγ for the AF sample was approximately half of that for the sample before the AF (ILγ =25 μm); it was reasonable to consider that the elongated PAG structure mainly evolved through the extension of the PAGs in the RD during the AF with a rolling reduction of 74%. The ILAV for the packets (ILp) was measured to be 7 μm in the non-AF samples, while the ILp along the SD (ILp_SD) and RD (ILp_RD) were measured to be 5 μm and 11 μm in the AF samples, respectively. The ILAV for the blocks (ILB) were measured to be approximately 0.8 μm in the non-AF samples, and 0.6 μm in the AF samples, respectively; the boundaries of the blocks were defined as grain boundaries with a misorientation angle of 10° or over.9,20) Furthermore, the inverse pole figure for the RD indicates that the AF sample contains a weak <110>//RD fiber texture (Fig. 3(c)). Hence, the AF sample has an anisotropic matrix structure prior to the WTF. The width of packets and the size of blocks were also finer in the AF sample than in the non-AF sample, although the initial PAG size was larger in the AF sample than in the non-AF sample. The AF is thus considered to be effective in developing a fine and highly anisotropic microstructure through the following WTF with lower rolling reduction.
Microstructures of the non-AF (a) and AF samples ((b),(c)) tempered at 773 K for 1 h. The dash lines denote the prior-austenite grain (PAG) boundaries in the inverse pole figure (IPF) maps for the SD. The black lines in the IPF maps indicate block boundaries, which were defined as grain boundaries with a misorientation angle of 10° or over.9,20) IPF for the RD in the AF sample (c) is also shown. (Online version in color.)
Figure 4 shows the microstructures of the samples subjected to the WTF at 773 K. The deformation of the blocks and packets during the WTF depends on their variations in the crystallographic orientation and their geometric arrangement, resulting in the evolution of hierarchical, heterogeneous microstructures in the AF and non-AF samples (Figs. 4(a)–4(c)). As expected, the matrix microstructure of the non-AF sample tempformed at 74% consists of the packet bands and UFEGs that are elongated along the RD (Figs. 4(c), 4(f)). The UFEG structures evolve inside the packet bands and the crystallographic orientation and geometric arrangement of the UFEGs are relatively similar within a packet band. It is characteristic in the non-AF sample tempformed at 46%, that relatively coarse block-shaped grains are observed in places (as indicated by arrows), in addition to elongated grains (Figs. 4(b), 4(e)). Such block-shaped grains are suggested to be formed through the deformation of the blocks whose long axes are aligned at high angles to the RD. The kinetics of the microstructure evolution from the coarse block-shaped grains into the UFEGs appears to be slow, and, as a result, relatively blocky grains remain even after 74% rolling (as indicated by arrows). On the other hand, the coarse block-shaped grains are hardly observed in the AF sample tempformed at 46% (Figs. 4(a), 4(d)). The grain structure appears to be more flattened and elongated in the AF sample than the non-AF sample, at the rolling reduction of 46%. As for the morphology and distribution of carbide particles, the TEM observation (Figs. 4(g)–4(i)) reveals that the non-AF and AF samples have an almost similar distribution of spheroidized carbide particles after the WTF with rolling reductions ranging 46% to 74%. It has been commonly observed that carbide particles grew slightly, becoming more spherical during the WTF at 773 K.8,9,10,11) Furthermore, strong <110>//RD fiber textures are developed through the WTF (Figs. 4(j)–4(l)).
Microstructures of the AF sample tempformed at 46% ((a), (d), (g), (j)), and non-AF samples tempformed at 46% ((b), (e), (h), (k)) and 74% ((c), (f), (i), (l)). The black lines in the IPF maps for the SD represent high angle boundaries (HABs) with a misorientation angle of 15° or over. The white and black arrows indicate block-shaped grains and carbide particles, respectively. IPFs for the RD are also shown. (Online version in color.)
