a. Macrostructure and Vickers Microhardness
The OM metallograph presented in Figure 4-1 shows the microstructure of the steel consisted of ferrite (white-etched phase) and pearlite (dark-etched phase) after isothermal transformation at 650oC for 1 h. A great number of intragranular ferritic grains are formed. The microstructure seems to include both allotriomorphic and idiomorphic ferrite.
Figure 4-1 The optical micrograph of the microstructure of the sample austenitized at 1200 oC subsequently isothermally transformed at 650 oC for 1h.
91
The term “allotriomorphic” means that the phase is crystalline in internal structure, but not in outward form [39]. Therefore, the allotriomorphic ferrite, as indicated in Figure 4-1, reveals non-equiaxed morphology, determined by bicrystallography of ferrite and austenite. The term “idiomorphic” implies that the concerned phase has faces belonging to its intrinsic crystalline form [39]. In steels, idiomorphic ferrite is understood to be that which has a roughly equiaxed morphology, as indicated in Figure 4-1. The volume fraction of each phase was determined by point counting method: 6 ± 0.5 % for intergranular ferrite; 22 ± 2.3 % for intragranular ferrite; and 72 ± 1.8 % for pearlite.
The measured values were Hv 254.2 ± 11.0 and Hv 356.0 ± 9.0 for ferrite and pearlite after isothermal transformation, respectively. Actually, there was little difference in the hardness of each ferrite isomorph (Hv 249.3 ± 8.2 for intergranular allotrimorphic ferrite, Hv 243.6 ± 12.1 for intragranular allotrimorphic ferrite, Hv 259.4 ± 11.7 for idiomorphic ferrite). There could be a significant contribution from the strengthening of dispersion of carbides to all ferrite isomorphs and maybe pearlite.
b. Precipitation in ferrite
Figure 4-2(a) shows the transition of vanadium carbide (VC) morphology from IP into fibrous form in a single ferritic grain during the austenite-to-ferrite transformation.
However, as shown in Figure 4-2(b), the selection area diffraction pattern (SADP) indicates that both interphase-precipitated VC and fibrous VC were NaCl-type carbides.
Furthermore, it should be highlighted that the interphase-precipitated VC and fibrous VC had two different variants of the Baker-Nutting (B-N) orientation relationship (OR):
VC ferrite
<1 1 0> || <1 0 0>
and{0 0 1} || {0 0 1}
VC ferrite with the ferritic matrix.There are three crystallography variants of B-N OR between NaCl-type carbides
92
and body-centered cubic (bcc) ferrite in steels. In the present analysis, one variant of B-N OR between interphase-precipitated VC and ferrite was [1 1 0]VC(IP) || [1 0 0]ferrite and (0 0 1)VC(IP) || (0 0 1)ferrite, as identified in Figure 4-2(c). In contrast, fibrous VC possessed another variant of B-N OR with ferrite: [1 1 0]VC(F) || [0 0 1]ferrite and
VC(F) ferrite
(0 0 1) || (1 0 0) , as identified in Figure 4-2(d). The two variants imply that
there is a 90o rotation in the orientations of the carbide broad planes,
(0 0 1)
VC. The center dark-field images shown in Figure 4-2(e) and Figure 4-2(f) were illuminated by individual 002 reflections in Figure 4-2(b) of the two B-N variants, respectively. This new finding implies that the nucleation and growth of fibrous VC might be different from those of interphase-precipitated VC because they hold obviously different variants of B-N ORs with ferrite on the ferrite/austenite interface and grow into distinctive carbide morphologies. Edmonds had initially proposed different growth mechanisms for interphase-precipitated and fibrous carbides [3]. However, there has been no direct and striking evidence for his proposition until the present crystallography characterizations by TEM.93
Figure 4-2 TEM analysis on the area containing two carbide aggregates in ferritic matrix. (a) a bright-field image. (b) the SADP, (c) the indexed SADP of the interphase-precipitated carbide, (d) the indexed SADP of the carbide fiber, (e) and (f) the dark-field image illuminated by carbide 002 reflection for the interphase-precipitated carbide and carbide fiber, respectively.