Figure 5 summarizes the microstructural parameters as a function of the equivalent strain (εeq). The KAM value, measured by EBSP analysis, reflects the density of the geometrically necessary dislocations (GNDs)21) which are associated with plastic strain gradients.22) The average KAM value for the non-AF sample shows a maximum at the εeq of 0.7 (=rolling reduction of 46%), followed by a slight decrease. A similar tendency was observed in the 0.4C-2Si-1Cr-1Mo steel subjected to the WTF at 773 K.9) This suggested that the GNDs might accumulate to form low angle boundaries and eventually develop into high angle boundaries (HABs), leading to grain subdivision during the WTF. Actually, the fraction of HABs with a misoriantation angle of 15° or over was confirmed to increase with increasing the εeq. In as-tempered state, the average KAM value tends to be higher in the AF sample than in the non-AF sample, and it also increases with the WTF. However, the increment in the average KAM value in the AF sample is smaller than that in the non-AF sample; their average KAM values become almost identical at the εeq of 0.7. The ILAV along the SD (ILHAB_SD) and RD (ILHAB_RD) for the HABs in the non-AF samples were measured to be 0.55 μm and 0.79 μm at the εeq of 0.7, and 0.33 μm and 0.92 μm at the εeq of 1.5 (=rolling reduction of 74%), respectively. The ILHAB_SD and ILHAB_RD in the AF samples were measured to be 0.42 μm and 0.88 μm at the εeq of 0.7. Hence, the AF sample has finer transverse grain size and higher aspect ratio than the non-AF sample, at the εeq of 0.7. A similar tendency is also observed for the packet structure. The ILp_SD and ILp_RD in the AF sample tempformed at 46% are comparable to those for the packet bands in the non-AF sample tempformed at 74%. Furthermore, the development of the <110>//RD fiber texture is shown in Fig. 6, in which the data points of the integrated (110) XRD peak intensity ratio (Im/Is) are plotted as a function of the εeq. The data points in the 0.4C-2Si-1Cr-1Mo steel are also present. The Im/Is in the non-AF samples increases linearly with increasing the εeq. Bourell23) also reported that the degree of {100}<110> texture became stronger with the rolling strain in multi-pass warm plate rolling of low-carbon steels and was independent of the rolling temperature between 813 K and 923 K. Although the AF sample has the advantage of having a <110>//RD fiber texture prior to the WTF, the degree of <110>//RD fiber texture becomes comparable between the non-AF and AF samples, at the εeq of 0.7. Therefore, the AF sample demonstrates the faster kinetics of the microstructural changes, i.e. refinement in the transverse grain size and development of highly elongated grain structure, than the non-AF sample, in the WTF with rolling reductions up to 50%.
Changes in average intercept lengths (ILAV) in the SD and RD for the HABs with a misorientation angle of 15° or over (ILHAB), packet boundaries (ILP), and average KAM value for a 1st neighbor rank (KAM-1st) during the WTF.
Change in integrated intensity ratio (Im/Is) of (110) XRD peak on the RD plane during the WTF. Data for the 0.4C-2Si-1Cr-1Mo steel9) are also plotted.
Figure 7 shows the nominal stress-strain curves. The non-AF and AF samples exhibit continuous yielding behaviors in as-tempered state. In contrast, the non-AF and AF samples tempformed at 46 and 74% exhibit discontinuous yielding behaviors. An obvious yield drop during discontinuous yielding is observed in the 74%-rolled sample. Such a yield drop was often observed in ultrafine grained material processed by severe plastic deformation, and it was usually accompanied by plastic instability.1,2) However, an adequate uniform elongation is maintained in the WTF sample, and this is a common feature in the medium-carbon steel with an UFEG structure.7,8,9,10,11)
Nominal stress-strain curves of the WTF samples as a function of rolling reduction.
Table 1 summarizes the tensile properties. In the non-AF samples, the WTF causes increases both in the σy and σB. The effect of the WTF is especially pronounced for the σy. This is commonly observed in the WTF steels.7,8,9,10,11) In as-tempered state, the AF sample has higher σy and σB than the non-AF sample. For the AF sample, the WTF with a rolling reduction of 46% causes an increase in the σy but not in the σB. There is no marked difference in the σy among the non-AF and the AF samples tempformed at 46% and 74%. The average uniform elongation (εu) and total elongation (εt) for the non-AF samples tend to increase with the WTF. The respective values of the εu and εt for the AF samples are almost comparable to those for the non-AF samples at the same rolling reduction. Furthermore, the WTF has no effect on the reduction of area (δ) both in the non-AF and AF samples. Hence, the AF with a rolling reduction of 70% results in an increase in strength without deteriorating elongation. As a result, higher strength is obtainable in the AF sample than in the non-AF sample in the following WTF at 773 K with rolling reductions of up to 50%.
| Sample | r (%) | Yield Strength σy (GPa) | Tensile Strength σB (GPa) | Uniform Elongation εu (%) | Total Elongation εt (%) | Reduction in Area δ (%) |
|---|---|---|---|---|---|---|
| non-AF | 0 | 1.33 | 1.57 | 4.9 | 12.6 | 52.3 |
| 46 | 1.66 | 1.72 | 6.0 | 15.2 | 49.1 | |
| 74 | 1.70 | 1.71 | 6.4 | 15.3 | 51.6 | |
| AF | 0 | 1.54 | 1.74 | 4.3 | 13.3 | 53.4 |
| 46 | 1.72 | 1.75 | 6.0 | 14.7 | 49.7 |
It has been reported that AF steels exhibit higher resistance to softening during tempering compared to non-AF steels.24,25,26) This can be explained in terms of higher density of dislocations,25) denser distribution of carbide particles,25,26) and finer block structure27) in the AF steels. The results in Figs. 3 and 5 can support it. The AF sample has finer block structure and higher density of GNDs than the non-AF sample in as-tempered state; however, the difference in carbide particle distribution between the non-AF and AF samples should be clarified. On the other hand, in the UFEG structure with a strong <110>//RD deformation texture, the microstructural parameters affecting the strength are considered to be its transverse grain size, GND density, and carbide particle distribution.9,10) A comparison among the changes in these microstructural parameters in Figs. 4 and 5 indicates that the σy might increase in response to the increase in the GND density and the decrease in the transverse grain size during the WTF with rolling reductions up to 50%.9) Furthermore, the grain subdivision accompanied by the evolution of strong texture evolution may become more significant in the final stage of the WTF process. Hence, the influence of the transverse grain size on the σy might be more significant through the evolution of an UFEG structure during the WTF, and the AF effect on the strength will become significantly smaller.
3.3. Charpy Impact PropertiesFigure 8 displays the representative fracture appearances of the Charpy V-notched (CVN) impact bars. Figure 9 shows the vE as a function of testing temperature and the rolling reduction. In as-tempered condition, the vE for the non-AF samples is low and it decreases with decreasing testing temperature. In addition, the AF treatment has little influence on the vE; there is no difference in the vE between the non-AF and AF samples at the same testing temperature. It has often been reported that the AF caused an increase in strength but not an increase in the vE.15) By contrast, the WTF causes an enhancement of the vE, in the non-AF and AF samples. The WTF samples exhibit an inverse temperature dependence of the vE; the vE markedly increased with decreasing temperature. This increase in the vE corresponds to the occurrence of crack-arrester type delamination fracture (Fig. 8). According to the variations in the fracture appearances of the CVN bars and the vE, the Charpy impact behavior of the WTF sample can be divided into three regions; 1) the ductile fracture region at a temperature of approximately 350 K or above, where ductile fracture was observed to be the primary mode of failure, 2) the delamination toughening (D.T.) region (as indicated by dash arrows), and 3) the delamination diminishing region. As expected, in the non-AF samples, the WTF at 74% results in much