The growth mechanisms of interphase-precipitated and fibrous carbides have been extensively studied and discussed [3, 43]. However, the transition of IP to carbide fiber growth associated with the variant changes of B-N OR has not been reported. Smith and Dunne claimed that the interphase-precipitated MC (TiC, VC, or NbC) plate-like
94
carbide would keep its broad plane as parallel as possible to the terrace plane of the ledged interface during austenite-to-ferrite phase transformation. This provides the shortest diffusion path and minimizes interfacial energy at the interface. However, this rule of variant selection seems not be applied to fibrous carbide, in which carbide also nucleate on the interface. The variant selection of carbide nuclei for fibrous carbide might be resulted from other mechanism. Actually, it has been reported that the fibrous carbide nuclei continue cooperative growth with ferrite into austenite, resulting in a fibrous morphology in ferritic matrix. Such a mechanism is similar to that of pearlite reaction: cementite (Fe3C) continues cooperative growth with pearlitic ferrite, and Fe3C follows only one variant of Bagayaski [81] or Pitsch-Petch [82]OR with pearlitic ferrite [83]. The B-N OR always causes MC carbide growth with the broad plane,
{
0 0 1}
MC|| 0 0 1{ }
ferrite, such that only one variant of interphase-precipitated carbide is generally observed in ferritic matrix. Figure 4-2(a) and Figure 4-2(d), the broad planes of VC fibres, (0 0 1)VC(F)|| (1 0 0)ferrite, were nearly parallel to its growth direction. However, if the fibrous carbides nucleated on the terrace plane of the interface, two contradictions would arise: (1) they could be interphase-precipitated carbides, and (2) no variant transition would be required. Evidence provided in Figure 4-2(b) implies that the nuclei of fibrous VC have to nucleate in different locations on the interface during the austenite-to-ferrite transformation but not on the terrace plane of the interface, and then grow into fibrous morphology.Incoherent interface had been suggested to be associated with the formation of fibrous carbides during the austenite-to-ferrite phase transformation [3, 43]. However, Berry et al. [42] reported that the VC fibers could hold a Kurdjumov-Sachs (K-S) OR with ferrite. They further concluded that the VC fibers formed on the interface
95
associated with ferrite and austenite by K-S OR. The SADPs given in Figure 4-2(b) show VC fibres follows B-N OR with respect to ferrite. It indicates VC fibers should form on the ferrite/austenite interface which is relatively incoherent. It is recently recognized that IP is related to the ledge mechanism on the terrace planes of the ferrite/austenite interface associated with irrational OR. The transformation ledge nucleates on a high-indexed incoherent interface and is composed of a terrace plane with relatively low interfacial energy and a high-energy step plane. Hence, the massive nucleations of transformation ledges have to be suppressed if the cooperative transformation of ferrite and carbide is expected to occur on the incoherent interface. It implies that the development of fibrous VC should occur when the driving force of austenite-to-ferrite transformation is low, such that fibrous VC should always take places after interphase-precipiated VC, as shown in Figure 4-3. Therefore, the embryos of VC fibers are supposed to nucleate on a slowed-down incoherent ferrite/austenite interface (not on the terrace plane of an interfacial ledge). The present suggestion rules out the possibility for carbides to form the interphase-precipitaed VC. It is presumed that the nucleation of VC fiber consumes local carbon contents and increases the driving force of austenite-to-ferrite transformation. At the same while, it induces ferrite formation then carbide continuously and cooperatively grows with ferrite, holding one new variant of B-N OR with respect to the ferritic matrix. The cooperative growth process of VC fibre and ferritic matrix is an eutectoid reaction which is controlled by the diffusion of carbide-forming elements, i.e. Ti, Nb, Mo, Cr, and V. Previous studies had suggested that high transformation temperatures, and increasing the addition of Mn, which slowed-down the growth rate of ferrite are the conditions for the development of fibrous carbide.
96
Figure 4-3 Transition of carbide aggregates (interphase precipitation, carbide fiber, and pearlite) during the austenite-to-ferrite transformation.
c. Precipitation in pearlitic ferrite
In addition to intragranular ferrite, both interphase-precipitated VC and fibrous VC were observed in pearlitic ferrite as shown in Figure 4-4. Once again, the interphase-precipitated VC, see Figure 4-4(a) to Figure 4-4(c), was related to pearlitic ferrite by the B-N OR. Parsons and Edmonds suggested that the interphase-precipitated VC nucleate on the terrace plane of the pearlite/austenite interface, which is associated with a rational OR between pearlitic ferrite and austenite during pearlitic transformation.