greater enhancement in the Charpy impact properties than the WTF at 46%. Firstly, the average value of vE (Upper Shelf Energy, vEUS) for the 74%-rolled sample exhibiting almost complete ductile fractures is much higher than that for the 46%-rolled sample; the vEUS for the 46% and 74%-rolled samples are approximately 50 J at 350 K and 140 J at 420 K, respectively. Secondly, the delamination toughening for the 74%-rolled sample occurs in a wider temperature range than that for the 46%-rolled sample. The vE for the 74%-rolled sample is especially enhanced as a consequence of the crack-arrester delamination, in which cracks branched and propagated in the longitudinal direction of impact bars, as indicated by + in Fig. 9. The crack-arrester delamination occurs under conditions of nearly uniaxial tension, which is an unfavorable cleavage.28) The stress fielding effect from the formation of microcracks along the RD can also reduce the driving force for fracture.29) As a result, the vE is enhanced through the occurrence of crack arrester-type delamination. When the temperature below which the vE for the WTF samples decrease to the level of its vEUS value or less was defined as the delamination finish temperature (TDFT),10,11) the respective TDFT values were estimated to be approximately 230 K for the 46%-rolled sample, and 170 K for the 74%-rolled sample. As for the AF samples, the WTF with a rolling reduction of 46% also results in the occurrence of crack arrester-type delamination, and the vEUS and TDFT are approximately 75 J at 420 K and 190 K, respectively. Therefore, the AF sample demonstrates superior Charpy impact properties to the non-AF samples, when compared at the rolling reduction of 46%.
Fracture appearances of the CVN specimens tested at 213 K: non-AF and AF samples tempformed at different rolling reductions. The macroscopic crack branching angle (β) was defined as the angle between the crack path and longitudinal direction (//RD) of the impact bar.
CVN absorbed energy (vE) as a function of testing temperature and rolling reduction in the WTF samples. Data points with + denote that the specimen that exhibited a delamination in which the β in Fig. 8 was below 15°. The arrow indicates that the specimen did not separate into two pieces during the impact test.
Figure 10 displays the representative fracture surfaces of the CVN specimens that were tested at 213 K, where a significant difference in the vE was observed between the non-AF and AF samples. The AF sample without the WTF exhibits a quasi-cleavage fracture (Fig. 10(d)). Quasi-cleavage fracture was also observed in the non-AF sample tested at 213 K. By contrast, in the WTF samples, “terraces” and “steps” are formed. The terraces correspond to delamination planes and are generally parallel to the RD (Figs. 10(b), 10(c), 10(f)). Such stepwise crack propagation is commonly observed in crack arrester-type delamination.14,30) In the non-AF sample tempformed at 74%, the fracture mode on the terraces occurs primarily by a quasi-cleavage (Fig. 10(c)), whereas a ductile fracture mode with fine dimple patterns occurs on the steps. Narrow, elongated quasi-cleavage facets that are aligned in the RD on the terraces appear to correspond to the morphology of the UFEG structures. The presence of many steps provides evidence for the extensive local plastic flow during delamination, generating a higher vE. On the other hand, the delamination crack propagation along the RD is less pronounced in the non-AF sample tempformed at 46%. The terraces of the 46%-rolled sample are wider and shorter (Fig. 10(b)) than those of the 74%-rolled sample and are primarily linked with quasi-cleavage planes (Fig. 10(a)). Wide, less-elongated quasi-cleavage facets appear to correspond to the morphology of the microstructure in the 46%-rolled sample. The linkage of the cleavage cracks in the transverse directions might, thus, result in the degradation of vE. The fracture surface of the AF sample tempformed at 46% (Figs. 10(e)–10(f)) appears to be close to that of the non-AF sample tempformed at 74% rather than that of the non-AF sample tempformed at 46%.
Fracture surfaces of the CVN specimens tested at 213 K: non-AF sample tempformed at 46% ((a)–(b)) and 74% (c), and AF samples tempformed at 0% (d) and 46% ((e)–(f)). The white and black arrows indicate ductile steps and transverse quasi-cleavage cracks, respectively.
When weak interfaces and/or weak planes are present parallel to the longitudinal direction (//RD) of the impact test bar, the interaction between the weak interfaces (and/or weak planes) and the tensile stress along the SD (σt//SD) can cause delaminations.14,30,31) The σt is generated by the localized plastic constrain at the notch and/or the crack tip, and it is proportional to the σy. If the Tresca yield criterion is assumed in the process zone within the plastically yielding region of the notch root, the σt//RD and σt//SD in the process zone can be approximated as 2.6×σy and 1.6×σy,32) respectively.
Bcc steel cleaves along the {100} planes, and the coherence length of the {100} cleavage planes corresponds to the effective grain size, Deff, for cleavage.33) The cleavage fracture stress (σc) is proportional to the inverse square root of the Deff.33) In lath martensitic steels, block size governs the Deff for cleavage; cleavage cracks were observed to deflect or blunt at the boundaries between blocks with different Bain axes.34) The present AF sample has an anisotropic microstructure consisting of PAGs and packets that are elongated along the RD (Figs. 2 and 3(b)). However, the blocks inside the packets are not aligned along the RD in as-tempered state. Also, the ILAV for the block was measured to be almost the same between the SD and RD; the block structure is considered to be an isotropic grain structure for cleavage fracture. As a result, once cracks initiated in the AF sample, cleavage cracks might easily propagate along the SD, leading to the failure of the sample (Fig. 8). Figure 11 shows the relationship between the microstructure and the crack propagation paths in the CVN specimens tested at 213 K. Quasi-cleavage cracks might penetrate the elongated PAG and packets without any significant branching along the RD (Fig. 11(c)). It appears that cleavage cracks might be on the {100} family of planes. Therefore, in bcc steel, the spatial distribution of {100} cleavage planes as well as weak interfaces such as grain boundaries (involving carbide particles) is important to control the occurrence of delamination.32)
SEM micrographs showing the relationship between the microstructure and the crack propagation paths in the CVN specimens tested at 213 K (3% Nital etch): non-AF sample tempformed at 46% (a) and 74% (b), and AF samples tempformed at 0% (c) and 46% (d).
The UFEG structure with a strong <110>//RD fiber texture provides many {100} cleavage planes both on the SD planes (which are normal to the SD) and on the 45° planes (which are inclined at an angle of 45° to the RD). The coherence length of the {100} cleavage planes, which corresponds to the Deff, is shorter in the 45° direction than in the RD; the cleavage fracture stress along the 45° direction (σc//45°) is thus greater than the cleavage fracture stress along the SD (σc//SD). Similarly, the intergranular fracture stress in the SD is low, as a result of the elongated grain shape and the good grain boundary continuity along the RD.19,35,36) Therefore, these provide the necessary conditions to cause the brittle delamination along the RD, on the basis of the Yoffee diagram, which indicates the relationship between the σt and σc, for the UFEG structure.8,9,10) Since the refinement in transverse grain size may enhance the σc//45° as well as the σy, the occurrence of the transverse cracking may be suppressed, leading to the occurrence of delamination.
In addition, higher toughness and ductility of the RD plane (which are normal to the RD) was considered to be a sufficient condition in order to enhance the delamination toughening. In the UFEG structure, an effective mechanism for this is to reduce the volume fraction and size of carbide particles, when the transverse grain size and the degree of the <110>//RD fiber texture are almost the identical.10)
Furthermore, microstructural inhomogeneity (which results in transverse crack propagation) should be minimized to enhance the delamination toughening.9,11) Since the crystallographic orientation and geometric arrangement of the UFEGs are relatively similar within a packet band,9,11) the packet band may have relatively good continuity of the {100} cleavage planes and weak UFEG boundaries (including carbide particles).11) Hence, cracks can penetrate the packet band without any significant deflection, and the size of packet bands may be considered to be an effective grain size for the propagation of delamination cracks. The reduction in the width of the packet bands can lead to the suppression of the transverse crack propagation, leading to the enhanced delamination. When compared to the (0.2-0.4)C-2Si-1Cr-1Mo steels with UFEG structures,10,11) the delamination toughening in the present 0.4C-2Cr-1Mo-2Ni steel with an UFEG structure is enhanced at lower temperatures. One of the reasons for this might be that the width of the packet bands was smaller in the present 0.4C-2Cr-1Mo-2Ni steel than in the (0.2-0.4)C-2Si-1Cr-1Mo steels.10,11) Furthermore, the AF with a rolling reduction of 70% is considered to be effective in reducing the width of the packets.