However, Ohmori had pointed out that the pearlite can grow into the austenite without rational OR with pearlitic ferrite [84]. The pearlitic ferrite can, by epitaxy-like growth, inherit the orientation of the original ferritic grain formed before the pearlite
97
transformation. Yen et al. [14] recently reported that IP can occur on the terrace plane of the ledged ferrite/austenite interface, related to an irrational OR between pearlitic ferrite and austenite. Therefore, it is supposed that the IP occurs by ledge mechanism on the interface corresponding to either rational or irrational ORs between pearlitic ferrite and austenite.
Figure 4-4 The observed interphase-precipitated carbide and carbide fiber in pearlitic ferrite. (a-c) the bright-field, dark-field images and associated SADP of the interphase-precipitaed carbide; (d-f) that for carbide fiber.
The present investigation also observed that the VC fibers precipitated in the pearlitic ferrite, as shown from Figure 4-4(d) to Figure 4-4(e). The fibrous VC was also related to pearlitic ferrite by the B-N OR, as shown in Figure 4-4(f). As elucidated for VC fiber growth in ferrite, the fibrous VC will orient to pearlitic ferrite by B-N OR and
98
keep growing cooperatively with pearlitic ferrite during the pearlitic transformation.
The present analysis shows that VC fiber keeps B-N OR with pearlitic ferrite and indicates that it should form on the interface associated with an irrational OR between pearlitic ferrite and austenite as well.
For carbide fiber growth in pearlitic ferrite, it can be expected that the thermodynamic driving force for Fe3C formation is lower than that for VC. As a consequence, the carbide fiber formation would precede the pearlite formation. It also implies if a slower transformation pearlite/austenite front presents, it is possible for VC fiber to form in the pearlitic ferrite between cementite lamella, as shown from Figure 4-4(d) to Figure 4-4(f). Moreover, Fe3C formation consumes great amounts of carbon; it further causes nucleation of ledges on the transformation front. In such a case, as displayed in Figure 4-4(a) and Figure 4-4(c), IP carbides can form in pearlitic ferrite.
The mechanisms to form interphase-precipitated VC and fibrous VC are similar to that one in ferrite. The above illustrations are well-consistent with the model proposed by Parsons and Edmonds [52] to describe the interphase-precipitated VC in pearlitic ferrite. However, not every pearlite colony was found to have alloy carbides (interphase-precipitated or fibrous VCs) in pearlitic ferrite. The criterion is determined by (1) the interfacial structure at pearlite/austenite interface, (2) the consumption of the carbon by the formation of Fe3C, and (3) the diffusion rate of carbide forming elements.
However, the proposition of diffusion and consumption of carbon and alloy elements during the transformation is beyond the scope of the present TEM investigation.
A high-resolution TEM (HRTEM) lattice image was taken from a region in which VC fibers precipitating in ferrite (Figure 4-5). The diffractogram generated by two-dimensional fast Fourier transform of the lattice image is given in Figure 4-5(b). It reveals, once again, that the VC fiber has B-N OR with respect to the ferritic matrix
99
along a zone axis
[
1 0 0]
ferrite|| [1 1 0]carbide[100]ferrite//[11�0]carbide. The overlap of carbide and ferrite lattices contributes the moiré fringes, which can be used to determine the size of carbides. The fibrous VC is about 3.06 nm in thickness and has a great aspect ratio (length/width), demonstrating a slender morphology unlike that of the plate-like interphase-precipitated carbides. For the interface of interphase precipitation or supersaturated precipitation, the broad plane of the MC carbide is the broad face of the carbide platelet. In other words, the carbide growth rates in directions on the broad face are approximately equivalent under B-N OR. However, the face of the fibrous VC is also the broad plane,(
1 0 0)
VC(F)|| 1 0 0( )
ferrite, as shown in Figure 4-5(a), strongly suggesting that the growth rates of the advancing and side directions of VC fiber are quite different, due to the nature of cooperative growth with ferrite. The Nano-Probe EDX spectrum of the VC fiber is displayed in Figure 4-5(c), and its chemical composition is mainly vanadium.Figure 4-5 The HRTEM image of the carbide fiber in ferrite. (a) the lattice image, (b) the FFT diffractogram, and (c) the EDS analysis of carbide fiber.