The AF sample tempformed at 46% has an elongated packet structure, in which the ILp_SD and ILp_RD are almost identical with those of the 74%-rolled sample with an UFEG structure, respectively (Fig. 5). However, the long axes of the elongated grains inside the packets are not parallel to the RD in many parts (Figs. 4(a), 4(d)). The degree of the <110>//RD fiber texture in the AF sample is weaker than that in the 74%-rolled sample (Fig. 6). Thus, this might be a reason why the delamination toughening in the AF sample tempformed at 46% was not striking compared to the non-AF sample tempformed at 74%. Figures 11(b) and 11(d) indicate that delamination cracks might preferentially propagate along the longitudinal axes of the elongated grains. When compared to the non-AF sample tempformed at 74% (Fig. 11(b)), the delamination crack in the AF sample (Fig. 11(d)) tends to propagate to the transverse directions, as a consequence of the elongated grains that are not oriented along the RD. On the other hand, at the same rolling reduction of 46%, the AF sample has more flattened and elongated grain structure than the non-AF sample, and coarse block-shaped grains are hardly existed (Fig. 4). This might result in the enhanced delamination toughening in the AF sample. It appears that more wide and less elongated grains that are not aligned to the RD might provide preferable paths for transverse crack propagation in the non-AF sample tempformed at 46% (Fig. 11(a)).
Figure 12 summarizes the vE at room temperature as a function of the σy. Data for the conventional steels37,38) and an ausformed 0.2C-3Mo-3Ni steel14) are also shown. The AF with a rolling reduction of 74% achieves an excellent combination of ultra-high strength and toughness through the WTF at 773 K with a rolling reduction of 46%. The strength and toughness balance of the AF sample tempformed at 46% is almost comparable to that of the non-AF samples tempformed 74%.
The influence of the AF using multi-pass caliber rolling with a rolling reduction of 74% on the microstructure, tensile and Charpy impact properties was investigated for a 0.4%C-2%Cr-1%Mo-2%Ni steel that was processed via WTF using multi-pass caliber rolling at 773 K. The results obtained are as follows;
(1) The AF resulted in the evolution of anisotropic microstructure consisting of elongated PAGs and packets. The AF sample had finer block structure than the non-AF sample.
(2) The AF sample demonstrated the faster kinetics of the microstructural changes during the WTF, i.e. refinement in the transverse grain size and development of highly elongated grain structure, than the non-AF sample. The degree of <110>//RD fiber texture that was developed through the WTF depended on the equivalent strain.
(3) The WTF caused an increase of the σy both in the non-AF and AF samples. The AF sample exhibited higher strength than the non-AF sample in the WTF with rolling reductions up to 46%.
(4) The WTF with the rolling reduction of 46% or over resulted in an enhanced toughness in the non-AF and AF samples. At the 46% rolling reduction, the AF sample exhibited superior Charpy impact properties to the non-AF samples; the delamination toughening is more pronounced in the AF sample than in the non-AF sample.
(5) It was confirmed that the AF with a rolling reduction of 74% was effective to produce an excellent combination of ultra-high strength and toughness through the following WTF with a rolling reduction of 46%.
The authors thank Mr. Kuroda and Mr. Iida for materials processing with caliber-rolling and Ms. Hirota for her assistance with the microstructural observation. The study was partly supported by the AMADA Foundation Grant Number AF-2013005, JSPS KAKENHI Grant Number 15H04150, and the Japan Science and Technology Agency (JST) under Collaborative Research Based on Industrial Demand “Heterogeneous Structure Control: Towards Innovative Development of Metallic Structural Materials”.
